git/sha1_name.c

1509 lines
37 KiB
C
Raw Normal View History

#include "cache.h"
#include "tag.h"
#include "commit.h"
#include "tree.h"
#include "blob.h"
#include "tree-walk.h"
#include "refs.h"
#include "remote.h"
#include "dir.h"
static int get_sha1_oneline(const char *, unsigned char *, struct commit_list *);
typedef int (*disambiguate_hint_fn)(const unsigned char *, void *);
struct disambiguate_state {
disambiguate_hint_fn fn;
void *cb_data;
unsigned char candidate[20];
unsigned candidate_exists:1;
unsigned candidate_checked:1;
unsigned candidate_ok:1;
unsigned disambiguate_fn_used:1;
unsigned ambiguous:1;
unsigned always_call_fn:1;
};
static void update_candidates(struct disambiguate_state *ds, const unsigned char *current)
{
if (ds->always_call_fn) {
ds->ambiguous = ds->fn(current, ds->cb_data) ? 1 : 0;
return;
}
if (!ds->candidate_exists) {
/* this is the first candidate */
hashcpy(ds->candidate, current);
ds->candidate_exists = 1;
return;
} else if (!hashcmp(ds->candidate, current)) {
/* the same as what we already have seen */
return;
}
if (!ds->fn) {
/* cannot disambiguate between ds->candidate and current */
ds->ambiguous = 1;
return;
}
if (!ds->candidate_checked) {
ds->candidate_ok = ds->fn(ds->candidate, ds->cb_data);
ds->disambiguate_fn_used = 1;
ds->candidate_checked = 1;
}
if (!ds->candidate_ok) {
/* discard the candidate; we know it does not satisfy fn */
hashcpy(ds->candidate, current);
ds->candidate_checked = 0;
return;
}
/* if we reach this point, we know ds->candidate satisfies fn */
if (ds->fn(current, ds->cb_data)) {
/*
* if both current and candidate satisfy fn, we cannot
* disambiguate.
*/
ds->candidate_ok = 0;
ds->ambiguous = 1;
}
/* otherwise, current can be discarded and candidate is still good */
}
static void find_short_object_filename(int len, const char *hex_pfx, struct disambiguate_state *ds)
{
struct alternate_object_database *alt;
char hex[40];
static struct alternate_object_database *fakeent;
if (!fakeent) {
/*
* Create a "fake" alternate object database that
* points to our own object database, to make it
* easier to get a temporary working space in
* alt->name/alt->base while iterating over the
* object databases including our own.
*/
const char *objdir = get_object_directory();
size_t objdir_len = strlen(objdir);
fakeent = xmalloc(st_add3(sizeof(*fakeent), objdir_len, 43));
memcpy(fakeent->base, objdir, objdir_len);
fakeent->name = fakeent->base + objdir_len + 1;
fakeent->name[-1] = '/';
}
fakeent->next = alt_odb_list;
xsnprintf(hex, sizeof(hex), "%.2s", hex_pfx);
for (alt = fakeent; alt && !ds->ambiguous; alt = alt->next) {
struct dirent *de;
DIR *dir;
/*
* every alt_odb struct has 42 extra bytes after the base
* for exactly this purpose
*/
xsnprintf(alt->name, 42, "%.2s/", hex_pfx);
dir = opendir(alt->base);
if (!dir)
continue;
while (!ds->ambiguous && (de = readdir(dir)) != NULL) {
unsigned char sha1[20];
if (strlen(de->d_name) != 38)
continue;
if (memcmp(de->d_name, hex_pfx + 2, len - 2))
continue;
memcpy(hex + 2, de->d_name, 38);
if (!get_sha1_hex(hex, sha1))
update_candidates(ds, sha1);
}
closedir(dir);
}
}
static int match_sha(unsigned len, const unsigned char *a, const unsigned char *b)
{
do {
if (*a != *b)
return 0;
a++;
b++;
len -= 2;
} while (len > 1);
if (len)
if ((*a ^ *b) & 0xf0)
return 0;
return 1;
}
static void unique_in_pack(int len,
const unsigned char *bin_pfx,
struct packed_git *p,
struct disambiguate_state *ds)
{
uint32_t num, last, i, first = 0;
const unsigned char *current = NULL;
open_pack_index(p);
num = p->num_objects;
last = num;
while (first < last) {
uint32_t mid = (first + last) / 2;
const unsigned char *current;
int cmp;
current = nth_packed_object_sha1(p, mid);
cmp = hashcmp(bin_pfx, current);
if (!cmp) {
first = mid;
break;
}
if (cmp > 0) {
first = mid+1;
continue;
}
last = mid;
}
/*
* At this point, "first" is the location of the lowest object
* with an object name that could match "bin_pfx". See if we have
* 0, 1 or more objects that actually match(es).
*/
for (i = first; i < num && !ds->ambiguous; i++) {
current = nth_packed_object_sha1(p, i);
if (!match_sha(len, bin_pfx, current))
break;
update_candidates(ds, current);
}
}
static void find_short_packed_object(int len, const unsigned char *bin_pfx,
struct disambiguate_state *ds)
{
struct packed_git *p;
prepare_packed_git();
for (p = packed_git; p && !ds->ambiguous; p = p->next)
unique_in_pack(len, bin_pfx, p, ds);
}
#define SHORT_NAME_NOT_FOUND (-1)
#define SHORT_NAME_AMBIGUOUS (-2)
static int finish_object_disambiguation(struct disambiguate_state *ds,
unsigned char *sha1)
{
if (ds->ambiguous)
return SHORT_NAME_AMBIGUOUS;
if (!ds->candidate_exists)
return SHORT_NAME_NOT_FOUND;
if (!ds->candidate_checked)
/*
* If this is the only candidate, there is no point
* calling the disambiguation hint callback.
*
* On the other hand, if the current candidate
* replaced an earlier candidate that did _not_ pass
* the disambiguation hint callback, then we do have
* more than one objects that match the short name
* given, so we should make sure this one matches;
* otherwise, if we discovered this one and the one
* that we previously discarded in the reverse order,
* we would end up showing different results in the
* same repository!
*/
ds->candidate_ok = (!ds->disambiguate_fn_used ||
ds->fn(ds->candidate, ds->cb_data));
if (!ds->candidate_ok)
return SHORT_NAME_AMBIGUOUS;
hashcpy(sha1, ds->candidate);
return 0;
}
static int disambiguate_commit_only(const unsigned char *sha1, void *cb_data_unused)
{
int kind = sha1_object_info(sha1, NULL);
return kind == OBJ_COMMIT;
}
static int disambiguate_committish_only(const unsigned char *sha1, void *cb_data_unused)
{
struct object *obj;
int kind;
kind = sha1_object_info(sha1, NULL);
if (kind == OBJ_COMMIT)
return 1;
if (kind != OBJ_TAG)
return 0;
/* We need to do this the hard way... */
get_short_sha1(): correctly disambiguate type-limited abbreviation One test in t1512 that expects a failure incorrectly passed. The test prepares a commit whose object name begins with ten "0"s, and also prepares a tag that points at the commit. The object name of the tag also begins with ten "0"s. There is no other commit-ish object in the repository whose name begins with such a prefix. Ideally, in such a repository: $ git rev-parse --verify 0000000000^{commit} should yield that commit. If 0000000000 is taken as the commit 0000000000e4f, peeling it to a commmit yields that commit itself, and if 0000000000 is taken as the tag 0000000000f8f, peeling it to a commit also yields the same commit, so in that twisted sense, the extended SHA-1 expression 0000000000^{commit} is unambigous. The test that expects a failure is to check the above command. The reason the test expects a failure is that we did not implement such a "unification" of two candidate objects. What we did (or at least, meant to) implement was to recognise that a commit-ish is required to expand 0000000000, and notice that there are two succh commit-ish, and diagnose the request as ambiguous. However, there was a bug in the logic to check the candidate objects. When the code saw 0000000000f8f (a tag) that shared the shortened prefix (ten "0"s), it tried to make sure that the tag is a commit-ish by looking at the tag object. Because it incorrectly used lookup_object() when the tag has not been parsed, however, we incorrectly declared that the tag is _not_ a commit-ish, leaving the sole commit in the repository, 0000000000e4f, that has the required prefix as "unique match", causing the test to pass when it shouldn't. This fixes the logic to inspect the type of the object a tag refers to, to make the test that is expected to fail correctly fail. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-07-02 04:54:45 +00:00
obj = deref_tag(parse_object(sha1), NULL, 0);
if (obj && obj->type == OBJ_COMMIT)
return 1;
return 0;
}
static int disambiguate_tree_only(const unsigned char *sha1, void *cb_data_unused)
{
int kind = sha1_object_info(sha1, NULL);
return kind == OBJ_TREE;
}
static int disambiguate_treeish_only(const unsigned char *sha1, void *cb_data_unused)
{
struct object *obj;
int kind;
kind = sha1_object_info(sha1, NULL);
if (kind == OBJ_TREE || kind == OBJ_COMMIT)
return 1;
if (kind != OBJ_TAG)
return 0;
/* We need to do this the hard way... */
obj = deref_tag(lookup_object(sha1), NULL, 0);
if (obj && (obj->type == OBJ_TREE || obj->type == OBJ_COMMIT))
return 1;
return 0;
}
static int disambiguate_blob_only(const unsigned char *sha1, void *cb_data_unused)
{
int kind = sha1_object_info(sha1, NULL);
return kind == OBJ_BLOB;
}
static int prepare_prefixes(const char *name, int len,
unsigned char *bin_pfx,
char *hex_pfx)
{
int i;
hashclr(bin_pfx);
memset(hex_pfx, 'x', 40);
for (i = 0; i < len ;i++) {
unsigned char c = name[i];
unsigned char val;
if (c >= '0' && c <= '9')
val = c - '0';
else if (c >= 'a' && c <= 'f')
val = c - 'a' + 10;
else if (c >= 'A' && c <='F') {
val = c - 'A' + 10;
c -= 'A' - 'a';
}
else
return -1;
hex_pfx[i] = c;
if (!(i & 1))
val <<= 4;
bin_pfx[i >> 1] |= val;
}
return 0;
}
static int get_short_sha1(const char *name, int len, unsigned char *sha1,
unsigned flags)
{
int status;
char hex_pfx[40];
unsigned char bin_pfx[20];
struct disambiguate_state ds;
int quietly = !!(flags & GET_SHA1_QUIETLY);
