git/diff.c

4769 lines
126 KiB
C
Raw Normal View History

/*
* Copyright (C) 2005 Junio C Hamano
*/
#include "cache.h"
#include "quote.h"
#include "diff.h"
#include "diffcore.h"
binary patch. This adds "binary patch" to the diff output and teaches apply what to do with them. On the diff generation side, traditionally, we said "Binary files differ\n" without giving anything other than the preimage and postimage object name on the index line. This was good enough for applying a patch generated from your own repository (very useful while rebasing), because the postimage would be available in such a case. However, this was not useful when the recipient of such a patch via e-mail were to apply it, even if the preimage was available. This patch allows the diff to generate "binary" patch when operating under --full-index option. The binary patch follows the usual extended git diff headers, and looks like this: "GIT binary patch\n" <length byte><data>"\n" ... "\n" Each line is prefixed with a "length-byte", whose value is upper or lowercase alphabet that encodes number of bytes that the data on the line decodes to (1..52 -- 'A' means 1, 'B' means 2, ..., 'Z' means 26, 'a' means 27, ...). <data> is 1 or more groups of 5-byte sequence, each of which encodes up to 4 bytes in base85 encoding. Because 52 / 4 * 5 = 65 and we have the length byte, an output line is capped to 66 characters. The payload is the same diff-delta as we use in the packfiles. On the consumption side, git-apply now can decode and apply the binary patch when --allow-binary-replacement is given, the diff was generated with --full-index, and the receiving repository has the preimage blob, which is the same condition as it always required when accepting an "Binary files differ\n" patch. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-04 23:51:44 +00:00
#include "delta.h"
#include "xdiff-interface.h"
#include "color.h"
#include "attr.h"
#include "run-command.h"
#include "utf8.h"
#include "userdiff.h"
#include "sigchain.h"
#include "submodule.h"
#include "ll-merge.h"
Avoid accessing a slow working copy during diffcore operations. The Cygwin folks have done a fine job at creating a POSIX layer on Windows That Just Works(tm). However it comes with a penalty; accessing files in the working tree by way of stat/open/mmap can be slower for diffcore than inflating the data from a blob which is stored in a packfile. This performance problem is especially an issue in merge-recursive when dealing with nearly 7000 added files, as we are loading each file's content from the working directory to perform rename detection. I have literally seen (and sadly watched) paint dry in less time than it takes for merge-recursive to finish such a merge. On the other hand this very same merge runs very fast on Solaris. If Git is compiled with NO_FAST_WORKING_DIRECTORY set then we will avoid looking at the working directory when the blob in question is available within a packfile and the caller doesn't need the data unpacked into a temporary file. We don't use loose objects as they have the same open/mmap/close costs as the working directory file access, but have the additional CPU overhead of needing to inflate the content before use. So it is still faster to use the working tree file over the loose object. If the caller needs the file data unpacked into a temporary file its likely because they are going to call an external diff program, passing the file as a parameter. In this case reusing the working tree file will be faster as we don't need to inflate the data and write it out to a temporary file. The NO_FAST_WORKING_DIRECTORY feature is enabled by default on Cygwin, as that is the platform which currently appears to benefit the most from this option. Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-12-14 11:15:57 +00:00
#ifdef NO_FAST_WORKING_DIRECTORY
#define FAST_WORKING_DIRECTORY 0
#else
#define FAST_WORKING_DIRECTORY 1
#endif
static int diff_detect_rename_default;
static int diff_rename_limit_default = 400;
static int diff_suppress_blank_empty;
int diff_use_color_default = -1;
static const char *diff_word_regex_cfg;
static const char *external_diff_cmd_cfg;
int diff_auto_refresh_index = 1;
static int diff_mnemonic_prefix;
static int diff_no_prefix;
static int diff_dirstat_permille_default = 30;
static struct diff_options default_diff_options;
static char diff_colors[][COLOR_MAXLEN] = {
GIT_COLOR_RESET,
GIT_COLOR_NORMAL, /* PLAIN */
GIT_COLOR_BOLD, /* METAINFO */
GIT_COLOR_CYAN, /* FRAGINFO */
GIT_COLOR_RED, /* OLD */
GIT_COLOR_GREEN, /* NEW */
GIT_COLOR_YELLOW, /* COMMIT */
GIT_COLOR_BG_RED, /* WHITESPACE */
GIT_COLOR_NORMAL, /* FUNCINFO */
};
static int parse_diff_color_slot(const char *var, int ofs)
{
if (!strcasecmp(var+ofs, "plain"))
return DIFF_PLAIN;
if (!strcasecmp(var+ofs, "meta"))
return DIFF_METAINFO;
if (!strcasecmp(var+ofs, "frag"))
return DIFF_FRAGINFO;
if (!strcasecmp(var+ofs, "old"))
return DIFF_FILE_OLD;
if (!strcasecmp(var+ofs, "new"))
return DIFF_FILE_NEW;
if (!strcasecmp(var+ofs, "commit"))
return DIFF_COMMIT;
if (!strcasecmp(var+ofs, "whitespace"))
return DIFF_WHITESPACE;
if (!strcasecmp(var+ofs, "func"))
return DIFF_FUNCINFO;
2009-12-12 12:25:24 +00:00
return -1;
}
Improve error handling when parsing dirstat parameters When encountering errors or unknown tokens while parsing parameters to the --dirstat option, it makes sense to die() with an error message informing the user of which parameter did not make sense. However, when parsing the diff.dirstat config variable, we cannot simply die(), but should instead (after warning the user) ignore the erroneous or unrecognized parameter. After all, future Git versions might add more dirstat parameters, and using two different Git versions on the same repo should not cripple the older Git version just because of a parameter that is only understood by a more recent Git version. This patch fixes the issue by refactoring the dirstat parameter parsing so that parse_dirstat_params() keeps on parsing parameters, even if an earlier parameter was not recognized. When parsing has finished, it returns zero if all parameters were successfully parsed, and non-zero if one or more parameters were not recognized (with appropriate error messages appended to the 'errmsg' argument). The parse_dirstat_params() callers then decide (based on the return value from parse_dirstat_params()) whether to warn and ignore (in case of diff.dirstat), or to warn and die (in case of --dirstat). The patch also adds a couple of tests verifying the correct behavior of --dirstat and diff.dirstat in the face of unknown (possibly future) dirstat parameters. Suggested-by: Junio C Hamano <gitster@pobox.com> Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:22 +00:00
static int parse_dirstat_params(struct diff_options *options, const char *params,
struct strbuf *errmsg)
{
const char *p = params;
Improve error handling when parsing dirstat parameters When encountering errors or unknown tokens while parsing parameters to the --dirstat option, it makes sense to die() with an error message informing the user of which parameter did not make sense. However, when parsing the diff.dirstat config variable, we cannot simply die(), but should instead (after warning the user) ignore the erroneous or unrecognized parameter. After all, future Git versions might add more dirstat parameters, and using two different Git versions on the same repo should not cripple the older Git version just because of a parameter that is only understood by a more recent Git version. This patch fixes the issue by refactoring the dirstat parameter parsing so that parse_dirstat_params() keeps on parsing parameters, even if an earlier parameter was not recognized. When parsing has finished, it returns zero if all parameters were successfully parsed, and non-zero if one or more parameters were not recognized (with appropriate error messages appended to the 'errmsg' argument). The parse_dirstat_params() callers then decide (based on the return value from parse_dirstat_params()) whether to warn and ignore (in case of diff.dirstat), or to warn and die (in case of --dirstat). The patch also adds a couple of tests verifying the correct behavior of --dirstat and diff.dirstat in the face of unknown (possibly future) dirstat parameters. Suggested-by: Junio C Hamano <gitster@pobox.com> Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:22 +00:00
int p_len, ret = 0;
while (*p) {
Improve error handling when parsing dirstat parameters When encountering errors or unknown tokens while parsing parameters to the --dirstat option, it makes sense to die() with an error message informing the user of which parameter did not make sense. However, when parsing the diff.dirstat config variable, we cannot simply die(), but should instead (after warning the user) ignore the erroneous or unrecognized parameter. After all, future Git versions might add more dirstat parameters, and using two different Git versions on the same repo should not cripple the older Git version just because of a parameter that is only understood by a more recent Git version. This patch fixes the issue by refactoring the dirstat parameter parsing so that parse_dirstat_params() keeps on parsing parameters, even if an earlier parameter was not recognized. When parsing has finished, it returns zero if all parameters were successfully parsed, and non-zero if one or more parameters were not recognized (with appropriate error messages appended to the 'errmsg' argument). The parse_dirstat_params() callers then decide (based on the return value from parse_dirstat_params()) whether to warn and ignore (in case of diff.dirstat), or to warn and die (in case of --dirstat). The patch also adds a couple of tests verifying the correct behavior of --dirstat and diff.dirstat in the face of unknown (possibly future) dirstat parameters. Suggested-by: Junio C Hamano <gitster@pobox.com> Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:22 +00:00
p_len = strchrnul(p, ',') - p;
if (!memcmp(p, "changes", p_len)) {
New --dirstat=lines mode, doing dirstat analysis based on diffstat This patch adds an alternative implementation of show_dirstat(), called show_dirstat_by_line(), which uses the more expensive diffstat analysis (as opposed to show_dirstat()'s own (relatively inexpensive) analysis) to derive the numbers from which the --dirstat output is computed. The alternative implementation is controlled by the new "lines" parameter to the --dirstat option (or the diff.dirstat config variable). For binary files, the diffstat analysis counts bytes instead of lines, so to prevent binary files from dominating the dirstat results, the byte counts for binary files are divided by 64 before being compared to their textual/line-based counterparts. This is a stupid and ugly - but very cheap - heuristic. In linux-2.6.git, running the three different --dirstat modes: time git diff v2.6.20..v2.6.30 --dirstat=changes > /dev/null vs. time git diff v2.6.20..v2.6.30 --dirstat=lines > /dev/null vs. time git diff v2.6.20..v2.6.30 --dirstat=files > /dev/null yields the following average runtimes on my machine: - "changes" (default): ~6.0 s - "lines": ~9.6 s - "files": ~0.1 s So, as expected, there's a considerable performance hit (~60%) by going through the full diffstat analysis as compared to the default "changes" analysis (obviously, "files" is much faster than both). As such, the "lines" mode is probably only useful if you really need the --dirstat numbers to be consistent with the numbers returned from the other --*stat options. The patch also includes documentation and tests for the new dirstat mode. Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:21 +00:00
DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
Improve error handling when parsing dirstat parameters When encountering errors or unknown tokens while parsing parameters to the --dirstat option, it makes sense to die() with an error message informing the user of which parameter did not make sense. However, when parsing the diff.dirstat config variable, we cannot simply die(), but should instead (after warning the user) ignore the erroneous or unrecognized parameter. After all, future Git versions might add more dirstat parameters, and using two different Git versions on the same repo should not cripple the older Git version just because of a parameter that is only understood by a more recent Git version. This patch fixes the issue by refactoring the dirstat parameter parsing so that parse_dirstat_params() keeps on parsing parameters, even if an earlier parameter was not recognized. When parsing has finished, it returns zero if all parameters were successfully parsed, and non-zero if one or more parameters were not recognized (with appropriate error messages appended to the 'errmsg' argument). The parse_dirstat_params() callers then decide (based on the return value from parse_dirstat_params()) whether to warn and ignore (in case of diff.dirstat), or to warn and die (in case of --dirstat). The patch also adds a couple of tests verifying the correct behavior of --dirstat and diff.dirstat in the face of unknown (possibly future) dirstat parameters. Suggested-by: Junio C Hamano <gitster@pobox.com> Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:22 +00:00
} else if (!memcmp(p, "lines", p_len)) {
New --dirstat=lines mode, doing dirstat analysis based on diffstat This patch adds an alternative implementation of show_dirstat(), called show_dirstat_by_line(), which uses the more expensive diffstat analysis (as opposed to show_dirstat()'s own (relatively inexpensive) analysis) to derive the numbers from which the --dirstat output is computed. The alternative implementation is controlled by the new "lines" parameter to the --dirstat option (or the diff.dirstat config variable). For binary files, the diffstat analysis counts bytes instead of lines, so to prevent binary files from dominating the dirstat results, the byte counts for binary files are divided by 64 before being compared to their textual/line-based counterparts. This is a stupid and ugly - but very cheap - heuristic. In linux-2.6.git, running the three different --dirstat modes: time git diff v2.6.20..v2.6.30 --dirstat=changes > /dev/null vs. time git diff v2.6.20..v2.6.30 --dirstat=lines > /dev/null vs. time git diff v2.6.20..v2.6.30 --dirstat=files > /dev/null yields the following average runtimes on my machine: - "changes" (default): ~6.0 s - "lines": ~9.6 s - "files": ~0.1 s So, as expected, there's a considerable performance hit (~60%) by going through the full diffstat analysis as compared to the default "changes" analysis (obviously, "files" is much faster than both). As such, the "lines" mode is probably only useful if you really need the --dirstat numbers to be consistent with the numbers returned from the other --*stat options. The patch also includes documentation and tests for the new dirstat mode. Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:21 +00:00
DIFF_OPT_SET(options, DIRSTAT_BY_LINE);
DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
Improve error handling when parsing dirstat parameters When encountering errors or unknown tokens while parsing parameters to the --dirstat option, it makes sense to die() with an error message informing the user of which parameter did not make sense. However, when parsing the diff.dirstat config variable, we cannot simply die(), but should instead (after warning the user) ignore the erroneous or unrecognized parameter. After all, future Git versions might add more dirstat parameters, and using two different Git versions on the same repo should not cripple the older Git version just because of a parameter that is only understood by a more recent Git version. This patch fixes the issue by refactoring the dirstat parameter parsing so that parse_dirstat_params() keeps on parsing parameters, even if an earlier parameter was not recognized. When parsing has finished, it returns zero if all parameters were successfully parsed, and non-zero if one or more parameters were not recognized (with appropriate error messages appended to the 'errmsg' argument). The parse_dirstat_params() callers then decide (based on the return value from parse_dirstat_params()) whether to warn and ignore (in case of diff.dirstat), or to warn and die (in case of --dirstat). The patch also adds a couple of tests verifying the correct behavior of --dirstat and diff.dirstat in the face of unknown (possibly future) dirstat parameters. Suggested-by: Junio C Hamano <gitster@pobox.com> Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:22 +00:00
} else if (!memcmp(p, "files", p_len)) {
New --dirstat=lines mode, doing dirstat analysis based on diffstat This patch adds an alternative implementation of show_dirstat(), called show_dirstat_by_line(), which uses the more expensive diffstat analysis (as opposed to show_dirstat()'s own (relatively inexpensive) analysis) to derive the numbers from which the --dirstat output is computed. The alternative implementation is controlled by the new "lines" parameter to the --dirstat option (or the diff.dirstat config variable). For binary files, the diffstat analysis counts bytes instead of lines, so to prevent binary files from dominating the dirstat results, the byte counts for binary files are divided by 64 before being compared to their textual/line-based counterparts. This is a stupid and ugly - but very cheap - heuristic. In linux-2.6.git, running the three different --dirstat modes: time git diff v2.6.20..v2.6.30 --dirstat=changes > /dev/null vs. time git diff v2.6.20..v2.6.30 --dirstat=lines > /dev/null vs. time git diff v2.6.20..v2.6.30 --dirstat=files > /dev/null yields the following average runtimes on my machine: - "changes" (default): ~6.0 s - "lines": ~9.6 s - "files": ~0.1 s So, as expected, there's a considerable performance hit (~60%) by going through the full diffstat analysis as compared to the default "changes" analysis (obviously, "files" is much faster than both). As such, the "lines" mode is probably only useful if you really need the --dirstat numbers to be consistent with the numbers returned from the other --*stat options. The patch also includes documentation and tests for the new dirstat mode. Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:21 +00:00
DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
DIFF_OPT_SET(options, DIRSTAT_BY_FILE);
Improve error handling when parsing dirstat parameters When encountering errors or unknown tokens while parsing parameters to the --dirstat option, it makes sense to die() with an error message informing the user of which parameter did not make sense. However, when parsing the diff.dirstat config variable, we cannot simply die(), but should instead (after warning the user) ignore the erroneous or unrecognized parameter. After all, future Git versions might add more dirstat parameters, and using two different Git versions on the same repo should not cripple the older Git version just because of a parameter that is only understood by a more recent Git version. This patch fixes the issue by refactoring the dirstat parameter parsing so that parse_dirstat_params() keeps on parsing parameters, even if an earlier parameter was not recognized. When parsing has finished, it returns zero if all parameters were successfully parsed, and non-zero if one or more parameters were not recognized (with appropriate error messages appended to the 'errmsg' argument). The parse_dirstat_params() callers then decide (based on the return value from parse_dirstat_params()) whether to warn and ignore (in case of diff.dirstat), or to warn and die (in case of --dirstat). The patch also adds a couple of tests verifying the correct behavior of --dirstat and diff.dirstat in the face of unknown (possibly future) dirstat parameters. Suggested-by: Junio C Hamano <gitster@pobox.com> Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:22 +00:00
} else if (!memcmp(p, "noncumulative", p_len)) {
DIFF_OPT_CLR(options, DIRSTAT_CUMULATIVE);
Improve error handling when parsing dirstat parameters When encountering errors or unknown tokens while parsing parameters to the --dirstat option, it makes sense to die() with an error message informing the user of which parameter did not make sense. However, when parsing the diff.dirstat config variable, we cannot simply die(), but should instead (after warning the user) ignore the erroneous or unrecognized parameter. After all, future Git versions might add more dirstat parameters, and using two different Git versions on the same repo should not cripple the older Git version just because of a parameter that is only understood by a more recent Git version. This patch fixes the issue by refactoring the dirstat parameter parsing so that parse_dirstat_params() keeps on parsing parameters, even if an earlier parameter was not recognized. When parsing has finished, it returns zero if all parameters were successfully parsed, and non-zero if one or more parameters were not recognized (with appropriate error messages appended to the 'errmsg' argument). The parse_dirstat_params() callers then decide (based on the return value from parse_dirstat_params()) whether to warn and ignore (in case of diff.dirstat), or to warn and die (in case of --dirstat). The patch also adds a couple of tests verifying the correct behavior of --dirstat and diff.dirstat in the face of unknown (possibly future) dirstat parameters. Suggested-by: Junio C Hamano <gitster@pobox.com> Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:22 +00:00
} else if (!memcmp(p, "cumulative", p_len)) {
DIFF_OPT_SET(options, DIRSTAT_CUMULATIVE);
} else if (isdigit(*p)) {
char *end;
Improve error handling when parsing dirstat parameters When encountering errors or unknown tokens while parsing parameters to the --dirstat option, it makes sense to die() with an error message informing the user of which parameter did not make sense. However, when parsing the diff.dirstat config variable, we cannot simply die(), but should instead (after warning the user) ignore the erroneous or unrecognized parameter. After all, future Git versions might add more dirstat parameters, and using two different Git versions on the same repo should not cripple the older Git version just because of a parameter that is only understood by a more recent Git version. This patch fixes the issue by refactoring the dirstat parameter parsing so that parse_dirstat_params() keeps on parsing parameters, even if an earlier parameter was not recognized. When parsing has finished, it returns zero if all parameters were successfully parsed, and non-zero if one or more parameters were not recognized (with appropriate error messages appended to the 'errmsg' argument). The parse_dirstat_params() callers then decide (based on the return value from parse_dirstat_params()) whether to warn and ignore (in case of diff.dirstat), or to warn and die (in case of --dirstat). The patch also adds a couple of tests verifying the correct behavior of --dirstat and diff.dirstat in the face of unknown (possibly future) dirstat parameters. Suggested-by: Junio C Hamano <gitster@pobox.com> Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:22 +00:00
int permille = strtoul(p, &end, 10) * 10;
if (*end == '.' && isdigit(*++end)) {
/* only use first digit */
Improve error handling when parsing dirstat parameters When encountering errors or unknown tokens while parsing parameters to the --dirstat option, it makes sense to die() with an error message informing the user of which parameter did not make sense. However, when parsing the diff.dirstat config variable, we cannot simply die(), but should instead (after warning the user) ignore the erroneous or unrecognized parameter. After all, future Git versions might add more dirstat parameters, and using two different Git versions on the same repo should not cripple the older Git version just because of a parameter that is only understood by a more recent Git version. This patch fixes the issue by refactoring the dirstat parameter parsing so that parse_dirstat_params() keeps on parsing parameters, even if an earlier parameter was not recognized. When parsing has finished, it returns zero if all parameters were successfully parsed, and non-zero if one or more parameters were not recognized (with appropriate error messages appended to the 'errmsg' argument). The parse_dirstat_params() callers then decide (based on the return value from parse_dirstat_params()) whether to warn and ignore (in case of diff.dirstat), or to warn and die (in case of --dirstat). The patch also adds a couple of tests verifying the correct behavior of --dirstat and diff.dirstat in the face of unknown (possibly future) dirstat parameters. Suggested-by: Junio C Hamano <gitster@pobox.com> Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:22 +00:00
permille += *end - '0';
/* .. and ignore any further digits */
Improve error handling when parsing dirstat parameters When encountering errors or unknown tokens while parsing parameters to the --dirstat option, it makes sense to die() with an error message informing the user of which parameter did not make sense. However, when parsing the diff.dirstat config variable, we cannot simply die(), but should instead (after warning the user) ignore the erroneous or unrecognized parameter. After all, future Git versions might add more dirstat parameters, and using two different Git versions on the same repo should not cripple the older Git version just because of a parameter that is only understood by a more recent Git version. This patch fixes the issue by refactoring the dirstat parameter parsing so that parse_dirstat_params() keeps on parsing parameters, even if an earlier parameter was not recognized. When parsing has finished, it returns zero if all parameters were successfully parsed, and non-zero if one or more parameters were not recognized (with appropriate error messages appended to the 'errmsg' argument). The parse_dirstat_params() callers then decide (based on the return value from parse_dirstat_params()) whether to warn and ignore (in case of diff.dirstat), or to warn and die (in case of --dirstat). The patch also adds a couple of tests verifying the correct behavior of --dirstat and diff.dirstat in the face of unknown (possibly future) dirstat parameters. Suggested-by: Junio C Hamano <gitster@pobox.com> Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:22 +00:00
while (isdigit(*++end))
; /* nothing */
}
Improve error handling when parsing dirstat parameters When encountering errors or unknown tokens while parsing parameters to the --dirstat option, it makes sense to die() with an error message informing the user of which parameter did not make sense. However, when parsing the diff.dirstat config variable, we cannot simply die(), but should instead (after warning the user) ignore the erroneous or unrecognized parameter. After all, future Git versions might add more dirstat parameters, and using two different Git versions on the same repo should not cripple the older Git version just because of a parameter that is only understood by a more recent Git version. This patch fixes the issue by refactoring the dirstat parameter parsing so that parse_dirstat_params() keeps on parsing parameters, even if an earlier parameter was not recognized. When parsing has finished, it returns zero if all parameters were successfully parsed, and non-zero if one or more parameters were not recognized (with appropriate error messages appended to the 'errmsg' argument). The parse_dirstat_params() callers then decide (based on the return value from parse_dirstat_params()) whether to warn and ignore (in case of diff.dirstat), or to warn and die (in case of --dirstat). The patch also adds a couple of tests verifying the correct behavior of --dirstat and diff.dirstat in the face of unknown (possibly future) dirstat parameters. Suggested-by: Junio C Hamano <gitster@pobox.com> Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:22 +00:00
if (end - p == p_len)
options->dirstat_permille = permille;
else {
strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%.*s'\n"),
Improve error handling when parsing dirstat parameters When encountering errors or unknown tokens while parsing parameters to the --dirstat option, it makes sense to die() with an error message informing the user of which parameter did not make sense. However, when parsing the diff.dirstat config variable, we cannot simply die(), but should instead (after warning the user) ignore the erroneous or unrecognized parameter. After all, future Git versions might add more dirstat parameters, and using two different Git versions on the same repo should not cripple the older Git version just because of a parameter that is only understood by a more recent Git version. This patch fixes the issue by refactoring the dirstat parameter parsing so that parse_dirstat_params() keeps on parsing parameters, even if an earlier parameter was not recognized. When parsing has finished, it returns zero if all parameters were successfully parsed, and non-zero if one or more parameters were not recognized (with appropriate error messages appended to the 'errmsg' argument). The parse_dirstat_params() callers then decide (based on the return value from parse_dirstat_params()) whether to warn and ignore (in case of diff.dirstat), or to warn and die (in case of --dirstat). The patch also adds a couple of tests verifying the correct behavior of --dirstat and diff.dirstat in the face of unknown (possibly future) dirstat parameters. Suggested-by: Junio C Hamano <gitster@pobox.com> Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:22 +00:00
p_len, p);
ret++;
}
} else {
strbuf_addf(errmsg, _(" Unknown dirstat parameter '%.*s'\n"),
Improve error handling when parsing dirstat parameters When encountering errors or unknown tokens while parsing parameters to the --dirstat option, it makes sense to die() with an error message informing the user of which parameter did not make sense. However, when parsing the diff.dirstat config variable, we cannot simply die(), but should instead (after warning the user) ignore the erroneous or unrecognized parameter. After all, future Git versions might add more dirstat parameters, and using two different Git versions on the same repo should not cripple the older Git version just because of a parameter that is only understood by a more recent Git version. This patch fixes the issue by refactoring the dirstat parameter parsing so that parse_dirstat_params() keeps on parsing parameters, even if an earlier parameter was not recognized. When parsing has finished, it returns zero if all parameters were successfully parsed, and non-zero if one or more parameters were not recognized (with appropriate error messages appended to the 'errmsg' argument). The parse_dirstat_params() callers then decide (based on the return value from parse_dirstat_params()) whether to warn and ignore (in case of diff.dirstat), or to warn and die (in case of --dirstat). The patch also adds a couple of tests verifying the correct behavior of --dirstat and diff.dirstat in the face of unknown (possibly future) dirstat parameters. Suggested-by: Junio C Hamano <gitster@pobox.com> Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:22 +00:00
p_len, p);
ret++;
}
Improve error handling when parsing dirstat parameters When encountering errors or unknown tokens while parsing parameters to the --dirstat option, it makes sense to die() with an error message informing the user of which parameter did not make sense. However, when parsing the diff.dirstat config variable, we cannot simply die(), but should instead (after warning the user) ignore the erroneous or unrecognized parameter. After all, future Git versions might add more dirstat parameters, and using two different Git versions on the same repo should not cripple the older Git version just because of a parameter that is only understood by a more recent Git version. This patch fixes the issue by refactoring the dirstat parameter parsing so that parse_dirstat_params() keeps on parsing parameters, even if an earlier parameter was not recognized. When parsing has finished, it returns zero if all parameters were successfully parsed, and non-zero if one or more parameters were not recognized (with appropriate error messages appended to the 'errmsg' argument). The parse_dirstat_params() callers then decide (based on the return value from parse_dirstat_params()) whether to warn and ignore (in case of diff.dirstat), or to warn and die (in case of --dirstat). The patch also adds a couple of tests verifying the correct behavior of --dirstat and diff.dirstat in the face of unknown (possibly future) dirstat parameters. Suggested-by: Junio C Hamano <gitster@pobox.com> Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:22 +00:00
p += p_len;
if (*p)
p++; /* more parameters, swallow separator */
}
Improve error handling when parsing dirstat parameters When encountering errors or unknown tokens while parsing parameters to the --dirstat option, it makes sense to die() with an error message informing the user of which parameter did not make sense. However, when parsing the diff.dirstat config variable, we cannot simply die(), but should instead (after warning the user) ignore the erroneous or unrecognized parameter. After all, future Git versions might add more dirstat parameters, and using two different Git versions on the same repo should not cripple the older Git version just because of a parameter that is only understood by a more recent Git version. This patch fixes the issue by refactoring the dirstat parameter parsing so that parse_dirstat_params() keeps on parsing parameters, even if an earlier parameter was not recognized. When parsing has finished, it returns zero if all parameters were successfully parsed, and non-zero if one or more parameters were not recognized (with appropriate error messages appended to the 'errmsg' argument). The parse_dirstat_params() callers then decide (based on the return value from parse_dirstat_params()) whether to warn and ignore (in case of diff.dirstat), or to warn and die (in case of --dirstat). The patch also adds a couple of tests verifying the correct behavior of --dirstat and diff.dirstat in the face of unknown (possibly future) dirstat parameters. Suggested-by: Junio C Hamano <gitster@pobox.com> Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:22 +00:00
return ret;
}
static int git_config_rename(const char *var, const char *value)
{
if (!value)
return DIFF_DETECT_RENAME;
if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
return DIFF_DETECT_COPY;
return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
}
/*
* These are to give UI layer defaults.
* The core-level commands such as git-diff-files should
* never be affected by the setting of diff.renames
* the user happens to have in the configuration file.
*/
int git_diff_ui_config(const char *var, const char *value, void *cb)
{
if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
diff_use_color_default = git_config_colorbool(var, value);
return 0;
}
if (!strcmp(var, "diff.renames")) {
diff_detect_rename_default = git_config_rename(var, value);
return 0;
}
if (!strcmp(var, "diff.autorefreshindex")) {
diff_auto_refresh_index = git_config_bool(var, value);
return 0;
}
if (!strcmp(var, "diff.mnemonicprefix")) {
diff_mnemonic_prefix = git_config_bool(var, value);
return 0;
}
if (!strcmp(var, "diff.noprefix")) {
diff_no_prefix = git_config_bool(var, value);
return 0;
}
if (!strcmp(var, "diff.external"))
return git_config_string(&external_diff_cmd_cfg, var, value);
if (!strcmp(var, "diff.wordregex"))
return git_config_string(&diff_word_regex_cfg, var, value);
if (!strcmp(var, "diff.ignoresubmodules"))
handle_ignore_submodules_arg(&default_diff_options, value);
if (git_color_config(var, value, cb) < 0)
return -1;
return git_diff_basic_config(var, value, cb);
}
int git_diff_basic_config(const char *var, const char *value, void *cb)
{
if (!strcmp(var, "diff.renamelimit")) {
diff_rename_limit_default = git_config_int(var, value);
return 0;
}
switch (userdiff_config(var, value)) {
case 0: break;
case -1: return -1;
default: return 0;
}
if (!prefixcmp(var, "diff.color.") || !prefixcmp(var, "color.diff.")) {
int slot = parse_diff_color_slot(var, 11);
2009-12-12 12:25:24 +00:00
if (slot < 0)
return 0;
if (!value)
return config_error_nonbool(var);
color_parse(value, var, diff_colors[slot]);
return 0;
}
/* like GNU diff's --suppress-blank-empty option */
if (!strcmp(var, "diff.suppressblankempty") ||
/* for backwards compatibility */
!strcmp(var, "diff.suppress-blank-empty")) {
diff_suppress_blank_empty = git_config_bool(var, value);
return 0;
}
if (!strcmp(var, "diff.dirstat")) {
Improve error handling when parsing dirstat parameters When encountering errors or unknown tokens while parsing parameters to the --dirstat option, it makes sense to die() with an error message informing the user of which parameter did not make sense. However, when parsing the diff.dirstat config variable, we cannot simply die(), but should instead (after warning the user) ignore the erroneous or unrecognized parameter. After all, future Git versions might add more dirstat parameters, and using two different Git versions on the same repo should not cripple the older Git version just because of a parameter that is only understood by a more recent Git version. This patch fixes the issue by refactoring the dirstat parameter parsing so that parse_dirstat_params() keeps on parsing parameters, even if an earlier parameter was not recognized. When parsing has finished, it returns zero if all parameters were successfully parsed, and non-zero if one or more parameters were not recognized (with appropriate error messages appended to the 'errmsg' argument). The parse_dirstat_params() callers then decide (based on the return value from parse_dirstat_params()) whether to warn and ignore (in case of diff.dirstat), or to warn and die (in case of --dirstat). The patch also adds a couple of tests verifying the correct behavior of --dirstat and diff.dirstat in the face of unknown (possibly future) dirstat parameters. Suggested-by: Junio C Hamano <gitster@pobox.com> Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:22 +00:00
struct strbuf errmsg = STRBUF_INIT;
default_diff_options.dirstat_permille = diff_dirstat_permille_default;
Improve error handling when parsing dirstat parameters When encountering errors or unknown tokens while parsing parameters to the --dirstat option, it makes sense to die() with an error message informing the user of which parameter did not make sense. However, when parsing the diff.dirstat config variable, we cannot simply die(), but should instead (after warning the user) ignore the erroneous or unrecognized parameter. After all, future Git versions might add more dirstat parameters, and using two different Git versions on the same repo should not cripple the older Git version just because of a parameter that is only understood by a more recent Git version. This patch fixes the issue by refactoring the dirstat parameter parsing so that parse_dirstat_params() keeps on parsing parameters, even if an earlier parameter was not recognized. When parsing has finished, it returns zero if all parameters were successfully parsed, and non-zero if one or more parameters were not recognized (with appropriate error messages appended to the 'errmsg' argument). The parse_dirstat_params() callers then decide (based on the return value from parse_dirstat_params()) whether to warn and ignore (in case of diff.dirstat), or to warn and die (in case of --dirstat). The patch also adds a couple of tests verifying the correct behavior of --dirstat and diff.dirstat in the face of unknown (possibly future) dirstat parameters. Suggested-by: Junio C Hamano <gitster@pobox.com> Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:22 +00:00
if (parse_dirstat_params(&default_diff_options, value, &errmsg))
warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
Improve error handling when parsing dirstat parameters When encountering errors or unknown tokens while parsing parameters to the --dirstat option, it makes sense to die() with an error message informing the user of which parameter did not make sense. However, when parsing the diff.dirstat config variable, we cannot simply die(), but should instead (after warning the user) ignore the erroneous or unrecognized parameter. After all, future Git versions might add more dirstat parameters, and using two different Git versions on the same repo should not cripple the older Git version just because of a parameter that is only understood by a more recent Git version. This patch fixes the issue by refactoring the dirstat parameter parsing so that parse_dirstat_params() keeps on parsing parameters, even if an earlier parameter was not recognized. When parsing has finished, it returns zero if all parameters were successfully parsed, and non-zero if one or more parameters were not recognized (with appropriate error messages appended to the 'errmsg' argument). The parse_dirstat_params() callers then decide (based on the return value from parse_dirstat_params()) whether to warn and ignore (in case of diff.dirstat), or to warn and die (in case of --dirstat). The patch also adds a couple of tests verifying the correct behavior of --dirstat and diff.dirstat in the face of unknown (possibly future) dirstat parameters. Suggested-by: Junio C Hamano <gitster@pobox.com> Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:22 +00:00
errmsg.buf);
strbuf_release(&errmsg);
diff_dirstat_permille_default = default_diff_options.dirstat_permille;
return 0;
}
Submodules: Add the new "ignore" config option for diff and status The new "ignore" config option controls the default behavior for "git status" and the diff family. It specifies under what circumstances they consider submodules as modified and can be set separately for each submodule. The command line option "--ignore-submodules=" has been extended to accept the new parameter "none" for both status and diff. Users that chose submodules to get rid of long work tree scanning times might want to set the "dirty" option for those submodules. This brings back the pre 1.7.0 behavior, where submodule work trees were never scanned for modifications. By using "--ignore-submodules=none" on the command line the status and diff commands can be told to do a full scan. This option can be set to the following values (which have the same name and meaning as for the "--ignore-submodules" option of status and diff): "all": All changes to the submodule will be ignored. "dirty": Only differences of the commit recorded in the superproject and the submodules HEAD will be considered modifications, all changes to the work tree of the submodule will be ignored. When using this value, the submodule will not be scanned for work tree changes at all, leading to a performance benefit on large submodules. "untracked": Only untracked files in the submodules work tree are ignored, a changed HEAD and/or modified files in the submodule will mark it as modified. "none" (which is the default): Either untracked or modified files in a submodules work tree or a difference between the subdmodules HEAD and the commit recorded in the superproject will make it show up as changed. This value is added as a new parameter for the "--ignore-submodules" option of the diff family and "git status" so the user can override the settings in the configuration. Signed-off-by: Jens Lehmann <Jens.Lehmann@web.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-08-05 22:39:25 +00:00
if (!prefixcmp(var, "submodule."))
return parse_submodule_config_option(var, value);
return git_default_config(var, value, cb);
}
static char *quote_two(const char *one, const char *two)
{
int need_one = quote_c_style(one, NULL, NULL, 1);
int need_two = quote_c_style(two, NULL, NULL, 1);
struct strbuf res = STRBUF_INIT;
if (need_one + need_two) {
strbuf_addch(&res, '"');
quote_c_style(one, &res, NULL, 1);
quote_c_style(two, &res, NULL, 1);
strbuf_addch(&res, '"');
} else {
strbuf_addstr(&res, one);
strbuf_addstr(&res, two);
}
return strbuf_detach(&res, NULL);
}
static const char *external_diff(void)
{
static const char *external_diff_cmd = NULL;
static int done_preparing = 0;
if (done_preparing)
return external_diff_cmd;
external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
if (!external_diff_cmd)
external_diff_cmd = external_diff_cmd_cfg;
done_preparing = 1;
return external_diff_cmd;
}
static struct diff_tempfile {
const char *name; /* filename external diff should read from */
char hex[41];
char mode[10];
char tmp_path[PATH_MAX];
} diff_temp[2];
typedef unsigned long (*sane_truncate_fn)(char *line, unsigned long len);
struct emit_callback {
int color_diff;
unsigned ws_rule;
int blank_at_eof_in_preimage;
int blank_at_eof_in_postimage;
int lno_in_preimage;
int lno_in_postimage;
sane_truncate_fn truncate;
const char **label_path;
struct diff_words_data *diff_words;
struct diff_options *opt;
int *found_changesp;
struct strbuf *header;
};
static int count_lines(const char *data, int size)
{
int count, ch, completely_empty = 1, nl_just_seen = 0;
count = 0;
while (0 < size--) {
ch = *data++;
if (ch == '\n') {
count++;
nl_just_seen = 1;
completely_empty = 0;
}
else {
nl_just_seen = 0;
completely_empty = 0;
}
}
if (completely_empty)
return 0;
if (!nl_just_seen)
count++; /* no trailing newline */
return count;
}
static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
{
if (!DIFF_FILE_VALID(one)) {
mf->ptr = (char *)""; /* does not matter */
mf->size = 0;
return 0;
}
else if (diff_populate_filespec(one, 0))
return -1;
mf->ptr = one->data;
mf->size = one->size;
return 0;
}
/* like fill_mmfile, but only for size, so we can avoid retrieving blob */
static unsigned long diff_filespec_size(struct diff_filespec *one)
{
if (!DIFF_FILE_VALID(one))
return 0;
diff_populate_filespec(one, 1);
return one->size;
}
static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
{
char *ptr = mf->ptr;
long size = mf->size;
int cnt = 0;
if (!size)
return cnt;
ptr += size - 1; /* pointing at the very end */
if (*ptr != '\n')
; /* incomplete line */
else
ptr--; /* skip the last LF */
while (mf->ptr < ptr) {
char *prev_eol;
for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
if (*prev_eol == '\n')
break;
if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
break;
cnt++;
ptr = prev_eol - 1;
}
return cnt;
}
static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
struct emit_callback *ecbdata)
{
int l1, l2, at;
unsigned ws_rule = ecbdata->ws_rule;
l1 = count_trailing_blank(mf1, ws_rule);
l2 = count_trailing_blank(mf2, ws_rule);
if (l2 <= l1) {
ecbdata->blank_at_eof_in_preimage = 0;
ecbdata->blank_at_eof_in_postimage = 0;
return;
}
at = count_lines(mf1->ptr, mf1->size);
ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
at = count_lines(mf2->ptr, mf2->size);
ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
}
static void emit_line_0(struct diff_options *o, const char *set, const char *reset,
int first, const char *line, int len)
{
int has_trailing_newline, has_trailing_carriage_return;
int nofirst;
FILE *file = o->file;
if (o->output_prefix) {
struct strbuf *msg = NULL;
msg = o->output_prefix(o, o->output_prefix_data);
assert(msg);
fwrite(msg->buf, msg->len, 1, file);
}
if (len == 0) {
has_trailing_newline = (first == '\n');
has_trailing_carriage_return = (!has_trailing_newline &&
(first == '\r'));
nofirst = has_trailing_newline || has_trailing_carriage_return;
} else {
has_trailing_newline = (len > 0 && line[len-1] == '\n');
if (has_trailing_newline)
len--;
has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
if (has_trailing_carriage_return)
len--;
nofirst = 0;
}
if (len || !nofirst) {
fputs(set, file);
if (!nofirst)
fputc(first, file);
fwrite(line, len, 1, file);
fputs(reset, file);
}
if (has_trailing_carriage_return)
fputc('\r', file);
if (has_trailing_newline)
fputc('\n', file);
}
static void emit_line(struct diff_options *o, const char *set, const char *reset,
const char *line, int len)
{
emit_line_0(o, set, reset, line[0], line+1, len-1);
}
static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
{
if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
ecbdata->blank_at_eof_in_preimage &&
ecbdata->blank_at_eof_in_postimage &&
ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
return 0;
return ws_blank_line(line, len, ecbdata->ws_rule);
}
static void emit_add_line(const char *reset,
struct emit_callback *ecbdata,
const char *line, int len)
{
const char *ws = diff_get_color(ecbdata->color_diff, DIFF_WHITESPACE);
const char *set = diff_get_color(ecbdata->color_diff, DIFF_FILE_NEW);
if (!*ws)
emit_line_0(ecbdata->opt, set, reset, '+', line, len);
else if (new_blank_line_at_eof(ecbdata, line, len))
/* Blank line at EOF - paint '+' as well */
emit_line_0(ecbdata->opt, ws, reset, '+', line, len);
else {
/* Emit just the prefix, then the rest. */
emit_line_0(ecbdata->opt, set, reset, '+', "", 0);
ws_check_emit(line, len, ecbdata->ws_rule,
ecbdata->opt->file, set, reset, ws);
}
}
static void emit_hunk_header(struct emit_callback *ecbdata,
const char *line, int len)
{
const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
static const char atat[2] = { '@', '@' };
const char *cp, *ep;
struct strbuf msgbuf = STRBUF_INIT;
int org_len = len;
int i = 1;
/*
* As a hunk header must begin with "@@ -<old>, +<new> @@",
* it always is at least 10 bytes long.
