[Releasing] Meeting minutes: Qt release team meeting 10.03.2014

Heikkinen Jani Jani.Heikkinen at digia.com
Tue Mar 11 07:16:12 CET 2014

Meeting minutes from Qt Release Team meeting 10.03.2014
* Qt 5.3 Alpha feedback
	- No major known issues, issues to be fixed before 5.3.0 Beta 1 release on  https://bugreports.qt-project.org/browse/QTBUG-36712 
* Qt 5.3 Beta schedule
	- Target date for Qt 5.3 Beta is 20th March 2014 (http://qt-project.org/wiki/Qt-5.3-release)
	- Snapshot build expected to be available starting of 11th March
* Mingw version change for 4.8.6 & 5.3.0:
	- Mingw-Builds i686-4.8.2-release-posix-dwarf-rt_v3-rev2 has issue with gdb packaged in there which crashes when used from Qt Creator.
	- Changing from Mingw-Builds i686-4.8.2-release-posix-dwarf-rt_v3-rev2 to Mingw-Builds i686-4.8.2-release-posix-dwarf-rt_v3-rev3 which has new gdb package should be minimal risk.
	- Plan B is to revert back to the known Mingw 4.8.0 version.
	- Kai will do additional tests and confirm version change
* Qt 4.8.6: 
	- New Qt 4.8.6 snapshot build available. See http://lists.qt-project.org/pipermail/development/2014-January/014910.html 
	- Some patches required for Qt 4.8.6 release. Unfortunately Qt 4.8 CI with Mac 10.6 configuration is broken (under investigation).
* Next meeting on 17th March 2014 16:00 CET

Irc log below:
(5:00:52 PM) akseli1: kkoehne: sahumada: thiago: fkleint: ZapB: tronical: wolfgang-b: vladimirM: aholza: peter-h: mapaaso: ankokko ping
(5:01:04 PM) peter-h: akseli1: pong
(5:01:06 PM) ankokko_: akseli1: pong
(5:01:13 PM) sahumada: akseli1: pong
(5:01:14 PM) mapaaso: akseli1: pong
(5:01:23 PM) fkleint: akseli1: pong
(5:01:33 PM) fkleint: No Jani today?
(5:01:50 PM) vladimirM: akseli1: pong
(5:01:54 PM) thiago: akseli1: pong
(5:01:56 PM) wolfgang-b: akseli1: pong
(5:02:16 PM) akseli1: fkleint: unfortunately no.. 
(5:02:40 PM) fkleint: Argh
(5:03:03 PM) akseli1: fkleint: jaheikki3 had to skip meeting on last minute 
(5:03:25 PM) fkleint: akseli1: I sent mail to Lars & him about the printer stuff , not sure if there was a conclusion
(5:03:50 PM) kkoehne_: akseli1: I think we need to update the Mingw version once more for 4.8.6/5.3.0
(5:04:19 PM) akseli1: fkleint: argh, and im not aware of that printer stuff or conclusion about it
(5:04:43 PM) fkleint: Hm, ok, lets go by agenda
(5:05:02 PM) akseli1: On agenda today
(5:05:11 PM) akseli1: Qt 5.3 Alpha release feedback
(5:05:11 PM) akseli1: Qt 5.3 schedule
(5:05:11 PM) akseli1: Mingw version for 4.8.6 & 5.3.0
(5:05:11 PM) akseli1: Qt 4.8.6 status
(5:05:11 PM) akseli1: Any additional items?
(5:06:19 PM) akseli1: starting with Qt 5.3 Alpha feedback .. 
(5:06:59 PM) akseli1: jaheikki3 has requested issues to be fixed before 5.3.0 Beta 1 release link to https://bugreports.qt-project.org/browse/QTBUG-36712
(5:07:14 PM) akseli1: No really major issues from alpha so far i think but does someone have additional items based on Qt 5.3 Alpha feedback which should be e.g. included to that metabug?
(5:09:47 PM) thiago: I guess not
(5:09:48 PM) akseli1: i guess that means no 
(5:10:15 PM) akseli1: that leads us to Qt 5.3 schedule then and especially 5.3 beta schedule
(5:10:27 PM) akseli1: Current planned schedule for Qt 5.3 beta is 20th Mar 2014 (http://qt-project.org/wiki/Qt-5.3-release)
ankokko ankokko_ 
(5:11:11 PM) akseli1: Qt5.3 beta installer configurations has been set up and we are expecting to have daily snapshot builds starting tomorrow, ankokko_: & mapaaso: please correct if i'm wrong
(5:11:27 PM) sahumada: akseli1: do we have snapshots already ?