if (len < MINIMUM_ABBREV || len > 40)
return -1;
if (prepare_prefixes(name, len, bin_pfx, hex_pfx) < 0)
return -1;
prepare_alt_odb();
memset(&ds, 0, sizeof(ds));
if (flags & GET_SHA1_COMMIT)
ds.fn = disambiguate_commit_only;
else if (flags & GET_SHA1_COMMITTISH)
ds.fn = disambiguate_committish_only;
else if (flags & GET_SHA1_TREE)
ds.fn = disambiguate_tree_only;
else if (flags & GET_SHA1_TREEISH)
ds.fn = disambiguate_treeish_only;
else if (flags & GET_SHA1_BLOB)
ds.fn = disambiguate_blob_only;
find_short_object_filename(len, hex_pfx, &ds);
find_short_packed_object(len, bin_pfx, &ds);
status = finish_object_disambiguation(&ds, sha1);
if (!quietly && (status == SHORT_NAME_AMBIGUOUS))
return error("short SHA1 %.*s is ambiguous.", len, hex_pfx);
return status;
}
int for_each_abbrev(const char *prefix, each_abbrev_fn fn, void *cb_data)
{
char hex_pfx[40];
unsigned char bin_pfx[20];
struct disambiguate_state ds;
int len = strlen(prefix);
if (len < MINIMUM_ABBREV || len > 40)
return -1;
if (prepare_prefixes(prefix, len, bin_pfx, hex_pfx) < 0)
return -1;
prepare_alt_odb();
memset(&ds, 0, sizeof(ds));
ds.always_call_fn = 1;
ds.cb_data = cb_data;
ds.fn = fn;
find_short_object_filename(len, hex_pfx, &ds);
find_short_packed_object(len, bin_pfx, &ds);
return ds.ambiguous;
}
int find_unique_abbrev_r(char *hex, const unsigned char *sha1, int len)
{
int status, exists;
sha1_to_hex_r(hex, sha1);
if (len == 40 || !len)
return 40;
exists = has_sha1_file(sha1);
while (len < 40) {
unsigned char sha1_ret[20];
status = get_short_sha1(hex, len, sha1_ret, GET_SHA1_QUIETLY);
if (exists
? !status
: status == SHORT_NAME_NOT_FOUND) {
hex[len] = 0;
return len;
}
len++;
}
return len;
}
const char *find_unique_abbrev(const unsigned char *sha1, int len)
{
static char hex[GIT_SHA1_HEXSZ + 1];
find_unique_abbrev_r(hex, sha1, len);
return hex;
}
static int ambiguous_path(const char *path, int len)
2005-10-28 19:41:49 +00:00
{
int slash = 1;
int cnt;
2005-10-28 19:41:49 +00:00
for (cnt = 0; cnt < len; cnt++) {
2005-10-28 19:41:49 +00:00
switch (*path++) {
case '\0':
break;
case '/':
if (slash)
break;
slash = 1;
continue;
case '.':
continue;
default:
slash = 0;
continue;
}
break;
2005-10-28 19:41:49 +00:00
}
return slash;
2005-10-28 19:41:49 +00:00
}
static inline int at_mark(const char *string, int len,
const char **suffix, int nr)
Teach @{upstream} syntax to strbuf_branchanme() This teaches @{upstream} syntax to interpret_branch_name(), instead of dwim_ref() machinery. There are places in git UI that behaves differently when you give a local branch name and when you give an extended SHA-1 expression that evaluates to the commit object name at the tip of the branch. The intent is that the special syntax such as @{-1} can stand in as if the user spelled the name of the branch in such places. The name of the branch "frotz" to switch to ("git checkout frotz"), and the name of the branch "nitfol" to fork a new branch "frotz" from ("git checkout -b frotz nitfol"), are examples of such places. These places take only the name of the branch (e.g. "frotz"), and they are supposed to act differently to an equivalent refname (e.g. "refs/heads/frotz"), so hooking the @{upstream} and @{-N} syntax to dwim_ref() is insufficient when we want to deal with cases a local branch is forked from another local branch and use "forked@{upstream}" to name the forkee branch. The "upstream" syntax "forked@{u}" is to specify the ref that "forked" is configured to merge with, and most often the forkee is a remote tracking branch, not a local branch. We cannot simply return a local branch name, but that does not necessarily mean we have to returns the full refname (e.g. refs/remotes/origin/frotz, when returning origin/frotz is enough). This update calls shorten_unambiguous_ref() to do so. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-20 07:17:11 +00:00
{
int i;
for (i = 0; i < nr; i++) {
Teach @{upstream} syntax to strbuf_branchanme() This teaches @{upstream} syntax to interpret_branch_name(), instead of dwim_ref() machinery. There are places in git UI that behaves differently when you give a local branch name and when you give an extended SHA-1 expression that evaluates to the commit object name at the tip of the branch. The intent is that the special syntax such as @{-1} can stand in as if the user spelled the name of the branch in such places. The name of the branch "frotz" to switch to ("git checkout frotz"), and the name of the branch "nitfol" to fork a new branch "frotz" from ("git checkout -b frotz nitfol"), are examples of such places. These places take only the name of the branch (e.g. "frotz"), and they are supposed to act differently to an equivalent refname (e.g. "refs/heads/frotz"), so hooking the @{upstream} and @{-N} syntax to dwim_ref() is insufficient when we want to deal with cases a local branch is forked from another local branch and use "forked@{upstream}" to name the forkee branch. The "upstream" syntax "forked@{u}" is to specify the ref that "forked" is configured to merge with, and most often the forkee is a remote tracking branch, not a local branch. We cannot simply return a local branch name, but that does not necessarily mean we have to returns the full refname (e.g. refs/remotes/origin/frotz, when returning origin/frotz is enough). This update calls shorten_unambiguous_ref() to do so. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-20 07:17:11 +00:00
int suffix_len = strlen(suffix[i]);
if (suffix_len <= len
&& !memcmp(string, suffix[i], suffix_len))
return suffix_len;
}
return 0;
}
static inline int upstream_mark(const char *string, int len)
{
const char *suffix[] = { "@{upstream}", "@{u}" };
return at_mark(string, len, suffix, ARRAY_SIZE(suffix));
}
static inline int push_mark(const char *string, int len)
{
const char *suffix[] = { "@{push}" };
return at_mark(string, len, suffix, ARRAY_SIZE(suffix));
}
static int get_sha1_1(const char *name, int len, unsigned char *sha1, unsigned lookup_flags);
interpret_branch_name: always respect "namelen" parameter interpret_branch_name gets passed a "name" buffer to parse, along with a "namelen" parameter representing its length. If "namelen" is zero, we fallback to the NUL-terminated string-length of "name". However, it does not necessarily follow that if we have gotten a non-zero "namelen", it is the NUL-terminated string-length of "name". E.g., when get_sha1() is parsing "foo:bar", we will be asked to operate only on the first three characters. Yet in interpret_branch_name and its helpers, we use string functions like strchr() to operate on "name", looking past the length we were given. This can result in us mis-parsing object names. We should instead be limiting our search to "namelen" bytes. There are three distinct types of object names this patch addresses: - The intrepret_empty_at helper uses strchr to find the next @-expression after our potential empty-at. In an expression like "@:foo@bar", it erroneously thinks that the second "@" is relevant, even if we were asked only to look at the first character. This case is easy to trigger (and we test it in this patch). - When finding the initial @-mark for @{upstream}, we use strchr. This means we might treat "foo:@{upstream}" as the upstream for "foo:", even though we were asked only to look at "foo". We cannot test this one in practice, because it is masked by another bug (which is fixed in the next patch). - The interpret_nth_prior_checkout helper did not receive the name length at all. This turns out not to be a problem in practice, though, because its parsing is so limited: it always starts from the far-left of the string, and will not tolerate a colon (which is currently the only way to get a smaller-than-strlen "namelen"). However, it's still worth fixing to make the code more obviously correct, and to future-proof us against callers with more exotic buffers. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-01-15 08:31:57 +00:00
static int interpret_nth_prior_checkout(const char *name, int namelen, struct strbuf *buf);
static int get_sha1_basic(const char *str, int len, unsigned char *sha1,
unsigned int flags)
{
static const char *warn_msg = "refname '%.*s' is ambiguous.";
static const char *object_name_msg = N_(
"Git normally never creates a ref that ends with 40 hex characters\n"
"because it will be ignored when you just specify 40-hex. These refs\n"
"may be created by mistake. For example,\n"
"\n"
" git checkout -b $br $(git rev-parse ...)\n"
"\n"
"where \"$br\" is somehow empty and a 40-hex ref is created. Please\n"
"examine these refs and maybe delete them. Turn this message off by\n"
"running \"git config advice.objectNameWarning false\"");
unsigned char tmp_sha1[20];
char *real_ref = NULL;
int refs_found = 0;
int at, reflog_len, nth_prior = 0;
if (len == 40 && !get_sha1_hex(str, sha1)) {
if (warn_ambiguous_refs && warn_on_object_refname_ambiguity) {
