When accessing a blob for a diff, we may try to reuse file
contents in the working tree, under the theory that it is
faster to mmap those file contents than it would be to
extract the content from the object database.
When we have to filter those contents, though, that
assumption does not hold. Even for our internal conversions
like CRLF, we have to allocate and fill a new buffer anyway.
But much worse, for external clean filters we have to exec
an arbitrary script, and we have no idea how expensive it
may be to run.
So let's skip this optimization when conversion into git's
"clean" form is required. This applies whenever the
"want_file" flag is false. When it's true, the caller
actually wants the smudged worktree contents, which the
reused file by definition already has (in fact, this is a
key optimization going the other direction, since reusing
the worktree file there lets us skip smudge filters).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
http://lkml.kernel.org/g/20160610075043.GA13411@sigill.intra.peff.net
reports that a change to add a new "function" with common ending
with the existing one at the end of the file is shown like this:
def foo
do_foo_stuff()
+ common_ending()
+end
+
+def bar
+ do_bar_stuff()
+
common_ending()
end
when the new heuristic is in use. In reality, the change is to add
the blank line before "def bar" and everything below, which is what
the code without the new heuristic shows.
Disable the heuristics by default, and resurrect the documentation
for the option and the configuration variables, while clearly
marking the feature as still experimental.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In order to produce the smallest possible diff and combine several diff
hunks together, we implement a heuristic from GNU Diff which moves diff
hunks forward as far as possible when we find common context above and
below a diff hunk. This sometimes produces less readable diffs when
writing C, Shell, or other programming languages, ie:
...
/*
+ *
+ *
+ */
+
+/*
...
instead of the more readable equivalent of
...
+/*
+ *
+ *
+ */
+
/*
...
Implement the following heuristic to (optionally) produce the desired
output.
If there are diff chunks which can be shifted around, shift each hunk
such that the last common empty line is below the chunk with the rest
of the context above.
This heuristic appears to resolve the above example and several other
common issues without producing significantly weird results. However, as
with any heuristic it is not really known whether this will always be
more optimal. Thus, it can be disabled via diff.compactionHeuristic.
Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Rename detection is a very convenient feature, and new users shouldn't
have to dig in the documentation to benefit from it.
Potential objections to activating rename detection are that it
sometimes fail, and it is sometimes slow. But rename detection is
already activated by default in several cases like "git status" and "git
merge", so activating diff.renames does not fundamentally change the
situation. When the rename detection fails, it now fails consistently
between "git diff" and "git status".
This setting does not affect plumbing commands, hence well-written
scripts will not be affected.
Signed-off-by: Matthieu Moy <Matthieu.Moy@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We allocate 100 bytes to hold the "Submodule commit ..."
text. This is enough, but it's not immediately obvious that
this is the case, and we have to repeat the magic 100 twice.
We could get away with xstrfmt here, but we want to know the
size, as well, so let's use a real strbuf. And while we're
here, we can clean up the logic around size_only. It
currently sets and clears the "data" field pointlessly, and
leaves the "should_free" flag on even after we have cleared
the data.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Using FLEX_ARRAY macros reduces the amount of manual
computation size we have to do. It also ensures we don't
overflow size_t, and it makes sure we write the same number
of bytes that we allocated.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The memory allocation scheme for the textconv interface is a
bit tricky, and not well documented. It was originally
designed as an internal part of diff.c (matching
fill_mmfile), but gradually was made public.
Refactoring it is difficult, but we can at least improve the
situation by documenting the intended flow and enforcing it
with an in-code assertion.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
After switching to use the tempfile module in commit 284098f1
(diff: use tempfile module), no declarations from sigchain.h are used in
diff.c anymore. Thus, remove the #include.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Before sha1_to_hex_r() existed, a simple way to get hex
sha1 into a buffer was with:
strcpy(buf, sha1_to_hex(sha1));
This isn't wrong (assuming the buf is 41 characters), but it
makes auditing the code base for bad strcpy() calls harder,
as these become false positives.