(5:11:40 PM) sahumada: ah .. you just answered :P
(5:11:52 PM) ZapB_: sorry, missed the ping. here now
(5:12:08 PM) mapaaso: akseli1: that's the plan.
(5:12:10 PM) fkleint: Oki..is RTA up and running?
(5:13:42 PM) akseli1: fkleint: yes, RTA up and running
(5:14:06 PM) fkleint: Cool..since there was some GL fiddling let's hope it did not break sth
(5:16:14 PM) akseli1: so based on current information there is no blocker so far for us to achieve qt5.3 beta on 20th Mar
(5:16:37 PM) ***thiago agrees
(5:16:48 PM) fkleint: akseli1: That depends on that stupid printing stuff
(5:17:16 PM) fkleint: akseli1: PLease have a chat with Jani & Lars about it..we should reach some conclusion
(5:17:45 PM) akseli1: fkleint: will call to jani after this meeting and have converstatioin about that subject
(5:17:50 PM) akseli1: *i will
(5:18:04 PM) fkleint: Thanks, please include Lars
(5:18:21 PM) akseli1: fkleint: yep
(5:18:44 PM) sahumada: akseli1: there is a P1 in the list for Beta 1 https://codereview.qt-project.org/77710 (Enginio source files missing) .. what happened with that ? I recall this email from fregl http://lists.qt-project.org/pipermail/development/2014-January/014910.html .. but I dont know what was the final outcome of it
(5:19:04 PM) sahumada: https://bugreports.qt-project.org/browse/QTBUG-37117 .. is the bug report
(5:19:22 PM) akseli1: sahumada: and there was already some patch for that if i recall correctly
(5:19:40 PM) sahumada: akseli1: so will enginio be part of the official set of add-ons modules ?
(5:20:08 PM) fregl: sahumada: I think the current understanding is that it will be next to the other modules as it was in the 5.2 online installer, but also in the offline installer
(5:22:55 PM) akseli1: hopefully we see enginio on offline installers are tomorrow
(5:23:10 PM) akseli1: *already tomorrow
(5:23:35 PM) akseli1: then Mingw version for 4.8.6 & 5.3.0 .. kkoehne_ ^
(5:24:13 PM) kkoehne_: akseli1: Well, yeah. So far we planned to ship both 4.8.6 and 5.3.0 binary packages with Mingw-Builds i686-4.8.2-release-posix-dwarf-rt_v3-rev2
(5:24:28 PM) kkoehne_: The issue is that the gdb packaged in there crashes when used from Qt Creator...
(5:24:34 PM) kkoehne_: Not exactly sure why.
(5:24:50 PM) kkoehne_: The good news is that Mingw-builds just released a new -rev3 today, with a newer gdb.
(5:25:00 PM) kkoehne_: Which seems to work. I'm still testing ...
(5:25:15 PM) kkoehne_: But it looks like we should move from  Mingw-Builds i686-4.8.2-release-posix-dwarf-rt_v3-rev2 ->  Mingw-Builds i686-4.8.2-release-posix-dwarf-rt_v3-rev3
(5:25:35 PM) kkoehne_: Apart from the gdb the Mingw-w64 headers were updated. But the gcc version etc is the same, so I think the risk is low.
(5:25:37 PM) thiago: makes sense
(5:25:40 PM) thiago: debugging is important
(5:25:43 PM) thiago: what's plan B?
(5:26:07 PM) kkoehne_: Well, Plan B would be to find out exactly why it crashes, and whether we can work around it from Qt Creator side.
(5:26:20 PM) kkoehne_: But honestly speaking just 'upgrading' sounds easier to me :)
(5:27:30 PM) thiago: right, but what happens if you run into more issues?
(5:27:40 PM) kkoehne_: Anybody feels strongly against switching to rev3 (given that I don't find any showstoppers still)? I know it's work to upgrade the build machines etc... In theory it should be actually enough 'just' to update the shipped package.
(5:27:55 PM) kkoehne_: thiago: Well ... if all else fails we could move back to the known 4.8.0 version.
