This project forked from hdp/git-wrapper
Perl wrapper for git(7) command line interface
Home Page: http://genehack.github.com/Git-Wrapper
Here are some other places you can find me on the web:
This makes using the --not flag to git rev-list impossible. There are likely other options this matters for as well.
Output from '/usr/local/bin/make test':
PERL_DL_NONLAZY=1 /export/home/cpant4/perl5/bin/perl5.15.0 "-MExtUtils::Command::MM" "-e" "test_harness(0, 'blib/lib', 'blib/arch')" t/*.t
# Testing Git::Wrapper 0.015
t/00-load.t ............. ok
open3: exec of git init failed at /export/home/cpant4/.cpan/build/Git-Wrapper-0.015-uyrDin/blib/lib/Git/Wrapper.pm line 68
t/basic.t ...............
Dubious, test returned 2 (wstat 512, 0x200)
No subtests run
t/pod-coverage.t ........ skipped: Test::Pod::Coverage 1.04 required for testing POD coverage
t/pod.t ................. skipped: Test::Pod 1.14 required for testing POD
t/release-pod-syntax.t .. skipped: these tests are for release candidate testing
Test Summary Report
-------------------
t/basic.t (Wstat: 512 Tests: 0 Failed: 0)
Non-zero exit status: 2
Parse errors: No plan found in TAP output
Files=5, Tests=1, 1 wallclock secs ( 0.02 usr 0.01 sys + 0.12 cusr 0.02 csys = 0.17 CPU)
Result: FAIL
Failed 1/5 test programs. 0/1 subtests failed.
make: *** [test_dynamic] Error 255
If $HOME is set to an invalid value on a Win32 system, the initial $git->init() call in t/basic.t fails (due to the underlying git init call not being able to load the global gitconfig file, because it's looking for it in at a path that isn't valid on Win32 systems).
One suggested option is to check for this in Makefile.PL and refuse to install if $HOME is set but the directory doesn't exist. I'm not sure that's the right solution.
An alternative is to check that the $git->init() call in t/basic.t succeeds, and issue a BAILOUT if it doesn't. This seems like a better solution in that it's more generic (and because if that first git init doesn't work the test suite is pretty well fucked for going forward).
See https://rt.cpan.org/Ticket/Display.html?id=64665 for gory details.
14:28 < frew> genehack: look at that commit I just pasted; that stuff is stuff that Git::Wrapper doesn't support that apparently wouldn't be hard to add 14:29 < frew> genehack: https://github.com/frioux/dist-zilla-plugin-git/blob/427260770320c48fcdf71394c794405c3f6a027c/lib/Dist/Zilla/Plugin/Git/CommitBuild.pm#L113 14:29 < dipsy> [ dist-zilla-plugin-git/lib/Dist/Zilla/Plugin/Git/CommitBuild.pm at 427260770320c48fcdf71394c794405c3f6a027c · frioux/dist-zilla-plugin-git · GitHub ] 14:29 < frew> genehack: that may be more helpful
The tests fail during install with:
# Testing git version: 2.9.2
# Failed test 'message'
# at t/basic.t line 77.
# got: 'FIRST
#
# BODY
# '
# expected: 'FIRST
#
# BODY
# '
# Looks like you failed 1 test of 26.
t/basic.t .................
Dubious, test returned 1 (wstat 256, 0x100)
Ubuntu 14.4, but I rather suspect the git version might be to blame. I checked some green CPAN testers output and it was using 1.8.3, e.g.
http://www.cpantesters.org/cpan/report/7f455f80-c654-11e4-b221-9e126cbd7f71
perl -v
This is perl 5, version 22, subversion 0 (v5.22.0) built for x86_64-linux
git --version
git version 2.9.2
If user have ~/.gitconfig with:
[commit]
gpgsign = true
then tests fail with:
t/basic.t .. # Testing git version: 2.10.2
ok 1
ok 2
ok 3 - repo is dirty
error: gpg failed to sign the data
fatal: failed to write commit object
# Tests were run but no plan was declared and done_testing() was not seen.
# Looks like your test exited with 128 just after 3.
Looks like you need to force use of own global config in tests.
echo -ne in sh on MacOS doesn't take args. So the CommitBuild doesn't work and the git mktree give format error because the "-ne" is send as a simple display.
A better way to do is simply use "open" command. But I have quick fix by surrouding the command by "bash -c 'cmd'" ...
so $cmd = qq{bash -c 'echo -ne "$cmd" | git mktree'};
See CPAN test results: http://www.cpantesters.org/distro/G/Git-Wrapper.html#Git-Wrapper-0.023
This code:
use Git::Wrapper;
my $git = Git::Wrapper->new($repo_path_here);
$git->log('--oneline','-1');
produces this:
$ ./test.pl
unhandled: $SHA $MSG at /opt/perl-5.14.2/lib/site_perl/5.14.2/Git/Wrapper.pm line 164.
when running git log --oneline -1 produces just:
$SHA $MSG
(Reported by @perigrin on IRC.)
