git
6 years agodiff: respect MIN_BLOCK_LENGTH for last block
Jonathan Tan [Wed, 16 Aug 2017 01:27:38 +0000 (18:27 -0700)] 
diff: respect MIN_BLOCK_LENGTH for last block

Currently, MIN_BLOCK_LENGTH is only checked when diff encounters a line
that does not belong to the current block. In particular, this means
that MIN_BLOCK_LENGTH is not checked after all lines are encountered.

Perform that check.

Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodiff: avoid redundantly clearing a flag
Jonathan Tan [Fri, 11 Aug 2017 22:49:14 +0000 (15:49 -0700)] 
diff: avoid redundantly clearing a flag

No code in diff.c sets DIFF_SYMBOL_MOVED_LINE except in
mark_color_as_moved(), so it is redundant to clear it for the current
line. Therefore, clear it only for previous lines.

This makes a refactoring in a subsequent patch easier.

Signed-off-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodiff: document the new --color-moved setting
Stefan Beller [Fri, 30 Jun 2017 20:53:10 +0000 (13:53 -0700)] 
diff: document the new --color-moved setting

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodiff.c: add dimming to moved line detection
Stefan Beller [Fri, 30 Jun 2017 20:53:09 +0000 (13:53 -0700)] 
diff.c: add dimming to moved line detection

Any lines inside a moved block of code are not interesting. Boundaries
of blocks are only interesting if they are next to another block of moved
code.

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodiff.c: color moved lines differently, plain mode
Stefan Beller [Fri, 30 Jun 2017 20:53:08 +0000 (13:53 -0700)] 
diff.c: color moved lines differently, plain mode

Add the 'plain' mode for move detection of code. This omits the checking
for adjacent blocks, so it is not as useful. If you have a lot of the
same blocks moved in the same patch, the 'Zebra' would end up slow as it
is O(n^2) (n is number of same blocks). So this may be useful there and
is generally easy to add. Instead be very literal at the move detection,
do not skip over short blocks here.

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodiff.c: color moved lines differently
Stefan Beller [Fri, 30 Jun 2017 20:53:07 +0000 (13:53 -0700)] 
diff.c: color moved lines differently

When a patch consists mostly of moving blocks of code around, it can
be quite tedious to ensure that the blocks are moved verbatim, and not
undesirably modified in the move. To that end, color blocks that are
moved within the same patch differently. For example (OM, del, add,
and NM are different colors):

    [OM]  -void sensitive_stuff(void)
    [OM]  -{
    [OM]  -        if (!is_authorized_user())
    [OM]  -                die("unauthorized");
    [OM]  -        sensitive_stuff(spanning,
    [OM]  -                        multiple,
    [OM]  -                        lines);
    [OM]  -}

           void another_function()
           {
    [del] -        printf("foo");
    [add] +        printf("bar");
           }

    [NM]  +void sensitive_stuff(void)
    [NM]  +{
    [NM]  +        if (!is_authorized_user())
    [NM]  +                die("unauthorized");
    [NM]  +        sensitive_stuff(spanning,
    [NM]  +                        multiple,
    [NM]  +                        lines);
    [NM]  +}

However adjacent blocks may be problematic. For example, in this
potentially malicious patch, the swapping of blocks can be spotted:

    [OM]  -void sensitive_stuff(void)
    [OM]  -{
    [OMA] -        if (!is_authorized_user())
    [OMA] -                die("unauthorized");
    [OM]  -        sensitive_stuff(spanning,
    [OM]  -                        multiple,
    [OM]  -                        lines);
    [OMA] -}

           void another_function()
           {
    [del] -        printf("foo");
    [add] +        printf("bar");
           }

    [NM]  +void sensitive_stuff(void)
    [NM]  +{
    [NMA] +        sensitive_stuff(spanning,
    [NMA] +                        multiple,
    [NMA] +                        lines);
    [NM]  +        if (!is_authorized_user())
    [NM]  +                die("unauthorized");
    [NMA] +}

If the moved code is larger, it is easier to hide some permutation in the
code, which is why some alternative coloring is needed.

This patch implements the first mode:
* basic alternating 'Zebra' mode
  This conveys all information needed to the user.  Defer customization to
  later patches.

First I implemented an alternative design, which would try to fingerprint
a line by its neighbors to detect if we are in a block or at the boundary.
This idea iss error prone as it inspected each line and its neighboring
lines to determine if the line was (a) moved and (b) if was deep inside
a hunk by having matching neighboring lines. This is unreliable as the
we can construct hunks which have equal neighbors that just exceed the
number of lines inspected. (Think of 'AXYZBXYZCXYZD..' with each letter
as a line, that is permutated to AXYZCXYZBXYZD..').

Instead this provides a dynamic programming greedy algorithm that finds
the largest moved hunk and then has several modes on highlighting bounds.

A note on the options '--submodule=diff' and '--color-words/--word-diff':
In the conversion to use emit_line in the prior patches both submodules
as well as word diff output carefully chose to call emit_line with sign=0.
All output with sign=0 is ignored for move detection purposes in this
patch, such that no weird looking output will be generated for these
cases. This leads to another thought: We could pass on '--color-moved' to
submodules such that they color up moved lines for themselves. If we'd do
so only line moves within a repository boundary are marked up.

It is useful to have moved lines colored, but there are annoying corner
cases, such as a single line moved, that is very common. For example
in a typical patch of C code, we have closing braces that end statement
blocks or functions.

While it is technically true that these lines are moved as they show up
elsewhere, it is harmful for the review as the reviewers attention is
drawn to such a minor side annoyance.

For now let's have a simple solution of hardcoding the number of
moved lines to be at least 3 before coloring them. Note, that the
length is applied across all blocks to find the 'lonely' blocks
that pollute new code, but do not interfere with a permutated
block where each permutation has less lines than 3.

Helped-by: Jonathan Tan <jonathantanmy@google.com>
Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodiff.c: buffer all output if asked to
Stefan Beller [Fri, 30 Jun 2017 00:07:06 +0000 (17:07 -0700)] 
diff.c: buffer all output if asked to

Introduce a new option 'emitted_symbols' in the struct diff_options which
controls whether all output is buffered up until all output is available.
It is set internally in diff.c when necessary.

We'll have a new struct 'emitted_string' in diff.c which will be used to
buffer each line.  The emitted_string will duplicate the memory of the
line to buffer as that is easiest to reason about for now. In a future
patch we may want to decrease the memory usage by not duplicating all
output for buffering but rather we may want to store offsets into the
file or in case of hunk descriptions such as the similarity score, we
could just store the relevant number and reproduce the text later on.

This approach was chosen as a first step because it is quite simple
compared to the alternative with less memory footprint.

emit_diff_symbol factors out the emission part and depending on the
diff_options->emitted_symbols the emission will be performed directly
when calling emit_diff_symbol or after the whole process is done, i.e.
by buffering we have add the possibility for a second pass over the
whole output before doing the actual output.

