1f2e05f0b7 ("wildmatch: fix exponential behavior", 2023-03-20)
introduced a new test with a background process. Backgrounding
necessarily gives a result of 0, so that a seemingly broken && chain is
not really broken.
Adjust t3070 slightly so that our chain lint test recognizes the
construct for what it is and does not raise a false positive.
Signed-off-by: Michael J Gruber <git@grubix.eu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When dowild() cannot match a '*' or '/**/' wildcard then it must return
WM_ABORT_TO_STARSTAR or WM_ABORT_ALL respectively. Failure to observe
this results in unnecessary backtracking and the time taken for a failed
match increases exponentially with the number of wildcards in the
pattern [1]. Unfortunately in some instances dowild() returns WM_NOMATCH
for a failed match resulting in long match times for patterns containing
multiple wildcards as can be seen in the following benchmark.
(Note that the timings in the Benchmark 1 are really measuring the time
to execute test-tool rather than the time to match the pattern)
Benchmark 1: t/helper/test-tool wildmatch wildmatch aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaab "*a"
Time (mean ± σ): 22.8 ms ± 1.7 ms [User: 12.1 ms, System: 10.6 ms]
Range (min … max): 19.4 ms … 26.9 ms 113 runs
Warning: Ignoring non-zero exit code.
Benchmark 2: t/helper/test-tool wildmatch wildmatch aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaab "*a*a*a*a*a*a*a*a*a"
Time (mean ± σ): 5.244 s ± 0.228 s [User: 5.229 s, System: 0.010 s]
Range (min … max): 4.969 s … 5.707 s 10 runs
Warning: Ignoring non-zero exit code.
Summary
't/helper/test-tool wildmatch wildmatch aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaab "*a"' ran
230.37 ± 20.04 times faster than 't/helper/test-tool wildmatch wildmatch aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaab "*a*a*a*a*a*a*a*a*a"'
The security implications are limited as it only affects operations that
are potentially DoS vectors. For example by creating a blob containing
such a pattern a malicious user can exploit this behavior to use large
amounts of CPU time on a remote server by pushing the blob and then
creating a new clone with --filter=sparse:oid. However this filter type
is usually disabled as it is known to consume large amounts of CPU time
even without this bug.
The WM_MATCH changed in the first hunk of this patch comes from the
original implementation imported from rsync in 5230f605e1 (Import
wildmatch from rsync, 2012-10-15). Compared to the others converted here
it is fairly harmless as it only triggers at the end of the pattern and
so will only cause a single unnecessary backtrack. The others introduced
by 6f1a31f0aa (wildmatch: advance faster in <asterisk> + <literal>
patterns, 2013-01-01) and 46983441ae (wildmatch: make a special case for
"*/" with FNM_PATHNAME, 2013-01-01) are more pernicious and will cause
exponential behavior.
A new test is added to protect against future regressions.
[1] https://research.swtch.com/glob
Helped-by: Derrick Stolee <derrickstolee@github.com>
Signed-off-by: Phillip Wood <phillip.wood@dunelm.org.uk>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
`test_run_` in test-lib.sh "lints" the body of a test by sending it down
a `sed chainlint.sed | grep` pipeline; this happens once for each test
run by a test script. Although this pipeline may seem relatively cheap
in isolation, it can become expensive when invoked 26800+ times by `make
test`, once for each test run, despite the existence of only 16500+ test
definitions across all tests scripts.
This difference in the number of tests defined in the scripts (16500+)
and the number of tests actually run by `make test` (26800+) is
explained by the fact that some test scripts run a very large number of
small tests, all driven by a series of functions/loops which fill in the
test bodies. This means that certain test definitions are being linted
repeatedly (tens or hundreds of times) unnecessarily. To avoid such
unnecessary work, 2d86a96220 (t: avoid sed-based chain-linting in some
expensive cases, 2021-05-13) added an optimization hack which allows
individual scripts to manually suppress the unnecessary repeated linting
of the same test definition.
However, unlike chainlint.sed which checks a test body as the test is
run, chainlint.pl checks each test definition just once, no matter how
many times the test is run, thus the sort of optimization hack
introduced by 2d86a96220 is no longer needed and can be retired.
Therefore, revert 2d86a96220.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The top-level &&-chain checker built into t/test-lib.sh causes tests to
magically exit with code 117 if the &&-chain is broken. However, it has
the shortcoming that the magic does not work within `{...}` groups,
`(...)` subshells, `$(...)` substitutions, or within bodies of compound
statements, such as `if`, `for`, `while`, `case`, etc. `chainlint.sed`
partly fills in the gap by catching broken &&-chains in `(...)`
subshells, but bugs can still lurk behind broken &&-chains in the other
cases.
