[06:43] <joelkraehemann> hi all
[06:49] <joelkraehemann> Is it possible to get bug-fixes into bionic for specific package?
[06:53] <joelkraehemann> https://launchpad.net/ubuntu/+source/gsequencer
[06:54] <joelkraehemann> ^^ Since I was busy for next major release 2.0.x didn't take care much about the issues
[06:54] <joelkraehemann> Anyhow, I created a sync request
[06:54] <joelkraehemann> but it was ignored
[06:55] <joelkraehemann> I found additional bugs as revising every single file for the 2.0.x release
[06:55] <joelkraehemann> At my opinion we should provide an updated 1.4.x version to fix the issues
[06:56] <joelkraehemann> What do you need in order to update the package?
[09:03] <RAOF> joelkraehemann: https://wiki.ubuntu.com/StableReleaseUpdates
[11:58] <ricotz> stgraber, hi, is there some discussion which could have been followed for decisions like https://launchpad.net/ubuntu/+source/lxd/1:0.3
[11:59] <tkamppeter> doko, I will be there.
[12:00] <rbasak> ricotz: isn't that https://lists.ubuntu.com/archives/ubuntu-devel/2018-August/040455.html ?
[12:00] <Unit193> ricotz: AFAIK, you can still use lxc.
[12:05] <ricotz> rbasak, Unit193, yes and yes, although I fearing more packages moving without a deb alternative
[12:05] <ricotz> e.g. same seems to be planned for chromium
[12:05] <rbasak> I believe that's the way things are going for fast moving stuff.
[12:06] <rbasak> If you're willing to put in the effort to maintain debs in parallel, perhaps take it up in ubuntu-devel@ on how to make that work technically?
[12:06] <Unit193> ricotz: Oh dear, really?  I was going to say, depending on what it is then one can simply backport from Debian into a PPA.
[12:07] <Unit193> Alternativly you can hold the last real version, but that's not generally the best to do in regards to security.
[12:08] <ricotz> rbasak, I understand but I am not interested in having snapd running here, in case of chromium: having a software release every 2 month seems not fast moving imho
[12:08] <rbasak> It becomes increasingly difficult to maintain for security in an LTS when upstreams add new dependencies etc.
[12:08] <ricotz> Unit193, a sane backport seems hard with introduced epochs
[12:09] <ricotz> rbasak, correct
[12:09] <rbasak> I don't know about Chromium but that's what is happening with Firefox, why it's becoming increasingly painful for people maintaining this stuff (security team etc), and why snaps are created as a solution to that problem.
[12:09] <Unit193> That is true, and pinnning for that doesn't seem ideal.
[12:09] <rbasak> That's why those doing the maintenance are moving to snaps.
[12:09] <Unit193> rbasak: Do you have a list of candidates?
[12:10] <rbasak> Unit193: I don't. That's a good and reasonable question for the mailing list though I think.
[12:11] <rbasak> Separately, if somebody wants to maintain eg. Firefox ESR in universe as a deb as an alternative, I don't think anyone would be opposed to that (provided it comes with a realistic commitment to maintain for security for LTS, etc). But that's a question for the ML too.
[12:36] <jbicha> https://community.ubuntu.com/t/7978 < Chromium as snap replacing .deb
[12:48] <tkamppeter> doko, hi
[13:19] <michagogo> Question: would an SRU to enable a feature via a build flag be appropriate?
[13:19] <michagogo> Specifically, the option to compile XRDP with support for vsock connections
[13:20] <rbasak> There's a "For Long Term Support releases we sometimes want to introduce new features" exception noted in https://wiki.ubuntu.com/StableReleaseUpdates
[13:20] <michagogo> The version in Bionic has it enabled (at Microsoft’s request, it was turned on in Debian, and then synced over)
[13:20] <rbasak> With some requirements.
[13:21] <michagogo> But Xenial doesn’t, so if you want to enable Hyper-V Enhanced Sessions there you need to build from source and install it yourself
[13:21] <rbasak> cking: see https://lists.ubuntu.com/archives/ubuntu-devel-discuss/2018-September/018138.html and the next two emails. Was that you?
[13:21] <rbasak> Am I right in thinking that this is from an accidental upload to Debian, which replied without checking the signature first? If so, maybe we can just ask for these all to be dropped from the ML?
[13:22] <cking> rbasak, yeh, I was on stupidly strong pain killers that day, and messed up. can indeed these can be dropped from the ML
[13:22] <rbasak> michagogo: that sounds like it might be acceptable to SRU in principle under a number of existing exceptions, platoform/HWE enablement included.
[13:22] <rbasak> hggdh: ^
[13:22] <michagogo> rbasak: I’m not sure I understand the second and third sentences there
[13:23] <michagogo> The feature is there in bionic, which is good
[13:23] <rbasak> michagogo: I mean that from your description here it sounds like it's acceptable to SRU in principle under existing established exceptions.
[13:24] <rbasak> I don't know the specifics though, and we still need to take care not to impact existing users by causing unexpected behaviour changes.
[13:24] <michagogo> I meant the second and third sentences in that section of the SRU wiki page
[13:24] <michagogo> “They must not change the behaviour on existing installations (e. g. entirely new packages are usually fine). If existing software needs to be modified to make use of the new feature”
[13:25] <rbasak> Enabling a feature via a build flags falls under "existing software needs to be modified".
[13:25] <rbasak> We need to make sure that doing so won't impact existing users.
[13:26] <rbasak> (eg. some existing user using some function that now enables a protocol flag not supported on the user's server, causing something that previously worked to fail following the SRU)
[13:26] <rbasak> That type of thing.