In 6440d34 (2012-03-14, diff: tweak a _copy_ of diff_options with
word-diff) we introduced a duplicate diff options struct for word
emissions as we may have different regex settings in there.
When buffering the output, we need to operate on just one buffer,
so we have to copy back the emissions of the word buffer into the
main buffer.

Unconditionally enable output via buffer in this patch as it yields
a great opportunity for testing, i.e. all the diff tests from the
test suite pass without having reordering issues (i.e. only parts
of the output got buffered, and we forgot to buffer other parts).
The test suite passes, which gives confidence that we converted all
functions to use emit_string for output.

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodiff.c: emit_diff_symbol learns about DIFF_SYMBOL_SUMMARY
Stefan Beller [Fri, 30 Jun 2017 00:07:05 +0000 (17:07 -0700)] 
diff.c: emit_diff_symbol learns about DIFF_SYMBOL_SUMMARY

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodiff.c: emit_diff_symbol learns about DIFF_SYMBOL_STAT_SEP
Stefan Beller [Fri, 30 Jun 2017 00:07:04 +0000 (17:07 -0700)] 
diff.c: emit_diff_symbol learns about DIFF_SYMBOL_STAT_SEP

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodiff.c: convert word diffing to use emit_diff_symbol
Stefan Beller [Fri, 30 Jun 2017 00:07:03 +0000 (17:07 -0700)] 
diff.c: convert word diffing to use emit_diff_symbol

The word diffing is not line oriented and would need some serious
effort to be transformed into a line oriented approach, so
just go with a symbol DIFF_SYMBOL_WORD_DIFF that is a partial line.

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodiff.c: convert show_stats to use emit_diff_symbol
Stefan Beller [Fri, 30 Jun 2017 00:07:02 +0000 (17:07 -0700)] 
diff.c: convert show_stats to use emit_diff_symbol

We call print_stat_summary from builtin/apply, so we still
need the version with a file pointer, so introduce
print_stat_summary_0 that uses emit_string machinery and
keep print_stat_summary with the same arguments around.

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodiff.c: convert emit_binary_diff_body to use emit_diff_symbol
Stefan Beller [Fri, 30 Jun 2017 00:07:01 +0000 (17:07 -0700)] 
diff.c: convert emit_binary_diff_body to use emit_diff_symbol

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agosubmodule.c: migrate diff output to use emit_diff_symbol
Stefan Beller [Fri, 30 Jun 2017 00:07:00 +0000 (17:07 -0700)] 
submodule.c: migrate diff output to use emit_diff_symbol

As the submodule process is no longer attached to the same file pointer
'o->file' as the superprojects process, there is a different result in
color.c::check_auto_color. That is why we need to pass coloring explicitly,
such that the submodule coloring decision will be made by the child process
processing the submodule. Only DIFF_SYMBOL_SUBMODULE_PIPETHROUGH contains
color, the other symbols are for embedding the submodule output into the
superprojects output.

Remove the colors from the function signatures, as all the coloring
decisions will be made either inside the child process or the final
emit_diff_symbol, but not in the functions driving the submodule diff.

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodiff.c: emit_diff_symbol learns DIFF_SYMBOL_REWRITE_DIFF
Stefan Beller [Fri, 30 Jun 2017 00:06:59 +0000 (17:06 -0700)] 
diff.c: emit_diff_symbol learns DIFF_SYMBOL_REWRITE_DIFF

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodiff.c: emit_diff_symbol learns about DIFF_SYMBOL_BINARY_FILES
Stefan Beller [Fri, 30 Jun 2017 00:06:58 +0000 (17:06 -0700)] 
diff.c: emit_diff_symbol learns about DIFF_SYMBOL_BINARY_FILES

we could save a little bit of memory when buffering in a later mode
by just passing the inner part ("%s and %s", file1, file 2), but
those a just a few bytes, so instead let's reuse the implementation from
DIFF_SYMBOL_HEADER and keep the whole line around.

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodiff.c: emit_diff_symbol learns DIFF_SYMBOL_HEADER
Stefan Beller [Fri, 30 Jun 2017 00:06:57 +0000 (17:06 -0700)] 
diff.c: emit_diff_symbol learns DIFF_SYMBOL_HEADER

The header is constructed lazily including line breaks, so just emit
the raw string as is.

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodiff.c: emit_diff_symbol learns DIFF_SYMBOL_FILEPAIR_{PLUS, MINUS}
Stefan Beller [Fri, 30 Jun 2017 00:06:56 +0000 (17:06 -0700)] 
diff.c: emit_diff_symbol learns DIFF_SYMBOL_FILEPAIR_{PLUS, MINUS}

We have to use fprintf instead of emit_line, because we want to emit the
tab after the color. This is important for ancient versions of gnu patch
AFAICT, although we probably do not want to feed colored output to the
patch utility, such that it would not matter if the trailing tab is
colored. Keep the corner case as-is though.

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodiff.c: emit_diff_symbol learns DIFF_SYMBOL_CONTEXT_INCOMPLETE
Stefan Beller [Fri, 30 Jun 2017 00:06:55 +0000 (17:06 -0700)] 
diff.c: emit_diff_symbol learns DIFF_SYMBOL_CONTEXT_INCOMPLETE

The context marker use the exact same output pattern, so reuse it.

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodiff.c: emit_diff_symbol learns DIFF_SYMBOL_WORDS[_PORCELAIN]
Stefan Beller [Fri, 30 Jun 2017 00:06:54 +0000 (17:06 -0700)] 
diff.c: emit_diff_symbol learns DIFF_SYMBOL_WORDS[_PORCELAIN]

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodiff.c: migrate emit_line_checked to use emit_diff_symbol
Stefan Beller [Fri, 30 Jun 2017 00:06:53 +0000 (17:06 -0700)] 
diff.c: migrate emit_line_checked to use emit_diff_symbol

Add a new flags field to emit_diff_symbol, that will be used by
context lines for:
* white space rules that are applicable (The first 12 bits)
  Take a note in cahe.c as well, when this ws rules are extended we have
  to fix the bits in the flags field.
* how the rules are evaluated (actually this double encodes the sign
  of the line, but the code is easier to keep this way, bits 13,14,15)
* if the line a blank line at EOF (bit 16)

The check if new lines need to be marked up as extra lines at the end of
file, is now done unconditionally. That should be ok, as
'new_blank_line_at_eof' has a quick early return.