Fix broken &&-chains in compound statements in order to reduce the
number of possible lurking bugs.
Signed-off-by: Eric Sunshine <sunshine@sunshineco.com>
Reviewed-by: Elijah Newren <newren@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Mark various "generic" tests as passing when git is compiled with
SANITIZE=leak. These tests were subjectively picked from the lists of
passing tests since they're all small, and test some generic feature
such as wildmatch(), commonly used environment variables, ident
parsing etc.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Commit 878f988350 (t/test-lib: teach --chain-lint to detect broken
&&-chains in subshells, 2018-07-11) introduced additional chain-lint
tests which add an extra "sed" pipeline to each test we run. This has a
measurable impact on runtime. Here are timings with and without a new
environment variable (added by this patch) that lets you disable just
the additional sed-based chain-lint tests:
Benchmark #1: GIT_TEST_CHAIN_LINT_HARDER=1 make test
Time (mean ± σ): 64.202 s ± 1.030 s [User: 622.469 s, System: 301.402 s]
Range (min … max): 61.571 s … 65.662 s 10 runs
Benchmark #2: GIT_TEST_CHAIN_LINT_HARDER=0 make test
Time (mean ± σ): 57.591 s ± 0.333 s [User: 529.368 s, System: 270.618 s]
Range (min … max): 57.143 s … 58.309 s 10 runs
Summary
'GIT_TEST_CHAIN_LINT_HARDER=0 make test' ran
1.11 ± 0.02 times faster than 'GIT_TEST_CHAIN_LINT_HARDER=1 make test'
Of course those extra lint checks are doing something useful, so paying
a few extra seconds (at least on Linux) isn't so bad (though note the
CPU time; we're bounded in our parallel run here by the slowest test, so
it really is ~120s of CPU improvement).
But we can observe that there are some test scripts where they produce a
much stronger effect, and provide less value. In t0027 and t3070 we run
a very large number of small tests, all driven by a series of
functions/loops which are filling in the test bodies. There we get much
less bang for our buck in terms of bug-finding versus CPU cost.
This patch introduces a mechanism for controlling when those extra
lint checks are run, at two levels:
- a user can ask to disable or to force-enable the checks by setting
GIT_TEST_CHAIN_LINT_HARDER
- if the user hasn't specified a preference, individual scripts can
disable the checks by setting GIT_TEST_CHAIN_LINT_HARDER_DEFAULT;
scripts which don't set that get the current behavior of enabling
them.
In addition, this patch flips the default for t0027 and t3070's
mass-generated sections to disable the extra checks. Here are the timing
results for t0027:
Benchmark #1: GIT_TEST_CHAIN_LINT_HARDER=1 ./t0027-auto-crlf.sh
Time (mean ± σ): 17.078 s ± 0.848 s [User: 14.878 s, System: 7.075 s]
Range (min … max): 15.952 s … 18.421 s 10 runs
Benchmark #2: GIT_TEST_CHAIN_LINT_HARDER=0 ./t0027-auto-crlf.sh
Time (mean ± σ): 9.063 s ± 0.759 s [User: 7.890 s, System: 3.362 s]
Range (min … max): 7.747 s … 10.619 s 10 runs
Benchmark #3: ./t0027-auto-crlf.sh
Time (mean ± σ): 9.186 s ± 0.881 s [User: 7.957 s, System: 3.427 s]
Range (min … max): 7.796 s … 10.498 s 10 runs
Summary
'GIT_TEST_CHAIN_LINT_HARDER=0 ./t0027-auto-crlf.sh' ran
1.01 ± 0.13 times faster than './t0027-auto-crlf.sh'
1.88 ± 0.18 times faster than 'GIT_TEST_CHAIN_LINT_HARDER=1 ./t0027-auto-crlf.sh'
We can see that disabling the checks for the whole script buys us an
almost 2x speedup. But the new default behavior, disabling them only for
the mass-generated part, gets us most of that speedup (but still leaves
the checks on for further manual tests people might write).
As a side note, I'd caution about comparing runtimes and CPU seconds
between this timing and the earlier "make test" one. In "make test",
we're running a lot of scripts in parallel, so the CPU is throttling
down (and thus a CPU second saved here would count for more during a
parallel run; the same work takes more CPU seconds there).
