<@adamwill:fedora.im>
16:00:02
!startmeeting F41-blocker-review
<@meetbot:fedora.im>
16:00:03
Meeting started at 2024-10-21 16:00:02 UTC
<@meetbot:fedora.im>
16:00:03
The Meeting name is 'F41-blocker-review'
<@adamwill:fedora.im>
16:00:05
!topic Roll Call
<@adamwill:fedora.im>
16:00:14
morning folks, who's around for some super fun blocker review times?
<@nielsenb:fedora.im>
16:00:39
!hi
<@zodbot:fedora.im>
16:00:41
Brandon Nielsen (nielsenb)
<@frantisekz:fedora.im>
16:00:46
!ho
<@frantisekz:fedora.im>
16:00:48
!hi
<@zodbot:fedora.im>
16:00:49
František Zatloukal (frantisekz)
<@adamwill:fedora.im>
16:01:28
!hi
<@zodbot:fedora.im>
16:01:29
Adam Williamson (adamwill) - he / him / his
<@adamwill:fedora.im>
16:01:29
!ho
<@adamwill:fedora.im>
16:01:33
!it'sofftoworkwego
<@derekenz:fedora.im>
16:02:35
!hi
<@zodbot:fedora.im>
16:02:36
Derek Enz (derekenz)
<@nirik:matrix.scrye.com>
16:02:46
🛳️ it
<@siosm:matrix.org>
16:03:07
!hi
<@zodbot:fedora.im>
16:03:08
Timothée Ravier (siosm) - he / him / his
<@adamwill:fedora.im>
16:04:09
i see nirik is ready for some calm and sober analysis
<@adamwill:fedora.im>
16:04:56
ok, boilerplate time
<@adamwill:fedora.im>
16:05:03
!info The bugs up for review today are available at:
<@adamwill:fedora.im>
16:05:03
!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.
<@adamwill:fedora.im>
16:05:03
Why are we here?
<@adamwill:fedora.im>
16:05:03
!topic Introduction
<@adamwill:fedora.im>
16:05:03
!info We'll be following the process outlined at:
<@adamwill:fedora.im>
16:05:03
<@adamwill:fedora.im>
16:05:03
<@adamwill:fedora.im>
16:05:03
!info The criteria for release blocking bugs can be found at:
<@adamwill:fedora.im>
16:05:03
<@adamwill:fedora.im>
16:05:03
<@adamwill:fedora.im>
16:05:03
<@adamwill:fedora.im>
16:05:16
!info for F41 Final, we have:
<@adamwill:fedora.im>
16:05:17
!info 1 Accepted Blockers
<@adamwill:fedora.im>
16:05:17
!info 1 Accepted 0-day Blockers
<@adamwill:fedora.im>
16:05:17
!info 6 Proposed Blockers
<@adamwill:fedora.im>
16:05:21
!info 3 Proposed Freeze Exceptions
<@adamwill:fedora.im>
16:05:21
!info 7 Accepted Freeze Exceptions
<@adamwill:fedora.im>
16:05:26
who wants to secretarialize?
<@frantisekz:fedora.im>
16:05:31
o|
<@adamwill:fedora.im>
16:05:37
thanks!
<@adamwill:fedora.im>
16:05:42
!info František Zatloukal will secretarialize
<@adamwill:fedora.im>
16:05:45
let's start with:
<@adamwill:fedora.im>
16:05:48
!topic Proposed Final blockers
<@zodbot:fedora.im>
16:06:00
kparal has already given cookies to frantisekz during the F40 timeframe
<@sgallagh:fedora.im>
16:06:17
!hi
<@adamwill:fedora.im>
16:06:18
!info several of these technically have enough ticket votes for a decision already, but I decided to keep them on the agenda as we're trying to pull together a candidate to make extra sure we're all on the same page
<@zodbot:fedora.im>
16:06:18
Stephen Gallagher (sgallagh) - he / him / his
<@adamwill:fedora.im>
16:06:30
<@adamwill:fedora.im>
16:06:30
!info Ticket vote: FinalBlocker (+8,1,-1) (+asciiwolf, +jmaybaum, +sumantrom, +ngompa, +catanzaro, +pbrobinson, +lruzicka, +thebeanogamer, geraldosimiao, -nielsenb)
<@adamwill:fedora.im>
16:06:30
!info Proposed Blocker, cups, POST
<@adamwill:fedora.im>
16:06:30
<@adamwill:fedora.im>
16:06:30
!topic (2316066) dbus activated apps including the welcome dialog suffer 30-45 sec delay and/or crash on Workstation Live image randomly
<@conan_kudo:matrix.org>
16:06:48
!hi
<@zodbot:fedora.im>
16:06:49
Neal Gompa (ngompa) - he / him / his
<@adamwill:fedora.im>
16:06:52
so, we have been kicking this one around for weeks, but thanks to excellent work by steve (hi steve, if you're lurking out there), we have a potential fix for it now
<@adamwill:fedora.im>
16:07:32
i'm still honestly on the fence about whether it's a blocker, especially now i think i know what caused it and i'm pretty confident it really is only ever going to be a ~1 minute delay before everything works ok
<@adamwill:fedora.im>
16:07:42
but it's an accepted FE, so we can fix it and make it go away
<@kparal:matrix.org>
16:07:47
even though we might be close to fixing this, I think overall this is rather not a blocker. We haven't found any big serious side effect, except for the initial startup delay
<@sgallagh:fedora.im>
16:08:03
I don't think it would pass muster at Go/No-Go, so I'm inclined to vote -1 blocker.
