[00:03] -queuebot:#ubuntu-release- Unapproved: xfce4-panel (disco-proposed/universe) [4.13.4-1 => 4.13.4-1ubuntu1] (mythbuntu, ubuntustudio, xubuntu)
[01:08] -queuebot:#ubuntu-release- Unapproved: gdb (disco-proposed/main) [8.2.90.20190311-0ubuntu1 => 8.2.91.20190405-0ubuntu1] (core)
[01:13] -queuebot:#ubuntu-release- Unapproved: accepted gdb [source] (disco-proposed) [8.2.91.20190405-0ubuntu1]
[01:51] -queuebot:#ubuntu-release- Unapproved: python2.7 (disco-proposed/main) [2.7.16-1 => 2.7.16-2] (ubuntu-desktop, ubuntu-server)
[01:56] -queuebot:#ubuntu-release- Unapproved: accepted python2.7 [source] (disco-proposed) [2.7.16-2]
[02:54] -queuebot:#ubuntu-release- Unapproved: gcc-defaults-ports (bionic-proposed/universe) [1.176ubuntu1.1 => 1.176ubuntu1.2] (no packageset)
[02:54] -queuebot:#ubuntu-release- Unapproved: gcc-defaults (bionic-proposed/main) [1.176ubuntu2.1 => 1.176ubuntu2.2] (core)
[08:11] -queuebot:#ubuntu-release- Unapproved: qtbase-opensource-src (disco-proposed/universe) [5.12.2+dfsg-2 => 5.12.2+dfsg-3] (kubuntu, qt5) (sync)
[08:13] <mitya57> ^^ infinity: this qtbase upload should fix the endianness issue we were talking about
[08:32] <infinity> mitya57: Excellent, I'll drop the kde hint and accept it.
[08:33] <mitya57> Thanks!
[08:33] <mitya57> (the package with failing tests was kdelibs4support)
[08:33]  * infinity nods.
[08:35] -queuebot:#ubuntu-release- Unapproved: accepted qtbase-opensource-src [sync] (disco-proposed) [5.12.2+dfsg-3]
[08:35] -queuebot:#ubuntu-release- Unapproved: accepted xfce4-panel [source] (disco-proposed) [4.13.4-1ubuntu1]
[08:36] <infinity> mitya57: Thanks for caring enough to look instead of just playing the "s390x no haz graphics, who cares" game.
[08:36]  * infinity really wishes we had more than one big-endian arch so s390x didn't look so "weird" to people.
[08:37] <mitya57> Actually here the problem was in a core function (related to string encoding). The graphics may still have bugs on s390x :-(
[08:37] <infinity> mitya57: Oh, I'm sure rendering is a mess, but yes, it was obviously an endian bug i nthe string encoding (and I mentioned as much when people asked me to ignore it :P)
[08:38] <infinity> mitya57: But I didn't know the code well enough to hunt down why and fix it, so thanks.
[08:38] <infinity> mitya57: Did you add a test upstream to go with the fix (does upstream Qt do deep nitpicking unit tests?)?
[08:39] <infinity> mitya57: Seems a bit odd that something like this would only be caught by a random test in a 3rd party consumer of the API.
[08:39] <infinity> Maybe Qt's testing strategy is "let KDE test it for us", I dunno.
[08:40] <mitya57> We are running tests in some Qt modules but not in qtbase.
[08:40] <mitya57> Because there are too many failures, and I don't have enough time to look at them all.
[08:40] <infinity> Someone needs to invent more hours.
[08:41] <mitya57> Yes!
[08:41] <mitya57> Maybe I will try to enable them at least for qtcore.
[08:41] <infinity> Would be nice if we could find a volunteer willing to go through the qtbase test suite on Debian/Ubuntu and XFAIL until it's green.
[08:41] <infinity> Then we'd have a (probably fairly short) list of stuff that needs looking at after that,.
[08:42] <infinity> Assuming the testsuite has a convenient way to twiddle that per-test.
[08:43] <mitya57> Some time ago (when Timo was the maintainer) we had a "110 files changed, 21678 insertions(+), 49 deletions(-)" patch.
[08:43] <mitya57> But we did not manage to port it from Qt 5.5 to 5.6 so it was dropped…
[08:43] <infinity> That's a lot of insertions.
[08:44] <infinity> Okay, I need to find a bed.
[08:44] <mitya57> https://salsa.debian.org/qt-kde-team/qt/qtbase/raw/ubuntu/artful/debian/patches/enable-tests.patch
[08:44] <mitya57> It starts with something one won't expect :)
[08:45] <mitya57> (We removed rfc3261.txt from the tarball because it is not DFSG compliant, and needed to replace it with something)
[08:45] <mitya57> But not counting this, it is still a large patch.
[08:45] <mitya57> Good night!
[08:47] <infinity> mitya57: Hahaha.  Okay, so the fake RFC is like 95% of the patch, that looks less scary with context. :P
[08:47] <mitya57> The rest is still 109 files…
[08:48] <infinity> mitya57: Having to QSKIP at the test level instead of some higher-level list of XFAILs looks a bit cumbersome to maintain, though.
[08:48] <mitya57> There are BLACKLIST files but I think they work only in upstream CI.
[08:49] <infinity> Compare with, say:
[08:49] <infinity> https://salsa.debian.org/glibc-team/glibc/blob/sid/debian/testsuite-xfail-debian.mk
[08:49] <infinity> Which is still not the most pleasant thing, but at least it's one file with no chance to conflict on every upstream import.
[08:49] <mitya57> That is not possible with Qt I'm afraid
[08:50] <infinity> Shame.
[08:50] <mitya57> There are files like https://code.qt.io/cgit/qt/qtbase.git/tree/tests/auto/corelib/thread/qthread/BLACKLIST, but they are per-directory, not global.
[08:50] <acheronuk> infinity: sorry. I'm so used to seeing KDE*something requiring a fix to adapt to new Qt, that I badly assumed :/
[08:51] <infinity> acheronuk: It's all good.  That's why we don't work in isolation.
[08:51]  * acheronuk nods
[08:52] <infinity> Right, bed.  For realz.
[08:53] <mitya57> Good night for realz! :)
[10:58] -queuebot:#ubuntu-release- Unapproved: ubuntu-mate-meta (disco-proposed/universe) [1.230 => 1.231] (ubuntu-mate)
[13:50] <doko> hmm, why does the gdb build hang?
[15:10] -queuebot:#ubuntu-release- Unapproved: gcc-8 (disco-proposed/main) [8.3.0-5ubuntu1 => 8.3.0-6ubuntu1] (core)
[15:13] -queuebot:#ubuntu-release- Unapproved: accepted gcc-8 [source] (disco-proposed) [8.3.0-6ubuntu1]
[15:13] -queuebot:#ubuntu-release- Unapproved: accepted ubuntu-mate-meta [source] (disco-proposed) [1.231]
[16:45] -queuebot:#ubuntu-release- Unapproved: cups-filters (disco-proposed/main) [1.22.3-1 => 1.22.4-1] (desktop-core, ubuntu-server) (sync)