Bug #81061 for Dist-Zilla-Plugin-Git: [Git::NextVersion] Bumping version from 0 to 1

Thu Nov 08 22:05:10 2012 shmuelfomberg [...] gmail.com - Ticket created

Subject:

[Git::NextVersion] Bumping version from 0 to 1

for some reason, Git::NextVersion insists on bumping the version to 1, and I have to always declare the version with V=0.007 Current git tags: v0.001 v0.002 v0.003 v0.004 v0.005 v0.006 in dist.ini, the plugin is declared without any parameter. Thanks.

Thu Nov 08 23:34:54 2012 cjm [...] cpan.org - Correspondence added

I was going to ask for a pointer to your Git repo, but when I tried building one of my dists (HTML-Tree) that uses Git::NextVersion, it did the same thing. The problem seems to be parsing the existing versions. When I added some debug logging, it reported every existing version was 0. I found a solution. Find the bit in Dist::Zilla::Plugin::Git::NextVersion that says 'version->parse($1)' and change that to 'version->parse("$1")'. Please try that and see if it fixes it for you, too. What I can't figure out is why the tests pass. Despite failing every time I try to build my dist, the tests work just fine.

Thu Nov 08 23:34:56 2012 The RT System itself - Status changed from 'new' to 'open'

Thu Nov 08 23:50:13 2012 shmuelfomberg [...] gmail.com - Correspondence added

Thanks, that fixed it for me. Good luck fixing the tests! :-) Shmuel.

Mon Nov 12 15:18:18 2012 cjm [...] cpan.org - Fixed in 2.002 added

Mon Nov 12 15:18:18 2012 cjm [...] cpan.org - Correspondence added

It's actually a bug in version->parse: https://rt.perl.org/rt3/Public/Bug/Display.html?id=115660 Basically, passing $1 to version->parse may or may not work, depending on what the rest of your program has been doing with $1. Unfortunately, loading Test::More tends to keep the bug from triggering. Matthew Horsfall discovered that adding "local $1;" in the right spot in the test will consistently trigger the bug, but only after I'd released 2.002. The improved test will be in the next release.

Mon Nov 12 15:18:19 2012 cjm [...] cpan.org - Status changed from 'open' to 'resolved'

Tue Nov 13 00:36:26 2012 shmuelfomberg [...] gmail.com - Correspondence added

On Mon Nov 12 15:18:18 2012, CJM wrote: Show quoted text

> It's actually a bug in version->parse: > https://rt.perl.org/rt3/Public/Bug/Display.html?id=115660 > > Basically, passing $1 to version->parse may or may not work, depending > on what the rest of your program has been doing with $1. Unfortunately, > loading Test::More tends to keep the bug from triggering. > > Matthew Horsfall discovered that adding "local $1;" in the right spot in > the test will consistently trigger the bug, but only after I'd released > 2.002. The improved test will be in the next release.

Thanks for the release. I don't think that you really need to bother testing that, as it is a known bug in Perl and surely will be closed soon. I wouldn't have. :-) About the but in the parse function - magic variables are really annoying to handle in XS. and you can never be sure where they will be injected from...

Tue Nov 13 00:36:27 2012 The RT System itself - Status changed from 'resolved' to 'open'

Tue Nov 13 00:51:00 2012 cjm [...] cpan.org - Status changed from 'open' to 'resolved'