Let's convert them to sha1_to_hex_r(), and likewise for
some calls to find_unique_abbrev(). While we're here, we'll
double-check that all of the buffers are correctly sized,
and use the more obvious GIT_SHA1_HEXSZ constant.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When we call into xdiff to perform a diff, we generally lose
the return code completely. Typically by ignoring the return
of our xdi_diff wrapper, but sometimes we even propagate
that return value up and then ignore it later. This can
lead to us silently producing incorrect diffs (e.g., "git
log" might produce no output at all, not even a diff header,
for a content-level diff).
In practice this does not happen very often, because the
typical reason for xdiff to report failure is that it
malloc() failed (it uses straight malloc, and not our
xmalloc wrapper). But it could also happen when xdiff
triggers one our callbacks, which returns an error (e.g.,
outf() in builtin/rerere.c tries to report a write failure
in this way). And the next patch also plans to add more
failure modes.
Let's notice an error return from xdiff and react
appropriately. In most of the diff.c code, we can simply
die(), which matches the surrounding code (e.g., that is
what we do if we fail to load a file for diffing in the
first place). This is not that elegant, but we are probably
better off dying to let the user know there was a problem,
rather than simply generating bogus output.
We could also just die() directly in xdi_diff, but the
callers typically have a bit more context, and can provide a
better message (and if we do later decide to pass errors up,
we're one step closer to doing so).
There is one interesting case, which is in diff_grep(). Here
if we cannot generate the diff, there is nothing to match,
and we silently return "no hits". This is actually what the
existing code does already, but we make it a little more
explicit.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We sometimes sprintf into fixed-size buffers when we know
that the buffer is large enough to fit the input (either
because it's a constant, or because it's numeric input that
is bounded in size). Likewise with strcpy of constant
strings.
However, these sites make it hard to audit sprintf and
strcpy calls for buffer overflows, as a reader has to
cross-reference the size of the array with the input. Let's
use xsnprintf instead, which communicates to a reader that
we don't expect this to overflow (and catches the mistake in
case we do).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We remove the extracted functions and directly parse into and read out
of the cache. This allows us to have one unified way of accessing
submodule configuration values specific to single submodules. Regardless
whether we need to access a configuration from history or from the
worktree.
Signed-off-by: Heiko Voigt <hvoigt@hvoigt.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Also add some code comments explaining how the fields in "struct
diff_tempfile" are used.
Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Check if a matched token is followed by a delimiter before advancing the
pointer arg. This avoids accepting composite words like "allnew" or
"defaultcontext" and misparsing them as "new" or "context".
Signed-off-by: Rene Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
People who work on projects with mostly linear history with frequent
whole file renames may want to always use "git log --follow" when
inspecting the life of the content that live in a single path.
Teach the command to behave as if "--follow" was given from the
command line when log.follow configuration variable is set *and*
there is one (and only one) path on the command line.
Signed-off-by: David Turner <dturner@twopensource.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The latter is a much more descriptive name (and we support
"color.diff.context" now). This also updates the name of any
local variables which were used to store the color.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The term "plain" is a bit ambiguous; let's allow the more
specific "context", but keep "plain" around for
compatibility.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Traditionally, we only cared about whitespace breakages introduced
in new lines. Some people want to paint whitespace breakages on old
lines, too. When they see a whitespace breakage on a new line, they
can spot the same kind of whitespace breakage on the corresponding
old line and want to say "Ah, those breakages are there but they
were inherited from the original, so let's not touch them for now."
Introduce `--ws-error-highlight=<kind>` option, that lets them pass
a comma separated list of `old`, `new`, and `context` to specify
what lines to highlight whitespace errors on.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Traditionally, we only had emit_add_line() helper, which knows how
to find and paint whitespace breakages on the given line, because we
only care about whitespace breakages introduced in new lines. The
context lines and old (i.e. deleted) lines are emitted with a
simpler emit_line_0() that paints the entire line in plain or old
colors.
Identify callers of emit_line_0() that show deleted lines and
context lines, have them call new helpers, emit_del_line() and
emit_context_line(), so that we can later tweak what is done to
these two classes of lines.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Clear the git_zstream variable at the start of git_deflate_init() etc.
so that callers don't have to do that.