<@adamwill:fedora.im>
16:08:03
if we just want to move along, i'm ok with taking the votes we have and making it a blocker, does anyone want to argue against?
<@frantisekz:fedora.im>
16:08:09
yeah, I was just figuring out how to word -1 FB +1 FE
<@adamwill:fedora.im>
16:08:16
i see, people do! excellent. :D
<@sgallagh:fedora.im>
16:08:16
But since we have FE+ and a proposed fix, this might be moot
<@adamwill:fedora.im>
16:08:37
Stephen Gallagher: well, if we make it FE and there turns out to be some kind of problem in my C coding (unpossible!) we can revert the patch and shipit
<@dustymabe:matrix.org>
16:08:46
👋
<@sgallagh:fedora.im>
16:08:50
adamw: That's why I said "might" :)
<@adamwill:fedora.im>
16:08:51
if we make it blocker and some such unaccountable consequence ensues, we would have to fix it some other way. or waive it.
<@frantisekz:fedora.im>
16:08:52
I mean, I was able to see it only in presence of Kamil Páral , never without him nearby
<@adamwill:fedora.im>
16:08:57
true
<@nirik:matrix.scrye.com>
16:09:11
-1 blocker, +1 FE
<@nielsenb:fedora.im>
16:09:19
Not a blocker, just annoying
<@kparal:matrix.org>
16:09:32
František Zatloukal: you must stop closing your eyes when I'm not nearby
<@adamwill:fedora.im>
16:09:45
is that why his car is so beat up?
<@frantisekz:fedora.im>
16:09:59
🤣🤣🤣
<@nielsenb:fedora.im>
16:10:32
I remain FinalBlocker -1, FinalFE +1
<@zodbot:fedora.im>
16:10:52
siosm gave a cookie to frantisekz. They now have 62 cookies, 13 of which were obtained in the Fedora 40 release cycle
<@adamwill:fedora.im>
16:11:30
so in the ticket we have a lot of +1s from folks who aren't here, which makes this tricky...asciiwolf, jmaybaum, Sumantro Mukherjee , Conan Kudo , catanzaro, Peter Robinson , lruzicka thebeanogamer all voted +1
<@adamwill:fedora.im>
16:12:31
so if i count votes now...we're at +8 / -4 , I think?
<@adamwill:fedora.im>
16:13:10
those 8 are +1, for -1 we have me, stephen, kamil, nirik, brandon , oops, that's 5. i fail counting
<@frantisekz:fedora.im>
16:13:21
and me -1!
<@adamwill:fedora.im>
16:13:28
+8/-5
<@adamwill:fedora.im>
16:13:42
if one more person votes -1 we can consider this vote split, move on, and make it go away as an FE. :P
<@sgallagh:fedora.im>
16:13:51
That sounds like a lack of consensus to reach blocker status to me
<@adamwill:fedora.im>
16:13:55
+8/-6
<@adamwill:fedora.im>
16:14:03
edit: failed counting again
<@adamwill:fedora.im>
16:14:06
yeah, let's call it
<@lbrabec:matrix.org>
16:15:01
if it helps to tip the scale... FinalFE +1, FinalBlocker -1 ;-)
<@adamwill:fedora.im>
16:15:07
thanks
<@zodbot:fedora.im>
16:15:22
No Fedora Accounts users have the @lbrabec:matrix.org Matrix Account defined
<@adamwill:fedora.im>
16:16:12
proposed !agreed 2316066 - punt (delay decision) - voting on this is at +8 / -7, so it's undecided. as it's an accepted FE and we have a probably-good fix for it, we're expecting to pull that fix in and make this go away that way.
<@frantisekz:fedora.im>
16:16:19
ack
<@kparal:matrix.org>
16:16:21
ack
<@nielsenb:fedora.im>
16:16:39
ack
<@adamwill:fedora.im>
16:16:48
!agreed 2316066 - punt (delay decision) - voting on this is at +8 / -7, so it's undecided. as it's an accepted FE and we have a probably-good fix for it, we're expecting to pull that fix in and make this go away that way.
