15:18:56 <itchka> #startmeeting
15:18:56 <chwido> Woof! Let's start the meeting. It's Wed Jun 15 15:18:56 2016 UTC. The chair is itchka. Information about me at https://wiki.openmandriva.org/en/Chwido.
15:18:56 <chwido> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:18:56 <chwido> You should add extra chair(s) just in case, with #chair nick.
15:18:56 <chwido> Have a good meeting, don't bark too much!
15:19:10 <itchka> #chair bero
15:19:10 <chwido> Current chairs: bero itchka
15:19:18 <itchka> Ok folks
15:19:32 <itchka> Only one thing on the agenda today.
15:19:40 <itchka> #item 1
15:20:00 <itchka> #Release Report
15:20:10 <itchka> #topic Release Report
15:20:30 <itchka> Where are we at bero?
15:21:46 <bero> Looking good from where I stand -- aside from UI not coming up on fedya's box, the only major problem seems to be calamares misbehaving in install mode (but working perfectly when started from live mode, so there's a workaround)
15:22:40 <bero> For everything else, we've already moved cooker to llvm 3.8.1rc1, mesa 12.0rc2, qt 5.6.1, KF 5.23 to run some initial sanity checks
15:22:54 <bero> probably Qt at least should be pulled into 3.0, no problems caused
15:23:10 <bero> for mesa, the reason for testing the update is hoping it'll resolve the problem with fedya's box
15:23:27 <itchka> I have seen several problems fixed in Libre Office  and Firefox which IO was going to report but others beat me too it.
15:23:59 <bero> but a cooker iso failed to build with dependency issues (being resolved) so we couldn't check for sure whether or not mesa changes anything on that box
15:25:00 <bero> that's pretty much all from my side -- I think we need to get that release out.
15:25:07 <itchka> There is more than one issue with Calamares. Going back after doing partitioning but not yet writing to disk results in a crash.
15:26:02 <itchka> i586 is still broken around udev in dracut. It's the most bizarre problem.
15:26:43 <bero> haven't checked i586 yet -- any details?
15:30:00 <itchka> bero: yes I can boot a minimal iso but if I make anything  else it will not boot. Dracut seems to get stuck in an endless loop with udev settle waiting for what appears to be the isofile.
15:30:18 <bero> weird...
15:30:20 <itchka> I'll post a screen shot hangon
15:32:25 <fedya> +
15:33:06 <fedya> bero: i tested 12.0.rc1 mesa
15:33:11 <fedya> issue not solved
15:33:16 <fedya> even with latest kernel firmware
15:33:30 <fedya> i think now that issue may be in libinput drivers
15:33:57 <itchka> https://s31.postimg.org/bjxdpg4sb/Virtual_Box_Lx3_15_06_2016_16_30_32.png
15:33:58 <bero> do you have any unusual input devices?
15:34:02 <itchka> bero: ^^^
15:34:21 <bero> everyone is using libinput, so unless you have anything weird attached, I don't see why that would happen only on your box...
15:34:38 <fedya> bero: i have mac keyboard!
15:34:54 <fedya> and second kebyboard just standart Genius
15:35:03 <fedya> ps\2 + usb mac keyboard
15:35:16 <bero> doesn't sound too unusual... does the crash still happen if you remove it?
15:35:18 <itchka> fedya: I have a Mac keyboard here I can test if you want.
15:36:24 <bero> PS/2 might be another issue... I haven't seen one of those for some time (but then again I think most laptop keyboards are PS/2 internally)
15:36:30 <bero> anyone else using a PS/2 device?
15:37:33 <bero> We could try dropping libinput in favor of evdev again
15:37:37 <bero> just to make sure
15:37:43 <bero> but it would be good to really know what's causing this
15:39:13 <itchka> My box has ps/2 so I could reproduce fedya's situation
15:41:05 <itchka> fedya: You are getting boot failure?
15:48:17 <bero> I think sddm crashes his X server
15:48:21 <christann> The system that I am having problems with has an AUSU P5K motherboard.
15:49:16 <bero> christann: what problems?
15:50:38 <christann> With ISO on June 9, I cannot boot into live mode. Stops when trying to get SDDM started up. Installing doesn't work either.
15:51:18 <christann> This is an 8 yr old PC.
15:51:31 <christann> Will try soon with the latest Cooker ISO.
15:51:55 <bero> hmm... might be the same as fedya's problem
15:52:03 <bero> any details on what happens there?
