The word-diff logic accumulates + and - lines until another line type
appears (normally [ @\]), at which point it generates the word diff.
This is usually correct, but it breaks when the preimage does not have
a newline at EOF:
$ printf "%s" "a a a" >a
$ printf "%s\n" "a ab a" >b
$ git diff --no-index --word-diff a b
diff --git 1/a 2/b
index 9f68e94..6a7c02f 100644
--- 1/a
+++ 2/b
@@ -1 +1 @@
[-a a a-]
No newline at end of file
{+a ab a+}
Because of the order of the lines in a unified diff
@@ -1 +1 @@
-a a a
\ No newline at end of file
+a ab a
the '\' line flushed the buffers, and the - and + lines were never
matched with each other.
A proper fix would defer such markers until the end of the hunk.
However, word-diff is inherently whitespace-ignoring, so as a cheap
fix simply ignore the marker (and hide it from the output).
We use a prefix match for '\ ' to parallel the logic in
apply.c:parse_fragment(). We currently do not localize this string
(just accept other variants of it in git-apply), but this should be
future-proof.
Noticed-by: Ivan Shirokoff <shirokoff@yandex-team.ru>
Signed-off-by: Thomas Rast <trast@student.ethz.ch>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add the option -W/--function-context to git diff. It is similar to
the same option of git grep and expands the context of change hunks
so that the whole surrounding function is shown. This "natural"
context can allow changes to be understood better.
Note: GNU patch doesn't like diffs generated with the new option;
it seems to expect context lines to be the same before and after
changes. git apply doesn't complain.
This implementation has the same shortcoming as the one in grep,
namely that there is no way to explicitly find the end of a
function. That means that a few lines of extra context are shown,
right up to the next recognized function begins. It's already
useful in its current form, though.
The function get_func_line() in xdiff/xemit.c is extended to work
forward as well as backward to find post-context as well as
pre-context. It returns the position of the first found matching
line. The func_line parameter is made optional, as we don't need
it for -W.
The enhanced function is then used in xdl_emit_diff() to extend
the context as needed. If the added context overlaps with the
next change, it is merged into the current hunk.
Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Earlier, 582aa00 (git diff too slow for a file, 2010-05-02)
unconditionally dropped XDF_NEED_MINIMAL option from the internal xdiff
invocation to help performance on pathological cases, while hinting that a
follow-up patch could reintroduce it with "--minimal" option from the
command line.
Make it so.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git diff -p" piped to external diffstat and "git diff --stat" may see
different patch text (both are valid and describe the same change
correctly) when counting the number of added and deleted lines, arriving
at different results to confuse the users, as --stat/--numstat codepath
always uses the hardcoded -U0 as the context length.
Make --stat/--numstat codepath to honor the context length the same way
as the textual patch codepath does to avoid this problem.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* jk/color-and-pager:
want_color: automatically fallback to color.ui
diff: don't load color config in plumbing
config: refactor get_colorbool function
color: delay auto-color decision until point of use
git_config_colorbool: refactor stdout_is_tty handling
diff: refactor COLOR_DIFF from a flag into an int
setup_pager: set GIT_PAGER_IN_USE
t7006: use test_config helpers
test-lib: add helper functions for config
t7006: modernize calls to unset
Conflicts:
builtin/commit.c
parse-options.c
The diff config callback is split into two functions: one
which loads "ui" config, and one which loads "basic" config.
The former chains to the latter, as the diff UI config is a
superset of the plumbing config.
The color.diff variable is only loaded in the UI config.
However, the basic config actually chains to
git_color_default_config, which loads color.ui. This doesn't
actually cause any bugs, because the plumbing diff code does
not actually look at the value of color.ui.
However, it is somewhat nonsensical, and it makes it
difficult to refactor the color code. It probably came about
because there is no git_color_config to load only color
config, but rather just git_color_default_config, which
loads color config and chains to git_default_config.