<@adamwill:fedora.im>
16:17:00
!info Proposed Blocker, kernel, ON_QA
<@adamwill:fedora.im>
16:17:00
!info Ticket vote: FinalFreezeException (+3,0,-0) (+asciiwolf, +adamwill, +kevin)
<@adamwill:fedora.im>
16:17:00
!topic (2319597) Bluetooth devices are not getting paired
<@adamwill:fedora.im>
16:17:00
<@adamwill:fedora.im>
16:17:00
<@adamwill:fedora.im>
16:17:00
!info Ticket vote: FinalBlocker (+4,0,-0) (+asciiwolf, +geraldosimiao, +lruzicka, +kparal)
<@sgallagh:fedora.im>
16:17:33
I thought I voted on the ticket, but it seems not. +1 Blocker
<@frantisekz:fedora.im>
16:17:40
yeah, I guess users may want to try out bluetooth on Lives, +1
<@adamwill:fedora.im>
16:17:40
this one feels a bit fuzzy to me, but upstream chatter does seem to indicate there's definitely some issues with bt in 6.11.3. so far nobody seems to be saying 6.11.4-301 is worse than 6.11.3, and it's now pending push to stable (as this is already an FE)
<@adamwill:fedora.im>
16:18:20
the original reporter seems to still have issues with 6.11.4-301, but nobody else has said they have, so far. when i tested it worked even with 6.11.3, so i can't help much. has anyone else had time to test much?
<@sgallagh:fedora.im>
16:18:39
FTR, I'm a blocker because using the Live for anything at all might be an issue if a person has only BT peripherals.
<@nielsenb:fedora.im>
16:19:08
I haven't had a chance to see if I can reproduce yet, which is why I haven't voted
<@nielsenb:fedora.im>
16:19:21
If it looks really widespread, I'm on team blocker
<@adamwill:fedora.im>
16:19:23
anyway, if we don't swap out the kernel at the last minute is it *really* even a fedora release candidate
<@frantisekz:fedora.im>
16:19:41
you know what, let's move it to 6.12
<@zodbot:fedora.im>
16:19:57
sgallagh gave a cookie to kevin. They now have 689 cookies, 30 of which were obtained in the Fedora 40 release cycle
<@kparal:matrix.org>
16:20:19
it affects my mouse, I'm also on team blocker 🙂
<@nielsenb:fedora.im>
16:20:49
Does 6.11.4 make it better?
<@kparal:matrix.org>
16:21:09
yes, it's fixed in -301
<@adamwill:fedora.im>
16:21:14
yay
<@nielsenb:fedora.im>
16:21:28
Well, I'm FinalFE +1
<@zodbot:fedora.im>
16:21:43
sgallagh gave a cookie to kparal. They now have 73 cookies, 3 of which were obtained in the Fedora 40 release cycle
<@nirik:matrix.scrye.com>
16:22:13
the bug is closed ERRATA now. ;)
<@adamwill:fedora.im>
16:22:38
proposed !agreed 2319597 - AcceptedBlocker (Final) - there definitely seems to be enough of an indication that bluetooth is broken enough in 6.11.3 (which was stable in F41) to affect live media use cases, even potentially making it unusable if you have to use bluetooth peripherals, so this is accepted as a violation of audio criteria and any criterion requiring input, in the case that bluetooth periperals are required and the hardware is affected by this bug
<@nirik:matrix.scrye.com>
16:23:06
ack
<@sgallagh:fedora.im>
16:23:07
ack
<@nielsenb:fedora.im>
16:23:09
ack
<@derekenz:fedora.im>
16:23:21
ack
<@adamwill:fedora.im>
16:23:23
!agreed 2319597 - AcceptedBlocker (Final) - there definitely seems to be enough of an indication that bluetooth is broken enough in 6.11.3 (which was stable in F41) to affect live media use cases, even potentially making it unusable if you have to use bluetooth peripherals, so this is accepted as a violation of audio criteria and any criterion requiring input, in the case that bluetooth periperals are required and the hardware is affected by this bug
<@adamwill:fedora.im>
16:24:12
!info we have three KDE screen reader issues coming up: 2320046 , 2320044 , 2320043 . let's have a bit of a pre-discussion about how to handle this
<@adamwill:fedora.im>
16:24:30
Conan Kudo: do the three separate bugs, each proposed as a blocker, still best represent the situation here?
<@conan_kudo:matrix.org>
16:25:01
2320046 is not new, it exists in f39, just checked
<@conan_kudo:matrix.org>
16:25:22
2320044 is unverifiable since orca was broken in f40 and we didn't have welcome in f39
<@adamwill:fedora.im>
16:26:09
so we're still not sure if the two bugs of the type 'it doesn't read something it seems like it should, but if you hit + enough times, it does' are the same bug or not, yet?