cat-file: disable object/refname ambiguity check for batch mode A common use of "cat-file --batch-check" is to feed a list of objects from "rev-list --objects" or a similar command. In this instance, all of our input objects are 40-byte sha1 ids. However, cat-file has always allowed arbitrary revision specifiers, and feeds the result to get_sha1(). Fortunately, get_sha1() recognizes a 40-byte sha1 before doing any hard work trying to look up refs, meaning this scenario should end up spending very little time converting the input into an object sha1. However, since 798c35f (get_sha1: warn about full or short object names that look like refs, 2013-05-29), when we encounter this case, we spend the extra effort to do a refname lookup anyway, just to print a warning. This is further exacerbated by ca91993 (get_packed_ref_cache: reload packed-refs file when it changes, 2013-06-20), which makes individual ref lookup more expensive by requiring a stat() of the packed-refs file for each missing ref. With no patches, this is the time it takes to run: $ git rev-list --objects --all >objects $ time git cat-file --batch-check='%(objectname)' <objects on the linux.git repository: real 1m13.494s user 0m25.924s sys 0m47.532s If we revert ca91993, the packed-refs up-to-date check, it gets a little better: real 0m54.697s user 0m21.692s sys 0m32.916s but we are still spending quite a bit of time on ref lookup (and we would not want to revert that patch, anyway, which has correctness issues). If we revert 798c35f, disabling the warning entirely, we get a much more reasonable time: real 0m7.452s user 0m6.836s sys 0m0.608s This patch does the moral equivalent of this final case (and gets similar speedups). We introduce a global flag that callers of get_sha1() can use to avoid paying the price for the warning. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-07-12 06:20:05 +00:00
refs_found = dwim_ref(str, len, tmp_sha1, &real_ref);
if (refs_found > 0) {
cat-file: disable object/refname ambiguity check for batch mode A common use of "cat-file --batch-check" is to feed a list of objects from "rev-list --objects" or a similar command. In this instance, all of our input objects are 40-byte sha1 ids. However, cat-file has always allowed arbitrary revision specifiers, and feeds the result to get_sha1(). Fortunately, get_sha1() recognizes a 40-byte sha1 before doing any hard work trying to look up refs, meaning this scenario should end up spending very little time converting the input into an object sha1. However, since 798c35f (get_sha1: warn about full or short object names that look like refs, 2013-05-29), when we encounter this case, we spend the extra effort to do a refname lookup anyway, just to print a warning. This is further exacerbated by ca91993 (get_packed_ref_cache: reload packed-refs file when it changes, 2013-06-20), which makes individual ref lookup more expensive by requiring a stat() of the packed-refs file for each missing ref. With no patches, this is the time it takes to run: $ git rev-list --objects --all >objects $ time git cat-file --batch-check='%(objectname)' <objects on the linux.git repository: real 1m13.494s user 0m25.924s sys 0m47.532s If we revert ca91993, the packed-refs up-to-date check, it gets a little better: real 0m54.697s user 0m21.692s sys 0m32.916s but we are still spending quite a bit of time on ref lookup (and we would not want to revert that patch, anyway, which has correctness issues). If we revert 798c35f, disabling the warning entirely, we get a much more reasonable time: real 0m7.452s user 0m6.836s sys 0m0.608s This patch does the moral equivalent of this final case (and gets similar speedups). We introduce a global flag that callers of get_sha1() can use to avoid paying the price for the warning. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-07-12 06:20:05 +00:00
warning(warn_msg, len, str);
if (advice_object_name_warning)
fprintf(stderr, "%s\n", _(object_name_msg));
}
free(real_ref);
}
return 0;
}
/* basic@{time or number or -number} format to query ref-log */
reflog_len = at = 0;
if (len && str[len-1] == '}') {
for (at = len-4; at >= 0; at--) {
if (str[at] == '@' && str[at+1] == '{') {
if (str[at+2] == '-') {
if (at != 0)
/* @{-N} not at start */
return -1;
nth_prior = 1;
continue;
}
if (!upstream_mark(str + at, len - at) &&
!push_mark(str + at, len - at)) {
reflog_len = (len-1) - (at+2);
len = at;
}
break;
}
}
}
2005-10-28 19:41:49 +00:00
/* Accept only unambiguous ref paths. */
if (len && ambiguous_path(str, len))
2005-10-28 19:41:49 +00:00
return -1;
if (nth_prior) {
struct strbuf buf = STRBUF_INIT;
int detached;
interpret_branch_name: always respect "namelen" parameter interpret_branch_name gets passed a "name" buffer to parse, along with a "namelen" parameter representing its length. If "namelen" is zero, we fallback to the NUL-terminated string-length of "name". However, it does not necessarily follow that if we have gotten a non-zero "namelen", it is the NUL-terminated string-length of "name". E.g., when get_sha1() is parsing "foo:bar", we will be asked to operate only on the first three characters. Yet in interpret_branch_name and its helpers, we use string functions like strchr() to operate on "name", looking past the length we were given. This can result in us mis-parsing object names. We should instead be limiting our search to "namelen" bytes. There are three distinct types of object names this patch addresses: - The intrepret_empty_at helper uses strchr to find the next @-expression after our potential empty-at. In an expression like "@:foo@bar", it erroneously thinks that the second "@" is relevant, even if we were asked only to look at the first character. This case is easy to trigger (and we test it in this patch). - When finding the initial @-mark for @{upstream}, we use strchr. This means we might treat "foo:@{upstream}" as the upstream for "foo:", even though we were asked only to look at "foo". We cannot test this one in practice, because it is masked by another bug (which is fixed in the next patch). - The interpret_nth_prior_checkout helper did not receive the name length at all. This turns out not to be a problem in practice, though, because its parsing is so limited: it always starts from the far-left of the string, and will not tolerate a colon (which is currently the only way to get a smaller-than-strlen "namelen"). However, it's still worth fixing to make the code more obviously correct, and to future-proof us against callers with more exotic buffers. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-01-15 08:31:57 +00:00
if (interpret_nth_prior_checkout(str, len, &buf) > 0) {
detached = (buf.len == 40 && !get_sha1_hex(buf.buf, sha1));
strbuf_release(&buf);
if (detached)
return 0;
}
}
if (!len && reflog_len)
/* allow "@{...}" to mean the current branch reflog */
refs_found = dwim_ref("HEAD", 4, sha1, &real_ref);
else if (reflog_len)
refs_found = dwim_log(str, len, sha1, &real_ref);
else
refs_found = dwim_ref(str, len, sha1, &real_ref);
if (!refs_found)
return -1;
if (warn_ambiguous_refs && !(flags & GET_SHA1_QUIETLY) &&
(refs_found > 1 ||
!get_short_sha1(str, len, tmp_sha1, GET_SHA1_QUIETLY)))
warning(warn_msg, len, str);
if (reflog_len) {
int nth, i;
unsigned long at_time;
unsigned long co_time;
int co_tz, co_cnt;
/* Is it asking for N-th entry, or approxidate? */
for (i = nth = 0; 0 <= nth && i < reflog_len; i++) {
char ch = str[at+2+i];
if ('0' <= ch && ch <= '9')
nth = nth * 10 + ch - '0';
else
nth = -1;
}
if (100000000 <= nth) {
at_time = nth;
nth = -1;
} else if (0 <= nth)
at_time = 0;
else {
int errors = 0;
char *tmp = xstrndup(str + at + 2, reflog_len);
at_time = approxidate_careful(tmp, &errors);
free(tmp);
if (errors) {
free(real_ref);
return -1;
}
}
if (read_ref_at(real_ref, flags, at_time, nth, sha1, NULL,
&co_time, &co_tz, &co_cnt)) {
if (!len) {
if (starts_with(real_ref, "refs/heads/")) {
str = real_ref + 11;
len = strlen(real_ref + 11);
} else {
/* detached HEAD */
str = "HEAD";
len = 4;
}
}
if (at_time) {
if (!(flags & GET_SHA1_QUIETLY)) {
warning("Log for '%.*s' only goes "
"back to %s.", len, str,
convert "enum date_mode" into a struct In preparation for adding date modes that may carry extra information beyond the mode itself, this patch converts the date_mode enum into a struct. Most of the conversion is fairly straightforward; we pass the struct as a pointer and dereference the type field where necessary. Locations that declare a date_mode can use a "{}" constructor. However, the tricky case is where we use the enum labels as constants, like: show_date(t, tz, DATE_NORMAL); Ideally we could say: show_date(t, tz, &{ DATE_NORMAL }); but of course C does not allow that. Likewise, we cannot cast the constant to a struct, because we need to pass an actual address. Our options are basically: 1. Manually add a "struct date_mode d = { DATE_NORMAL }" definition to each caller, and pass "&d". This makes the callers uglier, because they sometimes do not even have their own scope (e.g., they are inside a switch statement). 2. Provide a pre-made global "date_normal" struct that can be passed by address. We'd also need "date_rfc2822", "date_iso8601", and so forth. But at least the ugliness is defined in one place. 3. Provide a wrapper that generates the correct struct on the fly. The big downside is that we end up pointing to a single global, which makes our wrapper non-reentrant. But show_date is already not reentrant, so it does not matter. This patch implements 3, along with a minor macro to keep the size of the callers sane. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2015-06-25 16:55:02 +00:00
