Control: tag -1 pending On Wed, Jul 26, 2023 at 10:19:53PM +0200, Lucas Nussbaum wrote: > Source: devscripts > Version: 2.23.5 > Severity: serious > Justification: FTBFS > Tags: trixie sid ftbfs > User: lu...@debian.org > Usertags: ftbfs-20230726 ftbfs-trixie
> During a rebuild of all packages in sid, your package failed to build > on amd64. > > Undefined subroutine &JSON::to_json called at ./t/salsa.pm line 49. > > # Tests were run but no plan was declared and done_testing() was not seen. > > # Looks like your test exited with 255 just after 1. > > test_comments_in_quoted_strings2 > > t/salsa.t ......... > > Dubious, test returned 255 (wstat 65280, 0xff00) This regressed with libgitlab-api-v4-perl 0.27-1, which is not migrating to testing because of the same failure on the autopkgtest side. The problem is that t/salsa.pm uses JSON::to_json() without loading the JSON module. This was always wrong, but GitLab::API::v4::RESTClient used to bring it in transitively. That has now changed as libgitlab-api-v4-perl has switched from the JSON module to JSON::MaybeXS. Umh. After patching this and another thing locally I see now that it's also #1041220 / #1038486, which were incorrectly merged resulting in confused metadata that makes it look like it's all a thing of the past. FWIW I think #1041220 should not have been reassigned + merged. It was about libgitlab-api-v4-perl 0.26-3 being seriously buggy because it is out of sync with unstable. When devscripts is fixed, libgitlab-api-v4-perl 0.27-1 will then migrate and #1041220 would have stopped affecting testing without any other action on that bug. But undoing the merge now would probably make things even more confusing. Copying Yadd and Gregor who were involved. It looks like devscripts.git has the necessary patches (including a perltidy fix) but somebody just needs to upload it (and possibly fix the BTS mess afterwards if it becomes a problem.) -- Niko