15:52:38 <bero> fedya: is the box you're having problems with an older box as well?
15:53:16 <bero> I wonder if we have an SSSE3 instruction or something like that somewhere... Would give an Illegal Instruction on older CPUs while not being noticable on newer ones
15:53:41 <itchka> christann: Is this a 64bit machine.
15:54:12 <bero> If it's only 8 years old, it almost certainly is
15:54:19 <christann> Yes with a quad processor.
15:56:21 <itchka> I have an old PC of similar ilk I'll run it up later.
15:58:20 <itchka> bero: Did you take a look at the screenshot.
15:58:23 <bero> I have a 7yr old box here, ASRock P55M Pro with an i7-860... That one works normally
15:58:53 <bero> itchka: yes, can't really see what the problem there is though... No real indication of what's going wrong
15:59:56 <OnlyHuman> yarr
16:04:25 <itchka> Any ideas how I might narrow it down. It seems to me that there is a package being installed which is perhaps asking for a device to be created which is in conflict with the device being used for the iso image.
16:04:33 <itchka> bero: ^^
16:05:07 <itchka> There can't be many packages with that potential.
16:05:31 <bero> no idea... I don't think it's any package's fault because this looks like something that would happen way before the real system even boots up...
16:05:48 <bero> we should run this by TPG, he's the closest to an expert on dracut we have
16:06:05 <itchka> bero: Yes but a minimal iso boots so the dract initrd image must be ok.
16:06:38 <bero> is there any difference between how you set those up? (e.g. one attached as virtual CD drive and the other as usb stick or whatever)?
16:08:20 <itchka> bero: Identical methods of creation and test. the only difference being the packages installed in the live image.
16:08:50 <bero> that's really odd... packages should never affect what goes into dracut
16:09:44 <fedya> itchka not boot failure. only radeonsi driver issue
16:10:03 <itchka> Surely udev has to start devices in dracut? If a package adds a device to udevs device list then dracut initrd should honour it.
16:10:16 <fedya> itchka: https://issues.openmandriva.org/show_bug.cgi?id=1706
16:10:38 <itchka> bero: Or am I wrong in this?
16:12:33 <fedya> bero my pc is AMD FX(tm)-8320 Eight-Core Processor
16:12:40 <fedya> it's not an ancient hardware
16:14:12 <fedya> motherboard ASUS M5A97 EVO R2.0
16:14:42 <fedya> i tried latest iso on laptop, that i bought in 2010
16:14:44 <fedya> working well
16:14:51 <fedya> issue only with my PC
16:15:00 <fedya> i will disable ps\2 keyboard
16:15:02 <fedya> and look again
16:19:04 <itchka> fedya: Does the older radeon driver work or is that not up to date enough for your card?
16:36:40 <itchka> bero: I just checked there are different sets of udev rules for the minimal versus a basic X build iso. Do you think it's valid to remove the additional rules one by one and rebuild the mksquashfs and iso and then try booting again?.
16:49:23 <bero> sorry, was on the phone, parents calling
16:49:41 * bero wonders how libinput can cause a crash in the graphics driver as opposed to in the X server
16:50:11 <bero> but I guess we can take it out, unless we're dealing with multitouch displays, libinput doesn't really bring advantages compared to evdev
16:50:29 <bero> but of course wayland will require libinput, so we'd better sort this out by the time we're ready to move to it
16:53:56 <itchka> Just had a GP fault
16:55:19 <christann> Just tried cooker ISO 337 (yesterday) on my old PC and I can run live in basic-graphics mode. Trying to install in basic-graphics mode - The first calamares screen appears for about a second then the system reboots.
16:55:53 <christann> It has a NVIDIA 7300GT video card using a DVI-HDMI cable to a TV set.
16:56:12 <christann> Works fine with 2014.0 and nvidia 304 video drivers.
16:56:30 <bero> itchka: you missed:
16:56:34 <bero> [18:49:23] <bero> sorry, was on the phone, parents calling
16:56:34 <bero> [18:49:41] bero wonders how libinput can cause a crash in the graphics driver as opposed to in the X server
16:56:34 <bero> [18:50:11] <bero> but I guess we can take it out, unless we're dealing with multitouch displays, libinput doesn't really bring advantages compared to evdev
16:56:34 <bero> [18:50:29] <bero> but of course wayland will require libinput, so we'd better sort this out by the time we're ready to move to it
16:56:48 <bero> and as for udev rules, can't hurt, but I don't think the X udev rules actually get into the initrd
16:57:35 <bero> christann: install not working (but install from live working) is a known bug, not related to your box
16:57:47 <christann> bero: Thanks
16:59:42 <itchka> bero: I have found the issue in i586 I was right it is the udev rules. It's the 91-drm-modeset.rules. If I remove those the dracut hang goes away.