*/
if (len < 10 ||
memcmp(line, atat, 2) ||
!(ep = memmem(line + 2, len - 2, atat, 2))) {
emit_line(ecbdata->opt, plain, reset, line, len);
return;
}
ep += 2; /* skip over @@ */
/* The hunk header in fraginfo color */
strbuf_add(&msgbuf, frag, strlen(frag));
strbuf_add(&msgbuf, line, ep - line);
strbuf_add(&msgbuf, reset, strlen(reset));
/*
* trailing "\r\n"
*/
for ( ; i < 3; i++)
if (line[len - i] == '\r' || line[len - i] == '\n')
len--;
/* blank before the func header */
for (cp = ep; ep - line < len; ep++)
if (*ep != ' ' && *ep != '\t')
break;
if (ep != cp) {
strbuf_add(&msgbuf, plain, strlen(plain));
strbuf_add(&msgbuf, cp, ep - cp);
strbuf_add(&msgbuf, reset, strlen(reset));
}
if (ep < line + len) {
strbuf_add(&msgbuf, func, strlen(func));
strbuf_add(&msgbuf, ep, line + len - ep);
strbuf_add(&msgbuf, reset, strlen(reset));
}
strbuf_add(&msgbuf, line + len, org_len - len);
emit_line(ecbdata->opt, "", "", msgbuf.buf, msgbuf.len);
strbuf_release(&msgbuf);
}
diff: refactor tempfile cleanup handling There are two pieces of code that create tempfiles for diff: run_external_diff and run_textconv. The former cleans up its tempfiles in the face of premature death (i.e., by die() or by signal), but the latter does not. After this patch, they will both use the same cleanup routines. To make clear what the change is, let me first explain what happens now: - run_external_diff uses a static global array of 2 diff_tempfile structs (since it knows it will always need exactly 2 tempfiles). It calls prepare_temp_file (which doesn't know anything about the global array) on each of the structs, creating the tempfiles that need to be cleaned up. It then registers atexit and signal handlers to look through the global array and remove the tempfiles. If it succeeds, it calls the handler manually (which marks the tempfile structs as unused). - textconv has its own tempfile struct, which it allocates using prepare_temp_file and cleans up manually. No signal or atexit handlers. The new code moves the installation of cleanup handlers into the prepare_temp_file function. Which means that that function now has to understand that there is static tempfile storage. So what happens now is: - run_external_diff calls prepare_temp_file - prepare_temp_file calls claim_diff_tempfile, which allocates an unused slot from our global array - prepare_temp_file installs (if they have not already been installed) atexit and signal handlers for cleanup - prepare_temp_file sets up the tempfile as usual - prepare_temp_file returns a pointer to the allocated tempfile The advantage being that run_external_diff no longer has to care about setting up cleanup handlers. Now by virtue of calling prepare_temp_file, run_textconv gets the same benefit, as will any future users of prepare_temp_file. There are also a few side benefits to the specific implementation: - we now install cleanup handlers _before_ allocating the tempfile, closing a race which could leave temp cruft - when allocating a slot in the global array, we will now detect a situation where the old slots were not properly vacated (i.e., somebody forgot to call remove upon leaving the function). In the old code, such a situation would silently overwrite the tempfile names, meaning we would forget to clean them up. The new code dies with a bug warning. - we make sure only to install the signal handler once. This isn't a big deal, since we are just overwriting the old handler, but will become an issue when a later patch converts the code to use sigchain Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-01-22 05:59:56 +00:00
static struct diff_tempfile *claim_diff_tempfile(void) {
int i;
for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
if (!diff_temp[i].name)
return diff_temp + i;
die("BUG: diff is failing to clean up its tempfiles");
}
static int remove_tempfile_installed;
static void remove_tempfile(void)
{
int i;
for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
if (diff_temp[i].name == diff_temp[i].tmp_path)
unlink_or_warn(diff_temp[i].name);
diff_temp[i].name = NULL;
}
diff: refactor tempfile cleanup handling There are two pieces of code that create tempfiles for diff: run_external_diff and run_textconv. The former cleans up its tempfiles in the face of premature death (i.e., by die() or by signal), but the latter does not. After this patch, they will both use the same cleanup routines. To make clear what the change is, let me first explain what happens now: - run_external_diff uses a static global array of 2 diff_tempfile structs (since it knows it will always need exactly 2 tempfiles). It calls prepare_temp_file (which doesn't know anything about the global array) on each of the structs, creating the tempfiles that need to be cleaned up. It then registers atexit and signal handlers to look through the global array and remove the tempfiles. If it succeeds, it calls the handler manually (which marks the tempfile structs as unused). - textconv has its own tempfile struct, which it allocates using prepare_temp_file and cleans up manually. No signal or atexit handlers. The new code moves the installation of cleanup handlers into the prepare_temp_file function. Which means that that function now has to understand that there is static tempfile storage. So what happens now is: - run_external_diff calls prepare_temp_file - prepare_temp_file calls claim_diff_tempfile, which allocates an unused slot from our global array - prepare_temp_file installs (if they have not already been installed) atexit and signal handlers for cleanup - prepare_temp_file sets up the tempfile as usual - prepare_temp_file returns a pointer to the allocated tempfile The advantage being that run_external_diff no longer has to care about setting up cleanup handlers. Now by virtue of calling prepare_temp_file, run_textconv gets the same benefit, as will any future users of prepare_temp_file. There are also a few side benefits to the specific implementation: - we now install cleanup handlers _before_ allocating the tempfile, closing a race which could leave temp cruft - when allocating a slot in the global array, we will now detect a situation where the old slots were not properly vacated (i.e., somebody forgot to call remove upon leaving the function). In the old code, such a situation would silently overwrite the tempfile names, meaning we would forget to clean them up. The new code dies with a bug warning. - we make sure only to install the signal handler once. This isn't a big deal, since we are just overwriting the old handler, but will become an issue when a later patch converts the code to use sigchain Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-01-22 05:59:56 +00:00
}
static void remove_tempfile_on_signal(int signo)
{
remove_tempfile();
sigchain_pop(signo);
diff: refactor tempfile cleanup handling There are two pieces of code that create tempfiles for diff: run_external_diff and run_textconv. The former cleans up its tempfiles in the face of premature death (i.e., by die() or by signal), but the latter does not. After this patch, they will both use the same cleanup routines. To make clear what the change is, let me first explain what happens now: - run_external_diff uses a static global array of 2 diff_tempfile structs (since it knows it will always need exactly 2 tempfiles). It calls prepare_temp_file (which doesn't know anything about the global array) on each of the structs, creating the tempfiles that need to be cleaned up. It then registers atexit and signal handlers to look through the global array and remove the tempfiles. If it succeeds, it calls the handler manually (which marks the tempfile structs as unused). - textconv has its own tempfile struct, which it allocates using prepare_temp_file and cleans up manually. No signal or atexit handlers. The new code moves the installation of cleanup handlers into the prepare_temp_file function. Which means that that function now has to understand that there is static tempfile storage. So what happens now is: - run_external_diff calls prepare_temp_file - prepare_temp_file calls claim_diff_tempfile, which allocates an unused slot from our global array - prepare_temp_file installs (if they have not already been installed) atexit and signal handlers for cleanup - prepare_temp_file sets up the tempfile as usual - prepare_temp_file returns a pointer to the allocated tempfile The advantage being that run_external_diff no longer has to care about setting up cleanup handlers. Now by virtue of calling prepare_temp_file, run_textconv gets the same benefit, as will any future users of prepare_temp_file. There are also a few side benefits to the specific implementation: - we now install cleanup handlers _before_ allocating the tempfile, closing a race which could leave temp cruft - when allocating a slot in the global array, we will now detect a situation where the old slots were not properly vacated (i.e., somebody forgot to call remove upon leaving the function). In the old code, such a situation would silently overwrite the tempfile names, meaning we would forget to clean them up. The new code dies with a bug warning. - we make sure only to install the signal handler once. This isn't a big deal, since we are just overwriting the old handler, but will become an issue when a later patch converts the code to use sigchain Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-01-22 05:59:56 +00:00
raise(signo);
}
static void print_line_count(FILE *file, int count)
{
switch (count) {
case 0:
fprintf(file, "0,0");
break;
case 1:
fprintf(file, "1");
break;
default:
fprintf(file, "1,%d", count);
break;
}
}
static void emit_rewrite_lines(struct emit_callback *ecb,
int prefix, const char *data, int size)
{
const char *endp = NULL;
static const char *nneof = " No newline at end of file\n";
const char *old = diff_get_color(ecb->color_diff, DIFF_FILE_OLD);
const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
while (0 < size) {
int len;
endp = memchr(data, '\n', size);
len = endp ? (endp - data + 1) : size;
if (prefix != '+') {
ecb->lno_in_preimage++;
emit_line_0(ecb->opt, old, reset, '-',
data, len);
} else {
ecb->lno_in_postimage++;
emit_add_line(reset, ecb, data, len);
}
size -= len;
data += len;
}
if (!endp) {
const char *plain = diff_get_color(ecb->color_diff,
DIFF_PLAIN);
emit_line_0(ecb->opt, plain, reset, '\\',
nneof, strlen(nneof));
}
}
static void emit_rewrite_diff(const char *name_a,
const char *name_b,
struct diff_filespec *one,
struct diff_filespec *two,
struct userdiff_driver *textconv_one,
struct userdiff_driver *textconv_two,
struct diff_options *o)
{
int lc_a, lc_b;
const char *name_a_tab, *name_b_tab;
const char *metainfo = diff_get_color(o->use_color, DIFF_METAINFO);
const char *fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
const char *reset = diff_get_color(o->use_color, DIFF_RESET);
static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
const char *a_prefix, *b_prefix;
char *data_one, *data_two;
size_t size_one, size_two;
struct emit_callback ecbdata;
char *line_prefix = "";
struct strbuf *msgbuf;
if (o && o->output_prefix) {
msgbuf = o->output_prefix(o, o->output_prefix_data);
line_prefix = msgbuf->buf;
}
if (diff_mnemonic_prefix && DIFF_OPT_TST(o, REVERSE_DIFF)) {
a_prefix = o->b_prefix;
b_prefix = o->a_prefix;
} else {
a_prefix = o->a_prefix;
b_prefix = o->b_prefix;
}
name_a += (*name_a == '/');
name_b += (*name_b == '/');
name_a_tab = strchr(name_a, ' ') ? "\t" : "";
name_b_tab = strchr(name_b, ' ') ? "\t" : "";
strbuf_reset(&a_name);
strbuf_reset(&b_name);
quote_two_c_style(&a_name, a_prefix, name_a, 0);
quote_two_c_style(&b_name, b_prefix, name_b, 0);
size_one = fill_textconv(textconv_one, one, &data_one);
size_two = fill_textconv(textconv_two, two, &data_two);
memset(&ecbdata, 0, sizeof(ecbdata));
color: delay auto-color decision until point of use When we read a color value either from a config file or from the command line, we use git_config_colorbool to convert it from the tristate always/never/auto into a single yes/no boolean value. This has some timing implications with respect to starting a pager. If we start (or decide not to start) the pager before checking the colorbool, everything is fine. Either isatty(1) will give us the right information, or we will properly check for pager_in_use(). However, if we decide to start a pager after we have checked the colorbool, things are not so simple. If stdout is a tty, then we will have already decided to use color. However, the user may also have configured color.pager not to use color with the pager. In this case, we need to actually turn off color. Unfortunately, the pager code has no idea which color variables were turned on (and there are many of them throughout the code, and they may even have been manipulated after the colorbool selection by something like "--color" on the command line). This bug can be seen any time a pager is started after config and command line options are checked. This has affected "git diff" since 89d07f7 (diff: don't run pager if user asked for a diff style exit code, 2007-08-12). It has also affect the log family since 1fda91b (Fix 'git log' early pager startup error case, 2010-08-24). This patch splits the notion of parsing a colorbool and actually checking the configuration. The "use_color" variables now have an additional possible value, GIT_COLOR_AUTO. Users of the variable should use the new "want_color()" wrapper, which will lazily determine and cache the auto-color decision. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-08-18 05:04:23 +00:00
ecbdata.color_diff = want_color(o->use_color);
ecbdata.found_changesp = &o->found_changes;
ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
ecbdata.opt = o;
if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
mmfile_t mf1, mf2;
mf1.ptr = (char *)data_one;
mf2.ptr = (char *)data_two;
mf1.size = size_one;
mf2.size = size_two;
check_blank_at_eof(&mf1, &mf2, &ecbdata);
}
ecbdata.lno_in_preimage = 1;
ecbdata.lno_in_postimage = 1;
lc_a = count_lines(data_one, size_one);
lc_b = count_lines(data_two, size_two);
fprintf(o->file,
"%s%s--- %s%s%s\n%s%s+++ %s%s%s\n%s%s@@ -",
line_prefix, metainfo, a_name.buf, name_a_tab, reset,
line_prefix, metainfo, b_name.buf, name_b_tab, reset,
line_prefix, fraginfo);
if (!o->irreversible_delete)
print_line_count(o->file, lc_a);
else
fprintf(o->file, "?,?");
fprintf(o->file, " +");
print_line_count(o->file, lc_b);
fprintf(o->file, " @@%s\n", reset);
if (lc_a && !o->irreversible_delete)
emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
if (lc_b)
emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
if (textconv_one)
free((char *)data_one);
if (textconv_two)
free((char *)data_two);
}
struct diff_words_buffer {
mmfile_t text;
long alloc;
2009-01-17 16:29:44 +00:00
struct diff_words_orig {
const char *begin, *end;
} *orig;
int orig_nr, orig_alloc;
};
static void diff_words_append(char *line, unsigned long len,
struct diff_words_buffer *buffer)
{
ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
line++;
len--;
memcpy(buffer->text.ptr + buffer->text.size, line, len);
buffer->text.size += len;
buffer->text.ptr[buffer->text.size] = '\0';
}
struct diff_words_style_elem {
const char *prefix;
const char *suffix;
const char *color; /* NULL; filled in by the setup code if
* color is enabled */
};
struct diff_words_style {
enum diff_words_type type;
struct diff_words_style_elem new, old, ctx;
const char *newline;
};
Fix sparse warnings Fix warnings from 'make check'. - These files don't include 'builtin.h' causing sparse to complain that cmd_* isn't declared: builtin/clone.c:364, builtin/fetch-pack.c:797, builtin/fmt-merge-msg.c:34, builtin/hash-object.c:78, builtin/merge-index.c:69, builtin/merge-recursive.c:22 builtin/merge-tree.c:341, builtin/mktag.c:156, builtin/notes.c:426 builtin/notes.c:822, builtin/pack-redundant.c:596, builtin/pack-refs.c:10, builtin/patch-id.c:60, builtin/patch-id.c:149, builtin/remote.c:1512, builtin/remote-ext.c:240, builtin/remote-fd.c:53, builtin/reset.c:236, builtin/send-pack.c:384, builtin/unpack-file.c:25, builtin/var.c:75 - These files have symbols which should be marked static since they're only file scope: submodule.c:12, diff.c:631, replace_object.c:92, submodule.c:13, submodule.c:14, trace.c:78, transport.c:195, transport-helper.c:79, unpack-trees.c:19, url.c:3, url.c:18, url.c:104, url.c:117, url.c:123, url.c:129, url.c:136, thread-utils.c:21, thread-utils.c:48 - These files redeclare symbols to be different types: builtin/index-pack.c:210, parse-options.c:564, parse-options.c:571, usage.c:49, usage.c:58, usage.c:63, usage.c:72 - These files use a literal integer 0 when they really should use a NULL pointer: daemon.c:663, fast-import.c:2942, imap-send.c:1072, notes-merge.c:362 While we're in the area, clean up some unused #includes in builtin files (mostly exec_cmd.h). Signed-off-by: Stephen Boyd <bebarino@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-03-22 07:51:05 +00:00
static struct diff_words_style diff_words_styles[] = {
{ DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
{ DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
{ DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
};
struct diff_words_data {
struct diff_words_buffer minus, plus;
2009-01-17 16:29:44 +00:00
const char *current_plus;
int last_minus;
struct diff_options *opt;
regex_t *word_regex;
enum diff_words_type type;
struct diff_words_style *style;
};
static int fn_out_diff_words_write_helper(FILE *fp,
struct diff_words_style_elem *st_el,
const char *newline,
size_t count, const char *buf,
const char *line_prefix)
{
int print = 0;
while (count) {
char *p = memchr(buf, '\n', count);
if (print)
fputs(line_prefix, fp);
if (p != buf) {
if (st_el->color && fputs(st_el->color, fp) < 0)
return -1;
if (fputs(st_el->prefix, fp) < 0 ||
fwrite(buf, p ? p - buf : count, 1, fp) != 1 ||
fputs(st_el->suffix, fp) < 0)
return -1;
if (st_el->color && *st_el->color
&& fputs(GIT_COLOR_RESET, fp) < 0)
return -1;
}
if (!p)
return 0;
if (fputs(newline, fp) < 0)
return -1;
count -= p + 1 - buf;
buf = p + 1;
print = 1;
}
return 0;
}
/*
* '--color-words' algorithm can be described as:
*
* 1. collect a the minus/plus lines of a diff hunk, divided into
* minus-lines and plus-lines;
*
* 2. break both minus-lines and plus-lines into words and
* place them into two mmfile_t with one word for each line;
*
* 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
*
* And for the common parts of the both file, we output the plus side text.
* diff_words->current_plus is used to trace the current position of the plus file
* which printed. diff_words->last_minus is used to trace the last minus word
* printed.
*
* For '--graph' to work with '--color-words', we need to output the graph prefix
* on each line of color words output. Generally, there are two conditions on
* which we should output the prefix.
*
* 1. diff_words->last_minus == 0 &&
* diff_words->current_plus == diff_words->plus.text.ptr
*
* that is: the plus text must start as a new line, and if there is no minus
* word printed, a graph prefix must be printed.
*
* 2. diff_words->current_plus > diff_words->plus.text.ptr &&
* *(diff_words->current_plus - 1) == '\n'
*
* that is: a graph prefix must be printed following a '\n'
*/
static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
{
if ((diff_words->last_minus == 0 &&
diff_words->current_plus == diff_words->plus.text.ptr) ||
(diff_words->current_plus > diff_words->plus.text.ptr &&
*(diff_words->current_plus - 1) == '\n')) {
return 1;
} else {
return 0;
}
}
static void fn_out_diff_words_aux(void *priv, char *line, unsigned long len)
{
struct diff_words_data *diff_words = priv;
struct diff_words_style *style = diff_words->style;
2009-01-17 16:29:44 +00:00
int minus_first, minus_len, plus_first, plus_len;
const char *minus_begin, *minus_end, *plus_begin, *plus_end;
struct diff_options *opt = diff_words->opt;
struct strbuf *msgbuf;
char *line_prefix = "";
2009-01-17 16:29:44 +00:00
if (line[0] != '@' || parse_hunk_header(line, len,
&minus_first, &minus_len, &plus_first, &plus_len))
return;
assert(opt);
if (opt->output_prefix) {
msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
line_prefix = msgbuf->buf;
}
2009-01-17 16:29:44 +00:00
/* POSIX requires that first be decremented by one if len == 0... */
if (minus_len) {
minus_begin = diff_words->minus.orig[minus_first].begin;
minus_end =
diff_words->minus.orig[minus_first + minus_len - 1].end;
} else
minus_begin = minus_end =
diff_words->minus.orig[minus_first].end;
if (plus_len) {
plus_begin = diff_words->plus.orig[plus_first].begin;
plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
} else
plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
if (color_words_output_graph_prefix(diff_words)) {
fputs(line_prefix, diff_words->opt->file);
}
if (diff_words->current_plus != plus_begin) {
fn_out_diff_words_write_helper(diff_words->opt->file,
&style->ctx, style->newline,
plus_begin - diff_words->current_plus,
diff_words->current_plus, line_prefix);
if (*(plus_begin - 1) == '\n')
fputs(line_prefix, diff_words->opt->file);
}
if (minus_begin != minus_end) {
fn_out_diff_words_write_helper(diff_words->opt->file,
&style->old, style->newline,
minus_end - minus_begin, minus_begin,
line_prefix);
}
if (plus_begin != plus_end) {
fn_out_diff_words_write_helper(diff_words->opt->file,
&style->new, style->newline,
plus_end - plus_begin, plus_begin,
line_prefix);
}
2009-01-17 16:29:44 +00:00
diff_words->current_plus = plus_end;
diff_words->last_minus = minus_first;
}
/* This function starts looking at *begin, and returns 0 iff a word was found. */
static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
int *begin, int *end)
{
if (word_regex && *begin < buffer->size) {
regmatch_t match[1];
if (!regexec(word_regex, buffer->ptr + *begin, 1, match, 0)) {
char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
'\n', match[0].rm_eo - match[0].rm_so);
*end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
*begin += match[0].rm_so;
return *begin >= *end;
}
return -1;
}
/* find the next word */
while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
(*begin)++;
if (*begin >= buffer->size)
return -1;
/* find the end of the word */
*end = *begin + 1;
while (*end < buffer->size && !isspace(buffer->ptr[*end]))
(*end)++;
return 0;
}
/*
2009-01-17 16:29:44 +00:00
* This function splits the words in buffer->text, stores the list with
* newline separator into out, and saves the offsets of the original words
* in buffer->orig.
*/
static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
regex_t *word_regex)
{
2009-01-17 16:29:44 +00:00
int i, j;
long alloc = 0;
2009-01-17 16:29:44 +00:00
out->size = 0;
out->ptr = NULL;
2009-01-17 16:29:44 +00:00
/* fake an empty "0th" word */
ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
buffer->orig_nr = 1;
for (i = 0; i < buffer->text.size; i++) {
if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
return;
2009-01-17 16:29:44 +00:00
/* store original boundaries */
ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
buffer->orig_alloc);
buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
buffer->orig_nr++;
/* store one word */
ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
2009-01-17 16:29:44 +00:00
memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
out->ptr[out->size + j - i] = '\n';
out->size += j - i + 1;
i = j - 1;
}
}
/* this executes the word diff on the accumulated buffers */
static void diff_words_show(struct diff_words_data *diff_words)
{
xpparam_t xpp;
xdemitconf_t xecfg;
mmfile_t minus, plus;
struct diff_words_style *style = diff_words->style;
struct diff_options *opt = diff_words->opt;
struct strbuf *msgbuf;
char *line_prefix = "";
assert(opt);
if (opt->output_prefix) {
msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
line_prefix = msgbuf->buf;
}
2009-01-17 16:29:44 +00:00
/* special case: only removal */
if (!diff_words->plus.text.size) {
fputs(line_prefix, diff_words->opt->file);
fn_out_diff_words_write_helper(diff_words->opt->file,
&style->old, style->newline,
diff_words->minus.text.size,
diff_words->minus.text.ptr, line_prefix);
2009-01-17 16:29:44 +00:00
diff_words->minus.text.size = 0;
return;
}
diff_words->current_plus = diff_words->plus.text.ptr;
diff_words->last_minus = 0;
memset(&xpp, 0, sizeof(xpp));
memset(&xecfg, 0, sizeof(xecfg));
diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
xpp.flags = 0;
/* as only the hunk header will be parsed, we need a 0-context */
2009-01-17 16:29:44 +00:00
xecfg.ctxlen = 0;
xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, diff_words,
&xpp, &xecfg);
free(minus.ptr);
free(plus.ptr);
2009-01-17 16:29:44 +00:00
if (diff_words->current_plus != diff_words->plus.text.ptr +
diff_words->plus.text.size) {
if (color_words_output_graph_prefix(diff_words))
fputs(line_prefix, diff_words->opt->file);
fn_out_diff_words_write_helper(diff_words->opt->file,
&style->ctx, style->newline,
2009-01-17 16:29:44 +00:00
diff_words->plus.text.ptr + diff_words->plus.text.size
- diff_words->current_plus, diff_words->current_plus,
line_prefix);
}
diff_words->minus.text.size = diff_words->plus.text.size = 0;
}
2009-10-30 17:09:06 +00:00
/* In "color-words" mode, show word-diff of words accumulated in the buffer */
static void diff_words_flush(struct emit_callback *ecbdata)
{
if (ecbdata->diff_words->minus.text.size ||
ecbdata->diff_words->plus.text.size)
diff_words_show(ecbdata->diff_words);
}
static void free_diff_words_data(struct emit_callback *ecbdata)
{
if (ecbdata->diff_words) {
2009-10-30 17:09:06 +00:00
diff_words_flush(ecbdata);
Avoid unnecessary "if-before-free" tests. This change removes all obvious useless if-before-free tests. E.g., it replaces code like this: if (some_expression) free (some_expression); with the now-equivalent: free (some_expression); It is equivalent not just because POSIX has required free(NULL) to work for a long time, but simply because it has worked for so long that no reasonable porting target fails the test. Here's some evidence from nearly 1.5 years ago: http://www.winehq.org/pipermail/wine-patches/2006-October/031544.html FYI, the change below was prepared by running the following: git ls-files -z | xargs -0 \ perl -0x3b -pi -e \ 's/\bif\s*\(\s*(\S+?)(?:\s*!=\s*NULL)?\s*\)\s+(free\s*\(\s*\1\s*\))/$2/s' Note however, that it doesn't handle brace-enclosed blocks like "if (x) { free (x); }". But that's ok, since there were none like that in git sources. Beware: if you do use the above snippet, note that it can produce syntactically invalid C code. That happens when the affected "if"-statement has a matching "else". E.g., it would transform this if (x) free (x); else foo (); into this: free (x); else foo (); There were none of those here, either. If you're interested in automating detection of the useless tests, you might like the useless-if-before-free script in gnulib: [it *does* detect brace-enclosed free statements, and has a --name=S option to make it detect free-like functions with different names] http://git.sv.gnu.org/gitweb/?p=gnulib.git;a=blob;f=build-aux/useless-if-before-free Addendum: Remove one more (in imap-send.c), spotted by Jean-Luc Herren <jlh@gmx.ch>. Signed-off-by: Jim Meyering <meyering@redhat.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-01-31 17:26:32 +00:00
free (ecbdata->diff_words->minus.text.ptr);
2009-01-17 16:29:44 +00:00
free (ecbdata->diff_words->minus.orig);
Avoid unnecessary "if-before-free" tests. This change removes all obvious useless if-before-free tests. E.g., it replaces code like this: if (some_expression) free (some_expression); with the now-equivalent: free (some_expression); It is equivalent not just because POSIX has required free(NULL) to work for a long time, but simply because it has worked for so long that no reasonable porting target fails the test. Here's some evidence from nearly 1.5 years ago: http://www.winehq.org/pipermail/wine-patches/2006-October/031544.html FYI, the change below was prepared by running the following: git ls-files -z | xargs -0 \ perl -0x3b -pi -e \ 's/\bif\s*\(\s*(\S+?)(?:\s*!=\s*NULL)?\s*\)\s+(free\s*\(\s*\1\s*\))/$2/s' Note however, that it doesn't handle brace-enclosed blocks like "if (x) { free (x); }". But that's ok, since there were none like that in git sources. Beware: if you do use the above snippet, note that it can produce syntactically invalid C code. That happens when the affected "if"-statement has a matching "else". E.g., it would transform this if (x) free (x); else foo (); into this: free (x); else foo (); There were none of those here, either. If you're interested in automating detection of the useless tests, you might like the useless-if-before-free script in gnulib: [it *does* detect brace-enclosed free statements, and has a --name=S option to make it detect free-like functions with different names] http://git.sv.gnu.org/gitweb/?p=gnulib.git;a=blob;f=build-aux/useless-if-before-free Addendum: Remove one more (in imap-send.c), spotted by Jean-Luc Herren <jlh@gmx.ch>. Signed-off-by: Jim Meyering <meyering@redhat.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-01-31 17:26:32 +00:00
free (ecbdata->diff_words->plus.text.ptr);
2009-01-17 16:29:44 +00:00
free (ecbdata->diff_words->plus.orig);
if (ecbdata->diff_words->word_regex) {
regfree(ecbdata->diff_words->word_regex);
free(ecbdata->diff_words->word_regex);
}
free(ecbdata->diff_words);
ecbdata->diff_words = NULL;
}
}
const char *diff_get_color(int diff_use_color, enum color_diff ix)
{
color: delay auto-color decision until point of use When we read a color value either from a config file or from the command line, we use git_config_colorbool to convert it from the tristate always/never/auto into a single yes/no boolean value. This has some timing implications with respect to starting a pager. If we start (or decide not to start) the pager before checking the colorbool, everything is fine. Either isatty(1) will give us the right information, or we will properly check for pager_in_use(). However, if we decide to start a pager after we have checked the colorbool, things are not so simple. If stdout is a tty, then we will have already decided to use color. However, the user may also have configured color.pager not to use color with the pager. In this case, we need to actually turn off color. Unfortunately, the pager code has no idea which color variables were turned on (and there are many of them throughout the code, and they may even have been manipulated after the colorbool selection by something like "--color" on the command line). This bug can be seen any time a pager is started after config and command line options are checked. This has affected "git diff" since 89d07f7 (diff: don't run pager if user asked for a diff style exit code, 2007-08-12). It has also affect the log family since 1fda91b (Fix 'git log' early pager startup error case, 2010-08-24). This patch splits the notion of parsing a colorbool and actually checking the configuration. The "use_color" variables now have an additional possible value, GIT_COLOR_AUTO. Users of the variable should use the new "want_color()" wrapper, which will lazily determine and cache the auto-color decision. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-08-18 05:04:23 +00:00
if (want_color(diff_use_color))
return diff_colors[ix];
return "";
}
static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
{
const char *cp;
unsigned long allot;
size_t l = len;
if (ecb->truncate)
return ecb->truncate(line, len);
cp = line;
allot = l;
while (0 < l) {
(void) utf8_width(&cp, &l);
if (!cp)
break; /* truncated in the middle? */
}
return allot - l;
}
static void find_lno(const char *line, struct emit_callback *ecbdata)
{
const char *p;
ecbdata->lno_in_preimage = 0;
ecbdata->lno_in_postimage = 0;
p = strchr(line, '-');
if (!p)
return; /* cannot happen */
ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
p = strchr(p, '+');
if (!p)
return; /* cannot happen */
ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
}
static void fn_out_consume(void *priv, char *line, unsigned long len)
{
struct emit_callback *ecbdata = priv;
const char *meta = diff_get_color(ecbdata->color_diff, DIFF_METAINFO);
const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
struct diff_options *o = ecbdata->opt;
char *line_prefix = "";
struct strbuf *msgbuf;
if (o && o->output_prefix) {
msgbuf = o->output_prefix(o, o->output_prefix_data);
line_prefix = msgbuf->buf;
}
if (ecbdata->header) {
fprintf(ecbdata->opt->file, "%s", ecbdata->header->buf);
strbuf_reset(ecbdata->header);
ecbdata->header = NULL;
}
*(ecbdata->found_changesp) = 1;
if (ecbdata->label_path[0]) {
const char *name_a_tab, *name_b_tab;
name_a_tab = strchr(ecbdata->label_path[0], ' ') ? "\t" : "";
name_b_tab = strchr(ecbdata->label_path[1], ' ') ? "\t" : "";
fprintf(ecbdata->opt->file, "%s%s--- %s%s%s\n",
line_prefix, meta, ecbdata->label_path[0], reset, name_a_tab);
fprintf(ecbdata->opt->file, "%s%s+++ %s%s%s\n",
line_prefix, meta, ecbdata->label_path[1], reset, name_b_tab);
ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
}
if (diff_suppress_blank_empty
&& len == 2 && line[0] == ' ' && line[1] == '\n') {
line[0] = '\n';
len = 1;
}
if (line[0] == '@') {
2009-10-30 17:09:06 +00:00
if (ecbdata->diff_words)
diff_words_flush(ecbdata);
len = sane_truncate_line(ecbdata, line, len);
find_lno(line, ecbdata);
emit_hunk_header(ecbdata, line, len);
if (line[len-1] != '\n')
putc('\n', ecbdata->opt->file);
return;
}
if (len < 1) {
emit_line(ecbdata->opt, reset, reset, line, len);
if (ecbdata->diff_words
&& ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN)
fputs("~\n", ecbdata->opt->file);
return;
}
if (ecbdata->diff_words) {
if (line[0] == '-') {
diff_words_append(line, len,
&ecbdata->diff_words->minus);
return;
} else if (line[0] == '+') {
diff_words_append(line, len,
&ecbdata->diff_words->plus);
return;
} else if (!prefixcmp(line, "\\ ")) {
/*
* Eat the "no newline at eof" marker as if we
* saw a "+" or "-" line with nothing on it,
* and return without diff_words_flush() to
* defer processing. If this is the end of
* preimage, more "+" lines may come after it.
*/
return;
}
2009-10-30 17:09:06 +00:00
diff_words_flush(ecbdata);
if (ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN) {
emit_line(ecbdata->opt, plain, reset, line, len);
fputs("~\n", ecbdata->opt->file);
} else {
/*
* Skip the prefix character, if any. With
* diff_suppress_blank_empty, there may be
* none.
*/
if (line[0] != '\n') {
line++;
len--;
}
emit_line(ecbdata->opt, plain, reset, line, len);
}
return;
}
if (line[0] != '+') {
const char *color =
diff_get_color(ecbdata->color_diff,
line[0] == '-' ? DIFF_FILE_OLD : DIFF_PLAIN);
ecbdata->lno_in_preimage++;
if (line[0] == ' ')
ecbdata->lno_in_postimage++;
emit_line(ecbdata->opt, color, reset, line, len);
} else {
ecbdata->lno_in_postimage++;
emit_add_line(reset, ecbdata, line + 1, len - 1);
}
}
static char *pprint_rename(const char *a, const char *b)
{
const char *old = a;
const char *new = b;
struct strbuf name = STRBUF_INIT;
int pfx_length, sfx_length;
int len_a = strlen(a);
int len_b = strlen(b);
int a_midlen, b_midlen;
int qlen_a = quote_c_style(a, NULL, NULL, 0);
int qlen_b = quote_c_style(b, NULL, NULL, 0);
if (qlen_a || qlen_b) {
quote_c_style(a, &name, NULL, 0);
strbuf_addstr(&name, " => ");
quote_c_style(b, &name, NULL, 0);
return strbuf_detach(&name, NULL);
}
/* Find common prefix */
pfx_length = 0;
while (*old && *new && *old == *new) {
if (*old == '/')
pfx_length = old - a + 1;
old++;
new++;
}
/* Find common suffix */
old = a + len_a;
new = b + len_b;
sfx_length = 0;
while (a <= old && b <= new && *old == *new) {
if (*old == '/')
sfx_length = len_a - (old - a);
old--;
new--;
}
/*
* pfx{mid-a => mid-b}sfx
* {pfx-a => pfx-b}sfx
* pfx{sfx-a => sfx-b}
* name-a => name-b
*/
a_midlen = len_a - pfx_length - sfx_length;
b_midlen = len_b - pfx_length - sfx_length;
if (a_midlen < 0)
a_midlen = 0;
if (b_midlen < 0)
b_midlen = 0;
strbuf_grow(&name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
if (pfx_length + sfx_length) {
strbuf_add(&name, a, pfx_length);
strbuf_addch(&name, '{');
}
strbuf_add(&name, a + pfx_length, a_midlen);
strbuf_addstr(&name, " => ");
strbuf_add(&name, b + pfx_length, b_midlen);
if (pfx_length + sfx_length) {
strbuf_addch(&name, '}');
strbuf_add(&name, a + len_a - sfx_length, sfx_length);
}
return strbuf_detach(&name, NULL);
}
struct diffstat_t {
int nr;
int alloc;
struct diffstat_file {
char *from_name;
char *name;
char *print_name;
unsigned is_unmerged:1;
unsigned is_binary:1;
unsigned is_renamed:1;
uintmax_t added, deleted;
} **files;
};
static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
const char *name_a,
const char *name_b)
{
struct diffstat_file *x;
x = xcalloc(sizeof (*x), 1);
if (diffstat->nr == diffstat->alloc) {
diffstat->alloc = alloc_nr(diffstat->alloc);
diffstat->files = xrealloc(diffstat->files,
diffstat->alloc * sizeof(x));
}
diffstat->files[diffstat->nr++] = x;
if (name_b) {
x->from_name = xstrdup(name_a);
x->name = xstrdup(name_b);
x->is_renamed = 1;
}
else {
x->from_name = NULL;
x->name = xstrdup(name_a);
}
return x;
}
static void diffstat_consume(void *priv, char *line, unsigned long len)
{
struct diffstat_t *diffstat = priv;
struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
if (line[0] == '+')
x->added++;
else if (line[0] == '-')
x->deleted++;
}
const char mime_boundary_leader[] = "------------";
static int scale_linear(int it, int width, int max_change)
{
if (!it)
return 0;
/*
* make sure that at least one '-' or '+' is printed if
* there is any change to this path. The easiest way is to
* scale linearly as if the alloted width is one column shorter
* than it is, and then add 1 to the result.
*/
return 1 + (it * (width - 1) / max_change);
}
static void show_name(FILE *file,
const char *prefix, const char *name, int len)
{
fprintf(file, " %s%-*s |", prefix, len, name);
}
static void show_graph(FILE *file, char ch, int cnt, const char *set, const char *reset)
{
if (cnt <= 0)
return;
fprintf(file, "%s", set);
while (cnt--)
putc(ch, file);
fprintf(file, "%s", reset);
}
static void fill_print_name(struct diffstat_file *file)
{
char *pname;
if (file->print_name)
return;
if (!file->is_renamed) {
struct strbuf buf = STRBUF_INIT;
if (quote_c_style(file->name, &buf, NULL, 0)) {
pname = strbuf_detach(&buf, NULL);
} else {
pname = file->name;
strbuf_release(&buf);
}
} else {
pname = pprint_rename(file->from_name, file->name);
}
file->print_name = pname;
}
int print_stat_summary(FILE *fp, int files, int insertions, int deletions)
{
struct strbuf sb = STRBUF_INIT;
int ret;
if (!files) {
assert(insertions == 0 && deletions == 0);
return fputs(_(" 0 files changed\n"), fp);
}
strbuf_addf(&sb,
Q_(" %d file changed", " %d files changed", files),
files);
/*
* For binary diff, the caller may want to print "x files
* changed" with insertions == 0 && deletions == 0.
*
* Not omitting "0 insertions(+), 0 deletions(-)" in this case
* is probably less confusing (i.e skip over "2 files changed
* but nothing about added/removed lines? Is this a bug in Git?").
*/
if (insertions || deletions == 0) {
/*
* TRANSLATORS: "+" in (+) is a line addition marker;
* do not translate it.
*/
strbuf_addf(&sb,
Q_(", %d insertion(+)", ", %d insertions(+)",
insertions),
insertions);
}
if (deletions || insertions == 0) {
/*
* TRANSLATORS: "-" in (-) is a line removal marker;
* do not translate it.
*/
strbuf_addf(&sb,
Q_(", %d deletion(-)", ", %d deletions(-)",
deletions),
deletions);
}
strbuf_addch(&sb, '\n');
ret = fputs(sb.buf, fp);
strbuf_release(&sb);
return ret;
}
static void show_stats(struct diffstat_t *data, struct diff_options *options)
{
int i, len, add, del, adds = 0, dels = 0;
uintmax_t max_change = 0, max_len = 0;
int total_files = data->nr;
int width, name_width, count;
const char *reset, *add_c, *del_c;
const char *line_prefix = "";
int extra_shown = 0;
struct strbuf *msg = NULL;
if (data->nr == 0)
return;
if (options->output_prefix) {
msg = options->output_prefix(options, options->output_prefix_data);
line_prefix = msg->buf;
}
width = options->stat_width ? options->stat_width : 80;
name_width = options->stat_name_width ? options->stat_name_width : 50;
count = options->stat_count ? options->stat_count : data->nr;
/* Sanity: give at least 5 columns to the graph,
* but leave at least 10 columns for the name.
*/
if (width < 25)
width = 25;
if (name_width < 10)
name_width = 10;
else if (width < name_width + 15)
name_width = width - 15;
/* Find the longest filename and max number of changes */
reset = diff_get_color_opt(options, DIFF_RESET);
add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
for (i = 0; (i < count) && (i < data->nr); i++) {
struct diffstat_file *file = data->files[i];
uintmax_t change = file->added + file->deleted;
if (!data->files[i]->is_renamed &&
(change == 0)) {
count++; /* not shown == room for one more */
continue;
}
fill_print_name(file);
len = strlen(file->print_name);
if (max_len < len)
max_len = len;
if (file->is_binary || file->is_unmerged)
continue;
if (max_change < change)
max_change = change;
}
count = i; /* min(count, data->nr) */
/* Compute the width of the graph part;
* 10 is for one blank at the beginning of the line plus
* " | count " between the name and the graph.
*
* From here on, name_width is the width of the name area,
* and width is the width of the graph area.