show_date(co_time, co_tz, DATE_MODE(RFC2822)));
}
} else {
if (flags & GET_SHA1_QUIETLY) {
exit(128);
}
die("Log for '%.*s' only has %d entries.",
len, str, co_cnt);
}
}
}
free(real_ref);
return 0;
}
static int get_parent(const char *name, int len,
unsigned char *result, int idx)
{
unsigned char sha1[20];
int ret = get_sha1_1(name, len, sha1, GET_SHA1_COMMITTISH);
struct commit *commit;
struct commit_list *p;
if (ret)
return ret;
commit = lookup_commit_reference(sha1);
if (parse_commit(commit))
return -1;
if (!idx) {
hashcpy(result, commit->object.oid.hash);
return 0;
}
p = commit->parents;
while (p) {
if (!--idx) {
hashcpy(result, p->item->object.oid.hash);
return 0;
}
p = p->next;
}
return -1;
}
static int get_nth_ancestor(const char *name, int len,
unsigned char *result, int generation)
{
unsigned char sha1[20];
struct commit *commit;
int ret;
ret = get_sha1_1(name, len, sha1, GET_SHA1_COMMITTISH);
if (ret)
return ret;
commit = lookup_commit_reference(sha1);
if (!commit)
return -1;
while (generation--) {
if (parse_commit(commit) || !commit->parents)
return -1;
commit = commit->parents->item;
}
hashcpy(result, commit->object.oid.hash);
return 0;
}
struct object *peel_to_type(const char *name, int namelen,
struct object *o, enum object_type expected_type)
{
if (name && !namelen)
namelen = strlen(name);
while (1) {
if (!o || (!o->parsed && !parse_object(o->oid.hash)))
return NULL;
peel_onion(): teach $foo^{object} peeler A string that names an object can be suffixed with ^{type} peeler to say "I have this object name; peel it until you get this type. If you cannot do so, it is an error". v1.8.2^{commit} asks for a commit that is pointed at an annotated tag v1.8.2; v1.8.2^{tree} unwraps it further to the top-level tree object. A special suffix ^{} (i.e. no type specified) means "I do not care what it unwraps to; just peel annotated tag until you get something that is not a tag". When you have a random user-supplied string, you can turn it to a bare 40-hex object name, and cause it to error out if such an object does not exist, with: git rev-parse --verify "$userstring^{}" for most objects, but this does not yield the tag object name when $userstring refers to an annotated tag. Introduce a new suffix, ^{object}, that only makes sure the given name refers to an existing object. Then git rev-parse --verify "$userstring^{object}" becomes a way to make sure $userstring refers to an existing object. This is necessary because the plumbing "rev-parse --verify" is only about "make sure the argument is something we can feed to get_sha1() and turn it into a raw 20-byte object name SHA-1" and is not about "make sure that 20-byte object name SHA-1 refers to an object that exists in our object store". When the given $userstring is already a 40-hex, by definition "rev-parse --verify $userstring" can turn it into a raw 20-byte object name. With "$userstring^{object}", we can make sure that the 40-hex string names an object that exists in our object store before "--verify" kicks in. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-03-31 22:24:12 +00:00
if (expected_type == OBJ_ANY || o->type == expected_type)
return o;
if (o->type == OBJ_TAG)
o = ((struct tag*) o)->tagged;
else if (o->type == OBJ_COMMIT)
o = &(((struct commit *) o)->tree->object);
else {
if (name)
error("%.*s: expected %s type, but the object "
"dereferences to %s type",
namelen, name, typename(expected_type),
typename(o->type));
return NULL;
}
}
}
static int peel_onion(const char *name, int len, unsigned char *sha1)
{
unsigned char outer[20];
const char *sp;
unsigned int expected_type = 0;
unsigned lookup_flags = 0;
struct object *o;
/*
* "ref^{type}" dereferences ref repeatedly until you cannot
* dereference anymore, or you get an object of given type,
* whichever comes first. "ref^{}" means just dereference
* tags until you get a non-tag. "ref^0" is a shorthand for
* "ref^{commit}". "commit^{tree}" could be used to find the
* top-level tree of the given commit.
*/
if (len < 4 || name[len-1] != '}')
return -1;
for (sp = name + len - 1; name <= sp; sp--) {
int ch = *sp;
if (ch == '{' && name < sp && sp[-1] == '^')
break;
}
if (sp <= name)
return -1;
sp++; /* beginning of type name, or closing brace for empty */
if (starts_with(sp, "commit}"))
expected_type = OBJ_COMMIT;
else if (starts_with(sp, "tag}"))
expected_type = OBJ_TAG;
else if (starts_with(sp, "tree}"))
expected_type = OBJ_TREE;
else if (starts_with(sp, "blob}"))
expected_type = OBJ_BLOB;
else if (starts_with(sp, "object}"))
peel_onion(): teach $foo^{object} peeler A string that names an object can be suffixed with ^{type} peeler to say "I have this object name; peel it until you get this type. If you cannot do so, it is an error". v1.8.2^{commit} asks for a commit that is pointed at an annotated tag v1.8.2; v1.8.2^{tree} unwraps it further to the top-level tree object. A special suffix ^{} (i.e. no type specified) means "I do not care what it unwraps to; just peel annotated tag until you get something that is not a tag". When you have a random user-supplied string, you can turn it to a bare 40-hex object name, and cause it to error out if such an object does not exist, with: git rev-parse --verify "$userstring^{}" for most objects, but this does not yield the tag object name when $userstring refers to an annotated tag. Introduce a new suffix, ^{object}, that only makes sure the given name refers to an existing object. Then git rev-parse --verify "$userstring^{object}" becomes a way to make sure $userstring refers to an existing object. This is necessary because the plumbing "rev-parse --verify" is only about "make sure the argument is something we can feed to get_sha1() and turn it into a raw 20-byte object name SHA-1" and is not about "make sure that 20-byte object name SHA-1 refers to an object that exists in our object store". When the given $userstring is already a 40-hex, by definition "rev-parse --verify $userstring" can turn it into a raw 20-byte object name. With "$userstring^{object}", we can make sure that the 40-hex string names an object that exists in our object store before "--verify" kicks in. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-03-31 22:24:12 +00:00
expected_type = OBJ_ANY;
else if (sp[0] == '}')
expected_type = OBJ_NONE;
else if (sp[0] == '/')
expected_type = OBJ_COMMIT;
else
return -1;
if (expected_type == OBJ_COMMIT)
lookup_flags = GET_SHA1_COMMITTISH;
else if (expected_type == OBJ_TREE)
lookup_flags = GET_SHA1_TREEISH;
if (get_sha1_1(name, sp - name - 2, outer, lookup_flags))
return -1;
o = parse_object(outer);
if (!o)
return -1;
if (!expected_type) {
o = deref_tag(o, name, sp - name - 2);
if (!o || (!o->parsed && !parse_object(o->oid.hash)))
return -1;
hashcpy(sha1, o->oid.hash);
return 0;
}
/*
* At this point, the syntax look correct, so
* if we do not get the needed object, we should
* barf.
*/
o = peel_to_type(name, len, o, expected_type);
if (!o)
return -1;
hashcpy(sha1, o->oid.hash);
if (sp[0] == '/') {
/* "$commit^{/foo}" */
char *prefix;
int ret;
struct commit_list *list = NULL;
/*
* $commit^{/}. Some regex implementation may reject.
* We don't need regex anyway. '' pattern always matches.
*/
if (sp[1] == '}')
return 0;
prefix = xstrndup(sp + 1, name + len - 1 - (sp + 1));
commit_list_insert((struct commit *)o, &list);
ret = get_sha1_oneline(prefix, sha1, list);
free(prefix);
return ret;
}
return 0;
}
static int get_describe_name(const char *name, int len, unsigned char *sha1)
{
const char *cp;
unsigned flags = GET_SHA1_QUIETLY | GET_SHA1_COMMIT;
for (cp = name + len - 1; name + 2 <= cp; cp--) {
char ch = *cp;
if (!isxdigit(ch)) {
/* We must be looking at g in "SOMETHING-g"
* for it to be describe output.
*/
if (ch == 'g' && cp[-1] == '-') {
cp++;
len -= cp - name;
return get_short_sha1(cp, len, sha1, flags);
}
}
}
return -1;
}
static int get_sha1_1(const char *name, int len, unsigned char *sha1, unsigned lookup_flags)
{
int ret, has_suffix;
const char *cp;
/*
* "name~3" is "name^^^", "name~" is "name~1", and "name^" is "name^1".
*/
has_suffix = 0;
for (cp = name + len - 1; name <= cp; cp--) {
int ch = *cp;
if ('0' <= ch && ch <= '9')
continue;
if (ch == '~' || ch == '^')
has_suffix = ch;
break;
}
if (has_suffix) {
int num = 0;
int len1 = cp - name;
cp++;
while (cp < name + len)
num = num * 10 + *cp++ - '0';
if (!num && len1 == len - 1)
num = 1;
if (has_suffix == '^')
return get_parent(name, len1, sha1, num);
/* else if (has_suffix == '~') -- goes without saying */
return get_nth_ancestor(name, len1, sha1, num);
}
ret = peel_onion(name, len, sha1);
if (!ret)
return 0;
ret = get_sha1_basic(name, len, sha1, lookup_flags);
if (!ret)
return 0;
/* It could be describe output that is "SOMETHING-gXXXX" */
ret = get_describe_name(name, len, sha1);
if (!ret)
return 0;
return get_short_sha1(name, len, sha1, lookup_flags);
}
/*
* This interprets names like ':/Initial revision of "git"' by searching
* through history and returning the first commit whose message starts
* the given regular expression.
*
* For negative-matching, prefix the pattern-part with '!-', like: ':/!-WIP'.
*
* For a literal '!' character at the beginning of a pattern, you have to repeat
* that, like: ':/!!foo'
*
* For future extension, all other sequences beginning with ':/!' are reserved.