<@conan_kudo:matrix.org>
16:27:06
yeah we don't know
<@adamwill:fedora.im>
16:27:44
okay
<@adamwill:fedora.im>
16:27:57
i guess let's go ahead and consider them one at a time, then. but i'm gonna put the qt one first
<@adamwill:fedora.im>
16:28:07
!info Ticket vote: FinalBlocker (+4,0,-0) (+nielsenb, +geraldosimiao, +thebeanogamer, +lruzicka)
<@adamwill:fedora.im>
16:28:07
<@adamwill:fedora.im>
16:28:07
<@adamwill:fedora.im>
16:28:07
!info Proposed Blocker, qt6-qtbase, MODIFIED
<@adamwill:fedora.im>
16:28:07
!topic (2320043) Qt does not watch for screen reader being enabled properly, rendering Plasma invisible to Orca
<@adamwill:fedora.im>
16:28:26
this one seems pretty bad, fairly new, and fixable! the holy trifecta
<@adamwill:fedora.im>
16:28:31
so, i'm fine with a +1 here
<@frantisekz:fedora.im>
16:28:34
+1
<@conan_kudo:matrix.org>
16:28:40
+1
<@derekenz:fedora.im>
16:28:40
+1
<@adamwill:fedora.im>
16:28:50
note we kinda established a precedent for 'screen reading being completely broken is a blocker' with https://bugzilla.redhat.com/show_bug.cgi?id=2305763
<@conan_kudo:matrix.org>
16:28:51
and importantly, the update does fix it :)
<@nielsenb:fedora.im>
16:29:08
FinalBlocker +1
<@conan_kudo:matrix.org>
16:29:09
I did horrible things to test it in the right circumstances, but it works
<@nirik:matrix.scrye.com>
16:29:16
FinalBlocker +1
<@adamwill:fedora.im>
16:29:31
*one of us*
<@sgallagh:fedora.im>
16:29:39
FinalBlocker +1
<@zodbot:fedora.im>
16:29:47
sgallagh has already given cookies to ngompa during the F40 timeframe
<@conan_kudo:matrix.org>
16:29:54
hey I was once an SQA engineer :)
<@conan_kudo:matrix.org>
16:30:00
I have mad respect for QA
<@sgallagh:fedora.im>
16:30:14
I was once too. Definitely not twice ✌️
<@adamwill:fedora.im>
16:30:30
proposed !agreed 2320043 - AcceptedBlocker (Final) - this is accepted as a conditional violation of all desktop related criteria for KDE for all people who need a screen reader (same rationale on which we previously accepted 2305763)
<@frantisekz:fedora.im>
16:30:37
ack
<@nielsenb:fedora.im>
16:30:42
ack
<@derekenz:fedora.im>
16:30:43
ack
<@sgallagh:fedora.im>
16:30:48
ack
<@conan_kudo:matrix.org>
16:30:50
ack
<@adamwill:fedora.im>
16:30:58
!agreed 2320043 - AcceptedBlocker (Final) - this is accepted as a conditional violation of all desktop related criteria for KDE for all people who need a screen reader (same rationale on which we previously accepted 2305763)
<@adamwill:fedora.im>
16:31:13
<@adamwill:fedora.im>
16:31:13
<@adamwill:fedora.im>
16:31:13
!info Ticket vote: FinalBlocker (+3,0,-0) (+nielsenb, +thebeanogamer, +lruzicka)
<@adamwill:fedora.im>
16:31:13
!info Proposed Blocker, konsole, NEW
<@adamwill:fedora.im>
16:31:13
!topic (2320046) Konsole command output isn't read by the screen reader
<@conan_kudo:matrix.org>
16:31:42
so this issue is not new... it exists in f39 as well
<@adamwill:fedora.im>
16:31:43
so, my current understanding of this is basically that the reader doesn't read some output you'd expect it would, but if you hit the keypad plus key enough times, it does...right?
<@conan_kudo:matrix.org>
16:31:49
yes
<@conan_kudo:matrix.org>
16:31:58
only if you use the + key on the numpad though
<@conan_kudo:matrix.org>
16:32:03
the regular plus key does nothing
<@conan_kudo:matrix.org>
16:32:20
it's apparently an orca thing
<@adamwill:fedora.im>
16:32:39
i think the point is so you can *also* type a plus if you need to
<@conan_kudo:matrix.org>
16:32:44
yes
<@adamwill:fedora.im>
16:32:51
if both plus keys did screen reader things, you've got a problem when you want to do any math. :P
<@conan_kudo:matrix.org>
16:32:58
that's very true
<@adamwill:fedora.im>
16:33:19
presumably there is a standard way to handle this on keypad-less systems, for folks who use screen readers? I really don't know :/
<@sgallagh:fedora.im>
16:33:28
I'm leaning towards -1 Final Blocker here. General screen reader not working at all is a blocker, but I don't know that just the console having an issue is enough
<@conan_kudo:matrix.org>
16:33:32
I don't know either, but I assume there is
<@conan_kudo:matrix.org>
16:33:54
I'm leaning FE rather than Blocker now that I've discovered this issue existed forever
<@adamwill:fedora.im>
16:34:06
but i think we shouldn't worry about that wrinkle for the purpose of blocker evaluation - it seems weird for us to say 'oh we think this is worse because the regular plus key doesn't work', it seems clear we should assume the keyboard shortcuts are as they are for good reason
<@sgallagh:fedora.im>
16:34:07
I'm good with an FE if a fix is found
<@conan_kudo:matrix.org>
16:34:18
-1 FB +1 FE
<@nielsenb:fedora.im>
16:34:34
Yeah, the fact that we've shipped it twice makes me lean away from my blocker vote
<@nielsenb:fedora.im>
16:34:46
Doesn't me I like it...
<@frantisekz:fedora.im>
16:34:48
yeah, -1 FB, +1 FE sounds reasonable
<@nirik:matrix.scrye.com>
16:34:48
Does this only affect konsole?
<@adamwill:fedora.im>
16:34:52
yeah, based on it being pre-existing and the 'workaround' being 'use a normal screen reader function' it seems maybe less blocker-y? it would be great to have input from someone who actually uses a screen reader regularly, but not sure we have any such person handy :|
<@nirik:matrix.scrye.com>
16:34:55
or any terminal app?
