* jn/maint-plug-leak:
write-tree: Avoid leak when index refers to an invalid object
read-tree: stop leaking tree objects
core: Stop leaking ondisk_cache_entrys
Merge branch 'bc/use-more-hardlinks-in-install' into maint
* bc/use-more-hardlinks-in-install:
Makefile: make hard/symbolic links for non-builtins too
Makefile: link builtins residing in bin directory to main git binary too
* tr/rfc-reset-doc:
Documentation/reset: move "undo permanently" example behind "make topic"
Documentation/reset: reorder examples to match description
Documentation/reset: promote 'examples' one section up
Documentation/reset: separate options by mode
Documentation/git-reset: reorder modes for soft-mixed-hard progression
builtin/checkout: Fix message when switching to an existing branch
Fix "Switched to a new branch <name>" to read "Switched to branch
<name>" when <name> corresponds to an existing branch. This bug was
introduced in 02ac983 while introducing the `-B` switch.
Cc: Tay Ray Chuan <rctay89@gmail.com> Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
asciidoc already takes care of including a doctype for most of the
HTML documentation, but the user manual which is processed with
docbook-xsl directly lacks one (at least with Debian docbook-xsl
1.75.2+dfsg-5). This makes it harder to automatically validate the
HTML.
Reported-by: 積丹尼 <jidanni@jidanni.org> Signed-off-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Out-of-memory errors can either be actual lack of memory, or bugs (like
code trying to call xmalloc(-1) by mistake). A little more information
may help tracking bugs reported by users.
Signed-off-by: Matthieu Moy <Matthieu.Moy@imag.fr> Signed-off-by: Junio C Hamano <gitster@pobox.com>
merge-recursive: Avoid excessive output for and reprocessing of renames
In 5a2580d (merge_recursive: Fix renames across paths below D/F conflicts
2010-07-09) and ae74548 (merge-recursive: Fix multiple file rename across
D/F conflict 2010-08-17), renames across D/F conflicts were fixed by
making process_renames() consider as unprocessed renames whose dst_entry
"still" had higher stage entries. The assumption was that those higher
stage entries would have been cleared out of dst_entry by that point in
cases where the conflict could be resolved (normal renames with no D/F
conflicts). That is not the case -- higher stage entries will remain in
all cases.
Fix this by checking for higher stage entries corresponding to D/F
conflicts, namely that stages 2 and 3 have exactly one nonzero mode between
them. The nonzero mode stage corresponds to a file at the path, while the
stage with a zero mode will correspond to a directory at that path (since
rename/delete conflicts will have already been handled before this codepath
is reached.)
Signed-off-by: Elijah Newren <newren@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
t/t7008-grep-binary.sh: un-TODO a test that needs REG_STARTEND
Now that we have a regex engine that supports REG_STARTEND this test
should fail if "git grep" can't grep NULL characters.
Platforms that don't have a POSIX regex engine which supports
REG_STARTEND should always define NO_REGEX=YesPlease when compiling.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Acked-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
compat/regex: define out variables only used under RE_ENABLE_I18N
Wrap variables that were only used RE_ENABLE_I18N in `#ifdef
RE_ENABLE_I18N`. This eliminates compiler warnings when compiling with
NO_REGEX=YesPlease.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
git-gui: ensure correct application termination in git-gui--askpass
With Tk 8.5 the askpass utility can hang waiting for the wish shell
implicit event loop to exit. This patch uses an explicit event loop
to ensure correct application termination.
Reported-by: Anders Kaseorg <andersk@mit.edu> Tested-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Pat Thoyts <patthoyts@users.sourceforge.net>
Change regerror() declaration from K&R style to ANSI C (C89)
The MSVC headers typedef errcode as int, and thus confused the compiler in
the K&R style definition. ANSI style deconfuses it.
This patch was originally applied as v1.6.5-rc2~23 but needs to be
re-applied since compat/regex was overwritten by Ævar Arnfjörð
Bjarmason with the gawk regex engine.