We get similar results for t3070:
Benchmark #1: GIT_TEST_CHAIN_LINT_HARDER=1 ./t3070-wildmatch.sh
Time (mean ± σ): 20.054 s ± 3.967 s [User: 16.003 s, System: 8.286 s]
Range (min … max): 11.891 s … 23.671 s 10 runs
Benchmark #2: GIT_TEST_CHAIN_LINT_HARDER=0 ./t3070-wildmatch.sh
Time (mean ± σ): 12.399 s ± 2.256 s [User: 7.542 s, System: 5.342 s]
Range (min … max): 9.606 s … 15.727 s 10 runs
Benchmark #3: ./t3070-wildmatch.sh
Time (mean ± σ): 10.726 s ± 3.476 s [User: 6.790 s, System: 4.365 s]
Range (min … max): 5.444 s … 15.376 s 10 runs
Summary
'./t3070-wildmatch.sh' ran
1.16 ± 0.43 times faster than 'GIT_TEST_CHAIN_LINT_HARDER=0 ./t3070-wildmatch.sh'
1.87 ± 0.71 times faster than 'GIT_TEST_CHAIN_LINT_HARDER=1 ./t3070-wildmatch.sh'
Again, we get almost a 2x speedup disabling these. In this case, there
are no tests not covered by the script's "default to disable" behavior,
so the second two benchmarks should be the same (and while they do
differ, you can see the variance is quite high but they're within one
standard deviation).
So it seems like for these two scripts, at least, disabling the extra
checks is a reasonable tradeoff. Sadly, the overall runtime of "make
test" on my system doesn't get much faster. But that's because we're
mostly limited by the cost of the single biggest test. Here are the
top-5 tests by wall-clock time from a parallel run, before my patch:
57.9192368984222 t9001-send-email.sh
45.6329638957977 t0027-auto-crlf.sh
32.5278220176697 t3070-wildmatch.sh
22.2701289653778 t7610-mergetool.sh
20.8635759353638 t1701-racy-split-index.sh
And after:
57.1476998329163 t9001-send-email.sh
33.776211977005 t0027-auto-crlf.sh
21.3116669654846 t7610-mergetool.sh
20.7748689651489 t1701-racy-split-index.sh
19.6957249641418 t7112-reset-submodule.sh
We dropped 12s from t0027, and t3070 dropped off our list entirely at
around 16s. In both cases we're bound by t9001, but its slowness is
due to the actual tests, so we'll have to deal with it in a different
way. But this reduces overall CPU, and means that dealing with t9001 (by
improving the speed of send-email or splitting it apart) will let us
reduce our overall runtime even on multi-core machines.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A pattern with '**' that does not have a slash on either side used
to be an invalid one, but the code now treats such double-asterisks
the same way as two normal asterisks that happen to be adjacent to
each other.
* nd/wildmatch-double-asterisk:
wildmatch: change behavior of "foo**bar" in WM_PATHNAME mode
In WM_PATHNAME mode (or FNM_PATHNAME), '*' does not match '/' and '**'
can but only in three patterns:
- '**/' matches zero or more leading directories
- '/**/' matches zero or more directories in between
- '/**' matches zero or more trailing directories/files
When '**' is present but not in one of these patterns, the current
behavior is consider the pattern invalid and stop matching. In other
words, 'foo**bar' never matches anything, whatever you throw at it.
This behavior is arguably a bit confusing partly because we can't
really tell the user their pattern is invalid so that they can fix
it. So instead, tolerate it and make '**' act like two regular '*'s
(which is essentially the same as a single asterisk). This behavior
seems more predictable.
Noticed-by: dana <dana@dana.is>
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Change various tests that use an idiom of the form:
>expect &&
test_cmp expect actual
To instead use:
test_must_be_empty actual
The test_must_be_empty() wrapper was introduced in ca8d148daf ("test:
test_must_be_empty helper", 2013-06-09). Many of these tests have been
added after that time. This was mostly found with, and manually pruned
from:
git grep '^\s+>.*expect.* &&$' t
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There has never been any full roundtrip testing of what git-ls-files
and other commands that use wildmatch() actually do, rather we've been
satisfied with just testing the underlying C function.
Due to git-ls-files and friends having their own codepaths before they
call wildmatch() there's sometimes differences in the behavior between
the two. Even when we test for those (as with [1]), there was no one
place where you can review how these two modes differ.
Now there is. We now attempt to create a file called $haystack and
match $needle against it for each pair of $needle and $haystack that
we were passing to test-wildmatch.
If we can't create the file we skip the test. This ensures that we can
run this on all platforms and not maintain some infinitely growing
whitelist of e.g. platforms that don't support certain characters in
filenames.