There are just a couple and seem easy to hit.
Git::Wrapper fails horribly if the filename contains spaces. Yet such filenames are legal in a git repository.
Fixing this bug properly - so that all possible filenames are handled correctly - could be difficult. But it would be an improvement to handle just space characters, perhaps by using quotation marks in the git command line.
With git 1.7.0.4 the test basic.t fails with the following error:
t/basic.t .................. 21/? #
error: option `color' takes no value
usage: git branch [options] [-r | -a] [--merged | --no-merged]
or: git branch [options] [-l] [-f] <branchname> [<start-point>]
or: git branch [options] [-r] (-d | -D) <branchname>
or: git branch [options] (-m | -M) [<oldbranch>] <newbranch>
Generic options
-v, --verbose be verbose
-t, --track set up tracking mode (see git-pull(1))
--set-upstream change upstream info
--color use colored output
-r act on remote-tracking branches
--contains <commit> print only branches that contain the commit
--abbrev[=<n>] use <n> digits to display SHA-1s
Specific git-branch actions:
-a list both remote-tracking and local branches
-d delete fully merged branch
-D delete branch (even if not merged)
-m move/rename a branch and its reflog
-M move/rename a branch, even if target exists
-l create the branch's reflog
-f, --force force creation (when already exists)
--no-merged <commit> print only not merged branches
--merged <commit> print only merged branches
# Tests were run but no plan was declared and done_testing() was not seen.
This error is caused by the branch method in Git/Wrapper.pm . In older versions of Git the param --color takes no argument, is either --color or --no-color.
git branch [--color | --no-color] [-r | -a]
This is different in newer versions of Git
git branch [--color[=] | --no-color] [-r | -a]
It occurred to me that the lines returned by Git::Wrapper are octects, not characters. I understand that this is undertested, given that most of us write commit messages in English/ASCII only, however, looking at the source, it seems that no attempt to decode the strings is done.
This is not a bug per se, but I'd suggest to formalize this in the documentation, so people using the module are aware of what kind of strings to expect.
because meh.
this makes me a sad panda: http://matrix.cpantesters.org/?dist=Git-Wrapper+0.031
We seem to be losing a lot of test coverage, which has cascading effects on the distributions that use this module, because of a few small booboos in the tests. Would it be possible to detect the presence of git earlier, say in Makefile.PL, so a NA result could be returned instead of FAIL?
Name: Path::Class::File
Version: 0.25
Directory: /usr/share/perl5/vendor_perl
File: /usr/share/perl5/vendor_perl/Path/Class/File.pm
Core module: no
I had this version installed and a test was failing with ->spew missing, it must have been added between 0.25 and 0.32
The t/basic.t tests is failing on 0.046 for git versions 1.7.1 and 1.8.2.3. Here are the logs that I have:
# Testing git version: 1.7.1
# Failed test 'expected copy score'
# at t/basic.t line 150.
# got: undef
# expected: '100'
# Failed test 'expected copy type'
# at t/basic.t line 151.
# got: 'A'
# expected: 'C'
# Failed test 'using RUN('log','--format=%H') to get all four commit SHAs'
# at t/basic.t line 201.
# Looks like you failed 3 tests of 34.
t/basic.t .......
Dubious, test returned 3 (wstat 768, 0x300)
Failed 3/34 subtests
(less 2 skipped subtests: 29 okay)