*/
name_width = (name_width < max_len) ? name_width : max_len;
if (width < (name_width + 10) + max_change)
width = width - (name_width + 10);
else
width = max_change;
for (i = 0; i < count; i++) {
const char *prefix = "";
char *name = data->files[i]->print_name;
uintmax_t added = data->files[i]->added;
uintmax_t deleted = data->files[i]->deleted;
int name_len;
if (!data->files[i]->is_renamed &&
(added + deleted == 0)) {
total_files--;
continue;
}
/*
* "scale" the filename
*/
len = name_width;
name_len = strlen(name);
if (name_width < name_len) {
char *slash;
prefix = "...";
len -= 3;
name += name_len - len;
slash = strchr(name, '/');
if (slash)
name = slash;
}
if (data->files[i]->is_binary) {
fprintf(options->file, "%s", line_prefix);
show_name(options->file, prefix, name, len);
fprintf(options->file, " Bin ");
fprintf(options->file, "%s%"PRIuMAX"%s",
del_c, deleted, reset);
fprintf(options->file, " -> ");
fprintf(options->file, "%s%"PRIuMAX"%s",
add_c, added, reset);
fprintf(options->file, " bytes");
fprintf(options->file, "\n");
continue;
}
else if (data->files[i]->is_unmerged) {
fprintf(options->file, "%s", line_prefix);
show_name(options->file, prefix, name, len);
fprintf(options->file, " Unmerged\n");
continue;
}
/*
* scale the add/delete
*/
add = added;
del = deleted;
adds += add;
dels += del;
if (width <= max_change) {
int total = add + del;
total = scale_linear(add + del, width, max_change);
if (total < 2 && add && del)
/* width >= 2 due to the sanity check */
total = 2;
if (add < del) {
add = scale_linear(add, width, max_change);
del = total - add;
} else {
del = scale_linear(del, width, max_change);
add = total - del;
}
}
fprintf(options->file, "%s", line_prefix);
show_name(options->file, prefix, name, len);
fprintf(options->file, "%5"PRIuMAX"%s", added + deleted,
added + deleted ? " " : "");
show_graph(options->file, '+', add, add_c, reset);
show_graph(options->file, '-', del, del_c, reset);
fprintf(options->file, "\n");
}
for (i = count; i < data->nr; i++) {
uintmax_t added = data->files[i]->added;
uintmax_t deleted = data->files[i]->deleted;
if (!data->files[i]->is_renamed &&
(added + deleted == 0)) {
total_files--;
continue;
}
adds += added;
dels += deleted;
if (!extra_shown)
fprintf(options->file, "%s ...\n", line_prefix);
extra_shown = 1;
}
fprintf(options->file, "%s", line_prefix);
print_stat_summary(options->file, total_files, adds, dels);
}
static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
{
int i, adds = 0, dels = 0, total_files = data->nr;
if (data->nr == 0)
return;
for (i = 0; i < data->nr; i++) {
if (!data->files[i]->is_binary &&
!data->files[i]->is_unmerged) {
int added = data->files[i]->added;
int deleted= data->files[i]->deleted;
if (!data->files[i]->is_renamed &&
(added + deleted == 0)) {
total_files--;
} else {
adds += added;
dels += deleted;
}
}
}
if (options->output_prefix) {
struct strbuf *msg = NULL;
msg = options->output_prefix(options,
options->output_prefix_data);
fprintf(options->file, "%s", msg->buf);
}
print_stat_summary(options->file, total_files, adds, dels);
}
static void show_numstat(struct diffstat_t *data, struct diff_options *options)
{
int i;
if (data->nr == 0)
return;
for (i = 0; i < data->nr; i++) {
struct diffstat_file *file = data->files[i];
if (options->output_prefix) {
struct strbuf *msg = NULL;
msg = options->output_prefix(options,
options->output_prefix_data);
fprintf(options->file, "%s", msg->buf);
}
if (file->is_binary)
fprintf(options->file, "-\t-\t");
else
fprintf(options->file,
"%"PRIuMAX"\t%"PRIuMAX"\t",
file->added, file->deleted);
if (options->line_termination) {
fill_print_name(file);
if (!file->is_renamed)
write_name_quoted(file->name, options->file,
options->line_termination);
else {
fputs(file->print_name, options->file);
putc(options->line_termination, options->file);
}
} else {
if (file->is_renamed) {
putc('\0', options->file);
write_name_quoted(file->from_name, options->file, '\0');
}
write_name_quoted(file->name, options->file, '\0');
}
}
}
struct dirstat_file {
const char *name;
unsigned long changed;
Add "--dirstat" for some directory statistics This adds a new form of overview diffstat output, doing something that I have occasionally ended up doing manually (and badly, because it's actually pretty nasty to do), and that I think is very useful for an project like the kernel that has a fairly deep and well-separated directory structure with semantic meaning. What I mean by that is that it's often interesting to see exactly which sub-directories are impacted by a patch, and to what degree - even if you don't perhaps care so much about the individual files themselves. What makes the concept more interesting is that the "impact" is often hierarchical: in the kernel, for example, something could either have a very localized impact to "fs/ext3/" and then it's interesting to see that such a patch changes mostly that subdirectory, but you could have another patch that changes some generic VFS-layer issue which affects _many_ subdirectories that are all under "fs/", but none - or perhaps just a couple of them - of the individual filesystems are interesting in themselves. So what commonly happens is that you may have big changes in a specific sub-subdirectory, but still also significant separate changes to the subdirectory leading up to that - maybe you have significant VFS-level changes, but *also* changes under that VFS layer in the NFS-specific directories, for example. In that case, you do want the low-level parts that are significant to show up, but then the insignificant ones should show up as under the more generic top-level directory. This patch shows all of that with "--dirstat". The output can be either something simple like commit 81772fe... Author: Thomas Gleixner <tglx@linutronix.de> Date: Sun Feb 10 23:57:36 2008 +0100 x86: remove over noisy debug printk pageattr-test.c contains a noisy debug printk that people reported. The condition under which it prints (randomly tapping into a mem_map[] hole and not being able to c_p_a() there) is valid behavior and not interesting to report. Remove it. Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> 100.0% arch/x86/mm/ or something much more complex like commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 15.3% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ where that latter example is an example of significant work in some individual fs/*/ subdirectories (like the patches to reiserfs accounting for 7.6% of the whole), but then discounting those individual filesystems, there's also 15.3% other "random" things that weren't worth reporting on their oen left over under fs/ in general (either in that directory itself, or in subdirectories of fs/ that didn't have enough changes to be reported individually). I'd like to stress that the "15.3% fs/" mentioned above is the stuff that is under fs/ but that was _not_ significant enough to report on its own. So the above does _not_ mean that 15.3% of the work was under fs/ per se, because that 15.3% does *not* include the already-reported 7.6% of afs, 7.6% of fuse etc. If you want to enable "cumulative" directory statistics, you can use the "--cumulative" flag, which adds up percentages recursively even when they have been already reported for a sub-directory. That cumulative output is disabled if *all* of the changes in one subdirectory come from a deeper subdirectory, to avoid repeating subdirectories all the way to the root. For an example of the cumulative reporting, the above commit becomes commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 61.5% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ in which the commit percentages now obviously add up to much more than 100%: now the changes that were already reported for the sub-directories under fs/ are then cumulatively included in the whole percentage of fs/ (ie now shows 61.5% as opposed to the 15.3% without the cumulative reporting). The default reporting limit has been arbitrarily set at 3%, which seems to be a pretty good cut-off, but you can specify the cut-off manually by giving it as an option parameter (eg "--dirstat=5" makes the cut-off be at 5% instead) NOTE! The percentages are purely about the total lines added and removed, not anything smarter (or dumber) than that. Also note that you should not generally expect things to add up to 100%: not only does it round down, we don't report leftover scraps (they add up to the top-level change count, but we don't even bother reporting that, it only reports subdirectories). Quite frankly, as a top-level manager this is really convenient for me, but it's going to be very boring for git itself since there are few subdirectories. Also, don't expect things to make tons of sense if you combine this with "-M" and there are cross-directory renames etc. But even for git itself, you can get some fun statistics. Try out git log --dirstat and see the occasional mentions of things like Documentation/, git-gui/, gitweb/ and gitk-git/. Or try out something like git diff --dirstat v1.5.0..v1.5.4 which does kind of git an overview that shows *something*. But in general, the output is more exciting for big projects with deeper structure, and doing a git diff --dirstat v2.6.24..v2.6.25-rc1 on the kernel is what I actually wrote this for! Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 21:26:31 +00:00
};
struct dirstat_dir {
struct dirstat_file *files;
int alloc, nr, permille, cumulative;
};
static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
unsigned long changed, const char *base, int baselen)
Add "--dirstat" for some directory statistics This adds a new form of overview diffstat output, doing something that I have occasionally ended up doing manually (and badly, because it's actually pretty nasty to do), and that I think is very useful for an project like the kernel that has a fairly deep and well-separated directory structure with semantic meaning. What I mean by that is that it's often interesting to see exactly which sub-directories are impacted by a patch, and to what degree - even if you don't perhaps care so much about the individual files themselves. What makes the concept more interesting is that the "impact" is often hierarchical: in the kernel, for example, something could either have a very localized impact to "fs/ext3/" and then it's interesting to see that such a patch changes mostly that subdirectory, but you could have another patch that changes some generic VFS-layer issue which affects _many_ subdirectories that are all under "fs/", but none - or perhaps just a couple of them - of the individual filesystems are interesting in themselves. So what commonly happens is that you may have big changes in a specific sub-subdirectory, but still also significant separate changes to the subdirectory leading up to that - maybe you have significant VFS-level changes, but *also* changes under that VFS layer in the NFS-specific directories, for example. In that case, you do want the low-level parts that are significant to show up, but then the insignificant ones should show up as under the more generic top-level directory. This patch shows all of that with "--dirstat". The output can be either something simple like commit 81772fe... Author: Thomas Gleixner <tglx@linutronix.de> Date: Sun Feb 10 23:57:36 2008 +0100 x86: remove over noisy debug printk pageattr-test.c contains a noisy debug printk that people reported. The condition under which it prints (randomly tapping into a mem_map[] hole and not being able to c_p_a() there) is valid behavior and not interesting to report. Remove it. Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> 100.0% arch/x86/mm/ or something much more complex like commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 15.3% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ where that latter example is an example of significant work in some individual fs/*/ subdirectories (like the patches to reiserfs accounting for 7.6% of the whole), but then discounting those individual filesystems, there's also 15.3% other "random" things that weren't worth reporting on their oen left over under fs/ in general (either in that directory itself, or in subdirectories of fs/ that didn't have enough changes to be reported individually). I'd like to stress that the "15.3% fs/" mentioned above is the stuff that is under fs/ but that was _not_ significant enough to report on its own. So the above does _not_ mean that 15.3% of the work was under fs/ per se, because that 15.3% does *not* include the already-reported 7.6% of afs, 7.6% of fuse etc. If you want to enable "cumulative" directory statistics, you can use the "--cumulative" flag, which adds up percentages recursively even when they have been already reported for a sub-directory. That cumulative output is disabled if *all* of the changes in one subdirectory come from a deeper subdirectory, to avoid repeating subdirectories all the way to the root. For an example of the cumulative reporting, the above commit becomes commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 61.5% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ in which the commit percentages now obviously add up to much more than 100%: now the changes that were already reported for the sub-directories under fs/ are then cumulatively included in the whole percentage of fs/ (ie now shows 61.5% as opposed to the 15.3% without the cumulative reporting). The default reporting limit has been arbitrarily set at 3%, which seems to be a pretty good cut-off, but you can specify the cut-off manually by giving it as an option parameter (eg "--dirstat=5" makes the cut-off be at 5% instead) NOTE! The percentages are purely about the total lines added and removed, not anything smarter (or dumber) than that. Also note that you should not generally expect things to add up to 100%: not only does it round down, we don't report leftover scraps (they add up to the top-level change count, but we don't even bother reporting that, it only reports subdirectories). Quite frankly, as a top-level manager this is really convenient for me, but it's going to be very boring for git itself since there are few subdirectories. Also, don't expect things to make tons of sense if you combine this with "-M" and there are cross-directory renames etc. But even for git itself, you can get some fun statistics. Try out git log --dirstat and see the occasional mentions of things like Documentation/, git-gui/, gitweb/ and gitk-git/. Or try out something like git diff --dirstat v1.5.0..v1.5.4 which does kind of git an overview that shows *something*. But in general, the output is more exciting for big projects with deeper structure, and doing a git diff --dirstat v2.6.24..v2.6.25-rc1 on the kernel is what I actually wrote this for! Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 21:26:31 +00:00
{
unsigned long this_dir = 0;
unsigned int sources = 0;
const char *line_prefix = "";
struct strbuf *msg = NULL;
if (opt->output_prefix) {
msg = opt->output_prefix(opt, opt->output_prefix_data);
line_prefix = msg->buf;
}
Add "--dirstat" for some directory statistics This adds a new form of overview diffstat output, doing something that I have occasionally ended up doing manually (and badly, because it's actually pretty nasty to do), and that I think is very useful for an project like the kernel that has a fairly deep and well-separated directory structure with semantic meaning. What I mean by that is that it's often interesting to see exactly which sub-directories are impacted by a patch, and to what degree - even if you don't perhaps care so much about the individual files themselves. What makes the concept more interesting is that the "impact" is often hierarchical: in the kernel, for example, something could either have a very localized impact to "fs/ext3/" and then it's interesting to see that such a patch changes mostly that subdirectory, but you could have another patch that changes some generic VFS-layer issue which affects _many_ subdirectories that are all under "fs/", but none - or perhaps just a couple of them - of the individual filesystems are interesting in themselves. So what commonly happens is that you may have big changes in a specific sub-subdirectory, but still also significant separate changes to the subdirectory leading up to that - maybe you have significant VFS-level changes, but *also* changes under that VFS layer in the NFS-specific directories, for example. In that case, you do want the low-level parts that are significant to show up, but then the insignificant ones should show up as under the more generic top-level directory. This patch shows all of that with "--dirstat". The output can be either something simple like commit 81772fe... Author: Thomas Gleixner <tglx@linutronix.de> Date: Sun Feb 10 23:57:36 2008 +0100 x86: remove over noisy debug printk pageattr-test.c contains a noisy debug printk that people reported. The condition under which it prints (randomly tapping into a mem_map[] hole and not being able to c_p_a() there) is valid behavior and not interesting to report. Remove it. Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> 100.0% arch/x86/mm/ or something much more complex like commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 15.3% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ where that latter example is an example of significant work in some individual fs/*/ subdirectories (like the patches to reiserfs accounting for 7.6% of the whole), but then discounting those individual filesystems, there's also 15.3% other "random" things that weren't worth reporting on their oen left over under fs/ in general (either in that directory itself, or in subdirectories of fs/ that didn't have enough changes to be reported individually). I'd like to stress that the "15.3% fs/" mentioned above is the stuff that is under fs/ but that was _not_ significant enough to report on its own. So the above does _not_ mean that 15.3% of the work was under fs/ per se, because that 15.3% does *not* include the already-reported 7.6% of afs, 7.6% of fuse etc. If you want to enable "cumulative" directory statistics, you can use the "--cumulative" flag, which adds up percentages recursively even when they have been already reported for a sub-directory. That cumulative output is disabled if *all* of the changes in one subdirectory come from a deeper subdirectory, to avoid repeating subdirectories all the way to the root. For an example of the cumulative reporting, the above commit becomes commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 61.5% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ in which the commit percentages now obviously add up to much more than 100%: now the changes that were already reported for the sub-directories under fs/ are then cumulatively included in the whole percentage of fs/ (ie now shows 61.5% as opposed to the 15.3% without the cumulative reporting). The default reporting limit has been arbitrarily set at 3%, which seems to be a pretty good cut-off, but you can specify the cut-off manually by giving it as an option parameter (eg "--dirstat=5" makes the cut-off be at 5% instead) NOTE! The percentages are purely about the total lines added and removed, not anything smarter (or dumber) than that. Also note that you should not generally expect things to add up to 100%: not only does it round down, we don't report leftover scraps (they add up to the top-level change count, but we don't even bother reporting that, it only reports subdirectories). Quite frankly, as a top-level manager this is really convenient for me, but it's going to be very boring for git itself since there are few subdirectories. Also, don't expect things to make tons of sense if you combine this with "-M" and there are cross-directory renames etc. But even for git itself, you can get some fun statistics. Try out git log --dirstat and see the occasional mentions of things like Documentation/, git-gui/, gitweb/ and gitk-git/. Or try out something like git diff --dirstat v1.5.0..v1.5.4 which does kind of git an overview that shows *something*. But in general, the output is more exciting for big projects with deeper structure, and doing a git diff --dirstat v2.6.24..v2.6.25-rc1 on the kernel is what I actually wrote this for! Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 21:26:31 +00:00
while (dir->nr) {
struct dirstat_file *f = dir->files;
Add "--dirstat" for some directory statistics This adds a new form of overview diffstat output, doing something that I have occasionally ended up doing manually (and badly, because it's actually pretty nasty to do), and that I think is very useful for an project like the kernel that has a fairly deep and well-separated directory structure with semantic meaning. What I mean by that is that it's often interesting to see exactly which sub-directories are impacted by a patch, and to what degree - even if you don't perhaps care so much about the individual files themselves. What makes the concept more interesting is that the "impact" is often hierarchical: in the kernel, for example, something could either have a very localized impact to "fs/ext3/" and then it's interesting to see that such a patch changes mostly that subdirectory, but you could have another patch that changes some generic VFS-layer issue which affects _many_ subdirectories that are all under "fs/", but none - or perhaps just a couple of them - of the individual filesystems are interesting in themselves. So what commonly happens is that you may have big changes in a specific sub-subdirectory, but still also significant separate changes to the subdirectory leading up to that - maybe you have significant VFS-level changes, but *also* changes under that VFS layer in the NFS-specific directories, for example. In that case, you do want the low-level parts that are significant to show up, but then the insignificant ones should show up as under the more generic top-level directory. This patch shows all of that with "--dirstat". The output can be either something simple like commit 81772fe... Author: Thomas Gleixner <tglx@linutronix.de> Date: Sun Feb 10 23:57:36 2008 +0100 x86: remove over noisy debug printk pageattr-test.c contains a noisy debug printk that people reported. The condition under which it prints (randomly tapping into a mem_map[] hole and not being able to c_p_a() there) is valid behavior and not interesting to report. Remove it. Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> 100.0% arch/x86/mm/ or something much more complex like commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 15.3% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ where that latter example is an example of significant work in some individual fs/*/ subdirectories (like the patches to reiserfs accounting for 7.6% of the whole), but then discounting those individual filesystems, there's also 15.3% other "random" things that weren't worth reporting on their oen left over under fs/ in general (either in that directory itself, or in subdirectories of fs/ that didn't have enough changes to be reported individually). I'd like to stress that the "15.3% fs/" mentioned above is the stuff that is under fs/ but that was _not_ significant enough to report on its own. So the above does _not_ mean that 15.3% of the work was under fs/ per se, because that 15.3% does *not* include the already-reported 7.6% of afs, 7.6% of fuse etc. If you want to enable "cumulative" directory statistics, you can use the "--cumulative" flag, which adds up percentages recursively even when they have been already reported for a sub-directory. That cumulative output is disabled if *all* of the changes in one subdirectory come from a deeper subdirectory, to avoid repeating subdirectories all the way to the root. For an example of the cumulative reporting, the above commit becomes commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 61.5% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ in which the commit percentages now obviously add up to much more than 100%: now the changes that were already reported for the sub-directories under fs/ are then cumulatively included in the whole percentage of fs/ (ie now shows 61.5% as opposed to the 15.3% without the cumulative reporting). The default reporting limit has been arbitrarily set at 3%, which seems to be a pretty good cut-off, but you can specify the cut-off manually by giving it as an option parameter (eg "--dirstat=5" makes the cut-off be at 5% instead) NOTE! The percentages are purely about the total lines added and removed, not anything smarter (or dumber) than that. Also note that you should not generally expect things to add up to 100%: not only does it round down, we don't report leftover scraps (they add up to the top-level change count, but we don't even bother reporting that, it only reports subdirectories). Quite frankly, as a top-level manager this is really convenient for me, but it's going to be very boring for git itself since there are few subdirectories. Also, don't expect things to make tons of sense if you combine this with "-M" and there are cross-directory renames etc. But even for git itself, you can get some fun statistics. Try out git log --dirstat and see the occasional mentions of things like Documentation/, git-gui/, gitweb/ and gitk-git/. Or try out something like git diff --dirstat v1.5.0..v1.5.4 which does kind of git an overview that shows *something*. But in general, the output is more exciting for big projects with deeper structure, and doing a git diff --dirstat v2.6.24..v2.6.25-rc1 on the kernel is what I actually wrote this for! Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 21:26:31 +00:00
int namelen = strlen(f->name);
unsigned long this;
char *slash;
if (namelen < baselen)
break;
if (memcmp(f->name, base, baselen))
break;
slash = strchr(f->name + baselen, '/');
if (slash) {
int newbaselen = slash + 1 - f->name;
this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
Add "--dirstat" for some directory statistics This adds a new form of overview diffstat output, doing something that I have occasionally ended up doing manually (and badly, because it's actually pretty nasty to do), and that I think is very useful for an project like the kernel that has a fairly deep and well-separated directory structure with semantic meaning. What I mean by that is that it's often interesting to see exactly which sub-directories are impacted by a patch, and to what degree - even if you don't perhaps care so much about the individual files themselves. What makes the concept more interesting is that the "impact" is often hierarchical: in the kernel, for example, something could either have a very localized impact to "fs/ext3/" and then it's interesting to see that such a patch changes mostly that subdirectory, but you could have another patch that changes some generic VFS-layer issue which affects _many_ subdirectories that are all under "fs/", but none - or perhaps just a couple of them - of the individual filesystems are interesting in themselves. So what commonly happens is that you may have big changes in a specific sub-subdirectory, but still also significant separate changes to the subdirectory leading up to that - maybe you have significant VFS-level changes, but *also* changes under that VFS layer in the NFS-specific directories, for example. In that case, you do want the low-level parts that are significant to show up, but then the insignificant ones should show up as under the more generic top-level directory. This patch shows all of that with "--dirstat". The output can be either something simple like commit 81772fe... Author: Thomas Gleixner <tglx@linutronix.de> Date: Sun Feb 10 23:57:36 2008 +0100 x86: remove over noisy debug printk pageattr-test.c contains a noisy debug printk that people reported. The condition under which it prints (randomly tapping into a mem_map[] hole and not being able to c_p_a() there) is valid behavior and not interesting to report. Remove it. Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> 100.0% arch/x86/mm/ or something much more complex like commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 15.3% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ where that latter example is an example of significant work in some individual fs/*/ subdirectories (like the patches to reiserfs accounting for 7.6% of the whole), but then discounting those individual filesystems, there's also 15.3% other "random" things that weren't worth reporting on their oen left over under fs/ in general (either in that directory itself, or in subdirectories of fs/ that didn't have enough changes to be reported individually). I'd like to stress that the "15.3% fs/" mentioned above is the stuff that is under fs/ but that was _not_ significant enough to report on its own. So the above does _not_ mean that 15.3% of the work was under fs/ per se, because that 15.3% does *not* include the already-reported 7.6% of afs, 7.6% of fuse etc. If you want to enable "cumulative" directory statistics, you can use the "--cumulative" flag, which adds up percentages recursively even when they have been already reported for a sub-directory. That cumulative output is disabled if *all* of the changes in one subdirectory come from a deeper subdirectory, to avoid repeating subdirectories all the way to the root. For an example of the cumulative reporting, the above commit becomes commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 61.5% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ in which the commit percentages now obviously add up to much more than 100%: now the changes that were already reported for the sub-directories under fs/ are then cumulatively included in the whole percentage of fs/ (ie now shows 61.5% as opposed to the 15.3% without the cumulative reporting). The default reporting limit has been arbitrarily set at 3%, which seems to be a pretty good cut-off, but you can specify the cut-off manually by giving it as an option parameter (eg "--dirstat=5" makes the cut-off be at 5% instead) NOTE! The percentages are purely about the total lines added and removed, not anything smarter (or dumber) than that. Also note that you should not generally expect things to add up to 100%: not only does it round down, we don't report leftover scraps (they add up to the top-level change count, but we don't even bother reporting that, it only reports subdirectories). Quite frankly, as a top-level manager this is really convenient for me, but it's going to be very boring for git itself since there are few subdirectories. Also, don't expect things to make tons of sense if you combine this with "-M" and there are cross-directory renames etc. But even for git itself, you can get some fun statistics. Try out git log --dirstat and see the occasional mentions of things like Documentation/, git-gui/, gitweb/ and gitk-git/. Or try out something like git diff --dirstat v1.5.0..v1.5.4 which does kind of git an overview that shows *something*. But in general, the output is more exciting for big projects with deeper structure, and doing a git diff --dirstat v2.6.24..v2.6.25-rc1 on the kernel is what I actually wrote this for! Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 21:26:31 +00:00
sources++;
} else {
this = f->changed;
Add "--dirstat" for some directory statistics This adds a new form of overview diffstat output, doing something that I have occasionally ended up doing manually (and badly, because it's actually pretty nasty to do), and that I think is very useful for an project like the kernel that has a fairly deep and well-separated directory structure with semantic meaning. What I mean by that is that it's often interesting to see exactly which sub-directories are impacted by a patch, and to what degree - even if you don't perhaps care so much about the individual files themselves. What makes the concept more interesting is that the "impact" is often hierarchical: in the kernel, for example, something could either have a very localized impact to "fs/ext3/" and then it's interesting to see that such a patch changes mostly that subdirectory, but you could have another patch that changes some generic VFS-layer issue which affects _many_ subdirectories that are all under "fs/", but none - or perhaps just a couple of them - of the individual filesystems are interesting in themselves. So what commonly happens is that you may have big changes in a specific sub-subdirectory, but still also significant separate changes to the subdirectory leading up to that - maybe you have significant VFS-level changes, but *also* changes under that VFS layer in the NFS-specific directories, for example. In that case, you do want the low-level parts that are significant to show up, but then the insignificant ones should show up as under the more generic top-level directory. This patch shows all of that with "--dirstat". The output can be either something simple like commit 81772fe... Author: Thomas Gleixner <tglx@linutronix.de> Date: Sun Feb 10 23:57:36 2008 +0100 x86: remove over noisy debug printk pageattr-test.c contains a noisy debug printk that people reported. The condition under which it prints (randomly tapping into a mem_map[] hole and not being able to c_p_a() there) is valid behavior and not interesting to report. Remove it. Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> 100.0% arch/x86/mm/ or something much more complex like commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 15.3% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ where that latter example is an example of significant work in some individual fs/*/ subdirectories (like the patches to reiserfs accounting for 7.6% of the whole), but then discounting those individual filesystems, there's also 15.3% other "random" things that weren't worth reporting on their oen left over under fs/ in general (either in that directory itself, or in subdirectories of fs/ that didn't have enough changes to be reported individually). I'd like to stress that the "15.3% fs/" mentioned above is the stuff that is under fs/ but that was _not_ significant enough to report on its own. So the above does _not_ mean that 15.3% of the work was under fs/ per se, because that 15.3% does *not* include the already-reported 7.6% of afs, 7.6% of fuse etc. If you want to enable "cumulative" directory statistics, you can use the "--cumulative" flag, which adds up percentages recursively even when they have been already reported for a sub-directory. That cumulative output is disabled if *all* of the changes in one subdirectory come from a deeper subdirectory, to avoid repeating subdirectories all the way to the root. For an example of the cumulative reporting, the above commit becomes commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 61.5% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ in which the commit percentages now obviously add up to much more than 100%: now the changes that were already reported for the sub-directories under fs/ are then cumulatively included in the whole percentage of fs/ (ie now shows 61.5% as opposed to the 15.3% without the cumulative reporting). The default reporting limit has been arbitrarily set at 3%, which seems to be a pretty good cut-off, but you can specify the cut-off manually by giving it as an option parameter (eg "--dirstat=5" makes the cut-off be at 5% instead) NOTE! The percentages are purely about the total lines added and removed, not anything smarter (or dumber) than that. Also note that you should not generally expect things to add up to 100%: not only does it round down, we don't report leftover scraps (they add up to the top-level change count, but we don't even bother reporting that, it only reports subdirectories). Quite frankly, as a top-level manager this is really convenient for me, but it's going to be very boring for git itself since there are few subdirectories. Also, don't expect things to make tons of sense if you combine this with "-M" and there are cross-directory renames etc. But even for git itself, you can get some fun statistics. Try out git log --dirstat and see the occasional mentions of things like Documentation/, git-gui/, gitweb/ and gitk-git/. Or try out something like git diff --dirstat v1.5.0..v1.5.4 which does kind of git an overview that shows *something*. But in general, the output is more exciting for big projects with deeper structure, and doing a git diff --dirstat v2.6.24..v2.6.25-rc1 on the kernel is what I actually wrote this for! Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 21:26:31 +00:00
dir->files++;
dir->nr--;
sources += 2;
}
this_dir += this;
}
/*
* We don't report dirstat's for
* - the top level
* - or cases where everything came from a single directory
* under this directory (sources == 1).
*/
if (baselen && sources != 1) {
if (this_dir) {
int permille = this_dir * 1000 / changed;
if (permille >= dir->permille) {
fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
permille / 10, permille % 10, baselen, base);
Add "--dirstat" for some directory statistics This adds a new form of overview diffstat output, doing something that I have occasionally ended up doing manually (and badly, because it's actually pretty nasty to do), and that I think is very useful for an project like the kernel that has a fairly deep and well-separated directory structure with semantic meaning. What I mean by that is that it's often interesting to see exactly which sub-directories are impacted by a patch, and to what degree - even if you don't perhaps care so much about the individual files themselves. What makes the concept more interesting is that the "impact" is often hierarchical: in the kernel, for example, something could either have a very localized impact to "fs/ext3/" and then it's interesting to see that such a patch changes mostly that subdirectory, but you could have another patch that changes some generic VFS-layer issue which affects _many_ subdirectories that are all under "fs/", but none - or perhaps just a couple of them - of the individual filesystems are interesting in themselves. So what commonly happens is that you may have big changes in a specific sub-subdirectory, but still also significant separate changes to the subdirectory leading up to that - maybe you have significant VFS-level changes, but *also* changes under that VFS layer in the NFS-specific directories, for example. In that case, you do want the low-level parts that are significant to show up, but then the insignificant ones should show up as under the more generic top-level directory. This patch shows all of that with "--dirstat". The output can be either something simple like commit 81772fe... Author: Thomas Gleixner <tglx@linutronix.de> Date: Sun Feb 10 23:57:36 2008 +0100 x86: remove over noisy debug printk pageattr-test.c contains a noisy debug printk that people reported. The condition under which it prints (randomly tapping into a mem_map[] hole and not being able to c_p_a() there) is valid behavior and not interesting to report. Remove it. Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> 100.0% arch/x86/mm/ or something much more complex like commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 15.3% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ where that latter example is an example of significant work in some individual fs/*/ subdirectories (like the patches to reiserfs accounting for 7.6% of the whole), but then discounting those individual filesystems, there's also 15.3% other "random" things that weren't worth reporting on their oen left over under fs/ in general (either in that directory itself, or in subdirectories of fs/ that didn't have enough changes to be reported individually). I'd like to stress that the "15.3% fs/" mentioned above is the stuff that is under fs/ but that was _not_ significant enough to report on its own. So the above does _not_ mean that 15.3% of the work was under fs/ per se, because that 15.3% does *not* include the already-reported 7.6% of afs, 7.6% of fuse etc. If you want to enable "cumulative" directory statistics, you can use the "--cumulative" flag, which adds up percentages recursively even when they have been already reported for a sub-directory. That cumulative output is disabled if *all* of the changes in one subdirectory come from a deeper subdirectory, to avoid repeating subdirectories all the way to the root. For an example of the cumulative reporting, the above commit becomes commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 61.5% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ in which the commit percentages now obviously add up to much more than 100%: now the changes that were already reported for the sub-directories under fs/ are then cumulatively included in the whole percentage of fs/ (ie now shows 61.5% as opposed to the 15.3% without the cumulative reporting). The default reporting limit has been arbitrarily set at 3%, which seems to be a pretty good cut-off, but you can specify the cut-off manually by giving it as an option parameter (eg "--dirstat=5" makes the cut-off be at 5% instead) NOTE! The percentages are purely about the total lines added and removed, not anything smarter (or dumber) than that. Also note that you should not generally expect things to add up to 100%: not only does it round down, we don't report leftover scraps (they add up to the top-level change count, but we don't even bother reporting that, it only reports subdirectories). Quite frankly, as a top-level manager this is really convenient for me, but it's going to be very boring for git itself since there are few subdirectories. Also, don't expect things to make tons of sense if you combine this with "-M" and there are cross-directory renames etc. But even for git itself, you can get some fun statistics. Try out git log --dirstat and see the occasional mentions of things like Documentation/, git-gui/, gitweb/ and gitk-git/. Or try out something like git diff --dirstat v1.5.0..v1.5.4 which does kind of git an overview that shows *something*. But in general, the output is more exciting for big projects with deeper structure, and doing a git diff --dirstat v2.6.24..v2.6.25-rc1 on the kernel is what I actually wrote this for! Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 21:26:31 +00:00
if (!dir->cumulative)
return 0;
}
}
}
return this_dir;
}
Fix '--dirstat' with cross-directory renaming The dirstat code depends on the fact that we always generate diffs with the names sorted, since it then just does a single-pass walk-over of the sorted list of names and how many changes there were. The sorting means that all files are nicely grouped by directory. That all works fine. Except when we have rename detection, and suddenly the nicely sorted list of pathnames isn't all that sorted at all. And now the single-pass dirstat walk gets all confused, and you can get results like this: [torvalds@nehalem linux]$ git diff --dirstat=2 -M v2.6.27-rc4..v2.6.27-rc5 3.0% arch/powerpc/configs/ 6.8% arch/arm/configs/ 2.7% arch/powerpc/configs/ 4.2% arch/arm/configs/ 5.6% arch/powerpc/configs/ 8.4% arch/arm/configs/ 5.5% arch/powerpc/configs/ 23.3% arch/arm/configs/ 8.6% arch/powerpc/configs/ 4.0% arch/ 4.4% drivers/usb/musb/ 4.0% drivers/watchdog/ 7.6% drivers/ 3.5% fs/ The trivial fix is to add a sorting pass, fixing it to: [torvalds@nehalem linux]$ git diff --dirstat=2 -M v2.6.27-rc4..v2.6.27-rc5 43.0% arch/arm/configs/ 25.5% arch/powerpc/configs/ 5.3% arch/ 4.4% drivers/usb/musb/ 4.0% drivers/watchdog/ 7.6% drivers/ 3.5% fs/ Spot the difference. In case anybody wonders: it's because of a ton of renames from {include/asm-blackfin => arch/blackfin/include/asm} that just totally messed up the file ordering in between arch/arm and arch/powerpc. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-08-28 23:19:08 +00:00
static int dirstat_compare(const void *_a, const void *_b)
{
const struct dirstat_file *a = _a;
const struct dirstat_file *b = _b;
return strcmp(a->name, b->name);
}
static void show_dirstat(struct diff_options *options)
Add "--dirstat" for some directory statistics This adds a new form of overview diffstat output, doing something that I have occasionally ended up doing manually (and badly, because it's actually pretty nasty to do), and that I think is very useful for an project like the kernel that has a fairly deep and well-separated directory structure with semantic meaning. What I mean by that is that it's often interesting to see exactly which sub-directories are impacted by a patch, and to what degree - even if you don't perhaps care so much about the individual files themselves. What makes the concept more interesting is that the "impact" is often hierarchical: in the kernel, for example, something could either have a very localized impact to "fs/ext3/" and then it's interesting to see that such a patch changes mostly that subdirectory, but you could have another patch that changes some generic VFS-layer issue which affects _many_ subdirectories that are all under "fs/", but none - or perhaps just a couple of them - of the individual filesystems are interesting in themselves. So what commonly happens is that you may have big changes in a specific sub-subdirectory, but still also significant separate changes to the subdirectory leading up to that - maybe you have significant VFS-level changes, but *also* changes under that VFS layer in the NFS-specific directories, for example. In that case, you do want the low-level parts that are significant to show up, but then the insignificant ones should show up as under the more generic top-level directory. This patch shows all of that with "--dirstat". The output can be either something simple like commit 81772fe... Author: Thomas Gleixner <tglx@linutronix.de> Date: Sun Feb 10 23:57:36 2008 +0100 x86: remove over noisy debug printk pageattr-test.c contains a noisy debug printk that people reported. The condition under which it prints (randomly tapping into a mem_map[] hole and not being able to c_p_a() there) is valid behavior and not interesting to report. Remove it. Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> 100.0% arch/x86/mm/ or something much more complex like commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 15.3% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ where that latter example is an example of significant work in some individual fs/*/ subdirectories (like the patches to reiserfs accounting for 7.6% of the whole), but then discounting those individual filesystems, there's also 15.3% other "random" things that weren't worth reporting on their oen left over under fs/ in general (either in that directory itself, or in subdirectories of fs/ that didn't have enough changes to be reported individually). I'd like to stress that the "15.3% fs/" mentioned above is the stuff that is under fs/ but that was _not_ significant enough to report on its own. So the above does _not_ mean that 15.3% of the work was under fs/ per se, because that 15.3% does *not* include the already-reported 7.6% of afs, 7.6% of fuse etc. If you want to enable "cumulative" directory statistics, you can use the "--cumulative" flag, which adds up percentages recursively even when they have been already reported for a sub-directory. That cumulative output is disabled if *all* of the changes in one subdirectory come from a deeper subdirectory, to avoid repeating subdirectories all the way to the root. For an example of the cumulative reporting, the above commit becomes commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 61.5% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ in which the commit percentages now obviously add up to much more than 100%: now the changes that were already reported for the sub-directories under fs/ are then cumulatively included in the whole percentage of fs/ (ie now shows 61.5% as opposed to the 15.3% without the cumulative reporting). The default reporting limit has been arbitrarily set at 3%, which seems to be a pretty good cut-off, but you can specify the cut-off manually by giving it as an option parameter (eg "--dirstat=5" makes the cut-off be at 5% instead) NOTE! The percentages are purely about the total lines added and removed, not anything smarter (or dumber) than that. Also note that you should not generally expect things to add up to 100%: not only does it round down, we don't report leftover scraps (they add up to the top-level change count, but we don't even bother reporting that, it only reports subdirectories). Quite frankly, as a top-level manager this is really convenient for me, but it's going to be very boring for git itself since there are few subdirectories. Also, don't expect things to make tons of sense if you combine this with "-M" and there are cross-directory renames etc. But even for git itself, you can get some fun statistics. Try out git log --dirstat and see the occasional mentions of things like Documentation/, git-gui/, gitweb/ and gitk-git/. Or try out something like git diff --dirstat v1.5.0..v1.5.4 which does kind of git an overview that shows *something*. But in general, the output is more exciting for big projects with deeper structure, and doing a git diff --dirstat v2.6.24..v2.6.25-rc1 on the kernel is what I actually wrote this for! Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 21:26:31 +00:00
{
int i;
unsigned long changed;
struct dirstat_dir dir;
struct diff_queue_struct *q = &diff_queued_diff;
dir.files = NULL;
dir.alloc = 0;
dir.nr = 0;
dir.permille = options->dirstat_permille;
dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
Add "--dirstat" for some directory statistics This adds a new form of overview diffstat output, doing something that I have occasionally ended up doing manually (and badly, because it's actually pretty nasty to do), and that I think is very useful for an project like the kernel that has a fairly deep and well-separated directory structure with semantic meaning. What I mean by that is that it's often interesting to see exactly which sub-directories are impacted by a patch, and to what degree - even if you don't perhaps care so much about the individual files themselves. What makes the concept more interesting is that the "impact" is often hierarchical: in the kernel, for example, something could either have a very localized impact to "fs/ext3/" and then it's interesting to see that such a patch changes mostly that subdirectory, but you could have another patch that changes some generic VFS-layer issue which affects _many_ subdirectories that are all under "fs/", but none - or perhaps just a couple of them - of the individual filesystems are interesting in themselves. So what commonly happens is that you may have big changes in a specific sub-subdirectory, but still also significant separate changes to the subdirectory leading up to that - maybe you have significant VFS-level changes, but *also* changes under that VFS layer in the NFS-specific directories, for example. In that case, you do want the low-level parts that are significant to show up, but then the insignificant ones should show up as under the more generic top-level directory. This patch shows all of that with "--dirstat". The output can be either something simple like commit 81772fe... Author: Thomas Gleixner <tglx@linutronix.de> Date: Sun Feb 10 23:57:36 2008 +0100 x86: remove over noisy debug printk pageattr-test.c contains a noisy debug printk that people reported. The condition under which it prints (randomly tapping into a mem_map[] hole and not being able to c_p_a() there) is valid behavior and not interesting to report. Remove it. Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> 100.0% arch/x86/mm/ or something much more complex like commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 15.3% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ where that latter example is an example of significant work in some individual fs/*/ subdirectories (like the patches to reiserfs accounting for 7.6% of the whole), but then discounting those individual filesystems, there's also 15.3% other "random" things that weren't worth reporting on their oen left over under fs/ in general (either in that directory itself, or in subdirectories of fs/ that didn't have enough changes to be reported individually). I'd like to stress that the "15.3% fs/" mentioned above is the stuff that is under fs/ but that was _not_ significant enough to report on its own. So the above does _not_ mean that 15.3% of the work was under fs/ per se, because that 15.3% does *not* include the already-reported 7.6% of afs, 7.6% of fuse etc. If you want to enable "cumulative" directory statistics, you can use the "--cumulative" flag, which adds up percentages recursively even when they have been already reported for a sub-directory. That cumulative output is disabled if *all* of the changes in one subdirectory come from a deeper subdirectory, to avoid repeating subdirectories all the way to the root. For an example of the cumulative reporting, the above commit becomes commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 61.5% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ in which the commit percentages now obviously add up to much more than 100%: now the changes that were already reported for the sub-directories under fs/ are then cumulatively included in the whole percentage of fs/ (ie now shows 61.5% as opposed to the 15.3% without the cumulative reporting). The default reporting limit has been arbitrarily set at 3%, which seems to be a pretty good cut-off, but you can specify the cut-off manually by giving it as an option parameter (eg "--dirstat=5" makes the cut-off be at 5% instead) NOTE! The percentages are purely about the total lines added and removed, not anything smarter (or dumber) than that. Also note that you should not generally expect things to add up to 100%: not only does it round down, we don't report leftover scraps (they add up to the top-level change count, but we don't even bother reporting that, it only reports subdirectories). Quite frankly, as a top-level manager this is really convenient for me, but it's going to be very boring for git itself since there are few subdirectories. Also, don't expect things to make tons of sense if you combine this with "-M" and there are cross-directory renames etc. But even for git itself, you can get some fun statistics. Try out git log --dirstat and see the occasional mentions of things like Documentation/, git-gui/, gitweb/ and gitk-git/. Or try out something like git diff --dirstat v1.5.0..v1.5.4 which does kind of git an overview that shows *something*. But in general, the output is more exciting for big projects with deeper structure, and doing a git diff --dirstat v2.6.24..v2.6.25-rc1 on the kernel is what I actually wrote this for! Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 21:26:31 +00:00
changed = 0;
for (i = 0; i < q->nr; i++) {
struct diff_filepair *p = q->queue[i];
const char *name;
unsigned long copied, added, damage;
int content_changed;
name = p->two->path ? p->two->path : p->one->path;
if (p->one->sha1_valid && p->two->sha1_valid)
content_changed = hashcmp(p->one->sha1, p->two->sha1);
else
content_changed = 1;
if (!content_changed) {
/*
* The SHA1 has not changed, so pre-/post-content is
* identical. We can therefore skip looking at the
* file contents altogether.
*/
damage = 0;
goto found_damage;
}
if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE)) {
/*
* In --dirstat-by-file mode, we don't really need to
* look at the actual file contents at all.
* The fact that the SHA1 changed is enough for us to
* add this file to the list of results
* (with each file contributing equal damage).
*/
damage = 1;
goto found_damage;
}
if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
diff_populate_filespec(p->one, 0);
diff_populate_filespec(p->two, 0);
diffcore_count_changes(p->one, p->two, NULL, NULL, 0,
&copied, &added);
diff_free_filespec_data(p->one);
diff_free_filespec_data(p->two);
} else if (DIFF_FILE_VALID(p->one)) {
diff_populate_filespec(p->one, 1);
copied = added = 0;
diff_free_filespec_data(p->one);
} else if (DIFF_FILE_VALID(p->two)) {
diff_populate_filespec(p->two, 1);
copied = 0;
added = p->two->size;
diff_free_filespec_data(p->two);
} else
continue;
/*
* Original minus copied is the removed material,
* added is the new material. They are both damages
* made to the preimage.