*/
/* Remember to update object flag allocation in object.h */
#define ONELINE_SEEN (1u<<20)
static int handle_one_ref(const char *path, const struct object_id *oid,
int flag, void *cb_data)
{
struct commit_list **list = cb_data;
struct object *object = parse_object(oid->hash);
if (!object)
return 0;
if (object->type == OBJ_TAG) {
object = deref_tag(object, path, strlen(path));
if (!object)
return 0;
}
if (object->type != OBJ_COMMIT)
return 0;
commit_list_insert((struct commit *)object, list);
return 0;
}
static int get_sha1_oneline(const char *prefix, unsigned char *sha1,
struct commit_list *list)
{
struct commit_list *backup = NULL, *l;
int found = 0;
int negative = 0;
regex_t regex;
if (prefix[0] == '!') {
prefix++;
if (prefix[0] == '-') {
prefix++;
negative = 1;
} else if (prefix[0] != '!') {
return -1;
}
}
if (regcomp(&regex, prefix, REG_EXTENDED))
return -1;
for (l = list; l; l = l->next) {
l->item->object.flags |= ONELINE_SEEN;
commit_list_insert(l->item, &backup);
}
while (list) {
const char *p, *buf;
struct commit *commit;
int matches;
commit = pop_most_recent_commit(&list, ONELINE_SEEN);
if (!parse_object(commit->object.oid.hash))
continue;
buf = get_commit_buffer(commit, NULL);
p = strstr(buf, "\n\n");
matches = negative ^ (p && !regexec(&regex, p + 2, 0, NULL, 0));
unuse_commit_buffer(commit, buf);
if (matches) {
hashcpy(sha1, commit->object.oid.hash);
found = 1;
break;
}
}
regfree(&regex);
free_commit_list(list);
for (l = backup; l; l = l->next)
clear_commit_marks(l->item, ONELINE_SEEN);
free_commit_list(backup);
return found ? 0 : -1;
}
struct grab_nth_branch_switch_cbdata {
int remaining;
struct strbuf buf;
};
static int grab_nth_branch_switch(unsigned char *osha1, unsigned char *nsha1,
const char *email, unsigned long timestamp, int tz,
const char *message, void *cb_data)
{
struct grab_nth_branch_switch_cbdata *cb = cb_data;
const char *match = NULL, *target = NULL;
size_t len;
if (skip_prefix(message, "checkout: moving from ", &match))
target = strstr(match, " to ");
if (!match || !target)
return 0;
if (--(cb->remaining) == 0) {
len = target - match;
strbuf_reset(&cb->buf);
strbuf_add(&cb->buf, match, len);
return 1; /* we are done */
}
return 0;
}
/*
Teach @{upstream} syntax to strbuf_branchanme() This teaches @{upstream} syntax to interpret_branch_name(), instead of dwim_ref() machinery. There are places in git UI that behaves differently when you give a local branch name and when you give an extended SHA-1 expression that evaluates to the commit object name at the tip of the branch. The intent is that the special syntax such as @{-1} can stand in as if the user spelled the name of the branch in such places. The name of the branch "frotz" to switch to ("git checkout frotz"), and the name of the branch "nitfol" to fork a new branch "frotz" from ("git checkout -b frotz nitfol"), are examples of such places. These places take only the name of the branch (e.g. "frotz"), and they are supposed to act differently to an equivalent refname (e.g. "refs/heads/frotz"), so hooking the @{upstream} and @{-N} syntax to dwim_ref() is insufficient when we want to deal with cases a local branch is forked from another local branch and use "forked@{upstream}" to name the forkee branch. The "upstream" syntax "forked@{u}" is to specify the ref that "forked" is configured to merge with, and most often the forkee is a remote tracking branch, not a local branch. We cannot simply return a local branch name, but that does not necessarily mean we have to returns the full refname (e.g. refs/remotes/origin/frotz, when returning origin/frotz is enough). This update calls shorten_unambiguous_ref() to do so. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-20 07:17:11 +00:00
* Parse @{-N} syntax, return the number of characters parsed
* if successful; otherwise signal an error with negative value.
*/
interpret_branch_name: always respect "namelen" parameter interpret_branch_name gets passed a "name" buffer to parse, along with a "namelen" parameter representing its length. If "namelen" is zero, we fallback to the NUL-terminated string-length of "name". However, it does not necessarily follow that if we have gotten a non-zero "namelen", it is the NUL-terminated string-length of "name". E.g., when get_sha1() is parsing "foo:bar", we will be asked to operate only on the first three characters. Yet in interpret_branch_name and its helpers, we use string functions like strchr() to operate on "name", looking past the length we were given. This can result in us mis-parsing object names. We should instead be limiting our search to "namelen" bytes. There are three distinct types of object names this patch addresses: - The intrepret_empty_at helper uses strchr to find the next @-expression after our potential empty-at. In an expression like "@:foo@bar", it erroneously thinks that the second "@" is relevant, even if we were asked only to look at the first character. This case is easy to trigger (and we test it in this patch). - When finding the initial @-mark for @{upstream}, we use strchr. This means we might treat "foo:@{upstream}" as the upstream for "foo:", even though we were asked only to look at "foo". We cannot test this one in practice, because it is masked by another bug (which is fixed in the next patch). - The interpret_nth_prior_checkout helper did not receive the name length at all. This turns out not to be a problem in practice, though, because its parsing is so limited: it always starts from the far-left of the string, and will not tolerate a colon (which is currently the only way to get a smaller-than-strlen "namelen"). However, it's still worth fixing to make the code more obviously correct, and to future-proof us against callers with more exotic buffers. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-01-15 08:31:57 +00:00
static int interpret_nth_prior_checkout(const char *name, int namelen,
struct strbuf *buf)
{
long nth;
int retval;
struct grab_nth_branch_switch_cbdata cb;
const char *brace;
char *num_end;
interpret_branch_name: always respect "namelen" parameter interpret_branch_name gets passed a "name" buffer to parse, along with a "namelen" parameter representing its length. If "namelen" is zero, we fallback to the NUL-terminated string-length of "name". However, it does not necessarily follow that if we have gotten a non-zero "namelen", it is the NUL-terminated string-length of "name". E.g., when get_sha1() is parsing "foo:bar", we will be asked to operate only on the first three characters. Yet in interpret_branch_name and its helpers, we use string functions like strchr() to operate on "name", looking past the length we were given. This can result in us mis-parsing object names. We should instead be limiting our search to "namelen" bytes. There are three distinct types of object names this patch addresses: - The intrepret_empty_at helper uses strchr to find the next @-expression after our potential empty-at. In an expression like "@:foo@bar", it erroneously thinks that the second "@" is relevant, even if we were asked only to look at the first character. This case is easy to trigger (and we test it in this patch). - When finding the initial @-mark for @{upstream}, we use strchr. This means we might treat "foo:@{upstream}" as the upstream for "foo:", even though we were asked only to look at "foo". We cannot test this one in practice, because it is masked by another bug (which is fixed in the next patch). - The interpret_nth_prior_checkout helper did not receive the name length at all. This turns out not to be a problem in practice, though, because its parsing is so limited: it always starts from the far-left of the string, and will not tolerate a colon (which is currently the only way to get a smaller-than-strlen "namelen"). However, it's still worth fixing to make the code more obviously correct, and to future-proof us against callers with more exotic buffers. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-01-15 08:31:57 +00:00
if (namelen < 4)
return -1;
if (name[0] != '@' || name[1] != '{' || name[2] != '-')
return -1;
interpret_branch_name: always respect "namelen" parameter interpret_branch_name gets passed a "name" buffer to parse, along with a "namelen" parameter representing its length. If "namelen" is zero, we fallback to the NUL-terminated string-length of "name". However, it does not necessarily follow that if we have gotten a non-zero "namelen", it is the NUL-terminated string-length of "name". E.g., when get_sha1() is parsing "foo:bar", we will be asked to operate only on the first three characters. Yet in interpret_branch_name and its helpers, we use string functions like strchr() to operate on "name", looking past the length we were given. This can result in us mis-parsing object names. We should instead be limiting our search to "namelen" bytes. There are three distinct types of object names this patch addresses: - The intrepret_empty_at helper uses strchr to find the next @-expression after our potential empty-at. In an expression like "@:foo@bar", it erroneously thinks that the second "@" is relevant, even if we were asked only to look at the first character. This case is easy to trigger (and we test it in this patch). - When finding the initial @-mark for @{upstream}, we use strchr. This means we might treat "foo:@{upstream}" as the upstream for "foo:", even though we were asked only to look at "foo". We cannot test this one in practice, because it is masked by another bug (which is fixed in the next patch). - The interpret_nth_prior_checkout helper did not receive the name length at all. This turns out not to be a problem in practice, though, because its parsing is so limited: it always starts from the far-left of the string, and will not tolerate a colon (which is currently the only way to get a smaller-than-strlen "namelen"). However, it's still worth fixing to make the code more obviously correct, and to future-proof us against callers with more exotic buffers. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-01-15 08:31:57 +00:00
