16:00:35 #startmeeting F36-blocker-review 16:00:35 Meeting started Mon Apr 25 16:00:35 2022 UTC. 16:00:35 This meeting is logged and archived in a public location. 16:00:35 The chair is adamw. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions. 16:00:35 Useful Commands: #action #agreed #halp #info #idea #link #topic. 16:00:35 The meeting name has been set to 'f36-blocker-review' 16:00:39 #meetingname F36-blocker-review 16:00:39 The meeting name has been set to 'f36-blocker-review' 16:00:40 #topic Roll Call 16:00:48 ahoyhoy folks, who's around for a short blocker review meeting? 16:01:20 .hello2 16:01:21 bcotton: bcotton 'Ben Cotton' 16:01:41 .hello2 16:01:42 coremodule: coremodule 'Geoffrey Marr' 16:01:52 * coremodule is here, willing to act as secretary. 16:03:41 .hello 16:03:42 lruzicka: (hello ) -- Alias for "hellomynameis $1". 16:03:48 .hello2 16:03:49 lruzicka: lruzicka 'Lukáš Růžička' 16:04:47 .hello2 16:04:48 cmurf: cmurf 'Chris Murphy' 16:05:26 #chair lruzicka cmurf 16:05:26 Current chairs: adamw cmurf lruzicka 16:05:34 .hello geraldosimiao 16:05:35 geraldosimiao: geraldosimiao 'Geraldo S. Simião Kutz' 16:05:39 boilerplate alert! 16:05:41 .hello jbwillia 16:05:42 #topic Introduction 16:05:44 Southern_Gentlem: jbwillia 'Ben Williams' 16:05:46 Why are we here? 16:05:48 #info Our purpose in this meeting is to review proposed blocker and nice-to-have bugs and decide whether to accept them, and to monitor the progress of fixing existing accepted blocker and nice-to-have bugs. 16:05:52 #info We'll be following the process outlined at: 16:05:54 #link https://fedoraproject.org/wiki/QA:SOP_Blocker_Bug_Meeting 16:05:57 #info The bugs up for review today are available at: 16:06:00 #link http://qa.fedoraproject.org/blockerbugs/current 16:06:04 #info The criteria for release blocking bugs can be found at: 16:06:07 #link https://fedoraproject.org/wiki/Basic_Release_Criteria 16:06:09 #link https://fedoraproject.org/wiki/Fedora_36_Beta_Release_Criteria 16:06:12 #link https://fedoraproject.org/wiki/Fedora_36_Final_Release_Criteria 16:06:23 #info for F36 Final, we have: 16:06:31 #info 2 Proposed Blockers 16:06:31 #info 4 Accepted Blockers 16:06:35 #info 3 Proposed Freeze Exceptions 16:06:35 #info 13 Accepted Freeze Exceptions 16:06:47 #info coremodule will secretarialize 16:06:58 #topic Proposed Final Blockers 16:07:05 #topic (2078442) Searching iso crashes virtual machine manager 16:07:08 #link https://bugzilla.redhat.com/show_bug.cgi?id=2078442 16:07:34 #link https://pagure.io/fedora-qa/blocker-review/issue/778 16:07:35 #info Proposed Blocker, gtk3, NEW 16:07:35 so, sumantro hit a crash in the GTK file chooser. but nobody else could reproduce it yet. I found an upstream report from four months ago which looks a lot like the same thing. 16:08:11 seems gtk ends up passing null to a glib function that expects a string, but I don't think we know why yet. but it doesn't seem super common or you'd expect more reports/dupes. 16:09:09 didn't sumantro also say it's fixed when applying the latest updates from testing 16:09:40 yeah, he said that 16:10:05 yeah, i am actually looking at glib and gtk rn to see if anyone changed anything obvious there 16:13:12 on the whole i'm -1 if this isn't commonly encountered 16:13:23 yeah, i think i agree 16:13:37 particularly since running VMs from a live doesn't seem like a common use case (or maybe i'm wrong?) 16:14:12 well, i think this could potentially happen in anything that uses the file selector 16:14:16 We almost haven't space on a live to install vm 16:14:22 don't think it matters what the app is 16:14:58 that would make sense. i'd be willing to grant an FE in case we decide we need another RC 16:15:02 adamw: Good point 16:15:37 adamw: So we need more testing with this on mind 16:15:49 geraldosimiao: people could use the live to run a VM stored on other media (e.g. a USB stick or the local hard drive). i just don't think most people _do_ 16:15:51 * So we need more testing with this in mind 16:16:06 bcotton: Oh yes. You're right 16:16:51 bcotton, they could but I do not think anybody does as this would be terribly slow. 16:17:04 I am also thinking I'm a -1, I agree with bcotton that it's not a common usecase and that we've had several people try to reproduce to no avail. 