<@nielsenb:fedora.im>
16:34:56
Doesn't mean I like it...
<@conan_kudo:matrix.org>
16:35:38
I would assume other terminal apps may behave differently
<@conan_kudo:matrix.org>
16:36:04
I know vte and konsolepart both behave very differently for a11y
<@conan_kudo:matrix.org>
16:36:23
and I don't know anything about qmlkonsole or any of the other terminal emulator engines
<@nirik:matrix.scrye.com>
16:36:44
ok, just thinking about it for workarounds, but yeah...
<@derekenz:fedora.im>
16:37:24
-1 FB +FE
<@nielsenb:fedora.im>
16:37:31
FinalBlocker -1
<@nielsenb:fedora.im>
16:37:46
Not sure how I feel about a FE with no fix immediately apparent
<@conan_kudo:matrix.org>
16:38:12
well the bug upstream is being looked at
<@conan_kudo:matrix.org>
16:38:14
so who knows
<@sgallagh:fedora.im>
16:38:15
FE doesn't grant it automatic inclusion, just an assertion that we'd consider accepting it if a fix comes in in time
<@nielsenb:fedora.im>
16:38:39
Yeah, and it's easy enough to test I suppose
<@nielsenb:fedora.im>
16:38:44
FinalFE +1
<@nirik:matrix.scrye.com>
16:38:55
I'm definitely +1 for FinalFE
<@adamwill:fedora.im>
16:39:14
alrighty
<@adamwill:fedora.im>
16:41:08
proposed !agreed 2320046 - RejectedBlocker (Final), AcceptedFreezeException (Final) - this is rejected as a blocker as the problem is not too critical: screen reading works, and all content can be read, it just seems necessary to prompt the reader to continue (using the normal key for this purpose) at points where it seems like that shouldn't be necessary. we also note that this is not new behaviour (previous releases behaved the same)
<@nielsenb:fedora.im>
16:41:21
ack
<@sgallagh:fedora.im>
16:41:26
ack
<@frantisekz:fedora.im>
16:41:30
ack
<@derekenz:fedora.im>
16:41:30
ack
<@conan_kudo:matrix.org>
16:41:34
ack
<@adamwill:fedora.im>
16:41:46
!agreed 2320046 - RejectedBlocker (Final), AcceptedFreezeException (Final) - this is rejected as a blocker as the problem is not too critical: screen reading works, and all content can be read, it just seems necessary to prompt the reader to continue (using the normal key for this purpose) at points where it seems like that shouldn't be necessary. we also note that this is not new behaviour (previous releases behaved the same)
<@adamwill:fedora.im>
16:42:07
!info Proposed Blocker, plasma-welcome, NEW, depends on other bugs
<@adamwill:fedora.im>
16:42:07
<@adamwill:fedora.im>
16:42:07
!topic (2320044) Plasma Welcome isn't read by the screen reader when activated
<@adamwill:fedora.im>
16:42:07
!info Ticket vote: FinalBlocker (+4,0,-0) (+nielsenb, +geraldosimiao, +thebeanogamer, +lruzicka)
<@adamwill:fedora.im>
16:42:07
<@adamwill:fedora.im>
16:42:22
this is similar to the previous one, but in the welcome screen. we don't yet know if they share a cause or just the symptom
<@conan_kudo:matrix.org>
16:42:35
and the f39 welcome center exhibits the same behavior
<@conan_kudo:matrix.org>
16:42:39
so it's not new either
<@conan_kudo:matrix.org>
16:42:58
but you don't see it in the live environment because the live mode welcome center was added in f40
<@sgallagh:fedora.im>
16:43:14
Yeah, I'm -1 Blocker, +1 FE here. I *really* don't like that it fails to be spoken, but I can't see it passing the "last blocker" test
<@conan_kudo:matrix.org>
16:43:54
-1 FB +1 FE
<@conan_kudo:matrix.org>
16:44:11
and just in case people are wondering, it's the same on X11 and Wayland
<@conan_kudo:matrix.org>
16:44:25
Qt a11y is window system independent
<@nielsenb:fedora.im>
16:44:27
I guess same as before
<@nielsenb:fedora.im>
16:44:33
FinalBlocker -1, FinalFE +1
<@conan_kudo:matrix.org>
16:44:35
since AT-SPI does not care about X11 or Wayland
<@frantisekz:fedora.im>
16:44:52
-1 FB +1 FE
<@conan_kudo:matrix.org>
16:47:25
the good news is that at least with the qt fix, when you tab through the app, it does tell you things
<@conan_kudo:matrix.org>
16:47:30
it just won't read the body text
<@derekenz:fedora.im>
16:47:43
-1 FB
<@derekenz:fedora.im>
16:47:51
+1 FE
<@conan_kudo:matrix.org>
16:48:04
so you can navigate the app at least
<@conan_kudo:matrix.org>
16:48:20
it's not _great_, but at least it's not invisible
<@adamwill:fedora.im>
16:48:53
proposed !agreed 2320044 - RejectedBlocker (Final), AcceptedFreezeException (Final) - this is rejected as a blocker as the problem is not too critical: screen reading works, and all critical content can be read, it just seems necessary to prompt the reader to get body content read (using the normal key for this purpose) at points where it seems like that shouldn't be necessary. we also note that this is not new behaviour (previous releases behaved the same)
<@sgallagh:fedora.im>
16:48:56
If you're using a screen reader, there's a good chance that everything is invisible...