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodiff.c: emit_diff_symbol learns DIFF_SYMBOL_NO_LF_EOF
Stefan Beller [Fri, 30 Jun 2017 00:06:52 +0000 (17:06 -0700)] 
diff.c: emit_diff_symbol learns DIFF_SYMBOL_NO_LF_EOF

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodiff.c: emit_diff_symbol learns DIFF_SYMBOL_CONTEXT_FRAGINFO
Stefan Beller [Fri, 30 Jun 2017 00:06:51 +0000 (17:06 -0700)] 
diff.c: emit_diff_symbol learns DIFF_SYMBOL_CONTEXT_FRAGINFO

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodiff.c: emit_diff_symbol learns DIFF_SYMBOL_CONTEXT_MARKER
Stefan Beller [Fri, 30 Jun 2017 00:06:50 +0000 (17:06 -0700)] 
diff.c: emit_diff_symbol learns DIFF_SYMBOL_CONTEXT_MARKER

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodiff.c: introduce emit_diff_symbol
Stefan Beller [Fri, 30 Jun 2017 00:06:49 +0000 (17:06 -0700)] 
diff.c: introduce emit_diff_symbol

In a later patch we want to buffer all output before emitting it as a
new feature ("markup moved lines") conceptually cannot be implemented
in a single pass over the output.

There are different approaches to buffer all output such as:
* Buffering on the char level, i.e. we'd have a char[] which would
  grow at approximately 80 characters a line. This would keep the
  output completely unstructured, but might be very easy to implement,
  such as redirecting all output to a temporary file and working off
  that. The later passes over the buffer are quite complicated though,
  because we have to parse back any output and then decide if it should
  be modified.

* Buffer on a line level. As the output is mostly line oriented already,
  this would make sense, but it still is a bit awkward as we'd have to
  make sense of it again by looking at the first characters of a line
  to decide what part of a diff a line is.

* Buffer semantically. Imagine there is a formal grammar for the diff
  output and we'd keep the symbols of this grammar around. This keeps
  the highest level of structure in the buffered data, such that the
  actual memory requirements are less than say the first option. Instead
  of buffering the characters of the line, we'll buffer what we intend
  to do plus additional information for the specifics. An output of

    diff --git a/new.txt b/new.txt
    index fa69b07..412428c 100644
    Binary files a/new.txt and b/new.txt differ

  could be buffered as
     DIFF_SYMBOL_DIFF_START + new.txt
     DIFF_SYMBOL_INDEX_MODE + fa69b07 412428c "non-executable" flag
     DIFF_SYMBOL_BINARY_FILES + new.txt

This and the following patches introduce the third option of buffering
by first moving any output to emit_diff_symbol, and then introducing the
buffering in this function.

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodiff.c: factor out diff_flush_patch_all_file_pairs
Stefan Beller [Fri, 30 Jun 2017 00:06:48 +0000 (17:06 -0700)] 
diff.c: factor out diff_flush_patch_all_file_pairs

In a later patch we want to do more things before and after all filepairs
are flushed. So factor flushing out all file pairs into its own function
that the new code can be plugged in easily.

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodiff.c: move line ending check into emit_hunk_header
Stefan Beller [Fri, 30 Jun 2017 00:06:47 +0000 (17:06 -0700)] 
diff.c: move line ending check into emit_hunk_header

The emit_hunk_header() function is responsible for assembling a
hunk header and calling emit_line() to send the hunk header
to the output file.  Its only caller fn_out_consume() needs
to prepare for a case where the function emits an incomplete
line and add the terminating LF.

Instead make sure emit_hunk_header() to always send a
completed line to emit_line().

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodiff.c: readability fix
Stefan Beller [Fri, 30 Jun 2017 00:06:46 +0000 (17:06 -0700)] 
diff.c: readability fix

We already have dereferenced 'p->two' into a local variable 'two'.
Use that.

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agoMerge branch 'sb/hashmap-customize-comparison' into sb/diff-color-move
Junio C Hamano [Fri, 30 Jun 2017 20:12:34 +0000 (13:12 -0700)] 
Merge branch 'sb/hashmap-customize-comparison' into sb/diff-color-move

* sb/hashmap-customize-comparison: (566 commits)
  hashmap: migrate documentation from Documentation/technical into header
  patch-ids.c: use hashmap correctly
  hashmap.h: compare function has access to a data field
  Twelfth batch for 2.14
  Git 2.13.2
  Eleventh batch for 2.14
  Revert "split-index: add and use unshare_split_index()"
  Tenth batch for 2.14
  add--interactive: quote commentChar regex
  add--interactive: handle EOF in prompt_yesno
  auto-correct: tweak phrasing
  docs: update 64-bit core.packedGitLimit default
  t7508: fix a broken indentation
  grep: fix erroneously copy/pasted variable in check/assert pattern
  Ninth batch for 2.14
  glossary: define 'stash entry'
  status: add optional stash count information
  stash: update documentation to use 'stash entry'
  for_each_bisect_ref(): don't trim refnames
  mergetools/meld: improve compatibiilty with Meld on macOS X
  ...

6 years agohashmap: migrate documentation from Documentation/technical into header
Stefan Beller [Fri, 30 Jun 2017 19:14:07 +0000 (12:14 -0700)] 
hashmap: migrate documentation from Documentation/technical into header

While at it, clarify the use of `key`, `keydata`, `entry_or_key` as well
as documenting the new data pointer for the compare function.

Rework the example.

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agopatch-ids.c: use hashmap correctly
Stefan Beller [Fri, 30 Jun 2017 19:14:06 +0000 (12:14 -0700)] 
patch-ids.c: use hashmap correctly

As alluded to in the previous patch, the code in patch-ids.c is
using the hashmaps API wrong.

Luckily we do not have a bug, as all hashmap functionality that we use
here (hashmap_get) passes through the keydata.  If hashmap_get_next were
to be used, a bug would occur as that passes NULL for the key_data.

So instead use the hashmap API correctly and provide the caller required
data in the compare function via the first argument that always gets
passed and was setup via the hashmap_init function.

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agohashmap.h: compare function has access to a data field
Stefan Beller [Fri, 30 Jun 2017 19:14:05 +0000 (12:14 -0700)] 
hashmap.h: compare function has access to a data field

When using the hashmap a common need is to have access to caller provided
data in the compare function. A couple of times we abuse the keydata field
to pass in the data needed. This happens for example in patch-ids.c.

This patch changes the function signature of the compare function
to have one more void pointer available. The pointer given for each
invocation of the compare function must be defined in the init function
of the hashmap and is just passed through.

Documentation of this new feature is deferred to a later patch.
This is a rather mechanical conversion, just adding the new pass-through
parameter.  However while at it improve the naming of the fields of all
compare functions used by hashmaps by ensuring unused parameters are
prefixed with 'unused_' and naming the parameters what they are (instead
of 'unused' make it 'unused_keydata').

Signed-off-by: Stefan Beller <sbeller@google.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agoTwelfth batch for 2.14
Junio C Hamano [Mon, 26 Jun 2017 21:12:46 +0000 (14:12 -0700)] 
Twelfth batch for 2.14

Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agoMerge branch 'mb/reword-autocomplete-message'
Junio C Hamano [Mon, 26 Jun 2017 21:09:33 +0000 (14:09 -0700)] 
Merge branch 'mb/reword-autocomplete-message'

Message update.

* mb/reword-autocomplete-message:
  auto-correct: tweak phrasing

6 years agoMerge branch 'ks/t7508-indent-fix'
Junio C Hamano [Mon, 26 Jun 2017 21:09:32 +0000 (14:09 -0700)] 
Merge branch 'ks/t7508-indent-fix'

Cosmetic update to a test.

