[09:35] <mtj> hi folks, i have a bug i would like to fix, and would like some advice
[09:35] <mtj> https://bugs.launchpad.net/ubuntu/+source/libcgi-compile-perl/+bug/1927882
[09:35] -ubottu:#ubuntu-motu- Launchpad bug 1927882 in libcgi-compile-perl (Ubuntu) "Regression in libcgi-compile-perl 0.24-1 in Ubuntu Focal" [Undecided, Confirmed]
[09:39] <mtj> ..whats the best way to get libcgi-compile-perl 0.25 into focal?
[09:44] <mtj> good news is that the problem is fixed in jammy
[09:45] <mtj> ..but focal is still broken :/
[09:45] <mtj> https://packaging.ubuntu.com/html/packaging-new-software.html
[09:46] <mtj> https://wiki.ubuntu.com/UbuntuDevelopment/NewPackages
[09:48] <schopin> mtj: this is not a new package, but an update to an existing one, you're looking for this page: https://wiki.ubuntu.com/StableReleaseUpdates
[09:56] <mtj> schopin: many thanks
[09:56] <schopin> mtj: basically you need to justify the update, prepare the upload, and find a sponsor. I don't think we're very good at actually processing the sponsoring queue (I know I'm not, at least).
[09:58] <mtj> hmm, from reading the actual bug report, i dont have enough info to document a failure - other than making the test suite fail
[09:59] <mtj> perhaps thats a good enough demonstration?
[10:00] <mtj> https://packaging.ubuntu.com/html/fixing-ftbfs.html
[10:01] <schopin> If the package is FTBFS that's a fairly good argument to fix it, yes :)
[10:01] <mtj> i think this is a FTBFS situation
[10:01] <mtj> snap
[10:01] <schopin> Otherwise you could extract the test that demonstrates the failure into a self-contained perl script.
[10:01] <mtj> ah, yep
[10:03] <schopin> For impact and what could go wrong, it's often a good idea to look at reverse-dependencies of the package. There's a handy tool for that in ubuntu-dev-tools: reverse-depends
[10:07] <schopin> mtj: I just tried, and the package presently in focal currently builds fine, so it's not technically a FTBFS. I'm guessing what you meant was that along with the fix you added a test in the test suite?
[10:09] <mtj> hmm, ok
[10:10] <mtj> i now see i get a pass running ./Build test - rather than prove -v
[10:11] <mtj> that solves a mystery of how 0.24 got into debian unstable
[10:11] <mtj> https://tracker.debian.org/pkg/libcgi-compile-perl
[10:33] <mtj> hmm, i dont think i can find a trivial way to show a failure here - i might have to leave this
[10:48] <mtj> plan b, whats the chance of getting libcgi-compile-perl 0.25 backported to focal-backports?
[11:18] <schopin> mtj: can't you just compile two sample projects and assert that they're in separate packages?
[11:19] <schopin> If the package is broken, it should definitely be fixed in -updates IMHO. Backports should be to get new features.
[11:36] <mtj> schopin: hmm, my problem is finding a 'sample project' that fails on 0.24
[11:38] <mtj> i have one, but the steps to install, and provoke a failure is somewhat complex
[11:39] <mtj> https://github.com/Koha-Community/Koha/commit/4a3049a8be8dc2d22eac796521a4185f1248282a
[11:39] -ubottu:#ubuntu-motu- Commit 4a3049a in Koha-Community/Koha "Bug 28302: Forbid CGI::Compile 0.24"
[11:39] <mtj> https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=28302
[11:39] -ubottu:#ubuntu-motu- bugs.koha-community.org bug 28302 in Koha "Koha does not work with CGI::Compile 0.24" [Major, Closed: Fixed]
[11:41] <mtj> koha has provided 0.25 in its own apt repo, so probem is fixed for focal systems there
[11:48] <mtj> seems to be a horrible bug - test suite passes ok, yet bug subtley glitches your web app :/