* If the resulting damage is zero, we know that
* diffcore_count_changes() considers the two entries to
* be identical, but since content_changed is true, we
* know that there must have been _some_ kind of change,
* so we force all entries to have damage > 0.
*/
damage = (p->one->size - copied) + added;
if (!damage)
damage = 1;
found_damage:
ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
dir.files[dir.nr].name = name;
dir.files[dir.nr].changed = damage;
changed += damage;
dir.nr++;
Add "--dirstat" for some directory statistics This adds a new form of overview diffstat output, doing something that I have occasionally ended up doing manually (and badly, because it's actually pretty nasty to do), and that I think is very useful for an project like the kernel that has a fairly deep and well-separated directory structure with semantic meaning. What I mean by that is that it's often interesting to see exactly which sub-directories are impacted by a patch, and to what degree - even if you don't perhaps care so much about the individual files themselves. What makes the concept more interesting is that the "impact" is often hierarchical: in the kernel, for example, something could either have a very localized impact to "fs/ext3/" and then it's interesting to see that such a patch changes mostly that subdirectory, but you could have another patch that changes some generic VFS-layer issue which affects _many_ subdirectories that are all under "fs/", but none - or perhaps just a couple of them - of the individual filesystems are interesting in themselves. So what commonly happens is that you may have big changes in a specific sub-subdirectory, but still also significant separate changes to the subdirectory leading up to that - maybe you have significant VFS-level changes, but *also* changes under that VFS layer in the NFS-specific directories, for example. In that case, you do want the low-level parts that are significant to show up, but then the insignificant ones should show up as under the more generic top-level directory. This patch shows all of that with "--dirstat". The output can be either something simple like commit 81772fe... Author: Thomas Gleixner <tglx@linutronix.de> Date: Sun Feb 10 23:57:36 2008 +0100 x86: remove over noisy debug printk pageattr-test.c contains a noisy debug printk that people reported. The condition under which it prints (randomly tapping into a mem_map[] hole and not being able to c_p_a() there) is valid behavior and not interesting to report. Remove it. Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> 100.0% arch/x86/mm/ or something much more complex like commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 15.3% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ where that latter example is an example of significant work in some individual fs/*/ subdirectories (like the patches to reiserfs accounting for 7.6% of the whole), but then discounting those individual filesystems, there's also 15.3% other "random" things that weren't worth reporting on their oen left over under fs/ in general (either in that directory itself, or in subdirectories of fs/ that didn't have enough changes to be reported individually). I'd like to stress that the "15.3% fs/" mentioned above is the stuff that is under fs/ but that was _not_ significant enough to report on its own. So the above does _not_ mean that 15.3% of the work was under fs/ per se, because that 15.3% does *not* include the already-reported 7.6% of afs, 7.6% of fuse etc. If you want to enable "cumulative" directory statistics, you can use the "--cumulative" flag, which adds up percentages recursively even when they have been already reported for a sub-directory. That cumulative output is disabled if *all* of the changes in one subdirectory come from a deeper subdirectory, to avoid repeating subdirectories all the way to the root. For an example of the cumulative reporting, the above commit becomes commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 61.5% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ in which the commit percentages now obviously add up to much more than 100%: now the changes that were already reported for the sub-directories under fs/ are then cumulatively included in the whole percentage of fs/ (ie now shows 61.5% as opposed to the 15.3% without the cumulative reporting). The default reporting limit has been arbitrarily set at 3%, which seems to be a pretty good cut-off, but you can specify the cut-off manually by giving it as an option parameter (eg "--dirstat=5" makes the cut-off be at 5% instead) NOTE! The percentages are purely about the total lines added and removed, not anything smarter (or dumber) than that. Also note that you should not generally expect things to add up to 100%: not only does it round down, we don't report leftover scraps (they add up to the top-level change count, but we don't even bother reporting that, it only reports subdirectories). Quite frankly, as a top-level manager this is really convenient for me, but it's going to be very boring for git itself since there are few subdirectories. Also, don't expect things to make tons of sense if you combine this with "-M" and there are cross-directory renames etc. But even for git itself, you can get some fun statistics. Try out git log --dirstat and see the occasional mentions of things like Documentation/, git-gui/, gitweb/ and gitk-git/. Or try out something like git diff --dirstat v1.5.0..v1.5.4 which does kind of git an overview that shows *something*. But in general, the output is more exciting for big projects with deeper structure, and doing a git diff --dirstat v2.6.24..v2.6.25-rc1 on the kernel is what I actually wrote this for! Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 21:26:31 +00:00
}
/* This can happen even with many files, if everything was renames */
if (!changed)
return;
/* Show all directories with more than x% of the changes */
Fix '--dirstat' with cross-directory renaming The dirstat code depends on the fact that we always generate diffs with the names sorted, since it then just does a single-pass walk-over of the sorted list of names and how many changes there were. The sorting means that all files are nicely grouped by directory. That all works fine. Except when we have rename detection, and suddenly the nicely sorted list of pathnames isn't all that sorted at all. And now the single-pass dirstat walk gets all confused, and you can get results like this: [torvalds@nehalem linux]$ git diff --dirstat=2 -M v2.6.27-rc4..v2.6.27-rc5 3.0% arch/powerpc/configs/ 6.8% arch/arm/configs/ 2.7% arch/powerpc/configs/ 4.2% arch/arm/configs/ 5.6% arch/powerpc/configs/ 8.4% arch/arm/configs/ 5.5% arch/powerpc/configs/ 23.3% arch/arm/configs/ 8.6% arch/powerpc/configs/ 4.0% arch/ 4.4% drivers/usb/musb/ 4.0% drivers/watchdog/ 7.6% drivers/ 3.5% fs/ The trivial fix is to add a sorting pass, fixing it to: [torvalds@nehalem linux]$ git diff --dirstat=2 -M v2.6.27-rc4..v2.6.27-rc5 43.0% arch/arm/configs/ 25.5% arch/powerpc/configs/ 5.3% arch/ 4.4% drivers/usb/musb/ 4.0% drivers/watchdog/ 7.6% drivers/ 3.5% fs/ Spot the difference. In case anybody wonders: it's because of a ton of renames from {include/asm-blackfin => arch/blackfin/include/asm} that just totally messed up the file ordering in between arch/arm and arch/powerpc. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-08-28 23:19:08 +00:00
qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
gather_dirstat(options, &dir, changed, "", 0);
Add "--dirstat" for some directory statistics This adds a new form of overview diffstat output, doing something that I have occasionally ended up doing manually (and badly, because it's actually pretty nasty to do), and that I think is very useful for an project like the kernel that has a fairly deep and well-separated directory structure with semantic meaning. What I mean by that is that it's often interesting to see exactly which sub-directories are impacted by a patch, and to what degree - even if you don't perhaps care so much about the individual files themselves. What makes the concept more interesting is that the "impact" is often hierarchical: in the kernel, for example, something could either have a very localized impact to "fs/ext3/" and then it's interesting to see that such a patch changes mostly that subdirectory, but you could have another patch that changes some generic VFS-layer issue which affects _many_ subdirectories that are all under "fs/", but none - or perhaps just a couple of them - of the individual filesystems are interesting in themselves. So what commonly happens is that you may have big changes in a specific sub-subdirectory, but still also significant separate changes to the subdirectory leading up to that - maybe you have significant VFS-level changes, but *also* changes under that VFS layer in the NFS-specific directories, for example. In that case, you do want the low-level parts that are significant to show up, but then the insignificant ones should show up as under the more generic top-level directory. This patch shows all of that with "--dirstat". The output can be either something simple like commit 81772fe... Author: Thomas Gleixner <tglx@linutronix.de> Date: Sun Feb 10 23:57:36 2008 +0100 x86: remove over noisy debug printk pageattr-test.c contains a noisy debug printk that people reported. The condition under which it prints (randomly tapping into a mem_map[] hole and not being able to c_p_a() there) is valid behavior and not interesting to report. Remove it. Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> 100.0% arch/x86/mm/ or something much more complex like commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 15.3% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ where that latter example is an example of significant work in some individual fs/*/ subdirectories (like the patches to reiserfs accounting for 7.6% of the whole), but then discounting those individual filesystems, there's also 15.3% other "random" things that weren't worth reporting on their oen left over under fs/ in general (either in that directory itself, or in subdirectories of fs/ that didn't have enough changes to be reported individually). I'd like to stress that the "15.3% fs/" mentioned above is the stuff that is under fs/ but that was _not_ significant enough to report on its own. So the above does _not_ mean that 15.3% of the work was under fs/ per se, because that 15.3% does *not* include the already-reported 7.6% of afs, 7.6% of fuse etc. If you want to enable "cumulative" directory statistics, you can use the "--cumulative" flag, which adds up percentages recursively even when they have been already reported for a sub-directory. That cumulative output is disabled if *all* of the changes in one subdirectory come from a deeper subdirectory, to avoid repeating subdirectories all the way to the root. For an example of the cumulative reporting, the above commit becomes commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 61.5% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ in which the commit percentages now obviously add up to much more than 100%: now the changes that were already reported for the sub-directories under fs/ are then cumulatively included in the whole percentage of fs/ (ie now shows 61.5% as opposed to the 15.3% without the cumulative reporting). The default reporting limit has been arbitrarily set at 3%, which seems to be a pretty good cut-off, but you can specify the cut-off manually by giving it as an option parameter (eg "--dirstat=5" makes the cut-off be at 5% instead) NOTE! The percentages are purely about the total lines added and removed, not anything smarter (or dumber) than that. Also note that you should not generally expect things to add up to 100%: not only does it round down, we don't report leftover scraps (they add up to the top-level change count, but we don't even bother reporting that, it only reports subdirectories). Quite frankly, as a top-level manager this is really convenient for me, but it's going to be very boring for git itself since there are few subdirectories. Also, don't expect things to make tons of sense if you combine this with "-M" and there are cross-directory renames etc. But even for git itself, you can get some fun statistics. Try out git log --dirstat and see the occasional mentions of things like Documentation/, git-gui/, gitweb/ and gitk-git/. Or try out something like git diff --dirstat v1.5.0..v1.5.4 which does kind of git an overview that shows *something*. But in general, the output is more exciting for big projects with deeper structure, and doing a git diff --dirstat v2.6.24..v2.6.25-rc1 on the kernel is what I actually wrote this for! Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 21:26:31 +00:00
}
New --dirstat=lines mode, doing dirstat analysis based on diffstat This patch adds an alternative implementation of show_dirstat(), called show_dirstat_by_line(), which uses the more expensive diffstat analysis (as opposed to show_dirstat()'s own (relatively inexpensive) analysis) to derive the numbers from which the --dirstat output is computed. The alternative implementation is controlled by the new "lines" parameter to the --dirstat option (or the diff.dirstat config variable). For binary files, the diffstat analysis counts bytes instead of lines, so to prevent binary files from dominating the dirstat results, the byte counts for binary files are divided by 64 before being compared to their textual/line-based counterparts. This is a stupid and ugly - but very cheap - heuristic. In linux-2.6.git, running the three different --dirstat modes: time git diff v2.6.20..v2.6.30 --dirstat=changes > /dev/null vs. time git diff v2.6.20..v2.6.30 --dirstat=lines > /dev/null vs. time git diff v2.6.20..v2.6.30 --dirstat=files > /dev/null yields the following average runtimes on my machine: - "changes" (default): ~6.0 s - "lines": ~9.6 s - "files": ~0.1 s So, as expected, there's a considerable performance hit (~60%) by going through the full diffstat analysis as compared to the default "changes" analysis (obviously, "files" is much faster than both). As such, the "lines" mode is probably only useful if you really need the --dirstat numbers to be consistent with the numbers returned from the other --*stat options. The patch also includes documentation and tests for the new dirstat mode. Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:21 +00:00
static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
{
int i;
unsigned long changed;
struct dirstat_dir dir;
if (data->nr == 0)
return;
dir.files = NULL;
dir.alloc = 0;
dir.nr = 0;
dir.permille = options->dirstat_permille;
dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
changed = 0;
for (i = 0; i < data->nr; i++) {
struct diffstat_file *file = data->files[i];
unsigned long damage = file->added + file->deleted;
if (file->is_binary)
/*
* binary files counts bytes, not lines. Must find some
* way to normalize binary bytes vs. textual lines.
* The following heuristic assumes that there are 64
* bytes per "line".
* This is stupid and ugly, but very cheap...
*/
damage = (damage + 63) / 64;
ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
dir.files[dir.nr].name = file->name;
dir.files[dir.nr].changed = damage;
changed += damage;
dir.nr++;
}
/* This can happen even with many files, if everything was renames */
if (!changed)
return;
/* Show all directories with more than x% of the changes */
qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
gather_dirstat(options, &dir, changed, "", 0);
}
static void free_diffstat_info(struct diffstat_t *diffstat)
{
int i;
for (i = 0; i < diffstat->nr; i++) {
struct diffstat_file *f = diffstat->files[i];
if (f->name != f->print_name)
free(f->print_name);
free(f->name);
free(f->from_name);
free(f);
}
free(diffstat->files);
}
struct checkdiff_t {
const char *filename;
int lineno;
int conflict_marker_size;
struct diff_options *o;
unsigned ws_rule;
unsigned status;
};
static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
{
char firstchar;
int cnt;
if (len < marker_size + 1)
return 0;
firstchar = line[0];
switch (firstchar) {
case '=': case '>': case '<': case '|':
break;
default:
return 0;
}
for (cnt = 1; cnt < marker_size; cnt++)
if (line[cnt] != firstchar)
return 0;
/* line[1] thru line[marker_size-1] are same as firstchar */
if (len < marker_size + 1 || !isspace(line[marker_size]))
return 0;
return 1;
}
static void checkdiff_consume(void *priv, char *line, unsigned long len)
{
struct checkdiff_t *data = priv;
int marker_size = data->conflict_marker_size;
const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
char *err;
char *line_prefix = "";
struct strbuf *msgbuf;
assert(data->o);
if (data->o->output_prefix) {
msgbuf = data->o->output_prefix(data->o,
data->o->output_prefix_data);
line_prefix = msgbuf->buf;
}
if (line[0] == '+') {
unsigned bad;
data->lineno++;
if (is_conflict_marker(line + 1, marker_size, len - 1)) {
data->status |= 1;
fprintf(data->o->file,
"%s%s:%d: leftover conflict marker\n",
line_prefix, data->filename, data->lineno);
}
bad = ws_check(line + 1, len - 1, data->ws_rule);
if (!bad)
return;
data->status |= bad;
err = whitespace_error_string(bad);
fprintf(data->o->file, "%s%s:%d: %s.\n",
line_prefix, data->filename, data->lineno, err);
free(err);
emit_line(data->o, set, reset, line, 1);
ws_check_emit(line + 1, len - 1, data->ws_rule,
data->o->file, set, reset, ws);
} else if (line[0] == ' ') {
data->lineno++;
} else if (line[0] == '@') {
char *plus = strchr(line, '+');
if (plus)
data->lineno = strtol(plus, NULL, 10) - 1;
else
die("invalid diff");
}
}
static unsigned char *deflate_it(char *data,
unsigned long size,
unsigned long *result_size)
binary patch. This adds "binary patch" to the diff output and teaches apply what to do with them. On the diff generation side, traditionally, we said "Binary files differ\n" without giving anything other than the preimage and postimage object name on the index line. This was good enough for applying a patch generated from your own repository (very useful while rebasing), because the postimage would be available in such a case. However, this was not useful when the recipient of such a patch via e-mail were to apply it, even if the preimage was available. This patch allows the diff to generate "binary" patch when operating under --full-index option. The binary patch follows the usual extended git diff headers, and looks like this: "GIT binary patch\n" <length byte><data>"\n" ... "\n" Each line is prefixed with a "length-byte", whose value is upper or lowercase alphabet that encodes number of bytes that the data on the line decodes to (1..52 -- 'A' means 1, 'B' means 2, ..., 'Z' means 26, 'a' means 27, ...). <data> is 1 or more groups of 5-byte sequence, each of which encodes up to 4 bytes in base85 encoding. Because 52 / 4 * 5 = 65 and we have the length byte, an output line is capped to 66 characters. The payload is the same diff-delta as we use in the packfiles. On the consumption side, git-apply now can decode and apply the binary patch when --allow-binary-replacement is given, the diff was generated with --full-index, and the receiving repository has the preimage blob, which is the same condition as it always required when accepting an "Binary files differ\n" patch. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-04 23:51:44 +00:00
{
int bound;
unsigned char *deflated;
2011-06-10 18:52:15 +00:00
git_zstream stream;
memset(&stream, 0, sizeof(stream));
git_deflate_init(&stream, zlib_compression_level);
bound = git_deflate_bound(&stream, size);
deflated = xmalloc(bound);
stream.next_out = deflated;
stream.avail_out = bound;
stream.next_in = (unsigned char *)data;
stream.avail_in = size;
while (git_deflate(&stream, Z_FINISH) == Z_OK)
; /* nothing */
git_deflate_end(&stream);
*result_size = stream.total_out;
return deflated;
binary patch. This adds "binary patch" to the diff output and teaches apply what to do with them. On the diff generation side, traditionally, we said "Binary files differ\n" without giving anything other than the preimage and postimage object name on the index line. This was good enough for applying a patch generated from your own repository (very useful while rebasing), because the postimage would be available in such a case. However, this was not useful when the recipient of such a patch via e-mail were to apply it, even if the preimage was available. This patch allows the diff to generate "binary" patch when operating under --full-index option. The binary patch follows the usual extended git diff headers, and looks like this: "GIT binary patch\n" <length byte><data>"\n" ... "\n" Each line is prefixed with a "length-byte", whose value is upper or lowercase alphabet that encodes number of bytes that the data on the line decodes to (1..52 -- 'A' means 1, 'B' means 2, ..., 'Z' means 26, 'a' means 27, ...). <data> is 1 or more groups of 5-byte sequence, each of which encodes up to 4 bytes in base85 encoding. Because 52 / 4 * 5 = 65 and we have the length byte, an output line is capped to 66 characters. The payload is the same diff-delta as we use in the packfiles. On the consumption side, git-apply now can decode and apply the binary patch when --allow-binary-replacement is given, the diff was generated with --full-index, and the receiving repository has the preimage blob, which is the same condition as it always required when accepting an "Binary files differ\n" patch. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-04 23:51:44 +00:00
}
static void emit_binary_diff_body(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
binary patch. This adds "binary patch" to the diff output and teaches apply what to do with them. On the diff generation side, traditionally, we said "Binary files differ\n" without giving anything other than the preimage and postimage object name on the index line. This was good enough for applying a patch generated from your own repository (very useful while rebasing), because the postimage would be available in such a case. However, this was not useful when the recipient of such a patch via e-mail were to apply it, even if the preimage was available. This patch allows the diff to generate "binary" patch when operating under --full-index option. The binary patch follows the usual extended git diff headers, and looks like this: "GIT binary patch\n" <length byte><data>"\n" ... "\n" Each line is prefixed with a "length-byte", whose value is upper or lowercase alphabet that encodes number of bytes that the data on the line decodes to (1..52 -- 'A' means 1, 'B' means 2, ..., 'Z' means 26, 'a' means 27, ...). <data> is 1 or more groups of 5-byte sequence, each of which encodes up to 4 bytes in base85 encoding. Because 52 / 4 * 5 = 65 and we have the length byte, an output line is capped to 66 characters. The payload is the same diff-delta as we use in the packfiles. On the consumption side, git-apply now can decode and apply the binary patch when --allow-binary-replacement is given, the diff was generated with --full-index, and the receiving repository has the preimage blob, which is the same condition as it always required when accepting an "Binary files differ\n" patch. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-04 23:51:44 +00:00
{
void *cp;
void *delta;
void *deflated;
void *data;
unsigned long orig_size;
unsigned long delta_size;
unsigned long deflate_size;
unsigned long data_size;
binary patch. This adds "binary patch" to the diff output and teaches apply what to do with them. On the diff generation side, traditionally, we said "Binary files differ\n" without giving anything other than the preimage and postimage object name on the index line. This was good enough for applying a patch generated from your own repository (very useful while rebasing), because the postimage would be available in such a case. However, this was not useful when the recipient of such a patch via e-mail were to apply it, even if the preimage was available. This patch allows the diff to generate "binary" patch when operating under --full-index option. The binary patch follows the usual extended git diff headers, and looks like this: "GIT binary patch\n" <length byte><data>"\n" ... "\n" Each line is prefixed with a "length-byte", whose value is upper or lowercase alphabet that encodes number of bytes that the data on the line decodes to (1..52 -- 'A' means 1, 'B' means 2, ..., 'Z' means 26, 'a' means 27, ...). <data> is 1 or more groups of 5-byte sequence, each of which encodes up to 4 bytes in base85 encoding. Because 52 / 4 * 5 = 65 and we have the length byte, an output line is capped to 66 characters. The payload is the same diff-delta as we use in the packfiles. On the consumption side, git-apply now can decode and apply the binary patch when --allow-binary-replacement is given, the diff was generated with --full-index, and the receiving repository has the preimage blob, which is the same condition as it always required when accepting an "Binary files differ\n" patch. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-04 23:51:44 +00:00
/* We could do deflated delta, or we could do just deflated two,
* whichever is smaller.
binary patch. This adds "binary patch" to the diff output and teaches apply what to do with them. On the diff generation side, traditionally, we said "Binary files differ\n" without giving anything other than the preimage and postimage object name on the index line. This was good enough for applying a patch generated from your own repository (very useful while rebasing), because the postimage would be available in such a case. However, this was not useful when the recipient of such a patch via e-mail were to apply it, even if the preimage was available. This patch allows the diff to generate "binary" patch when operating under --full-index option. The binary patch follows the usual extended git diff headers, and looks like this: "GIT binary patch\n" <length byte><data>"\n" ... "\n" Each line is prefixed with a "length-byte", whose value is upper or lowercase alphabet that encodes number of bytes that the data on the line decodes to (1..52 -- 'A' means 1, 'B' means 2, ..., 'Z' means 26, 'a' means 27, ...). <data> is 1 or more groups of 5-byte sequence, each of which encodes up to 4 bytes in base85 encoding. Because 52 / 4 * 5 = 65 and we have the length byte, an output line is capped to 66 characters. The payload is the same diff-delta as we use in the packfiles. On the consumption side, git-apply now can decode and apply the binary patch when --allow-binary-replacement is given, the diff was generated with --full-index, and the receiving repository has the preimage blob, which is the same condition as it always required when accepting an "Binary files differ\n" patch. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-04 23:51:44 +00:00
*/
delta = NULL;
deflated = deflate_it(two->ptr, two->size, &deflate_size);
if (one->size && two->size) {
delta = diff_delta(one->ptr, one->size,
two->ptr, two->size,
&delta_size, deflate_size);
if (delta) {
void *to_free = delta;
orig_size = delta_size;
delta = deflate_it(delta, delta_size, &delta_size);
free(to_free);
binary patch. This adds "binary patch" to the diff output and teaches apply what to do with them. On the diff generation side, traditionally, we said "Binary files differ\n" without giving anything other than the preimage and postimage object name on the index line. This was good enough for applying a patch generated from your own repository (very useful while rebasing), because the postimage would be available in such a case. However, this was not useful when the recipient of such a patch via e-mail were to apply it, even if the preimage was available. This patch allows the diff to generate "binary" patch when operating under --full-index option. The binary patch follows the usual extended git diff headers, and looks like this: "GIT binary patch\n" <length byte><data>"\n" ... "\n" Each line is prefixed with a "length-byte", whose value is upper or lowercase alphabet that encodes number of bytes that the data on the line decodes to (1..52 -- 'A' means 1, 'B' means 2, ..., 'Z' means 26, 'a' means 27, ...). <data> is 1 or more groups of 5-byte sequence, each of which encodes up to 4 bytes in base85 encoding. Because 52 / 4 * 5 = 65 and we have the length byte, an output line is capped to 66 characters. The payload is the same diff-delta as we use in the packfiles. On the consumption side, git-apply now can decode and apply the binary patch when --allow-binary-replacement is given, the diff was generated with --full-index, and the receiving repository has the preimage blob, which is the same condition as it always required when accepting an "Binary files differ\n" patch. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-04 23:51:44 +00:00
}
}
if (delta && delta_size < deflate_size) {
fprintf(file, "%sdelta %lu\n", prefix, orig_size);
free(deflated);
data = delta;
data_size = delta_size;
}
else {
fprintf(file, "%sliteral %lu\n", prefix, two->size);
free(delta);
data = deflated;
data_size = deflate_size;
}
binary patch. This adds "binary patch" to the diff output and teaches apply what to do with them. On the diff generation side, traditionally, we said "Binary files differ\n" without giving anything other than the preimage and postimage object name on the index line. This was good enough for applying a patch generated from your own repository (very useful while rebasing), because the postimage would be available in such a case. However, this was not useful when the recipient of such a patch via e-mail were to apply it, even if the preimage was available. This patch allows the diff to generate "binary" patch when operating under --full-index option. The binary patch follows the usual extended git diff headers, and looks like this: "GIT binary patch\n" <length byte><data>"\n" ... "\n" Each line is prefixed with a "length-byte", whose value is upper or lowercase alphabet that encodes number of bytes that the data on the line decodes to (1..52 -- 'A' means 1, 'B' means 2, ..., 'Z' means 26, 'a' means 27, ...). <data> is 1 or more groups of 5-byte sequence, each of which encodes up to 4 bytes in base85 encoding. Because 52 / 4 * 5 = 65 and we have the length byte, an output line is capped to 66 characters. The payload is the same diff-delta as we use in the packfiles. On the consumption side, git-apply now can decode and apply the binary patch when --allow-binary-replacement is given, the diff was generated with --full-index, and the receiving repository has the preimage blob, which is the same condition as it always required when accepting an "Binary files differ\n" patch. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-04 23:51:44 +00:00
/* emit data encoded in base85 */
cp = data;
while (data_size) {
int bytes = (52 < data_size) ? 52 : data_size;
binary patch. This adds "binary patch" to the diff output and teaches apply what to do with them. On the diff generation side, traditionally, we said "Binary files differ\n" without giving anything other than the preimage and postimage object name on the index line. This was good enough for applying a patch generated from your own repository (very useful while rebasing), because the postimage would be available in such a case. However, this was not useful when the recipient of such a patch via e-mail were to apply it, even if the preimage was available. This patch allows the diff to generate "binary" patch when operating under --full-index option. The binary patch follows the usual extended git diff headers, and looks like this: "GIT binary patch\n" <length byte><data>"\n" ... "\n" Each line is prefixed with a "length-byte", whose value is upper or lowercase alphabet that encodes number of bytes that the data on the line decodes to (1..52 -- 'A' means 1, 'B' means 2, ..., 'Z' means 26, 'a' means 27, ...). <data> is 1 or more groups of 5-byte sequence, each of which encodes up to 4 bytes in base85 encoding. Because 52 / 4 * 5 = 65 and we have the length byte, an output line is capped to 66 characters. The payload is the same diff-delta as we use in the packfiles. On the consumption side, git-apply now can decode and apply the binary patch when --allow-binary-replacement is given, the diff was generated with --full-index, and the receiving repository has the preimage blob, which is the same condition as it always required when accepting an "Binary files differ\n" patch. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-04 23:51:44 +00:00
char line[70];
data_size -= bytes;
binary patch. This adds "binary patch" to the diff output and teaches apply what to do with them. On the diff generation side, traditionally, we said "Binary files differ\n" without giving anything other than the preimage and postimage object name on the index line. This was good enough for applying a patch generated from your own repository (very useful while rebasing), because the postimage would be available in such a case. However, this was not useful when the recipient of such a patch via e-mail were to apply it, even if the preimage was available. This patch allows the diff to generate "binary" patch when operating under --full-index option. The binary patch follows the usual extended git diff headers, and looks like this: "GIT binary patch\n" <length byte><data>"\n" ... "\n" Each line is prefixed with a "length-byte", whose value is upper or lowercase alphabet that encodes number of bytes that the data on the line decodes to (1..52 -- 'A' means 1, 'B' means 2, ..., 'Z' means 26, 'a' means 27, ...). <data> is 1 or more groups of 5-byte sequence, each of which encodes up to 4 bytes in base85 encoding. Because 52 / 4 * 5 = 65 and we have the length byte, an output line is capped to 66 characters. The payload is the same diff-delta as we use in the packfiles. On the consumption side, git-apply now can decode and apply the binary patch when --allow-binary-replacement is given, the diff was generated with --full-index, and the receiving repository has the preimage blob, which is the same condition as it always required when accepting an "Binary files differ\n" patch. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-04 23:51:44 +00:00
if (bytes <= 26)
line[0] = bytes + 'A' - 1;
else
line[0] = bytes - 26 + 'a' - 1;
encode_85(line + 1, cp, bytes);
cp = (char *) cp + bytes;
fprintf(file, "%s", prefix);
fputs(line, file);
fputc('\n', file);
binary patch. This adds "binary patch" to the diff output and teaches apply what to do with them. On the diff generation side, traditionally, we said "Binary files differ\n" without giving anything other than the preimage and postimage object name on the index line. This was good enough for applying a patch generated from your own repository (very useful while rebasing), because the postimage would be available in such a case. However, this was not useful when the recipient of such a patch via e-mail were to apply it, even if the preimage was available. This patch allows the diff to generate "binary" patch when operating under --full-index option. The binary patch follows the usual extended git diff headers, and looks like this: "GIT binary patch\n" <length byte><data>"\n" ... "\n" Each line is prefixed with a "length-byte", whose value is upper or lowercase alphabet that encodes number of bytes that the data on the line decodes to (1..52 -- 'A' means 1, 'B' means 2, ..., 'Z' means 26, 'a' means 27, ...). <data> is 1 or more groups of 5-byte sequence, each of which encodes up to 4 bytes in base85 encoding. Because 52 / 4 * 5 = 65 and we have the length byte, an output line is capped to 66 characters. The payload is the same diff-delta as we use in the packfiles. On the consumption side, git-apply now can decode and apply the binary patch when --allow-binary-replacement is given, the diff was generated with --full-index, and the receiving repository has the preimage blob, which is the same condition as it always required when accepting an "Binary files differ\n" patch. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-04 23:51:44 +00:00
}
fprintf(file, "%s\n", prefix);
free(data);
binary patch. This adds "binary patch" to the diff output and teaches apply what to do with them. On the diff generation side, traditionally, we said "Binary files differ\n" without giving anything other than the preimage and postimage object name on the index line. This was good enough for applying a patch generated from your own repository (very useful while rebasing), because the postimage would be available in such a case. However, this was not useful when the recipient of such a patch via e-mail were to apply it, even if the preimage was available. This patch allows the diff to generate "binary" patch when operating under --full-index option. The binary patch follows the usual extended git diff headers, and looks like this: "GIT binary patch\n" <length byte><data>"\n" ... "\n" Each line is prefixed with a "length-byte", whose value is upper or lowercase alphabet that encodes number of bytes that the data on the line decodes to (1..52 -- 'A' means 1, 'B' means 2, ..., 'Z' means 26, 'a' means 27, ...). <data> is 1 or more groups of 5-byte sequence, each of which encodes up to 4 bytes in base85 encoding. Because 52 / 4 * 5 = 65 and we have the length byte, an output line is capped to 66 characters. The payload is the same diff-delta as we use in the packfiles. On the consumption side, git-apply now can decode and apply the binary patch when --allow-binary-replacement is given, the diff was generated with --full-index, and the receiving repository has the preimage blob, which is the same condition as it always required when accepting an "Binary files differ\n" patch. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-04 23:51:44 +00:00
}
static void emit_binary_diff(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
{
fprintf(file, "%sGIT binary patch\n", prefix);
emit_binary_diff_body(file, one, two, prefix);
emit_binary_diff_body(file, two, one, prefix);
}
static void diff_filespec_load_driver(struct diff_filespec *one)
{
/* Use already-loaded driver */
if (one->driver)
return;
if (S_ISREG(one->mode))
diff: introduce diff.<driver>.binary The "diff" gitattribute is somewhat overloaded right now. It can say one of three things: 1. this file is definitely binary, or definitely not (i.e., diff or !diff) 2. this file should use an external diff engine (i.e., diff=foo, diff.foo.command = custom-script) 3. this file should use particular funcname patterns (i.e., diff=foo, diff.foo.(x?)funcname = some-regex) Most of the time, there is no conflict between these uses, since using one implies that the other is irrelevant (e.g., an external diff engine will decide for itself whether the file is binary). However, there is at least one conflicting situation: there is no way to say "use the regular rules to determine whether this file is binary, but if we do diff it textually, use this funcname pattern." That is, currently setting diff=foo indicates that the file is definitely text. This patch introduces a "binary" config option for a diff driver, so that one can explicitly set diff.foo.binary. We default this value to "don't know". That is, setting a diff attribute to "foo" and using "diff.foo.funcname" will have no effect on the binaryness of a file. To get the current behavior, one can set diff.foo.binary to true. This patch also has one additional advantage: it cleans up the interface to the userdiff code a bit. Before, calling code had to know more about whether attributes were false, true, or unset to determine binaryness. Now that binaryness is a property of a driver, we can represent these situations just by passing back a driver struct. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2008-10-05 21:43:36 +00:00
one->driver = userdiff_find_by_path(one->path);
/* Fallback to default settings */
diff: introduce diff.<driver>.binary The "diff" gitattribute is somewhat overloaded right now. It can say one of three things: 1. this file is definitely binary, or definitely not (i.e., diff or !diff) 2. this file should use an external diff engine (i.e., diff=foo, diff.foo.command = custom-script) 3. this file should use particular funcname patterns (i.e., diff=foo, diff.foo.(x?)funcname = some-regex) Most of the time, there is no conflict between these uses, since using one implies that the other is irrelevant (e.g., an external diff engine will decide for itself whether the file is binary). However, there is at least one conflicting situation: there is no way to say "use the regular rules to determine whether this file is binary, but if we do diff it textually, use this funcname pattern." That is, currently setting diff=foo indicates that the file is definitely text. This patch introduces a "binary" config option for a diff driver, so that one can explicitly set diff.foo.binary. We default this value to "don't know". That is, setting a diff attribute to "foo" and using "diff.foo.funcname" will have no effect on the binaryness of a file. To get the current behavior, one can set diff.foo.binary to true. This patch also has one additional advantage: it cleans up the interface to the userdiff code a bit. Before, calling code had to know more about whether attributes were false, true, or unset to determine binaryness. Now that binaryness is a property of a driver, we can represent these situations just by passing back a driver struct. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2008-10-05 21:43:36 +00:00
if (!one->driver)
one->driver = userdiff_find_by_name("default");
}
int diff_filespec_is_binary(struct diff_filespec *one)
{
diff: introduce diff.<driver>.binary The "diff" gitattribute is somewhat overloaded right now. It can say one of three things: 1. this file is definitely binary, or definitely not (i.e., diff or !diff) 2. this file should use an external diff engine (i.e., diff=foo, diff.foo.command = custom-script) 3. this file should use particular funcname patterns (i.e., diff=foo, diff.foo.(x?)funcname = some-regex) Most of the time, there is no conflict between these uses, since using one implies that the other is irrelevant (e.g., an external diff engine will decide for itself whether the file is binary). However, there is at least one conflicting situation: there is no way to say "use the regular rules to determine whether this file is binary, but if we do diff it textually, use this funcname pattern." That is, currently setting diff=foo indicates that the file is definitely text. This patch introduces a "binary" config option for a diff driver, so that one can explicitly set diff.foo.binary. We default this value to "don't know". That is, setting a diff attribute to "foo" and using "diff.foo.funcname" will have no effect on the binaryness of a file. To get the current behavior, one can set diff.foo.binary to true. This patch also has one additional advantage: it cleans up the interface to the userdiff code a bit. Before, calling code had to know more about whether attributes were false, true, or unset to determine binaryness. Now that binaryness is a property of a driver, we can represent these situations just by passing back a driver struct. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2008-10-05 21:43:36 +00:00
if (one->is_binary == -1) {
diff_filespec_load_driver(one);
if (one->driver->binary != -1)
one->is_binary = one->driver->binary;
else {
if (!one->data && DIFF_FILE_VALID(one))
diff_populate_filespec(one, 0);
if (one->data)
one->is_binary = buffer_is_binary(one->data,
one->size);
if (one->is_binary == -1)
one->is_binary = 0;
}
}
return one->is_binary;
}
static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
{
diff: introduce diff.<driver>.binary The "diff" gitattribute is somewhat overloaded right now. It can say one of three things: 1. this file is definitely binary, or definitely not (i.e., diff or !diff) 2. this file should use an external diff engine (i.e., diff=foo, diff.foo.command = custom-script) 3. this file should use particular funcname patterns (i.e., diff=foo, diff.foo.(x?)funcname = some-regex) Most of the time, there is no conflict between these uses, since using one implies that the other is irrelevant (e.g., an external diff engine will decide for itself whether the file is binary). However, there is at least one conflicting situation: there is no way to say "use the regular rules to determine whether this file is binary, but if we do diff it textually, use this funcname pattern." That is, currently setting diff=foo indicates that the file is definitely text. This patch introduces a "binary" config option for a diff driver, so that one can explicitly set diff.foo.binary. We default this value to "don't know". That is, setting a diff attribute to "foo" and using "diff.foo.funcname" will have no effect on the binaryness of a file. To get the current behavior, one can set diff.foo.binary to true. This patch also has one additional advantage: it cleans up the interface to the userdiff code a bit. Before, calling code had to know more about whether attributes were false, true, or unset to determine binaryness. Now that binaryness is a property of a driver, we can represent these situations just by passing back a driver struct. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2008-10-05 21:43:36 +00:00
diff_filespec_load_driver(one);
return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
}
static const char *userdiff_word_regex(struct diff_filespec *one)
{
diff_filespec_load_driver(one);
return one->driver->word_regex;
}
void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
{
if (!options->a_prefix)
options->a_prefix = a;
if (!options->b_prefix)
options->b_prefix = b;
}
struct userdiff_driver *get_textconv(struct diff_filespec *one)
{
if (!DIFF_FILE_VALID(one))
return NULL;
diff_filespec_load_driver(one);
return userdiff_get_textconv(one->driver);
}
static void builtin_diff(const char *name_a,
const char *name_b,
struct diff_filespec *one,
struct diff_filespec *two,
const char *xfrm_msg,
int must_show_header,
binary patch. This adds "binary patch" to the diff output and teaches apply what to do with them. On the diff generation side, traditionally, we said "Binary files differ\n" without giving anything other than the preimage and postimage object name on the index line. This was good enough for applying a patch generated from your own repository (very useful while rebasing), because the postimage would be available in such a case. However, this was not useful when the recipient of such a patch via e-mail were to apply it, even if the preimage was available. This patch allows the diff to generate "binary" patch when operating under --full-index option. The binary patch follows the usual extended git diff headers, and looks like this: "GIT binary patch\n" <length byte><data>"\n" ... "\n" Each line is prefixed with a "length-byte", whose value is upper or lowercase alphabet that encodes number of bytes that the data on the line decodes to (1..52 -- 'A' means 1, 'B' means 2, ..., 'Z' means 26, 'a' means 27, ...). <data> is 1 or more groups of 5-byte sequence, each of which encodes up to 4 bytes in base85 encoding. Because 52 / 4 * 5 = 65 and we have the length byte, an output line is capped to 66 characters. The payload is the same diff-delta as we use in the packfiles. On the consumption side, git-apply now can decode and apply the binary patch when --allow-binary-replacement is given, the diff was generated with --full-index, and the receiving repository has the preimage blob, which is the same condition as it always required when accepting an "Binary files differ\n" patch. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-04 23:51:44 +00:00
struct diff_options *o,
int complete_rewrite)
{
mmfile_t mf1, mf2;
const char *lbl[2];
char *a_one, *b_two;
const char *set = diff_get_color_opt(o, DIFF_METAINFO);
const char *reset = diff_get_color_opt(o, DIFF_RESET);
const char *a_prefix, *b_prefix;
struct userdiff_driver *textconv_one = NULL;
struct userdiff_driver *textconv_two = NULL;
struct strbuf header = STRBUF_INIT;
struct strbuf *msgbuf;
char *line_prefix = "";
if (o->output_prefix) {
msgbuf = o->output_prefix(o, o->output_prefix_data);
line_prefix = msgbuf->buf;
}
if (DIFF_OPT_TST(o, SUBMODULE_LOG) &&
(!one->mode || S_ISGITLINK(one->mode)) &&
(!two->mode || S_ISGITLINK(two->mode))) {
const char *del = diff_get_color_opt(o, DIFF_FILE_OLD);
const char *add = diff_get_color_opt(o, DIFF_FILE_NEW);
show_submodule_summary(o->file, one ? one->path : two->path,
one->sha1, two->sha1, two->dirty_submodule,
del, add, reset);
return;
}
if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
textconv_one = get_textconv(one);
textconv_two = get_textconv(two);
}
diff_set_mnemonic_prefix(o, "a/", "b/");
if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
a_prefix = o->b_prefix;
b_prefix = o->a_prefix;
} else {
a_prefix = o->a_prefix;
b_prefix = o->b_prefix;
}
fix bogus "diff --git" header from "diff --no-index" When "git diff --no-index" is given an absolute pathname, it would generate a diff header with the absolute path prepended by the prefix, like: diff --git a/dev/null b/foo Not only is this nonsensical, and not only does it violate the description of diffs given in git-diff(1), but it would produce broken binary diffs. Unlike text diffs, the binary diffs don't contain the filenames anywhere else, and so "git apply" relies on this header to figure out the filename. This patch just refuses to use an invalid name for anything visible in the diff. Now, this fixes the "git diff --no-index --binary a /dev/null" kind of case (and we'll end up using "a" as the basename), but some other insane cases are impossible to handle. If you do git diff --no-index --binary a /bin/echo you'll still get a patch like diff --git a/a b/bin/echo old mode 100644 new mode 100755 index ... and "git apply" will refuse to apply it for a couple of reasons, and the diff is simply bogus. And that, btw, is no longer a bug, I think. It's impossible to know whethe the user meant for the patch to be a rename or not. And as such, refusing to apply it because you don't know what name you should use is probably _exactly_ the right thing to do! Original problem reported by Imre Deak. Test script and problem description by Jeff King. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2008-10-05 19:35:15 +00:00
/* Never use a non-valid filename anywhere if at all possible */
name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, set, a_one, b_two, reset);
if (lbl[0][0] == '/') {
/* /dev/null */
strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, set, two->mode, reset);
if (xfrm_msg)
strbuf_addstr(&header, xfrm_msg);
must_show_header = 1;
}
else if (lbl[1][0] == '/') {
strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, set, one->mode, reset);
if (xfrm_msg)
strbuf_addstr(&header, xfrm_msg);
must_show_header = 1;
}
else {
if (one->mode != two->mode) {
strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, set, one->mode, reset);
strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, set, two->mode, reset);
must_show_header = 1;
}
if (xfrm_msg)
strbuf_addstr(&header, xfrm_msg);
/*
* we do not run diff between different kind
* of objects.
