Yes, systemd_system_unitdir helps with the systemd files, but the other files: /usr/lib/snapper/systemd-helper /usr/lib/snapper/installation-helper /usr/lib/snapper/plugins
don't use systemd_system_unitdir and as the path to them is hardcoded in the source, I guess we should either hardcode the path in FILES as well (so that they are still packaged and found in runtime even if someone changes libdir value) or change both the code and the packaging to respect libdir (which might be too much to ask after looking at the Makefile.am). I have a .patch file to change Makefile.am to allow passing PREFIX, LIBDIR, DATADIR, SYSTEMDDIR values, but then I've noticed plugins and helpers and stopped there (as I changing that as well to respect passed LIBDIR) should be at least partially tested in runtime which I wasn't planning to. Sending PR with just Makefile.am changes and asking the developers to "finish it" also probably won't help much, so I've just sent previous e-mail and added SKIP_RECIPE in our DISTRO with multilib. I can share the draft PR if someone finds that useful. Cheers, On Thu, Oct 31, 2024 at 6:43 PM Markus Volk <[email protected]> wrote: > > On Thu, Oct 24 2024 at 12:47:46 PM +02:00:00, Martin Jansa > <[email protected]> wrote: > > should the packaging be changed to use nonarch_libdir and let lib32-snapper > and snapper conflict with each other? Or is there some better way to resolve > this? > > > Using systemd_system_unitdir will also basically just use nonarch_libdir. But > I wouldn't have a problem with them conflicting each other. I only need > snapper for some experiments. Thanks for looking into this
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#113376): https://lists.openembedded.org/g/openembedded-devel/message/113376 Mute This Topic: https://lists.openembedded.org/mt/109116183/21656 Group Owner: [email protected] Unsubscribe: https://lists.openembedded.org/g/openembedded-devel/unsub [[email protected]] -=-=-=-=-=-=-=-=-=-=-=-