Signed-off-by: Rene Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When --shortstat is used in conjunction with --dirstat=changes, git diff will
output the dirstat information twice: first as calculated by the 'lines'
algorithm, then as calculated by the 'changes' algorithm:
$ git diff --dirstat=changes,10 --shortstat v2.2.0..v2.2.1
23 files changed, 453 insertions(+), 54 deletions(-)
33.5% Documentation/RelNotes/
26.2% t/
46.6% Documentation/RelNotes/
16.6% t/
The same duplication happens for --shortstat together with --dirstat=files, but
not for --shortstat together with --dirstat=lines.
Limit output to only include one dirstat part, calculated as specified
by the --dirstat parameter. Also, add test for this.
Signed-off-by: Mårten Kongstad <marten.kongstad@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Originally the color-parsing function was used only for
config variables. It made sense to pass the variable name so
that the die() message could be something like:
$ git -c color.branch.plain=bogus branch
fatal: bad color value 'bogus' for variable 'color.branch.plain'
These days we call it in other contexts, and the resulting
error messages are a little confusing:
$ git log --pretty='%C(bogus)'
fatal: bad color value 'bogus' for variable '--pretty format'
$ git config --get-color foo.bar bogus
fatal: bad color value 'bogus' for variable 'command line'
This patch teaches color_parse to complain only about the
value, and then return an error code. Config callers can
then propagate that up to the config parser, which mentions
the variable name. Other callers can provide a custom
message. After this patch these three cases now look like:
$ git -c color.branch.plain=bogus branch
error: invalid color value: bogus
fatal: unable to parse 'color.branch.plain' from command-line config
$ git log --pretty='%C(bogus)'
error: invalid color value: bogus
fatal: unable to parse --pretty format
$ git config --get-color foo.bar bogus
error: invalid color value: bogus
fatal: unable to parse default color value
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Most struct child_process variables are cleared using memset first after
declaration. Provide a macro, CHILD_PROCESS_INIT, that can be used to
initialize them statically instead. That's shorter, doesn't require a
function call and is slightly more readable (especially given that we
already have STRBUF_INIT, ARGV_ARRAY_INIT etc.).
Helped-by: Johannes Sixt <j6t@kdbg.org>
Signed-off-by: Rene Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
If we are given two SHA-1 and asked to determine if they are different
(but not _what_ differences), we know right away by comparing SHA-1.
A side effect of this patch is, because large files are marked binary,
diff-tree will not need to unpack them. 'diff-index --cached' will not
either. But 'diff-files' still does.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Too large files may lead to failure to allocate memory. If it happens
here, it could impact quite a few commands that involve
diff. Moreover, too large files are inefficient to compare anyway (and
most likely non-text), so mark them binary and skip looking at their
content.
Noticed-by: Dale R. Worley <worley@alum.mit.edu>
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Avoid code duplication and let strbuf_addstr() call strlen() for us.
Signed-off-by: Rene Scharfe <l.s.r@web.de>
Reviewed-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
As in earlier commits, the diff option parser uses
starts_with to find that an argument starts with "--stat-",
and then adds strlen("stat-") to find the rest of the
option.
However, in this case the starts_with and the strlen are
separated across functions, making it easy to call the
latter without the former. Let's use skip_prefix instead of
raw pointer arithmetic to catch such a case.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
It's a common idiom to match a prefix and then skip past it
with strlen, like:
if (starts_with(foo, "bar"))
foo += strlen("bar");
This avoids magic numbers, but means we have to repeat the
string (and there is no compiler check that we didn't make a
typo in one of the strings).
We can use skip_prefix to handle this case without repeating
ourselves.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
It's a common idiom to match a prefix and then skip past it
with a magic number, like:
if (starts_with(foo, "bar"))
foo += 3;
This is easy to get wrong, since you have to count the
prefix string yourself, and there's no compiler check if the
string changes. We can use skip_prefix to avoid the magic
numbers here.
Note that some of these conversions could be much shorter.
For example:
if (starts_with(arg, "--foo=")) {
bar = arg + 6;
continue;
}
could become:
if (skip_prefix(arg, "--foo=", &bar))
continue;
However, I have left it as:
if (skip_prefix(arg, "--foo=", &v)) {
bar = v;
continue;
}
to visually match nearby cases which need to actually
process the string. Like:
if (skip_prefix(arg, "--foo=", &v)) {
bar = atoi(v);
continue;
}
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This function originally took a whole config variable name
("var") and an offset ("ofs"). It checked "var+ofs" against
each color slot, but reported errors using the whole "var".
