https://bitbucket.org/lazka/quodlibet-files/raw/default/releases/quodlibet-3.0.0.tar.gz Reproducible: Always 3.0.0 (2013-06-15) - THE NEMESIS HYPOTHESIS Requirements & Packaging Changes Python 2.6+ GTK+ 3.2+ & GIR (instead of 2.x) GStreamer 1.0+ & GIR (instead of 0.10) PyGObject 3.2+ (3.4 highly recommended) (instead of PyGTK) PyGObject cairo support libgpod 0.5+ (instead of python-gpod) libkeybinder-3.0 & GIR (instead of python-keybinder) HAL support removed New operon script + man page New .ini file for registering QL as a GNOME Shell Search Provider Translations New: Czech translation (Honza Hejzl) Russian translation update (Anton Shestakov) Lithuanian translation update [1079] (Naglis Jonaitis) Swedish translation update [1117] (Daniel Nyberg) Spanish, Basque and Galician translation update (Johám-Luís Miguéns Vila) Greek translation update [1175] (Dimitris Papageorgiou) General Improved rating visibility [1070] (Nick Boultbee) File system view: DnD directories to external targets (Nick Boultbee) Support GNOME Notification Sources Bayesian averaging for set (album) ratings [1085] (Nick Boultbee) New command line tagger: operon (see man operon) Hide songs on not-mounted drives on start without library refresh [984] Preferences UI for managing masked mount points [984] Support all patterns as song list headers [507, 1121] (Nick Boultbee) Save/restore queue position Documentation is now Sphinx/reST based and hosted on readthedocs.org Fixes Fix unwanted re-filtering of all open browsers if the search history changes Fix crash when re-adding devices while QL is running [1120] Remove EF directory mime type again (too many problems with file managers) Tagging APEv2: Add disc<->discnumber mapping Plugins New: Custom Commands plugin (Nick Boultbee) New: GNOME Shell Search Provider plugin [1147] ReplayGain plugin: Parallel processing [807] CD burn plugin: Add Xfburn support [1173] Known Regressions GStreamer output device selection is no longer supported. (GStreamer 1.0 has removed the property probing interface) No multimedia keys support with non-GNOME DEs in some distributions due to missing packaging: https://bugzilla.redhat.com/show_bug.cgi?id=963476 http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=710254 Animated On-Screen Display plugin is missing transparency support No Windows build (work in progress)
Created attachment 355504 [details] quodlibet 3.0.2 ebuild Based on quodlibet-2.5.ebuild. It`s not perfect but works on my gnome-3.8 system.
Quodlibet 3.0.2 installs and runs, but won't play any song here: === STACK TRACE Traceback (most recent call last): File "/usr/lib64/python2.7/site-packages/quodlibet/qltk/quodlibetwindow.py", line 113, in __select_song if player.go_to(iter, True): File "/usr/lib64/python2.7/site-packages/quodlibet/player/_base.py", line 179, in go_to self._end(True) File "/usr/lib64/python2.7/site-packages/quodlibet/player/gstbe.py", line 545, in _end if not self.__init_pipeline(): File "/usr/lib64/python2.7/site-packages/quodlibet/player/gstbe.py", line 208, in __init_pipeline map(bufbin.add, pipeline) File "/usr/lib64/python2.7/site-packages/gi/types.py", line 43, in function return info.invoke(*args, **kwargs) TypeError: Argument 1 does not allow None as a value This error was also reported at: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=724067 https://code.google.com/p/quodlibet/issues/detail?id=1238 In the second report, the bug was fixed by upgrading some gnome/gtk-related packages. I'm on a mostly stable system still on gnome 2.23 (and don't plan to upgrade until gnome-shell goes stable). So maybe there is some version dependency not listed in the ebuild.
The dependency on >=dev-python/pygobject-3.2 is not enough to get the gtk3 python bindings. I had to add x11-libs/gtk+:3[introspection] to $RDEPEND. Since now quodlibet is built on gtk3, I think that the dependency for dev-libs/keybinder should be changed to dev-libs/keybinder:3[python], but I have not tested if it fails otherwise. It could also be nice to add a dependency for media-plugins/gst-plugins-meta to force pulling some common decoders.
*** Bug 500290 has been marked as a duplicate of this bug. ***
*** Bug 500068 has been marked as a duplicate of this bug. ***
*** Bug 475160 has been marked as a duplicate of this bug. ***
This bug report is missing the attachment with the working ebuild for 2.6.3. This attachment is available at bug #500290. Please make it visible here, so people know that all that’s needed to fix the currently not usable quodlibet 2.5 is to test and deploy the ebuild. I’ll try to keep out of this bug-report here to avoid getting non-constructive out of frustration.
Note that there also is a working ebuild for quodlibet-plugins in bug #500294.