17:00:57 <bero> Hmm... can you reproduce that on real hardware then? DRM+modeset going wrong is something that could essentially come down to "VirtualBox doesn't have a proper drm driver"
17:01:25 <bero> And there's nothing in those rules that should be able to cause problems...
17:01:34 <bero> all that file says is
17:01:37 <bero> KERNEL=="controlD[0-9]*", SUBSYSTEM=="drm", MODE="0666"
17:02:09 <bero> so in other words "if drm tries to generate a controlD[0-9]* device, make it mode world-writable"
17:02:47 <cc^mint> bero!
17:03:16 <bero> hi cc^mint
17:03:25 <cc^mint> :-)
17:03:28 <itchka> Well yes but would they have it working on x86_64 and not on i586 seems unlikely anyways I'm not sure there's a drm driver for x86_64 otherwise wouldn't plymouth work properly?
17:04:44 <itchka> bero: I have an idea for another test it will only take five minutes
17:04:45 <bero> itchka: true, the architecture shouldn't make a difference there
17:05:03 <itchka> kernel perhaps?
17:06:40 <bero> that can always be the problem...
17:07:23 <itchka> I'm building the minimal iso with the drm rules included
17:08:20 <christann> On my old PC, started up in live mode, ran calamares from root. After setting disk partitions and entering names and passwords, calarames crashes.
17:09:02 <christann> Last message is: Cannot mix incompatible Qt library (version 0x50600) with this library (0x50601).
17:11:35 <itchka> That sounds a bit dire
17:26:31 <itchka> Well we need to get back to our next course of action. Do we release a BETA I honestly don't think we are at rc status. What do others think?
17:28:59 <christann> I have no problem with releasing a BETA
17:29:19 <bero> "Cannot mix incompatible Qt library (version 0x50600) with this library (0x50601)" means you somehow got qt from cooker
17:30:03 <bero> let's release a beta, but we can't delay things forever
17:30:07 <bero> we're already a year late...
17:30:55 <christann> I am using the latest cooker ISO - 337 because I wanted to use the new version of Mesa.
17:31:53 <ben79> Where is iso 337 latest I see is 331
17:32:56 <itchka> bero: So do we push the cooker updates or at least some of them? Plasma for me is sometimes goes crazy in quite an unpredicatle manner. Is it any better in cooker?
17:33:03 <ben79> Oh, I see, it's a Cooker iso
17:34:13 <itchka> bero: More bizarreness minimal iso boots with the drm rules but doesn't actually create the device nodes....
17:37:00 <christann> Could be related to the fact that I am running in basic graphics mode.
17:37:42 <christann> When I try to start up in live mode - regular graphics, kwin crashes. I have the core dump and will post a bug report.
17:41:23 <cc^mint> Mandrake!
17:43:34 <HisShadow> yeah, Mandrake my first linux distro
17:44:57 <itchka> Mine too
17:45:47 <christann> Mine also.
17:46:13 <rugyada> et's release Beta2 to have more test & feedback
17:46:25 <rugyada> let's*
17:46:50 <bero> itchka: My cooker box is behaving well, but I haven't seen any craziness before we did the updates, so I'm not sure if the fixes are related...
17:47:10 <bero> Let's backport some of the cooker changes after we've made the official beta2 iso
17:47:54 <bero> Qt 5.6.1 is definitely a good idea to have, it fixes around 70 bugs and hasn't caused any regressions (except for that "Cannot mix incompatible Qt library" thing which is caused by the fact that the iso was built in the middle of updating)
17:50:42 <itchka> Alright but I really think we should make some sort of plan to concentrate resources on some of the more major issues left. Particularly the Calamares install issue and the i586 booting. We need to get these sorted out before any rc imho.
17:51:08 <bero> agreed
17:53:51 <christann> I will wait for another cooker ISO and try my old PC again. Probably early next week since we are going to Ottawa (big city) for the weekend.