# Testing git version: 1.8.2.3
ok 1
ok 2
ok 3 - repo is dirty
ok 4 - repo is clean
ok 5
ok 6
ok 7 - one log entry
ok 8 - id
ok 9 - message
ok 10 - threw Git::Wrapper::Exception
ok 11 - date
ok 12 - id
ok 13 - log(--oneline) dies
ok 14 - RUN(log --oneline) lives
ok 15 - one log entry
ok 16 - one raw log entry
ok 17 - expected raw modification object
ok 18 - still one raw log entry
ok 19 - expected score
ok 20 - expected type
ok 21 - expected filename
ok 22 - one for the third time
ok 23 - expected copy score
ok 24 - expected copy type
ok 25 - expected copy filename
ok 26 - Attempt to commit interactively fails quickly
ok 27 # skip : .git/COMMIT_EDITMSG: Permission denied
# error: T
not ok 28 - using RUN('log','--format=%H') to get all four commit SHAs
# Failed test 'using RUN('log','--format=%H') to get all four commit SHAs'
# at t/basic.t line 201.
ok 29 - argument test: long arg, no spaces in val
ok 30 - argument test: long arg, spaces in val
ok 31 - argument test: short arg, no spaces in val
ok 32 - argument test: short arg, spaces in val
ok 33 - new branch name is correct
ok 34 - passing content with -STDIN option
1..34
# Looks like you failed 1 test of 34.
Dubious, test returned 1 (wstat 256, 0x100)
Failed 1/34 subtests
(less 1 skipped subtest: 32 okay)
Test Summary Report
-------------------
t/basic.t (Wstat: 256 Tests: 34 Failed: 1)
Failed test: 28
Non-zero exit status: 1
Files=1, Tests=34, 1 wallclock secs ( 0.03 usr 0.00 sys + 0.17 cusr 0.23 csys = 0.43 CPU)
Result: FAIL
It would be handy for log entries to have a filenames method which returns a list of files changed. This would involve changing the git log call to a different format such as git log --stat and parsing it accordingly.
Update docs to show how these work.
This Git::Wrapper newbie spent time untangling two Git::Wrapper behaviours to understand why git->clone wasn't doing what he expected.
Number 1: I wasn't sure how to add non-option parameters, i.e., a parameter on the command line that is not associated with an --option. An example is the repo url in git clone $repo_url. I blundered around trying every combination of hashs and lists before stumbling across #97 which made it clear in an example. The cpan/module documentation only covers options added with or without values. Perhaps you could help newbies out by explicitly explaining there how to add a non-option parameter.
These teeting problems aside, thanks for a great module.
t/basic.t skips creating a second commit if the installed git is so old that it does not support --allow-empty-message, but then it always expects to find 2 commits.
Simple fix:
@@ -143,6 +143,7 @@ sub _timeout (&) {
return $timeout;
}
+my $commit_count=1;
SKIP: {
if ( versioncmp( $git->version , '1.7.0.5') eq -1 ) {
skip 'testing old git without commit --allow-empty-message support' , 1;
@@ -168,10 +169,11 @@ SKIP: {
@log = $git->log();
is(@log, 2, 'two log entries, one with empty commit message');
+ ++$commit_count;
};
my @out = $git->RUN('log','--format=%H');
-ok scalar @out == 2, q{using RUN('log','--format=%H') to get all 2 commit SHAs}
+ok scalar @out == $commit_count, qq{using RUN('log','--format=%H') to get all $
# test --message vs. -m
my @arg_tests = (Hi, sorry I was not able to find a better way to get in touch. I maintain the AUR package at https://aur.archlinux.org/packages/perl-git-wrapper for Arch Linux. I am using a tarball from cpan.org, but I am curious if this is the better "source of truth" for the source code, as I see there was a recent PR accepted.
Thank you.
My git config includes this:
[color]
diff = auto
status = auto
branch = auto
interactive = auto
ui = true
And this:
perl -MGit::Wrapper -wle'print Git::Wrapper->new(".")->branch'
...has "master" in green.
One side effect of this is that any code that compares branch strings literally (for example, [Git::CheckFor::CorrectBranch]) will fail.
All git commands need to be run with colors turned off, or alternatively the output needs to be cleaned up before returning it.
Below is the error stack from stage 'make test':
[MSG] [Sun Jul 17 10:46:21 2011] cp lib/Git/Wrapper.pm blib/lib/Git/Wrapper.pm
Manifying blib/man3/Git::Wrapper.3
[ERROR] [Sun Jul 17 10:46:35 2011] MAKE TEST failed: PERL_DL_NONLAZY=1 /Volumes/Media/smoke/bare/perl-5.12.0/bin/perl "-MExtUtils::Command::MM" "-e" "test_harness(0, 'blib/lib', 'blib/arch')" t/*.t
# Testing Git::Wrapper 0.015
t/00-load.t ............. ok
error: unknown option `porcelain'
usage: git status [options] [--] <filepattern>...
-q, --quiet be quiet
-v, --verbose be verbose
Commit message options
-F, --file <FILE> read log from file
--author <AUTHOR> override author for commit
-m, --message <MESSAGE>
specify commit message
-c, --reedit-message <COMMIT>
reuse and edit message from specified commit
-C, --reuse-message <COMMIT>
reuse message from specified commit
-s, --signoff add Signed-off-by:
-t, --template <FILE>
use specified template file
-e, --edit force edit of commit
Commit contents options
-a, --all commit all changed files
-i, --include add specified files to index for commit
--interactive interactively add files
-o, --only commit only specified files
-n, --no-verify bypass pre-commit hook
--amend amend previous commit
-u, --untracked-files[=<mode>]
show untracked files, optional modes: all, normal, no. (Default: all)
--allow-empty ok to record an empty change
--cleanup <default> how to strip spaces and #comments from message
# Tests were run but no plan was declared and done_testing() was not seen.
t/basic.t ...............