A notable exception to this is Windows, where due to the reasons
explained in [2] the shellscript emulation layer might fake the
creation of a file such as "*", and "test -e" for it will succeed
since it just got created with some character that maps to "*", but
git ls-files won't be fooled by this.
Thus we need to skip creating certain filenames entirely on Windows,
the list here might be overly aggressive. I don't have access to a
Windows system to test this.
As a result of doing these tests we can now see the cases where these
two ways of testing wildmatch differ:
* Creating a file called 'a[]b' and running ls-files 'a[]b' will show
that file, but wildmatch("a[]b", "a[]b") will not match
* wildmatch() won't match a file called \ against \, but ls-files
will.
* `git --glob-pathspecs ls-files 'foo**'` will match a file
'foo/bba/arr', but wildmatch won't, however pathmatch will.
This seems like a bug to me, the two are otherwise equivalent as
these tests show.
This also reveals the case discussed in [1], since 2.16.0 '' is now an
error as far as ls-files is concerned, but wildmatch() itself happily
accepts it.
1. 9e4e8a64c2 ("pathspec: die on empty strings as pathspec",
2017-06-06)
2. nycvar.QRO.7.76.6.1801052133380.1337@wbunaarf-fpuvaqryva.tvgsbejvaqbjf.bet
(https://public-inbox.org/git/?q=nycvar.QRO.7.76.6.1801052133380.1337%40wbunaarf-fpuvaqryva.tvgsbejvaqbjf.bet)
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Rewrite the wildmatch() test suite so that each test now tests all
combinations of the wildmatch() WM_CASEFOLD and WM_PATHNAME flags.
Before this change some test inputs were not tested on
e.g. WM_PATHNAME. Now the function is stress tested on all possible
inputs, and for each input we declare what the result should be if the
mode is case-insensitive, or pathname matching, or case-sensitive or
not matching pathnames.
Also before this change, nothing was testing case-insensitive
non-pathname matching, so I've added that to test-wildmatch.c and made
use of it.
This yields a rather scary patch, but there are no functional changes
here, just more test coverage. Some now-redundant tests were deleted
as a result of this change, since they were now duplicating an earlier
test.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Use of ! should be reserved for non-git programs that are assumed not
to fail, see README. With this change only
t/t0110-urlmatch-normalization.sh is still using this anti-pattern.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Remove the unused fnmatch() test parameter from the wildmatch
test. The code that used to test this was removed in 70a8fc999d ("stop
using fnmatch (either native or compat)", 2014-02-15).
As a --word-diff shows the only change to the body of the tests is the
removal of the second out of four parameters passed to match().
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Use a pattern from the nul_match() function in t7008-grep-binary.sh to
make sure that we don't just fall through to the "else" if there's an
unknown parameter.
This is something I added in commit 77f6f4406f ("grep: add a test
helper function for less verbose -f \0 tests", 2017-05-20) to grep
tests, which were modeled on these wildmatch tests, and I'm now
porting back to the original wildmatch tests.
I am not using the "say '...'; exit 1" pattern from t0000-basic.sh
because if I fail I want to run the rest of the tests (unless under
-i), and doing this makes sure we do that and don't exit right away
without fully reporting our errors.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Don't try to vertically align the test output, which is futile anyway
under the TAP output where we're going to be emitting a number for
each test without aligning the test count.
This makes subsequent changes of mine where I'm not going to be
aligning this output as I add new tests easier.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Change the wildmatch test to use more standard shell style, usually we
use "if test" not "if [".
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Replace the 4-width mixed space & tab indentation in this file with
indentation with tabs as we do in most of the rest of our tests.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
A negated character class that does not include '/', e.g. [^a-z]:
- Should match '/' when doing "wildmatch"
- Should not match '/' when doing "pathmatch"
Add two tests to cover these cases.
Helped-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Remove a test line that's exactly the same as the preceding
line.
This was brought in in commit feabcc173b ("Integrate wildmatch to
git", 2012-10-15), these tests are originally copied from rsync.git,
but the duplicate line was never present there, so must have just
snuck in during integration with git by accident.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since v1.8.4 (about six months ago) wildmatch is used as default
replacement for fnmatch. We have seen only one fix since so wildmatch
probably has done a good job as fnmatch replacement. This concludes
the fnmatch->wildmatch transition by no longer relying on fnmatch.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Case folding is not done correctly when matching against the [:upper:]
character class and uppercased character ranges (e.g. A-Z).
Specifically, an uppercase letter fails to match against any of them
when case folding is requested because plain characters in the pattern
and the whole string are preemptively lowercased to handle the base case
fast.