*/
if ((one->mode ^ two->mode) & S_IFMT)
goto free_ab_and_return;
if (complete_rewrite &&
(textconv_one || !diff_filespec_is_binary(one)) &&
(textconv_two || !diff_filespec_is_binary(two))) {
fprintf(o->file, "%s", header.buf);
strbuf_reset(&header);
emit_rewrite_diff(name_a, name_b, one, two,
textconv_one, textconv_two, o);
o->found_changes = 1;
goto free_ab_and_return;
}
}
if (o->irreversible_delete && lbl[1][0] == '/') {
fprintf(o->file, "%s", header.buf);
strbuf_reset(&header);
goto free_ab_and_return;
} else if (!DIFF_OPT_TST(o, TEXT) &&
diff: avoid useless filespec population builtin_diff calls fill_mmfile fairly early, which in turn calls diff_populate_filespec, which actually retrieves the file's blob contents into a buffer. Long ago, this was sensible as we would need to look at the blobs eventually. These days, however, we may not ever want those blobs if we end up using a textconv cache, and for large binary files (exactly the sort for which you might have a textconv cache), just retrieving the objects can be costly. This patch just pushes the fill_mmfile call a bit later, so we can avoid populating the filespec in some cases. There is one thing to note that looks like a bug but isn't. We push the fill_mmfile down into the first branch of a conditional. It seems like we would need it on the other branch, too, but we don't; fill_textconv does it for us (in fact, before this, we were just writing over the results of the fill_mmfile on that branch). Here's a timing sample on a commit with 45 changed jpgs and avis. The result is fully textconv cached, but we still wasted a lot of time just pulling the blobs from storage. The total size of the blobs (source and dest) is about 180M. [before] $ time git show >/dev/null real 0m0.352s user 0m0.148s sys 0m0.200s [after] $ time git show >/dev/null real 0m0.009s user 0m0.004s sys 0m0.004s And that's on a warm cache. On a cold cache, the "after" case is not much worse, but the "before" case has to do an extra 180M of I/O. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-04-02 00:14:24 +00:00
( (!textconv_one && diff_filespec_is_binary(one)) ||
(!textconv_two && diff_filespec_is_binary(two)) )) {
if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
die("unable to read files to diff");
/* Quite common confusing case */
if (mf1.size == mf2.size &&
!memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
if (must_show_header)
fprintf(o->file, "%s", header.buf);
goto free_ab_and_return;
}
fprintf(o->file, "%s", header.buf);
strbuf_reset(&header);
if (DIFF_OPT_TST(o, BINARY))
emit_binary_diff(o->file, &mf1, &mf2, line_prefix);
binary patch. This adds "binary patch" to the diff output and teaches apply what to do with them. On the diff generation side, traditionally, we said "Binary files differ\n" without giving anything other than the preimage and postimage object name on the index line. This was good enough for applying a patch generated from your own repository (very useful while rebasing), because the postimage would be available in such a case. However, this was not useful when the recipient of such a patch via e-mail were to apply it, even if the preimage was available. This patch allows the diff to generate "binary" patch when operating under --full-index option. The binary patch follows the usual extended git diff headers, and looks like this: "GIT binary patch\n" <length byte><data>"\n" ... "\n" Each line is prefixed with a "length-byte", whose value is upper or lowercase alphabet that encodes number of bytes that the data on the line decodes to (1..52 -- 'A' means 1, 'B' means 2, ..., 'Z' means 26, 'a' means 27, ...). <data> is 1 or more groups of 5-byte sequence, each of which encodes up to 4 bytes in base85 encoding. Because 52 / 4 * 5 = 65 and we have the length byte, an output line is capped to 66 characters. The payload is the same diff-delta as we use in the packfiles. On the consumption side, git-apply now can decode and apply the binary patch when --allow-binary-replacement is given, the diff was generated with --full-index, and the receiving repository has the preimage blob, which is the same condition as it always required when accepting an "Binary files differ\n" patch. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-04 23:51:44 +00:00
else
fprintf(o->file, "%sBinary files %s and %s differ\n",
line_prefix, lbl[0], lbl[1]);
o->found_changes = 1;
} else {
/* Crazy xdl interfaces.. */
const char *diffopts = getenv("GIT_DIFF_OPTS");
xpparam_t xpp;
xdemitconf_t xecfg;
struct emit_callback ecbdata;
const struct userdiff_funcname *pe;
diff -p: squelch "diff --git" header for stat-dirty paths The plumbing "diff" commands look at the working tree files without refreshing the index themselves for performance reasons (the calling script is expected to do that upfront just once, before calling one or more of them). In the early days of git, they showed the "diff --git" header before they actually ask the xdiff machinery to produce patches, and ended up showing only these headers if the real contents are the same and the difference they noticed was only because the stat info cached in the index did not match that of the working tree. It was too late for the implementation to take the header that it already emitted back. But 3e97c7c (No diff -b/-w output for all-whitespace changes, 2009-11-19) introduced necessary logic to keep the meta-information headers in a strbuf and delay their output until the xdiff machinery noticed actual changes. This was primarily in order to generate patches that ignore whitespaces. When operating under "-w" mode, we wouldn't know if the header is needed until we actually look at the resulting patch, so it was a sensible thing to do, but we did not realize that the same reasoning applies to stat-dirty paths. Later, 296c6bb (diff: fix "git show -C -C" output when renaming a binary file, 2010-05-26) generalized this machinery and added must_show_header toggle. This is turned on when the header must be shown even when there is no patch to be produced, e.g. only the mode was changed, or the path was renamed, without changing the contents. However, when it did so, it still kept the special case for the "-w" mode, which meant that the plumbing would keep showing these phantom changes. This corrects this historical inconsistency by allowing the plumbing to omit paths that are only stat-dirty from its output in the same way as it handles whitespace only changes under "-w" option. The change in the behaviour can be seen in the updated test. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2012-03-01 02:14:16 +00:00
if (must_show_header) {
fprintf(o->file, "%s", header.buf);
strbuf_reset(&header);
}
mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
pe = diff_funcname_pattern(one);
if (!pe)
pe = diff_funcname_pattern(two);
memset(&xpp, 0, sizeof(xpp));
memset(&xecfg, 0, sizeof(xecfg));
memset(&ecbdata, 0, sizeof(ecbdata));
ecbdata.label_path = lbl;
color: delay auto-color decision until point of use When we read a color value either from a config file or from the command line, we use git_config_colorbool to convert it from the tristate always/never/auto into a single yes/no boolean value. This has some timing implications with respect to starting a pager. If we start (or decide not to start) the pager before checking the colorbool, everything is fine. Either isatty(1) will give us the right information, or we will properly check for pager_in_use(). However, if we decide to start a pager after we have checked the colorbool, things are not so simple. If stdout is a tty, then we will have already decided to use color. However, the user may also have configured color.pager not to use color with the pager. In this case, we need to actually turn off color. Unfortunately, the pager code has no idea which color variables were turned on (and there are many of them throughout the code, and they may even have been manipulated after the colorbool selection by something like "--color" on the command line). This bug can be seen any time a pager is started after config and command line options are checked. This has affected "git diff" since 89d07f7 (diff: don't run pager if user asked for a diff style exit code, 2007-08-12). It has also affect the log family since 1fda91b (Fix 'git log' early pager startup error case, 2010-08-24). This patch splits the notion of parsing a colorbool and actually checking the configuration. The "use_color" variables now have an additional possible value, GIT_COLOR_AUTO. Users of the variable should use the new "want_color()" wrapper, which will lazily determine and cache the auto-color decision. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-08-18 05:04:23 +00:00
ecbdata.color_diff = want_color(o->use_color);
ecbdata.found_changesp = &o->found_changes;
ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
check_blank_at_eof(&mf1, &mf2, &ecbdata);
ecbdata.opt = o;
ecbdata.header = header.len ? &header : NULL;
xpp.flags = o->xdl_opts;
xecfg.ctxlen = o->context;
xecfg.interhunkctxlen = o->interhunkcontext;
xecfg.flags = XDL_EMIT_FUNCNAMES;
if (DIFF_OPT_TST(o, FUNCCONTEXT))
xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
if (pe)
xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
if (!diffopts)
;
else if (!prefixcmp(diffopts, "--unified="))
xecfg.ctxlen = strtoul(diffopts + 10, NULL, 10);
else if (!prefixcmp(diffopts, "-u"))
xecfg.ctxlen = strtoul(diffopts + 2, NULL, 10);
if (o->word_diff) {
int i;
ecbdata.diff_words =
xcalloc(1, sizeof(struct diff_words_data));
ecbdata.diff_words->type = o->word_diff;
ecbdata.diff_words->opt = o;
if (!o->word_regex)
o->word_regex = userdiff_word_regex(one);
if (!o->word_regex)
o->word_regex = userdiff_word_regex(two);
if (!o->word_regex)
o->word_regex = diff_word_regex_cfg;
if (o->word_regex) {
ecbdata.diff_words->word_regex = (regex_t *)
xmalloc(sizeof(regex_t));
if (regcomp(ecbdata.diff_words->word_regex,
o->word_regex,
REG_EXTENDED | REG_NEWLINE))
die ("Invalid regular expression: %s",
o->word_regex);
}
for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
if (o->word_diff == diff_words_styles[i].type) {
ecbdata.diff_words->style =
&diff_words_styles[i];
break;
}
}
color: delay auto-color decision until point of use When we read a color value either from a config file or from the command line, we use git_config_colorbool to convert it from the tristate always/never/auto into a single yes/no boolean value. This has some timing implications with respect to starting a pager. If we start (or decide not to start) the pager before checking the colorbool, everything is fine. Either isatty(1) will give us the right information, or we will properly check for pager_in_use(). However, if we decide to start a pager after we have checked the colorbool, things are not so simple. If stdout is a tty, then we will have already decided to use color. However, the user may also have configured color.pager not to use color with the pager. In this case, we need to actually turn off color. Unfortunately, the pager code has no idea which color variables were turned on (and there are many of them throughout the code, and they may even have been manipulated after the colorbool selection by something like "--color" on the command line). This bug can be seen any time a pager is started after config and command line options are checked. This has affected "git diff" since 89d07f7 (diff: don't run pager if user asked for a diff style exit code, 2007-08-12). It has also affect the log family since 1fda91b (Fix 'git log' early pager startup error case, 2010-08-24). This patch splits the notion of parsing a colorbool and actually checking the configuration. The "use_color" variables now have an additional possible value, GIT_COLOR_AUTO. Users of the variable should use the new "want_color()" wrapper, which will lazily determine and cache the auto-color decision. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-08-18 05:04:23 +00:00
if (want_color(o->use_color)) {
struct diff_words_style *st = ecbdata.diff_words->style;
st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
st->ctx.color = diff_get_color_opt(o, DIFF_PLAIN);
}
}
xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
&xpp, &xecfg);
if (o->word_diff)
free_diff_words_data(&ecbdata);
if (textconv_one)
free(mf1.ptr);
if (textconv_two)
free(mf2.ptr);
xdiff_clear_find_func(&xecfg);
}
free_ab_and_return:
strbuf_release(&header);
diff_free_filespec_data(one);
diff_free_filespec_data(two);
free(a_one);
free(b_two);
return;
}
static void builtin_diffstat(const char *name_a, const char *name_b,
struct diff_filespec *one,
struct diff_filespec *two,
struct diffstat_t *diffstat,
struct diff_options *o,
int complete_rewrite)
{
mmfile_t mf1, mf2;
struct diffstat_file *data;
data = diffstat_add(diffstat, name_a, name_b);
if (!one || !two) {
data->is_unmerged = 1;
return;
}
if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
data->is_binary = 1;
data->added = diff_filespec_size(two);
data->deleted = diff_filespec_size(one);
}
else if (complete_rewrite) {
diff_populate_filespec(one, 0);
diff_populate_filespec(two, 0);
data->deleted = count_lines(one->data, one->size);
data->added = count_lines(two->data, two->size);
}
else {
/* Crazy xdl interfaces.. */
xpparam_t xpp;
xdemitconf_t xecfg;
if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
die("unable to read files to diff");
memset(&xpp, 0, sizeof(xpp));
memset(&xecfg, 0, sizeof(xecfg));
xpp.flags = o->xdl_opts;
xecfg.ctxlen = o->context;
xecfg.interhunkctxlen = o->interhunkcontext;
xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
&xpp, &xecfg);
}
diff_free_filespec_data(one);
diff_free_filespec_data(two);
}
static void builtin_checkdiff(const char *name_a, const char *name_b,
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
const char *attr_path,
struct diff_filespec *one,
struct diff_filespec *two,
struct diff_options *o)
{
mmfile_t mf1, mf2;
struct checkdiff_t data;
if (!two)
return;
memset(&data, 0, sizeof(data));
data.filename = name_b ? name_b : name_a;
data.lineno = 0;
data.o = o;
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
data.ws_rule = whitespace_rule(attr_path);
data.conflict_marker_size = ll_merge_marker_size(attr_path);
if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
die("unable to read files to diff");
/*
* All the other codepaths check both sides, but not checking
* the "old" side here is deliberate. We are checking the newly
* introduced changes, and as long as the "new" side is text, we
* can and should check what it introduces.
*/
if (diff_filespec_is_binary(two))
goto free_and_return;
else {
/* Crazy xdl interfaces.. */
xpparam_t xpp;
xdemitconf_t xecfg;
memset(&xpp, 0, sizeof(xpp));
memset(&xecfg, 0, sizeof(xecfg));
xecfg.ctxlen = 1; /* at least one context line */
xpp.flags = 0;
xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
&xpp, &xecfg);
if (data.ws_rule & WS_BLANK_AT_EOF) {
struct emit_callback ecbdata;
int blank_at_eof;
ecbdata.ws_rule = data.ws_rule;
check_blank_at_eof(&mf1, &mf2, &ecbdata);
blank_at_eof = ecbdata.blank_at_eof_in_postimage;
if (blank_at_eof) {
static char *err;
if (!err)
err = whitespace_error_string(WS_BLANK_AT_EOF);
fprintf(o->file, "%s:%d: %s.\n",
data.filename, blank_at_eof, err);
data.status = 1; /* report errors */
}
}
}
free_and_return:
diff_free_filespec_data(one);
diff_free_filespec_data(two);
if (data.status)
DIFF_OPT_SET(o, CHECK_FAILED);
}
struct diff_filespec *alloc_filespec(const char *path)
{
int namelen = strlen(path);
struct diff_filespec *spec = xmalloc(sizeof(*spec) + namelen + 1);
memset(spec, 0, sizeof(*spec));
spec->path = (char *)(spec + 1);
memcpy(spec->path, path, namelen+1);
spec->count = 1;
diff: introduce diff.<driver>.binary The "diff" gitattribute is somewhat overloaded right now. It can say one of three things: 1. this file is definitely binary, or definitely not (i.e., diff or !diff) 2. this file should use an external diff engine (i.e., diff=foo, diff.foo.command = custom-script) 3. this file should use particular funcname patterns (i.e., diff=foo, diff.foo.(x?)funcname = some-regex) Most of the time, there is no conflict between these uses, since using one implies that the other is irrelevant (e.g., an external diff engine will decide for itself whether the file is binary). However, there is at least one conflicting situation: there is no way to say "use the regular rules to determine whether this file is binary, but if we do diff it textually, use this funcname pattern." That is, currently setting diff=foo indicates that the file is definitely text. This patch introduces a "binary" config option for a diff driver, so that one can explicitly set diff.foo.binary. We default this value to "don't know". That is, setting a diff attribute to "foo" and using "diff.foo.funcname" will have no effect on the binaryness of a file. To get the current behavior, one can set diff.foo.binary to true. This patch also has one additional advantage: it cleans up the interface to the userdiff code a bit. Before, calling code had to know more about whether attributes were false, true, or unset to determine binaryness. Now that binaryness is a property of a driver, we can represent these situations just by passing back a driver struct. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2008-10-05 21:43:36 +00:00
spec->is_binary = -1;
return spec;
}
void free_filespec(struct diff_filespec *spec)
{
if (!--spec->count) {
diff_free_filespec_data(spec);
free(spec);
}
}
void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
unsigned short mode)
{
if (mode) {
spec->mode = canon_mode(mode);
hashcpy(spec->sha1, sha1);
spec->sha1_valid = !is_null_sha1(sha1);
}
}
/*
* Given a name and sha1 pair, if the index tells us the file in
* the work tree has that object contents, return true, so that
* prepare_temp_file() does not have to inflate and extract.
*/
Avoid accessing a slow working copy during diffcore operations. The Cygwin folks have done a fine job at creating a POSIX layer on Windows That Just Works(tm). However it comes with a penalty; accessing files in the working tree by way of stat/open/mmap can be slower for diffcore than inflating the data from a blob which is stored in a packfile. This performance problem is especially an issue in merge-recursive when dealing with nearly 7000 added files, as we are loading each file's content from the working directory to perform rename detection. I have literally seen (and sadly watched) paint dry in less time than it takes for merge-recursive to finish such a merge. On the other hand this very same merge runs very fast on Solaris. If Git is compiled with NO_FAST_WORKING_DIRECTORY set then we will avoid looking at the working directory when the blob in question is available within a packfile and the caller doesn't need the data unpacked into a temporary file. We don't use loose objects as they have the same open/mmap/close costs as the working directory file access, but have the additional CPU overhead of needing to inflate the content before use. So it is still faster to use the working tree file over the loose object. If the caller needs the file data unpacked into a temporary file its likely because they are going to call an external diff program, passing the file as a parameter. In this case reusing the working tree file will be faster as we don't need to inflate the data and write it out to a temporary file. The NO_FAST_WORKING_DIRECTORY feature is enabled by default on Cygwin, as that is the platform which currently appears to benefit the most from this option. Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-12-14 11:15:57 +00:00
static int reuse_worktree_file(const char *name, const unsigned char *sha1, int want_file)
{
struct cache_entry *ce;
struct stat st;
int pos, len;
/*
* We do not read the cache ourselves here, because the
* benchmark with my previous version that always reads cache
* shows that it makes things worse for diff-tree comparing
* two linux-2.6 kernel trees in an already checked out work
* tree. This is because most diff-tree comparisons deal with
* only a small number of files, while reading the cache is
* expensive for a large project, and its cost outweighs the
* savings we get by not inflating the object to a temporary
* file. Practically, this code only helps when we are used
* by diff-cache --cached, which does read the cache before
* calling us.
*/
if (!active_cache)
return 0;
Avoid accessing a slow working copy during diffcore operations. The Cygwin folks have done a fine job at creating a POSIX layer on Windows That Just Works(tm). However it comes with a penalty; accessing files in the working tree by way of stat/open/mmap can be slower for diffcore than inflating the data from a blob which is stored in a packfile. This performance problem is especially an issue in merge-recursive when dealing with nearly 7000 added files, as we are loading each file's content from the working directory to perform rename detection. I have literally seen (and sadly watched) paint dry in less time than it takes for merge-recursive to finish such a merge. On the other hand this very same merge runs very fast on Solaris. If Git is compiled with NO_FAST_WORKING_DIRECTORY set then we will avoid looking at the working directory when the blob in question is available within a packfile and the caller doesn't need the data unpacked into a temporary file. We don't use loose objects as they have the same open/mmap/close costs as the working directory file access, but have the additional CPU overhead of needing to inflate the content before use. So it is still faster to use the working tree file over the loose object. If the caller needs the file data unpacked into a temporary file its likely because they are going to call an external diff program, passing the file as a parameter. In this case reusing the working tree file will be faster as we don't need to inflate the data and write it out to a temporary file. The NO_FAST_WORKING_DIRECTORY feature is enabled by default on Cygwin, as that is the platform which currently appears to benefit the most from this option. Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-12-14 11:15:57 +00:00
/* We want to avoid the working directory if our caller
* doesn't need the data in a normal file, this system
* is rather slow with its stat/open/mmap/close syscalls,
* and the object is contained in a pack file. The pack
* is probably already open and will be faster to obtain
* the data through than the working directory. Loose
* objects however would tend to be slower as they need
* to be individually opened and inflated.
*/
if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(sha1))
Avoid accessing a slow working copy during diffcore operations. The Cygwin folks have done a fine job at creating a POSIX layer on Windows That Just Works(tm). However it comes with a penalty; accessing files in the working tree by way of stat/open/mmap can be slower for diffcore than inflating the data from a blob which is stored in a packfile. This performance problem is especially an issue in merge-recursive when dealing with nearly 7000 added files, as we are loading each file's content from the working directory to perform rename detection. I have literally seen (and sadly watched) paint dry in less time than it takes for merge-recursive to finish such a merge. On the other hand this very same merge runs very fast on Solaris. If Git is compiled with NO_FAST_WORKING_DIRECTORY set then we will avoid looking at the working directory when the blob in question is available within a packfile and the caller doesn't need the data unpacked into a temporary file. We don't use loose objects as they have the same open/mmap/close costs as the working directory file access, but have the additional CPU overhead of needing to inflate the content before use. So it is still faster to use the working tree file over the loose object. If the caller needs the file data unpacked into a temporary file its likely because they are going to call an external diff program, passing the file as a parameter. In this case reusing the working tree file will be faster as we don't need to inflate the data and write it out to a temporary file. The NO_FAST_WORKING_DIRECTORY feature is enabled by default on Cygwin, as that is the platform which currently appears to benefit the most from this option. Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-12-14 11:15:57 +00:00
return 0;
len = strlen(name);
pos = cache_name_pos(name, len);
if (pos < 0)
return 0;
ce = active_cache[pos];
/*
* This is not the sha1 we are looking for, or
* unreusable because it is not a regular file.
*/
if (hashcmp(sha1, ce->sha1) || !S_ISREG(ce->ce_mode))
return 0;
/*
* If ce is marked as "assume unchanged", there is no
* guarantee that work tree matches what we are looking for.
*/
if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
return 0;
/*
* If ce matches the file in the work tree, we can reuse it.
*/
if (ce_uptodate(ce) ||
(!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
return 1;
return 0;
}
static int populate_from_stdin(struct diff_filespec *s)
{
struct strbuf buf = STRBUF_INIT;
size_t size = 0;
if (strbuf_read(&buf, 0, 0) < 0)
return error("error while reading from stdin %s",
strerror(errno));
s->should_munmap = 0;
s->data = strbuf_detach(&buf, &size);
s->size = size;
s->should_free = 1;
return 0;
}
static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
{
int len;
char *data = xmalloc(100), *dirty = "";
/* Are we looking at the work tree? */
if (s->dirty_submodule)
dirty = "-dirty";
len = snprintf(data, 100,
"Subproject commit %s%s\n", sha1_to_hex(s->sha1), dirty);
s->data = data;
s->size = len;
s->should_free = 1;
if (size_only) {
s->data = NULL;
free(data);
}
return 0;
}
/*
* While doing rename detection and pickaxe operation, we may need to
* grab the data for the blob (or file) for our own in-core comparison.
* diff_filespec has data and size fields for this purpose.
*/
int diff_populate_filespec(struct diff_filespec *s, int size_only)
{
int err = 0;
if (!DIFF_FILE_VALID(s))
die("internal error: asking to populate invalid file.");
if (S_ISDIR(s->mode))
return -1;
if (s->data)
return 0;
if (size_only && 0 < s->size)
return 0;
if (S_ISGITLINK(s->mode))
return diff_populate_gitlink(s, size_only);
if (!s->sha1_valid ||
Avoid accessing a slow working copy during diffcore operations. The Cygwin folks have done a fine job at creating a POSIX layer on Windows That Just Works(tm). However it comes with a penalty; accessing files in the working tree by way of stat/open/mmap can be slower for diffcore than inflating the data from a blob which is stored in a packfile. This performance problem is especially an issue in merge-recursive when dealing with nearly 7000 added files, as we are loading each file's content from the working directory to perform rename detection. I have literally seen (and sadly watched) paint dry in less time than it takes for merge-recursive to finish such a merge. On the other hand this very same merge runs very fast on Solaris. If Git is compiled with NO_FAST_WORKING_DIRECTORY set then we will avoid looking at the working directory when the blob in question is available within a packfile and the caller doesn't need the data unpacked into a temporary file. We don't use loose objects as they have the same open/mmap/close costs as the working directory file access, but have the additional CPU overhead of needing to inflate the content before use. So it is still faster to use the working tree file over the loose object. If the caller needs the file data unpacked into a temporary file its likely because they are going to call an external diff program, passing the file as a parameter. In this case reusing the working tree file will be faster as we don't need to inflate the data and write it out to a temporary file. The NO_FAST_WORKING_DIRECTORY feature is enabled by default on Cygwin, as that is the platform which currently appears to benefit the most from this option. Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-12-14 11:15:57 +00:00
reuse_worktree_file(s->path, s->sha1, 0)) {
struct strbuf buf = STRBUF_INIT;
struct stat st;
int fd;
Lazy man's auto-CRLF It currently does NOT know about file attributes, so it does its conversion purely based on content. Maybe that is more in the "git philosophy" anyway, since content is king, but I think we should try to do the file attributes to turn it off on demand. Anyway, BY DEFAULT it is off regardless, because it requires a [core] AutoCRLF = true in your config file to be enabled. We could make that the default for Windows, of course, the same way we do some other things (filemode etc). But you can actually enable it on UNIX, and it will cause: - "git update-index" will write blobs without CRLF - "git diff" will diff working tree files without CRLF - "git checkout" will write files to the working tree _with_ CRLF and things work fine. Funnily, it actually shows an odd file in git itself: git clone -n git test-crlf cd test-crlf git config core.autocrlf true git checkout git diff shows a diff for "Documentation/docbook-xsl.css". Why? Because we have actually checked in that file *with* CRLF! So when "core.autocrlf" is true, we'll always generate a *different* hash for it in the index, because the index hash will be for the content _without_ CRLF. Is this complete? I dunno. It seems to work for me. It doesn't use the filename at all right now, and that's probably a deficiency (we could certainly make the "is_binary()" heuristics also take standard filename heuristics into account). I don't pass in the filename at all for the "index_fd()" case (git-update-index), so that would need to be passed around, but this actually works fine. NOTE NOTE NOTE! The "is_binary()" heuristics are totally made-up by yours truly. I will not guarantee that they work at all reasonable. Caveat emptor. But it _is_ simple, and it _is_ safe, since it's all off by default. The patch is pretty simple - the biggest part is the new "convert.c" file, but even that is really just basic stuff that anybody can write in "Teaching C 101" as a final project for their first class in programming. Not to say that it's bug-free, of course - but at least we're not talking about rocket surgery here. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2007-02-13 19:07:23 +00:00
if (!strcmp(s->path, "-"))
return populate_from_stdin(s);
if (lstat(s->path, &st) < 0) {
if (errno == ENOENT) {
err_empty:
err = -1;
empty:
s->data = (char *)"";
s->size = 0;
return err;
}
}
s->size = xsize_t(st.st_size);
if (!s->size)
goto empty;
if (S_ISLNK(st.st_mode)) {
struct strbuf sb = STRBUF_INIT;
if (strbuf_readlink(&sb, s->path, s->size))
goto err_empty;
s->size = sb.len;
s->data = strbuf_detach(&sb, NULL);
s->should_free = 1;
return 0;
}
if (size_only)
return 0;
fd = open(s->path, O_RDONLY);
if (fd < 0)
goto err_empty;
s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
close(fd);
s->should_munmap = 1;
Lazy man's auto-CRLF It currently does NOT know about file attributes, so it does its conversion purely based on content. Maybe that is more in the "git philosophy" anyway, since content is king, but I think we should try to do the file attributes to turn it off on demand. Anyway, BY DEFAULT it is off regardless, because it requires a [core] AutoCRLF = true in your config file to be enabled. We could make that the default for Windows, of course, the same way we do some other things (filemode etc). But you can actually enable it on UNIX, and it will cause: - "git update-index" will write blobs without CRLF - "git diff" will diff working tree files without CRLF - "git checkout" will write files to the working tree _with_ CRLF and things work fine. Funnily, it actually shows an odd file in git itself: git clone -n git test-crlf cd test-crlf git config core.autocrlf true git checkout git diff shows a diff for "Documentation/docbook-xsl.css". Why? Because we have actually checked in that file *with* CRLF! So when "core.autocrlf" is true, we'll always generate a *different* hash for it in the index, because the index hash will be for the content _without_ CRLF. Is this complete? I dunno. It seems to work for me. It doesn't use the filename at all right now, and that's probably a deficiency (we could certainly make the "is_binary()" heuristics also take standard filename heuristics into account). I don't pass in the filename at all for the "index_fd()" case (git-update-index), so that would need to be passed around, but this actually works fine. NOTE NOTE NOTE! The "is_binary()" heuristics are totally made-up by yours truly. I will not guarantee that they work at all reasonable. Caveat emptor. But it _is_ simple, and it _is_ safe, since it's all off by default. The patch is pretty simple - the biggest part is the new "convert.c" file, but even that is really just basic stuff that anybody can write in "Teaching C 101" as a final project for their first class in programming. Not to say that it's bug-free, of course - but at least we're not talking about rocket surgery here. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2007-02-13 19:07:23 +00:00
/*
* Convert from working tree format to canonical git format
*/
safecrlf: Add mechanism to warn about irreversible crlf conversions CRLF conversion bears a slight chance of corrupting data. autocrlf=true will convert CRLF to LF during commit and LF to CRLF during checkout. A file that contains a mixture of LF and CRLF before the commit cannot be recreated by git. For text files this is the right thing to do: it corrects line endings such that we have only LF line endings in the repository. But for binary files that are accidentally classified as text the conversion can corrupt data. If you recognize such corruption early you can easily fix it by setting the conversion type explicitly in .gitattributes. Right after committing you still have the original file in your work tree and this file is not yet corrupted. You can explicitly tell git that this file is binary and git will handle the file appropriately. Unfortunately, the desired effect of cleaning up text files with mixed line endings and the undesired effect of corrupting binary files cannot be distinguished. In both cases CRLFs are removed in an irreversible way. For text files this is the right thing to do because CRLFs are line endings, while for binary files converting CRLFs corrupts data. This patch adds a mechanism that can either warn the user about an irreversible conversion or can even refuse to convert. The mechanism is controlled by the variable core.safecrlf, with the following values: - false: disable safecrlf mechanism - warn: warn about irreversible conversions - true: refuse irreversible conversions The default is to warn. Users are only affected by this default if core.autocrlf is set. But the current default of git is to leave core.autocrlf unset, so users will not see warnings unless they deliberately chose to activate the autocrlf mechanism. The safecrlf mechanism's details depend on the git command. The general principles when safecrlf is active (not false) are: - we warn/error out if files in the work tree can modified in an irreversible way without giving the user a chance to backup the original file. - for read-only operations that do not modify files in the work tree we do not not print annoying warnings. There are exceptions. Even though... - "git add" itself does not touch the files in the work tree, the next checkout would, so the safety triggers; - "git apply" to update a text file with a patch does touch the files in the work tree, but the operation is about text files and CRLF conversion is about fixing the line ending inconsistencies, so the safety does not trigger; - "git diff" itself does not touch the files in the work tree, it is often run to inspect the changes you intend to next "git add". To catch potential problems early, safety triggers. The concept of a safety check was originally proposed in a similar way by Linus Torvalds. Thanks to Dimitry Potapov for insisting on getting the naked LF/autocrlf=true case right. Signed-off-by: Steffen Prohaska <prohaska@zib.de>
2008-02-06 11:25:58 +00:00
if (convert_to_git(s->path, s->data, s->size, &buf, safe_crlf)) {
size_t size = 0;
Lazy man's auto-CRLF It currently does NOT know about file attributes, so it does its conversion purely based on content. Maybe that is more in the "git philosophy" anyway, since content is king, but I think we should try to do the file attributes to turn it off on demand. Anyway, BY DEFAULT it is off regardless, because it requires a [core] AutoCRLF = true in your config file to be enabled. We could make that the default for Windows, of course, the same way we do some other things (filemode etc). But you can actually enable it on UNIX, and it will cause: - "git update-index" will write blobs without CRLF - "git diff" will diff working tree files without CRLF - "git checkout" will write files to the working tree _with_ CRLF and things work fine. Funnily, it actually shows an odd file in git itself: git clone -n git test-crlf cd test-crlf git config core.autocrlf true git checkout git diff shows a diff for "Documentation/docbook-xsl.css". Why? Because we have actually checked in that file *with* CRLF! So when "core.autocrlf" is true, we'll always generate a *different* hash for it in the index, because the index hash will be for the content _without_ CRLF. Is this complete? I dunno. It seems to work for me. It doesn't use the filename at all right now, and that's probably a deficiency (we could certainly make the "is_binary()" heuristics also take standard filename heuristics into account). I don't pass in the filename at all for the "index_fd()" case (git-update-index), so that would need to be passed around, but this actually works fine. NOTE NOTE NOTE! The "is_binary()" heuristics are totally made-up by yours truly. I will not guarantee that they work at all reasonable. Caveat emptor. But it _is_ simple, and it _is_ safe, since it's all off by default. The patch is pretty simple - the biggest part is the new "convert.c" file, but even that is really just basic stuff that anybody can write in "Teaching C 101" as a final project for their first class in programming. Not to say that it's bug-free, of course - but at least we're not talking about rocket surgery here. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2007-02-13 19:07:23 +00:00
munmap(s->data, s->size);
s->should_munmap = 0;
s->data = strbuf_detach(&buf, &size);
s->size = size;
Lazy man's auto-CRLF It currently does NOT know about file attributes, so it does its conversion purely based on content. Maybe that is more in the "git philosophy" anyway, since content is king, but I think we should try to do the file attributes to turn it off on demand. Anyway, BY DEFAULT it is off regardless, because it requires a [core] AutoCRLF = true in your config file to be enabled. We could make that the default for Windows, of course, the same way we do some other things (filemode etc). But you can actually enable it on UNIX, and it will cause: - "git update-index" will write blobs without CRLF - "git diff" will diff working tree files without CRLF - "git checkout" will write files to the working tree _with_ CRLF and things work fine. Funnily, it actually shows an odd file in git itself: git clone -n git test-crlf cd test-crlf git config core.autocrlf true git checkout git diff shows a diff for "Documentation/docbook-xsl.css". Why? Because we have actually checked in that file *with* CRLF! So when "core.autocrlf" is true, we'll always generate a *different* hash for it in the index, because the index hash will be for the content _without_ CRLF. Is this complete? I dunno. It seems to work for me. It doesn't use the filename at all right now, and that's probably a deficiency (we could certainly make the "is_binary()" heuristics also take standard filename heuristics into account). I don't pass in the filename at all for the "index_fd()" case (git-update-index), so that would need to be passed around, but this actually works fine. NOTE NOTE NOTE! The "is_binary()" heuristics are totally made-up by yours truly. I will not guarantee that they work at all reasonable. Caveat emptor. But it _is_ simple, and it _is_ safe, since it's all off by default. The patch is pretty simple - the biggest part is the new "convert.c" file, but even that is really just basic stuff that anybody can write in "Teaching C 101" as a final project for their first class in programming. Not to say that it's bug-free, of course - but at least we're not talking about rocket surgery here. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2007-02-13 19:07:23 +00:00
s->should_free = 1;
}
}
else {
enum object_type type;
if (size_only) {
type = sha1_object_info(s->sha1, &s->size);
if (type < 0)
die("unable to read %s", sha1_to_hex(s->sha1));
} else {
s->data = read_sha1_file(s->sha1, &type, &s->size);
if (!s->data)
die("unable to read %s", sha1_to_hex(s->sha1));
s->should_free = 1;
}
}
return 0;
}
void diff_free_filespec_blob(struct diff_filespec *s)
{
if (s->should_free)
free(s->data);
else if (s->should_munmap)
munmap(s->data, s->size);
if (s->should_free || s->should_munmap) {
s->should_free = s->should_munmap = 0;
s->data = NULL;
}
}
void diff_free_filespec_data(struct diff_filespec *s)
{
diff_free_filespec_blob(s);
free(s->cnt_data);
s->cnt_data = NULL;
}
static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
void *blob,
unsigned long size,
const unsigned char *sha1,
int mode)
{
int fd;
struct strbuf buf = STRBUF_INIT;
struct strbuf template = STRBUF_INIT;
char *path_dup = xstrdup(path);
const char *base = basename(path_dup);
/* Generate "XXXXXX_basename.ext" */
strbuf_addstr(&template, "XXXXXX_");
strbuf_addstr(&template, base);
fd = git_mkstemps(temp->tmp_path, PATH_MAX, template.buf,
strlen(base) + 1);
if (fd < 0)
die_errno("unable to create temp-file");
if (convert_to_working_tree(path,
(const char *)blob, (size_t)size, &buf)) {
blob = buf.buf;
size = buf.len;
}
if (write_in_full(fd, blob, size) != size)
die_errno("unable to write temp-file");
close(fd);
temp->name = temp->tmp_path;
strcpy(temp->hex, sha1_to_hex(sha1));
temp->hex[40] = 0;
sprintf(temp->mode, "%06o", mode);
strbuf_release(&buf);
strbuf_release(&template);
free(path_dup);
}
diff: refactor tempfile cleanup handling There are two pieces of code that create tempfiles for diff: run_external_diff and run_textconv. The former cleans up its tempfiles in the face of premature death (i.e., by die() or by signal), but the latter does not. After this patch, they will both use the same cleanup routines. To make clear what the change is, let me first explain what happens now: - run_external_diff uses a static global array of 2 diff_tempfile structs (since it knows it will always need exactly 2 tempfiles). It calls prepare_temp_file (which doesn't know anything about the global array) on each of the structs, creating the tempfiles that need to be cleaned up. It then registers atexit and signal handlers to look through the global array and remove the tempfiles. If it succeeds, it calls the handler manually (which marks the tempfile structs as unused). - textconv has its own tempfile struct, which it allocates using prepare_temp_file and cleans up manually. No signal or atexit handlers. The new code moves the installation of cleanup handlers into the prepare_temp_file function. Which means that that function now has to understand that there is static tempfile storage. So what happens now is: - run_external_diff calls prepare_temp_file - prepare_temp_file calls claim_diff_tempfile, which allocates an unused slot from our global array - prepare_temp_file installs (if they have not already been installed) atexit and signal handlers for cleanup - prepare_temp_file sets up the tempfile as usual - prepare_temp_file returns a pointer to the allocated tempfile The advantage being that run_external_diff no longer has to care about setting up cleanup handlers. Now by virtue of calling prepare_temp_file, run_textconv gets the same benefit, as will any future users of prepare_temp_file. There are also a few side benefits to the specific implementation: - we now install cleanup handlers _before_ allocating the tempfile, closing a race which could leave temp cruft - when allocating a slot in the global array, we will now detect a situation where the old slots were not properly vacated (i.e., somebody forgot to call remove upon leaving the function). In the old code, such a situation would silently overwrite the tempfile names, meaning we would forget to clean them up. The new code dies with a bug warning. - we make sure only to install the signal handler once. This isn't a big deal, since we are just overwriting the old handler, but will become an issue when a later patch converts the code to use sigchain Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-01-22 05:59:56 +00:00
static struct diff_tempfile *prepare_temp_file(const char *name,
struct diff_filespec *one)
{
diff: refactor tempfile cleanup handling There are two pieces of code that create tempfiles for diff: run_external_diff and run_textconv. The former cleans up its tempfiles in the face of premature death (i.e., by die() or by signal), but the latter does not. After this patch, they will both use the same cleanup routines. To make clear what the change is, let me first explain what happens now: - run_external_diff uses a static global array of 2 diff_tempfile structs (since it knows it will always need exactly 2 tempfiles). It calls prepare_temp_file (which doesn't know anything about the global array) on each of the structs, creating the tempfiles that need to be cleaned up. It then registers atexit and signal handlers to look through the global array and remove the tempfiles. If it succeeds, it calls the handler manually (which marks the tempfile structs as unused). - textconv has its own tempfile struct, which it allocates using prepare_temp_file and cleans up manually. No signal or atexit handlers. The new code moves the installation of cleanup handlers into the prepare_temp_file function. Which means that that function now has to understand that there is static tempfile storage. So what happens now is: - run_external_diff calls prepare_temp_file - prepare_temp_file calls claim_diff_tempfile, which allocates an unused slot from our global array - prepare_temp_file installs (if they have not already been installed) atexit and signal handlers for cleanup - prepare_temp_file sets up the tempfile as usual - prepare_temp_file returns a pointer to the allocated tempfile The advantage being that run_external_diff no longer has to care about setting up cleanup handlers. Now by virtue of calling prepare_temp_file, run_textconv gets the same benefit, as will any future users of prepare_temp_file. There are also a few side benefits to the specific implementation: - we now install cleanup handlers _before_ allocating the tempfile, closing a race which could leave temp cruft - when allocating a slot in the global array, we will now detect a situation where the old slots were not properly vacated (i.e., somebody forgot to call remove upon leaving the function). In the old code, such a situation would silently overwrite the tempfile names, meaning we would forget to clean them up. The new code dies with a bug warning. - we make sure only to install the signal handler once. This isn't a big deal, since we are just overwriting the old handler, but will become an issue when a later patch converts the code to use sigchain Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-01-22 05:59:56 +00:00
struct diff_tempfile *temp = claim_diff_tempfile();
if (!DIFF_FILE_VALID(one)) {
not_a_valid_file:
/* A '-' entry produces this for file-2, and
* a '+' entry produces this for file-1.
