15:01:30 <rdieter> #startmeeting kde-sig
15:01:30 <zodbot> Meeting started Tue Nov 28 15:01:30 2017 UTC.  The chair is rdieter. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:30 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:01:30 <zodbot> The meeting name has been set to 'kde-sig'
15:01:32 <rdieter> #meetingname kde-sig
15:01:32 <zodbot> The meeting name has been set to 'kde-sig'
15:01:36 <rdieter> #topic roll call
15:01:44 <rdieter> hi all, who's present for friendly kde-sig meeting?
15:01:52 * than is present
15:01:57 <pino|work> o/
15:02:01 <lupinix> .hello2
15:02:02 <zodbot> lupinix: lupinix 'Christian Dersch' <lupinix.fedora@gmail.com>
15:03:01 <rdieter> #info rdieter than pino|work lupinix present
15:03:04 <rdieter> #chair than pino|work lupinix
15:03:04 <zodbot> Current chairs: lupinix pino|work rdieter than
15:03:34 <Southern_Gentlem> .hello jbwillia
15:03:35 <zodbot> Southern_Gentlem: jbwillia 'Ben Williams' <vaioof@gmail.com>
15:03:47 <rdieter> #info Southern_Gentlem present
15:03:47 <tosky> o/
15:03:48 * jreznik_ is here
15:03:52 <rdieter> #info jreznik_ present
15:04:36 <mbriza> hello
15:04:57 <rdieter> #info mbriza present
15:05:06 <rdieter> #info tosky present
15:05:15 <rdieter> #topic agenda
15:05:19 <rdieter> ok, what to discuss today?
15:05:31 <rdieter> Kevin_Kofler had mentioned qt-5.9.3 in #fedora-kde earlier
15:05:53 <rdieter> I can include that as well as other recent updates I've been working on
15:07:49 <tosky> no other points from me - thanks for the steady stream of updates
15:08:21 <lupinix> nothing else here
15:08:38 <rdieter> ok
15:08:43 <rdieter> #topic qt-5.9.3 status
15:09:16 <rdieter> so, thanks to jgrulich doing much of the heavy lifting late last week, got qt-5.9.3 imported into rawhide
15:09:41 <rdieter> and I helped with all the associated rebuilds (the list of private-api consumers is growing :( )
15:10:26 <rdieter> did most of the work in a side tag to avoid any intermediate broken deps in rawhide (f28-kde)
15:10:28 <lupinix> thanks
15:10:38 <lupinix> rdieter++
15:10:44 <lupinix> jgrulich++
15:10:45 <zodbot> lupinix: Karma for jgrulich changed to 1 (for the f27 release cycle):  https://badges.fedoraproject.org/tags/cookie/any
15:10:49 <pino|work> jgrulich++
15:10:54 <rdieter> had f26 kf5-5.40.0 builds ready for a little while, finally submitted to bodhi yesterday
15:11:35 <rdieter> oh, and got pim stuff from kde-apps-17.11.80 imported into rawhide too (needed since libical-3.0 had been introduced)
15:11:58 <rdieter> one more thing to mention for transparency...
15:12:22 <rdieter> f28+ kf5-related macros updated %_kf5_metainfodir from /usr/share/appdata to /usr/share/metainfo
15:12:40 <pino|work> go for it
15:12:47 <rdieter> so... now's the time to check any package hard-coding .../appdata instead of using %_kf5_metainfodir
15:13:07 <rdieter> since it will fail now :-P
15:13:45 <rdieter> sorry, I guess I through in all updates there, not just qt 5.9.3
15:13:52 <rdieter> threw even
15:14:01 <tosky> there was a note about qt 5.9.3 and kio on kde-distro-packagers
15:14:03 * tosky checks
15:14:14 <rdieter> I may have missed that
15:14:19 <tosky> https://mail.kde.org/pipermail/kde-distro-packagers/2017-November/000297.html
15:14:25 <tosky> kio needs a fix
15:14:34 <tosky> and I think I've seen another relevant change $somewhere
15:15:24 <rdieter> thanks, that probably means we don't want to rush qt-5.9.3 to -updates
15:15:45 <rdieter> Kevin_Kofler: ^^ , fyi
15:15:47 <tosky> and there is probably https://phabricator.kde.org/D9029 too
15:16:29 <tosky> I've seen also Arch packagers mentioning an issue with digikam and Qt 5.9.3 (as if digikam didn't have enough)
15:16:47 <tosky> https://bugs.kde.org/show_bug.cgi?id=387373
15:17:20 <rdieter> also think it unfortunate Qt introduced such a change into a minor release, but I'm sure they have their "reasons(tm)"
15:17:47 <rdieter> tosky: thanks
15:17:57 <rdieter> jgrulich++
15:17:57 <zodbot> rdieter: Karma for jgrulich changed to 2 (for the f27 release cycle):  https://badges.fedoraproject.org/tags/cookie/any
15:18:01 <rdieter> (overdue cookies)
15:18:04 <rdieter> tosky++
15:18:05 <zodbot> rdieter: Karma for tosky changed to 1 (for the f27 release cycle):  https://badges.fedoraproject.org/tags/cookie/any
15:18:48 <jgrulich> hi
15:19:08 <rdieter> #info jgrulich present
15:19:09 <rdieter> hi
15:19:24 <Kevin_Kofler> rdieter: OK
15:19:42 <rdieter> I'll review this qt-5.9.3 fixes/patches, etc... after meeting, and import into packaging as appropriate
15:19:45 <Kevin_Kofler> WTF at Qt deliberately breaking POSIX compliance…
15:19:51 <Kevin_Kofler> foo//bar == foo/bar in POSIX
15:20:15 <pino|work> iirc not
15:21:20 <Kevin_Kofler> Maybe not in the letter of the standard, but all the *nix OSes implement things that way.
