Created attachment 335794 [details] build.log dep should be: cdda? ( || ( dev-libs/libcdio-paranoia <dev-libs/libcdio-0.90[-minimal] ) ) and you need this to get to the actual errors: if has_version dev-libs/libcdio-paranoia; then sed -i \ -e 's:cdio/cdda.h:cdio/paranoia/cdda.h:' \ -e 's:cdio/paranoia.h:cdio/paranoia/paranoia.h:' \ src/*.{c,h} src/decoder/dec_cdda.h || die fi
I have made the changes to the ebuild but a patch to make it build is still missing. Will report this upstream if i have time.
(In reply to comment #0) > Created attachment 335794 [details] > build.log > > dep should be: > > cdda? ( || ( dev-libs/libcdio-paranoia <dev-libs/libcdio-0.90[-minimal] ) ) > > and you need this to get to the actual errors: > > if has_version dev-libs/libcdio-paranoia; then > sed -i \ > -e 's:cdio/cdda.h:cdio/paranoia/cdda.h:' \ > -e 's:cdio/paranoia.h:cdio/paranoia/paranoia.h:' \ > src/*.{c,h} src/decoder/dec_cdda.h || die > fi Shouldn't this be if has_version dev-libs/libcdio-paranoia; then sed -i \ -e 's:cdio/cdda.h:cdio/paranoia/cdda.h:' \ -e 's:cdio/paranoia.h:cdio/paranoia/paranoia.h:' \ src/*.{c,h} src/decoder/dec_cdda.{c,h} || die fi There is another include in src/decoder/dec_cdda.c.
(In reply to comment #2) > (In reply to comment #0) > > Created attachment 335794 [details] > > build.log > > > > dep should be: > > > > cdda? ( || ( dev-libs/libcdio-paranoia <dev-libs/libcdio-0.90[-minimal] ) ) > > > > and you need this to get to the actual errors: > > > > if has_version dev-libs/libcdio-paranoia; then > > sed -i \ > > -e 's:cdio/cdda.h:cdio/paranoia/cdda.h:' \ > > -e 's:cdio/paranoia.h:cdio/paranoia/paranoia.h:' \ > > src/*.{c,h} src/decoder/dec_cdda.h || die > > fi > > Shouldn't this be > > if has_version dev-libs/libcdio-paranoia; then > sed -i \ > -e 's:cdio/cdda.h:cdio/paranoia/cdda.h:' \ > -e 's:cdio/paranoia.h:cdio/paranoia/paranoia.h:' \ > src/*.{c,h} src/decoder/dec_cdda.{c,h} || die > fi > > There is another include in src/decoder/dec_cdda.c. If it's there, definately. I didn't really want any of this committed to Portage before the full fix is available. Daniel, I think you can point upstream to example patches, like: http://sources.gentoo.org/cgi-bin/viewvc.cgi/gentoo-x86/media-plugins/gst-plugins-cdio/files/gst-plugins-cdio-0.10.19-cdio-cd-text-api.patch?view=markup http://sources.gentoo.org/cgi-bin/viewvc.cgi/gentoo-x86/media-plugins/audacious-plugins/files/audacious-plugins-3.3.3-libcdio.patch?view=markup http://sources.gentoo.org/cgi-bin/viewvc.cgi/gentoo-x86/media-video/mplayer2/files/mplayer2-2.0_p20121128-cdio-api-fixes.patch?view=markup Thanks.
And I would temporarily disable entire libcdio support from aqualung in ~arch pending on the fix. But that's just my personal opinion.
(In reply to comment #4) > And I would temporarily disable entire libcdio support from aqualung in > ~arch pending on the fix. But that's just my personal opinion. I added the links to the patches on the upstream tracker. Stable version of aqualung now restricts to <dev-libs/libcdio-0.90[-minimal] and the arch version has cdda (libcdio) support disabled).
This is the only incompatible package left in Portage. I'd like to lastrite it, or remove the cdda USE flag so we can proceed with the stabilization of the new libcdio-paranoia package and the eselect module.
(In reply to comment #6) > This is the only incompatible package left in Portage. > I'd like to lastrite it, or remove the cdda USE flag so we can proceed with > the stabilization of the new libcdio-paranoia package and the eselect module. Oh, nevermind, I see -r2 has the USE flag removed! I'll just file -r2 stabilization bug once I'll open the stablereq. Sorry for the noise.
A patch seems to be here: http://sourceforge.net/mailarchive/forum.php?thread_name=1361777053-5781-1-git-send-email-jamie%40audible.transient.net&forum_name=aqualung-friends Didn't test it, though since I don't get compile errors in the first place and didn't really understood your comments about "needing s.th. to trigger the error". I vote for dropping the CDDA USE flag instead of dropping the whole package
If someone backports the patch to the version in tree I will apply it else we stabilize with the use flag removed and add it back with the next version bump.
No release since 2010-01-31 and no patch backport effort for 2/3 of a year, I guess it wouldn't be unreasonable to simply drop the USE flag and continue stabilisation?
(In reply to Andreas Sturmlechner from comment #10) > No release since 2010-01-31 and no patch backport effort for 2/3 of a year, > I guess it wouldn't be unreasonable to simply drop the USE flag and continue > stabilisation? I'm opening libcdio-paranoia stabilization bug now, so yes, aqualung will have to be stabilized to drop the libcdio support.