brace = memchr(name, '}', namelen);
if (!brace)
return -1;
nth = strtol(name + 3, &num_end, 10);
if (num_end != brace)
return -1;
if (nth <= 0)
return -1;
cb.remaining = nth;
strbuf_init(&cb.buf, 20);
retval = 0;
if (0 < for_each_reflog_ent_reverse("HEAD", grab_nth_branch_switch, &cb)) {
strbuf_reset(buf);
strbuf_addbuf(buf, &cb.buf);
retval = brace - name + 1;
}
strbuf_release(&cb.buf);
return retval;
}
int get_sha1_mb(const char *name, unsigned char *sha1)
{
struct commit *one, *two;
struct commit_list *mbs;
unsigned char sha1_tmp[20];
const char *dots;
int st;
dots = strstr(name, "...");
if (!dots)
return get_sha1(name, sha1);
if (dots == name)
st = get_sha1("HEAD", sha1_tmp);
else {
struct strbuf sb;
strbuf_init(&sb, dots - name);
strbuf_add(&sb, name, dots - name);
st = get_sha1_committish(sb.buf, sha1_tmp);
strbuf_release(&sb);
}
if (st)
return st;
one = lookup_commit_reference_gently(sha1_tmp, 0);
if (!one)
return -1;
if (get_sha1_committish(dots[3] ? (dots + 3) : "HEAD", sha1_tmp))
return -1;
two = lookup_commit_reference_gently(sha1_tmp, 0);
if (!two)
return -1;
mbs = get_merge_bases(one, two);
if (!mbs || mbs->next)
st = -1;
else {
st = 0;
hashcpy(sha1, mbs->item->object.oid.hash);
}
free_commit_list(mbs);
return st;
}
/* parse @something syntax, when 'something' is not {.*} */
static int interpret_empty_at(const char *name, int namelen, int len, struct strbuf *buf)
{
const char *next;
if (len || name[1] == '{')
return -1;
/* make sure it's a single @, or @@{.*}, not @foo */
interpret_branch_name: always respect "namelen" parameter interpret_branch_name gets passed a "name" buffer to parse, along with a "namelen" parameter representing its length. If "namelen" is zero, we fallback to the NUL-terminated string-length of "name". However, it does not necessarily follow that if we have gotten a non-zero "namelen", it is the NUL-terminated string-length of "name". E.g., when get_sha1() is parsing "foo:bar", we will be asked to operate only on the first three characters. Yet in interpret_branch_name and its helpers, we use string functions like strchr() to operate on "name", looking past the length we were given. This can result in us mis-parsing object names. We should instead be limiting our search to "namelen" bytes. There are three distinct types of object names this patch addresses: - The intrepret_empty_at helper uses strchr to find the next @-expression after our potential empty-at. In an expression like "@:foo@bar", it erroneously thinks that the second "@" is relevant, even if we were asked only to look at the first character. This case is easy to trigger (and we test it in this patch). - When finding the initial @-mark for @{upstream}, we use strchr. This means we might treat "foo:@{upstream}" as the upstream for "foo:", even though we were asked only to look at "foo". We cannot test this one in practice, because it is masked by another bug (which is fixed in the next patch). - The interpret_nth_prior_checkout helper did not receive the name length at all. This turns out not to be a problem in practice, though, because its parsing is so limited: it always starts from the far-left of the string, and will not tolerate a colon (which is currently the only way to get a smaller-than-strlen "namelen"). However, it's still worth fixing to make the code more obviously correct, and to future-proof us against callers with more exotic buffers. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-01-15 08:31:57 +00:00
next = memchr(name + len + 1, '@', namelen - len - 1);
if (next && next[1] != '{')
return -1;
if (!next)
next = name + namelen;
if (next != name + 1)
return -1;
strbuf_reset(buf);
strbuf_add(buf, "HEAD", 4);
return 1;
}
static int reinterpret(const char *name, int namelen, int len, struct strbuf *buf)
{
/* we have extra data, which might need further processing */
struct strbuf tmp = STRBUF_INIT;
int used = buf->len;
int ret;
strbuf_add(buf, name + len, namelen - len);
ret = interpret_branch_name(buf->buf, buf->len, &tmp);
/* that data was not interpreted, remove our cruft */
if (ret < 0) {
strbuf_setlen(buf, used);
return len;
}
strbuf_reset(buf);
strbuf_addbuf(buf, &tmp);
strbuf_release(&tmp);
/* tweak for size of {-N} versus expanded ref name */
return ret - used + len;
}
static void set_shortened_ref(struct strbuf *buf, const char *ref)
{
char *s = shorten_unambiguous_ref(ref, 0);
strbuf_reset(buf);
strbuf_addstr(buf, s);
free(s);
}
static int interpret_branch_mark(const char *name, int namelen,
int at, struct strbuf *buf,
int (*get_mark)(const char *, int),
const char *(*get_data)(struct branch *,
struct strbuf *))
{
int len;
struct branch *branch;
struct strbuf err = STRBUF_INIT;
const char *value;
len = get_mark(name + at, namelen - at);
if (!len)
return -1;
interpret_branch_name: avoid @{upstream} past colon get_sha1() cannot currently parse a valid object name like "HEAD:@{upstream}" (assuming that such an oddly named file exists in the HEAD commit). It takes two passes to parse the string: 1. It first considers the whole thing as a ref, which results in looking for the upstream of "HEAD:". 2. It finds the colon, parses "HEAD" as a tree-ish, and then finds the path "@{upstream}" in the tree. For a path that looks like a normal reflog (e.g., "HEAD:@{yesterday}"), the first pass is a no-op. We try to dwim_ref("HEAD:"), that returns zero refs, and we proceed with colon-parsing. For "HEAD:@{upstream}", though, the first pass ends up in interpret_upstream_mark, which tries to find the branch "HEAD:". When it sees that the branch does not exist, it actually dies rather than returning an error to the caller. As a result, we never make it to the second pass. One obvious way of fixing this would be to teach interpret_upstream_mark to simply report "no, this isn't an upstream" in such a case. However, that would make the error-reporting for legitimate upstream cases significantly worse. Something like "bogus@{upstream}" would simply report "unknown revision: bogus@{upstream}", while the current code diagnoses a wide variety of possible misconfigurations (no such branch, branch exists but does not have upstream, etc). However, we can take advantage of the fact that a branch name cannot contain a colon. Therefore even if we find an upstream mark, any prefix with a colon must mean that the upstream mark we found is actually a pathname, and should be disregarded completely. This patch implements that logic. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-01-15 08:37:23 +00:00
if (memchr(name, ':', at))
return -1;
if (at) {
char *name_str = xmemdupz(name, at);
branch = branch_get(name_str);
free(name_str);
} else
branch = branch_get(NULL);
value = get_data(branch, &err);
if (!value)
die("%s", err.buf);
set_shortened_ref(buf, value);
return len + at;
}
Teach @{upstream} syntax to strbuf_branchanme() This teaches @{upstream} syntax to interpret_branch_name(), instead of dwim_ref() machinery. There are places in git UI that behaves differently when you give a local branch name and when you give an extended SHA-1 expression that evaluates to the commit object name at the tip of the branch. The intent is that the special syntax such as @{-1} can stand in as if the user spelled the name of the branch in such places. The name of the branch "frotz" to switch to ("git checkout frotz"), and the name of the branch "nitfol" to fork a new branch "frotz" from ("git checkout -b frotz nitfol"), are examples of such places. These places take only the name of the branch (e.g. "frotz"), and they are supposed to act differently to an equivalent refname (e.g. "refs/heads/frotz"), so hooking the @{upstream} and @{-N} syntax to dwim_ref() is insufficient when we want to deal with cases a local branch is forked from another local branch and use "forked@{upstream}" to name the forkee branch. The "upstream" syntax "forked@{u}" is to specify the ref that "forked" is configured to merge with, and most often the forkee is a remote tracking branch, not a local branch. We cannot simply return a local branch name, but that does not necessarily mean we have to returns the full refname (e.g. refs/remotes/origin/frotz, when returning origin/frotz is enough). This update calls shorten_unambiguous_ref() to do so. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-20 07:17:11 +00:00
/*
* This reads short-hand syntax that not only evaluates to a commit
* object name, but also can act as if the end user spelled the name
* of the branch from the command line.
*
* - "@{-N}" finds the name of the Nth previous branch we were on, and
* places the name of the branch in the given buf and returns the
* number of characters parsed if successful.
*
* - "<branch>@{upstream}" finds the name of the other ref that
* <branch> is configured to merge with (missing <branch> defaults
* to the current branch), and places the name of the branch in the
* given buf and returns the number of characters parsed if
* successful.
*
* If the input is not of the accepted format, it returns a negative
* number to signal an error.
*
* If the input was ok but there are not N branch switches in the
* reflog, it returns 0.