However, since 8b8e862 (ignore unknown color configuration,
2009-12-12), it returns -1 rather than printing its own
error, and therefore only cares about var+ofs. We can drop
the ofs parameter and teach its sole caller to derive the
pointer itself.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
xcalloc() takes two arguments: the number of elements and their size.
diffstat_add() passes the arguments in reverse order, passing the
size of a diffstat_file*, followed by the number of diffstat_file* to
be allocated.
Rearrange them so they are in the correct order.
Signed-off-by: Brian Gesiak <modocache@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Because of the way "--follow" is implemented, we must have
exactly one pathspec. "git log" enforces this restriction,
but other users of the revision traversal code do not. For
example, "git format-patch --follow" will segfault during
try_to_follow_renames, as we have no pathspecs at all.
We can push this check down into diff_setup_done, which is
probably a better place anyway. It is the diff code that
introduces this restriction, so other parts of the code
should not need to care themselves.
Reported-by: "Michael S. Tsirkin" <mst@redhat.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The current logic makes it hard to see what gets put onto
the command line in which cases. Pulling out a helper
function lets us see that we have two sets of file data, and
the second set either uses the original name, or the "other"
renamed/copy name.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Whether we have diff_filespecs to give to the diff command
or not, we always are going to run the program and pass it
the pathname. Let's pull that duplicated part out of the
conditional to make it more obvious.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This fflush was added in d5535ec (Use run_command() to spawn
external diff programs instead of fork/exec., 2007-10-19),
because flushing buffers before forking is a good habit.
But later, 7d0b18a (Add output flushing before fork(),
2008-08-04) added it to the generic run-command interface,
meaning that our flush here is redundant.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When the external diff reports an error, we try to clean up
and die. However, we can make this process a bit simpler:
1. We do not need to bother freeing memory, since we are
about to exit. Nor do we need to clean up our
tempfiles, since the atexit() handler will do it for
us. So we can die as soon as we see the error.
3. We can just call die() rather than fprintf/exit. This
does technically change our exit code, but the exit
code of "1" is not meaningful here. In fact, it is
probably wrong, since "1" from diff usually means
"completed successfully, but there were differences".
And while we're there, we can mark the error message for
translation, and drop the full stop at the end to make it
more like our other messages.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We currently use static buffers and a static array for
formatting the environment passed to the external diff.
There's nothing wrong in the code, but it is much easier to
verify that it is correct if we use a dynamic argv_array.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We currently generate the command-line for the external
command using a fixed-length array of size 10. But if there
is a rename, we actually need 11 elements (10 items, plus a
NULL), and end up writing a random NULL onto the stack.
Rather than bump the limit, let's just use an argv_array, which
makes this sort of error impossible.
Noticed-by: Max L <infthi.inbox@gmail.com>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since we do not translate diffstat any more, remove the obsolete comments.
Signed-off-by: Jiang Xin <worldhello.net@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
As was recently shown in "combine-diff: optimize
combine_diff_path sets intersection", combine-diff runs very slowly. In
that commit we optimized paths sets intersection, but that accounted
only for ~ 25% of the slowness, and as my tracing showed, for linux.git
v3.10..v3.11, for merges a lot of time is spent computing
diff(commit,commit^2) just to only then intersect that huge diff to
almost small set of files from diff(commit,commit^1).
In previous commit, we described the problem in more details, and
reworked the diff tree-walker to be general one - i.e. to work in
multiple parent case too. Now is the time to take advantage of it for
finding paths for combine diff.
The implementation is straightforward - if we know, we can get generated
diff paths directly, and at present that means no diff filtering or
rename/copy detection was requested(*), we can call multiparent tree-walker
directly and get ready paths.
(*) because e.g. at present, all diffcore transformations work on
diff_filepair queues, but in the future, that limitation can be
lifted, if filters would operate directly on combine_diff_paths.