Signed-off-by: Frank Li <lznuaa@gmail.com> Signed-off-by: Marius Storm-Olsen <mstormo@gmail.com> Acked-by: Johannes Sixt <j6t@kdbg.org> Signed-off-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>
compat/regex: get the gawk regex engine to compile within git
We need to define -DGAWK -DNO_MBSUPPORT so that the gawk regex engine
will compile, and include stdio.h and stddef.h in regex.h. Gawk itself
includes these headers before it includes the regex.h header.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Acked-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
compat/regex: use the regex engine from gawk for compat
Change the regex engine in compat to use the gawk engine from the
gawk-devel module in gawk CVS. This engine supports the REG_STARTEND
flag, which was optionally available in Git since v1.7.2-rc0~77^2~1.
The source was grabbed from cvs.savannah.gnu.org:/sources/gawk, and
these are the upstream versions of the files being included:
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Acked-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
In a --no-ff merge with conflicts, "git commit" used to forget the
--no-ff when used to complete the merge. That was fixed by
v1.6.1-rc1~134^2 (builtin-commit: use reduce_heads() only when
appropriate, 2008-10-03) for the builtin merge. Port the change to
the merge script in contrib/examples.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Based on v1.7.1.1~23^2 (merge: --log appends shortlog to message if
specified, 2010-05-11). Without this change, the scripted
(non-builtin) merge does not pass t7604.
Cc: Tay Ray Chuan <rctay89@gmail.com> Signed-off-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Some git-merge-* commands are not merge strategies. This is based on
v1.6.1-rc1~294^2~7 (builtin-merge: allow using a custom strategy,
2008-07-30) but it is less smart: we just use a hard-coded list of
forbidden strategy names. It is okay if this falls out of date, since
the code is just an example.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
- point out remote-tracking branches as "remote branch
'upstream/master'";
- avoid misleading log messages when a tag and branch
share a name.
This approximates the builtin merge command's behavior well
enough to pass the relevant tests.
Based roughly on v1.6.4.2~10^2 (merge: indicate remote tracking
branches in merge message, 2009-08-09) and v1.6.4.2~10^2~1 (merge: fix
incorrect merge message for ambiguous tag/branch, 2009-08-09).
Cc: Jeff King <peff@peff.net> Signed-off-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Check MERGE_HEAD and bail out if it exists. Based on v1.6.3.3~3^2
(refuse to merge during a merge, 2009-06-01). Without this change,
the scripted merge does not pass t3030.
Cc: Clemens Buchacher <drizzd@aon.at> Signed-off-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
merge script: tweak unmerged files message to match builtin
Before:
You are in the middle of a conflicted merge.
After:
Merge is not possible because you have unmerged files.
I prefer the old message, but the new one is more consistent with
other commands and tests expect it. In particular, without this
change the scripted merge does not pass t3030.
Based on v1.7.0-rc0~66^2 (Be more user-friendly when refusing to do
something because of conflict., 2010-01-12).
Cc: Matthieu Moy <Matthieu.Moy@imag.fr> Signed-off-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since v1.7.1.1~23^2 (merge: --log appends shortlog to message if
specified, 2010-05-11), the fmt-merge-msg backend supports custom text
to override the merge title "Merge <foo> into <bar>".
Expose this functionality for scripted callers. Example:
git fmt-merge-msg --log -m \
"$(printf '%s\n' \
"Merge branch 'api-cleanup' into feature" \
'' \
'This is to use a few functions refactored for this purpose.'
)" <.git/FETCH_HEAD
Cc: Tay Ray Chuan <rctay89@gmail.com> Signed-off-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
merge-base --independent to print reduced parent list in a merge
While show-branch --independent does not support more than MAX_REVS
revs, git internally supports more with a different algorithm.
Expose that functionality as "git merge-base --independent".
This should help scripts to catch up with builtin merge in supporting
dodecapus.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
merge-base --octopus to mimic show-branch --merge-base
While show-branch --merge-base does not support more than MAX_REVS
revs, git supports more with a different algorithm
(v1.6.0-rc0~51^2~13, Introduce get_octopus_merge_bases() in commit.c,
2008-06-27). Expose that functionality.
This should help scripts to catch up with builtin merge in supporting
dodecapus.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Documentation: add a SEE ALSO section for merge-base
For example, a person reading the merge-base man page might wonder
about the fastest way to check if one commit is an ancestor of
another (which would require rev-list).
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Guard setup with test_expect_success. Use test_might_fail
instead of ignoring the exit code from git config --unset.