17:53:59 <itchka> Ok lets go then I think though we need to look at the release notes before we push it just to warn people that there are still issues. We don't to dissappoint people more than we have to.
17:54:20 <itchka> lib/udev
17:55:53 <rugyada> write beta2 release notes the known issue, maybe with links to bugs in bugzilla
17:56:01 <itchka> bero I have put all the udev rules from the udev lock iso into the minimal iso and the minimal iso still boots so It has to be a seconary effect.
17:56:15 <itchka> secondary
17:56:52 <bero> that's what I thought
17:57:28 <bero> maybe it's plymouth acting up
17:57:56 <itchka> I was booting in simple graphic mode and there plymouth=0
17:58:56 <itchka> to be precise plymouth.enable=0
18:09:21 <rugyada> what's still missing for beta2 release?
18:10:31 <itchka> How do you mean missing?
18:11:06 <rugyada> what still has to be done before release
18:12:50 <itchka> The release notes need some work I guess though I haven't looked at them in a while so I'm not sure how much. There will need to be a blog anouncement written.
18:15:05 <rugyada> I mean the technical side. Just a new build (or not even that) ?
18:15:44 <itchka> A new build basically as far as I'm aware.
18:15:54 <rugyada> ok
18:16:49 <ben79> OK, a new .iso needs to be built. Any time for testing QA approval?
18:16:53 <rugyada> install doesn's start from menu but it does from live, so people can install. just need to put a note.
18:17:52 <rugyada> let's try to build asap and to release after a couple of days testing. we could be ready for this weekend then
18:17:55 <rugyada> ?
18:19:05 <rugyada> QA approval - keeping in mind beta status.
18:19:40 <ben79> There would need to  be a regression from .iso 331 for me to not approve
18:20:11 <rugyada> ben79: ?
18:21:07 <rugyada> if new iso is not worst than 331 will get approval from you, you mean?
18:21:39 <ben79> rugyada: Meaning if we were voting on .iso 331 right now for Beta2 release I'd vote Go
18:21:44 <rugyada> ok
18:21:50 <rugyada> agree
18:21:59 <christann> +1
18:22:03 <rugyada> we are testing ISOs all the time now
18:22:10 <fedya> itchka omv2014.0 with x11-driver-video-ati-7.5.0-0.1-omv2014.0.x86_64 working fine
18:22:13 <rugyada> guess we know them very well :)
18:23:28 <rugyada> that's why I can think we can have beta2 release before/at time of weekend
18:23:46 <ben79> So assuming we build a new .iso with Qt 5.6.1 and perhaps KFramework 5.23 and nothing gets borked then it would be OK to release.
18:23:59 <HisShadow> I've returned the product build lists page so you can pick latest stuff right away
18:24:04 <HisShadow> https://abf.openmandriva.org/product_build_lists
18:24:26 <ben79> Can we get a new .iso today?
18:24:29 <itchka> ben79: we would not be shipping those yet
18:24:41 <rugyada> Qt 5.6.1 after beta2 said bero
18:24:45 <rugyada> iirc
18:24:51 <ben79> Then why don't we use .iso 331?
18:25:15 <itchka> There may have been other updates..
18:26:08 <ben79> There have been to python, libreoffice, among others but nothing I'd care enough to build a new .iso
18:26:52 <itchka> ben79: There was some bad stuff in Firefox
18:26:53 <rugyada> better a fresh build imho
18:27:10 <itchka> It didn't remember it's tabs
18:27:37 <itchka> really, really annoying bug. I wouldn't ship without that fix.
18:28:01 <bero> we should probably try to get the new mesa tested
18:28:02 <ben79> OK, then can we get a new .iso today?
18:28:23 <bero> it doesn't seem to have any regressions in Cooker and while it doesn't fix 1706, it seems to be a good step forward
18:28:36 <bero> since it's only 1 package, won't take long to backport
18:28:37 <ben79> And schedule Go/NoGo vote and release for when?
18:28:46 <bero> should be able to build an iso right after that's in
18:29:06 <christann> Monday or Sunday evenng?
18:29:19 <rugyada> Friday?
18:29:28 <itchka> bero: Sorry which package are we talking about here?
18:29:39 <bero> itchka: mesa
18:30:17 <rugyada> bero: new mesa in beta2 ?
18:30:29 <bero> yes
18:31:09 <itchka> but you say it's not fixing anything?
18:34:16 <rugyada> bero: leave new mesa in cooker, so its update won't delay beta release. people can test cooker iso.