Dubious, test returned 129 (wstat 33024, 0x8100)
All 2 subtests passed
t/pod-coverage.t ........ skipped: Test::Pod::Coverage 1.04 required for testing POD coverage
t/pod.t ................. skipped: Test::Pod 1.14 required for testing POD
t/release-pod-syntax.t .. skipped: these tests are for release candidate testing
Test Summary Report
-------------------
t/basic.t (Wstat: 33024 Tests: 2 Failed: 0)
Non-zero exit status: 129
Parse errors: No plan found in TAP output
Files=5, Tests=3, 11 wallclock secs ( 0.57 usr 0.76 sys + 2.54 cusr 6.35 csys = 10.22 CPU)
Result: FAIL
Failed 1/5 test programs. 0/3 subtests failed.
make: *** [test_dynamic] Error 255
How would you feel about a way to have Git::Wrapper only print the command it would run, as opposed to actually doing it?
My use case: I'm working on a packaging tool at $work (currently implemented as a moderately thin wrapper around Dist::Zilla). At certain points, I want the tool to commit things to git automatically. Naturally, I want my tests to be able to verify that these commits is happening. There are of course a few options for doing this:
Git::Wrapper itself is tested.) This doesn't work that well for me for a few reasons.git stash to make sure I don't blow anything away I can't get back, which means that every time I run my test suite, I end up littering my repo with stashes.Git::Wrapper that says, you know what, in this instance, don't really do the commit, just tell me what git command you're about to run (print it to STDERR or somesuch) and then I can check for that message and feel confident that the commit really would have taken place.Obviously I'm angling for option # 3 here. :-)
Assuming you are amenable to such an option, what would be your preference for how to activate it? I see, again, 3 decent options:
Test::Git::Wrapper) [not my favorite--most intrusive]$Git::Wrapper::TEST_MODE = 1)$ENV{GIT_WRAPPER_TEST_MODE} = 1) [least intrusive]If this sounds good to you, pick an option and let me know and I'll put together a pull request. From looking at the _cmd method, it should be pretty easy.
Run perlcritic, fix easy and reasonable complaints; create additional Issues for the more invasive/questionable objections by perlcritic.
@toddr, any ideas? I'm already chomp()ing the output and that doesn't seem to be making a difference. Also seems very specific to this one CPAN tester.
# Failed test 'Wrapper runs what ever binary we tell it to'
# at t/git_binary.t line 19.
# Structures begin differing at:
# $got->[0] = 'marco
'
# $expected->[0] = 'marco'
# Looks like you failed 1 test of 5.
t/git_binary.t .............
Dubious, test returned 1 (wstat 256, 0x100)
Failed 1/5 subtests
Failing test report -> http://www.cpantesters.org/cpan/report/ddf3e31a-6ce9-1014-bac9-c776d871b27e
In Debian we are currently applying the following patch to
Git-Wrapper.
We thought you might be interested in it too.
Description: spelling fix
Origin: vendor
Author: gregor herrmann <[email protected]>
Last-Update: 2016-08-19
The patch is tracked in our Git repository at
https://anonscm.debian.org/cgit/pkg-perl/packages/libgit-wrapper-perl.git/plain/debian/patches/spelling.patch
Thanks for considering,
gregor herrmann,
Debian Perl Group
Replace it with File::chdir or something, it's buggy on Windows
See, e.g., http://www.cpantesters.org/cpan/report/62c988bc-8f3a-1014-8acc-31012b825c07
I couldn't find this in the documentation, sorry.
How do I run something like this:
git pull --rebase=preserve origin master
In this case, I can handle the rebase=preserve because it's key/value, but the other two parameters, which I tend to use when wanting to pull data from only one ref, I have no idea how to write.
It's a very minor thing, but I can't simulate git log -n1 with Git::Wrapper. That is, if I do this:
my ($last_log) = $git->log({ max_count => 1 });
I get:
fatal: Option '--max-count' requires a value
I see what's going on here--your code sees a value of "1", decides it's a boolean switch, and buiilds the command line appropriately. But I'm not sure how to fix it. I even tried:
my ($last_log) = $git->log({ max_count => 1.0 });
thinking that would squeak me by the eq '1' test, but that didn't work either. Happily, the workaround is simple enough:
my ($last_log) = $git->log({ max_count => 2 });
which works fine. Technically, it's wasteful, since we're retrieving a log that we're just throwing away, but I sincerely doubt that's ever going to make any measurable performance difference.
Anyway, if you have a suggestion on how to fix it, I'd be glad to patch it, but I don't really see a clean way around it ATM. Maybe just give up and do a doco patch warning people this could be a problem ... ?