16:17:17 geraldosimiao: this is the sort of thing i'd kinda expect people to run across 'naturally' 16:17:17 -1 from me, also 16:17:36 -1 blocker, +1 FE 16:17:41 +1 fe 16:18:03 -1 Blocker 16:18:03 +1 FE 16:19:08 +1 FE 16:20:00 +1 FE 16:20:21 Kvm works fine for me 16:20:33 On F36 no such issues 16:20:43 BETA 1.4 iso 16:21:31 NishantMishra[m], BETA 1.4 iso is old as a dinosaur. 16:21:45 I updated it 16:21:52 With all the new stuff 16:22:13 lruzicka: ,😂😂😂 16:22:37 i am not seeing anything in udpates-testing that would fix this. but hey. 16:22:43 I have only that iso if anyone is willing to link me to RC1 16:22:43 NishantMishra[m], the problem is that you could skip an erroneous version that might be stuck on one of the composes where the problem happens. 16:23:32 adamw, sumantro claims that: the 16:23:33 xdg-desktop-portal-1.12.4-1.fc36.x86_64 16:23:33 fixed it 16:23:49 NishantMishra[m], all right, just wait a mom 16:24:03 proposed #agreed 2078442 - RejectedBlocker (Final) AcceptedFreezeException (Final) - this seems to be triggered by some specific local file or directory and doesn't seem like it happens often enough to be worth blocking on, but any crash in the file chooser is worth fixing on the live image, so FE granted 16:24:13 ack 16:24:16 ack 16:24:19 NishantMishra[m]: https://fedoraproject.org/wiki/Test_Results:Fedora_36_RC_1.1_Summary 16:24:19 lruzicka: if you read the whole bug, that seems not the whole story 16:24:21 NishantMishra[m], https://kojipkgs.fedoraproject.org/compose/36/Fedora-36-20220422.0/compose/Workstation/x86_64/iso/Fedora-Workstation-Live-x86_64-36-1.1.iso 16:24:30 lruzicka: what sumantro actually did is update from updates-testing and say it went away. with several hundred packages. 16:24:39 virt-manager doesn't even use the portal, i don't think. 16:24:52 adamw, yeah but he sort of swore on that package in a discussion with kparal 16:24:56 but there's no gtk3 update in updates-testing. there's a glib update, but i don't see how it could possibly fix this bug. 16:25:08 lruzicka: i think that's not right, though. :D 16:25:24 adamw, I am just saying .... 16:25:29 it could, i guess, be some sort of odd quirk where the search no longer happens to hit the broken path after the update for whatever reason 16:25:31 this is a long codepath 16:26:05 adamw, but yeah, I did not have the feeling the files were opened through a portal cause the portal file selecter opens differently 16:27:40 anyhoo 16:27:45 one more ack? 16:27:56 .hello ngompa 16:27:57 Eighth_Doctor: ngompa 'Neal Gompa' 16:28:00 ack 16:28:19 Ack 16:28:36 and my ax^W ack! 16:28:36 for virt-manager to use the portal, the API usage needs to change for opening the file selector 16:28:46 that is unlikely to happen before it gets ported to gtk4 anyway 16:30:15 #agreed 2078442 - RejectedBlocker (Final) AcceptedFreezeException (Final) - this seems to be triggered by some specific local file or directory and doesn't seem like it happens often enough to be worth blocking on, but any crash in the file chooser is worth fixing on the live image, so FE granted 16:30:22 #topic (2076596) The KDE ibus panel is not able to switch keyboard layouts on Wayland 16:30:24 #link https://bugzilla.redhat.com/show_bug.cgi?id=2076596 16:30:28 #link https://pagure.io/fedora-qa/blocker-review/issue/766 16:30:30 #info Proposed Blocker, ibus, ASSIGNED 16:31:13 so i re-proposed this as a blocker because geraldosimiao made a good point: the warning that was included in the fix is, of course, in english. and it doesn't have any translations 16:31:40 this is a bit of a shame for a warning that's likely gonna be mainly relevant to people whose first language isn't english 16:31:47 that would be... painful 16:31:58 but also, how quickly can we reasonably get it translated? 