18:34:21 <rugyada> imho :)
18:34:44 <bero> rugyada: it won't get delayed, the package is already building
18:34:54 <bero> itchka: it's fixing stuff, just not 1706 (which we hoped it would fix)
18:37:14 <rugyada> providing it will not cause any regression nor any delay then it could be ok.
18:37:32 <christann> I agree. Include the mesa update.
18:38:05 <itchka> bero: Ok as long as it's bringing something to the party. We have to concentrate on stabilising things somehow and I know that just building the next release is sometimes a quick fix but sometimes it brings with another complely different world of grief.
18:38:48 <bero> that's why I'd rather have stuff like that going into the beta so we can see whether or not it breaks anything
18:39:08 <bero> I'd also say qt 5.6.1 should go into the beta, but that would delay it given how long it takes to build
18:39:13 <bero> so let's keep that for the rc
18:39:35 <rugyada> better...
18:40:46 <itchka> If mesa is going in we will have to do some basic testing to ensure there aren't any issues with the graphics side of things.
18:40:54 <rugyada> bero: we have cooker for experiments
18:41:33 <bero> yes, it has been in cooker for some time now and nobody detected regressions
18:41:53 <rugyada> is not this the point to have unstable cooker and rather stable cooker ? :))
18:42:49 <rugyada> the "thrilling cooker" LOL
18:44:19 <rugyada> bero: tester are focusing in testing 3.0 branch right now
18:44:38 <bero> precisely
18:44:51 <bero> that's why something that should go into 3.0 final needs to be tested on the 3.0 branch now
18:45:23 <bero> much better to get it in now than to get a no-go on the rc because it has some undetected regression
18:45:52 <rugyada> sure. but beta2 has to be released in few days
18:46:53 <rugyada> there will be the time to test new (for 3.0) stuff soon after release, imho.
18:47:11 <rugyada> just my opinion yet.
18:48:25 <rugyada> ben79: christann: ?
18:49:41 <christann> I am happy to release now (ie. 331) or with the mesa update (after a day or so of testing. But would like to see beta2 as soon as convenient.
18:50:25 <ben79> Yes
18:50:46 <ben79> I'm getting tired of testing a never ending series of .iso's
18:50:54 <ben79> have other things that need my time
18:51:10 <bero> mesa should finish building in 20 minutes, then another 30 or so to build a new iso...
18:51:25 <bero> I'd say let's try that and if there's any problem whatsoever let's just release 331 so we don't risk delaying
18:51:38 <ben79> Also part of why we're taking so long on this is that we never stop adding new packages, most distros have a freeze as part of release process
18:51:38 <bero> but if it builds fine, let's get it some testing
18:52:11 <bero> we have one too
18:52:34 <christann> sounds good to me!
18:52:44 <bero> but when something is likely to fix a lot of things without risking much, we make exceptions
18:52:47 <bero> the way it should be
18:53:11 <bero> the main reason why we're taking so long is that there's a few things that need to be fixed before the final and that aren't fixed yet
18:53:17 <rugyada> gotta go for now, family calls. will read the log later. bye.
18:53:47 <itchka> Yes if we have a fallback situation where we are not going backwards I guess it's ok.
18:53:50 <bero> that causes some stuff to be added despite the freeze (because we've seen on cooker that the new version is better), not vice versa
18:54:19 <itchka> So it's all hands on deck to check the next build.
18:54:31 <ben79> There is always and always will be something or other that fixes a lot of things, it never ends
18:55:14 <ben79> Again, I'm getting tired of testing a never ending series of .iso's, feel like my time is being abused
18:57:39 <itchka> I know ben79 This has been a tough release but we are nearly there we have forked the repos so we are well on the way.
18:57:48 <christann> I have to go. I still am content with updating mesa, then testing a new ISO.
18:57:56 <itchka> Let's not give up now
19:03:39 <itchka> bero: will you post on cooker list when iso is ready please.
19:03:48 <bero> sure
19:05:31 <itchka> Ok I think I'll call the meeting to a close now. Just do a quick share.
19:08:29 <itchka> #share This weeks meeting brought us to the point where there will be a BETA2 as soon as abasic testing is done. There will be annoucement about availablility soon. This will be an x86_64 arch release we hope to overcome the boot issues with i586 soon. If all goes well with testing this BETA2 will bring the latest mesa.
19:08:40 <itchka> #endmeeting