So, let's make a 0-day bump. Seriously, though, the purpose of this bug is (mostly) to remind the maintainers that in media-libs/gst-plugins-* (including media-plugins/gst-plugins-libav) upstream has fixed an upstream problem of gtk-doc install, so now adding Gentoo workarounds for out-of-tree builds will produce the desired result (meaning, now the docs will get installed). For most, it's nearly just a copy-paste from -base, for -ugly a bit more care is needed (as it doesn't define multilib_src_configure - IMHO, the simplest way of handling that would be to bring it in sync with the others). The hardest part will obviously be figuring out the correct egl/gles2/wayland/etc. combination for the gl plugin changes.
Still haven't seen much help in figuring out gst-plugins-bad opengl stuff. I did receive a patch for my overlay though, but it just adds the USE flags, not REQUIRED_USE stuff, etc. 1.4.1 is out now, and that's my cue to hurry up finally and get these bumps done this week (perhaps without opengl support initially), hopefully by Friday, but if not then over weekend for sure.
Any update?
Gstreamer 1.4.3 is available now.
Created attachment 387504 [details, diff] changes to gst-plugins-bad ebuild adding GL plugin support Seeing as things are going nowhere anyway, I'm attaching what I did with gst-plugins-bad ebuild. I'm nearly certain that this particular useflag/configure switch setup will be somehow invalid outside my personal usecase (mesa/glx), but it's still a starting point, ain't it ?
Gstreamer 1.4.4 is out[1] 1. http://gstreamer.freedesktop.org/releases/gstreamer/1.4.4.html
latest is 1.4.5: http://lists.freedesktop.org/archives/gstreamer-devel/2014-December/050904.html
> upstream has fixed an upstream problem of gtk-doc install, so now adding Gentoo workarounds for out-of-tree builds will produce the desired result (meaning, now the docs will get installed). ...and now, let's take a look at gst-plugins-good-1.4.5.ebuild; what's still missing there ?
(In reply to Rafał Mużyło from comment #7) > > upstream has fixed an upstream problem of gtk-doc install, so now adding Gentoo workarounds for out-of-tree builds will produce the desired result (meaning, now the docs will get installed). > > ...and now, let's take a look at gst-plugins-good-1.4.5.ebuild; what's still > missing there ? The patch from you that you haven't submitted?
> The patch from you that you haven't submitted? ...and how about about taking a look at any of atk/pango/glib/gtk+ ebuilds, or I don't know, gst-plugins-base-1.4.5.ebuild and simply rechecking it with the content of the tarballs in question ? Yes, I'm being snippy here, but I've mentioned this back when mgorny was rewriting gstreamer classes for multilib suport, only at the time it was still broken upstream, so he didn't add it to *all* master ebuilds, only for those that were working. I suggested to make it future-proof, as the fix was already in the repo, but ...oh,well.
What patch is missing? Rafal, I guess your ebuild diff is to try to handle bug 525936?
(In reply to Pacho Ramos from comment #10) > What patch is missing? > > Rafal, I guess your ebuild diff is to try to handle bug 525936? Not really. Snippet from comment 4 was simply something that seemed to be minimal required change to adjust that ebuild to gst-plugins-gl being merged into -bad. Most likely, it's broken in a similar way to bug 525936 (I only needed it to work for me and that was covered by mesa/glx). As for the bitch fest from comment 7 onwards, it's just me complaining about a minor issue with a near trivial fix, that I consider somewhat important and that wouldn't even happen if mgorny put that fix in *all* the master ebuilds, instead of just the working ones, (the standard gtk-doc out-of-tree workaround) back when he was rewriting the eclasses and those tarballs where still broken (IRC our discussion did take place when upstream fixed the repos, waiting for 1.4.0 release).
all should be bumped now (or at least the stuff that was bumped to 1.2.x)
(In reply to Pacho Ramos from comment #12) > all should be bumped now (or at least the stuff that was bumped to 1.2.x) ...this might not be the same as bug 529788 (as that was a clash of divas, both of them devs), yet...`thank` you very much..