Point out setup commands that are shared by multiple tests,
to make it easy to write GIT_SKIP_TESTS specifications that
work.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Guard setup with test_expect_success, put the opening quote
starting each test on the same line as the test_expect_* invocation,
and combine related actions into single tests.
While at it:
- use test_cmp instead of expr or test $foo = $bar, for more helpful
output with -v when tests fail;
- use test_commit for brevity.
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Probably as a development aid, this test script runs gitk --all
to allow the driver to inspect history between tests when run
with --debug. As a result, running all tests with --debug
requires closing a long series of gitk displays, one at a time.
Use git log --graph --oneline instead. This way, the history is
available for viewing with "git show" but the test script finishes
without interaction.
Longer term, it would be nice to have an option to run a
user-specified command between tests. This patch does not do
that.
Cc: Lars Hjemli <hjemli@gmail.com> Signed-off-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Guard setup commands with test_expect_success, so they are easier
to visually skip over and get to the good part. While at it:
- use "printf '%s\n' a b ..." instead of "cat <<EOF" for test
vectors with short lines;
- use test_cmp instead of test foo = bar where possible, for
better output with -v on failure;
- do not go to extraordinary lengths to print a relevant message
when test commands fail. There is a patch in flight that could be
used to restore the nice error messages in a cleaner way.
Cc: Lars Hjemli <hjemli@gmail.com> Cc: Matthieu Moy <Matthieu.Moy@imag.fr> Signed-off-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
When "git submodule sync" synchronizes the repository URLs
it only updates submodules' .git/config. However, the old
URLs still exist in the super-project's .git/config.
Update the super-project's configuration so that commands
such as "git submodule update" use the URLs from .gitmodules.
Signed-off-by: David Aguilar <davvid@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
tests: fix syntax error in "Use advise() for hints" test
Change the test introduced in the "Use advise() for hints" patch by
Jonathan Nieder not to use '' for quotes inside '' delimited code. It
ended up introducing a file called <paths> to the main git repository.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
* jc/maint-follow-rename-fix:
log: test for regression introduced in v1.7.2-rc0~103^2~2
diff --follow: do call diffcore_std() as necessary
diff --follow: do not waste cycles while recursing
* cc/find-commit-subject:
blame: use find_commit_subject() instead of custom code
merge-recursive: use find_commit_subject() instead of custom code
bisect: use find_commit_subject() instead of custom code
revert: rename variables related to subject in get_message()
revert: refactor code to find commit subject in find_commit_subject()
revert: fix off by one read when searching the end of a commit subject
cvs tests: do not touch test CVS repositories shipped with source
Some tests in t96xx series (cvsimport) want to write into the control area
(CVSROOT) of their test CVS repositories, but this does not work well when
the source area is made read-only (test trash directories are moved via
--root=else/where option).
Copy the supplied test CVS repository to a scratch place at the beginning
of these tests.
t/t9602-cvsimport-branches-tags.sh: Add a PERL prerequisite
Change this test to declare a PERL prerequisite. These tests use the
-p switch, so they implicitly depend on Perl code, but nothing was
declaring this.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
t/t9601-cvsimport-vendor-branch.sh: Add a PERL prerequisite
Change this test to declare a PERL prerequisite. These tests use the
-p switch, so they implicitly depend on Perl code, but nothing was
declaring this.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Change this test to declare a PERL prerequisite. These tests use the
-p switch, so they implicitly depend on Perl code, but nothing was
declaring this.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
t/t9001-send-email.sh: convert setup code to tests
Change the setup code in t/t9001-send-email.sh to use
test_expect_success. This way it isn't needlessly run in environments
where the test prerequisites aren't met.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
t/t9001-send-email.sh: change from skip_all=* to prereq skip
Change this test to skip test with test prerequisites, and to do setup
work in tests. This improves the skipped statistics on platforms where
the test isn't run.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Remove the PROG=* assignment from t9001-send-email.sh. It's been there
since v1.4.0-rc1~30 when the test was originally added, but only tests
that source annotate-tests.sh need it, it was seemingly introduced to
this test via copy/paste coding.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
t/t9600-cvsimport.sh: change from skip_all=* to prereq skip
Change this test to skip test with test prerequisites, and to do setup
work in tests. This improves the skipped statistics on platforms where
the test isn't run.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
lib-patch-mode tests: change from skip_all=* to prereq skip
Change this test to skip test with test prerequisites, and to do setup
work in tests. This improves the skipped statistics on platforms where
the test isn't run.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
t/t3701-add-interactive.sh: change from skip_all=* to prereq skip
Change this test to skip test with test prerequisites, and to do setup
work in tests. This improves the skipped statistics on platforms where
the test isn't run.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
t/Makefile: Create test-results dir for smoke target
Change the smoke target to create a test-results directory. This was
done implicitly by the test-lib before my "test-lib: Don't write
test-results when HARNESS_ACTIVE" patch, but after that smoking from
the pu branch hasn't worked.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
The git-notes expensive timing test is only expensive because it
either did 10,100,1k and 10k iterations or nothing.