That optimization is kept and ISLOWER() is used in the [:upper:] case
when case folding is requested, while matching against a character range
is retried with toupper() if the character was lowercase, as the bounds
of the range itself cannot be modified (in a case-insensitive context,
[A-_] is not equivalent to [a-_]).
Signed-off-by: Anthony Ramine <n.oxyde@gmail.com>
Reviewed-by: Duy Nguyen <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Normally when we match "*X" on "abcX", we call dowild("X", "abcX"),
dowild("X", "bcX"), dowild("X", "cX") and dowild("X", "X"). Only the
last call may have a chance of matching. By skipping the text before
"X", we can eliminate the first three useless calls.
compat, '*/*/*' on linux-2.6.git file list 2000 times, before:
wildmatch 7s 985049us
fnmatch 2s 735541us or 34.26% faster
and after:
wildmatch 4s 492549us
fnmatch 0s 888263us or 19.77% slower
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Normally we need recursion for "*". In this case we know that it
matches everything until "/" so we can skip the recursion.
glibc, '*/*/*' on linux-2.6.git file list 2000 times
before:
wildmatch 8s 74513us
fnmatch 1s 97042us or 13.59% faster
after:
wildmatch 3s 521862us
fnmatch 3s 488616us or 99.06% slower
Same test with compat/fnmatch:
wildmatch 8s 110763us
fnmatch 2s 980845us or 36.75% faster
wildmatch 3s 522156us
fnmatch 1s 544487us or 43.85% slower
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
So far, wildmatch() has always honoured directory boundary and there
was no way to turn it off. Make it behave more like fnmatch() by
requiring all callers that want the FNM_PATHNAME behaviour to pass
that in the equivalent flag WM_PATHNAME. Callers that do not specify
WM_PATHNAME will get wildcards like ? and * in their patterns matched
against '/', just like not passing FNM_PATHNAME to fnmatch().
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"**" is adjusted to only be effective when surrounded by slashes, in
40bbee0 (wildmatch: adjust "**" behavior - 2012-10-15). Except that
the commit did it wrong:
1. when it checks for "the preceding slash unless ** is at the
beginning", it compares to wrong pointer. It should have compared
to the beginning of the pattern, not the text.
2. prev_p points to the character before "**", not the first "*". The
correct comparison must be "prev_p < pattern" or
"prev_p + 1 == pattern", not "prev_p == pattern".
3. The pattern must be surrounded by slashes unless it's at the
beginning or the end of the pattern. We do two checks: one for the
preceding slash and one the trailing slash. Both checks must be
met. The use of "||" is wrong.
This patch fixes all above.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The failing tests make use of a POSIX character class, '[:xdigit:]'
in this case, which some versions of the fnmatch() library function
do not support. In the spirit of commit f1cf7b79 ("t3070: disable
unreliable fnmatch tests", 15-10-2012), we disable the fnmatch() half
of these tests.
Signed-off-by: Ramsay Jones <ramsay@ramsay1.demon.co.uk>
Acked-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The MSYS bash mangles arguments that begin with a forward slash
when they are passed to test-wildmatch. This causes tests to fail.
Avoid mangling by prepending "XXX", which is removed by
test-wildmatch before further processing.
[J6t: reworded commit message]
Reported-by: Johannes Sixt <j6t@kdbg.org>
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Johannes Sixt <j6t@kdbg.org>
"foo/**/bar" matches "foo/x/bar", "foo/x/y/bar"... but not
"foo/bar". We make a special case, when foo/**/ is detected (and
"foo/" part is already matched), try matching "bar" with the rest of
the string.
"Match one or more directories" semantics can be easily achieved using
"foo/*/**/bar".
This also makes "**/foo" match "foo" in addition to "x/foo",
"x/y/foo"..
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Standard wildmatch() sees consecutive asterisks as "*" that can also
match slashes. But that may be hard to explain to users as
"abc/**/def" can match "abcdef", "abcxyzdef", "abc/def", "abc/x/def",
"abc/x/y/def"...
This patch changes wildmatch so that users can do
- "**/def" -> all paths ending with file/directory 'def'
- "abc/**" - equivalent to "/abc/"
- "abc/**/def" -> "abc/x/def", "abc/x/y/def"...
- otherwise consider the pattern malformed if "**" is found
Basically the magic of "**" only remains if it's wrapped around by
slashes.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
These tests show different results on different fnmatch() versions. We
don't want to test fnmatch here. We want to make sure wildmatch
behavior matches fnmatch and that only makes sense in cases when
fnmatch() behaves consistently.
Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