* ks/t7508-indent-fix:
  t7508: fix a broken indentation

6 years agoMerge branch 'jk/add-p-commentchar-fix'
Junio C Hamano [Mon, 26 Jun 2017 21:09:31 +0000 (14:09 -0700)] 
Merge branch 'jk/add-p-commentchar-fix'

"git add -p" were updated in 2.12 timeframe to cope with custom
core.commentchar but the implementation was buggy and a
metacharacter like $ and * did not work.

* jk/add-p-commentchar-fix:
  add--interactive: quote commentChar regex
  add--interactive: handle EOF in prompt_yesno

6 years agoMerge branch 'dt/raise-core-packed-git-limit'
Junio C Hamano [Mon, 26 Jun 2017 21:09:30 +0000 (14:09 -0700)] 
Merge branch 'dt/raise-core-packed-git-limit'

Doc update for a topic already in 'master'.

* dt/raise-core-packed-git-limit:
  docs: update 64-bit core.packedGitLimit default

6 years agoMerge branch 'mh/packed-ref-store-prep'
Junio C Hamano [Mon, 26 Jun 2017 21:09:29 +0000 (14:09 -0700)] 
Merge branch 'mh/packed-ref-store-prep'

Bugfix for a topic that is (only) in 'master'.

* mh/packed-ref-store-prep:
  for_each_bisect_ref(): don't trim refnames
  lock_packed_refs(): fix cache validity check

6 years agoMerge branch 'lb/status-stash-count'
Junio C Hamano [Mon, 26 Jun 2017 21:09:29 +0000 (14:09 -0700)] 
Merge branch 'lb/status-stash-count'

"git status" learned to optionally give how many stash entries the
user has in its output.

* lb/status-stash-count:
  glossary: define 'stash entry'
  status: add optional stash count information
  stash: update documentation to use 'stash entry'

6 years agoSync with 2.13.2
Junio C Hamano [Sat, 24 Jun 2017 22:34:14 +0000 (15:34 -0700)] 
Sync with 2.13.2

Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agoGit 2.13.2 v2.13.2
Junio C Hamano [Sat, 24 Jun 2017 22:31:36 +0000 (15:31 -0700)] 
Git 2.13.2

Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agoMerge branch 'sn/reset-doc-typofix' into maint
Junio C Hamano [Sat, 24 Jun 2017 22:29:35 +0000 (15:29 -0700)] 
Merge branch 'sn/reset-doc-typofix' into maint

Doc update.

* sn/reset-doc-typofix:
  doc: git-reset: fix a trivial typo

6 years agoMerge branch 'sg/doc-pretty-formats' into maint
Junio C Hamano [Sat, 24 Jun 2017 22:29:35 +0000 (15:29 -0700)] 
Merge branch 'sg/doc-pretty-formats' into maint

Doc update.

* sg/doc-pretty-formats:
  docs/pretty-formats: stress that %- removes all preceding line-feeds

6 years agoMerge branch 'sd/t3200-branch-m-test' into maint
Junio C Hamano [Sat, 24 Jun 2017 22:29:34 +0000 (15:29 -0700)] 
Merge branch 'sd/t3200-branch-m-test' into maint

New test.

* sd/t3200-branch-m-test:
  t3200: add test for single parameter passed to -m option

6 years agoMerge branch 'sg/revision-parser-skip-prefix' into maint
Junio C Hamano [Sat, 24 Jun 2017 22:29:34 +0000 (15:29 -0700)] 
Merge branch 'sg/revision-parser-skip-prefix' into maint

Code clean-up.

* sg/revision-parser-skip-prefix:
  revision.c: use skip_prefix() in handle_revision_pseudo_opt()
  revision.c: use skip_prefix() in handle_revision_opt()
  revision.c: stricter parsing of '--early-output'
  revision.c: stricter parsing of '--no-{min,max}-parents'
  revision.h: turn rev_info.early_output back into an unsigned int

6 years agoMerge branch 'km/test-mailinfo-b-failure' into maint
Junio C Hamano [Sat, 24 Jun 2017 22:29:33 +0000 (15:29 -0700)] 
Merge branch 'km/test-mailinfo-b-failure' into maint

New tests.

* km/test-mailinfo-b-failure:
  t5100: add some more mailinfo tests

6 years agoMerge branch 'sb/submodule-rm-absorb' into maint
Junio C Hamano [Sat, 24 Jun 2017 22:29:32 +0000 (15:29 -0700)] 
Merge branch 'sb/submodule-rm-absorb' into maint

Doc update to a recently graduated topic.

* sb/submodule-rm-absorb:
  Documentation/git-rm: correct submodule description

6 years agoMerge branch 'jc/diff-tree-stale-comment' into maint
Junio C Hamano [Sat, 24 Jun 2017 22:29:31 +0000 (15:29 -0700)] 
Merge branch 'jc/diff-tree-stale-comment' into maint

Comment fix.

* jc/diff-tree-stale-comment:
  diff-tree: update stale in-code comments

6 years agoMerge branch 'ps/stash-push-pathspec-fix' into maint
Junio C Hamano [Sat, 24 Jun 2017 22:29:30 +0000 (15:29 -0700)] 
Merge branch 'ps/stash-push-pathspec-fix' into maint

"git stash push <pathspec>" did not work from a subdirectory at all.
Bugfix for a topic in v2.13

* ps/stash-push-pathspec-fix:
  git-stash: fix pushing stash with pathspec from subdir

6 years agoMerge branch 'ls/github' into maint
Junio C Hamano [Sat, 24 Jun 2017 22:29:29 +0000 (15:29 -0700)] 
Merge branch 'ls/github' into maint

Help contributors that visit us at GitHub.

* ls/github:
  Configure Git contribution guidelines for github.com

6 years agoMerge branch 'jk/pack-idx-corruption-safety' into maint
Junio C Hamano [Sat, 24 Jun 2017 22:29:29 +0000 (15:29 -0700)] 
Merge branch 'jk/pack-idx-corruption-safety' into maint

A flaky test has been corrected.

* jk/pack-idx-corruption-safety:
  t5313: make extended-table test more deterministic

6 years agoMerge branch 'jk/diff-blob' into maint
Junio C Hamano [Sat, 24 Jun 2017 22:29:28 +0000 (15:29 -0700)] 
Merge branch 'jk/diff-blob' into maint

The result from "git diff" that compares two blobs, e.g. "git diff
$commit1:$path $commit2:$path", used to be shown with the full
object name as given on the command line, but it is more natural to
use the $path in the output and use it to look up .gitattributes.