*/
temp->name = "/dev/null";
strcpy(temp->hex, ".");
strcpy(temp->mode, ".");
diff: refactor tempfile cleanup handling There are two pieces of code that create tempfiles for diff: run_external_diff and run_textconv. The former cleans up its tempfiles in the face of premature death (i.e., by die() or by signal), but the latter does not. After this patch, they will both use the same cleanup routines. To make clear what the change is, let me first explain what happens now: - run_external_diff uses a static global array of 2 diff_tempfile structs (since it knows it will always need exactly 2 tempfiles). It calls prepare_temp_file (which doesn't know anything about the global array) on each of the structs, creating the tempfiles that need to be cleaned up. It then registers atexit and signal handlers to look through the global array and remove the tempfiles. If it succeeds, it calls the handler manually (which marks the tempfile structs as unused). - textconv has its own tempfile struct, which it allocates using prepare_temp_file and cleans up manually. No signal or atexit handlers. The new code moves the installation of cleanup handlers into the prepare_temp_file function. Which means that that function now has to understand that there is static tempfile storage. So what happens now is: - run_external_diff calls prepare_temp_file - prepare_temp_file calls claim_diff_tempfile, which allocates an unused slot from our global array - prepare_temp_file installs (if they have not already been installed) atexit and signal handlers for cleanup - prepare_temp_file sets up the tempfile as usual - prepare_temp_file returns a pointer to the allocated tempfile The advantage being that run_external_diff no longer has to care about setting up cleanup handlers. Now by virtue of calling prepare_temp_file, run_textconv gets the same benefit, as will any future users of prepare_temp_file. There are also a few side benefits to the specific implementation: - we now install cleanup handlers _before_ allocating the tempfile, closing a race which could leave temp cruft - when allocating a slot in the global array, we will now detect a situation where the old slots were not properly vacated (i.e., somebody forgot to call remove upon leaving the function). In the old code, such a situation would silently overwrite the tempfile names, meaning we would forget to clean them up. The new code dies with a bug warning. - we make sure only to install the signal handler once. This isn't a big deal, since we are just overwriting the old handler, but will become an issue when a later patch converts the code to use sigchain Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-01-22 05:59:56 +00:00
return temp;
}
if (!remove_tempfile_installed) {
atexit(remove_tempfile);
sigchain_push_common(remove_tempfile_on_signal);
diff: refactor tempfile cleanup handling There are two pieces of code that create tempfiles for diff: run_external_diff and run_textconv. The former cleans up its tempfiles in the face of premature death (i.e., by die() or by signal), but the latter does not. After this patch, they will both use the same cleanup routines. To make clear what the change is, let me first explain what happens now: - run_external_diff uses a static global array of 2 diff_tempfile structs (since it knows it will always need exactly 2 tempfiles). It calls prepare_temp_file (which doesn't know anything about the global array) on each of the structs, creating the tempfiles that need to be cleaned up. It then registers atexit and signal handlers to look through the global array and remove the tempfiles. If it succeeds, it calls the handler manually (which marks the tempfile structs as unused). - textconv has its own tempfile struct, which it allocates using prepare_temp_file and cleans up manually. No signal or atexit handlers. The new code moves the installation of cleanup handlers into the prepare_temp_file function. Which means that that function now has to understand that there is static tempfile storage. So what happens now is: - run_external_diff calls prepare_temp_file - prepare_temp_file calls claim_diff_tempfile, which allocates an unused slot from our global array - prepare_temp_file installs (if they have not already been installed) atexit and signal handlers for cleanup - prepare_temp_file sets up the tempfile as usual - prepare_temp_file returns a pointer to the allocated tempfile The advantage being that run_external_diff no longer has to care about setting up cleanup handlers. Now by virtue of calling prepare_temp_file, run_textconv gets the same benefit, as will any future users of prepare_temp_file. There are also a few side benefits to the specific implementation: - we now install cleanup handlers _before_ allocating the tempfile, closing a race which could leave temp cruft - when allocating a slot in the global array, we will now detect a situation where the old slots were not properly vacated (i.e., somebody forgot to call remove upon leaving the function). In the old code, such a situation would silently overwrite the tempfile names, meaning we would forget to clean them up. The new code dies with a bug warning. - we make sure only to install the signal handler once. This isn't a big deal, since we are just overwriting the old handler, but will become an issue when a later patch converts the code to use sigchain Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-01-22 05:59:56 +00:00
remove_tempfile_installed = 1;
}
if (!one->sha1_valid ||
Avoid accessing a slow working copy during diffcore operations. The Cygwin folks have done a fine job at creating a POSIX layer on Windows That Just Works(tm). However it comes with a penalty; accessing files in the working tree by way of stat/open/mmap can be slower for diffcore than inflating the data from a blob which is stored in a packfile. This performance problem is especially an issue in merge-recursive when dealing with nearly 7000 added files, as we are loading each file's content from the working directory to perform rename detection. I have literally seen (and sadly watched) paint dry in less time than it takes for merge-recursive to finish such a merge. On the other hand this very same merge runs very fast on Solaris. If Git is compiled with NO_FAST_WORKING_DIRECTORY set then we will avoid looking at the working directory when the blob in question is available within a packfile and the caller doesn't need the data unpacked into a temporary file. We don't use loose objects as they have the same open/mmap/close costs as the working directory file access, but have the additional CPU overhead of needing to inflate the content before use. So it is still faster to use the working tree file over the loose object. If the caller needs the file data unpacked into a temporary file its likely because they are going to call an external diff program, passing the file as a parameter. In this case reusing the working tree file will be faster as we don't need to inflate the data and write it out to a temporary file. The NO_FAST_WORKING_DIRECTORY feature is enabled by default on Cygwin, as that is the platform which currently appears to benefit the most from this option. Signed-off-by: Shawn O. Pearce <spearce@spearce.org> Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-12-14 11:15:57 +00:00
reuse_worktree_file(name, one->sha1, 1)) {
struct stat st;
if (lstat(name, &st) < 0) {
if (errno == ENOENT)
goto not_a_valid_file;
die_errno("stat(%s)", name);
}
if (S_ISLNK(st.st_mode)) {
struct strbuf sb = STRBUF_INIT;
if (strbuf_readlink(&sb, name, st.st_size) < 0)
die_errno("readlink(%s)", name);
prep_temp_blob(name, temp, sb.buf, sb.len,
(one->sha1_valid ?
one->sha1 : null_sha1),
(one->sha1_valid ?
one->mode : S_IFLNK));
strbuf_release(&sb);
}
else {
/* we can borrow from the file in the work tree */
temp->name = name;
if (!one->sha1_valid)
strcpy(temp->hex, sha1_to_hex(null_sha1));
else
strcpy(temp->hex, sha1_to_hex(one->sha1));
/* Even though we may sometimes borrow the
* contents from the work tree, we always want
* one->mode. mode is trustworthy even when
* !(one->sha1_valid), as long as
* DIFF_FILE_VALID(one).
*/
sprintf(temp->mode, "%06o", one->mode);
}
diff: refactor tempfile cleanup handling There are two pieces of code that create tempfiles for diff: run_external_diff and run_textconv. The former cleans up its tempfiles in the face of premature death (i.e., by die() or by signal), but the latter does not. After this patch, they will both use the same cleanup routines. To make clear what the change is, let me first explain what happens now: - run_external_diff uses a static global array of 2 diff_tempfile structs (since it knows it will always need exactly 2 tempfiles). It calls prepare_temp_file (which doesn't know anything about the global array) on each of the structs, creating the tempfiles that need to be cleaned up. It then registers atexit and signal handlers to look through the global array and remove the tempfiles. If it succeeds, it calls the handler manually (which marks the tempfile structs as unused). - textconv has its own tempfile struct, which it allocates using prepare_temp_file and cleans up manually. No signal or atexit handlers. The new code moves the installation of cleanup handlers into the prepare_temp_file function. Which means that that function now has to understand that there is static tempfile storage. So what happens now is: - run_external_diff calls prepare_temp_file - prepare_temp_file calls claim_diff_tempfile, which allocates an unused slot from our global array - prepare_temp_file installs (if they have not already been installed) atexit and signal handlers for cleanup - prepare_temp_file sets up the tempfile as usual - prepare_temp_file returns a pointer to the allocated tempfile The advantage being that run_external_diff no longer has to care about setting up cleanup handlers. Now by virtue of calling prepare_temp_file, run_textconv gets the same benefit, as will any future users of prepare_temp_file. There are also a few side benefits to the specific implementation: - we now install cleanup handlers _before_ allocating the tempfile, closing a race which could leave temp cruft - when allocating a slot in the global array, we will now detect a situation where the old slots were not properly vacated (i.e., somebody forgot to call remove upon leaving the function). In the old code, such a situation would silently overwrite the tempfile names, meaning we would forget to clean them up. The new code dies with a bug warning. - we make sure only to install the signal handler once. This isn't a big deal, since we are just overwriting the old handler, but will become an issue when a later patch converts the code to use sigchain Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-01-22 05:59:56 +00:00
return temp;
}
else {
if (diff_populate_filespec(one, 0))
die("cannot read data blob for %s", one->path);
prep_temp_blob(name, temp, one->data, one->size,
one->sha1, one->mode);
}
diff: refactor tempfile cleanup handling There are two pieces of code that create tempfiles for diff: run_external_diff and run_textconv. The former cleans up its tempfiles in the face of premature death (i.e., by die() or by signal), but the latter does not. After this patch, they will both use the same cleanup routines. To make clear what the change is, let me first explain what happens now: - run_external_diff uses a static global array of 2 diff_tempfile structs (since it knows it will always need exactly 2 tempfiles). It calls prepare_temp_file (which doesn't know anything about the global array) on each of the structs, creating the tempfiles that need to be cleaned up. It then registers atexit and signal handlers to look through the global array and remove the tempfiles. If it succeeds, it calls the handler manually (which marks the tempfile structs as unused). - textconv has its own tempfile struct, which it allocates using prepare_temp_file and cleans up manually. No signal or atexit handlers. The new code moves the installation of cleanup handlers into the prepare_temp_file function. Which means that that function now has to understand that there is static tempfile storage. So what happens now is: - run_external_diff calls prepare_temp_file - prepare_temp_file calls claim_diff_tempfile, which allocates an unused slot from our global array - prepare_temp_file installs (if they have not already been installed) atexit and signal handlers for cleanup - prepare_temp_file sets up the tempfile as usual - prepare_temp_file returns a pointer to the allocated tempfile The advantage being that run_external_diff no longer has to care about setting up cleanup handlers. Now by virtue of calling prepare_temp_file, run_textconv gets the same benefit, as will any future users of prepare_temp_file. There are also a few side benefits to the specific implementation: - we now install cleanup handlers _before_ allocating the tempfile, closing a race which could leave temp cruft - when allocating a slot in the global array, we will now detect a situation where the old slots were not properly vacated (i.e., somebody forgot to call remove upon leaving the function). In the old code, such a situation would silently overwrite the tempfile names, meaning we would forget to clean them up. The new code dies with a bug warning. - we make sure only to install the signal handler once. This isn't a big deal, since we are just overwriting the old handler, but will become an issue when a later patch converts the code to use sigchain Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-01-22 05:59:56 +00:00
return temp;
}
/* An external diff command takes:
*
* diff-cmd name infile1 infile1-sha1 infile1-mode \
* infile2 infile2-sha1 infile2-mode [ rename-to ]
*
*/
static void run_external_diff(const char *pgm,
const char *name,
const char *other,
struct diff_filespec *one,
struct diff_filespec *two,
const char *xfrm_msg,
int complete_rewrite)
{
const char *spawn_arg[10];
int retval;
const char **arg = &spawn_arg[0];
if (one && two) {
diff: refactor tempfile cleanup handling There are two pieces of code that create tempfiles for diff: run_external_diff and run_textconv. The former cleans up its tempfiles in the face of premature death (i.e., by die() or by signal), but the latter does not. After this patch, they will both use the same cleanup routines. To make clear what the change is, let me first explain what happens now: - run_external_diff uses a static global array of 2 diff_tempfile structs (since it knows it will always need exactly 2 tempfiles). It calls prepare_temp_file (which doesn't know anything about the global array) on each of the structs, creating the tempfiles that need to be cleaned up. It then registers atexit and signal handlers to look through the global array and remove the tempfiles. If it succeeds, it calls the handler manually (which marks the tempfile structs as unused). - textconv has its own tempfile struct, which it allocates using prepare_temp_file and cleans up manually. No signal or atexit handlers. The new code moves the installation of cleanup handlers into the prepare_temp_file function. Which means that that function now has to understand that there is static tempfile storage. So what happens now is: - run_external_diff calls prepare_temp_file - prepare_temp_file calls claim_diff_tempfile, which allocates an unused slot from our global array - prepare_temp_file installs (if they have not already been installed) atexit and signal handlers for cleanup - prepare_temp_file sets up the tempfile as usual - prepare_temp_file returns a pointer to the allocated tempfile The advantage being that run_external_diff no longer has to care about setting up cleanup handlers. Now by virtue of calling prepare_temp_file, run_textconv gets the same benefit, as will any future users of prepare_temp_file. There are also a few side benefits to the specific implementation: - we now install cleanup handlers _before_ allocating the tempfile, closing a race which could leave temp cruft - when allocating a slot in the global array, we will now detect a situation where the old slots were not properly vacated (i.e., somebody forgot to call remove upon leaving the function). In the old code, such a situation would silently overwrite the tempfile names, meaning we would forget to clean them up. The new code dies with a bug warning. - we make sure only to install the signal handler once. This isn't a big deal, since we are just overwriting the old handler, but will become an issue when a later patch converts the code to use sigchain Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-01-22 05:59:56 +00:00
struct diff_tempfile *temp_one, *temp_two;
const char *othername = (other ? other : name);
temp_one = prepare_temp_file(name, one);
temp_two = prepare_temp_file(othername, two);
*arg++ = pgm;
*arg++ = name;
diff: refactor tempfile cleanup handling There are two pieces of code that create tempfiles for diff: run_external_diff and run_textconv. The former cleans up its tempfiles in the face of premature death (i.e., by die() or by signal), but the latter does not. After this patch, they will both use the same cleanup routines. To make clear what the change is, let me first explain what happens now: - run_external_diff uses a static global array of 2 diff_tempfile structs (since it knows it will always need exactly 2 tempfiles). It calls prepare_temp_file (which doesn't know anything about the global array) on each of the structs, creating the tempfiles that need to be cleaned up. It then registers atexit and signal handlers to look through the global array and remove the tempfiles. If it succeeds, it calls the handler manually (which marks the tempfile structs as unused). - textconv has its own tempfile struct, which it allocates using prepare_temp_file and cleans up manually. No signal or atexit handlers. The new code moves the installation of cleanup handlers into the prepare_temp_file function. Which means that that function now has to understand that there is static tempfile storage. So what happens now is: - run_external_diff calls prepare_temp_file - prepare_temp_file calls claim_diff_tempfile, which allocates an unused slot from our global array - prepare_temp_file installs (if they have not already been installed) atexit and signal handlers for cleanup - prepare_temp_file sets up the tempfile as usual - prepare_temp_file returns a pointer to the allocated tempfile The advantage being that run_external_diff no longer has to care about setting up cleanup handlers. Now by virtue of calling prepare_temp_file, run_textconv gets the same benefit, as will any future users of prepare_temp_file. There are also a few side benefits to the specific implementation: - we now install cleanup handlers _before_ allocating the tempfile, closing a race which could leave temp cruft - when allocating a slot in the global array, we will now detect a situation where the old slots were not properly vacated (i.e., somebody forgot to call remove upon leaving the function). In the old code, such a situation would silently overwrite the tempfile names, meaning we would forget to clean them up. The new code dies with a bug warning. - we make sure only to install the signal handler once. This isn't a big deal, since we are just overwriting the old handler, but will become an issue when a later patch converts the code to use sigchain Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-01-22 05:59:56 +00:00
*arg++ = temp_one->name;
*arg++ = temp_one->hex;
*arg++ = temp_one->mode;
*arg++ = temp_two->name;
*arg++ = temp_two->hex;
*arg++ = temp_two->mode;
if (other) {
*arg++ = other;
*arg++ = xfrm_msg;
}
} else {
*arg++ = pgm;
*arg++ = name;
}
*arg = NULL;
fflush(NULL);
retval = run_command_v_opt(spawn_arg, RUN_USING_SHELL);
remove_tempfile();
if (retval) {
fprintf(stderr, "external diff died, stopping at %s.\n", name);
exit(1);
}
}
static int similarity_index(struct diff_filepair *p)
{
return p->score * 100 / MAX_SCORE;
}
static void fill_metainfo(struct strbuf *msg,
const char *name,
const char *other,
struct diff_filespec *one,
struct diff_filespec *two,
struct diff_options *o,
struct diff_filepair *p,
int *must_show_header,
int use_color)
{
const char *set = diff_get_color(use_color, DIFF_METAINFO);
const char *reset = diff_get_color(use_color, DIFF_RESET);
struct strbuf *msgbuf;
char *line_prefix = "";
*must_show_header = 1;
if (o->output_prefix) {
msgbuf = o->output_prefix(o, o->output_prefix_data);
line_prefix = msgbuf->buf;
}
strbuf_init(msg, PATH_MAX * 2 + 300);
switch (p->status) {
case DIFF_STATUS_COPIED:
strbuf_addf(msg, "%s%ssimilarity index %d%%",
line_prefix, set, similarity_index(p));
strbuf_addf(msg, "%s\n%s%scopy from ",
reset, line_prefix, set);
quote_c_style(name, msg, NULL, 0);
strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
quote_c_style(other, msg, NULL, 0);
strbuf_addf(msg, "%s\n", reset);
break;
case DIFF_STATUS_RENAMED:
strbuf_addf(msg, "%s%ssimilarity index %d%%",
line_prefix, set, similarity_index(p));
strbuf_addf(msg, "%s\n%s%srename from ",
reset, line_prefix, set);
quote_c_style(name, msg, NULL, 0);
strbuf_addf(msg, "%s\n%s%srename to ",
reset, line_prefix, set);
quote_c_style(other, msg, NULL, 0);
strbuf_addf(msg, "%s\n", reset);
break;
case DIFF_STATUS_MODIFIED:
if (p->score) {
strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
line_prefix,
set, similarity_index(p), reset);
break;
}
/* fallthru */
default:
*must_show_header = 0;
}
if (one && two && hashcmp(one->sha1, two->sha1)) {
int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
if (DIFF_OPT_TST(o, BINARY)) {
mmfile_t mf;
if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
(!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
abbrev = 40;
}
strbuf_addf(msg, "%s%sindex %s..", line_prefix, set,
find_unique_abbrev(one->sha1, abbrev));
strbuf_addstr(msg, find_unique_abbrev(two->sha1, abbrev));
if (one->mode == two->mode)
strbuf_addf(msg, " %06o", one->mode);
strbuf_addf(msg, "%s\n", reset);
}
}
static void run_diff_cmd(const char *pgm,
const char *name,
const char *other,
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
const char *attr_path,
struct diff_filespec *one,
struct diff_filespec *two,
struct strbuf *msg,
binary patch. This adds "binary patch" to the diff output and teaches apply what to do with them. On the diff generation side, traditionally, we said "Binary files differ\n" without giving anything other than the preimage and postimage object name on the index line. This was good enough for applying a patch generated from your own repository (very useful while rebasing), because the postimage would be available in such a case. However, this was not useful when the recipient of such a patch via e-mail were to apply it, even if the preimage was available. This patch allows the diff to generate "binary" patch when operating under --full-index option. The binary patch follows the usual extended git diff headers, and looks like this: "GIT binary patch\n" <length byte><data>"\n" ... "\n" Each line is prefixed with a "length-byte", whose value is upper or lowercase alphabet that encodes number of bytes that the data on the line decodes to (1..52 -- 'A' means 1, 'B' means 2, ..., 'Z' means 26, 'a' means 27, ...). <data> is 1 or more groups of 5-byte sequence, each of which encodes up to 4 bytes in base85 encoding. Because 52 / 4 * 5 = 65 and we have the length byte, an output line is capped to 66 characters. The payload is the same diff-delta as we use in the packfiles. On the consumption side, git-apply now can decode and apply the binary patch when --allow-binary-replacement is given, the diff was generated with --full-index, and the receiving repository has the preimage blob, which is the same condition as it always required when accepting an "Binary files differ\n" patch. Signed-off-by: Junio C Hamano <junkio@cox.net>
2006-05-04 23:51:44 +00:00
struct diff_options *o,
struct diff_filepair *p)
{
const char *xfrm_msg = NULL;
int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
int must_show_header = 0;
if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
pgm = NULL;
else {
struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
if (drv && drv->external)
pgm = drv->external;
}
if (msg) {
/*
* don't use colors when the header is intended for an
* external diff driver
*/
fill_metainfo(msg, name, other, one, two, o, p,
&must_show_header,
color: delay auto-color decision until point of use When we read a color value either from a config file or from the command line, we use git_config_colorbool to convert it from the tristate always/never/auto into a single yes/no boolean value. This has some timing implications with respect to starting a pager. If we start (or decide not to start) the pager before checking the colorbool, everything is fine. Either isatty(1) will give us the right information, or we will properly check for pager_in_use(). However, if we decide to start a pager after we have checked the colorbool, things are not so simple. If stdout is a tty, then we will have already decided to use color. However, the user may also have configured color.pager not to use color with the pager. In this case, we need to actually turn off color. Unfortunately, the pager code has no idea which color variables were turned on (and there are many of them throughout the code, and they may even have been manipulated after the colorbool selection by something like "--color" on the command line). This bug can be seen any time a pager is started after config and command line options are checked. This has affected "git diff" since 89d07f7 (diff: don't run pager if user asked for a diff style exit code, 2007-08-12). It has also affect the log family since 1fda91b (Fix 'git log' early pager startup error case, 2010-08-24). This patch splits the notion of parsing a colorbool and actually checking the configuration. The "use_color" variables now have an additional possible value, GIT_COLOR_AUTO. Users of the variable should use the new "want_color()" wrapper, which will lazily determine and cache the auto-color decision. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-08-18 05:04:23 +00:00
want_color(o->use_color) && !pgm);
xfrm_msg = msg->len ? msg->buf : NULL;
}
if (pgm) {
run_external_diff(pgm, name, other, one, two, xfrm_msg,
complete_rewrite);
return;
}
if (one && two)
builtin_diff(name, other ? other : name,
one, two, xfrm_msg, must_show_header,
o, complete_rewrite);
else
fprintf(o->file, "* Unmerged path %s\n", name);
}
static void diff_fill_sha1_info(struct diff_filespec *one)
{
if (DIFF_FILE_VALID(one)) {
if (!one->sha1_valid) {
struct stat st;
if (!strcmp(one->path, "-")) {
hashcpy(one->sha1, null_sha1);
return;
}
if (lstat(one->path, &st) < 0)
die_errno("stat '%s'", one->path);
if (index_path(one->sha1, one->path, &st, 0))
die("cannot hash %s", one->path);
}
}
else
hashclr(one->sha1);
}
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
{
/* Strip the prefix but do not molest /dev/null and absolute paths */
if (*namep && **namep != '/') {
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
*namep += prefix_length;
if (**namep == '/')
++*namep;
}
if (*otherp && **otherp != '/') {
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
*otherp += prefix_length;
if (**otherp == '/')
++*otherp;
}
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
}
static void run_diff(struct diff_filepair *p, struct diff_options *o)
{
const char *pgm = external_diff();
struct strbuf msg;
struct diff_filespec *one = p->one;
struct diff_filespec *two = p->two;
const char *name;
const char *other;
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
const char *attr_path;
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
name = p->one->path;
other = (strcmp(name, p->two->path) ? p->two->path : NULL);
attr_path = name;
if (o->prefix_length)
strip_prefix(o->prefix_length, &name, &other);
if (DIFF_PAIR_UNMERGED(p)) {
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
run_diff_cmd(pgm, name, NULL, attr_path,
NULL, NULL, NULL, o, p);
return;
}
diff_fill_sha1_info(one);
diff_fill_sha1_info(two);
if (!pgm &&
DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
(S_IFMT & one->mode) != (S_IFMT & two->mode)) {
/*
* a filepair that changes between file and symlink
* needs to be split into deletion and creation.
*/
struct diff_filespec *null = alloc_filespec(two->path);
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
run_diff_cmd(NULL, name, other, attr_path,
one, null, &msg, o, p);
free(null);
strbuf_release(&msg);
null = alloc_filespec(one->path);
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
run_diff_cmd(NULL, name, other, attr_path,
null, two, &msg, o, p);
free(null);
}
else
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
run_diff_cmd(pgm, name, other, attr_path,
one, two, &msg, o, p);
strbuf_release(&msg);
}
static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
struct diffstat_t *diffstat)
{
const char *name;
const char *other;
int complete_rewrite = 0;
if (DIFF_PAIR_UNMERGED(p)) {
/* unmerged */
builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, 0);
return;
}
name = p->one->path;
other = (strcmp(name, p->two->path) ? p->two->path : NULL);
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
if (o->prefix_length)
strip_prefix(o->prefix_length, &name, &other);
diff_fill_sha1_info(p->one);
diff_fill_sha1_info(p->two);
if (p->status == DIFF_STATUS_MODIFIED && p->score)
complete_rewrite = 1;
builtin_diffstat(name, other, p->one, p->two, diffstat, o, complete_rewrite);
}
static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
{
const char *name;
const char *other;
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
const char *attr_path;
if (DIFF_PAIR_UNMERGED(p)) {
/* unmerged */
return;
}
name = p->one->path;
other = (strcmp(name, p->two->path) ? p->two->path : NULL);
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
attr_path = other ? other : name;
if (o->prefix_length)
strip_prefix(o->prefix_length, &name, &other);
diff_fill_sha1_info(p->one);
diff_fill_sha1_info(p->two);
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
}
void diff_setup(struct diff_options *options)
{
memcpy(options, &default_diff_options, sizeof(*options));
options->file = stdout;
options->line_termination = '\n';
options->break_opt = -1;
options->rename_limit = -1;
options->dirstat_permille = diff_dirstat_permille_default;
options->context = 3;
options->change = diff_change;
options->add_remove = diff_addremove;
options->use_color = diff_use_color_default;
options->detect_rename = diff_detect_rename_default;
if (diff_no_prefix) {
options->a_prefix = options->b_prefix = "";
} else if (!diff_mnemonic_prefix) {
options->a_prefix = "a/";
options->b_prefix = "b/";
}
}
int diff_setup_done(struct diff_options *options)
{
int count = 0;
if (options->output_format & DIFF_FORMAT_NAME)
count++;
if (options->output_format & DIFF_FORMAT_NAME_STATUS)
count++;
if (options->output_format & DIFF_FORMAT_CHECKDIFF)
count++;
if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
count++;
if (count > 1)
die("--name-only, --name-status, --check and -s are mutually exclusive");
/*
* Most of the time we can say "there are changes"
* only by checking if there are changed paths, but
* --ignore-whitespace* options force us to look
* inside contents.
*/
if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
else
DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
options->detect_rename = DIFF_DETECT_COPY;
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
if (!DIFF_OPT_TST(options, RELATIVE_NAME))
options->prefix = NULL;
if (options->prefix)
options->prefix_length = strlen(options->prefix);
else
options->prefix_length = 0;
if (options->output_format & (DIFF_FORMAT_NAME |
DIFF_FORMAT_NAME_STATUS |
DIFF_FORMAT_CHECKDIFF |
DIFF_FORMAT_NO_OUTPUT))
options->output_format &= ~(DIFF_FORMAT_RAW |
DIFF_FORMAT_NUMSTAT |
DIFF_FORMAT_DIFFSTAT |
DIFF_FORMAT_SHORTSTAT |
Add "--dirstat" for some directory statistics This adds a new form of overview diffstat output, doing something that I have occasionally ended up doing manually (and badly, because it's actually pretty nasty to do), and that I think is very useful for an project like the kernel that has a fairly deep and well-separated directory structure with semantic meaning. What I mean by that is that it's often interesting to see exactly which sub-directories are impacted by a patch, and to what degree - even if you don't perhaps care so much about the individual files themselves. What makes the concept more interesting is that the "impact" is often hierarchical: in the kernel, for example, something could either have a very localized impact to "fs/ext3/" and then it's interesting to see that such a patch changes mostly that subdirectory, but you could have another patch that changes some generic VFS-layer issue which affects _many_ subdirectories that are all under "fs/", but none - or perhaps just a couple of them - of the individual filesystems are interesting in themselves. So what commonly happens is that you may have big changes in a specific sub-subdirectory, but still also significant separate changes to the subdirectory leading up to that - maybe you have significant VFS-level changes, but *also* changes under that VFS layer in the NFS-specific directories, for example. In that case, you do want the low-level parts that are significant to show up, but then the insignificant ones should show up as under the more generic top-level directory. This patch shows all of that with "--dirstat". The output can be either something simple like commit 81772fe... Author: Thomas Gleixner <tglx@linutronix.de> Date: Sun Feb 10 23:57:36 2008 +0100 x86: remove over noisy debug printk pageattr-test.c contains a noisy debug printk that people reported. The condition under which it prints (randomly tapping into a mem_map[] hole and not being able to c_p_a() there) is valid behavior and not interesting to report. Remove it. Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> 100.0% arch/x86/mm/ or something much more complex like commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 15.3% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ where that latter example is an example of significant work in some individual fs/*/ subdirectories (like the patches to reiserfs accounting for 7.6% of the whole), but then discounting those individual filesystems, there's also 15.3% other "random" things that weren't worth reporting on their oen left over under fs/ in general (either in that directory itself, or in subdirectories of fs/ that didn't have enough changes to be reported individually). I'd like to stress that the "15.3% fs/" mentioned above is the stuff that is under fs/ but that was _not_ significant enough to report on its own. So the above does _not_ mean that 15.3% of the work was under fs/ per se, because that 15.3% does *not* include the already-reported 7.6% of afs, 7.6% of fuse etc. If you want to enable "cumulative" directory statistics, you can use the "--cumulative" flag, which adds up percentages recursively even when they have been already reported for a sub-directory. That cumulative output is disabled if *all* of the changes in one subdirectory come from a deeper subdirectory, to avoid repeating subdirectories all the way to the root. For an example of the cumulative reporting, the above commit becomes commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 61.5% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ in which the commit percentages now obviously add up to much more than 100%: now the changes that were already reported for the sub-directories under fs/ are then cumulatively included in the whole percentage of fs/ (ie now shows 61.5% as opposed to the 15.3% without the cumulative reporting). The default reporting limit has been arbitrarily set at 3%, which seems to be a pretty good cut-off, but you can specify the cut-off manually by giving it as an option parameter (eg "--dirstat=5" makes the cut-off be at 5% instead) NOTE! The percentages are purely about the total lines added and removed, not anything smarter (or dumber) than that. Also note that you should not generally expect things to add up to 100%: not only does it round down, we don't report leftover scraps (they add up to the top-level change count, but we don't even bother reporting that, it only reports subdirectories). Quite frankly, as a top-level manager this is really convenient for me, but it's going to be very boring for git itself since there are few subdirectories. Also, don't expect things to make tons of sense if you combine this with "-M" and there are cross-directory renames etc. But even for git itself, you can get some fun statistics. Try out git log --dirstat and see the occasional mentions of things like Documentation/, git-gui/, gitweb/ and gitk-git/. Or try out something like git diff --dirstat v1.5.0..v1.5.4 which does kind of git an overview that shows *something*. But in general, the output is more exciting for big projects with deeper structure, and doing a git diff --dirstat v2.6.24..v2.6.25-rc1 on the kernel is what I actually wrote this for! Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 21:26:31 +00:00
DIFF_FORMAT_DIRSTAT |
DIFF_FORMAT_SUMMARY |
DIFF_FORMAT_PATCH);
/*
* These cases always need recursive; we do not drop caller-supplied
* recursive bits for other formats here.
*/
if (options->output_format & (DIFF_FORMAT_PATCH |
DIFF_FORMAT_NUMSTAT |
DIFF_FORMAT_DIFFSTAT |
DIFF_FORMAT_SHORTSTAT |
Add "--dirstat" for some directory statistics This adds a new form of overview diffstat output, doing something that I have occasionally ended up doing manually (and badly, because it's actually pretty nasty to do), and that I think is very useful for an project like the kernel that has a fairly deep and well-separated directory structure with semantic meaning. What I mean by that is that it's often interesting to see exactly which sub-directories are impacted by a patch, and to what degree - even if you don't perhaps care so much about the individual files themselves. What makes the concept more interesting is that the "impact" is often hierarchical: in the kernel, for example, something could either have a very localized impact to "fs/ext3/" and then it's interesting to see that such a patch changes mostly that subdirectory, but you could have another patch that changes some generic VFS-layer issue which affects _many_ subdirectories that are all under "fs/", but none - or perhaps just a couple of them - of the individual filesystems are interesting in themselves. So what commonly happens is that you may have big changes in a specific sub-subdirectory, but still also significant separate changes to the subdirectory leading up to that - maybe you have significant VFS-level changes, but *also* changes under that VFS layer in the NFS-specific directories, for example. In that case, you do want the low-level parts that are significant to show up, but then the insignificant ones should show up as under the more generic top-level directory. This patch shows all of that with "--dirstat". The output can be either something simple like commit 81772fe... Author: Thomas Gleixner <tglx@linutronix.de> Date: Sun Feb 10 23:57:36 2008 +0100 x86: remove over noisy debug printk pageattr-test.c contains a noisy debug printk that people reported. The condition under which it prints (randomly tapping into a mem_map[] hole and not being able to c_p_a() there) is valid behavior and not interesting to report. Remove it. Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Ingo Molnar <mingo@elte.hu> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> 100.0% arch/x86/mm/ or something much more complex like commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 15.3% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ where that latter example is an example of significant work in some individual fs/*/ subdirectories (like the patches to reiserfs accounting for 7.6% of the whole), but then discounting those individual filesystems, there's also 15.3% other "random" things that weren't worth reporting on their oen left over under fs/ in general (either in that directory itself, or in subdirectories of fs/ that didn't have enough changes to be reported individually). I'd like to stress that the "15.3% fs/" mentioned above is the stuff that is under fs/ but that was _not_ significant enough to report on its own. So the above does _not_ mean that 15.3% of the work was under fs/ per se, because that 15.3% does *not* include the already-reported 7.6% of afs, 7.6% of fuse etc. If you want to enable "cumulative" directory statistics, you can use the "--cumulative" flag, which adds up percentages recursively even when they have been already reported for a sub-directory. That cumulative output is disabled if *all* of the changes in one subdirectory come from a deeper subdirectory, to avoid repeating subdirectories all the way to the root. For an example of the cumulative reporting, the above commit becomes commit e231c2e... Author: David Howells <dhowells@redhat.com> Date: Thu Feb 7 00:15:26 2008 -0800 Convert ERR_PTR(PTR_ERR(p)) instances to ERR_CAST(p) 20.5% crypto/ 7.6% fs/afs/ 7.6% fs/fuse/ 7.6% fs/gfs2/ 5.1% fs/jffs2/ 5.1% fs/nfs/ 5.1% fs/nfsd/ 7.6% fs/reiserfs/ 61.5% fs/ 7.6% net/rxrpc/ 10.2% security/keys/ in which the commit percentages now obviously add up to much more than 100%: now the changes that were already reported for the sub-directories under fs/ are then cumulatively included in the whole percentage of fs/ (ie now shows 61.5% as opposed to the 15.3% without the cumulative reporting). The default reporting limit has been arbitrarily set at 3%, which seems to be a pretty good cut-off, but you can specify the cut-off manually by giving it as an option parameter (eg "--dirstat=5" makes the cut-off be at 5% instead) NOTE! The percentages are purely about the total lines added and removed, not anything smarter (or dumber) than that. Also note that you should not generally expect things to add up to 100%: not only does it round down, we don't report leftover scraps (they add up to the top-level change count, but we don't even bother reporting that, it only reports subdirectories). Quite frankly, as a top-level manager this is really convenient for me, but it's going to be very boring for git itself since there are few subdirectories. Also, don't expect things to make tons of sense if you combine this with "-M" and there are cross-directory renames etc. But even for git itself, you can get some fun statistics. Try out git log --dirstat and see the occasional mentions of things like Documentation/, git-gui/, gitweb/ and gitk-git/. Or try out something like git diff --dirstat v1.5.0..v1.5.4 which does kind of git an overview that shows *something*. But in general, the output is more exciting for big projects with deeper structure, and doing a git diff --dirstat v2.6.24..v2.6.25-rc1 on the kernel is what I actually wrote this for! Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 21:26:31 +00:00
DIFF_FORMAT_DIRSTAT |
DIFF_FORMAT_SUMMARY |
DIFF_FORMAT_CHECKDIFF))
DIFF_OPT_SET(options, RECURSIVE);
/*
* Also pickaxe would not work very well if you do not say recursive
*/
if (options->pickaxe)
DIFF_OPT_SET(options, RECURSIVE);
/*
* When patches are generated, submodules diffed against the work tree
* must be checked for dirtiness too so it can be shown in the output
*/
if (options->output_format & DIFF_FORMAT_PATCH)
DIFF_OPT_SET(options, DIRTY_SUBMODULES);
if (options->detect_rename && options->rename_limit < 0)
options->rename_limit = diff_rename_limit_default;
if (options->setup & DIFF_SETUP_USE_CACHE) {
if (!active_cache)
/* read-cache does not die even when it fails
* so it is safe for us to do this here. Also
* it does not smudge active_cache or active_nr
* when it fails, so we do not have to worry about
* cleaning it up ourselves either.
*/
read_cache();
}
if (options->abbrev <= 0 || 40 < options->abbrev)
options->abbrev = 40; /* full */
/*
* It does not make sense to show the first hit we happened
* to have found. It does not make sense not to return with
* exit code in such a case either.