(5:28:09 PM) kkoehne_: thiago: The one we shipped already with 5.2.1 etc
(5:28:26 PM) thiago: ok, that's plan B
(5:28:31 PM) thiago: looks good to me
ankokko ankokko_ 
(5:29:45 PM) fkleint: What is the story with the Qt 4.8 integration wrt Mac 10.6. Apparently, some breakage crept in?
(5:29:59 PM) akseli1: upgrading sounds like a plan .. kkoehne_ do you need additional time with testing or should upgrade tomorrow already?
(5:30:48 PM) kkoehne_: akseli1: I'd like to do some more minor testing , but be done tomorrow morning.
(5:31:01 PM) kkoehne_: akseli1: I can give you a heads-up then .
(5:31:09 PM) akseli1: kkoehne_: ok, sounds good, thanks
(5:31:51 PM) akseli1: fkleint: status is a bit uncertain .. we can't get ci 48 integrating on any ci nodes but
(5:32:32 PM) fkleint: akseli1: Does that mean regressions crept in and the release is basically broken? I still don't quite understand the issue?
(5:32:35 PM) akseli1: with some machines outside ci system can't find anything wrong .. problem seems to be quite random anyway
(5:32:52 PM) fkleint: Hm
(5:33:06 PM) akseli1: and guys investigatin are not sure what is wrong or where is the root cause
(5:34:52 PM) akseli1: i'm not convinced that release is broken but we havent found solution for ci either .. investigation is ongoing
(5:35:26 PM) akseli1: and that pretty much summarizes qt 4.8.6 status except that there is new snapshot available (http://lists.qt-project.org/pipermail/development/2014-March/016011.html)
(5:35:51 PM) thiago: good
(5:35:56 PM) akseli1: and which might require new mingw version tomorrow:)
(5:36:06 PM) thiago: fair enough
(5:36:09 PM) thiago: btw, kkoehne_:
(5:36:24 PM) thiago: it would be really nice if there was a big link to where one should download the MinGW version
(5:36:36 PM) akseli1: thiago: link available on installer
(5:36:41 PM) thiago: it's surprisingly hard to find out which one we had used
(5:36:53 PM) thiago: to the exact tarball?
(5:37:02 PM) thiago: not a download side. We need exact version and build.
(5:37:05 PM) kkoehne_: thiago: True. We've a link in the installer already, but putting it on the download page too doesn't hurt.
(5:37:15 PM) akseli1: thiago: but i agree it could be documented properly
(5:37:50 PM) thiago: I'm not sure I understood what the problem is though
(5:37:52 PM) thiago: the Mac packages are there
(5:37:56 PM) thiago: what's the poblem?
(5:38:06 PM) sahumada: thiago: the CI is broken since Feb 25th
(5:38:21 PM) sahumada: not the packaging machines
(5:38:46 PM) fkleint: thiago: Check https://codereview.qt-project.org/#change,77936 
(5:38:54 PM) fkleint: thiago: the failures there..Mac again and again
(5:39:37 PM) fkleint: thiago: apparently it is only 10.6, but it is not clear whether breakage crept in ..Simo said it only happens in virtual machines or sth (or vv)
(5:39:56 PM) thiago: so we have changes we should have in 4.8.6 but are failing to integrate
(5:39:57 PM) thiago: correct?
(5:40:01 PM) fkleint: yes
(5:40:11 PM) fkleint: I mean, not super high prio
(5:40:13 PM) fregl: I think the issue is that we only rarely try to build this - we should have regular builds that show that the system is working
(5:40:15 PM) fkleint: but it should work
(5:40:25 PM) thiago: are those changes fixing issues that are blockers for 4.8.6?
(5:40:37 PM) fkleint: Support would say so, but no.
(5:40:40 PM) fkleint: ;-)
(5:40:54 PM) fkleint: but 4.8 integration should just work
(5:40:56 PM) thiago: if we can't reproduce the issues outside the Mac VMs, on 10.6, then we should be reasonably sure it's not a breakage we caused
(5:40:58 PM) fregl: well, it's pretty bad if we can't get our changes in, blockers or not
(5:40:59 PM) akseli1: thiago: one of the fixes is addressing carbon build .. cant remember how many fixes tried
(5:41:00 PM) thiago: agreed
(5:41:14 PM) fkleint: OTOH, https://codereview.qt-project.org/#change,80434 is a kind of blocker
(5:41:18 PM) fkleint: regression crash
(5:41:50 PM) thiago: ok, then we wait for the fix
(5:41:58 PM) sahumada: the problem with Mac 10.6 is that the farm was removed and we left only one machine for Qt 4.8 (a second machine was added afterwards) .. and it seems that we left the faulty ones 
(5:42:29 PM) fkleint: Urgs
(5:43:10 PM) thiago: are we not testing 10.6 for Qt 5?
