Forums before death by AOL, social media and spammers... "We can't have nice things"
|    linux.debian.bugs.dist    |    Ohh some weird Debian bug report thing    |    28,835 messages    |
[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]
|    Message 28,518 of 28,835    |
|    =?utf-8?Q?=C3=89tienne?= Mollier to All    |
|    Bug#1128659: trixie-pu: package plastima    |
|    22 Feb 26 12:20:01    |
      XPost: linux.debian.devel.release       From: emollier@debian.org              --QRjuw5p4Lc0+DcYA       Content-Type: text/plain; charset=utf-8       Content-Disposition: inline       Content-Transfer-Encoding: quoted-printable              Package: release.debian.org       Severity: normal       Tags: trixie       X-Debbugs-Cc: plastimatch@packages.debian.org       Control: affects -1 + src:plastimatch       User: release.debian.org@packages.debian.org       Usertags: pu              Hi Stable Release Managers,              I would like to proceed to a stable update (and oldstable update       in a separate entry) of plastimatch following a recent need for       repacking the upstream source code to exclude files.              [ Reason ]       As initially identified in #1124959, the source package       plastimatch 1.10.0+dfsg.1-1 in trixie is shipping several files       that are not suitable for commercial use since at least version       1.9.3+dfsg.1-1 shipped in bullseye.              [ Impact ]       If the update is not approved, the source will continue shipping       non-free files, which would eventually require a removal of the       package from stable releases.              [ Tests ]       After removal of affected source files, plastimatch kept       building properly. The package ships an autopkgtest, which I       have ensured remains in working conditions.              [ Risks ]       The files src/register/cuda/*.cu are Cuda kernels that aren't       strictly needed and can run only on specific hardware. The       header files src/register/viscous_convolution.h and       src/register/viscous_global.h seemed more risky to remove, as       they could be needed at build time, but this turned out to not       be a problem; upon investigation of their content, they seemed       related to adding support to Cuda kernels, which may explain how       they end up optional. The issue was brought upstream and they       indicated that affected files were never in working conditions       in the first place[1]. The package is only available on amd64       release architecture due to missing dependencies on other       platforms. Therefore, I believe that the changes are not too       risky.              [1]: https://gitlab.com/plastimatch/plastimatch/-/issues/95              [ Checklist ]        [*] *all* changes are documented in the d/changelog        [*] I reviewed all changes and I approve them        [*] attach debdiff against the package in (old)stable        [*] the issue is verified as fixed in unstable              [ Changes ]       Hunk by hunk from the debdiff, changes are:               * .gitignore and .gitlab-ci.yml files have been caught in the        blast radius of the repacking procedure, so went missing,        but they are not needed downstream anyways, so not a great        loss;               * d/copyright adds exclusion of non-free cuda kernels and        header files binding them to plastimatch;               * d/copyright removes also a leftover paragraph reflecting        past upstream releases;               * d/salsa-ci.yml is adjusted to avoid tripping unnecessarily        the i386 build, which is going to be impossible, and such        change should have no impact downstream anyways;               * d/watch is adjusted to reflect the +dfsg version bump;               * all the remaining hunks and big bulk of the debdiff reflect        the removal of Cuda kernel and header files.              [ Other info ]       I was slightly confused by changelog-file-missing-explicit-entry       lintian warning and I believe that this may render the whole       update harder to read. If you thing that this needs to be       simplified, I'm happy to make amendments.              I'm not too happy with my accidental removal of .git* files       during my repack in unstable, but reintroducing them to minimize       the debdiff is going to involve another round trip through sid       IIUC.              Have a nice day, :)       --         .''`. Étienne Mollier |
[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]
(c) 1994, bbs@darkrealms.ca