Change it to do 10 by default, with an option to run the expensive
version with the old GIT_NOTES_TIMING_TESTS=ZomgYesPlease variable.
Since nobody was ostensibly running this test under TAP the code had
bitrotted so that it emitted invalid TAP. This change fixes that.
The old version would also mysteriously fail on systems without
/usr/bin/time, there's now a check for that using the multiple test
prerequisite facility.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
t/t3300-funny-names: change from skip_all=* to prereq skip
Change this test to skip test with test prerequisites, and to do setup
work in tests. This improves the skipped statistics on platforms where
the test isn't run.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
t/t3902-quoted: change from skip_all=* to prereq skip
Change this test to skip test with test prerequisites, and to do setup
work in tests. This improves the skipped statistics on platforms where
the test isn't run.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
t/t4016-diff-quote: change from skip_all=* to prereq skip
Change this test to skip test with test prerequisites, and to do setup
work in tests. This improves the skipped statistics on platforms where
the test isn't run.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
t/t5503-tagfollow: change from skip_all=* to prereq skip
Change this test to skip test with test prerequisites, and to do setup
work in tests. This improves the skipped statistics on platforms where
the test isn't run.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
t/t7005-editor: change from skip_all=* to prereq skip
Change this test to skip test with test prerequisites, and to do setup
work in tests. This improves the skipped statistics on platforms where
the test isn't run.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
t/t5705-clone-2gb: change from skip_all=* to prereq skip
Change this test to skip test with test prerequisites, and to do setup
work in tests. This improves the skipped statistics on platforms where
the test isn't run.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
t/t1304-default-acl: change from skip_all=* to prereq skip
Change this test to skip test with test prerequisites, and to do setup
work in tests. This improves the skipped statistics on platforms where
the test isn't run.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
t/README: Update "Skipping tests" to align with best practices
The example I initially added to "Skipping tests" wasn't very
good. We'd rather skip tests using the three-arg prereq form to the
test_* functions, not bail out with a skip message.
Change the documentation to reflect that, but retain the bailout
example under a disclaimer which explains that it's probably not a
good idea to use it.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
t/t4004-diff-rename-symlink.sh: use three-arg <prereq>
Change the tests that skipped due to unavailable SYMLINKS support to
use the three-arg prereq form of test_expect_success.
This is like the "tests: implicitly skip SYMLINKS tests using
<prereq>" change, but I needed to create an additional test for some
setup code. It's in a separate change as suggested by Jonathan Nieder
for ease of reviewing.
> +++ b/t/t4004-diff-rename-symlink.sh
> @@ -40,8 +34,9 @@ test_expect_success \
> # rezrov and nitfol are rename/copy of frotz and bozbar should be
> # a new creation.
>
> -GIT_DIFF_OPTS=--unified=0 git diff-index -M -p $tree >current
> -cat >expected <<\EOF
> +test_expect_success SYMLINKS 'setup diff output' "
> + GIT_DIFF_OPTS=--unified=0 git diff-index -M -p $tree >current
> + cat >expected <<\EOF
> diff --git a/bozbar b/bozbar
> new file mode 120000
> --- /dev/null
Probably belongs in a separate patch. More importantly, it is missing
&&-chaining (not a regression, but it is best to set a good example).
Reviewed-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
t/README: Add a note about the dangers of coverage chasing
Having no coverage at all is almost always a bad sign, but trying to
attain 100% coverage everywhere is usually a waste of time. Add a
paragraph to explain this to future test writers.
