Needs a release?

[karenetheridge]

It looks like there's been enough changes (IMO anyway) to justify a release -- this will also allow an update of the bugtracker and repository information in metadata.

[wchristian]

@karenetheridge is right, we need a new release of PPI, and my dev release at https://metacpan.org/release/MITHALDU/PPI-1.216_01 has not had a single test failure. I actually meant to ask for a new release after Easter, but got side-tracked. @adamkennedy, can you please cut a new release? Alternatively, i would still very much appreciate COMAINT, so i could switch to a "release early, release often" schedule.

[adamkennedy]

I'll set up COMAINT, since you're switched build systems and my release automation likely won't work with PPI any more.

Please note that a release-often approach for PPI is strongly discouraged, as it's common for even fairly innocent changes to break test cases for things like Perl Critic and have a major impact on the downstream ecosystem.

Each production release needs to be carefully coordinated with the major downstream dependencies, so at the very least they have a chance to prepare a new release of their own which matches the changes.

You should, at minimum, ensure you have a "release candidate" dev release tarball on CPAN, then reach out to the Perl Critic release manager and have him test against their current release, their dev releases, and potentially trunk, and see if they need to do anything.

Adam

On 30 May 2014 09:07, Christian Walde notifications@github.com wrote:

@karenetheridge https://github.com/karenetheridge is right, we need a new release of PPI, and my dev release at https://metacpan.org/release/MITHALDU/PPI-1.216_01 has not had a single test failure. I actually meant to ask for a new release after Easter, but got side-tracked. @adamkennedy https://github.com/adamkennedy, can you please cut a new release? Alternatively, i would still very much appreciate COMAINT, so i could switch to a "release early, release often" schedule.

— Reply to this email directly or view it on GitHub https://github.com/adamkennedy/PPI/issues/81#issuecomment-44668413.

[wchristian]

Caveats duly noted and agreed on. By "often" i simply mean more than every two years. :)

Notes for the future: Any PPI release will be preceded by a -TRIAL release, a bpo blog post, twitter post and direct mail to major downstreams with an attempt to get their explicit go-ahead. Speaking of which, do you have a current list of major downstreams?

As for automation, can you give me a list of things it should do, as far as you remember?

[wchristian]

@adamkennedy The dev release i did shows all green: http://matrix.cpantesters.org/?dist=PPI+1.216_01

I do however not seem to have COMAINT yet: https://metacpan.org/release/MITHALDU/PPI-1.216_01

Can you please take care of that?

[moregan]

+1 on getting a release out there

[wchristian]

Adam has agreed to let mst assign me comaint rights, so i can now do a release as soon as necessary and practical, and am in fact planning to do one on the 15th august, since i managed to satisfy, with the latest commits to master, my confidence in being able to release a proper tar ball, by diffing the one i generated for 1.217_01 with the one for 1.215 and not finding any unexpected changes.

To this effect i've written a blog post:

http://blogs.perl.org/users/mithaldu/2014/08/please-test-ppi-1217-01---bug-fixes-speed-optimizations-tests-doc-improvements.html

And emailed the owners of the dists pinpointed as depending on PPI:

https://gist.github.com/wchristian/0af3361e3459af7b5086

As bounces come in i will try to figure out the correct addresses and resend.

[wchristian]

And the release has been done: http://blogs.perl.org/users/mithaldu/2014/08/ppi-1218-has-been-released---bug-fixes-speed-optimizations-tests-doc-improvements.html

Thanks to everyone involved. :)