This patch splits out the color-specific portion of
git_color_default_config so that the diff UI config can call
it directly. This is perhaps better explained by the
chaining of callbacks. Before we had:
git_diff_ui_config
-> git_diff_basic_config
-> git_color_default_config
-> git_default_config
Now we have:
git_diff_ui_config
-> git_color_config
-> git_diff_basic_config
-> git_default_config
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
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>
Usually this function figures out for itself whether stdout
is a tty. However, it has an extra parameter just to allow
git-config to override the auto-detection for its
--get-colorbool option.
Instead of an extra parameter, let's just use a global
variable. This makes calling easier in the common case, and
will make refactoring the colorbool code much simpler.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This lets us store more than just a bit flag for whether we
want color; we can also store whether we want automatic
colors. This can be useful for making the automatic-color
decision closer to the point of use.
This mostly just involves replacing DIFF_OPT_* calls with
manipulations of the flag. The biggest exception is that
calls to DIFF_OPT_TST must check for "o->use_color > 0",
which lets an "unknown" value (i.e., the default) stay at
"no color". In the previous code, a value of "-1" was not
propagated at all.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* rc/histogram-diff:
xdiff/xhistogram: drop need for additional variable
xdiff/xhistogram: rely on xdl_trim_ends()
xdiff/xhistogram: rework handling of recursed results
xdiff: do away with xdl_mmfile_next()
Make test number unique
xdiff/xprepare: use a smaller sample size for histogram diff
xdiff/xprepare: skip classification
teach --histogram to diff
t4033-diff-patience: factor out tests
xdiff/xpatience: factor out fall-back-diff function
xdiff/xprepare: refactor abort cleanups
xdiff/xprepare: use memset()
* jc/zlib-wrap:
zlib: allow feeding more than 4GB in one go
zlib: zlib can only process 4GB at a time
zlib: wrap deflateBound() too
zlib: wrap deflate side of the API
zlib: wrap inflateInit2 used to accept only for gzip format
zlib: wrap remaining calls to direct inflate/inflateEnd
zlib wrapper: refactor error message formatter
* jk/combine-diff-binary-etc:
combine-diff: respect textconv attributes
refactor get_textconv to not require diff_filespec
combine-diff: handle binary files as binary
combine-diff: calculate mode_differs earlier
combine-diff: split header printing into its own function
* jc/zlib-wrap:
zlib: allow feeding more than 4GB in one go
zlib: zlib can only process 4GB at a time
zlib: wrap deflateBound() too
zlib: wrap deflate side of the API
zlib: wrap inflateInit2 used to accept only for gzip format
zlib: wrap remaining calls to direct inflate/inflateEnd
zlib wrapper: refactor error message formatter
Conflicts:
sha1_file.c
Port JGit's HistogramDiff algorithm over to C. Rough numbers (TODO) show
that it is faster than its --patience cousin, as well as the default
Meyers algorithm.
The implementation has been reworked to use structs and pointers,
instead of bitmasks, thus doing away with JGit's 2^28 line limit.
We also use xdiff's default hash table implementation (xdl_hash_bits()
with XDL_HASHLONG()) for convenience.
Signed-off-by: Tay Ray Chuan <rctay89@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* jk/combine-diff-binary-etc:
combine-diff: respect textconv attributes
refactor get_textconv to not require diff_filespec
combine-diff: handle binary files as binary
combine-diff: calculate mode_differs earlier
combine-diff: split header printing into its own function
The size of objects we read from the repository and data we try to put
into the repository are represented in "unsigned long", so that on larger
architectures we can handle objects that weigh more than 4GB.
But the interface defined in zlib.h to communicate with inflate/deflate
limits avail_in (how many bytes of input are we calling zlib with) and
avail_out (how many bytes of output from zlib are we ready to accept)
fields effectively to 4GB by defining their type to be uInt.