A dirty trick would be to make $Git::Wrapper::git an our variable.
Ideally, it would be nice if this could be passed via new.
The hash-object test added by 8f52f8c fails -- need to figure out when the 'no-filters' option was added to the hash_object subcmd and skip the test on git versions earlier than that, the same way we skip other stuff in a version based way (i.e., add a supports_hash_object_filters method a la supports_log_raw_dates and supports_status_porcelain).
Here's the error report from CPAN testers, 1.5.4.3 clearly fails but some Git changelog digging is in order.
Below is the error stack from stage 'make test':
PERL_DL_NONLAZY=1 /export/home/cpan/pit/thr/perl-5.12.2/bin/perl "-MExtUtils::Command::MM" "-e" "test_harness(0, 'blib/lib', 'blib/arch')" t/*.t
# Testing Git::Wrapper 0.022
t/00-load.t ............. ok
t/author-err.t .......... skipped: these tests are for testing by the author
# Testing git version: 1.5.4.3
usage: git-hash-object [-t <type>] [-w] [--stdin] <file>...
# Tests were run but no plan was declared and done_testing() was not seen.
t/basic.t ...............
Dubious, test returned 129 (wstat 33024, 0x8100)
All 21 subtests passed
(less 7 skipped subtests: 14 okay)
t/pod-coverage.t ........ skipped: Test::Pod::Coverage 1.04 required for testing POD coverage
t/pod.t ................. skipped: Test::Pod 1.14 required for testing POD
t/release-pod-syntax.t .. skipped: these tests are for release candidate testing
Test Summary Report
-------------------
t/basic.t (Wstat: 33024 Tests: 21 Failed: 0)
Non-zero exit status: 129
Parse errors: No plan found in TAP output
Files=6, Tests=27, 6 wallclock secs ( 0.08 usr 0.06 sys + 0.47 cusr 0.84 csys = 1.45 CPU)
Result: FAIL
Failed 1/6 test programs. 0/27 subtests failed.
make: *** [test_dynamic] Error 255
here's an example test: http://www.cpantesters.org/cpan/report/12ed2af8-6bf4-1014-8695-9e662ad63fef
the test: https://metacpan.org/source/ETHER/Dist-Zilla-Plugin-Git-Contributors-0.004/t/01-basic.t#L50
and the prep work done for MSWin32: https://metacpan.org/source/ETHER/Dist-Zilla-Plugin-Git-Contributors-0.004/t/lib/GitSetup.pm#L37
sadly, this doesn't seem to be enough. But I'm not sure what's missing.
Was added in 1.7.6, meaning that forcing no-abbrev-commit on breaks not only the test suite, but also the tool.
Looks like two repos got their wires crossed - the entire history of App-GitGot is also in this repo.
Check this out: https://github.com/genehack/Git-Wrapper/tree/v1.333
The following thing happens on this line: https://github.com/genehack/Git-Wrapper/blob/master/t/basic.t#L225 I got the output by adding Devel::Confess and the STDOUT output to the error message.
$git->add('argument_testfile');
$git->commit({ $flag => $msg }); # here
ok 21 - argument test: long arg, no spaces in val
ok 22 - argument test: long arg, spaces in val
git exited non-zero but had no output to stderr, stdout: On branch master
Changes not staged for commit:
modified: argument_testfile
no changes added to commit
at lib/Git/Wrapper.pm line 136.
Git::Wrapper::RUN(Git::Wrapper=HASH(0x632fc54), "commit", HASH(0x63359b4)) called at lib/Git/Wrapper.pm line 65
Git::Wrapper::AUTOLOAD(Git::Wrapper=HASH(0x632fc54), HASH(0x63359b4)) called at D:\cpan\Git-Wrapper\t\basic.t line 228
# Tests were run but no plan was declared and done_testing() was not seen.
# Looks like your test exited with 1 just after 22.
What happens here is that the add returns with success because windows has assured it of success but the write to the drive is still pending, and the commit then ends up seeing the index pre-write and fails with a "nothing to do" message.
I'm not entirely sure how to fix this. One way might be to just put a sleep at the end, another might be to try and get the add to only return when absolutely done.
Your thoughts?
To reproduce this, please clone the git repository at [email protected]:petdance/ack2.git. (At the time the latest revision was 06d47c8f.) Then in that checkout run
git log --name-only IsPathGroup.pm
This prints a single revision which affects that file. I would therefore expect to get back that revision from the log method, and moreover the revision would have a single modification which is for that file. But:
perl -MGit::Wrapper -E '$g = new Git::Wrapper("."); @l = $g->log({ raw => 1, follow => 1, date => "iso" }, "--", "IsPathGroup.pm"); foreach (@l) { @m = $_->modifications; say scalar @m }'
You will see that 0 is printed - so the revision found has no modifications.