3.0.2 in Portage, will stabilize in about 30 days if no regressions are reported
(In reply to Samuli Suominen from comment #10) > 3.0.2 in Portage, will stabilize in about 30 days if no regressions are > reported Do I understand correctly, that you intentionally only added 3.0.2 to the tree, but not 2.6.3? Upstream maintains both. Why do you leave out 2.6.3, even though there is an ebuild. At this point, contributing to quodlibet in Gentoo has been a very frustrating experience for me.
(In reply to Arne Babenhauserheide from comment #11) > (In reply to Samuli Suominen from comment #10) > > 3.0.2 in Portage, will stabilize in about 30 days if no regressions are > > reported > > Do I understand correctly, that you intentionally only added 3.0.2 to the > tree, but not 2.6.3? > > Upstream maintains both. Why do you leave out 2.6.3, even though there is an > ebuild. At this point, contributing to quodlibet in Gentoo has been a very > frustrating experience for me. 3.0.2 doesn't work for you as a replacement for 2.6.3? can you tell why?
(In reply to Samuli Suominen from comment #12) > 3.0.2 doesn't work for you as a replacement for 2.6.3? can you tell why? Because it has completely different requirements. Are all gtk2 themes available for gtk3? If not, this is a regression. Can you ensure me that there will be *no* regressions on non-Gnome systems? That custom-plugins will keep working?
First regression: quodlibet 3.0.2 does not adhere my inverse mousewheel direction setting. 3 minutes into testing 3.0.2 and I hit the first regression, while 2.6.3 worked flawlessly. This is the kind of issue why I want the smooth update which makes the program work again and not the big jump with lots of changes which breaks stuff. And please don’t start to call it small: The update to 2.6.3 fixes a crash (critical bug). The update to 3.0.2 changes the audio-backend, the graphical libraries, and so forth, which is bound to create some problems that first need to be ironed out before it can be called anything close to stable. So please deploy 2.6.3 as incremental update which makes QuodLibet 2.x work again in Gentoo.
And now I see that the inverted mouse regression is an unfixed bug in gnome-control-center… https://bugzilla.gnome.org/show_bug.cgi?id=674716 https://bugzilla.gnome.org/show_bug.cgi?id=682457 And the window does not react when it does not have focus — this starts to get even more frustrating. Please just get 2.6.3 into the tree. I provided the ebuilds for that, I can’t do more. The user interface of 3.0.2 is broken due to the bugs I linked you. 2.6.3 works. What reason is there for not including quodlibet 2.6.3?
A further bug-report which shows the breakage, this time for ubuntu: https://bugs.launchpad.net/ubuntu/+source/gtk+3.0/+bug/966237
(In reply to Arne Babenhauserheide from comment #13) > (In reply to Samuli Suominen from comment #12) > > 3.0.2 doesn't work for you as a replacement for 2.6.3? can you tell why? > > Because it has completely different requirements. Are all gtk2 themes > available for gtk3? If not, this is a regression. that's bs. there is no such magical requirement for gtk2 themes to work on gtk3. > Can you ensure me that there will be *no* regressions on non-Gnome systems? ensure? no, nobody can. all software has bugs, but that's why we have the 30 days period to report bugs that might block 3.0.x from going stable. > That custom-plugins will keep working? of course not. portage is not responsible what you have installed outside of it.
(In reply to Samuli Suominen from comment #17) > ensure? no, nobody can. all software has bugs, but that's why we have the 30 > days period to report bugs that might block 3.0.x from going stable. Can I mark the upstream gnome bugs as blockers?
(In reply to Samuli Suominen from comment #17) > that's bs. there is no such magical requirement for gtk2 themes to work on > gtk3. … > ensure? no, nobody can. all software has bugs, but that's why we have the 30 > days period to report bugs that might block 3.0.x from going stable. … > of course not. portage is not responsible what you have installed outside > of it. 2.6.3 is an incremental update, same requirements as 2.5 which is already in the tree and stable - but broken due to a recent gtk2 update. I do *not* argue for keeping 3.0.2 out of the tree, just to *also* add 2.6.3. The bugs I linked in #c15 make Quodlibet 3.0.2 close to unusable with the mouse wheel set to natural scrolling (an option in KDE systemsettings - and the option I‘ve been using for half a year: Ever since a friend told me that Apple switched to that and I thought that it could be interesting to test it). And this should definitely block stabilization (though not getting 3.0.2 into the tree).
PS: I’m sorry for my harsh tone. It’s very frustrating to see 2.6.3 blocked and to not understand why. Can you explain me the reasons, why you do not want 2.6.3 in the tree, even though it is a maintained upstream release?
committed 2.6.3 to stable.
Thank you! I downgraded and everything works again.