<@frantisekz:fedora.im>
16:49:07
ack
<@derekenz:fedora.im>
16:49:12
ack
<@nielsenb:fedora.im>
16:49:16
ack
<@conan_kudo:matrix.org>
16:49:39
ack
<@sgallagh:fedora.im>
16:49:46
ack
<@adamwill:fedora.im>
16:50:49
!agreed 2320044 - RejectedBlocker (Final), AcceptedFreezeException (Final) - this is rejected as a blocker as the problem is not too critical: screen reading works, and all critical content can be read, it just seems necessary to prompt the reader to get body content read (using the normal key for this purpose) at points where it seems like that shouldn't be necessary. we also note that this is not new behaviour (previous releases behaved the same)
<@adamwill:fedora.im>
16:51:05
!info Proposed Blocker, qt6-qtwayland, NEW
<@adamwill:fedora.im>
16:51:05
!info Ticket vote: FinalBlocker (+1,0,-1) (+lruzicka, -nielsenb)
<@adamwill:fedora.im>
16:51:05
!topic (2318535) KDE apps like KInfoCenter and System Settings sometimes don't start
<@adamwill:fedora.im>
16:51:05
<@adamwill:fedora.im>
16:51:05
<@adamwill:fedora.im>
16:51:18
so, I think this one is largely mitigated by the VM performance fix, right?
<@conan_kudo:matrix.org>
16:51:26
seems like it
<@adamwill:fedora.im>
16:51:27
it still appears to be a real bug but it's probably very hard to encounter now
<@conan_kudo:matrix.org>
16:51:38
I couldn't replicate it as much as I tried
<@derekenz:fedora.im>
16:51:47
Could not repro
<@adamwill:fedora.im>
16:52:34
you know it's release week when: you realize you have the same openqa test open in two tabs, that are only one tab apart in your tab bar
<@conan_kudo:matrix.org>
16:52:42
haha
<@sgallagh:fedora.im>
16:54:06
And the tab between them is a real-estate posting for a yak farm?
<@adamwill:fedora.im>
16:54:08
so, i'm probably -1/-1 at this point
<@adamwill:fedora.im>
16:54:27
how is there no yak emoji yet
<@frantisekz:fedora.im>
16:54:32
yeah, -1 {FE,FB}
<@adamwill:fedora.im>
16:54:41
brb writing to the unicode consortium
<@frantisekz:fedora.im>
16:54:53
rfc time!
<@nirik:matrix.scrye.com>
16:55:01
only 🐃
<@adamwill:fedora.im>
16:56:24
proposed !agreed 2318535 - RejectedBlocker (Final) RejectedFreezeException (Final) - this seems to be pretty much entirely mitigated by the fix for KDE performance under software acceleration (https://bugzilla.redhat.com/show_bug.cgi?id=2312900 ). there probably is still technically a bug here, and upstream is working on it, but it doesn't seem like we need to rush to pull it in for release
<@nielsenb:fedora.im>
16:56:35
ack
<@frantisekz:fedora.im>
16:56:38
ack
<@derekenz:fedora.im>
16:56:40
ack
<@sgallagh:fedora.im>
16:56:43
ack
<@adamwill:fedora.im>
16:56:57
!agreed 2318535 - RejectedBlocker (Final) RejectedFreezeException (Final) - this seems to be pretty much entirely mitigated by the fix for KDE performance under software acceleration (https://bugzilla.redhat.com/show_bug.cgi?id=2312900 ). there probably is still technically a bug here, and upstream is working on it, but it doesn't seem like we need to rush to pull it in for release
<@adamwill:fedora.im>
16:57:11
okay, that's all the blockers, moving on to:
<@adamwill:fedora.im>
16:57:15
!topic Proposed Final freeze exceptions
<@adamwill:fedora.im>
16:57:27
!info Ticket vote: FinalFreezeException (+1,0,-0) (+augenauf)
<@adamwill:fedora.im>
16:57:27
!info Proposed Freeze Exceptions, golang, NEW
<@adamwill:fedora.im>
16:57:27
<@adamwill:fedora.im>
16:57:27
<@adamwill:fedora.im>
16:57:27
!topic (2310555) CVE-2024-34156 golang: Calling Decoder.Decode on a message which contains deeply nested structures can cause a panic due to stack exhaustion [fedora-all]
<@adamwill:fedora.im>
16:57:27
!info Ticket vote: FinalBlocker (+1,0,-0) (+augenauf)
<@adamwill:fedora.im>
16:57:45
oh, huh, i closed this but it's still in the list? weird
<@adamwill:fedora.im>
16:57:49
i thought closed bugs got dropped
<@adamwill:fedora.im>
16:58:05
why does blockerbugs still think it's open?
