Bug#1005068: transition (unplanned?): libwacom9
On Mon, 7 Feb 2022 13:52:48 +0200 Timo Aaltonen wrote: > Sebastian Ramacher kirjoitti 7.2.2022 klo 11.38: > > On 2022-02-07 10:48:01, Timo Aaltonen wrote: > >> Simon McVittie kirjoitti 7.2.2022 klo 3.17: > >>> On Mon, 07 Feb 2022 at 00:51:08 +, Marius Gripsgard wrote: > The ABI/API breakage is not *too* bad, see [0]. just hope none use those > deprecated APIs > > [0] https://github.com/linuxwacom/libwacom/commit/b4f3c3f855a68e35b5ebfc305bc67c565f9146c7 > >>> > >>> Thanks. Looks like nobody should have been using those symbols anyway > >>> (the commit message says they were private and not mentioned in any public > >>> header files), and codesearch.debian.net does not find any references > >>> outside src:libwacom, so it should be safe to binNMU dependent packages. > >>> > >>> smcv > >>> > >> > >> Right, should've asked for a transition. > >> > >> Filed nmu's for wacomtablet and cinnamon-settings-daemon, libinput rebuild > >> upload I did myself. > > > > If there's a transition bug, there is no need to file individual > > requests for nmus. > > > > Cheers > > Oh, sorry about those then. I thought there was something I had to do to > fix this mess.. > > I'll upload a source-only libwacom later, thanks. > > > -- > t > > Can't wait for this one to be fixed - it is a blocker on my Sid apt upgrade :)
Bug#1005068: transition (unplanned?): libwacom9
Sebastian Ramacher kirjoitti 7.2.2022 klo 11.38: On 2022-02-07 10:48:01, Timo Aaltonen wrote: Simon McVittie kirjoitti 7.2.2022 klo 3.17: On Mon, 07 Feb 2022 at 00:51:08 +, Marius Gripsgard wrote: The ABI/API breakage is not *too* bad, see [0]. just hope none use those deprecated APIs [0] https://github.com/linuxwacom/libwacom/commit/b4f3c3f855a68e35b5ebfc305bc67c565f9146c7 Thanks. Looks like nobody should have been using those symbols anyway (the commit message says they were private and not mentioned in any public header files), and codesearch.debian.net does not find any references outside src:libwacom, so it should be safe to binNMU dependent packages. smcv Right, should've asked for a transition. Filed nmu's for wacomtablet and cinnamon-settings-daemon, libinput rebuild upload I did myself. If there's a transition bug, there is no need to file individual requests for nmus. Cheers Oh, sorry about those then. I thought there was something I had to do to fix this mess.. I'll upload a source-only libwacom later, thanks. -- t
Bug#1005068: transition (unplanned?): libwacom9
Hi Timo I've scheduled binNMUs for cinnamon-settings-daemon and wacomtablet. Please remember that libwacom needs a source-only upload for migration. Regards Graham
Bug#1005068: transition (unplanned?): libwacom9
On 2022-02-07 10:48:01, Timo Aaltonen wrote: > Simon McVittie kirjoitti 7.2.2022 klo 3.17: > > On Mon, 07 Feb 2022 at 00:51:08 +, Marius Gripsgard wrote: > > > The ABI/API breakage is not *too* bad, see [0]. just hope none use those > > > deprecated APIs > > > > > > [0] > > > https://github.com/linuxwacom/libwacom/commit/b4f3c3f855a68e35b5ebfc305bc67c565f9146c7 > > > > Thanks. Looks like nobody should have been using those symbols anyway > > (the commit message says they were private and not mentioned in any public > > header files), and codesearch.debian.net does not find any references > > outside src:libwacom, so it should be safe to binNMU dependent packages. > > > > smcv > > > > Right, should've asked for a transition. > > Filed nmu's for wacomtablet and cinnamon-settings-daemon, libinput rebuild > upload I did myself. If there's a transition bug, there is no need to file individual requests for nmus. Cheers > > > -- > t > -- Sebastian Ramacher
Bug#1005068: transition (unplanned?): libwacom9
Simon McVittie kirjoitti 7.2.2022 klo 3.17: On Mon, 07 Feb 2022 at 00:51:08 +, Marius Gripsgard wrote: The ABI/API breakage is not *too* bad, see [0]. just hope none use those deprecated APIs [0] https://github.com/linuxwacom/libwacom/commit/b4f3c3f855a68e35b5ebfc305bc67c565f9146c7 Thanks. Looks like nobody should have been using those symbols anyway (the commit message says they were private and not mentioned in any public header files), and codesearch.debian.net does not find any references outside src:libwacom, so it should be safe to binNMU dependent packages. smcv Right, should've asked for a transition. Filed nmu's for wacomtablet and cinnamon-settings-daemon, libinput rebuild upload I did myself. -- t
Bug#1005068: transition (unplanned?): libwacom9
On Mon, 07 Feb 2022 at 00:51:08 +, Marius Gripsgard wrote: > The ABI/API breakage is not *too* bad, see [0]. just hope none use those > deprecated APIs > > [0] > https://github.com/linuxwacom/libwacom/commit/b4f3c3f855a68e35b5ebfc305bc67c565f9146c7 Thanks. Looks like nobody should have been using those symbols anyway (the commit message says they were private and not mentioned in any public header files), and codesearch.debian.net does not find any references outside src:libwacom, so it should be safe to binNMU dependent packages. smcv
Bug#1005068: transition (unplanned?): libwacom9
Package: release.debian.org Severity: normal User: release.debian@packages.debian.org Usertags: transition X-Debbugs-Cc: libwa...@packages.debian.org Control: forwarded -1 https://release.debian.org/transitions/html/auto-libwacom.html src:libwacom was recently updated from libwacom2 to libwacom9, starting a transition. I don't see any indication that this was approved by the release team, so I'm opening this transition tracking bug (my apologies if there is already a transition-tracker that I missed). https://release.debian.org/transitions/html/auto-libwacom.html looks like it is tracking this transition correctly. Because libwacom2 and libwacom9 both depend on a matching libwacom-common, this makes affected packages such as mutter uninstallable in unstable. How big are the API/ABI breaks? Are dependent packages expected to compile and work successfully without code changes? In addition to rebuilding reverse-dependencis, src:libwacom will need a source-only upload to unstable so that it will be allowed to migrate. It is usually best for transitions to happen by uploading NEW packages to experimental, then opening a transition-tracking bug like this one and getting release team approval, and finally re-uploading to unstable after approval has been received, particularly if the transition is an all-or-nothing one like this. smcv