In many places in our code, we allocate a large buffer (e.g. mmap'ing a
large loose object file) and tell zlib its size by assigning the size to
avail_in field of the stream, but that will truncate the high octets of
the real size. The worst part of this story is that we often pass around
z_stream (the state object used by zlib) to keep track of the number of
used bytes in input/output buffer by inspecting these two fields, which
practically limits our callchain to the same 4GB limit.
Wrap z_stream in another structure git_zstream that can express avail_in
and avail_out in unsigned long. For now, just die() when the caller gives
a size that cannot be given to a single zlib call. In later patches in the
series, we would make git_inflate() and git_deflate() internally loop to
give callers an illusion that our "improved" version of zlib interface can
operate on a buffer larger than 4GB in one go.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Wrap deflateInit, deflate, and deflateEnd for everybody, and the sole use
of deflateInit2 in remote-curl.c to tell the library to use gzip header
and trailer in git_deflate_init_gzip().
There is only one caller that cares about the status from deflateEnd().
Introduce git_deflate_end_gently() to let that sole caller retrieve the
status and act on it (i.e. die) for now, but we would probably want to
make inflate_end/deflate_end die when they ran out of memory and get
rid of the _gently() kind.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* jc/rename-degrade-cc-to-c:
diffcore-rename: fall back to -C when -C -C busts the rename limit
diffcore-rename: record filepair for rename src
diffcore-rename: refactor "too many candidates" logic
builtin/diff.c: remove duplicated call to diff_result_code()
Refactor the "do not stop feeding the backend early" logic into a small
helper function and use it in both run_diff_files() and diff_tree() that
has the stop-early optimization. We may later add other types of diffcore
transformation that require to look at the whole result like diff-filter
does, and having the logic in a single place is essential for longer term
maintainability.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Often one is interested in the full --stat output only for commits which
change a few files, but not others, because larger restructuring gives a
--stat which fills a few screens.
Introduce a new option --stat-count=<count> which limits the --stat output
to the first <count> lines, followed by a "..." line. It can
also be given as the third parameter in
--stat=<width>,<name-width>,<count>.
Also, the unstuck form is supported analogous to the other two stat
parameters.
Signed-off-by: Michael J Gruber <git@drmicha.warpmail.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Currently, --stat calculates the longest name from all items but then
drops some (mode changes) from the output later on.
Instead, drop them from the namelen generation and calculation.
Signed-off-by: Michael J Gruber <git@drmicha.warpmail.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This function actually does two things:
1. Load the userdiff driver for the filespec.
2. Decide whether the driver has a textconv component, and
initialize the textconv cache if applicable.
Only part (1) requires the filespec object, and some callers
may not have a filespec at all. So let's split them it into
two functions, and put part (2) with the userdiff code,
which is a better fit.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
With diff.suppress-blank-empty=true, "git diff --word-diff" would
output data that had been read from uninitialized heap memory.
The problem was that fn_out_consume did not account for the
possibility of a line with length 1, i.e., the empty context line
that diff.suppress-blank-empty=true converts from " \n" to "\n".
Since it assumed there would always be a prefix character (the space),
it decremented "len" unconditionally, thus passing len=0 to emit_line,
which would then blindly call emit_line_0 with len=-1 which would
pass that value on to fwrite as SIZE_MAX. Boom.
Signed-off-by: Jim Meyering <meyering@redhat.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* jh/dirstat-lines:
Mark dirstat error messages for translation
Improve error handling when parsing dirstat parameters
New --dirstat=lines mode, doing dirstat analysis based on diffstat
Allow specifying --dirstat cut-off percentage as a floating point number
Add config variable for specifying default --dirstat behavior
Refactor --dirstat parsing; deprecate --cumulative and --dirstat-by-file
Make --dirstat=0 output directories that contribute < 0.1% of changes
Add several testcases for --dirstat and friends
* jc/fix-diff-files-unmerged:
diff-files: show unmerged entries correctly
diff: remove often unused parameters from diff_unmerge()
diff.c: return filepair from diff_unmerge()
test: use $_z40 from test-lib
* jc/fix-diff-files-unmerged:
diff-files: show unmerged entries correctly
diff: remove often unused parameters from diff_unmerge()
diff.c: return filepair from diff_unmerge()
test: use $_z40 from test-lib
* jh/dirstat:
--dirstat: In case of renames, use target filename instead of source filename
Teach --dirstat not to completely ignore rearranged lines within a file
--dirstat-by-file: Make it faster and more correct
--dirstat: Describe non-obvious differences relative to --stat or regular diff
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>
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>
Only the first digit after the decimal point is kept, as the dirstat
calculations all happen in permille.
