emerge --depclean wants to remove dev-qt/qtsql-5.15.16 as not required and keeps the lib as kaffeine still uses it. Recompiling kaffeine then fails due to not finding qtsql. Seems like the ebuild needs a dependency on it. Looks like it used to be one of the prior qt-base version. Reproducible: Always Steps to Reproduce: 1. emerge kaffeine-2.0.19 2. emerge -C qtsql 3. try to emerge kaffeine again It's obviously simple to just keep qtsql around, I use --depclean -p and manually remove anyway. But I neither want qtsql in my world file nor have a false positive all the time.
FWIW, the dep is there in 2.0.19 (which also has Qt6).
I don't know what you mean by that. I just synced to be sure and my kaffeine-2.0.19.ebuild does depend on qt6 but not on dev-qt/qtsql. Just on sql use flag in qtbase.
Invalid, then.
What I meant was that the version in ~arch doesn't have this problem, just the stable one does. I.e. it got fixed when porting to qt6. That version is correct. It was just a note. The stable one is still wrong.
emerge --info states: ACCEPT_KEYWORDS="amd64 ~amd64" To my knowledge this means im on unstable? And it doesn't change the fact that I use the ebuild from portage which does miss a dependency. Building kaffeine fails in the configure phase without qtsql. Just performed a rebuild, ldd /usr/bin/kaffeine also still links to Qt5 while the ebuild mandates QTMIN=6.7.2 Doesn't look correct to me.
Then that is a completely different issue again? The deps for the ~arch version look correct as qtbase[sql] is the correct dependency for qt6 (the splitting of packages/components changed). If the ~arch ebuild is using qt5 wrongly, then that is its own bug that should be filed separately. Having no logs or output isn't helping the confusion here either.
Ah, but your original bug was about 2.0.19. Okay, so it's the same issue, but your analysis is wrong.
sigh... they really tricked me there
*** Bug 945134 has been marked as a duplicate of this bug. ***
The bug has been closed via the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=a6f27914abcafa1fcb554272da2b9ae5811eb4f7 commit a6f27914abcafa1fcb554272da2b9ae5811eb4f7 Author: Andreas Sturmlechner <asturm@gentoo.org> AuthorDate: 2024-11-27 17:36:36 +0000 Commit: Andreas Sturmlechner <asturm@gentoo.org> CommitDate: 2024-11-27 17:38:57 +0000 media-video/kaffeine: drop 2.0.19 Closes: https://bugs.gentoo.org/945125 Signed-off-by: Andreas Sturmlechner <asturm@gentoo.org> media-video/kaffeine/Manifest | 1 - media-video/kaffeine/kaffeine-2.0.19.ebuild | 58 ----------------------------- 2 files changed, 59 deletions(-) Additionally, it has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=2d6285258e672aa88572d2eeeb3c048d259c777e commit 2d6285258e672aa88572d2eeeb3c048d259c777e Author: Andreas Sturmlechner <asturm@gentoo.org> AuthorDate: 2024-11-27 17:36:16 +0000 Commit: Andreas Sturmlechner <asturm@gentoo.org> CommitDate: 2024-11-27 17:38:57 +0000 media-video/kaffeine: add 2.1.0_pre20241125 Bug: https://bugs.gentoo.org/945125 Signed-off-by: Andreas Sturmlechner <asturm@gentoo.org> media-video/kaffeine/Manifest | 1 + .../kaffeine/kaffeine-2.1.0_pre20241125.ebuild | 53 ++++++++++++++++++++++ 2 files changed, 54 insertions(+)
So I attributed it to the wrong source and the ebuild was wrongly converted to qt6?
(In reply to David Philippi from comment #11) > So I attributed it to the wrong source and the ebuild was wrongly converted > to qt6? Yes, a packaging mistake due to live ebuild already synced to KF6-based git master, but v2.0.19 (tag) still made from a 2.0.x stable branch. Thanks for filing the bug.