Inspired-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Reviewed-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Document how test writers can generate coverage reports, to ensure
that their tests are really testing the code they think they're
testing.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Reviewed-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add a target to generate a detailed HTML report for the entire Git
codebase using Devel::Cover's cover(1) tool. Output it in
cover_db_html instead of the default cover_db, so that it isn't mixed
up with our raw report files.
The target depends on the coverage-report-cover-db target, it may be
run redundantly if it was previously run. But the HTML output won't be
affected by running gcov2perl twice, so I didn't try to avoid that
small redundancy.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Reviewed-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add a target to convert the *.gcov files to a Devel::Cover
database. That database can subsequently be formatted by the cover(1)
tool which is included with Devel::Cover.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Reviewed-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Change the coverage-report target so that it doesn't generate the
coverage-untested-functions file by default. I'm adding more targets
for doing various things with the gcov files, and they shouldn't all
run by default.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Reviewed-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Makefile: Include subdirectories in "make cover" reports
We generate profiling files in all the $(OBJECTS) dirs. Aggregate
results from there, and add them to the corresponding clean target.
Also expand the gcov arguments. Generate reports for things like "x()
|| y()" using --all-blocks, and add --preserve-paths since we're
profiling in subdirectories now.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Reviewed-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
gitignore: Ignore files generated by "make coverage"
The "make coverage" support added by Thomas Rast in 901c369af5 didn't
contain a corresponding patch to patch .gitignore.
Change gitignore to ignore the *.gcda, *.gcno and *.gcov files
generated by GCC and our coverage invocations.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Reviewed-by: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
t/README: Add SMOKE_{COMMENT,TAGS}= to smoke_report target
The smoke server supports a free form text field with comments about a
report, and a comma delimited list of tags. Change the smoke_report
target to expose this functionality. Now smokers can send more data
that explains and categorizes the reports they're submitting.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
t/Makefile: Can't include GIT-BUILD-OPTIONS, it's a .sh
Change the smoke testing portion of t/Makefile not to include
GIT-BUILD-OPTIONS. It's a shellscript, not a Makefile snippet, so it
had the nasty side-effect of sneaking e.g. SHELL_PATH = '/bin/sh'
(with quotes) everywhere.
Just add our own PERL_PATH variable as a workaround. The t/Makefile
already has e.g. an equivalent SHELL_PATH and TAR option which
duplicate the definitions in GIT-BUILD-OPTIONS.
Reported-by: Thomas Rast <trast@student.ethz.ch> Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Add the capability to send smoke reports from the Git test suite.
Currently we only notice bugs in the test suite when it's run
manually. Bugs in Git that only occur on obscure platforms or setups
that the core developers aren't using can thus go unnoticed.
This series aims to change that. With it, anyone that's interested in
avoiding bitrot in Git can volunteer to run a smoke tester. A smoke
tester periodically compiles the latest version of Git, runs the test
suite, and submits a report to a central server indicating how the
test run went.
A smoke tester might run something like this in cron:
#!/bin/sh
cd ~/g/git
git fetch
for branch in maint master next pu; do
git checkout origin/$i &&
make clean all &&
cd t &&
make smoke_report
done
The smoker might want to compile git with non-default flags, include
bisecting functionality or run the tests under valgrind. Doing that is
outside the scope of this patch, this just adds a report submission
mechanism. But including a canonical smoke runner is something we'll
want to include eventually.
What this does now is add smoke and smoke_report targets to t/Makefile
(this example only uses a few tests for demonstration):
$ make clean smoke
rm -f -r 'trash directory'.* test-results
rm -f t????/cvsroot/CVSROOT/?*
rm -f -r valgrind/bin
rm -f .prove
perl ./harness --git-version="1.7.2.1.173.gc9b40" \
--no-verbose \
--archive="test-results/git-smoke.tar.gz" \
t0000-basic.sh t0001-init.sh t0002-gitfile.sh t0003-attributes.sh t0004-unwritable.sh t0005-signals.sh t0006-date.sh
t0000-basic.sh ....... ok
t0001-init.sh ........ ok
t0002-gitfile.sh ..... ok
t0003-attributes.sh .. ok
t0004-unwritable.sh .. ok
t0005-signals.sh ..... ok
t0006-date.sh ........ ok
All tests successful.