* jk/diff-blob:
  diff: use blob path for blob/file diffs
  diff: use pending "path" if it is available
  diff: use the word "path" instead of "name" for blobs
  diff: pass whole pending entry in blobinfo
  handle_revision_arg: record paths for pending objects
  handle_revision_arg: record modes for "a..b" endpoints
  t4063: add tests of direct blob diffs
  get_sha1_with_context: dynamically allocate oc->path
  get_sha1_with_context: always initialize oc->symlink_path
  sha1_name: consistently refer to object_context as "oc"
  handle_revision_arg: add handle_dotdot() helper
  handle_revision_arg: hoist ".." check out of range parsing
  handle_revision_arg: stop using "dotdot" as a generic pointer
  handle_revision_arg: simplify commit reference lookups
  handle_revision_arg: reset "dotdot" consistently

6 years agoMerge branch 'jc/name-rev-lw-tag' into maint
Junio C Hamano [Sat, 24 Jun 2017 22:29:27 +0000 (15:29 -0700)] 
Merge branch 'jc/name-rev-lw-tag' into maint

"git describe --contains" penalized light-weight tags so much that
they were almost never considered.  Instead, give them about the
same chance to be considered as an annotated tag that is the same
age as the underlying commit would.

* jc/name-rev-lw-tag:
  name-rev: favor describing with tags and use committer date to tiebreak
  name-rev: refactor logic to see if a new candidate is a better name

6 years agoEleventh batch for 2.14
Junio C Hamano [Sat, 24 Jun 2017 21:34:11 +0000 (14:34 -0700)] 
Eleventh batch for 2.14

Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agoMerge branch 'ab/free-and-null'
Junio C Hamano [Sat, 24 Jun 2017 21:28:41 +0000 (14:28 -0700)] 
Merge branch 'ab/free-and-null'

A common pattern to free a piece of memory and assign NULL to the
pointer that used to point at it has been replaced with a new
FREE_AND_NULL() macro.

* ab/free-and-null:
  *.[ch] refactoring: make use of the FREE_AND_NULL() macro
  coccinelle: make use of the "expression" FREE_AND_NULL() rule
  coccinelle: add a rule to make "expression" code use FREE_AND_NULL()
  coccinelle: make use of the "type" FREE_AND_NULL() rule
  coccinelle: add a rule to make "type" code use FREE_AND_NULL()
  git-compat-util: add a FREE_AND_NULL() wrapper around free(ptr); ptr = NULL

6 years agoMerge branch 'jk/warn-add-gitlink'
Junio C Hamano [Sat, 24 Jun 2017 21:28:41 +0000 (14:28 -0700)] 
Merge branch 'jk/warn-add-gitlink'

Using "git add d/i/r" when d/i/r is the top of the working tree of
a separate repository would create a gitlink in the index, which
would appear as a not-quite-initialized submodule to others.  We
learned to give warnings when this happens.

* jk/warn-add-gitlink:
  t: move "git add submodule" into test blocks
  add: warn when adding an embedded repository

6 years agoMerge branch 'bw/config-h'
Junio C Hamano [Sat, 24 Jun 2017 21:28:40 +0000 (14:28 -0700)] 
Merge branch 'bw/config-h'

Fix configuration codepath to pay proper attention to commondir
that is used in multi-worktree situation, and isolate config API
into its own header file.

* bw/config-h:
  config: don't implicitly use gitdir or commondir
  config: respect commondir
  setup: teach discover_git_directory to respect the commondir
  config: don't include config.h by default
  config: remove git_config_iter
  config: create config.h

6 years agoMerge branch 'bw/ls-files-sans-the-index'
Junio C Hamano [Sat, 24 Jun 2017 21:28:40 +0000 (14:28 -0700)] 
Merge branch 'bw/ls-files-sans-the-index'

Code clean-up.

* bw/ls-files-sans-the-index:
  ls-files: factor out tag calculation
  ls-files: factor out debug info into a function
  ls-files: convert show_files to take an index
  ls-files: convert show_ce_entry to take an index
  ls-files: convert prune_cache to take an index
  ls-files: convert ce_excluded to take an index
  ls-files: convert show_ru_info to take an index
  ls-files: convert show_other_files to take an index
  ls-files: convert show_killed_files to take an index
  ls-files: convert write_eolinfo to take an index
  ls-files: convert overlay_tree_on_cache to take an index
  tree: convert read_tree to take an index parameter
  convert: convert renormalize_buffer to take an index
  convert: convert convert_to_git to take an index
  convert: convert convert_to_git_filter_fd to take an index
  convert: convert crlf_to_git to take an index
  convert: convert get_cached_convert_stats_ascii to take an index

6 years agoMerge branch 'js/alias-early-config'
Junio C Hamano [Sat, 24 Jun 2017 21:28:39 +0000 (14:28 -0700)] 
Merge branch 'js/alias-early-config'

The code to pick up and execute command alias definition from the
configuration used to switch to the top of the working tree and
then come back when the expanded alias was executed, which was
unnecessarilyl complex.  Attempt to simplify the logic by using the
early-config mechanism that does not chdir around.

* js/alias-early-config:
  alias: use the early config machinery to expand aliases
  t7006: demonstrate a problem with aliases in subdirectories
  t1308: relax the test verifying that empty alias values are disallowed
  help: use early config when autocorrecting aliases
  config: report correct line number upon error
  discover_git_directory(): avoid setting invalid git_dir

6 years agoMerge branch 'sn/reset-doc-typofix'
Junio C Hamano [Sat, 24 Jun 2017 21:28:39 +0000 (14:28 -0700)] 
Merge branch 'sn/reset-doc-typofix'

Doc update.

* sn/reset-doc-typofix:
  doc: git-reset: fix a trivial typo

6 years agoMerge branch 'sg/doc-pretty-formats'
Junio C Hamano [Sat, 24 Jun 2017 21:28:39 +0000 (14:28 -0700)] 
Merge branch 'sg/doc-pretty-formats'

Doc update.

* sg/doc-pretty-formats:
  docs/pretty-formats: stress that %- removes all preceding line-feeds

6 years agoMerge branch 'rs/pretty-add-again'
Junio C Hamano [Sat, 24 Jun 2017 21:28:38 +0000 (14:28 -0700)] 
Merge branch 'rs/pretty-add-again'

The pretty-format specifiers like '%h', '%t', etc. had an
optimization that no longer works correctly.  In preparation/hope
of getting it correctly implemented, first discard the optimization
that is broken.

* rs/pretty-add-again:
  pretty: recalculate duplicate short hashes

6 years agoMerge branch 'jk/diff-highlight-module'
Junio C Hamano [Sat, 24 Jun 2017 21:28:37 +0000 (14:28 -0700)] 
Merge branch 'jk/diff-highlight-module'

The 'diff-highlight' program (in contrib/) has been restructured
for easier reuse by an external project 'diff-so-fancy'.

* jk/diff-highlight-module:
  diff-highlight: split code into module

6 years agoMerge branch 'ah/doc-gitattributes-empty-index'
Junio C Hamano [Sat, 24 Jun 2017 21:28:37 +0000 (14:28 -0700)] 
Merge branch 'ah/doc-gitattributes-empty-index'

An example in documentation that does not work in multi worktree
configuration has been corrected.