If you can't reproduce this I can send a tarball of the git clone directory.
Testing against a pretty plain vanilla git on perl 5.24.1-RC2.
I'm still investigating the exact nature of the white space differences.
t/basic.t ..
# Testing git version: 2.9.2
ok 1
ok 2
ok 3 - repo is dirty
ok 4 - repo is clean
ok 5
ok 6
ok 7 - one log entry
ok 8 - id
not ok 9 - message
# Failed test 'message'
# at t/basic.t line 77.
# got: 'FIRST
#
# BODY
# '
# expected: 'FIRST
#
# BODY
# '
ok 10 - threw Git::Wrapper::Exception
ok 11 - date
ok 12 - id
ok 13 - log(--oneline) dies
ok 14 - RUN(log --oneline) lives
ok 15 - one log entry
ok 16 - one raw log entry
ok 17
ok 18 - Attempt to commit interactively fails quickly
ok 19 - two log entries, one with empty commit message
ok 20 - using RUN('log','--format=%H') to get all 2 commit SHAs
ok 21 - argument test: long arg, no spaces in val
ok 22 - argument test: long arg, spaces in val
ok 23 - argument test: short arg, no spaces in val
ok 24 - argument test: short arg, spaces in val
ok 25 - new branch name is correct
ok 26 - passing content with -STDIN option
1..26
# Looks like you failed 1 test of 26.
Dubious, test returned 1 (wstat 256, 0x100)
Failed 1/26 subtests
Test Summary Report
-------------------
t/basic.t (Wstat: 256 Tests: 26 Failed: 1)
Failed test: 9
Non-zero exit status: 1
Files=1, Tests=26, 1 wallclock secs ( 0.03 usr 0.00 sys + 0.13 cusr 0.11 csys = 0.27 CPU)
Result: FAIL
Makefile.PL dies at line 14 when it runs git config --list on CentOS 6 and the user doesn't have a .gitconfig in their home directory directory. It dies with a misleading error message about "git seems broken; maybe check your HOME environment variable?"
If you run git config --list from the command line, it returns a value of 128 and the error string "fatal: error processing config file(s)" Git with CentOS 6.8 is 1.7.1 - and this appears to be an issue with this version of Git, but it would be nice to at least have a better error message when Git::Wrapper refuses to install.
Running with git 1.8.2.1 and perl 5.16.3
The test hangs while executing the following line:
$git->commit({ message => "", 'allow-empty-message' => 1 });
pstree shows that the underlying git process has run interactively:
$ pstree -p 32148
prove(32148)───perl(32152)───git(32185)───editor(32186)───{editor}(32187)
$ ps -o args= 32185
git commit --allow-empty-message --message=
I assume this is caused by a change of behaviour in git, in this version the only way I've found to create an empty commit message is to use --file=/dev/null, no variation of the --message switch triggers the expected behaviour.
Below is the error stack from stage 'make test':
[MSG] [Sun Jul 17 17:26:01 2011] cp lib/Git/Wrapper.pm blib/lib/Git/Wrapper.pm
Manifying blib/man3/Git::Wrapper.3
[ERROR] [Sun Jul 17 17:26:02 2011] MAKE TEST failed: PERL_DL_NONLAZY=1 /home/cpan/pit/bare/perl-5.8.9/bin/perl "-MExtUtils::Command::MM" "-e" "test_harness(0, 'blib/lib', 'blib/arch')" t/*.t
# Testing Git::Wrapper 0.015
t/00-load.t ............. ok
*** Please tell me who you are.
Run
git config --global user.email "[email protected]"
git config --global user.name "Your Name"
to set your account's default identity.
Omit --global to set the identity only in this repository.
fatal: empty ident <[email protected]> not allowed
t/basic.t ...............
Dubious, test returned 2 (wstat 512, 0x200)
All 3 subtests passed
# Tests were run but no plan was declared and done_testing() was not seen.
t/pod-coverage.t ........ skipped: Test::Pod::Coverage 1.04 required for testing POD coverage
t/pod.t ................. skipped: Test::Pod 1.14 required for testing POD
t/release-pod-syntax.t .. skipped: these tests are for release candidate testing
Test Summary Report
-------------------
t/basic.t (Wstat: 512 Tests: 3 Failed: 0)
Non-zero exit status: 2
Parse errors: No plan found in TAP output
Files=5, Tests=4, 1 wallclock secs ( 0.03 usr 0.01 sys + 0.13 cusr 0.20 csys = 0.37 CPU)
Result: FAIL
*** Error code 255
So, based on this:
15:39 (genehack) how about something like: ->clone( $repo , [ $path | %options ]);
15:39 * perigrin nods
15:39 (perigrin) that would work
15:39 (perigrin) $options{prefix} = /tmp
15:39 (genehack) i'm trying to think if there's anything in '%options' other than root_dir
15:39 (perigrin) would *certainly* solve my problem.