TAP Archive created at /home/avar/g/git/t/test-results/git-smoke.tar.gz
The smoke target uses TAP::Harness::Archive to aggregate the test
results into a tarball. The tarball contains two things, the output of
every test file that was run, and a metadata file:
The "extra_properties" hash is where we'll stick Git-specific info,
like whether Git was compiled with gettext or the fallback regex
engine, and what branch we're compiling. Currently no metadata like
this is included.
The entire tarball is then submitted to a central smokebox at
smoke.git.nix.is. This is done with curl(1) via the "smoke_report"
target:
$ make smoke_report
curl \
-H "Expect: " \
-F project=Git \
-F architecture=x86_64 \
-F platform=Linux \
-F revision="1.7.2.1.173.gc9b40" \
-F report_file=@test-results/git-smoke.tar.gz \
http://smoke.git.nix.is/app/projects/process_add_report/1 \
| grep -v ^Redirecting
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 117k 100 63 100 117k 3 6430 0:00:21 0:00:18 0:00:03 0
Reported #8 added.
Reports are then made available on the smokebox via a web interface:
test-lib: Multi-prereq support only checked the last prereq
The support for multiple test prerequisites added by me in "test-lib:
Add support for multiple test prerequisites" was broken.
The for iterated over each prerequisite and returned true/false within
a case statement, but since it missed a return statement only the last
prerequisite in the list of prerequisites was ever considered, the
rest were ignored.
Fix that by changing the test_have_prereq code to something less
clever that keeps a count of the total prereqs and the ones we have
and compares the count at the end.
This comes with the added advantage that it's easy to list the missing
prerequisites in the test output, implement that while I'm at it.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
tests: A SANITY test prereq for testing if we're root
Some tests depend on not being able to write to files after chmod
-w. This doesn't work when running the tests as root.
Change test-lib.sh to test if this works, and if so it sets a new
SANITY test prerequisite. The tests that use this previously failed
when run under root.
There was already a test for this in t3600-rm.sh, added by Junio C
Hamano in 2283645 in 2006. That check now uses the new SANITY
prerequisite.
Some of this was resurrected from the "Tests in Cygwin" thread in May
2009:
t/README: Document the predefined test prerequisites
The README for the test library suggested that you grep the
test-lib.sh for test_set_prereq to see what the preset prerequisites
were.
Remove that bit, and write a section explaining all the preset
prerequisites. Most of the text was lifted from from Junio C Hamano
and Johannes Sixt, See the "Tests in Cygwin" thread in May 2009 for
the originals:
test-lib: Print missing prerequisites in test output
Change the test output to print needed prerequisites as part of the
TAP. This makes it easy to see at a glance why a test was
skipped. Before:
ok 7 # skip <message>
ok 9 # skip <message>
After:
ok 7 # skip <message> (prereqs: DONTHAVEIT)
ok 9 # skip <message> (prereqs: HAVEIT,DONTHAVEIT)
This'll also be useful for smoke testing output, where the developer
reading the output may not be familiar with the system where tests are
being skipped.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
test-lib: Add support for multiple test prerequisites
Change the test_have_prereq function in test-lib.sh to support a
comma-separated list of prerequisites. This is useful for tests that
need e.g. both POSIXPERM and SANITY.
The implementation was stolen from Junio C Hamano and Johannes Sixt,
the tests and documentation were not. See the "Tests in Cygwin" thread
in May 2009 for the originals:
test-lib: Don't write test-results when HARNESS_ACTIVE
TAP harnesses don't need to read test-results/*, since they keep track
of the number of passing/failing tests internally. Skip the generation
of these files when HARNESS_ACTIVE is set.
It's now possible to run the Git test suite without writing anything
to the t/ directory at all if you use a TAP harness and the --root
switch:
cd t
sudo mount -t tmpfs none /tmp/memory -o size=300m
prove -j9 ./t[0-9]*.sh :: --root=/tmp/memory
The I/O that the ~500 test-results/* files contributed was very
minimal, but I thought this was worth mentioning.
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
* jn/maint-plug-leak:
write-tree: Avoid leak when index refers to an invalid object
read-tree: stop leaking tree objects
core: Stop leaking ondisk_cache_entrys
* jl/submodule-ignore-diff:
Add tests for the diff.ignoreSubmodules config option
Add the 'diff.ignoreSubmodules' config setting
Submodules: Use "ignore" settings from .gitmodules too for diff and status
Submodules: Add the new "ignore" config option for diff and status