16:32:01 Sorry for that peeps lol 16:32:09 so, to recap briefly: the problem is that the ibus input switcher cannot switch keyboard layouts (only input methods) in KDE-on-Wayland 16:32:27 Let me check my friends at ln 10 pt-br 16:32:54 the "fix" we went with is to disable this functionality in the ibus switcher entirely, and add a message that says "Layout changes do not work in Plasma Wayland. Please use systemsettings5." 16:32:56 when this is not translated it will affect all foreign languages, won't it? 16:33:04 Conan Kudo: yeah, this is a concern 16:33:23 I don't know if we can block on something we can't reasonably resolve 16:33:26 lruzicka: this would primarily matter to languages which 'natively' use input switching 16:33:30 russian is the major one there 16:33:35 there's a few others 16:33:40 * marcdeop[m] is late to.the party but can help with Spanish, Catala a d German translations if needed 16:33:50 it's additionally tricky because this was done in a downstream patch 16:33:53 it's not merged upstream 16:34:09 I think this is going to be one of those "too difficult to fix" things then :( 16:34:20 adamw, it depends ... I do not use input switching, for instance, but some Czech users do. 16:34:28 like most of i18n/l10n stuff depends on upstream integration 16:34:28 so i don't know how hard the mechanism would be to get translations for a downstream patch into the package build, without the string actually being in upstream git 16:35:43 and also: what criterion does this violate? afaict, we only require available translations to be displayed. we don't require them to exist 16:35:58 Actually, the icon for IBUS switchers is made invisible in the GUI so users might not click on it. However, it leaves an empty spot in the lower bar and if you click into it, you get the message. 16:36:28 there also doesn't seem to be any kinda-sorta similar existing string we could use instead, sadly. 16:36:54 I think I would probably not block on this simply because of how difficult it would be to fix it 16:37:09 lruzicka: yes, the 'disablement' part of the fix is still 'valid' for all languages, of course 16:37:23 Ibus phantom menace 😂😂 16:37:35 Ben Cotton (he/him): well, it's a case of whether the initial bug is fully 'resolved' if the explanation we added to 'resolve' it isn't translated, basically 16:37:47 to be clear, I'm speaking specifically on the translation stuff 16:38:09 I just don't know how we'd generate translations for a patched thing reliably 16:38:10 What's the criteria about translating warnings? 16:38:15 And release 16:38:33 practically speaking we may just have to say 'well this sucks' and be ok with it, i think :| 16:38:41 yup 16:38:48 we could of course document it as a common bug. in english. haaaaah. 16:38:53 we can commonbugs it with translations :P 16:39:05 I assume someone translates commonbugs 16:39:08 we could hardwire all the translations into the patch, but I am ok with leaving it as is. 16:39:09 "this issue you can't understands explains why you can't understand this message" 16:39:11 adamw: Put the corrects translation on the common bugs page, in others languages 16:39:21 or maybe that doesn't happen anymore since mattdm's whole forum thing makes that harder 16:39:24 adamw: Yeah lol 16:39:29 common bugs is on ask.fp.o now. i dunno if there's a translation mechanism there. we, uh, should've thought about that. 16:39:40 there isn't 16:39:58 you could plug in google translate widget 16:40:07 but that actually doesn't help in this specific instance 16:40:39 well, we could advise users to translate it for themselves using the google translation tool 16:40:46 (joke intended) 16:40:49 it doesn't look like the f35 common bugs page was translated to any other languages either, so, hey, at least we're consistently bad! 16:40:56 anyhoo. 