15:39 (genehack) or prefix
15:39 (perigrin) well there could be
15:39 (perigrin) bare => 1
15:39 (perigrin) would set --bare
15:39 (perigrin) etc.
15:40 (genehack) right, i need to review what options 'clone' can take
also add a 'humanish' filter to turn foo/bar/baz.git into baz; make that the default $path when it's not given (just like git clone)
I usually stringify my Path::Class objects when passing them to things that expect strings,
but I forgot, and got a bunch of failing test reports for smokers using the latest Git::Wrapper dev release:
Can't bless non-reference value at /home/src/perl/repoperls/installed-perls/perl/v5.17.8/165a/lib/site_perl/5.17.8/Git/Wrapper.pm line 49.
While the changes to new() are partly backward-compatible, they are not for any argument that previously stringified to a dir (instead the dir argument silently gets ignored).
I'm not sure how much code is out there that might pass instances of Path::Class::Dir (or any other similar classes) but you may wish to consider that.
A quick git-grep shows that Dist::Zilla::Plugin::Git passes objects extensively.
Would you consider swapping the new if logic to say if it's a HASH do the new style otherwise assume the first arg is a dir?
If you desire to deprecate the old behavior a warning could certainly be added.
What do you think?
I'd be happy to write up some tests if you'd like.
I couldn't figure out how to coerce ExtUtils::MakeMaker to output the "provides" section in the META.yml/json.
CPAN.pm: Building G/GE/GENEHACK/Git-Wrapper-0.030.tar.gz
Checking if your kit is complete...
Looks good
Writing Makefile for Git::Wrapper
Writing MYMETA.yml and MYMETA.json
cp lib/Git/Wrapper/File/RawModification.pm blib/lib/Git/Wrapper/File/RawModification.pm
cp lib/Git/Wrapper.pm blib/lib/Git/Wrapper.pm
cp lib/Git/Wrapper/Statuses.pm blib/lib/Git/Wrapper/Statuses.pm
cp lib/Git/Wrapper/Log.pm blib/lib/Git/Wrapper/Log.pm
cp lib/Git/Wrapper/Exception.pm blib/lib/Git/Wrapper/Exception.pm
cp lib/Git/Wrapper/Status.pm blib/lib/Git/Wrapper/Status.pm
Manifying blib/man3/Git::Wrapper::Exception.3
Manifying blib/man3/Git::Wrapper::Log.3
Manifying blib/man3/Git::Wrapper.3
Manifying blib/man3/Git::Wrapper::File::RawModification.3
Manifying blib/man3/Git::Wrapper::Statuses.3
Manifying blib/man3/Git::Wrapper::Status.3
GENEHACK/Git-Wrapper-0.030.tar.gz
/usr/bin/make -- OK
Running make test
PERL_DL_NONLAZY=1 /usr/local/bin/perl "-MExtUtils::Command::MM" "-e" "test_harness(0, 'blib/lib', 'blib/arch')" t/*.t
t/00-load.t ................ 6/6 # Testing Git::Wrapper 0.030
t/00-load.t ................ ok
t/000-requires_external.t .. ok
t/author-err.t ............. skipped: these tests are for testing by the author
t/basic.t .................. # Testing git version: 1.6.3.2
t/basic.t .................. ok
t/git_binary.t ............. ok
t/parse_args.t ............. ok
t/path_class.t ............. # Testing git version: 1.6.3.2
t/path_class.t ............. 2/? fatal: '/tmp/FRA4lgN3mK/baz' is outside repository
# Tests were run but no plan was declared and done_testing() was not seen.
t/path_class.t ............. Dubious, test returned 128 (wstat 32768, 0x8000)
All 2 subtests passed
t/release-pod-coverage.t ... skipped: these tests are for release candidate testing
t/release-pod-syntax.t ..... skipped: these tests are for release candidate testing
Test Summary Report
-------------------
t/path_class.t (Wstat: 32768 Tests: 2 Failed: 0)
Non-zero exit status: 128
Parse errors: No plan found in TAP output
Files=9, Tests=62, 21 wallclock secs ( 0.24 usr 0.14 sys + 3.21 cusr 2.91 csys = 6.50 CPU)
Result: FAIL
Failed 1/9 test programs. 0/62 subtests failed.