[13:28] <doko> apw, xnox: I have a probembuilding the 4.18 cross kernel (includes)
[13:29] <doko>   CC      kernel/bounds.s
[13:29] <doko> gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
[13:29] <doko> gcc: warning: '-mcpu=' is deprecated; use '-mtune=' or '-march=' instead
[13:29] <doko> gcc: error: unrecognized command line option '-mmultiple'; did you mean '-imultilib'?
[13:29] <doko> make[5]: *** [Kbuild:21: kernel/bounds.s] Error 1
[13:29] <doko> does this ring a bell?
[13:30] <michagogo> rbasak: I see, I think. Maybe some time soon I’ll find time to look into it. Thanks!
[13:32] <doko> that goes into linux-source-4.18.0/arch/powerpc/Makefile in the big endian config
[13:39] <rbasak> tsimonq2, Skuggen, cpaelzer: mysql-5.7 5.7.23-2 uploaded to Debian. This should fix the ppc64el test issues. It will fail to migrate though due to component mismatch - we still need to upload a merge that drops the mecab b-d to undo the sync.
[13:39] <rbasak> I believe that's being coordinated between Skuggen and tsimonq2.
[13:43] <andrewsh> guys, is that a known issue that Unity is broken in the current cosmic?
[13:43] <andrewsh> (at least on my laptop)
[13:45] <andrewsh> or, better:
[13:45] <andrewsh> any pointers how to debug compiz[19701]: WARN  2018-09-14 15:37:46 unity.iconloader IconLoader.cpp:264 Unable to load icon . GThemedIcon .%20GThemedIcon%20unity-online-accounts%20unity-online-accounts-symbolic .%20GThemedIcon%20unity-online-accounts%20unity-online-accounts at size 64
[13:46] <Skuggen> rbasak: Right. I'll prepare an upload with the delta that drops the mecab dependency
[13:51] <jbicha> andrewsh: that sounds like Debian bug 908705
[13:52] <andrewsh> hmm
[13:52] <andrewsh> jbicha: should I try to revert glib?
[13:52] <jbicha> andrewsh: no, that's not very practical at all
[13:52] <andrewsh> I mean locally
[13:53] <andrewsh> I wonder why this doesn't affect GNOME
[13:53] <jbicha> andrewsh: what I suggest is building with this commit: https://gitlab.gnome.org/GNOME/glib/merge_requests/305
[13:54] <andrewsh> hmm
[13:57] <jbicha> Philip (who is upstream glib) did comment on the Debian bug so my instinct is to wait for him to review before pushing that patch to Debian/Ubuntu
[13:57] <andrewsh> okay, let's build and test it locally
[13:58] <andrewsh> oh noes, the OBS is busy rebuilding the world
[13:58] <andrewsh> I should use a local sbuild then
[13:58] <jbicha> I believe ppa's still work, the rebuild has low priority
[13:59] <jbicha> I'm assuming you don't mean OpenSUSE?
[14:00] <andrewsh> Open Build System
[14:00] <andrewsh> our (Collabora's)
[14:00] <jbicha> I assumed wrong then :)
[14:01] <jbicha> andrewsh: see our rebuild the world: https://launchpad.net/builders/
[14:02] <andrewsh> I don't think OBS distinguishes between first builds and rebuilds
[14:02] <andrewsh> and I don't think there are multiple priorities
[14:04] <andrewsh> I’ve started a rebuild of a Ubuntu-derived distro yesterday, and it's still on-going…
[14:04] <teward> doko: before I go running additional tests, why are we demoting pcre3 in favor of pcre2?
[14:04] <teward> seems to me like that'd be a backwards step
[14:05] <teward> unless I
[14:05] <teward> I've missed something*
[14:05] <jbicha> teward: pcre3 is poorly named, pcre2 is newer
[14:05] <teward> jbicha: that explains that one.  Can we update the bug to make a note of that so that we can clarify?  (Just for documentation purposes)
[14:07] <jbicha> teward: please do :)
[14:08] <teward> bug text updated.
[14:08] <teward> running build tests for nginx now
[14:21] <hggdh> rbasak: I understand, then,  that I can block emails from ftpmaster@ftp-master.d.o from being received by ubuntu-discuss. Correct?
[14:23] <tsimonq2> rbasak: ack
[14:24] <tsimonq2> Skuggen: If you don't have upload access though, I would prefer to sponsor your patch rather than Just Doing It.
[14:26] <rbasak> hggdh: agreed
[14:27] <hggdh> thank you, will do
[14:28] <teward> jbicha: doko: NGINX is good to go for pcre2
[14:32] <joelkraehemann> RAOF, thank you
[15:14] <andrewsh> jbicha: running on that patch now, works fine here
[20:35] <andrewsh> jbicha: uh-oh, compiz now broken too :/
[20:35] <andrewsh> shadows etc
[20:37] <andrewsh> hmm, I’m told a related issue is already reported as https://bugs.launchpad.net/bugs/1725443 but it is private, apparently
[20:39] <jbicha> bug 1725443 is kinda old
[20:42] <andrewsh> so, in Appearance → Behaviour, I now have "Low" enabled instead of "High"
[20:43] <andrewsh> if I toggle it back, the launcher renders incorrectly, and if I restart X, it never appears, presumably Unity crashes
[20:43] <andrewsh> or rather compiz
[20:44] <andrewsh> by incorrectly I mean missing or inverted colours, black background instead of transparency etc etc
[20:44] <jbicha> sorry, I don't use compiz
[20:44] <andrewsh> well, it's part of Unity, that’s why I use it