[04:13] <RAOF[m]> xnox: huh! The Mir symbols mismatch is an interesting one!
[04:14] <RAOF[m]> xnox: I think it's plausibly a bug/missing feature in dpkg-gensymbols.
[04:14] <RAOF[m]> But it's also weird behaviour by the linker.
[04:17] <RAOF[m]> With LTO enabled, the linker emits debug symbols (but not the function symbols) for some otherwise-unexported symbols.
[04:18] <RAOF[m]> dpkg-gensymbols then complains that we're introducing new symbols.
[04:20] <RAOF[m]> But you can't actually *link* against those symbols; they're just the debug info. So they're irrelevant for the symbols file.
[04:20] <RAOF[m]> Hmmmm
[04:54] <RAOF[m]> And, of course, dh_strip gets them out of the DSO *anyway*, and the problem is only that the `check-miral-symbols` target gets run before `dh_strip` is.
[06:12] <mvo> infinity: hey, good morning, sorry for not getting back last night, real life interfered :/
[06:13] <mvo> infinity: I tried to reproduce in various ways yesterday but no luck, I try harder now
[06:40] <mvo> jibel: fwiw, I can reproduce it now with both snapd 2.33 and 2.34 :/
[06:40] <mvo> jibel: we are looking into it
[06:47] <mvo> jibel: after a bit of poking I think this is "just" a side-effect of the seed order issue
[06:47] <mvo> jibel: i.e. once the seed order is correct and the initial snap seeding works then gedit should install again
[08:59] <LocutusOfBorg> Laney, good morning, quick question: if gst-plugins-good1.0 builds correctly, does this mean I can sync gst-plugins-base?
[08:59] <LocutusOfBorg> see https://launchpad.net/~costamagnagianfranco/+archive/ubuntu/locutusofborg-ppa for progresses :p
[09:04] <Laney> LocutusOfBorg: No, we have a change in -good to keep the qt5 stuff on armhf which requires that change in -base.
[09:04] <Laney> If you wanted to, you could check if all of those changes can be dropped (if the current mesa is fixed)
[09:04] <Laney> https://github.com/KhronosGroup/OpenGL-Registry/issues/162 this one was closed
[09:10] <Laney> are you doing the point release update?
[09:25] <xnox> RAOF[m], indeed the symbols to me looked weird.
[09:39] <LocutusOfBorg> Laney, I'm taking wrt cosmic, and the fact that we need to merge/sync gst...
[09:39] <Laney> I know you are
[09:40] <LocutusOfBorg> the -base plugin (new version) might have fixed that, that was the origin of the question
[09:40] <Laney> It'll be a fix in the headers that come with mesa
[09:40] <LocutusOfBorg> mmm they might have fixed in the plugin...
[09:40] <LocutusOfBorg> -#ifdef USE_EGL_RPI
[09:40] <LocutusOfBorg> +#if !GST_GL_HAVE_EGLUINT64KHR
[09:40] <LocutusOfBorg>  typedef khronos_uint64_t EGLuint64KHR;
[09:40] <LocutusOfBorg> this is what I'm seeing in the diff
[09:41] <LocutusOfBorg> maybe they workarounded it, without having to wait for mesa
[09:41] <LocutusOfBorg> so, this is the question: can I upload a sync if -good builds on arm*?
[09:41] <LocutusOfBorg> (this requires stealing your merge btw)
[09:42] <Laney> no, you have to keep my reverts in there
[09:43] <Laney> if it builds with the synced base and those reverts, that's the good state
[09:44] <Laney> (there -> -good)
[09:46] <LocutusOfBorg> Laney, I was talking about syncing base, I know the good needs a merge (and yes, here I kept the revert of course)
[09:47] <Laney> ok then
[09:47] <LocutusOfBorg> and it failed to build, so can I merge both? :)
[09:48] <LocutusOfBorg> tjaalton, mesa merge ping :p
[09:48] <Laney> sure, and -bad plz
[09:48] <Laney> if you could use the git repositories that would make me happy
[09:49] <LocutusOfBorg> I will
[09:51] <Laney> not sure newest mesa is fixed yet
[09:51] <Laney> https://salsa.debian.org/xorg-team/lib/mesa/blob/debian-unstable/include/GL/glext.h#L468 should be khronos_ssize_t etc
[09:53] <LocutusOfBorg> according to the debian bug seems not
[09:53] <LocutusOfBorg> I fail to see pull requests...
[09:54] <RAOF[m]> xnox: they're *meant* to be local:*'d away in the linker script (and, indeed, the actual text of the functions *is* not in the DSO), so it's *additionally* a ld.bfd bug. Which I'll file tomorrow.
[09:54] <Laney> I dunno what the process for those headers getting from khronos into mesa is; hopefully there is one
[10:15] <Laney> LocutusOfBorg: it'd be better if you would git merge properly rather than import-dsc
[10:16] <Laney> up until now those branches shared proper history with debian
[10:18] <LocutusOfBorg> oh... ok
[10:18] <LocutusOfBorg> should I force push?