Selftests verifying floating-point percentage input has been added.
Improved-by: Junio C Hamano <gitster@pobox.com>
Improved-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Johan Herland <johan@herland.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The new diff.dirstat config variable takes the same arguments as
'--dirstat=<args>', and specifies the default arguments for --dirstat.
The config is obviously overridden by --dirstat arguments passed on the
command line.
When not specified, the --dirstat defaults are 'changes,noncumulative,3'.
The patch also adds several tests verifying the interaction between the
diff.dirstat config variable, and the --dirstat command line option.
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>
Instead of having multiple interconnected dirstat-related options, teach
the --dirstat option itself to accept all behavior modifiers as parameters.
- Preserve the current --dirstat=<limit> (where <limit> is an integer
specifying a cut-off percentage)
- Add --dirstat=cumulative, replacing --cumulative
- Add --dirstat=files, replacing --dirstat-by-file
- Also add --dirstat=changes and --dirstat=noncumulative for specifying the
current default behavior. These allow the user to reset other --dirstat
parameters (e.g. 'cumulative' and 'files') occuring earlier on the
command line.
The deprecated options (--cumulative and --dirstat-by-file) are still
functional, although they have been removed from the documentation.
Allow multiple parameters to be separated by commas, e.g.:
--dirstat=files,10,cumulative
Update the documentation accordingly, and add testcases verifying the
behavior of the new syntax.
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>
The expected output from --dirstat=0, is to include any directory with
changes, even if those changes contribute a minuscule portion of the total
changes. However, currently, directories that contribute less than 0.1% are
not included, since their 'permille' value is 0, and there is an
'if (permille)' check in gather_dirstat() that causes them to be ignored.
This test is obviously intended to exclude directories that contribute no
changes whatsoever, but in this case, it hits too broadly. The correct
check is against 'this_dir' from which the permille is calculated. Only if
this value is 0 does the directory truly contribute no changes, and should
be skipped from the output.
This patches fixes this issue, and updates corresponding testcases to
expect the new behvaior.
Signed-off-by: Johan Herland <johan@herland.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* jc/rename-degrade-cc-to-c:
diffcore-rename: fall back to -C when -C -C busts the rename limit
diffcore-rename: record filepair for rename src
diffcore-rename: refactor "too many candidates" logic
builtin/diff.c: remove duplicated call to diff_result_code()
* jh/dirstat:
--dirstat: In case of renames, use target filename instead of source filename
Teach --dirstat not to completely ignore rearranged lines within a file
--dirstat-by-file: Make it faster and more correct
--dirstat: Describe non-obvious differences relative to --stat or regular diff
e9c8409 (diff-index --cached --raw: show tree entry on the LHS for
unmerged entries., 2007-01-05) added a <mode, object name> pair as
parameters to this function, to store them in the pre-image side of an
unmerged file pair. Now the function is fixed to return the filepair it
queued, we can make the caller on the special case codepath to do so.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The underlying diff_queue() returns diff_filepair so that the caller can
further add information to it, and the helper function diff_unmerge()
utilizes the feature itself, but does not expose it to its callers, which
was kind of selfish.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This changes --dirstat analysis to count "damage" toward the target filename,
rather than the source filename. For renames within a directory, this won't
matter to the final output, but when moving files between diretories, the
output now lists the target directory rather than the source directory.
Signed-off-by: Johan Herland <johan@herland.net>
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