<@adamwill:fedora.im>
16:58:20
!info this bug is actually closed, seems to be a blip in blockerbugs. moving on
<@decathorpe:fedora.im>
16:58:52
just to clarify - wouldn't a fix in golang necessitate rebuilding all applications written in Go too?
<@adamwill:fedora.im>
16:59:08
hum. maybe, i guess?
<@decathorpe:fedora.im>
16:59:39
AFAIK the last Golang mini-mass-rebuild happened quite a long while ago, so I don't think applications have actually been rebuilt with the fixed golang version in this case
<@conan_kudo:matrix.org>
17:00:39
yeah go apps have to be rebuilt for the fix to mean anything
<@adamwill:fedora.im>
17:00:44
well, we could reopen it, but i'd probably be -1 to a mass go rebuild for a CVE that only possibly causes DoS
<@adamwill:fedora.im>
17:00:56
well, we could reopen it, but i'd probably be -1 FE to a mass go rebuild for a CVE that only possibly causes DoS
<@decathorpe:fedora.im>
17:01:24
I asked in the Golang room, but yeah, I'd say that's not high priority
<@sgallagh:fedora.im>
17:03:20
Does this mean we'd need a post-release mass rebuild?
<@adamwill:fedora.im>
17:03:20
i reopened it, so i guess let's vote
<@sgallagh:fedora.im>
17:03:27
Because I'm not keen on that either...
<@decathorpe:fedora.im>
17:03:54
Stephen Gallagher: those happen more or less regularly for Golang CVEs anyway
<@adamwill:fedora.im>
17:04:00
with the evaluation at https://access.redhat.com/security/cve/CVE-2024-34156 in mind: do we want to grant an FE to rebuild...all go apps? all go apps verified to use Decoder.Decoding?...with the fixed compiler?
<@sgallagh:fedora.im>
17:04:19
This does not endear me to Golang...
<@decathorpe:fedora.im>
17:04:36
yup, but all languages that do static linking need this :(
<@nirik:matrix.scrye.com>
17:04:48
yuck
<@sgallagh:fedora.im>
17:04:58
Well, I guess if GCC had a serious code-generation error, we'd be in the same boat
<@adamwill:fedora.im>
17:05:05
Fabio Valentini: how many builds would we be talking here? how long?
<@conan_kudo:matrix.org>
17:05:21
that has happened before, it sucked
<@adamwill:fedora.im>
17:05:22
actually trying to *do* this for release sounds a bit...slip-y
<@decathorpe:fedora.im>
17:05:24
no idea. the Go mini-mass-rebuilds were handled by ... gotmax23 I think
<@nirik:matrix.scrye.com>
17:05:31
I don't think this should block release...
<@sgallagh:fedora.im>
17:06:15
-1 Blocker, -1 FE
<@adamwill:fedora.im>
17:06:37
i could really use a fuzzy search for bodhi. 'find me all updates with more than 5 packages and 'go' in at least one of their names'
<@nirik:matrix.scrye.com>
17:07:01
-1 Blocker, -1 FE
<@nielsenb:fedora.im>
17:07:21
Yeah, fix it in post
<@nielsenb:fedora.im>
17:07:24
FinalBlocker -1
<@nielsenb:fedora.im>
17:07:26
FinalFE - 1
<@nielsenb:fedora.im>
17:07:34
FinalFE -1
<@conan_kudo:matrix.org>
17:08:01
-1 FB -1 FE
<@decathorpe:fedora.im>
17:08:14
adamw: here's one https://bodhi.fedoraproject.org/updates/FEDORA-2022-3969b64d4b
<@decathorpe:fedora.im>
17:09:18
I don't think all of them were this big though
<@adamwill:fedora.im>
17:09:45
yoiks
<@decathorpe:fedora.im>
17:10:30
yeah I don't think we'd want to do this so soon before the release :)
<@conan_kudo:matrix.org>
17:10:52
post-GA for that
<@sgallagh:fedora.im>
17:12:04
We can definitely handle at least 315 packages :)
<@adamwill:fedora.im>
17:12:06
proposed !agreed 2310555 - RejectedBlocker (Final) RejectedFreezeException (Final) - this is rejected on the basis that the issues caused can be "satisfactorily resolved by a package update" (per the criteria). anything built with Go and persistently parsing untrusted input is probably some kind of system service which should be deployed only on an installed system. we are not aware of any likely context in which this vulnerability would be exposed in a live or installer environment
<@adamwill:fedora.im>
17:12:20
Stephen Gallagher: i think the biggest KDE one I found was already bigger
<@sgallagh:fedora.im>
17:12:24
ack
<@conan_kudo:matrix.org>
17:12:30
yeah that was ~700
<@decathorpe:fedora.im>
17:12:30
ack
<@conan_kudo:matrix.org>
17:12:33
ack
<@conan_kudo:matrix.org>
17:12:46
we can get up to 900 if a Qt upgrade is included
<@adamwill:fedora.im>
17:13:08
!agreed 2310555 - RejectedBlocker (Final) RejectedFreezeException (Final) - this is rejected on the basis that the issues caused can be "satisfactorily resolved by a package update" (per the criteria). anything built with Go and persistently parsing untrusted input is probably some kind of system service which should be deployed only on an installed system. we are not aware of any likely context in which this vulnerability would be exposed in a live or installer environment
<@adamwill:fedora.im>
17:13:37
!info Proposed Freeze Exceptions, rust-cursive0.16, MODIFIED
<@adamwill:fedora.im>
17:13:37
!topic (2320157) F41FailsToInstall: rust-cursive0.16+crossterm-devel, rust-cursive0.16+pancurses-devel
<@adamwill:fedora.im>
17:13:37
<@adamwill:fedora.im>
17:13:37
!info Ticket vote: FinalFreezeException (+1,0,-0) (+frantisekz)
<@adamwill:fedora.im>
17:13:37
<@decathorpe:fedora.im>
17:13:51
oh! oh! this is mine
<@adamwill:fedora.im>
17:13:58
FTI fix, sure. +1 if it's straightforward.