Timings for `git log --raw --no-abbrev --no-renames` without `-c` ("git log")
and with `-c` ("git log -c") and with `-c --merges` ("git log -c --merges")
before and after the patch are as follows:
linux.git v3.10..v3.11
log log -c log -c --merges
before 1.9s 16.4s 15.2s
after 1.9s 2.4s 1.1s
The result stayed the same.
Signed-off-by: Kirill Smelkov <kirr@mns.spb.ru>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Previously diff_tree(), which is now named ll_diff_tree_sha1(), was
generating diff_filepair(s) for two trees t1 and t2, and that was
usually used for a commit as t1=HEAD~, and t2=HEAD - i.e. to see changes
a commit introduces.
In Git, however, we have fundamentally built flexibility in that a
commit can have many parents - 1 for a plain commit, 2 for a simple merge,
but also more than 2 for merging several heads at once.
For merges there is a so called combine-diff, which shows diff, a merge
introduces by itself, omitting changes done by any parent. That works
through first finding paths, that are different to all parents, and then
showing generalized diff, with separate columns for +/- for each parent.
The code lives in combine-diff.c .
There is an impedance mismatch, however, in that a commit could
generally have any number of parents, and that while diffing trees, we
divide cases for 2-tree diffs and more-than-2-tree diffs. I mean there
is no special casing for multiple parents commits in e.g.
revision-walker .
That impedance mismatch *hurts* *performance* *badly* for generating
combined diffs - in "combine-diff: optimize combine_diff_path
sets intersection" I've already removed some slowness from it, but from
the timings provided there, it could be seen, that combined diffs still
cost more than an order of magnitude more cpu time, compared to diff for
usual commits, and that would only be an optimistic estimate, if we take
into account that for e.g. linux.git there is only one merge for several
dozens of plain commits.
That slowness comes from the fact that currently, while generating
combined diff, a lot of time is spent computing diff(commit,commit^2)
just to only then intersect that huge diff to almost small set of files
from diff(commit,commit^1).
That's because at present, to compute combine-diff, for first finding
paths, that "every parent touches", we use the following combine-diff
property/definition:
D(A,P1...Pn) = D(A,P1) ^ ... ^ D(A,Pn) (w.r.t. paths)
where
D(A,P1...Pn) is combined diff between commit A, and parents Pi
and
D(A,Pi) is usual two-tree diff Pi..A
So if any of that D(A,Pi) is huge, tracting 1 n-parent combine-diff as n
1-parent diffs and intersecting results will be slow.
And usually, for linux.git and other topic-based workflows, that
D(A,P2) is huge, because, if merge-base of A and P2, is several dozens
of merges (from A, via first parent) below, that D(A,P2) will be diffing
sum of merges from several subsystems to 1 subsystem.
The solution is to avoid computing n 1-parent diffs, and to find
changed-to-all-parents paths via scanning A's and all Pi's trees
simultaneously, at each step comparing their entries, and based on that
comparison, populate paths result, and deduce we could *skip*
*recursing* into subdirectories, if at least for 1 parent, sha1 of that
dir tree is the same as in A. That would save us from doing significant
amount of needless work.
Such approach is very similar to what diff_tree() does, only there we
deal with scanning only 2 trees simultaneously, and for n+1 tree, the
logic is a bit more complex:
D(T,P1...Pn) calculation scheme
-------------------------------
D(T,P1...Pn) = D(T,P1) ^ ... ^ D(T,Pn) (regarding resulting paths set)
D(T,Pj) - diff between T..Pj
D(T,P1...Pn) - combined diff from T to parents P1,...,Pn
We start from all trees, which are sorted, and compare their entries in
lock-step:
T P1 Pn
- - -
|t| |p1| |pn|
|-| |--| ... |--| imin = argmin(p1...pn)
| | | | | |
|-| |--| |--|
|.| |. | |. |
. . .
. . .
at any time there could be 3 cases:
1) t < p[imin];
2) t > p[imin];
3) t = p[imin].