16:40:56 yay? 16:41:13 we should consider moving commonbugs to docs and setting up translations there 16:41:20 but that's another thing for another problem for another time :) 16:41:28 lruzicka: And call it an Easter Egg 16:42:05 so yeah, i think this is sufficiently resolved for the purposes of the release criteria 16:42:17 proposed #agreed 2076596 - consider sufficiently resolved (remove proposed blocker status) - we discussed whether the existing 'fix' is sufficient since the note it added explanining the problem is not translated. we agreed this was unfortunate but we have no realistic way to resolve it in an acceptable timeframe, and at least the panel element is indeed disabled, so the criterion is satisfied 16:42:28 adamw: +1 16:42:28 i would like it to be more fixed, but i won't require it 16:42:32 ack 16:42:44 ack 16:42:49 I guess ack 16:43:40 Easter Egg: if you find the hidden icon and click it, you get an English warning... 16:43:53 😁 16:44:14 Who is saying Fedora is not an adventure game? 16:44:19 #agreed 2076596 - consider sufficiently resolved (remove proposed blocker status) - we discussed whether the existing 'fix' is sufficient since the note it added explanining the problem is not translated. we agreed this was unfortunate but we have no realistic way to resolve it in an acceptable timeframe, and at least the panel element is indeed disabled, so the criterion is satisfied 16:44:24 * Easter Egg: if you find the hidden icon and click it, you get an warning in English. .. 16:44:26 People pay for them, you know? 16:44:30 plot twist: the new monkey island game is actually fedora 36 16:44:45 🎉 16:45:00 #topic Proposed Freeze Exceptions 16:45:06 let's do these just in case we do have to respin 16:45:08 there's only 3 16:45:23 oh, and two of them have sufficient votes in ticket, so i'll skip those and clean them up after th emeeting 16:45:31 #topic (2078230) Missing VA-API encode support for VP8 and VP9 16:45:34 #link https://bugzilla.redhat.com/show_bug.cgi?id=2078230 16:45:35 +1 16:45:38 #link https://pagure.io/fedora-qa/blocker-review/issue/775 16:45:40 #info Proposed Freeze Exceptions, ffmpeg, ON_QA 16:45:46 #info Ticket vote: FinalFreezeException (+2,2,-0) (+ngompa, +lruzicka, geraldosimiao, kparal) 16:45:49 #info Ticket vote: 0Day (+1,0,-0) (+geraldosimiao) 16:46:42 actually, no 16:46:43 not sure we should really encourage people to do complex things with the web browser on the live images since it'll inevitably have a zillion known security bugs in two months, but...uh...i guess sure 16:47:28 the package in question isn't shipped on lives, is it? 16:47:36 ffmpeg ? 16:47:49 I'm not really sure on this here 16:47:52 not yet no 16:48:00 bcotton: no 16:48:19 if it's not shipped on anything, it doesn't need an FE... 16:48:21 then it does not need to be FE, I guess 16:48:36 I wasn't even going to bother until someone asked me to make sure that was part of the GA repos 16:48:43 so I shrugged and figured why not 16:49:10 is there a compelling reason it shoudl be in the GA repos and not in updates? 16:49:40 * this here (the proposal as FE or Day0) 16:49:47 yeah, i'm fine with this being 0-day. 16:49:48 -1 16:50:04 mostly in the "hardware (useful) enablement" category 16:50:16 but it really doesn't matter if it's zero day or GA repos that much 16:50:29 okay, -1 FE from me then 16:50:35 since it's not on the media right now, it's pretty academic either way 16:50:46 well, if somebody is installing from netinst and wants to install it, it could come handy 16:50:51 yeah 16:51:01 and it ensures that the base functionality is in the buildroot 16:51:02 but this will only happen on Day1 and then it will go away 16:51:21 especially for folks who like building from GA repos (which is somewhat common) 16:51:32 but again, meh 16:51:33 netinst uses updates repo by default anyhow 16:53:10 lruzicka: are you -1 now? 16:53:15 yeah 16:53:19 -1 16:53:31 that puts us at -3 ... neal, you're the last +1 16:54:03 well, I'll keep my votes then: -1 FE and +1 0-day 16:54:10 I'd prefer FE, merely because it's not a huge deal to ship as FE 16:54:18 between FE vs 0Day, I'd prefer FE 16:54:41 the policy is things get an FE for a specific reason, because we don't just want to push everything through the freeze. 