*/
int interpret_branch_name(const char *name, int namelen, struct strbuf *buf)
Teach @{upstream} syntax to strbuf_branchanme() This teaches @{upstream} syntax to interpret_branch_name(), instead of dwim_ref() machinery. There are places in git UI that behaves differently when you give a local branch name and when you give an extended SHA-1 expression that evaluates to the commit object name at the tip of the branch. The intent is that the special syntax such as @{-1} can stand in as if the user spelled the name of the branch in such places. The name of the branch "frotz" to switch to ("git checkout frotz"), and the name of the branch "nitfol" to fork a new branch "frotz" from ("git checkout -b frotz nitfol"), are examples of such places. These places take only the name of the branch (e.g. "frotz"), and they are supposed to act differently to an equivalent refname (e.g. "refs/heads/frotz"), so hooking the @{upstream} and @{-N} syntax to dwim_ref() is insufficient when we want to deal with cases a local branch is forked from another local branch and use "forked@{upstream}" to name the forkee branch. The "upstream" syntax "forked@{u}" is to specify the ref that "forked" is configured to merge with, and most often the forkee is a remote tracking branch, not a local branch. We cannot simply return a local branch name, but that does not necessarily mean we have to returns the full refname (e.g. refs/remotes/origin/frotz, when returning origin/frotz is enough). This update calls shorten_unambiguous_ref() to do so. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-20 07:17:11 +00:00
{
char *at;
const char *start;
interpret_branch_name: always respect "namelen" parameter interpret_branch_name gets passed a "name" buffer to parse, along with a "namelen" parameter representing its length. If "namelen" is zero, we fallback to the NUL-terminated string-length of "name". However, it does not necessarily follow that if we have gotten a non-zero "namelen", it is the NUL-terminated string-length of "name". E.g., when get_sha1() is parsing "foo:bar", we will be asked to operate only on the first three characters. Yet in interpret_branch_name and its helpers, we use string functions like strchr() to operate on "name", looking past the length we were given. This can result in us mis-parsing object names. We should instead be limiting our search to "namelen" bytes. There are three distinct types of object names this patch addresses: - The intrepret_empty_at helper uses strchr to find the next @-expression after our potential empty-at. In an expression like "@:foo@bar", it erroneously thinks that the second "@" is relevant, even if we were asked only to look at the first character. This case is easy to trigger (and we test it in this patch). - When finding the initial @-mark for @{upstream}, we use strchr. This means we might treat "foo:@{upstream}" as the upstream for "foo:", even though we were asked only to look at "foo". We cannot test this one in practice, because it is masked by another bug (which is fixed in the next patch). - The interpret_nth_prior_checkout helper did not receive the name length at all. This turns out not to be a problem in practice, though, because its parsing is so limited: it always starts from the far-left of the string, and will not tolerate a colon (which is currently the only way to get a smaller-than-strlen "namelen"). However, it's still worth fixing to make the code more obviously correct, and to future-proof us against callers with more exotic buffers. Signed-off-by: Jeff King <peff@peff.net> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2014-01-15 08:31:57 +00:00
int len = interpret_nth_prior_checkout(name, namelen, buf);
Teach @{upstream} syntax to strbuf_branchanme() This teaches @{upstream} syntax to interpret_branch_name(), instead of dwim_ref() machinery. There are places in git UI that behaves differently when you give a local branch name and when you give an extended SHA-1 expression that evaluates to the commit object name at the tip of the branch. The intent is that the special syntax such as @{-1} can stand in as if the user spelled the name of the branch in such places. The name of the branch "frotz" to switch to ("git checkout frotz"), and the name of the branch "nitfol" to fork a new branch "frotz" from ("git checkout -b frotz nitfol"), are examples of such places. These places take only the name of the branch (e.g. "frotz"), and they are supposed to act differently to an equivalent refname (e.g. "refs/heads/frotz"), so hooking the @{upstream} and @{-N} syntax to dwim_ref() is insufficient when we want to deal with cases a local branch is forked from another local branch and use "forked@{upstream}" to name the forkee branch. The "upstream" syntax "forked@{u}" is to specify the ref that "forked" is configured to merge with, and most often the forkee is a remote tracking branch, not a local branch. We cannot simply return a local branch name, but that does not necessarily mean we have to returns the full refname (e.g. refs/remotes/origin/frotz, when returning origin/frotz is enough). This update calls shorten_unambiguous_ref() to do so. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-20 07:17:11 +00:00
if (!namelen)
namelen = strlen(name);
if (!len) {
Teach @{upstream} syntax to strbuf_branchanme() This teaches @{upstream} syntax to interpret_branch_name(), instead of dwim_ref() machinery. There are places in git UI that behaves differently when you give a local branch name and when you give an extended SHA-1 expression that evaluates to the commit object name at the tip of the branch. The intent is that the special syntax such as @{-1} can stand in as if the user spelled the name of the branch in such places. The name of the branch "frotz" to switch to ("git checkout frotz"), and the name of the branch "nitfol" to fork a new branch "frotz" from ("git checkout -b frotz nitfol"), are examples of such places. These places take only the name of the branch (e.g. "frotz"), and they are supposed to act differently to an equivalent refname (e.g. "refs/heads/frotz"), so hooking the @{upstream} and @{-N} syntax to dwim_ref() is insufficient when we want to deal with cases a local branch is forked from another local branch and use "forked@{upstream}" to name the forkee branch. The "upstream" syntax "forked@{u}" is to specify the ref that "forked" is configured to merge with, and most often the forkee is a remote tracking branch, not a local branch. We cannot simply return a local branch name, but that does not necessarily mean we have to returns the full refname (e.g. refs/remotes/origin/frotz, when returning origin/frotz is enough). This update calls shorten_unambiguous_ref() to do so. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-20 07:17:11 +00:00
return len; /* syntax Ok, not enough switches */
} else if (len > 0) {
if (len == namelen)
return len; /* consumed all */
else
return reinterpret(name, namelen, len, buf);
}
for (start = name;
(at = memchr(start, '@', namelen - (start - name)));
start = at + 1) {
len = interpret_empty_at(name, namelen, at - name, buf);
if (len > 0)
return reinterpret(name, namelen, len, buf);
len = interpret_branch_mark(name, namelen, at - name, buf,
upstream_mark, branch_get_upstream);
if (len > 0)
return len;
len = interpret_branch_mark(name, namelen, at - name, buf,
push_mark, branch_get_push);
if (len > 0)
return len;
}
return -1;
Teach @{upstream} syntax to strbuf_branchanme() This teaches @{upstream} syntax to interpret_branch_name(), instead of dwim_ref() machinery. There are places in git UI that behaves differently when you give a local branch name and when you give an extended SHA-1 expression that evaluates to the commit object name at the tip of the branch. The intent is that the special syntax such as @{-1} can stand in as if the user spelled the name of the branch in such places. The name of the branch "frotz" to switch to ("git checkout frotz"), and the name of the branch "nitfol" to fork a new branch "frotz" from ("git checkout -b frotz nitfol"), are examples of such places. These places take only the name of the branch (e.g. "frotz"), and they are supposed to act differently to an equivalent refname (e.g. "refs/heads/frotz"), so hooking the @{upstream} and @{-N} syntax to dwim_ref() is insufficient when we want to deal with cases a local branch is forked from another local branch and use "forked@{upstream}" to name the forkee branch. The "upstream" syntax "forked@{u}" is to specify the ref that "forked" is configured to merge with, and most often the forkee is a remote tracking branch, not a local branch. We cannot simply return a local branch name, but that does not necessarily mean we have to returns the full refname (e.g. refs/remotes/origin/frotz, when returning origin/frotz is enough). This update calls shorten_unambiguous_ref() to do so. Signed-off-by: Junio C Hamano <gitster@pobox.com>
2010-01-20 07:17:11 +00:00
}
int strbuf_branchname(struct strbuf *sb, const char *name)
{
int len = strlen(name);
int used = interpret_branch_name(name, len, sb);
if (used == len)
return 0;
if (used < 0)
used = 0;
strbuf_add(sb, name + used, len - used);
return len;
}
int strbuf_check_branch_ref(struct strbuf *sb, const char *name)
{
strbuf_branchname(sb, name);
if (name[0] == '-')
return -1;
strbuf_splice(sb, 0, 0, "refs/heads/", 11);
return check_refname_format(sb->buf, 0);
}
/*
* This is like "get_sha1_basic()", except it allows "sha1 expressions",
* notably "xyz^" for "parent of xyz"
*/
int get_sha1(const char *name, unsigned char *sha1)
{
struct object_context unused;
return get_sha1_with_context(name, 0, sha1, &unused);
}
/*
* Many callers know that the user meant to name a commit-ish by
* syntactical positions where the object name appears. Calling this
* function allows the machinery to disambiguate shorter-than-unique
* abbreviated object names between commit-ish and others.
*
* Note that this does NOT error out when the named object is not a
* commit-ish. It is merely to give a hint to the disambiguation
* machinery.