* ah/doc-gitattributes-empty-index:
  doc: do not use `rm .git/index` when normalizing line endings

6 years agoMerge branch 'ab/wildmatch-glob-slash-test'
Junio C Hamano [Sat, 24 Jun 2017 21:28:37 +0000 (14:28 -0700)] 
Merge branch 'ab/wildmatch-glob-slash-test'

A new test to show the interaction between the pattern [^a-z]
(which matches '/') and a slash in a path has been added.  The
pattern should not match the slash with "pathmatch", but should
with "wildmatch".

* ab/wildmatch-glob-slash-test:
  wildmatch test: cover a blind spot in "/" matching

6 years agoMerge branch 'ab/pcre-v2'
Junio C Hamano [Sat, 24 Jun 2017 21:28:36 +0000 (14:28 -0700)] 
Merge branch 'ab/pcre-v2'

Hotfix for a topic already in 'master'.

* ab/pcre-v2:
  grep: fix erroneously copy/pasted variable in check/assert pattern

6 years agoMerge branch 'da/mergetools-meld-output-opt-on-macos'
Junio C Hamano [Sat, 24 Jun 2017 21:28:36 +0000 (14:28 -0700)] 
Merge branch 'da/mergetools-meld-output-opt-on-macos'

"git mergetool" learned to work around a wrapper MacOS X adds
around underlying meld.

* da/mergetools-meld-output-opt-on-macos:
  mergetools/meld: improve compatibiilty with Meld on macOS X

6 years agoMerge branch 'nd/split-index-unshare'
Junio C Hamano [Sat, 24 Jun 2017 19:04:25 +0000 (12:04 -0700)] 
Merge branch 'nd/split-index-unshare'

* nd/split-index-unshare:
  Revert "split-index: add and use unshare_split_index()"

6 years agoRevert "split-index: add and use unshare_split_index()"
Junio C Hamano [Sat, 24 Jun 2017 19:02:39 +0000 (12:02 -0700)] 
Revert "split-index: add and use unshare_split_index()"

This reverts commit f9d7abec2ad2f9eb3d8873169cc28c34273df082;
see public-inbox.org/git/CAP8UFD0bOfzY-_hBDKddOcJdPUpP2KEVaX_SrCgvAMYAHtseiQ@mail.gmail.com

6 years agoTenth batch for 2.14
Junio C Hamano [Thu, 22 Jun 2017 21:18:02 +0000 (14:18 -0700)] 
Tenth batch for 2.14

Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agoMerge branch 'rs/strbuf-addftime-zZ'
Junio C Hamano [Thu, 22 Jun 2017 21:15:25 +0000 (14:15 -0700)] 
Merge branch 'rs/strbuf-addftime-zZ'

As there is no portable way to pass timezone information to
strftime, some output format from "git log" and friends are
impossible to produce.  Teach our own strbuf_addftime to replace %z
and %Z with caller-supplied values to help working around this.

* rs/strbuf-addftime-zZ:
  date: use localtime() for "-local" time formats
  t0006: check --date=format zone offsets
  strbuf: let strbuf_addftime handle %z and %Z itself

6 years agoMerge branch 'sd/t3200-branch-m-test'
Junio C Hamano [Thu, 22 Jun 2017 21:15:25 +0000 (14:15 -0700)] 
Merge branch 'sd/t3200-branch-m-test'

New test.

* sd/t3200-branch-m-test:
  t3200: add test for single parameter passed to -m option

6 years agoMerge branch 'ps/stash-push-pathspec-fix'
Junio C Hamano [Thu, 22 Jun 2017 21:15:24 +0000 (14:15 -0700)] 
Merge branch 'ps/stash-push-pathspec-fix'

"git stash push <pathspec>" did not work from a subdirectory at all.
Bugfix for a topic in v2.13

* ps/stash-push-pathspec-fix:
  git-stash: fix pushing stash with pathspec from subdir

6 years agoMerge branch 'ls/github'
Junio C Hamano [Thu, 22 Jun 2017 21:15:24 +0000 (14:15 -0700)] 
Merge branch 'ls/github'

Help contributors that visit us at GitHub.

* ls/github:
  Configure Git contribution guidelines for github.com

6 years agoMerge branch 'sg/revision-parser-skip-prefix'
Junio C Hamano [Thu, 22 Jun 2017 21:15:23 +0000 (14:15 -0700)] 
Merge branch 'sg/revision-parser-skip-prefix'

Code clean-up.

* sg/revision-parser-skip-prefix:
  revision.c: use skip_prefix() in handle_revision_pseudo_opt()
  revision.c: use skip_prefix() in handle_revision_opt()
  revision.c: stricter parsing of '--early-output'
  revision.c: stricter parsing of '--no-{min,max}-parents'
  revision.h: turn rev_info.early_output back into an unsigned int

6 years agoMerge branch 'mh/fast-import-raise-default-depth'
Junio C Hamano [Thu, 22 Jun 2017 21:15:23 +0000 (14:15 -0700)] 
Merge branch 'mh/fast-import-raise-default-depth'

"fast-import" uses a default pack chain depth that is consistent
with other parts of the system.

* mh/fast-import-raise-default-depth:
  fast-import: increase the default pack depth to 50

6 years agoMerge branch 'km/test-mailinfo-b-failure'
Junio C Hamano [Thu, 22 Jun 2017 21:15:22 +0000 (14:15 -0700)] 
Merge branch 'km/test-mailinfo-b-failure'

New tests.

* km/test-mailinfo-b-failure:
  t5100: add some more mailinfo tests

6 years agoMerge branch 'ah/filter-branch-setup'
Junio C Hamano [Thu, 22 Jun 2017 21:15:21 +0000 (14:15 -0700)] 
Merge branch 'ah/filter-branch-setup'

"filter-branch" learned a pseudo filter "--setup" that can be used
to define a common function/variable that can be used by other
filters.

* ah/filter-branch-setup:
  filter-branch: add [--] to usage
  filter-branch: add `--setup` step

6 years agoMerge branch 'pc/dir-count-slashes'
Junio C Hamano [Thu, 22 Jun 2017 21:15:21 +0000 (14:15 -0700)] 
Merge branch 'pc/dir-count-slashes'

Three instances of the same helper function have been consolidated
to one.

* pc/dir-count-slashes:
  dir: create function count_slashes()

6 years agoMerge branch 'sb/t4005-modernize'
Junio C Hamano [Thu, 22 Jun 2017 21:15:21 +0000 (14:15 -0700)] 
Merge branch 'sb/t4005-modernize'

Test clean-up.

* sb/t4005-modernize:
  t4005: modernize style and drop hard coded sha1

6 years agoMerge branch 'nd/fopen-errors'
Junio C Hamano [Thu, 22 Jun 2017 21:15:20 +0000 (14:15 -0700)] 
Merge branch 'nd/fopen-errors'

Hotfix for a topic that is already in 'master'.