(5:43:12 PM) fregl: yes, afaict there are two 10.6 machines in jenkins, one of them disabled, leaving one machine
(5:43:16 PM) sahumada: thiago: no
(5:43:31 PM) fregl: we stopped with 5.3, right?
(5:44:48 PM) akseli1: yep, with 5.3
(5:45:30 PM) fregl: looks like the 10.6 machine has the same failures always. last successful build was feb 25th
(5:46:11 PM) sahumada: fregl: yes, because is the same faulty machine :)
(5:46:14 PM) akseli1: and replacing that 10.6 with another ones didn't help either
(5:46:45 PM) sahumada: I think that's a topic for tomorrow's CI meeting :)
(5:46:53 PM) akseli1: sahumada: +1
(5:47:02 PM) fregl: sounds sensible
(5:47:16 PM) akseli1: so i think that leaves us one item which is next meeting
(5:47:41 PM) fregl: have you considered updating on how we plan to make parallel releases in the future?
(5:47:41 PM) akseli1: which is on next Monday at the same time if no one has anything against that?
(5:47:52 PM) fregl: there was quite some internal discussion inside Digia
(5:47:55 PM) fkleint: sounds good
(5:48:05 PM) fregl: I guess this might be a good forum to get more people up to speed
(5:48:31 PM) akseli1: and that could be a perfect topic for next meeting:)
(5:48:33 PM) fregl: how do I get on the list of people pinged for this meeting?
(5:48:46 PM) sahumada: fregl: the calendar invite ?
(5:48:55 PM) fregl: sahumada: nope, the ping
(5:48:57 PM) fregl: on irc
(5:49:00 PM) thiago: wait
(5:49:02 PM) ***fregl never notices
(5:49:11 PM) sahumada: fregl: <akseli1> kkoehne: sahumada: thiago: fkleint: ZapB: tronical: wolfgang-b: vladimirM: aholza: peter-h: mapaaso: ankokko ping
(5:49:23 PM) thiago: I can't join next week, since I have to travel to California
(5:49:29 PM) fregl: sahumada: yes, I'm not in there
(5:49:37 PM) thiago: feel free to schedule it at a different time
(5:50:07 PM) akseli1: fregl: i'll make certain that you will be on next time
(5:50:15 PM) fregl: akseli1: thanks :)
(5:50:54 PM) akseli1: thiago: you are not able to join next week?
(5:51:01 PM) thiago: not at this hour
(5:51:15 PM) akseli1: thiago: sorry , missed your previous commet
(5:51:15 PM) thiago: I need to be at the airport at 7:00 am
(5:51:49 PM) thiago: actually, I might be able to join at this time, if I can get the flight on the Intel shuttle
(5:51:53 PM) thiago: then I can join from the plane
(5:52:00 PM) thiago: but I'm still waitlisted...
(5:52:43 PM) akseli1: so meeting next week at the same time:)
(5:52:53 PM) thiago: when do you guys enter DST?
(5:52:58 PM) thiago: two weeks from now?
(5:53:00 PM) fkleint: March 21..?
(5:53:01 PM) fkleint: one sec
(5:53:31 PM) fkleint: No
(5:54:14 PM) fkleint: 30.3.
(5:54:35 PM) thiago: ok, so next week it's 8 am for me. We're fine.
(5:54:56 PM) sahumada: /etc/localtime  Sun Mar 30 00:59:59 2014 UTC = Sun Mar 30 01:59:59 2014 CET isdst=0 gmtoff=3600
(5:54:56 PM) sahumada: /etc/localtime  Sun Mar 30 01:00:00 2014 UTC = Sun Mar 30 03:00:00 2014 CEST isdst=1 gmtoff=7200
(5:55:08 PM) fkleint: hopes QTimeZone represents  it correctly
(5:56:37 PM) akseli1: so .. if nothing else .. i would like to thank you all for the meeting:)

More information about the Releasing mailing list