16:55:00 you could say we can fix any bug 'for people who want to build from the frozen repos' or something. 16:55:14 well, since updates-testing got turned off in the repo, nobody gets these until later now 16:55:19 otherwise I pretty much wouldn't care 16:56:44 adamw: for things that aren't on the media, I wish we were more permissive for FEs 16:56:59 but meh 16:57:04 anyhow, i think we're at -4 / +1 now 16:57:05 so 16:57:45 proposed #agreed 2078230 - RejectedFreezeException (Final) - there is no specific justification for giving this a freeze exception, it is not on any media so far as we're aware. It can go out as a 0-day update. 16:58:01 ack 16:58:05 ack 16:58:09 ack 16:59:05 ack 16:59:59 #agreed 2078230 - RejectedFreezeException (Final) - there is no specific justification for giving this a freeze exception, it is not on any media so far as we're aware. It can go out as a 0-day update. 17:00:03 I expect that we will have ffmpeg preloaded on lives next cycle though, as we onboard software that depends on it 17:00:09 #topic Accepted Blockers 17:00:42 #info 2073708 and 2071259 are verified 17:01:21 Eighth_Doctor: that would be nice, no more need to use the rpmfusion for that 17:02:12 #topic (2070764) selinux-policy is preventing flatpak from updating / installing / removing flatpaks 17:02:15 #link https://bugzilla.redhat.com/show_bug.cgi?id=2070764 17:02:19 #link https://pagure.io/fedora-qa/blocker-review/issue/702 17:02:22 #info Accepted Blocker, container-selinux, ON_QA 17:03:05 how reproducible this really is? 17:03:06 so on the whole i'm treating this as 'sufficiently resolved for release purposes', especially now we pushed https://bodhi.fedoraproject.org/updates/FEDORA-2022-c5bee6b70f stable for F35 today 17:03:12 if anyone thinks otherwise, please say so 17:03:31 same applies to the other similar bug, 2056303 17:03:44 I canot reproduce that on the RC1.1 iso 17:04:36 * bcotton forever holds his peace 17:04:39 well, these are both upgrade bugs, aiui, so to reproduce you need to do an f35 install and upgrade 17:04:46 yeah I saw it 17:04:47 I lost my speech. 17:05:09 I consider it sufficiently resolved now\ 17:06:22 rgr 17:07:02 #info on current information we believe this bug and the similar 2056303 to be 'sufficiently resolved' as far as the criteria go - they should not be affecting upgrades from f34 or f35 run at release time. if anyone believes otherwise, please add information to the bug report(s) 17:07:18 #topic Open floor 17:07:27 ok, that's everything, and it means right now we're still on track to release rc1 17:07:31 please continue to test it! 17:07:34 adamw: ack 17:07:52 have we started the AD tests yet? 17:07:56 adamw: 🎉 17:08:10 Ben Cotton (he/him): tflink volunteered to do it 17:08:15 today 17:08:39 in the process of running them 17:09:14 tflink++ 17:09:27 wow, running AD tests on Monday. what a time to be alive :-) 17:09:40 there are a few other blanks to fill - some arm tests, cloud tests on real hw, some dual boot tests - but i think we should be able to knock 'em out by thursday 17:09:43 bcotton, let's hope they will be fine :D 17:09:44 i know, craziness 17:10:01 lruzicka: shhhhhh 17:10:08 so we can declare it GO rite? 17:10:27 not now. :D 17:10:29 go rite naugh 17:10:42 adamw: a RC1.1 being released!!!!! 17:10:42 I ran them right before final freeze, has something changed drastically since then? 17:10:50 or are you talking about something other than the AD tests 17:11:02 tflink[m], no, I meant those 17:11:05 lruzicka: have we primed the kparal containment device? 17:11:18 tflink: no, nothing should've changed 17:11:19 😂😂😂😂 17:11:19 adamw, he is in the cellar 17:11:28 if they passed then they ought to pass now 17:12:42 alrighty, looks like we're done here 17:18:26 thanks for coming, folks! 17:18:36 #endmeeting