Schematic deduction of what every case means, and what to do, follows:
1) t < p[imin] -> ∀j t ∉ Pj -> "+t" ∈ D(T,Pj) -> D += "+t"; t↓
2) t > p[imin]
2.1) ∃j: pj > p[imin] -> "-p[imin]" ∉ D(T,Pj) -> D += ø; ∀ pi=p[imin] pi↓
2.2) ∀i pi = p[imin] -> pi ∉ T -> "-pi" ∈ D(T,Pi) -> D += "-p[imin]"; ∀i pi↓
3) t = p[imin]
3.1) ∃j: pj > p[imin] -> "+t" ∈ D(T,Pj) -> only pi=p[imin] remains to investigate
3.2) pi = p[imin] -> investigate δ(t,pi)
|
|
v
3.1+3.2) looking at δ(t,pi) ∀i: pi=p[imin] - if all != ø ->
⎧δ(t,pi) - if pi=p[imin]
-> D += ⎨
⎩"+t" - if pi>p[imin]
in any case t↓ ∀ pi=p[imin] pi↓
~
For comparison, here is how diff_tree() works:
D(A,B) calculation scheme
-------------------------
A B
- -
|a| |b| a < b -> a ∉ B -> D(A,B) += +a a↓
|-| |-| a > b -> b ∉ A -> D(A,B) += -b b↓
| | | | a = b -> investigate δ(a,b) a↓ b↓
|-| |-|
|.| |.|
. .
. .
~~~~~~~~
This patch generalizes diff tree-walker to work with arbitrary number of
parents as described above - i.e. now there is a resulting tree t, and
some parents trees tp[i] i=[0..nparent). The generalization builds on
the fact that usual diff
D(A,B)
is by definition the same as combined diff
D(A,[B]),
so if we could rework the code for common case and make it be not slower
for nparent=1 case, usual diff(t1,t2) generation will not be slower, and
multiparent diff tree-walker would greatly benefit generating
combine-diff.
What we do is as follows:
1) diff tree-walker ll_diff_tree_sha1() is internally reworked to be
a paths generator (new name diff_tree_paths()), with each generated path
being `struct combine_diff_path` with info for path, new sha1,mode and for
every parent which sha1,mode it was in it.
2) From that info, we can still generate usual diff queue with
struct diff_filepairs, via "exporting" generated
combine_diff_path, if we know we run for nparent=1 case.
(see emit_diff() which is now named emit_diff_first_parent_only())
3) In order for diff_can_quit_early(), which checks
DIFF_OPT_TST(opt, HAS_CHANGES))
to work, that exporting have to be happening not in bulk, but
incrementally, one diff path at a time.
For such consumers, there is a new callback in diff_options
introduced:
->pathchange(opt, struct combine_diff_path *)
which, if set to !NULL, is called for every generated path.
(see new compat ll_diff_tree_sha1() wrapper around new paths
generator for setup)
4) The paths generation itself, is reworked from previous
ll_diff_tree_sha1() code according to "D(A,P1...Pn) calculation
scheme" provided above:
On the start we allocate [nparent] arrays in place what was
earlier just for one parent tree.
then we just generalize loops, and comparison according to the
algorithm.
Some notes(*):
1) alloca(), for small arrays, is used for "runs not slower for
nparent=1 case than before" goal - if we change it to xmalloc()/free()
the timings get ~1% worse. For alloca() we use just-introduced
xalloca/xalloca_free compatibility wrappers, so it should not be a
portability problem.
2) For every parent tree, we need to keep a tag, whether entry from that
parent equals to entry from minimal parent. For performance reasons I'm
keeping that tag in entry's mode field in unused bit - see S_IFXMIN_NEQ.
Not doing so, we'd need to alloca another [nparent] array, which hurts
performance.
3) For emitted paths, memory could be reused, if we know the path was
processed via callback and will not be needed later. We use efficient
hand-made realloc-style path_appendnew(), that saves us from ~1-1.5%
of potential additional slowdown.
4) goto(s) are used in several places, as the code executes a little bit
faster with lowered register pressure.
Also
- we should now check for FIND_COPIES_HARDER not only when two entries
names are the same, and their hashes are equal, but also for a case,
when a path was removed from some of all parents having it.
The reason is, if we don't, that path won't be emitted at all (see
"a > xi" case), and we'll just skip it, and FIND_COPIES_HARDER wants
all paths - with diff or without - to be emitted, to be later analyzed
for being copies sources.