*/
if (DIFF_OPT_TST(options, QUICK)) {
options->output_format = DIFF_FORMAT_NO_OUTPUT;
DIFF_OPT_SET(options, EXIT_WITH_STATUS);
}
return 0;
}
static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
{
char c, *eq;
int len;
if (*arg != '-')
return 0;
c = *++arg;
if (!c)
return 0;
if (c == arg_short) {
c = *++arg;
if (!c)
return 1;
if (val && isdigit(c)) {
char *end;
int n = strtoul(arg, &end, 10);
if (*end)
return 0;
*val = n;
return 1;
}
return 0;
}
if (c != '-')
return 0;
arg++;
eq = strchr(arg, '=');
if (eq)
len = eq - arg;
else
len = strlen(arg);
if (!len || strncmp(arg, arg_long, len))
return 0;
if (eq) {
int n;
char *end;
if (!isdigit(*++eq))
return 0;
n = strtoul(eq, &end, 10);
if (*end)
return 0;
*val = n;
}
return 1;
}
static int diff_scoreopt_parse(const char *opt);
static inline int short_opt(char opt, const char **argv,
const char **optarg)
{
const char *arg = argv[0];
if (arg[0] != '-' || arg[1] != opt)
return 0;
if (arg[2] != '\0') {
*optarg = arg + 2;
return 1;
}
if (!argv[1])
die("Option '%c' requires a value", opt);
*optarg = argv[1];
return 2;
}
int parse_long_opt(const char *opt, const char **argv,
const char **optarg)
{
const char *arg = argv[0];
if (arg[0] != '-' || arg[1] != '-')
return 0;
arg += strlen("--");
if (prefixcmp(arg, opt))
return 0;
arg += strlen(opt);
if (*arg == '=') { /* sticked form: --option=value */
*optarg = arg + 1;
return 1;
}
if (*arg != '\0')
return 0;
/* separate form: --option value */
if (!argv[1])
die("Option '--%s' requires a value", opt);
*optarg = argv[1];
return 2;
}
static int stat_opt(struct diff_options *options, const char **av)
{
const char *arg = av[0];
char *end;
int width = options->stat_width;
int name_width = options->stat_name_width;
int count = options->stat_count;
int argcount = 1;
arg += strlen("--stat");
end = (char *)arg;
switch (*arg) {
case '-':
if (!prefixcmp(arg, "-width")) {
arg += strlen("-width");
if (*arg == '=')
width = strtoul(arg + 1, &end, 10);
else if (!*arg && !av[1])
die("Option '--stat-width' requires a value");
else if (!*arg) {
width = strtoul(av[1], &end, 10);
argcount = 2;
}
} else if (!prefixcmp(arg, "-name-width")) {
arg += strlen("-name-width");
if (*arg == '=')
name_width = strtoul(arg + 1, &end, 10);
else if (!*arg && !av[1])
die("Option '--stat-name-width' requires a value");
else if (!*arg) {
name_width = strtoul(av[1], &end, 10);
argcount = 2;
}
} else if (!prefixcmp(arg, "-count")) {
arg += strlen("-count");
if (*arg == '=')
count = strtoul(arg + 1, &end, 10);
else if (!*arg && !av[1])
die("Option '--stat-count' requires a value");
else if (!*arg) {
count = strtoul(av[1], &end, 10);
argcount = 2;
}
}
break;
case '=':
width = strtoul(arg+1, &end, 10);
if (*end == ',')
name_width = strtoul(end+1, &end, 10);
if (*end == ',')
count = strtoul(end+1, &end, 10);
}
/* Important! This checks all the error cases! */
if (*end)
return 0;
options->output_format |= DIFF_FORMAT_DIFFSTAT;
options->stat_name_width = name_width;
options->stat_width = width;
options->stat_count = count;
return argcount;
}
static int parse_dirstat_opt(struct diff_options *options, const char *params)
{
Improve error handling when parsing dirstat parameters When encountering errors or unknown tokens while parsing parameters to the --dirstat option, it makes sense to die() with an error message informing the user of which parameter did not make sense. However, when parsing the diff.dirstat config variable, we cannot simply die(), but should instead (after warning the user) ignore the erroneous or unrecognized parameter. After all, future Git versions might add more dirstat parameters, and using two different Git versions on the same repo should not cripple the older Git version just because of a parameter that is only understood by a more recent Git version. This patch fixes the issue by refactoring the dirstat parameter parsing so that parse_dirstat_params() keeps on parsing parameters, even if an earlier parameter was not recognized. When parsing has finished, it returns zero if all parameters were successfully parsed, and non-zero if one or more parameters were not recognized (with appropriate error messages appended to the 'errmsg' argument). The parse_dirstat_params() callers then decide (based on the return value from parse_dirstat_params()) whether to warn and ignore (in case of diff.dirstat), or to warn and die (in case of --dirstat). The patch also adds a couple of tests verifying the correct behavior of --dirstat and diff.dirstat in the face of unknown (possibly future) dirstat parameters. Suggested-by: Junio C Hamano <gitster@pobox.com> Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:22 +00:00
struct strbuf errmsg = STRBUF_INIT;
if (parse_dirstat_params(options, params, &errmsg))
die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
Improve error handling when parsing dirstat parameters When encountering errors or unknown tokens while parsing parameters to the --dirstat option, it makes sense to die() with an error message informing the user of which parameter did not make sense. However, when parsing the diff.dirstat config variable, we cannot simply die(), but should instead (after warning the user) ignore the erroneous or unrecognized parameter. After all, future Git versions might add more dirstat parameters, and using two different Git versions on the same repo should not cripple the older Git version just because of a parameter that is only understood by a more recent Git version. This patch fixes the issue by refactoring the dirstat parameter parsing so that parse_dirstat_params() keeps on parsing parameters, even if an earlier parameter was not recognized. When parsing has finished, it returns zero if all parameters were successfully parsed, and non-zero if one or more parameters were not recognized (with appropriate error messages appended to the 'errmsg' argument). The parse_dirstat_params() callers then decide (based on the return value from parse_dirstat_params()) whether to warn and ignore (in case of diff.dirstat), or to warn and die (in case of --dirstat). The patch also adds a couple of tests verifying the correct behavior of --dirstat and diff.dirstat in the face of unknown (possibly future) dirstat parameters. Suggested-by: Junio C Hamano <gitster@pobox.com> Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:22 +00:00
errmsg.buf);
strbuf_release(&errmsg);
/*
* The caller knows a dirstat-related option is given from the command
* line; allow it to say "return this_function();"
*/
options->output_format |= DIFF_FORMAT_DIRSTAT;
return 1;
}
int diff_opt_parse(struct diff_options *options, const char **av, int ac)
{
const char *arg = av[0];
const char *optarg;
int argcount;
/* Output format options */
if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch"))
options->output_format |= DIFF_FORMAT_PATCH;
else if (opt_arg(arg, 'U', "unified", &options->context))
options->output_format |= DIFF_FORMAT_PATCH;
else if (!strcmp(arg, "--raw"))
options->output_format |= DIFF_FORMAT_RAW;
else if (!strcmp(arg, "--patch-with-raw"))
options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW;
else if (!strcmp(arg, "--numstat"))
options->output_format |= DIFF_FORMAT_NUMSTAT;
else if (!strcmp(arg, "--shortstat"))
options->output_format |= DIFF_FORMAT_SHORTSTAT;
else if (!strcmp(arg, "-X") || !strcmp(arg, "--dirstat"))
return parse_dirstat_opt(options, "");
else if (!prefixcmp(arg, "-X"))
return parse_dirstat_opt(options, arg + 2);
else if (!prefixcmp(arg, "--dirstat="))
return parse_dirstat_opt(options, arg + 10);
else if (!strcmp(arg, "--cumulative"))
return parse_dirstat_opt(options, "cumulative");
else if (!strcmp(arg, "--dirstat-by-file"))
return parse_dirstat_opt(options, "files");
else if (!prefixcmp(arg, "--dirstat-by-file=")) {
parse_dirstat_opt(options, "files");
return parse_dirstat_opt(options, arg + 18);
}
else if (!strcmp(arg, "--check"))
options->output_format |= DIFF_FORMAT_CHECKDIFF;
else if (!strcmp(arg, "--summary"))
options->output_format |= DIFF_FORMAT_SUMMARY;
else if (!strcmp(arg, "--patch-with-stat"))
options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT;
else if (!strcmp(arg, "--name-only"))
options->output_format |= DIFF_FORMAT_NAME;
else if (!strcmp(arg, "--name-status"))
options->output_format |= DIFF_FORMAT_NAME_STATUS;
else if (!strcmp(arg, "-s"))
options->output_format |= DIFF_FORMAT_NO_OUTPUT;
else if (!prefixcmp(arg, "--stat"))
/* --stat, --stat-width, --stat-name-width, or --stat-count */
return stat_opt(options, av);
/* renames options */
else if (!prefixcmp(arg, "-B") || !prefixcmp(arg, "--break-rewrites=") ||
!strcmp(arg, "--break-rewrites")) {
if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
return error("invalid argument to -B: %s", arg+2);
}
else if (!prefixcmp(arg, "-M") || !prefixcmp(arg, "--find-renames=") ||
!strcmp(arg, "--find-renames")) {
if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
return error("invalid argument to -M: %s", arg+2);
options->detect_rename = DIFF_DETECT_RENAME;
}
else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
options->irreversible_delete = 1;
}
else if (!prefixcmp(arg, "-C") || !prefixcmp(arg, "--find-copies=") ||
!strcmp(arg, "--find-copies")) {
if (options->detect_rename == DIFF_DETECT_COPY)
DIFF_OPT_SET(options, FIND_COPIES_HARDER);
if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
return error("invalid argument to -C: %s", arg+2);
options->detect_rename = DIFF_DETECT_COPY;
}
else if (!strcmp(arg, "--no-renames"))
options->detect_rename = 0;
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
else if (!strcmp(arg, "--relative"))
DIFF_OPT_SET(options, RELATIVE_NAME);
else if (!prefixcmp(arg, "--relative=")) {
DIFF_OPT_SET(options, RELATIVE_NAME);
options->prefix = arg + 11;
}
/* xdiff options */
else if (!strcmp(arg, "--minimal"))
DIFF_XDL_SET(options, NEED_MINIMAL);
else if (!strcmp(arg, "--no-minimal"))
DIFF_XDL_CLR(options, NEED_MINIMAL);
else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
DIFF_XDL_SET(options, IGNORE_WHITESPACE);
else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
else if (!strcmp(arg, "--ignore-space-at-eol"))
DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
else if (!strcmp(arg, "--patience"))
DIFF_XDL_SET(options, PATIENCE_DIFF);
else if (!strcmp(arg, "--histogram"))
DIFF_XDL_SET(options, HISTOGRAM_DIFF);
/* flags options */
else if (!strcmp(arg, "--binary")) {
options->output_format |= DIFF_FORMAT_PATCH;
DIFF_OPT_SET(options, BINARY);
}
else if (!strcmp(arg, "--full-index"))
DIFF_OPT_SET(options, FULL_INDEX);
else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
DIFF_OPT_SET(options, TEXT);
else if (!strcmp(arg, "-R"))
DIFF_OPT_SET(options, REVERSE_DIFF);
else if (!strcmp(arg, "--find-copies-harder"))
DIFF_OPT_SET(options, FIND_COPIES_HARDER);
Finally implement "git log --follow" Ok, I've really held off doing this too damn long, because I'm lazy, and I was always hoping that somebody else would do it. But no, people keep asking for it, but nobody actually did anything, so I decided I might as well bite the bullet, and instead of telling people they could add a "--follow" flag to "git log" to do what they want to do, I decided that it looks like I just have to do it for them.. The code wasn't actually that complicated, in that the diffstat for this patch literally says "70 insertions(+), 1 deletions(-)", but I will have to admit that in order to get to this fairly simple patch, you did have to know and understand the internal git diff generation machinery pretty well, and had to really be able to follow how commit generation interacts with generating patches and generating the log. So I suspect that while I was right that it wasn't that hard, I might have been expecting too much of random people - this patch does seem to be firmly in the core "Linus or Junio" territory. To make a long story short: I'm sorry for it taking so long until I just did it. I'm not going to guarantee that this works for everybody, but you really can just look at the patch, and after the appropriate appreciative noises ("Ooh, aah") over how clever I am, you can then just notice that the code itself isn't really that complicated. All the real new code is in the new "try_to_follow_renames()" function. It really isn't rocket science: we notice that the pathname we were looking at went away, so we start a full tree diff and try to see if we can instead make that pathname be a rename or a copy from some other previous pathname. And if we can, we just continue, except we show *that* particular diff, and ever after we use the _previous_ pathname. One thing to look out for: the "rename detection" is considered to be a singular event in the _linear_ "git log" output! That's what people want to do, but I just wanted to point out that this patch is *not* carrying around a "commit,pathname" kind of pair and it's *not* going to be able to notice the file coming from multiple *different* files in earlier history. IOW, if you use "git log --follow", then you get the stupid CVS/SVN kind of "files have single identities" kind of semantics, and git log will just pick the identity based on the normal move/copy heuristics _as_if_ the history could be linearized. Put another way: I think the model is broken, but given the broken model, I think this patch does just about as well as you can do. If you have merges with the same "file" having different filenames over the two branches, git will just end up picking _one_ of the pathnames at the point where the newer one goes away. It never looks at multiple pathnames in parallel. And if you understood all that, you probably didn't need it explained, and if you didn't understand the above blathering, it doesn't really mtter to you. What matters to you is that you can now do git log -p --follow builtin-rev-list.c and it will find the point where the old "rev-list.c" got renamed to "builtin-rev-list.c" and show it as such. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2007-06-19 21:22:46 +00:00
else if (!strcmp(arg, "--follow"))
DIFF_OPT_SET(options, FOLLOW_RENAMES);
else if (!strcmp(arg, "--color"))
options->use_color = 1;
else if (!prefixcmp(arg, "--color=")) {
int value = git_config_colorbool(NULL, arg+8);
color: delay auto-color decision until point of use When we read a color value either from a config file or from the command line, we use git_config_colorbool to convert it from the tristate always/never/auto into a single yes/no boolean value. This has some timing implications with respect to starting a pager. If we start (or decide not to start) the pager before checking the colorbool, everything is fine. Either isatty(1) will give us the right information, or we will properly check for pager_in_use(). However, if we decide to start a pager after we have checked the colorbool, things are not so simple. If stdout is a tty, then we will have already decided to use color. However, the user may also have configured color.pager not to use color with the pager. In this case, we need to actually turn off color. Unfortunately, the pager code has no idea which color variables were turned on (and there are many of them throughout the code, and they may even have been manipulated after the colorbool selection by something like "--color" on the command line). This bug can be seen any time a pager is started after config and command line options are checked. This has affected "git diff" since 89d07f7 (diff: don't run pager if user asked for a diff style exit code, 2007-08-12). It has also affect the log family since 1fda91b (Fix 'git log' early pager startup error case, 2010-08-24). This patch splits the notion of parsing a colorbool and actually checking the configuration. The "use_color" variables now have an additional possible value, GIT_COLOR_AUTO. Users of the variable should use the new "want_color()" wrapper, which will lazily determine and cache the auto-color decision. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-08-18 05:04:23 +00:00
if (value < 0)
return error("option `color' expects \"always\", \"auto\", or \"never\"");
color: delay auto-color decision until point of use When we read a color value either from a config file or from the command line, we use git_config_colorbool to convert it from the tristate always/never/auto into a single yes/no boolean value. This has some timing implications with respect to starting a pager. If we start (or decide not to start) the pager before checking the colorbool, everything is fine. Either isatty(1) will give us the right information, or we will properly check for pager_in_use(). However, if we decide to start a pager after we have checked the colorbool, things are not so simple. If stdout is a tty, then we will have already decided to use color. However, the user may also have configured color.pager not to use color with the pager. In this case, we need to actually turn off color. Unfortunately, the pager code has no idea which color variables were turned on (and there are many of them throughout the code, and they may even have been manipulated after the colorbool selection by something like "--color" on the command line). This bug can be seen any time a pager is started after config and command line options are checked. This has affected "git diff" since 89d07f7 (diff: don't run pager if user asked for a diff style exit code, 2007-08-12). It has also affect the log family since 1fda91b (Fix 'git log' early pager startup error case, 2010-08-24). This patch splits the notion of parsing a colorbool and actually checking the configuration. The "use_color" variables now have an additional possible value, GIT_COLOR_AUTO. Users of the variable should use the new "want_color()" wrapper, which will lazily determine and cache the auto-color decision. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-08-18 05:04:23 +00:00
options->use_color = value;
}
else if (!strcmp(arg, "--no-color"))
options->use_color = 0;
else if (!strcmp(arg, "--color-words")) {
options->use_color = 1;
options->word_diff = DIFF_WORDS_COLOR;
}
else if (!prefixcmp(arg, "--color-words=")) {
options->use_color = 1;
options->word_diff = DIFF_WORDS_COLOR;
options->word_regex = arg + 14;
}
else if (!strcmp(arg, "--word-diff")) {
if (options->word_diff == DIFF_WORDS_NONE)
options->word_diff = DIFF_WORDS_PLAIN;
}
else if (!prefixcmp(arg, "--word-diff=")) {
const char *type = arg + 12;
if (!strcmp(type, "plain"))
options->word_diff = DIFF_WORDS_PLAIN;
else if (!strcmp(type, "color")) {
options->use_color = 1;
options->word_diff = DIFF_WORDS_COLOR;
}
else if (!strcmp(type, "porcelain"))
options->word_diff = DIFF_WORDS_PORCELAIN;
else if (!strcmp(type, "none"))
options->word_diff = DIFF_WORDS_NONE;
else
die("bad --word-diff argument: %s", type);
}
else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
if (options->word_diff == DIFF_WORDS_NONE)
options->word_diff = DIFF_WORDS_PLAIN;
options->word_regex = optarg;
return argcount;
}
else if (!strcmp(arg, "--exit-code"))
DIFF_OPT_SET(options, EXIT_WITH_STATUS);
else if (!strcmp(arg, "--quiet"))
DIFF_OPT_SET(options, QUICK);
else if (!strcmp(arg, "--ext-diff"))
DIFF_OPT_SET(options, ALLOW_EXTERNAL);
else if (!strcmp(arg, "--no-ext-diff"))
DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
else if (!strcmp(arg, "--textconv"))
DIFF_OPT_SET(options, ALLOW_TEXTCONV);
else if (!strcmp(arg, "--no-textconv"))
DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
Submodules: Add the new "ignore" config option for diff and status The new "ignore" config option controls the default behavior for "git status" and the diff family. It specifies under what circumstances they consider submodules as modified and can be set separately for each submodule. The command line option "--ignore-submodules=" has been extended to accept the new parameter "none" for both status and diff. Users that chose submodules to get rid of long work tree scanning times might want to set the "dirty" option for those submodules. This brings back the pre 1.7.0 behavior, where submodule work trees were never scanned for modifications. By using "--ignore-submodules=none" on the command line the status and diff commands can be told to do a full scan. This option can be set to the following values (which have the same name and meaning as for the "--ignore-submodules" option of status and diff): "all": All changes to the submodule will be ignored. "dirty": Only differences of the commit recorded in the superproject and the submodules HEAD will be considered modifications, all changes to the work tree of the submodule will be ignored. When using this value, the submodule will not be scanned for work tree changes at all, leading to a performance benefit on large submodules. "untracked": Only untracked files in the submodules work tree are ignored, a changed HEAD and/or modified files in the submodule will mark it as modified. "none" (which is the default): Either untracked or modified files in a submodules work tree or a difference between the subdmodules HEAD and the commit recorded in the superproject will make it show up as changed. This value is added as a new parameter for the "--ignore-submodules" option of the diff family and "git status" so the user can override the settings in the configuration. Signed-off-by: Jens Lehmann <Jens.Lehmann@web.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-08-05 22:39:25 +00:00
else if (!strcmp(arg, "--ignore-submodules")) {
DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
Add the option "--ignore-submodules" to "git status" In some use cases it is not desirable that "git status" considers submodules that only contain untracked content as dirty. This may happen e.g. when the submodule is not under the developers control and not all build generated files have been added to .gitignore by the upstream developers. Using the "untracked" parameter for the "--ignore-submodules" option disables checking for untracked content and lets git diff report them as changed only when they have new commits or modified content. Sometimes it is not wanted to have submodules show up as changed when they just contain changes to their work tree (this was the behavior before 1.7.0). An example for that are scripts which just want to check for submodule commits while ignoring any changes to the work tree. Also users having large submodules known not to change might want to use this option, as the - sometimes substantial - time it takes to scan the submodule work tree(s) is saved when using the "dirty" parameter. And if you want to ignore any changes to submodules, you can now do that by using this option without parameters or with "all" (when the config option status.submodulesummary is set, using "all" will also suppress the output of the submodule summary). A new function handle_ignore_submodules_arg() is introduced to parse this option new to "git status" in a single location, as "git diff" already knew it. Signed-off-by: Jens Lehmann <Jens.Lehmann@web.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-06-25 14:56:47 +00:00
handle_ignore_submodules_arg(options, "all");
Submodules: Add the new "ignore" config option for diff and status The new "ignore" config option controls the default behavior for "git status" and the diff family. It specifies under what circumstances they consider submodules as modified and can be set separately for each submodule. The command line option "--ignore-submodules=" has been extended to accept the new parameter "none" for both status and diff. Users that chose submodules to get rid of long work tree scanning times might want to set the "dirty" option for those submodules. This brings back the pre 1.7.0 behavior, where submodule work trees were never scanned for modifications. By using "--ignore-submodules=none" on the command line the status and diff commands can be told to do a full scan. This option can be set to the following values (which have the same name and meaning as for the "--ignore-submodules" option of status and diff): "all": All changes to the submodule will be ignored. "dirty": Only differences of the commit recorded in the superproject and the submodules HEAD will be considered modifications, all changes to the work tree of the submodule will be ignored. When using this value, the submodule will not be scanned for work tree changes at all, leading to a performance benefit on large submodules. "untracked": Only untracked files in the submodules work tree are ignored, a changed HEAD and/or modified files in the submodule will mark it as modified. "none" (which is the default): Either untracked or modified files in a submodules work tree or a difference between the subdmodules HEAD and the commit recorded in the superproject will make it show up as changed. This value is added as a new parameter for the "--ignore-submodules" option of the diff family and "git status" so the user can override the settings in the configuration. Signed-off-by: Jens Lehmann <Jens.Lehmann@web.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-08-05 22:39:25 +00:00
} else if (!prefixcmp(arg, "--ignore-submodules=")) {
DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
Add the option "--ignore-submodules" to "git status" In some use cases it is not desirable that "git status" considers submodules that only contain untracked content as dirty. This may happen e.g. when the submodule is not under the developers control and not all build generated files have been added to .gitignore by the upstream developers. Using the "untracked" parameter for the "--ignore-submodules" option disables checking for untracked content and lets git diff report them as changed only when they have new commits or modified content. Sometimes it is not wanted to have submodules show up as changed when they just contain changes to their work tree (this was the behavior before 1.7.0). An example for that are scripts which just want to check for submodule commits while ignoring any changes to the work tree. Also users having large submodules known not to change might want to use this option, as the - sometimes substantial - time it takes to scan the submodule work tree(s) is saved when using the "dirty" parameter. And if you want to ignore any changes to submodules, you can now do that by using this option without parameters or with "all" (when the config option status.submodulesummary is set, using "all" will also suppress the output of the submodule summary). A new function handle_ignore_submodules_arg() is introduced to parse this option new to "git status" in a single location, as "git diff" already knew it. Signed-off-by: Jens Lehmann <Jens.Lehmann@web.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-06-25 14:56:47 +00:00
handle_ignore_submodules_arg(options, arg + 20);
Submodules: Add the new "ignore" config option for diff and status The new "ignore" config option controls the default behavior for "git status" and the diff family. It specifies under what circumstances they consider submodules as modified and can be set separately for each submodule. The command line option "--ignore-submodules=" has been extended to accept the new parameter "none" for both status and diff. Users that chose submodules to get rid of long work tree scanning times might want to set the "dirty" option for those submodules. This brings back the pre 1.7.0 behavior, where submodule work trees were never scanned for modifications. By using "--ignore-submodules=none" on the command line the status and diff commands can be told to do a full scan. This option can be set to the following values (which have the same name and meaning as for the "--ignore-submodules" option of status and diff): "all": All changes to the submodule will be ignored. "dirty": Only differences of the commit recorded in the superproject and the submodules HEAD will be considered modifications, all changes to the work tree of the submodule will be ignored. When using this value, the submodule will not be scanned for work tree changes at all, leading to a performance benefit on large submodules. "untracked": Only untracked files in the submodules work tree are ignored, a changed HEAD and/or modified files in the submodule will mark it as modified. "none" (which is the default): Either untracked or modified files in a submodules work tree or a difference between the subdmodules HEAD and the commit recorded in the superproject will make it show up as changed. This value is added as a new parameter for the "--ignore-submodules" option of the diff family and "git status" so the user can override the settings in the configuration. Signed-off-by: Jens Lehmann <Jens.Lehmann@web.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-08-05 22:39:25 +00:00
} else if (!strcmp(arg, "--submodule"))
DIFF_OPT_SET(options, SUBMODULE_LOG);
else if (!prefixcmp(arg, "--submodule=")) {
if (!strcmp(arg + 12, "log"))
DIFF_OPT_SET(options, SUBMODULE_LOG);
}
/* misc options */
else if (!strcmp(arg, "-z"))
options->line_termination = 0;
else if ((argcount = short_opt('l', av, &optarg))) {
options->rename_limit = strtoul(optarg, NULL, 10);
return argcount;
}
else if ((argcount = short_opt('S', av, &optarg))) {
options->pickaxe = optarg;
options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
return argcount;
} else if ((argcount = short_opt('G', av, &optarg))) {
options->pickaxe = optarg;
options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
return argcount;
}
else if (!strcmp(arg, "--pickaxe-all"))
options->pickaxe_opts |= DIFF_PICKAXE_ALL;
else if (!strcmp(arg, "--pickaxe-regex"))
options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
else if ((argcount = short_opt('O', av, &optarg))) {
options->orderfile = optarg;
return argcount;
}
else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
options->filter = optarg;
return argcount;
}
else if (!strcmp(arg, "--abbrev"))
options->abbrev = DEFAULT_ABBREV;
else if (!prefixcmp(arg, "--abbrev=")) {
options->abbrev = strtoul(arg + 9, NULL, 10);
if (options->abbrev < MINIMUM_ABBREV)
options->abbrev = MINIMUM_ABBREV;
else if (40 < options->abbrev)
options->abbrev = 40;
}
else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
options->a_prefix = optarg;
return argcount;
}
else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
options->b_prefix = optarg;
return argcount;
}
else if (!strcmp(arg, "--no-prefix"))
options->a_prefix = options->b_prefix = "";
else if (opt_arg(arg, '\0', "inter-hunk-context",
&options->interhunkcontext))
;
else if (!strcmp(arg, "-W"))
DIFF_OPT_SET(options, FUNCCONTEXT);
else if (!strcmp(arg, "--function-context"))
DIFF_OPT_SET(options, FUNCCONTEXT);
else if (!strcmp(arg, "--no-function-context"))
DIFF_OPT_CLR(options, FUNCCONTEXT);
else if ((argcount = parse_long_opt("output", av, &optarg))) {
options->file = fopen(optarg, "w");
if (!options->file)
die_errno("Could not open '%s'", optarg);
options->close_file = 1;
return argcount;
} else
return 0;
return 1;
}
int parse_rename_score(const char **cp_p)
{
unsigned long num, scale;
int ch, dot;
const char *cp = *cp_p;
num = 0;
scale = 1;
dot = 0;
for (;;) {
ch = *cp;
if ( !dot && ch == '.' ) {
scale = 1;
dot = 1;
} else if ( ch == '%' ) {
scale = dot ? scale*100 : 100;
cp++; /* % is always at the end */
break;
} else if ( ch >= '0' && ch <= '9' ) {
if ( scale < 100000 ) {
scale *= 10;
num = (num*10) + (ch-'0');
}
} else {
break;
}
cp++;
}
*cp_p = cp;
/* user says num divided by scale and we say internally that
* is MAX_SCORE * num / scale.
*/
return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
}
static int diff_scoreopt_parse(const char *opt)
{
int opt1, opt2, cmd;
if (*opt++ != '-')
return -1;
cmd = *opt++;
if (cmd == '-') {
/* convert the long-form arguments into short-form versions */
if (!prefixcmp(opt, "break-rewrites")) {
opt += strlen("break-rewrites");
if (*opt == 0 || *opt++ == '=')
cmd = 'B';
} else if (!prefixcmp(opt, "find-copies")) {
opt += strlen("find-copies");
if (*opt == 0 || *opt++ == '=')
cmd = 'C';
} else if (!prefixcmp(opt, "find-renames")) {
opt += strlen("find-renames");
if (*opt == 0 || *opt++ == '=')
cmd = 'M';
}
}
if (cmd != 'M' && cmd != 'C' && cmd != 'B')
return -1; /* that is not a -M, -C nor -B option */
opt1 = parse_rename_score(&opt);
if (cmd != 'B')
opt2 = 0;
else {
if (*opt == 0)
opt2 = 0;
else if (*opt != '/')
return -1; /* we expect -B80/99 or -B80 */
else {
opt++;
opt2 = parse_rename_score(&opt);
}
}
if (*opt != 0)
return -1;
return opt1 | (opt2 << 16);
}
struct diff_queue_struct diff_queued_diff;
void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
{
if (queue->alloc <= queue->nr) {
queue->alloc = alloc_nr(queue->alloc);
queue->queue = xrealloc(queue->queue,
sizeof(dp) * queue->alloc);
}
queue->queue[queue->nr++] = dp;
}
struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
struct diff_filespec *one,
struct diff_filespec *two)
{
struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
dp->one = one;
dp->two = two;
if (queue)
diff_q(queue, dp);
return dp;
}
void diff_free_filepair(struct diff_filepair *p)
{
free_filespec(p->one);
free_filespec(p->two);
free(p);
}
/* This is different from find_unique_abbrev() in that
* it stuffs the result with dots for alignment.
*/
const char *diff_unique_abbrev(const unsigned char *sha1, int len)
{
int abblen;
const char *abbrev;
if (len == 40)
return sha1_to_hex(sha1);
abbrev = find_unique_abbrev(sha1, len);
abblen = strlen(abbrev);
if (abblen < 37) {
static char hex[41];
if (len < abblen && abblen <= len + 2)
sprintf(hex, "%s%.*s", abbrev, len+3-abblen, "..");
else
sprintf(hex, "%s...", abbrev);
return hex;
}
return sha1_to_hex(sha1);
}
static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
{
int line_termination = opt->line_termination;
int inter_name_termination = line_termination ? '\t' : '\0';
if (opt->output_prefix) {
struct strbuf *msg = NULL;
msg = opt->output_prefix(opt, opt->output_prefix_data);
fprintf(opt->file, "%s", msg->buf);
}
if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
diff_unique_abbrev(p->one->sha1, opt->abbrev));
fprintf(opt->file, "%s ", diff_unique_abbrev(p->two->sha1, opt->abbrev));
}
if (p->score) {
fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
inter_name_termination);
} else {
fprintf(opt->file, "%c%c", p->status, inter_name_termination);
}
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
if (p->status == DIFF_STATUS_COPIED ||
p->status == DIFF_STATUS_RENAMED) {
const char *name_a, *name_b;
name_a = p->one->path;
name_b = p->two->path;
strip_prefix(opt->prefix_length, &name_a, &name_b);
write_name_quoted(name_a, opt->file, inter_name_termination);
write_name_quoted(name_b, opt->file, line_termination);
} else {
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
const char *name_a, *name_b;
name_a = p->one->mode ? p->one->path : p->two->path;
name_b = NULL;
strip_prefix(opt->prefix_length, &name_a, &name_b);
write_name_quoted(name_a, opt->file, line_termination);
}
}
int diff_unmodified_pair(struct diff_filepair *p)
{
/* This function is written stricter than necessary to support
* the currently implemented transformers, but the idea is to
* let transformers to produce diff_filepairs any way they want,
* and filter and clean them up here before producing the output.
*/
struct diff_filespec *one = p->one, *two = p->two;
if (DIFF_PAIR_UNMERGED(p))
return 0; /* unmerged is interesting */
/* deletion, addition, mode or type change
* and rename are all interesting.
*/
if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
DIFF_PAIR_MODE_CHANGED(p) ||
strcmp(one->path, two->path))
return 0;
/* both are valid and point at the same path. that is, we are
* dealing with a change.
*/
if (one->sha1_valid && two->sha1_valid &&
!hashcmp(one->sha1, two->sha1) &&
!one->dirty_submodule && !two->dirty_submodule)
return 1; /* no change */
if (!one->sha1_valid && !two->sha1_valid)
return 1; /* both look at the same file on the filesystem. */
return 0;
}
static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
{
if (diff_unmodified_pair(p))
return;
if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
(DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
return; /* no tree diffs in patch format */
run_diff(p, o);
}
static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
struct diffstat_t *diffstat)
{
if (diff_unmodified_pair(p))
return;
if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
(DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
return; /* no useful stat for tree diffs */
run_diffstat(p, o, diffstat);
}
static void diff_flush_checkdiff(struct diff_filepair *p,
struct diff_options *o)
{
if (diff_unmodified_pair(p))
return;
if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
(DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
return; /* nothing to check in tree diffs */
run_checkdiff(p, o);
}
int diff_queue_is_empty(void)
{
struct diff_queue_struct *q = &diff_queued_diff;
int i;
for (i = 0; i < q->nr; i++)
if (!diff_unmodified_pair(q->queue[i]))
return 0;
return 1;
}
#if DIFF_DEBUG
void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
{
fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
x, one ? one : "",
s->path,
DIFF_FILE_VALID(s) ? "valid" : "invalid",
s->mode,
s->sha1_valid ? sha1_to_hex(s->sha1) : "");
fprintf(stderr, "queue[%d] %s size %lu flags %d\n",
x, one ? one : "",
s->size, s->xfrm_flags);
}
void diff_debug_filepair(const struct diff_filepair *p, int i)
{
diff_debug_filespec(p->one, i, "one");
diff_debug_filespec(p->two, i, "two");
fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
p->score, p->status ? p->status : '?',
p->one->rename_used, p->broken_pair);
}
void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
{
int i;
if (msg)
fprintf(stderr, "%s\n", msg);
fprintf(stderr, "q->nr = %d\n", q->nr);
for (i = 0; i < q->nr; i++) {
struct diff_filepair *p = q->queue[i];
diff_debug_filepair(p, i);
}
}
#endif
static void diff_resolve_rename_copy(void)
{
int i;
struct diff_filepair *p;
struct diff_queue_struct *q = &diff_queued_diff;
diff_debug_queue("resolve-rename-copy", q);
for (i = 0; i < q->nr; i++) {
p = q->queue[i];
p->status = 0; /* undecided */
if (DIFF_PAIR_UNMERGED(p))
p->status = DIFF_STATUS_UNMERGED;
else if (!DIFF_FILE_VALID(p->one))
p->status = DIFF_STATUS_ADDED;
else if (!DIFF_FILE_VALID(p->two))
p->status = DIFF_STATUS_DELETED;
else if (DIFF_PAIR_TYPE_CHANGED(p))
p->status = DIFF_STATUS_TYPE_CHANGED;
/* from this point on, we are dealing with a pair
* whose both sides are valid and of the same type, i.e.
* either in-place edit or rename/copy edit.
*/
else if (DIFF_PAIR_RENAME(p)) {
/*
* A rename might have re-connected a broken
* pair up, causing the pathnames to be the
* same again. If so, that's not a rename at
* all, just a modification..
*
* Otherwise, see if this source was used for
* multiple renames, in which case we decrement
* the count, and call it a copy.
*/
if (!strcmp(p->one->path, p->two->path))
p->status = DIFF_STATUS_MODIFIED;
else if (--p->one->rename_used > 0)
p->status = DIFF_STATUS_COPIED;
else
p->status = DIFF_STATUS_RENAMED;
}
else if (hashcmp(p->one->sha1, p->two->sha1) ||
p->one->mode != p->two->mode ||
p->one->dirty_submodule ||
p->two->dirty_submodule ||
is_null_sha1(p->one->sha1))
p->status = DIFF_STATUS_MODIFIED;
else {
/* This is a "no-change" entry and should not
* happen anymore, but prepare for broken callers.