15:21:46 <Kevin_Kofler> Declaring the construct invalid is just the worst possible thing to do, it will never do what the user expects.
15:22:50 <rdieter> I think we can debate the merits of the Qt change outside of this meeting, point is... there's a problem, and we need to take things slowly and deliberately
15:25:00 <rdieter> and including the workarounds should be a safe thing to do either way
15:25:48 <rdieter> this feels similar to the changes introduced in QUrl awhile back
15:25:57 <rdieter> that broke a lot of stuff
15:26:28 <lupinix> and that for a bugfix release…
15:26:34 <lupinix> they should read https://semver.org/
15:27:16 <rdieter> I'm sure it's not that simple
15:28:02 <lupinix> i am
15:28:13 <lupinix> but thats an imho
15:28:47 <rdieter> does anyone know offhand why the change was introduced?
15:29:05 <rdieter> otherwise, we can discuss it more at length later
15:29:39 <rdieter> without that information, we're only speculating
15:31:30 <rdieter> #topic open discussion
15:31:42 <rdieter> anything else to mention or discuss today?
15:33:00 <Kevin_Kofler> In other news, 5.9.3 made it out to the Rawhide master mirror, so I can build qt5-qtwebengine-freeworld-5.9.3 there now.
15:36:09 <Kevin_Kofler> qt5-doc is outdated in Rawhide, by the way.
15:36:38 <Kevin_Kofler> Any plans to get rid of that monolithic thing that does not match the actual qt5-* packages?
15:37:12 <rdieter> I think it's mostly harmless, nothing depends on it matching exactly, and I doubt the docs changed significantly in a minor release
15:37:45 <rdieter> so we have the luxery of updating at our convenience, or even not at all if needed
15:37:49 <Kevin_Kofler> I want at least the QtWebEngine docs removed from qt5-doc, you disabled them in QtWebEngine without my permission (I explicitly told you not to do it), and I want QtWebEngine to come with matching docs.
15:38:39 <rdieter> Kevin_Kofler: please file a bug detailing exactly what you want omitted, and I'll take care of it
15:38:53 <rdieter> since qtwebengine releases independantly, it makes less/no sense to include in qt5-doc
15:39:31 <rdieter> probably same for qtwebkit, but there may have more bootstrapping consequences
15:39:49 <Kevin_Kofler> I think I'll reenable docs for F27 at least, since I'm going to push 5.9.3 before you push the rest of Qt, this will end up having both versions on the mirrors, but 5.9.3>5.9.2.
15:40:10 <Kevin_Kofler> (reenable in qt5-qtwebengine.spec, I mean)
15:40:37 <Kevin_Kofler> And I am going to work on 5.10.0 soon, it should be released within the next few days.
15:40:51 <rdieter> <nod>, yay
15:40:57 <Kevin_Kofler> But I want 5.9.3 out first, so I will likely not rush to 5.10.0 on the day of the release.
15:41:37 <rdieter> will have to think on that a bit too, since 5.9.x is LTS probably want to continue tracking that *somewhere*
15:41:59 <rdieter> f26 branch maybe
15:42:29 <Kevin_Kofler> My experience with the previous LTS branch (5.9) was horrible.
15:42:35 <Kevin_Kofler> (5.6, I mean)
15:42:56 <rdieter> that was their first try, hopefully better this time
15:42:56 <Kevin_Kofler> It did not get security fixes for months and I had to ship a snapshot to get security fixes out to F24 before its EOL.
15:43:26 <Kevin_Kofler> F25 is sticking on 5.9 LTS. Anything newer, just no.
15:43:27 <rdieter> again, I'm only speaking of core qt5 modules, we already treat qtwebengine separately
15:43:37 <Kevin_Kofler> That's for QtWebEngine.
15:43:59 <Kevin_Kofler> If you stick to Qt 5.9 LTS on F26, that should be workable for QtWebEngine, it should keep building against 5.9 LTS for a while.
15:44:40 <Kevin_Kofler> 5.10 dropped support for building against 5.7, I think it even still builds against 5.8.
15:45:05 <Kevin_Kofler> Things are likely to keep building against LTS Qt.
15:45:13 <lupinix> well, also depends on how newer versions work… e.g. qt 5.8 was a mess (which is why we had non lts 5.7 for very long time…)
15:45:21 <lupinix> *how well
15:46:22 <Kevin_Kofler> Well, my Konversation on F25 still eats characters in links when it wordwraps them, IIRC that was fixed in Qt 5.8, but that was never pushed to F25. :-/
15:46:51 <Kevin_Kofler> There are some annoying bugs in Qt 5.7 that did not get fixed because Qt 5.8 was not pushed, this is one of them.
15:47:29 <lupinix> well because 5.8 introduced new annoying bugs… but we discussed that several times ;)
15:48:51 <Kevin_Kofler> qt5-qtwebengine-freeworld-5.9.3 in Rawhide will have to wait a bit, the RPM Fusion builders have outdated metadata cached and are trying and failing to fetch outdated packages. (I need to file a bug for that, they should really not be caching metadata at all.)
15:51:59 <rdieter> ok, looks like we're winding down, thanks everyone
15:52:04 <rdieter> #endmeeting