*/
int get_sha1_committish(const char *name, unsigned char *sha1)
{
struct object_context unused;
return get_sha1_with_context(name, GET_SHA1_COMMITTISH,
sha1, &unused);
}
int get_sha1_treeish(const char *name, unsigned char *sha1)
{
struct object_context unused;
return get_sha1_with_context(name, GET_SHA1_TREEISH,
sha1, &unused);
}
int get_sha1_commit(const char *name, unsigned char *sha1)
{
struct object_context unused;
return get_sha1_with_context(name, GET_SHA1_COMMIT,
sha1, &unused);
}
int get_sha1_tree(const char *name, unsigned char *sha1)
{
struct object_context unused;
return get_sha1_with_context(name, GET_SHA1_TREE,
sha1, &unused);
}
int get_sha1_blob(const char *name, unsigned char *sha1)
{
struct object_context unused;
return get_sha1_with_context(name, GET_SHA1_BLOB,
sha1, &unused);
}
/* Must be called only when object_name:filename doesn't exist. */
static void diagnose_invalid_sha1_path(const char *prefix,
const char *filename,
const unsigned char *tree_sha1,
const char *object_name,
int object_name_len)
{
unsigned char sha1[20];
unsigned mode;
if (!prefix)
prefix = "";
if (file_exists(filename))
die("Path '%s' exists on disk, but not in '%.*s'.",
filename, object_name_len, object_name);
if (errno == ENOENT || errno == ENOTDIR) {
char *fullname = xstrfmt("%s%s", prefix, filename);
if (!get_tree_entry(tree_sha1, fullname,
sha1, &mode)) {
die("Path '%s' exists, but not '%s'.\n"
"Did you mean '%.*s:%s' aka '%.*s:./%s'?",
fullname,
filename,
object_name_len, object_name,
fullname,
object_name_len, object_name,
filename);
}
die("Path '%s' does not exist in '%.*s'",
filename, object_name_len, object_name);
}
}
/* Must be called only when :stage:filename doesn't exist. */
static void diagnose_invalid_index_path(int stage,
const char *prefix,
const char *filename)
{
Convert "struct cache_entry *" to "const ..." wherever possible I attempted to make index_state->cache[] a "const struct cache_entry **" to find out how existing entries in index are modified and where. The question I have is what do we do if we really need to keep track of on-disk changes in the index. The result is - diff-lib.c: setting CE_UPTODATE - name-hash.c: setting CE_HASHED - preload-index.c, read-cache.c, unpack-trees.c and builtin/update-index: obvious - entry.c: write_entry() may refresh the checked out entry via fill_stat_cache_info(). This causes "non-const struct cache_entry *" in builtin/apply.c, builtin/checkout-index.c and builtin/checkout.c - builtin/ls-files.c: --with-tree changes stagemask and may set CE_UPDATE Of these, write_entry() and its call sites are probably most interesting because it modifies on-disk info. But this is stat info and can be retrieved via refresh, at least for porcelain commands. Other just uses ce_flags for local purposes. So, keeping track of "dirty" entries is just a matter of setting a flag in index modification functions exposed by read-cache.c. Except unpack-trees, the rest of the code base does not do anything funny behind read-cache's back. The actual patch is less valueable than the summary above. But if anyone wants to re-identify the above sites. Applying this patch, then this: diff --git a/cache.h b/cache.h index 430d021..1692891 100644 --- a/cache.h +++ b/cache.h @@ -267,7 +267,7 @@ static inline unsigned int canon_mode(unsigned int mode) #define cache_entry_size(len) (offsetof(struct cache_entry,name) + (len) + 1) struct index_state { - struct cache_entry **cache; + const struct cache_entry **cache; unsigned int version; unsigned int cache_nr, cache_alloc, cache_changed; struct string_list *resolve_undo; will help quickly identify them without bogus warnings. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-07-09 15:29:00 +00:00
const struct cache_entry *ce;
int pos;
unsigned namelen = strlen(filename);
struct strbuf fullname = STRBUF_INIT;
if (!prefix)
prefix = "";
/* Wrong stage number? */
pos = cache_name_pos(filename, namelen);
if (pos < 0)
pos = -pos - 1;
if (pos < active_nr) {
ce = active_cache[pos];
if (ce_namelen(ce) == namelen &&
!memcmp(ce->name, filename, namelen))
die("Path '%s' is in the index, but not at stage %d.\n"
"Did you mean ':%d:%s'?",
filename, stage,
ce_stage(ce), filename);
}
/* Confusion between relative and absolute filenames? */
strbuf_addstr(&fullname, prefix);
strbuf_addstr(&fullname, filename);
pos = cache_name_pos(fullname.buf, fullname.len);
if (pos < 0)
pos = -pos - 1;
if (pos < active_nr) {
ce = active_cache[pos];
if (ce_namelen(ce) == fullname.len &&
!memcmp(ce->name, fullname.buf, fullname.len))
die("Path '%s' is in the index, but not '%s'.\n"
"Did you mean ':%d:%s' aka ':%d:./%s'?",
fullname.buf, filename,
ce_stage(ce), fullname.buf,
ce_stage(ce), filename);
}
if (file_exists(filename))
die("Path '%s' exists on disk, but not in the index.", filename);
if (errno == ENOENT || errno == ENOTDIR)
die("Path '%s' does not exist (neither on disk nor in the index).",
filename);
strbuf_release(&fullname);
}
static char *resolve_relative_path(const char *rel)
{
if (!starts_with(rel, "./") && !starts_with(rel, "../"))
return NULL;
if (!startup_info)
die("BUG: startup_info struct is not initialized.");
if (!is_inside_work_tree())
die("relative path syntax can't be used outside working tree.");
/* die() inside prefix_path() if resolved path is outside worktree */
return prefix_path(startup_info->prefix,
startup_info->prefix ? strlen(startup_info->prefix) : 0,
rel);
}
static int get_sha1_with_context_1(const char *name,
unsigned flags,
const char *prefix,
unsigned char *sha1,
struct object_context *oc)
{
int ret, bracket_depth;
int namelen = strlen(name);
const char *cp;
int only_to_die = flags & GET_SHA1_ONLY_TO_DIE;
memset(oc, 0, sizeof(*oc));
oc->mode = S_IFINVALID;
ret = get_sha1_1(name, namelen, sha1, flags);
if (!ret)
return ret;
/*
* sha1:path --> object name of path in ent sha1
* :path -> object name of absolute path in index
* :./path -> object name of path relative to cwd in index
* :[0-3]:path -> object name of path in index at stage
* :/foo -> recent commit matching foo
*/
if (name[0] == ':') {
int stage = 0;
Convert "struct cache_entry *" to "const ..." wherever possible I attempted to make index_state->cache[] a "const struct cache_entry **" to find out how existing entries in index are modified and where. The question I have is what do we do if we really need to keep track of on-disk changes in the index. The result is - diff-lib.c: setting CE_UPTODATE - name-hash.c: setting CE_HASHED - preload-index.c, read-cache.c, unpack-trees.c and builtin/update-index: obvious - entry.c: write_entry() may refresh the checked out entry via fill_stat_cache_info(). This causes "non-const struct cache_entry *" in builtin/apply.c, builtin/checkout-index.c and builtin/checkout.c - builtin/ls-files.c: --with-tree changes stagemask and may set CE_UPDATE Of these, write_entry() and its call sites are probably most interesting because it modifies on-disk info. But this is stat info and can be retrieved via refresh, at least for porcelain commands. Other just uses ce_flags for local purposes. So, keeping track of "dirty" entries is just a matter of setting a flag in index modification functions exposed by read-cache.c. Except unpack-trees, the rest of the code base does not do anything funny behind read-cache's back. The actual patch is less valueable than the summary above. But if anyone wants to re-identify the above sites. Applying this patch, then this: diff --git a/cache.h b/cache.h index 430d021..1692891 100644 --- a/cache.h +++ b/cache.h @@ -267,7 +267,7 @@ static inline unsigned int canon_mode(unsigned int mode) #define cache_entry_size(len) (offsetof(struct cache_entry,name) + (len) + 1) struct index_state { - struct cache_entry **cache; + const struct cache_entry **cache; unsigned int version; unsigned int cache_nr, cache_alloc, cache_changed; struct string_list *resolve_undo; will help quickly identify them without bogus warnings. Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
2013-07-09 15:29:00 +00:00
const struct cache_entry *ce;
char *new_path = NULL;
int pos;
if (!only_to_die && namelen > 2 && name[1] == '/') {
struct commit_list *list = NULL;
for_each_ref(handle_one_ref, &list);
commit_list_sort_by_date(&list);
return get_sha1_oneline(name + 2, sha1, list);
}
if (namelen < 3 ||
name[2] != ':' ||
name[1] < '0' || '3' < name[1])
cp = name + 1;
else {
stage = name[1] - '0';
cp = name + 3;
}
new_path = resolve_relative_path(cp);
if (!new_path) {
namelen = namelen - (cp - name);
} else {
cp = new_path;
namelen = strlen(cp);
}
strlcpy(oc->path, cp, sizeof(oc->path));
if (!active_cache)
read_cache();
pos = cache_name_pos(cp, namelen);
if (pos < 0)
pos = -pos - 1;
while (pos < active_nr) {
ce = active_cache[pos];
if (ce_namelen(ce) != namelen ||
memcmp(ce->name, cp, namelen))
break;
if (ce_stage(ce) == stage) {
hashcpy(sha1, ce->sha1);
oc->mode = ce->ce_mode;
free(new_path);
return 0;
}
pos++;
}
if (only_to_die && name[1] && name[1] != '/')
diagnose_invalid_index_path(stage, prefix, cp);
free(new_path);
return -1;
}
for (cp = name, bracket_depth = 0; *cp; cp++) {
if (*cp == '{')
bracket_depth++;
else if (bracket_depth && *cp == '}')
bracket_depth--;
else if (!bracket_depth && *cp == ':')
break;
}
if (*cp == ':') {
unsigned char tree_sha1[20];
int len = cp - name;
if (!get_sha1_1(name, len, tree_sha1, GET_SHA1_TREEISH)) {
const char *filename = cp+1;
char *new_filename = NULL;
new_filename = resolve_relative_path(filename);
if (new_filename)
filename = new_filename;
if (flags & GET_SHA1_FOLLOW_SYMLINKS) {
ret = get_tree_entry_follow_symlinks(tree_sha1,
filename, sha1, &oc->symlink_path,
&oc->mode);
} else {
ret = get_tree_entry(tree_sha1, filename,
sha1, &oc->mode);
if (ret && only_to_die) {
diagnose_invalid_sha1_path(prefix,
filename,
tree_sha1,
name, len);
}
}
hashcpy(oc->tree, tree_sha1);
strlcpy(oc->path, filename, sizeof(oc->path));
free(new_filename);
return ret;
} else {
if (only_to_die)
die("Invalid object name '%.*s'.", len, name);
}
}
return ret;
}
/*
* Call this function when you know "name" given by the end user must
* name an object but it doesn't; the function _may_ die with a better
* diagnostic message than "no such object 'name'", e.g. "Path 'doc' does not
* exist in 'HEAD'" when given "HEAD:doc", or it may return in which case
* you have a chance to diagnose the error further.
*/
void maybe_die_on_misspelt_object_name(const char *name, const char *prefix)
{
struct object_context oc;
unsigned char sha1[20];
get_sha1_with_context_1(name, GET_SHA1_ONLY_TO_DIE, prefix, sha1, &oc);
}
int get_sha1_with_context(const char *str, unsigned flags, unsigned char *sha1, struct object_context *orc)
{
if (flags & GET_SHA1_FOLLOW_SYMLINKS && flags & GET_SHA1_ONLY_TO_DIE)
die("BUG: incompatible flags for get_sha1_with_context");
return get_sha1_with_context_1(str, flags, NULL, sha1, orc);
}