[10:19]  * Laney looks the other way :-)
[10:19] <Laney> (yes, it's probably not been too long)
[10:24] <LocutusOfBorg> Laney, I don't get why you should look anywhere, looks like the history has never been pushed remotely
[10:25] <Laney> what?
[10:26] <LocutusOfBorg> the remote git repository is "clean" :)
[10:26]  * LocutusOfBorg feels dirty, as much as the remote is now clean
[10:27] <LocutusOfBorg> now I have time to properly do things...
[10:27] <LocutusOfBorg> in order to do it, I should: git merge debian/upstream pristine-tar and master, and then apply changes on top, right?
[10:29] <Laney> right
[10:29] <Laney> well upstream and pristine-tar will just be the same as what debian has, no commits there
[10:29] <Laney> you just push those branches and the upstream/xxx tag to launchpad
[10:30] <Laney> for the packaging branch, git merge the tag, which should keep our delta (may be some conflicts)
[10:30] <Laney> then check the various diffs
[10:30] <LocutusOfBorg> ack
[10:30] <LocutusOfBorg> will do after lunch maybe :D
[10:35] <LocutusOfBorg> Laney, can you please double-check?
[10:36] <Laney> LocutusOfBorg: ok, where should I look?
[10:38] <LocutusOfBorg> https://git.launchpad.net/~ubuntu-desktop/ubuntu/+source/gst-plugins-base1.0?h=master
[10:38] <LocutusOfBorg> and so on
[10:38] <LocutusOfBorg> I did base good and bad
[10:38] <Laney> ok, I would expect pristine-tar to be 6f933d105fc6d371a55bf516eab60725fac43861 and upstream to be 248418e33dcd27b99493cc56cc92872d662db276
[10:38] <Laney> for -base
[10:39] <LocutusOfBorg> and they are?
[10:39] <LocutusOfBorg> https://git.launchpad.net/~ubuntu-desktop/ubuntu/+source/gst-plugins-base1.0/log/?h=pristine-tar
[10:39] <Laney> let me check
[10:40] <Laney> looks good!
[10:41] <LocutusOfBorg> nice thanks! probably the force-push messed up a little bit the graphical page, it will refresh in some hours hopefully
[10:41] <Laney> then gbp buildpackage --git-tag-only
[10:41] <Laney> should give you the ubuntu/XXX tag to push once it's uploaded
[10:41] <LocutusOfBorg> I did that already
[10:42] <LocutusOfBorg> except for the good, because gbp.conf was badly set https://git.launchpad.net/~ubuntu-desktop/ubuntu/+source/gst-plugins-good1.0/commit/?id=8cf9b8fc1672cba9ad6690272f97f6dfa6c82f6f
[10:42] <LocutusOfBorg> I fixed but won't upload just because of that change
[10:42] <Laney> thx, you can make the tag manually
[10:43] <Laney> I think I only started using ubuntu/ tags on those repositories recently
[10:45] <LocutusOfBorg> it is already tagged as debian/1.14.2-1ubuntu1
[10:45] <LocutusOfBorg> so... ok lets copy the tag
[10:45] <Laney> :3
[10:46] <Laney> doesn't matter too much though
[10:47] <LocutusOfBorg> it matters, because I did in a bash loop the following "git diff debian/1.14.2-1 ubuntu/1.14.2-1ubuntu1"
[10:47] <LocutusOfBorg> and it failed :p
[10:47] <LocutusOfBorg> so, more automation with that change!
[10:47] <Laney> heh
[12:39] <tjaalton> LocutusOfBorg: next week when I'm back
[12:58] <LocutusOfBorg> sure thanks themill
[12:59] <LocutusOfBorg> s/themill/tjalton
[12:59] <LocutusOfBorg> :)
[12:59] <themill> 1 out of 7
[14:30] <abeato> sil2100, hi, I have an implementation for the ubuntu-image changes to be able to use a roorfs instead of lb: https://github.com/CanonicalLtd/ubuntu-image/pull/155
[14:30] <abeato> sil2100, there are probably some tests missing, but it would be good if you can take a look toi give thumbs up/down to the general approach
[14:45] <sil2100> abeato: sure, but only after the point-releases
[14:45] <sil2100> Since we're busy with those right now
[14:46] <sil2100> Thanks!
[15:03] <abeato> sil2100, sure, np
[17:15] <LocutusOfBorg> Laney, I don't really understand the new gl failure for plugins-good... any hint?
[17:47] <coreycb> jamespage: doko_ : fyi the py3.7 threading deadlock has been narrowed down to monkeypatching of stdlib thread modules combined with use of ThreadPoolExecutor.  more details at: https://bugs.python.org/issue34173
[18:41] <mvo> infinity: just in case you don't know already - your seed.yaml reorder workaround works and snap install gedit now works as expected as well
[20:37] <caravena> https://bugs.launchpad.net/bugs/1783363
[23:56] <Unit193> mwhudson: Hrm, did the retry not fix it?