*/
error("feeding unmodified %s to diffcore",
p->one->path);
p->status = DIFF_STATUS_UNKNOWN;
}
}
diff_debug_queue("resolve-rename-copy done", q);
}
static int check_pair_status(struct diff_filepair *p)
{
switch (p->status) {
case DIFF_STATUS_UNKNOWN:
return 0;
case 0:
die("internal error in diff-resolve-rename-copy");
default:
return 1;
}
}
static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
{
int fmt = opt->output_format;
if (fmt & DIFF_FORMAT_CHECKDIFF)
diff_flush_checkdiff(p, opt);
else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
diff_flush_raw(p, opt);
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
else if (fmt & DIFF_FORMAT_NAME) {
const char *name_a, *name_b;
name_a = p->two->path;
name_b = NULL;
strip_prefix(opt->prefix_length, &name_a, &name_b);
write_name_quoted(name_a, opt->file, opt->line_termination);
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
}
}
static void show_file_mode_name(FILE *file, const char *newdelete, struct diff_filespec *fs)
{
if (fs->mode)
fprintf(file, " %s mode %06o ", newdelete, fs->mode);
else
fprintf(file, " %s ", newdelete);
write_name_quoted(fs->path, file, '\n');
}
static void show_mode_change(FILE *file, struct diff_filepair *p, int show_name,
const char *line_prefix)
{
if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
fprintf(file, "%s mode change %06o => %06o%c", line_prefix, p->one->mode,
p->two->mode, show_name ? ' ' : '\n');
if (show_name) {
write_name_quoted(p->two->path, file, '\n');
}
}
}
static void show_rename_copy(FILE *file, const char *renamecopy, struct diff_filepair *p,
const char *line_prefix)
{
char *names = pprint_rename(p->one->path, p->two->path);
fprintf(file, " %s %s (%d%%)\n", renamecopy, names, similarity_index(p));
free(names);
show_mode_change(file, p, 0, line_prefix);
}
static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
{
FILE *file = opt->file;
char *line_prefix = "";
if (opt->output_prefix) {
struct strbuf *buf = opt->output_prefix(opt, opt->output_prefix_data);
line_prefix = buf->buf;
}
switch(p->status) {
case DIFF_STATUS_DELETED:
fputs(line_prefix, file);
show_file_mode_name(file, "delete", p->one);
break;
case DIFF_STATUS_ADDED:
fputs(line_prefix, file);
show_file_mode_name(file, "create", p->two);
break;
case DIFF_STATUS_COPIED:
fputs(line_prefix, file);
show_rename_copy(file, "copy", p, line_prefix);
break;
case DIFF_STATUS_RENAMED:
fputs(line_prefix, file);
show_rename_copy(file, "rename", p, line_prefix);
break;
default:
if (p->score) {
fprintf(file, "%s rewrite ", line_prefix);
write_name_quoted(p->two->path, file, ' ');
fprintf(file, "(%d%%)\n", similarity_index(p));
}
show_mode_change(file, p, !p->score, line_prefix);
break;
}
}
struct patch_id_t {
fix openssl headers conflicting with custom SHA1 implementations On ARM I have the following compilation errors: CC fast-import.o In file included from cache.h:8, from builtin.h:6, from fast-import.c:142: arm/sha1.h:14: error: conflicting types for 'SHA_CTX' /usr/include/openssl/sha.h:105: error: previous declaration of 'SHA_CTX' was here arm/sha1.h:16: error: conflicting types for 'SHA1_Init' /usr/include/openssl/sha.h:115: error: previous declaration of 'SHA1_Init' was here arm/sha1.h:17: error: conflicting types for 'SHA1_Update' /usr/include/openssl/sha.h:116: error: previous declaration of 'SHA1_Update' was here arm/sha1.h:18: error: conflicting types for 'SHA1_Final' /usr/include/openssl/sha.h:117: error: previous declaration of 'SHA1_Final' was here make: *** [fast-import.o] Error 1 This is because openssl header files are always included in git-compat-util.h since commit 684ec6c63c whenever NO_OPENSSL is not set, which somehow brings in <openssl/sha1.h> clashing with the custom ARM version. Compilation of git is probably broken on PPC too for the same reason. Turns out that the only file requiring openssl/ssl.h and openssl/err.h is imap-send.c. But only moving those problematic includes there doesn't solve the issue as it also includes cache.h which brings in the conflicting local SHA1 header file. As suggested by Jeff King, the best solution is to rename our references to SHA1 functions and structure to something git specific, and define those according to the implementation used. Signed-off-by: Nicolas Pitre <nico@cam.org> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2008-10-01 18:05:20 +00:00
git_SHA_CTX *ctx;
int patchlen;
};
static int remove_space(char *line, int len)
{
int i;
char *dst = line;
unsigned char c;
for (i = 0; i < len; i++)
if (!isspace((c = line[i])))
*dst++ = c;
return dst - line;
}
static void patch_id_consume(void *priv, char *line, unsigned long len)
{
struct patch_id_t *data = priv;
int new_len;
/* Ignore line numbers when computing the SHA1 of the patch */
if (!prefixcmp(line, "@@ -"))
return;
new_len = remove_space(line, len);
fix openssl headers conflicting with custom SHA1 implementations On ARM I have the following compilation errors: CC fast-import.o In file included from cache.h:8, from builtin.h:6, from fast-import.c:142: arm/sha1.h:14: error: conflicting types for 'SHA_CTX' /usr/include/openssl/sha.h:105: error: previous declaration of 'SHA_CTX' was here arm/sha1.h:16: error: conflicting types for 'SHA1_Init' /usr/include/openssl/sha.h:115: error: previous declaration of 'SHA1_Init' was here arm/sha1.h:17: error: conflicting types for 'SHA1_Update' /usr/include/openssl/sha.h:116: error: previous declaration of 'SHA1_Update' was here arm/sha1.h:18: error: conflicting types for 'SHA1_Final' /usr/include/openssl/sha.h:117: error: previous declaration of 'SHA1_Final' was here make: *** [fast-import.o] Error 1 This is because openssl header files are always included in git-compat-util.h since commit 684ec6c63c whenever NO_OPENSSL is not set, which somehow brings in <openssl/sha1.h> clashing with the custom ARM version. Compilation of git is probably broken on PPC too for the same reason. Turns out that the only file requiring openssl/ssl.h and openssl/err.h is imap-send.c. But only moving those problematic includes there doesn't solve the issue as it also includes cache.h which brings in the conflicting local SHA1 header file. As suggested by Jeff King, the best solution is to rename our references to SHA1 functions and structure to something git specific, and define those according to the implementation used. Signed-off-by: Nicolas Pitre <nico@cam.org> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2008-10-01 18:05:20 +00:00
git_SHA1_Update(data->ctx, line, new_len);
data->patchlen += new_len;
}
/* returns 0 upon success, and writes result into sha1 */
static int diff_get_patch_id(struct diff_options *options, unsigned char *sha1)
{
struct diff_queue_struct *q = &diff_queued_diff;
int i;
fix openssl headers conflicting with custom SHA1 implementations On ARM I have the following compilation errors: CC fast-import.o In file included from cache.h:8, from builtin.h:6, from fast-import.c:142: arm/sha1.h:14: error: conflicting types for 'SHA_CTX' /usr/include/openssl/sha.h:105: error: previous declaration of 'SHA_CTX' was here arm/sha1.h:16: error: conflicting types for 'SHA1_Init' /usr/include/openssl/sha.h:115: error: previous declaration of 'SHA1_Init' was here arm/sha1.h:17: error: conflicting types for 'SHA1_Update' /usr/include/openssl/sha.h:116: error: previous declaration of 'SHA1_Update' was here arm/sha1.h:18: error: conflicting types for 'SHA1_Final' /usr/include/openssl/sha.h:117: error: previous declaration of 'SHA1_Final' was here make: *** [fast-import.o] Error 1 This is because openssl header files are always included in git-compat-util.h since commit 684ec6c63c whenever NO_OPENSSL is not set, which somehow brings in <openssl/sha1.h> clashing with the custom ARM version. Compilation of git is probably broken on PPC too for the same reason. Turns out that the only file requiring openssl/ssl.h and openssl/err.h is imap-send.c. But only moving those problematic includes there doesn't solve the issue as it also includes cache.h which brings in the conflicting local SHA1 header file. As suggested by Jeff King, the best solution is to rename our references to SHA1 functions and structure to something git specific, and define those according to the implementation used. Signed-off-by: Nicolas Pitre <nico@cam.org> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2008-10-01 18:05:20 +00:00
git_SHA_CTX ctx;
struct patch_id_t data;
char buffer[PATH_MAX * 4 + 20];
fix openssl headers conflicting with custom SHA1 implementations On ARM I have the following compilation errors: CC fast-import.o In file included from cache.h:8, from builtin.h:6, from fast-import.c:142: arm/sha1.h:14: error: conflicting types for 'SHA_CTX' /usr/include/openssl/sha.h:105: error: previous declaration of 'SHA_CTX' was here arm/sha1.h:16: error: conflicting types for 'SHA1_Init' /usr/include/openssl/sha.h:115: error: previous declaration of 'SHA1_Init' was here arm/sha1.h:17: error: conflicting types for 'SHA1_Update' /usr/include/openssl/sha.h:116: error: previous declaration of 'SHA1_Update' was here arm/sha1.h:18: error: conflicting types for 'SHA1_Final' /usr/include/openssl/sha.h:117: error: previous declaration of 'SHA1_Final' was here make: *** [fast-import.o] Error 1 This is because openssl header files are always included in git-compat-util.h since commit 684ec6c63c whenever NO_OPENSSL is not set, which somehow brings in <openssl/sha1.h> clashing with the custom ARM version. Compilation of git is probably broken on PPC too for the same reason. Turns out that the only file requiring openssl/ssl.h and openssl/err.h is imap-send.c. But only moving those problematic includes there doesn't solve the issue as it also includes cache.h which brings in the conflicting local SHA1 header file. As suggested by Jeff King, the best solution is to rename our references to SHA1 functions and structure to something git specific, and define those according to the implementation used. Signed-off-by: Nicolas Pitre <nico@cam.org> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2008-10-01 18:05:20 +00:00
git_SHA1_Init(&ctx);
memset(&data, 0, sizeof(struct patch_id_t));
data.ctx = &ctx;
for (i = 0; i < q->nr; i++) {
xpparam_t xpp;
xdemitconf_t xecfg;
mmfile_t mf1, mf2;
struct diff_filepair *p = q->queue[i];
int len1, len2;
memset(&xpp, 0, sizeof(xpp));
memset(&xecfg, 0, sizeof(xecfg));
if (p->status == 0)
return error("internal diff status error");
if (p->status == DIFF_STATUS_UNKNOWN)
continue;
if (diff_unmodified_pair(p))
continue;
if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
(DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
continue;
if (DIFF_PAIR_UNMERGED(p))
continue;
diff_fill_sha1_info(p->one);
diff_fill_sha1_info(p->two);
if (fill_mmfile(&mf1, p->one) < 0 ||
fill_mmfile(&mf2, p->two) < 0)
return error("unable to read files to diff");
len1 = remove_space(p->one->path, strlen(p->one->path));
len2 = remove_space(p->two->path, strlen(p->two->path));
if (p->one->mode == 0)
len1 = snprintf(buffer, sizeof(buffer),
"diff--gita/%.*sb/%.*s"
"newfilemode%06o"
"---/dev/null"
"+++b/%.*s",
len1, p->one->path,
len2, p->two->path,
p->two->mode,
len2, p->two->path);
else if (p->two->mode == 0)
len1 = snprintf(buffer, sizeof(buffer),
"diff--gita/%.*sb/%.*s"
"deletedfilemode%06o"
"---a/%.*s"
"+++/dev/null",
len1, p->one->path,
len2, p->two->path,
p->one->mode,
len1, p->one->path);
else
len1 = snprintf(buffer, sizeof(buffer),
"diff--gita/%.*sb/%.*s"
"---a/%.*s"
"+++b/%.*s",
len1, p->one->path,
len2, p->two->path,
len1, p->one->path,
len2, p->two->path);
fix openssl headers conflicting with custom SHA1 implementations On ARM I have the following compilation errors: CC fast-import.o In file included from cache.h:8, from builtin.h:6, from fast-import.c:142: arm/sha1.h:14: error: conflicting types for 'SHA_CTX' /usr/include/openssl/sha.h:105: error: previous declaration of 'SHA_CTX' was here arm/sha1.h:16: error: conflicting types for 'SHA1_Init' /usr/include/openssl/sha.h:115: error: previous declaration of 'SHA1_Init' was here arm/sha1.h:17: error: conflicting types for 'SHA1_Update' /usr/include/openssl/sha.h:116: error: previous declaration of 'SHA1_Update' was here arm/sha1.h:18: error: conflicting types for 'SHA1_Final' /usr/include/openssl/sha.h:117: error: previous declaration of 'SHA1_Final' was here make: *** [fast-import.o] Error 1 This is because openssl header files are always included in git-compat-util.h since commit 684ec6c63c whenever NO_OPENSSL is not set, which somehow brings in <openssl/sha1.h> clashing with the custom ARM version. Compilation of git is probably broken on PPC too for the same reason. Turns out that the only file requiring openssl/ssl.h and openssl/err.h is imap-send.c. But only moving those problematic includes there doesn't solve the issue as it also includes cache.h which brings in the conflicting local SHA1 header file. As suggested by Jeff King, the best solution is to rename our references to SHA1 functions and structure to something git specific, and define those according to the implementation used. Signed-off-by: Nicolas Pitre <nico@cam.org> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2008-10-01 18:05:20 +00:00
git_SHA1_Update(&ctx, buffer, len1);
if (diff_filespec_is_binary(p->one) ||
diff_filespec_is_binary(p->two)) {
git_SHA1_Update(&ctx, sha1_to_hex(p->one->sha1), 40);
git_SHA1_Update(&ctx, sha1_to_hex(p->two->sha1), 40);
continue;
}
xpp.flags = 0;
xecfg.ctxlen = 3;
xecfg.flags = 0;
xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
&xpp, &xecfg);
}
fix openssl headers conflicting with custom SHA1 implementations On ARM I have the following compilation errors: CC fast-import.o In file included from cache.h:8, from builtin.h:6, from fast-import.c:142: arm/sha1.h:14: error: conflicting types for 'SHA_CTX' /usr/include/openssl/sha.h:105: error: previous declaration of 'SHA_CTX' was here arm/sha1.h:16: error: conflicting types for 'SHA1_Init' /usr/include/openssl/sha.h:115: error: previous declaration of 'SHA1_Init' was here arm/sha1.h:17: error: conflicting types for 'SHA1_Update' /usr/include/openssl/sha.h:116: error: previous declaration of 'SHA1_Update' was here arm/sha1.h:18: error: conflicting types for 'SHA1_Final' /usr/include/openssl/sha.h:117: error: previous declaration of 'SHA1_Final' was here make: *** [fast-import.o] Error 1 This is because openssl header files are always included in git-compat-util.h since commit 684ec6c63c whenever NO_OPENSSL is not set, which somehow brings in <openssl/sha1.h> clashing with the custom ARM version. Compilation of git is probably broken on PPC too for the same reason. Turns out that the only file requiring openssl/ssl.h and openssl/err.h is imap-send.c. But only moving those problematic includes there doesn't solve the issue as it also includes cache.h which brings in the conflicting local SHA1 header file. As suggested by Jeff King, the best solution is to rename our references to SHA1 functions and structure to something git specific, and define those according to the implementation used. Signed-off-by: Nicolas Pitre <nico@cam.org> Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
2008-10-01 18:05:20 +00:00
git_SHA1_Final(sha1, &ctx);
return 0;
}
int diff_flush_patch_id(struct diff_options *options, unsigned char *sha1)
{
struct diff_queue_struct *q = &diff_queued_diff;
int i;
int result = diff_get_patch_id(options, sha1);
for (i = 0; i < q->nr; i++)
diff_free_filepair(q->queue[i]);
free(q->queue);
DIFF_QUEUE_CLEAR(q);
return result;
}
static int is_summary_empty(const struct diff_queue_struct *q)
{
int i;
for (i = 0; i < q->nr; i++) {
const struct diff_filepair *p = q->queue[i];
switch (p->status) {
case DIFF_STATUS_DELETED:
case DIFF_STATUS_ADDED:
case DIFF_STATUS_COPIED:
case DIFF_STATUS_RENAMED:
return 0;
default:
if (p->score)
return 0;
if (p->one->mode && p->two->mode &&
p->one->mode != p->two->mode)
return 0;
break;
}
}
return 1;
}
static const char rename_limit_warning[] =
"inexact rename detection was skipped due to too many files.";
static const char degrade_cc_to_c_warning[] =
"only found copies from modified paths due to too many files.";
static const char rename_limit_advice[] =
"you may want to set your %s variable to at least "
"%d and retry the command.";
void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
{
if (degraded_cc)
warning(degrade_cc_to_c_warning);
else if (needed)
warning(rename_limit_warning);
else
return;
if (0 < needed && needed < 32767)
warning(rename_limit_advice, varname, needed);
}
void diff_flush(struct diff_options *options)
{
struct diff_queue_struct *q = &diff_queued_diff;
int i, output_format = options->output_format;
int separator = 0;
New --dirstat=lines mode, doing dirstat analysis based on diffstat This patch adds an alternative implementation of show_dirstat(), called show_dirstat_by_line(), which uses the more expensive diffstat analysis (as opposed to show_dirstat()'s own (relatively inexpensive) analysis) to derive the numbers from which the --dirstat output is computed. The alternative implementation is controlled by the new "lines" parameter to the --dirstat option (or the diff.dirstat config variable). For binary files, the diffstat analysis counts bytes instead of lines, so to prevent binary files from dominating the dirstat results, the byte counts for binary files are divided by 64 before being compared to their textual/line-based counterparts. This is a stupid and ugly - but very cheap - heuristic. In linux-2.6.git, running the three different --dirstat modes: time git diff v2.6.20..v2.6.30 --dirstat=changes > /dev/null vs. time git diff v2.6.20..v2.6.30 --dirstat=lines > /dev/null vs. time git diff v2.6.20..v2.6.30 --dirstat=files > /dev/null yields the following average runtimes on my machine: - "changes" (default): ~6.0 s - "lines": ~9.6 s - "files": ~0.1 s So, as expected, there's a considerable performance hit (~60%) by going through the full diffstat analysis as compared to the default "changes" analysis (obviously, "files" is much faster than both). As such, the "lines" mode is probably only useful if you really need the --dirstat numbers to be consistent with the numbers returned from the other --*stat options. The patch also includes documentation and tests for the new dirstat mode. Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:21 +00:00
int dirstat_by_line = 0;
/*
* Order: raw, stat, summary, patch
* or: name/name-status/checkdiff (other bits clear)
*/
if (!q->nr)
goto free_queue;
if (output_format & (DIFF_FORMAT_RAW |
DIFF_FORMAT_NAME |
DIFF_FORMAT_NAME_STATUS |
DIFF_FORMAT_CHECKDIFF)) {
for (i = 0; i < q->nr; i++) {
struct diff_filepair *p = q->queue[i];
if (check_pair_status(p))
flush_one_pair(p, options);
}
separator++;
}
New --dirstat=lines mode, doing dirstat analysis based on diffstat This patch adds an alternative implementation of show_dirstat(), called show_dirstat_by_line(), which uses the more expensive diffstat analysis (as opposed to show_dirstat()'s own (relatively inexpensive) analysis) to derive the numbers from which the --dirstat output is computed. The alternative implementation is controlled by the new "lines" parameter to the --dirstat option (or the diff.dirstat config variable). For binary files, the diffstat analysis counts bytes instead of lines, so to prevent binary files from dominating the dirstat results, the byte counts for binary files are divided by 64 before being compared to their textual/line-based counterparts. This is a stupid and ugly - but very cheap - heuristic. In linux-2.6.git, running the three different --dirstat modes: time git diff v2.6.20..v2.6.30 --dirstat=changes > /dev/null vs. time git diff v2.6.20..v2.6.30 --dirstat=lines > /dev/null vs. time git diff v2.6.20..v2.6.30 --dirstat=files > /dev/null yields the following average runtimes on my machine: - "changes" (default): ~6.0 s - "lines": ~9.6 s - "files": ~0.1 s So, as expected, there's a considerable performance hit (~60%) by going through the full diffstat analysis as compared to the default "changes" analysis (obviously, "files" is much faster than both). As such, the "lines" mode is probably only useful if you really need the --dirstat numbers to be consistent with the numbers returned from the other --*stat options. The patch also includes documentation and tests for the new dirstat mode. Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:21 +00:00
if (output_format & DIFF_FORMAT_DIRSTAT && DIFF_OPT_TST(options, DIRSTAT_BY_LINE))
dirstat_by_line = 1;
if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
dirstat_by_line) {
struct diffstat_t diffstat;
memset(&diffstat, 0, sizeof(struct diffstat_t));
for (i = 0; i < q->nr; i++) {
struct diff_filepair *p = q->queue[i];
if (check_pair_status(p))
diff_flush_stat(p, options, &diffstat);
}
if (output_format & DIFF_FORMAT_NUMSTAT)
show_numstat(&diffstat, options);
if (output_format & DIFF_FORMAT_DIFFSTAT)
show_stats(&diffstat, options);
if (output_format & DIFF_FORMAT_SHORTSTAT)
show_shortstats(&diffstat, options);
New --dirstat=lines mode, doing dirstat analysis based on diffstat This patch adds an alternative implementation of show_dirstat(), called show_dirstat_by_line(), which uses the more expensive diffstat analysis (as opposed to show_dirstat()'s own (relatively inexpensive) analysis) to derive the numbers from which the --dirstat output is computed. The alternative implementation is controlled by the new "lines" parameter to the --dirstat option (or the diff.dirstat config variable). For binary files, the diffstat analysis counts bytes instead of lines, so to prevent binary files from dominating the dirstat results, the byte counts for binary files are divided by 64 before being compared to their textual/line-based counterparts. This is a stupid and ugly - but very cheap - heuristic. In linux-2.6.git, running the three different --dirstat modes: time git diff v2.6.20..v2.6.30 --dirstat=changes > /dev/null vs. time git diff v2.6.20..v2.6.30 --dirstat=lines > /dev/null vs. time git diff v2.6.20..v2.6.30 --dirstat=files > /dev/null yields the following average runtimes on my machine: - "changes" (default): ~6.0 s - "lines": ~9.6 s - "files": ~0.1 s So, as expected, there's a considerable performance hit (~60%) by going through the full diffstat analysis as compared to the default "changes" analysis (obviously, "files" is much faster than both). As such, the "lines" mode is probably only useful if you really need the --dirstat numbers to be consistent with the numbers returned from the other --*stat options. The patch also includes documentation and tests for the new dirstat mode. Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:21 +00:00
if (output_format & DIFF_FORMAT_DIRSTAT)
show_dirstat_by_line(&diffstat, options);
free_diffstat_info(&diffstat);
separator++;
}
New --dirstat=lines mode, doing dirstat analysis based on diffstat This patch adds an alternative implementation of show_dirstat(), called show_dirstat_by_line(), which uses the more expensive diffstat analysis (as opposed to show_dirstat()'s own (relatively inexpensive) analysis) to derive the numbers from which the --dirstat output is computed. The alternative implementation is controlled by the new "lines" parameter to the --dirstat option (or the diff.dirstat config variable). For binary files, the diffstat analysis counts bytes instead of lines, so to prevent binary files from dominating the dirstat results, the byte counts for binary files are divided by 64 before being compared to their textual/line-based counterparts. This is a stupid and ugly - but very cheap - heuristic. In linux-2.6.git, running the three different --dirstat modes: time git diff v2.6.20..v2.6.30 --dirstat=changes > /dev/null vs. time git diff v2.6.20..v2.6.30 --dirstat=lines > /dev/null vs. time git diff v2.6.20..v2.6.30 --dirstat=files > /dev/null yields the following average runtimes on my machine: - "changes" (default): ~6.0 s - "lines": ~9.6 s - "files": ~0.1 s So, as expected, there's a considerable performance hit (~60%) by going through the full diffstat analysis as compared to the default "changes" analysis (obviously, "files" is much faster than both). As such, the "lines" mode is probably only useful if you really need the --dirstat numbers to be consistent with the numbers returned from the other --*stat options. The patch also includes documentation and tests for the new dirstat mode. Improved-by: Junio C Hamano <gitster@pobox.com> Signed-off-by: Johan Herland <johan@herland.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2011-04-29 09:36:21 +00:00
if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
show_dirstat(options);
if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
for (i = 0; i < q->nr; i++) {
diff_summary(options, q->queue[i]);
}
separator++;
}
if (output_format & DIFF_FORMAT_NO_OUTPUT &&
DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
/*
* run diff_flush_patch for the exit status. setting
* options->file to /dev/null should be safe, becaue we
* aren't supposed to produce any output anyway.
*/
if (options->close_file)
fclose(options->file);
options->file = fopen("/dev/null", "w");
if (!options->file)
die_errno("Could not open /dev/null");
options->close_file = 1;
for (i = 0; i < q->nr; i++) {
struct diff_filepair *p = q->queue[i];
if (check_pair_status(p))
diff_flush_patch(p, options);
if (options->found_changes)
break;
}
}
if (output_format & DIFF_FORMAT_PATCH) {
if (separator) {
putc(options->line_termination, options->file);
if (options->stat_sep) {
/* attach patch instead of inline */
fputs(options->stat_sep, options->file);
}
}
for (i = 0; i < q->nr; i++) {
struct diff_filepair *p = q->queue[i];
if (check_pair_status(p))
diff_flush_patch(p, options);
}
}
if (output_format & DIFF_FORMAT_CALLBACK)
options->format_callback(q, options, options->format_callback_data);
for (i = 0; i < q->nr; i++)
diff_free_filepair(q->queue[i]);
free_queue:
free(q->queue);
DIFF_QUEUE_CLEAR(q);
if (options->close_file)
fclose(options->file);
/*
* Report the content-level differences with HAS_CHANGES;
* diff_addremove/diff_change does not set the bit when
* DIFF_FROM_CONTENTS is in effect (e.g. with -w).
*/
if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
if (options->found_changes)
DIFF_OPT_SET(options, HAS_CHANGES);
else
DIFF_OPT_CLR(options, HAS_CHANGES);
}
}
static void diffcore_apply_filter(const char *filter)
{
int i;
struct diff_queue_struct *q = &diff_queued_diff;
struct diff_queue_struct outq;
DIFF_QUEUE_CLEAR(&outq);
if (!filter)
return;
if (strchr(filter, DIFF_STATUS_FILTER_AON)) {
int found;
for (i = found = 0; !found && i < q->nr; i++) {
struct diff_filepair *p = q->queue[i];
if (((p->status == DIFF_STATUS_MODIFIED) &&
((p->score &&
strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
(!p->score &&
strchr(filter, DIFF_STATUS_MODIFIED)))) ||
((p->status != DIFF_STATUS_MODIFIED) &&
strchr(filter, p->status)))
found++;
}
if (found)
return;
/* otherwise we will clear the whole queue
* by copying the empty outq at the end of this
* function, but first clear the current entries
* in the queue.
*/
for (i = 0; i < q->nr; i++)
diff_free_filepair(q->queue[i]);
}
else {
/* Only the matching ones */
for (i = 0; i < q->nr; i++) {
struct diff_filepair *p = q->queue[i];
if (((p->status == DIFF_STATUS_MODIFIED) &&
((p->score &&
strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
(!p->score &&
strchr(filter, DIFF_STATUS_MODIFIED)))) ||
((p->status != DIFF_STATUS_MODIFIED) &&
strchr(filter, p->status)))
diff_q(&outq, p);
else
diff_free_filepair(p);
}
}
free(q->queue);
*q = outq;
}
/* Check whether two filespecs with the same mode and size are identical */
static int diff_filespec_is_identical(struct diff_filespec *one,
struct diff_filespec *two)
{
if (S_ISGITLINK(one->mode))
return 0;
if (diff_populate_filespec(one, 0))
return 0;
if (diff_populate_filespec(two, 0))
return 0;
return !memcmp(one->data, two->data, one->size);
}
static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
{
int i;
struct diff_queue_struct *q = &diff_queued_diff;
struct diff_queue_struct outq;
DIFF_QUEUE_CLEAR(&outq);
for (i = 0; i < q->nr; i++) {
struct diff_filepair *p = q->queue[i];
/*
* 1. Entries that come from stat info dirtiness
* always have both sides (iow, not create/delete),
* one side of the object name is unknown, with
* the same mode and size. Keep the ones that
* do not match these criteria. They have real
* differences.
*
* 2. At this point, the file is known to be modified,
* with the same mode and size, and the object
* name of one side is unknown. Need to inspect
* the identical contents.
*/
if (!DIFF_FILE_VALID(p->one) || /* (1) */
!DIFF_FILE_VALID(p->two) ||
(p->one->sha1_valid && p->two->sha1_valid) ||
(p->one->mode != p->two->mode) ||
diff_populate_filespec(p->one, 1) ||
diff_populate_filespec(p->two, 1) ||
(p->one->size != p->two->size) ||
!diff_filespec_is_identical(p->one, p->two)) /* (2) */
diff_q(&outq, p);
else {
/*
* The caller can subtract 1 from skip_stat_unmatch
* to determine how many paths were dirty only
* due to stat info mismatch.
*/
if (!DIFF_OPT_TST(diffopt, NO_INDEX))
diffopt->skip_stat_unmatch++;
diff_free_filepair(p);
}
}
free(q->queue);
*q = outq;
}
unpack-trees.c: look ahead in the index This makes the traversal of index be in sync with the tree traversal. When unpack_callback() is fed a set of tree entries from trees, it inspects the name of the entry and checks if the an index entry with the same name could be hiding behind the current index entry, and (1) if the name appears in the index as a leaf node, it is also fed to the n_way_merge() callback function; (2) if the name is a directory in the index, i.e. there are entries in that are underneath it, then nothing is fed to the n_way_merge() callback function; (3) otherwise, if the name comes before the first eligible entry in the index, the index entry is first unpacked alone. When traverse_trees_recursive() descends into a subdirectory, the cache_bottom pointer is moved to walk index entries within that directory. All of these are omitted for diff-index, which does not even want to be fed an index entry and a tree entry with D/F conflicts. This fixes 3-way read-tree and exposes a bug in other parts of the system in t6035, test #5. The test prepares these three trees: O = HEAD^ 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x A = HEAD 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b/c/d 100644 blob 587be6b4c3f93f93c489c0111bba5596147a26cb a/x B = master 120000 blob a36b77384451ea1de7bd340ffca868249626bc52 a/b 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/b-2/c/d 100644 blob e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 a/x With a clean index that matches HEAD, running git read-tree -m -u --aggressive $O $A $B now yields 120000 a36b77384451ea1de7bd340ffca868249626bc52 3 a/b 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 0 a/b-2/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 1 a/b/c/d 100644 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391 2 a/b/c/d 100644 587be6b4c3f93f93c489c0111bba5596147a26cb 0 a/x which is correct. "master" created "a/b" symlink that did not exist, and removed "a/b/c/d" while HEAD did not do touch either path. Before this series, read-tree did not notice the situation and resolved addition of "a/b" and removal of "a/b/c/d" independently. If A = HEAD had another path "a/b/c/e" added, this merge should conflict but instead it silently resolved "a/b" and then immediately overwrote it to add "a/b/c/e", which was quite bogus. Tests in t1012 start to work with this. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-09-20 07:03:39 +00:00
static int diffnamecmp(const void *a_, const void *b_)
{
const struct diff_filepair *a = *((const struct diff_filepair **)a_);
const struct diff_filepair *b = *((const struct diff_filepair **)b_);
const char *name_a, *name_b;
name_a = a->one ? a->one->path : a->two->path;
name_b = b->one ? b->one->path : b->two->path;
return strcmp(name_a, name_b);
}
void diffcore_fix_diff_index(struct diff_options *options)
{
struct diff_queue_struct *q = &diff_queued_diff;
qsort(q->queue, q->nr, sizeof(q->queue[0]), diffnamecmp);
}
void diffcore_std(struct diff_options *options)
{
if (options->skip_stat_unmatch)
diffcore_skip_stat_unmatch(options);
if (!options->found_follow) {
/* See try_to_follow_renames() in tree-diff.c */
if (options->break_opt != -1)
diffcore_break(options->break_opt);
if (options->detect_rename)
diffcore_rename(options);
if (options->break_opt != -1)
diffcore_merge_broken();
}
if (options->pickaxe)
diffcore_pickaxe(options);
if (options->orderfile)
diffcore_order(options->orderfile);
if (!options->found_follow)
/* See try_to_follow_renames() in tree-diff.c */
diff_resolve_rename_copy();
diffcore_apply_filter(options->filter);
if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
DIFF_OPT_SET(options, HAS_CHANGES);
else
DIFF_OPT_CLR(options, HAS_CHANGES);
options->found_follow = 0;
}
int diff_result_code(struct diff_options *opt, int status)
{
int result = 0;
diff_warn_rename_limit("diff.renamelimit",
opt->needed_rename_limit,
opt->degraded_cc_to_c);
if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
!(opt->output_format & DIFF_FORMAT_CHECKDIFF))
return status;
if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
DIFF_OPT_TST(opt, HAS_CHANGES))
result |= 01;
if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
DIFF_OPT_TST(opt, CHECK_FAILED))
result |= 02;
return result;
}
int diff_can_quit_early(struct diff_options *opt)
{
return (DIFF_OPT_TST(opt, QUICK) &&
!opt->filter &&
DIFF_OPT_TST(opt, HAS_CHANGES));
}
Submodules: Add the new "ignore" config option for diff and status The new "ignore" config option controls the default behavior for "git status" and the diff family. It specifies under what circumstances they consider submodules as modified and can be set separately for each submodule. The command line option "--ignore-submodules=" has been extended to accept the new parameter "none" for both status and diff. Users that chose submodules to get rid of long work tree scanning times might want to set the "dirty" option for those submodules. This brings back the pre 1.7.0 behavior, where submodule work trees were never scanned for modifications. By using "--ignore-submodules=none" on the command line the status and diff commands can be told to do a full scan. This option can be set to the following values (which have the same name and meaning as for the "--ignore-submodules" option of status and diff): "all": All changes to the submodule will be ignored. "dirty": Only differences of the commit recorded in the superproject and the submodules HEAD will be considered modifications, all changes to the work tree of the submodule will be ignored. When using this value, the submodule will not be scanned for work tree changes at all, leading to a performance benefit on large submodules. "untracked": Only untracked files in the submodules work tree are ignored, a changed HEAD and/or modified files in the submodule will mark it as modified. "none" (which is the default): Either untracked or modified files in a submodules work tree or a difference between the subdmodules HEAD and the commit recorded in the superproject will make it show up as changed. This value is added as a new parameter for the "--ignore-submodules" option of the diff family and "git status" so the user can override the settings in the configuration. Signed-off-by: Jens Lehmann <Jens.Lehmann@web.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-08-05 22:39:25 +00:00
/*
* Shall changes to this submodule be ignored?
*
* Submodule changes can be configured to be ignored separately for each path,
* but that configuration can be overridden from the command line.
*/
static int is_submodule_ignored(const char *path, struct diff_options *options)
{
int ignored = 0;
unsigned orig_flags = options->flags;
if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
set_diffopt_flags_from_submodule_config(options, path);
if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
ignored = 1;
options->flags = orig_flags;
return ignored;
}
void diff_addremove(struct diff_options *options,
int addremove, unsigned mode,
const unsigned char *sha1,
const char *concatpath, unsigned dirty_submodule)
{
struct diff_filespec *one, *two;
Submodules: Add the new "ignore" config option for diff and status The new "ignore" config option controls the default behavior for "git status" and the diff family. It specifies under what circumstances they consider submodules as modified and can be set separately for each submodule. The command line option "--ignore-submodules=" has been extended to accept the new parameter "none" for both status and diff. Users that chose submodules to get rid of long work tree scanning times might want to set the "dirty" option for those submodules. This brings back the pre 1.7.0 behavior, where submodule work trees were never scanned for modifications. By using "--ignore-submodules=none" on the command line the status and diff commands can be told to do a full scan. This option can be set to the following values (which have the same name and meaning as for the "--ignore-submodules" option of status and diff): "all": All changes to the submodule will be ignored. "dirty": Only differences of the commit recorded in the superproject and the submodules HEAD will be considered modifications, all changes to the work tree of the submodule will be ignored. When using this value, the submodule will not be scanned for work tree changes at all, leading to a performance benefit on large submodules. "untracked": Only untracked files in the submodules work tree are ignored, a changed HEAD and/or modified files in the submodule will mark it as modified. "none" (which is the default): Either untracked or modified files in a submodules work tree or a difference between the subdmodules HEAD and the commit recorded in the superproject will make it show up as changed. This value is added as a new parameter for the "--ignore-submodules" option of the diff family and "git status" so the user can override the settings in the configuration. Signed-off-by: Jens Lehmann <Jens.Lehmann@web.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-08-05 22:39:25 +00:00
if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
return;
/* This may look odd, but it is a preparation for
* feeding "there are unchanged files which should
* not produce diffs, but when you are doing copy
* detection you would need them, so here they are"
* entries to the diff-core. They will be prefixed
* with something like '=' or '*' (I haven't decided
* which but should not make any difference).
* Feeding the same new and old to diff_change()
* also has the same effect.
* Before the final output happens, they are pruned after
* merged into rename/copy pairs as appropriate.
*/
if (DIFF_OPT_TST(options, REVERSE_DIFF))
addremove = (addremove == '+' ? '-' :
addremove == '-' ? '+' : addremove);
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
if (options->prefix &&
strncmp(concatpath, options->prefix, options->prefix_length))
return;
one = alloc_filespec(concatpath);
two = alloc_filespec(concatpath);
if (addremove != '+')
fill_filespec(one, sha1, mode);
if (addremove != '-') {
fill_filespec(two, sha1, mode);
two->dirty_submodule = dirty_submodule;
}
diff_queue(&diff_queued_diff, one, two);
if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
DIFF_OPT_SET(options, HAS_CHANGES);
}
void diff_change(struct diff_options *options,
unsigned old_mode, unsigned new_mode,
const unsigned char *old_sha1,
const unsigned char *new_sha1,
const char *concatpath,
unsigned old_dirty_submodule, unsigned new_dirty_submodule)
{
struct diff_filespec *one, *two;
Submodules: Add the new "ignore" config option for diff and status The new "ignore" config option controls the default behavior for "git status" and the diff family. It specifies under what circumstances they consider submodules as modified and can be set separately for each submodule. The command line option "--ignore-submodules=" has been extended to accept the new parameter "none" for both status and diff. Users that chose submodules to get rid of long work tree scanning times might want to set the "dirty" option for those submodules. This brings back the pre 1.7.0 behavior, where submodule work trees were never scanned for modifications. By using "--ignore-submodules=none" on the command line the status and diff commands can be told to do a full scan. This option can be set to the following values (which have the same name and meaning as for the "--ignore-submodules" option of status and diff): "all": All changes to the submodule will be ignored. "dirty": Only differences of the commit recorded in the superproject and the submodules HEAD will be considered modifications, all changes to the work tree of the submodule will be ignored. When using this value, the submodule will not be scanned for work tree changes at all, leading to a performance benefit on large submodules. "untracked": Only untracked files in the submodules work tree are ignored, a changed HEAD and/or modified files in the submodule will mark it as modified. "none" (which is the default): Either untracked or modified files in a submodules work tree or a difference between the subdmodules HEAD and the commit recorded in the superproject will make it show up as changed. This value is added as a new parameter for the "--ignore-submodules" option of the diff family and "git status" so the user can override the settings in the configuration. Signed-off-by: Jens Lehmann <Jens.Lehmann@web.de> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-08-05 22:39:25 +00:00
if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
is_submodule_ignored(concatpath, options))
return;
if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
unsigned tmp;
const unsigned char *tmp_c;
tmp = old_mode; old_mode = new_mode; new_mode = tmp;
tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
tmp = old_dirty_submodule; old_dirty_submodule = new_dirty_submodule;
new_dirty_submodule = tmp;
}
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
if (options->prefix &&
strncmp(concatpath, options->prefix, options->prefix_length))
return;
one = alloc_filespec(concatpath);
two = alloc_filespec(concatpath);
fill_filespec(one, old_sha1, old_mode);
fill_filespec(two, new_sha1, new_mode);
one->dirty_submodule = old_dirty_submodule;
two->dirty_submodule = new_dirty_submodule;
diff_queue(&diff_queued_diff, one, two);
if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
DIFF_OPT_SET(options, HAS_CHANGES);
}
struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
{
struct diff_filepair *pair;
struct diff_filespec *one, *two;
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
if (options->prefix &&
strncmp(path, options->prefix, options->prefix_length))
return NULL;
diff --relative: output paths as relative to the current subdirectory This adds --relative option to the diff family. When you start from a subdirectory: $ git diff --relative shows only the diff that is inside your current subdirectory, and without $prefix part. People who usually live in subdirectories may like it. There are a few things I should also mention about the change: - This works not just with diff but also works with the log family of commands, but the history pruning is not affected. In other words, if you go to a subdirectory, you can say: $ git log --relative -p but it will show the log message even for commits that do not touch the current directory. You can limit it by giving pathspec yourself: $ git log --relative -p . This originally was not a conscious design choice, but we have a way to affect diff pathspec and pruning pathspec independently. IOW "git log --full-diff -p ." tells it to prune history to commits that affect the current subdirectory but show the changes with full context. I think it makes more sense to leave pruning independent from --relative than the obvious alternative of always pruning with the current subdirectory, which would break the symmetry. - Because this works also with the log family, you could format-patch a single change, limiting the effect to your subdirectory, like so: $ cd gitk-git $ git format-patch -1 --relative 911f1eb But because that is a special purpose usage, this option will never become the default, with or without repository or user preference configuration. The risk of producing a partial patch and sending it out by mistake is too great if we did so. - This is inherently incompatible with --no-index, which is a bolted-on hack that does not have much to do with git itself. I didn't bother checking and erroring out on the combined use of the options, but probably I should. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2008-02-12 22:26:02 +00:00
one = alloc_filespec(path);
two = alloc_filespec(path);
pair = diff_queue(&diff_queued_diff, one, two);
pair->is_unmerged = 1;
return pair;
}
static char *run_textconv(const char *pgm, struct diff_filespec *spec,
size_t *outsize)
{
diff: refactor tempfile cleanup handling There are two pieces of code that create tempfiles for diff: run_external_diff and run_textconv. The former cleans up its tempfiles in the face of premature death (i.e., by die() or by signal), but the latter does not. After this patch, they will both use the same cleanup routines. To make clear what the change is, let me first explain what happens now: - run_external_diff uses a static global array of 2 diff_tempfile structs (since it knows it will always need exactly 2 tempfiles). It calls prepare_temp_file (which doesn't know anything about the global array) on each of the structs, creating the tempfiles that need to be cleaned up. It then registers atexit and signal handlers to look through the global array and remove the tempfiles. If it succeeds, it calls the handler manually (which marks the tempfile structs as unused). - textconv has its own tempfile struct, which it allocates using prepare_temp_file and cleans up manually. No signal or atexit handlers. The new code moves the installation of cleanup handlers into the prepare_temp_file function. Which means that that function now has to understand that there is static tempfile storage. So what happens now is: - run_external_diff calls prepare_temp_file - prepare_temp_file calls claim_diff_tempfile, which allocates an unused slot from our global array - prepare_temp_file installs (if they have not already been installed) atexit and signal handlers for cleanup - prepare_temp_file sets up the tempfile as usual - prepare_temp_file returns a pointer to the allocated tempfile The advantage being that run_external_diff no longer has to care about setting up cleanup handlers. Now by virtue of calling prepare_temp_file, run_textconv gets the same benefit, as will any future users of prepare_temp_file. There are also a few side benefits to the specific implementation: - we now install cleanup handlers _before_ allocating the tempfile, closing a race which could leave temp cruft - when allocating a slot in the global array, we will now detect a situation where the old slots were not properly vacated (i.e., somebody forgot to call remove upon leaving the function). In the old code, such a situation would silently overwrite the tempfile names, meaning we would forget to clean them up. The new code dies with a bug warning. - we make sure only to install the signal handler once. This isn't a big deal, since we are just overwriting the old handler, but will become an issue when a later patch converts the code to use sigchain Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-01-22 05:59:56 +00:00
struct diff_tempfile *temp;
const char *argv[3];
const char **arg = argv;
struct child_process child;
struct strbuf buf = STRBUF_INIT;
int err = 0;
diff: refactor tempfile cleanup handling There are two pieces of code that create tempfiles for diff: run_external_diff and run_textconv. The former cleans up its tempfiles in the face of premature death (i.e., by die() or by signal), but the latter does not. After this patch, they will both use the same cleanup routines. To make clear what the change is, let me first explain what happens now: - run_external_diff uses a static global array of 2 diff_tempfile structs (since it knows it will always need exactly 2 tempfiles). It calls prepare_temp_file (which doesn't know anything about the global array) on each of the structs, creating the tempfiles that need to be cleaned up. It then registers atexit and signal handlers to look through the global array and remove the tempfiles. If it succeeds, it calls the handler manually (which marks the tempfile structs as unused). - textconv has its own tempfile struct, which it allocates using prepare_temp_file and cleans up manually. No signal or atexit handlers. The new code moves the installation of cleanup handlers into the prepare_temp_file function. Which means that that function now has to understand that there is static tempfile storage. So what happens now is: - run_external_diff calls prepare_temp_file - prepare_temp_file calls claim_diff_tempfile, which allocates an unused slot from our global array - prepare_temp_file installs (if they have not already been installed) atexit and signal handlers for cleanup - prepare_temp_file sets up the tempfile as usual - prepare_temp_file returns a pointer to the allocated tempfile The advantage being that run_external_diff no longer has to care about setting up cleanup handlers. Now by virtue of calling prepare_temp_file, run_textconv gets the same benefit, as will any future users of prepare_temp_file. There are also a few side benefits to the specific implementation: - we now install cleanup handlers _before_ allocating the tempfile, closing a race which could leave temp cruft - when allocating a slot in the global array, we will now detect a situation where the old slots were not properly vacated (i.e., somebody forgot to call remove upon leaving the function). In the old code, such a situation would silently overwrite the tempfile names, meaning we would forget to clean them up. The new code dies with a bug warning. - we make sure only to install the signal handler once. This isn't a big deal, since we are just overwriting the old handler, but will become an issue when a later patch converts the code to use sigchain Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-01-22 05:59:56 +00:00
temp = prepare_temp_file(spec->path, spec);
*arg++ = pgm;
diff: refactor tempfile cleanup handling There are two pieces of code that create tempfiles for diff: run_external_diff and run_textconv. The former cleans up its tempfiles in the face of premature death (i.e., by die() or by signal), but the latter does not. After this patch, they will both use the same cleanup routines. To make clear what the change is, let me first explain what happens now: - run_external_diff uses a static global array of 2 diff_tempfile structs (since it knows it will always need exactly 2 tempfiles). It calls prepare_temp_file (which doesn't know anything about the global array) on each of the structs, creating the tempfiles that need to be cleaned up. It then registers atexit and signal handlers to look through the global array and remove the tempfiles. If it succeeds, it calls the handler manually (which marks the tempfile structs as unused). - textconv has its own tempfile struct, which it allocates using prepare_temp_file and cleans up manually. No signal or atexit handlers. The new code moves the installation of cleanup handlers into the prepare_temp_file function. Which means that that function now has to understand that there is static tempfile storage. So what happens now is: - run_external_diff calls prepare_temp_file - prepare_temp_file calls claim_diff_tempfile, which allocates an unused slot from our global array - prepare_temp_file installs (if they have not already been installed) atexit and signal handlers for cleanup - prepare_temp_file sets up the tempfile as usual - prepare_temp_file returns a pointer to the allocated tempfile The advantage being that run_external_diff no longer has to care about setting up cleanup handlers. Now by virtue of calling prepare_temp_file, run_textconv gets the same benefit, as will any future users of prepare_temp_file. There are also a few side benefits to the specific implementation: - we now install cleanup handlers _before_ allocating the tempfile, closing a race which could leave temp cruft - when allocating a slot in the global array, we will now detect a situation where the old slots were not properly vacated (i.e., somebody forgot to call remove upon leaving the function). In the old code, such a situation would silently overwrite the tempfile names, meaning we would forget to clean them up. The new code dies with a bug warning. - we make sure only to install the signal handler once. This isn't a big deal, since we are just overwriting the old handler, but will become an issue when a later patch converts the code to use sigchain Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-01-22 05:59:56 +00:00
*arg++ = temp->name;
*arg = NULL;
memset(&child, 0, sizeof(child));
child.use_shell = 1;
child.argv = argv;
child.out = -1;
if (start_command(&child)) {
diff: refactor tempfile cleanup handling There are two pieces of code that create tempfiles for diff: run_external_diff and run_textconv. The former cleans up its tempfiles in the face of premature death (i.e., by die() or by signal), but the latter does not. After this patch, they will both use the same cleanup routines. To make clear what the change is, let me first explain what happens now: - run_external_diff uses a static global array of 2 diff_tempfile structs (since it knows it will always need exactly 2 tempfiles). It calls prepare_temp_file (which doesn't know anything about the global array) on each of the structs, creating the tempfiles that need to be cleaned up. It then registers atexit and signal handlers to look through the global array and remove the tempfiles. If it succeeds, it calls the handler manually (which marks the tempfile structs as unused). - textconv has its own tempfile struct, which it allocates using prepare_temp_file and cleans up manually. No signal or atexit handlers. The new code moves the installation of cleanup handlers into the prepare_temp_file function. Which means that that function now has to understand that there is static tempfile storage. So what happens now is: - run_external_diff calls prepare_temp_file - prepare_temp_file calls claim_diff_tempfile, which allocates an unused slot from our global array - prepare_temp_file installs (if they have not already been installed) atexit and signal handlers for cleanup - prepare_temp_file sets up the tempfile as usual - prepare_temp_file returns a pointer to the allocated tempfile The advantage being that run_external_diff no longer has to care about setting up cleanup handlers. Now by virtue of calling prepare_temp_file, run_textconv gets the same benefit, as will any future users of prepare_temp_file. There are also a few side benefits to the specific implementation: - we now install cleanup handlers _before_ allocating the tempfile, closing a race which could leave temp cruft - when allocating a slot in the global array, we will now detect a situation where the old slots were not properly vacated (i.e., somebody forgot to call remove upon leaving the function). In the old code, such a situation would silently overwrite the tempfile names, meaning we would forget to clean them up. The new code dies with a bug warning. - we make sure only to install the signal handler once. This isn't a big deal, since we are just overwriting the old handler, but will become an issue when a later patch converts the code to use sigchain Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-01-22 05:59:56 +00:00
remove_tempfile();
return NULL;
}
if (strbuf_read(&buf, child.out, 0) < 0)
err = error("error reading from textconv command '%s'", pgm);
close(child.out);
if (finish_command(&child) || err) {
strbuf_release(&buf);
remove_tempfile();
return NULL;
}
diff: refactor tempfile cleanup handling There are two pieces of code that create tempfiles for diff: run_external_diff and run_textconv. The former cleans up its tempfiles in the face of premature death (i.e., by die() or by signal), but the latter does not. After this patch, they will both use the same cleanup routines. To make clear what the change is, let me first explain what happens now: - run_external_diff uses a static global array of 2 diff_tempfile structs (since it knows it will always need exactly 2 tempfiles). It calls prepare_temp_file (which doesn't know anything about the global array) on each of the structs, creating the tempfiles that need to be cleaned up. It then registers atexit and signal handlers to look through the global array and remove the tempfiles. If it succeeds, it calls the handler manually (which marks the tempfile structs as unused). - textconv has its own tempfile struct, which it allocates using prepare_temp_file and cleans up manually. No signal or atexit handlers. The new code moves the installation of cleanup handlers into the prepare_temp_file function. Which means that that function now has to understand that there is static tempfile storage. So what happens now is: - run_external_diff calls prepare_temp_file - prepare_temp_file calls claim_diff_tempfile, which allocates an unused slot from our global array - prepare_temp_file installs (if they have not already been installed) atexit and signal handlers for cleanup - prepare_temp_file sets up the tempfile as usual - prepare_temp_file returns a pointer to the allocated tempfile The advantage being that run_external_diff no longer has to care about setting up cleanup handlers. Now by virtue of calling prepare_temp_file, run_textconv gets the same benefit, as will any future users of prepare_temp_file. There are also a few side benefits to the specific implementation: - we now install cleanup handlers _before_ allocating the tempfile, closing a race which could leave temp cruft - when allocating a slot in the global array, we will now detect a situation where the old slots were not properly vacated (i.e., somebody forgot to call remove upon leaving the function). In the old code, such a situation would silently overwrite the tempfile names, meaning we would forget to clean them up. The new code dies with a bug warning. - we make sure only to install the signal handler once. This isn't a big deal, since we are just overwriting the old handler, but will become an issue when a later patch converts the code to use sigchain Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2009-01-22 05:59:56 +00:00
remove_tempfile();
return strbuf_detach(&buf, outsize);
}
size_t fill_textconv(struct userdiff_driver *driver,
struct diff_filespec *df,
char **outbuf)
{
size_t size;
if (!driver || !driver->textconv) {
if (!DIFF_FILE_VALID(df)) {
*outbuf = "";
return 0;
}
if (diff_populate_filespec(df, 0))
die("unable to read files to diff");
*outbuf = df->data;
return df->size;
}
if (driver->textconv_cache && df->sha1_valid) {
*outbuf = notes_cache_get(driver->textconv_cache, df->sha1,
&size);
if (*outbuf)
return size;
}
*outbuf = run_textconv(driver->textconv, df, &size);
if (!*outbuf)
die("unable to read files to diff");
if (driver->textconv_cache && df->sha1_valid) {
/* ignore errors, as we might be in a readonly repository */
notes_cache_put(driver->textconv_cache, df->sha1, *outbuf,
size);
/*
* we could save up changes and flush them all at the end,
* but we would need an extra call after all diffing is done.
* Since generating a cache entry is the slow path anyway,
* this extra overhead probably isn't a big deal.
*/
notes_cache_write(driver->textconv_cache);
}
return size;
}