* nd/fopen-errors:
  configure.ac: loosen FREAD_READS_DIRECTORIES test program

6 years agoadd--interactive: quote commentChar regex
Jeff King [Wed, 21 Jun 2017 19:28:59 +0000 (15:28 -0400)] 
add--interactive: quote commentChar regex

Since c9d961647 (i18n: add--interactive: mark
edit_hunk_manually message for translation, 2016-12-14),
when the user asks to edit a hunk manually, we respect
core.commentChar in generating the edit instructions.
However, when we then strip out comment lines, we use a
simple regex like:

  /^$commentChar/

If your chosen comment character is a regex metacharacter,
then that will behave in a confusing manner ("$", for
instance, would only eliminate blank lines, not actual
comment lines).

We can fix that by telling perl not to respect
metacharacters.

Reported-by: Christian Rösch <christian@croesch.de>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agoadd--interactive: handle EOF in prompt_yesno
Jeff King [Wed, 21 Jun 2017 19:26:36 +0000 (15:26 -0400)] 
add--interactive: handle EOF in prompt_yesno

The prompt_yesno function loops indefinitely waiting for a
"y" or "n" response. But it doesn't handle EOF, meaning
that we can end up in an infinite loop of reading EOF from
stdin. One way to simulate that is with:

  echo e | GIT_EDITOR='echo corrupt >' git add -p

Let's break out of the loop and propagate the undef to the
caller. Without modifying the callers that effectively turns
it into a "no" response. This is reasonable for both of the
current callers, and it leaves room for any future caller to
check for undef explicitly.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agoauto-correct: tweak phrasing
Marc Branchaud [Wed, 21 Jun 2017 13:57:38 +0000 (09:57 -0400)] 
auto-correct: tweak phrasing

When help.autoCorrect is enabled, an invalid git command prints a
warning and a continuation message, which differs depending on
whether or not the value of help.autoCorrect is positive or
negative.

With help.autoCorrect = 15:

   WARNING: You called a Git command named 'lgo', which does not exist.
   Continuing under the assumption that you meant 'log'
   in 1.5 seconds automatically...

With help.autoCorrect < 0:

   WARNING: You called a Git command named 'lgo', which does not exist.
   Continuing under the assumption that you meant 'log'

The continuation message's phrasing is awkward.  This commit cleans it up.
As a bonus, we now use full-sentence strings which make translation easier.

With help.autoCorrect = 15:

   WARNING: You called a Git command named 'lgo', which does not exist.
   Continuing in 1.5 seconds, assuming that you meant 'log'.

With help.autoCorrect < 0:

   WARNING: You called a Git command named 'lgo', which does not exist.
   Continuing under the assumption that you meant 'log'.

Signed-off-by: Marc Branchaud <marcnarc@xiplink.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agodocs: update 64-bit core.packedGitLimit default
Jeff King [Wed, 21 Jun 2017 13:51:30 +0000 (09:51 -0400)] 
docs: update 64-bit core.packedGitLimit default

We bumped the default in be4ca2905 (Increase
core.packedGitLimit, 2017-04-20) but never adjusted the
documentation to match.

Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agot7508: fix a broken indentation
Kaartic Sivaraam [Wed, 21 Jun 2017 12:31:45 +0000 (18:01 +0530)] 
t7508: fix a broken indentation

Change the indentation from "\t " to "\t". This indenting issue was
introduced when the test was added in commit 1d2f393ac9
("status/commit: show staged submodules regardless of ignore
config", 2014-04-05).

Helped-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Kaartic Sivaraam <kaarticsivaraam91196@gmail.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agogrep: fix erroneously copy/pasted variable in check/assert pattern
Ævar Arnfjörð Bjarmason [Mon, 19 Jun 2017 22:01:48 +0000 (22:01 +0000)] 
grep: fix erroneously copy/pasted variable in check/assert pattern

Fix an erroneously copy/pasted check for the pcre2_jit_stack variable
to check pcre2_match_context instead. The former was already checked
in the preceding "if" statement.

This is a trivial and obvious error introduced in my commit
94da9193a6 ("grep: add support for PCRE v2", 2017-06-01).

In practice if pcre2_match_context_create() returned NULL we were
likely in a situation where malloc() was returning NULL, and were thus
screwed anyway, but if only the pcre2_match_context_create() call
returned NULL (through some transitory bug) PCRE v2 would just
allocate and supply its own context object when matching, and we'd run
normally at the trivial expense of not getting a slight speedup by
sharing the context object between successive matches.

Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agoNinth batch for 2.14
Junio C Hamano [Mon, 19 Jun 2017 19:41:12 +0000 (12:41 -0700)] 
Ninth batch for 2.14

Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agoMerge branch 'jk/consistent-h'
Junio C Hamano [Mon, 19 Jun 2017 19:38:45 +0000 (12:38 -0700)] 
Merge branch 'jk/consistent-h'

"git $cmd -h" for builtin commands calls the implementation of the
command (i.e. cmd_$cmd() function) without doing any repository
set-up, and the commands that expect RUN_SETUP is done by the Git
potty needs to be prepared to show the help text without barfing.

* jk/consistent-h:
  t0012: test "-h" with builtins
  git: add hidden --list-builtins option
  version: convert to parse-options
  diff- and log- family: handle "git cmd -h" early
  submodule--helper: show usage for "-h"
  remote-{ext,fd}: print usage message on invalid arguments
  upload-archive: handle "-h" option early
  credential: handle invalid arguments earlier

6 years agoMerge branch 'ab/perf-remove-index-lock'
Junio C Hamano [Mon, 19 Jun 2017 19:38:44 +0000 (12:38 -0700)] 
Merge branch 'ab/perf-remove-index-lock'

When an existing repository is used for t/perf testing, we first
create bit-for-bit copy of it, which may grab a transient state of
the repository and freeze it into the repository used for testing,
which then may cause Git operations to fail.  Single out "the index
being locked" case and forcibly drop the lock from the copy.

* ab/perf-remove-index-lock:
  perf: work around the tested repo having an index.lock

6 years agoMerge branch 'bw/object-id'
Junio C Hamano [Mon, 19 Jun 2017 19:38:44 +0000 (12:38 -0700)] 
Merge branch 'bw/object-id'

Conversion from uchar[20] to struct object_id continues.

* bw/object-id: (33 commits)
  diff: rename diff_fill_sha1_info to diff_fill_oid_info
  diffcore-rename: use is_empty_blob_oid
  tree-diff: convert path_appendnew to object_id
  tree-diff: convert diff_tree_paths to struct object_id
  tree-diff: convert try_to_follow_renames to struct object_id
  builtin/diff-tree: cleanup references to sha1
  diff-tree: convert diff_tree_sha1 to struct object_id
  notes-merge: convert write_note_to_worktree to struct object_id
  notes-merge: convert verify_notes_filepair to struct object_id
  notes-merge: convert find_notes_merge_pair_ps to struct object_id
  notes-merge: convert merge_from_diffs to struct object_id
  notes-merge: convert notes_merge* to struct object_id
  tree-diff: convert diff_root_tree_sha1 to struct object_id
  combine-diff: convert find_paths_* to struct object_id
  combine-diff: convert diff_tree_combined to struct object_id
  diff: convert diff_flush_patch_id to struct object_id
  patch-ids: convert to struct object_id
  diff: finish conversion for prepare_temp_file to struct object_id
  diff: convert reuse_worktree_file to struct object_id
  diff: convert fill_filespec to struct object_id
  ...