The new check is only necessary for nparent >1, as for nparent=1 case
xmin_eqtotal always =1 =nparent, and a path is always added to diff as
removal.
~~~~~~~~
Timings for
# without -c, i.e. testing only nparent=1 case
`git log --raw --no-abbrev --no-renames`
before and after the patch are as follows:
navy.git linux.git v3.10..v3.11
before 0.611s 1.889s
after 0.619s 1.907s
slowdown 1.3% 0.9%
This timings show we did no harm to usual diff(tree1,tree2) generation.
From the table we can see that we actually did ~1% slowdown, but I think
I've "earned" that 1% in the previous patch ("tree-diff: reuse base
str(buf) memory on sub-tree recursion", HEAD~~) so for nparent=1 case,
net timings stays approximately the same.
The output also stayed the same.
(*) If we revert 1)-4) to more usual techniques, for nparent=1 case,
we'll get ~2-2.5% of additional slowdown, which I've tried to avoid, as
"do no harm for nparent=1 case" rule.
For linux.git, combined diff will run an order of magnitude faster and
appropriate timings will be provided in the next commit, as we'll be
taking advantage of the new diff tree-walker for combined-diff
generation there.
P.S. and combined diff is not some exotic/for-play-only stuff - for
example for a program I write to represent Git archives as readonly
filesystem, there is initial scan with
`git log --reverse --raw --no-abbrev --no-renames -c`
to extract log of what was created/changed when, as a result building a
map
{} sha1 -> in which commit (and date) a content was added
that `-c` means also show combined diff for merges, and without them, if
a merge is non-trivial (merges changes from two parents with both having
separate changes to a file), or an evil one, the map will not be full,
i.e. some valid sha1 would be absent from it.
That case was my initial motivation for combined diffs speedup.
Signed-off-by: Kirill Smelkov <kirr@mns.spb.ru>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Avoid scanning strings twice, once with strchr() and then with
strlen(), by using strchrnul().
Helped-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Rohit Mani <rohit.mani@outlook.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Use ALLOC_GROW() instead of open-coding it in diffstat_add() and
diff_q().
Signed-off-by: Dmitry S. Dolzhenko <dmitrys.dolzhenko@yandex.ru>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When QUICK is set (i.e. with --quiet) we try to do as little work as
possible, stopping after seeing the first change. stat-dirty is
considered a "change" but it may turn out not, if no actual content is
changed. The actual content test is performed too late in the process
and the shortcut may be taken prematurely, leading to incorrect return
code.
Assume we do "git diff --quiet". If we have a stat-dirty file "a" and
a really dirty file "b". We break the loop in run_diff_files() and
stop after "a" because we have got a "change". Later in
diffcore_skip_stat_unmatch() we find out "a" is actually not
changed. But there's nothing else in the diff queue, we incorrectly
declare "no change", ignoring the fact that "b" is changed.
This also happens to "git diff --quiet HEAD" when it hits
diff_can_quit_early() in oneway_diff().
This patch does the content test earlier in order to keep going if "a"
is unchanged. The test result is cached so that when
diffcore_skip_stat_unmatch() is done in the end, we spend no cycles on
re-testing "a".
Reported-by: IWAMOTO Toshihiro <iwamoto@valinux.co.jp>
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The GIT_EXTERNAL_DIFF calling code attempts to reuse existing worktree
files for the worktree side of diffs, for performance reasons.
However, that code also tries to do the same with submodules. This
results in calls to $GIT_EXTERNAL_DIFF where the old-file is a file of
the form "Submodule commit $sha1", but the new-file is a directory in
the worktree.
Fix it by never reusing a worktree "file" in the submodule case.
Reported-by: Grégory Pakosz <gregory.pakosz@gmail.com>
Signed-off-by: Thomas Rast <tr@thomasrast.ch>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The only mention of this field in the code is by some
debugging code which prints it out (and it will always be
zero, since we never touch it otherwise). It was obsoleted
very early on by 25d5ea4 ([PATCH] Redo rename/copy detection
logic., 2005-05-24).
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Jeff King