make: *** [test_dynamic] Error 255
GENEHACK/Git-Wrapper-0.030.tar.gz
/usr/bin/make test -- NOT OK
//hint// to see the cpan-testers results for installing this module, try:
reports GENEHACK/Git-Wrapper-0.030.tar.gz
Running make install
make test had returned bad status, won't install without force
Failed during this command:
GENEHACK/Git-Wrapper-0.030.tar.gz : make_test NO
install
$ perl -V
Summary of my perl5 (revision 5 version 18 subversion 0) configuration:
Platform:
osname=darwin, osvers=8.11.0, archname=darwin-2level
uname='darwin macintosh-9.local 8.11.0 darwin kernel version 8.11.0: wed oct 10 18:26:00 pdt 2007; root:xnu-792.24.17~1release_ppc power macintosh powerpc '
config_args='-des'
hint=recommended, useposix=true, d_sigaction=define
useithreads=undef, usemultiplicity=undef
useperlio=define, d_sfio=undef, uselargefiles=define, usesocks=undef
use64bitint=undef, use64bitall=undef, uselongdouble=undef
usemymalloc=n, bincompat5005=undef
Compiler:
cc='cc', ccflags ='-fno-common -DPERL_DARWIN -fno-strict-aliasing -pipe -I/usr/local/include -I/opt/local/include',
optimize='-O3',
cppflags='-fno-common -DPERL_DARWIN -fno-strict-aliasing -pipe -I/usr/local/include -I/opt/local/include'
ccversion='', gccversion='4.0.1 (Apple Computer, Inc. build 5250)', gccosandvers=''
intsize=4, longsize=4, ptrsize=4, doublesize=8, byteorder=4321
d_longlong=define, longlongsize=8, d_longdbl=define, longdblsize=16
ivtype='long', ivsize=4, nvtype='double', nvsize=8, Off_t='off_t', lseeksize=8
alignbytes=8, prototype=define
Linker and Libraries:
ld='env MACOSX_DEPLOYMENT_TARGET=10.3 cc', ldflags =' -L/usr/local/lib -L/opt/local/lib'
libpth=/usr/local/lib /opt/local/lib /usr/lib
libs=-ldbm -ldl -lm -lc
perllibs=-ldl -lm -lc
libc=, so=dylib, useshrplib=false, libperl=libperl.a
gnulibc_version=''
Dynamic Linking:
dlsrc=dl_dlopen.xs, dlext=bundle, d_dlsymun=undef, ccdlflags=' '
cccdlflags=' ', lddlflags=' -bundle -undefined dynamic_lookup -L/usr/local/lib -L/opt/local/lib'
Characteristics of this binary (from libperl):
Compile-time options: HAS_TIMES PERLIO_LAYERS PERL_DONT_CREATE_GVSV
PERL_HASH_FUNC_ONE_AT_A_TIME_HARD PERL_MALLOC_WRAP
PERL_PRESERVE_IVUV PERL_SAWAMPERSAND USE_LARGE_FILES
USE_LOCALE USE_LOCALE_COLLATE USE_LOCALE_CTYPE
USE_LOCALE_NUMERIC USE_PERLIO USE_PERL_ATOF
Built under darwin
Compiled at May 18 2013 12:31:26
%ENV:
PERLBREW_BASHRC_VERSION="0.59"
PERLBREW_HOME="/Users/jimk/.perlbrew"
PERLBREW_MANPATH=""
PERLBREW_PATH="/Users/jimk/perl5/perlbrew/bin"
PERLBREW_ROOT="/Users/jimk/perl5/perlbrew"
PERLBREW_VERSION="0.59"
@INC:
/usr/local/lib/perl5/site_perl/5.18.0/darwin-2level
/usr/local/lib/perl5/site_perl/5.18.0
/usr/local/lib/perl5/5.18.0/darwin-2level
/usr/local/lib/perl5/5.18.0
/usr/local/lib/perl5/site_perl/5.16.0
/usr/local/lib/perl5/site_perl/5.14.2
/usr/local/lib/perl5/site_perl/5.14.0
/usr/local/lib/perl5/site_perl/5.12.0
/usr/local/lib/perl5/site_perl/5.10.1
/usr/local/lib/perl5/site_perl/5.10.0
/usr/local/lib/perl5/site_perl/5.8.6
/usr/local/lib/perl5/site_perl
.
This is largely system dependent, but because we write stdin before reading stdout/stderr in Git::Wrapper::RUN and read stdout before stderr, we can end up with deadlocks when pipe buffers fill up and we are blocked on another descriptor. Solving this with Open3 gets pretty icky with select() but there are several alternative modules to help with this.
You need to add a [Git::Push] to the bottom of dist.ini, for the release commit and tag.
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
Django
The Web framework for perfectionists with deadlines.
Laravel
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
Microsoft
Open source projects and samples from Microsoft.
Google
Google ❤️ Open Source for everyone.
Alibaba
Alibaba Open Source for everyone
D3
Data-Driven Documents codes.
Tencent
China tencent open source team.