6 years agoMerge branch 'sb/submodule-rm-absorb'
Junio C Hamano [Mon, 19 Jun 2017 19:38:44 +0000 (12:38 -0700)] 
Merge branch 'sb/submodule-rm-absorb'

Doc update to a recently graduated topic.

* sb/submodule-rm-absorb:
  Documentation/git-rm: correct submodule description

6 years agoMerge branch 'ab/pcre-v2'
Junio C Hamano [Mon, 19 Jun 2017 19:38:43 +0000 (12:38 -0700)] 
Merge branch 'ab/pcre-v2'

Update "perl-compatible regular expression" support to enable JIT
and also allow linking with the newer PCRE v2 library.

* ab/pcre-v2:
  grep: add support for PCRE v2
  grep: un-break building with PCRE >= 8.32 without --enable-jit
  grep: un-break building with PCRE < 8.20
  grep: un-break building with PCRE < 8.32
  grep: add support for the PCRE v1 JIT API
  log: add -P as a synonym for --perl-regexp
  grep: skip pthreads overhead when using one thread
  grep: don't redundantly compile throwaway patterns under threading

6 years agoMerge branch 'jk/pathspec-magic-disambiguation'
Junio C Hamano [Mon, 19 Jun 2017 19:38:42 +0000 (12:38 -0700)] 
Merge branch 'jk/pathspec-magic-disambiguation'

The convention for a command line is to follow "git cmdname
--options" with revisions followed by an optional "--"
disambiguator and then finally pathspecs.  When "--" is not there,
we make sure early ones are all interpretable as revs (and do not
look like paths) and later ones are the other way around.  A
pathspec with "magic" (e.g. ":/p/a/t/h" that matches p/a/t/h from
the top-level of the working tree, no matter what subdirectory you
are working from) are conservatively judged as "not a path", which
required disambiguation more often.  The command line parser
learned to say "it's a pathspec" a bit more often when the syntax
looks like so.

* jk/pathspec-magic-disambiguation:
  verify_filename(): flip order of checks
  verify_filename(): treat ":(magic)" as a pathspec
  check_filename(): handle ":^" path magic
  check_filename(): use skip_prefix
  check_filename(): refactor ":/" handling
  t4208: add check for ":/" without matching file

6 years agoglossary: define 'stash entry'
Liam Beguin [Sat, 17 Jun 2017 22:30:52 +0000 (18:30 -0400)] 
glossary: define 'stash entry'

Signed-off-by: Liam Beguin <liambeguin@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agostatus: add optional stash count information
Liam Beguin [Sat, 17 Jun 2017 22:30:51 +0000 (18:30 -0400)] 
status: add optional stash count information

Introduce '--show-stash' and its configuration option 'status.showStash'
to allow git-status to show information about currently stashed entries.

Signed-off-by: Liam Beguin <liambeguin@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agostash: update documentation to use 'stash entry'
Liam Beguin [Sat, 17 Jun 2017 22:30:50 +0000 (18:30 -0400)] 
stash: update documentation to use 'stash entry'

Most of the time, a 'stash entry' is called a 'stash'. Lets try to make
this more consistent and use 'stash entry' instead.

Signed-off-by: Liam Beguin <liambeguin@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agofor_each_bisect_ref(): don't trim refnames
Michael Haggerty [Sun, 18 Jun 2017 13:39:41 +0000 (15:39 +0200)] 
for_each_bisect_ref(): don't trim refnames

`for_each_bisect_ref()` is called by `for_each_bad_bisect_ref()` with
a term "bad". This used to make it call `for_each_ref_in_submodule()`
with a prefix "refs/bisect/bad". But the latter is the name of the
reference that is being sought, so the empty string was being passed
to the callback as the trimmed refname. Moreover, this questionable
practice was turned into an error by

    b9c8e7f2fb prefix_ref_iterator: don't trim too much, 2017-05-22

It makes more sense (and agrees better with the documentation of
`--bisect`) for the callers to receive the full reference names. So

* Add a new function, `for_each_fullref_in_submodule()`, to the refs
  API. This plugs a gap in the existing functionality, analogous to
  `for_each_fullref_in()` but accepting a `submodule` argument.

* Change `for_each_bad_bisect_ref()` to call the new function rather
  than `for_each_ref_in_submodule()`.

* Add a test.

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agomergetools/meld: improve compatibiilty with Meld on macOS X
David Aguilar [Mon, 19 Jun 2017 02:10:33 +0000 (19:10 -0700)] 
mergetools/meld: improve compatibiilty with Meld on macOS X

The macOS X fork of Meld[1] requires a "=" in the "--output"
argument, as it uses a wrapper[2] script that munges the
"--output" argument before calling into the common "meld"
script.

The macOS X wrapper script[2] accepts "--output=<filename>"
only, despite the fact that the underlying meld code accepts
both "--output <filename" and "--output=<filename>"[3].

All versions of meld which accept "--output" accept it in
the "--output=<filename>" form, so use "--output=<file>" for
maximum compatibility.

[1] https://github.com/yousseb/meld
[2] https://github.com/yousseb/meld/blob/master/osx/Meld
[3] https://github.com/yousseb/meld/issues/42

Reported-by: Matthew Groth <mgroth49@gmail.com>
Helped-by: Samuel Lijin <sxlijin@gmail.com>
Signed-off-by: David Aguilar <davvid@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years ago*.[ch] refactoring: make use of the FREE_AND_NULL() macro
Ævar Arnfjörð Bjarmason [Thu, 15 Jun 2017 23:15:49 +0000 (23:15 +0000)] 
*.[ch] refactoring: make use of the FREE_AND_NULL() macro

Replace occurrences of `free(ptr); ptr = NULL` which weren't caught by
the coccinelle rule. These fall into two categories:

 - free/NULL assignments one after the other which coccinelle all put
   on one line, which is functionally equivalent code, but very ugly.

 - manually spotted occurrences where the NULL assignment isn't right
   after the free() call.

Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
6 years agococcinelle: make use of the "expression" FREE_AND_NULL() rule
Ævar Arnfjörð Bjarmason [Thu, 15 Jun 2017 23:15:48 +0000 (23:15 +0000)] 
coccinelle: make use of the "expression" FREE_AND_NULL() rule

A follow-up to the existing "expression" rule added in an earlier
change. This manually excludes a few occurrences, mostly things that
resulted in many FREE_AND_NULL() on one line, that'll be manually
fixed in a subsequent change.

Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>