<@decathorpe:fedora.im>
17:14:22
it is. the update contains only one change - drop the two unused subpackages that have broken deps
<@nielsenb:fedora.im>
17:14:29
FinalFE +1
<@adamwill:fedora.im>
17:14:43
does something obsolete them, so they will be removed on systems that already have them installed?
<@nirik:matrix.scrye.com>
17:14:51
FinalFE +1
<@sgallagh:fedora.im>
17:14:55
We're on FEs now, no more blockers?
<@adamwill:fedora.im>
17:15:13
Stephen Gallagher: yes, we did all the blockers (at least all that were proposed at start of meeting)
<@decathorpe:fedora.im>
17:15:21
no system should have them installed. but *if* somebody managed to install them (they're pretty useless outside mock) an update with "--allowerasing" will remove them
<@sgallagh:fedora.im>
17:15:31
Thanks
<@adamwill:fedora.im>
17:15:39
if somebody updated while the bad update was in u-t, did they get them installed?
<@adamwill:fedora.im>
17:16:09
we generally try to have obsoletes in place so people don't have to know about/use --allowerasing. i don't like that arg (and i especially don't like that gnome-software does it by default). it can cause problems.
<@decathorpe:fedora.im>
17:16:13
there was no bad update, the broken deps were caused by accidental retirement after final freeze went into effect
<@decathorpe:fedora.im>
17:16:34
there was no bad update, the broken deps were caused a package retirement
<@decathorpe:fedora.im>
17:16:41
there was no bad update, the broken deps were caused by a package retirement
<@adamwill:fedora.im>
17:16:57
okay. so you're saying the packages existed but you don't think anyone would have installed them?
<@decathorpe:fedora.im>
17:17:10
nobody should 😅
<@decathorpe:fedora.im>
17:17:20
we explicitly document that users shouldn't install rust-*-devel packages
<@decathorpe:fedora.im>
17:17:37
(that might change in the future, but for now that's the best we can do)
<@adamwill:fedora.im>
17:18:07
okay.
<@adamwill:fedora.im>
17:18:46
proposed !agreed 2320157 - AcceptedFreezeException (Final) - this is accepted as we generally accept FTI fixes in the interest of keeping the frozen final repo as consistent as possible
<@sgallagh:fedora.im>
17:19:26
ack
<@nielsenb:fedora.im>
17:19:44
ack
<@conan_kudo:matrix.org>
17:19:51
ack
<@nirik:matrix.scrye.com>
17:19:56
ack
<@adamwill:fedora.im>
17:20:16
!agreed 2320157 - AcceptedFreezeException (Final) - this is accepted as we generally accept FTI fixes in the interest of keeping the frozen final repo as consistent as possible
<@frantisekz:fedora.im>
17:20:17
ack
<@adamwill:fedora.im>
17:20:28
<@adamwill:fedora.im>
17:20:28
!info Ticket vote: FinalFreezeException (+2,0,-0) (+lruzicka, +frantisekz)
<@adamwill:fedora.im>
17:20:28
!topic (2320158) F41FailsToInstall: rust-termbg-devel
<@adamwill:fedora.im>
17:20:28
<@adamwill:fedora.im>
17:20:28
!info Proposed Freeze Exceptions, rust-termbg, VERIFIED
<@adamwill:fedora.im>
17:20:29
same?
<@decathorpe:fedora.im>
17:20:46
more or less ... this one was actually only 50% my fault
<@decathorpe:fedora.im>
17:21:01
the other 50% is that package retirements weren't blocked by releng after final freeze went into effect :)
<@decathorpe:fedora.im>
17:21:15
and the fix is more straightforward- no removed packages, only updated deps.
<@nielsenb:fedora.im>
17:21:26
FinalFE +1
<@nirik:matrix.scrye.com>
17:22:31
FinalFE +1
<@adamwill:fedora.im>
17:24:20
FinalFE +1
<@adamwill:fedora.im>
17:24:37
proposed !agreed 2320158 - AcceptedFreezeException (Final) - this is accepted as we generally accept FTI fixes in the interest of keeping the frozen final repo as consistent as possible
<@nirik:matrix.scrye.com>
17:24:53
ack
<@nielsenb:fedora.im>
17:25:08
ack
<@conan_kudo:matrix.org>
17:25:40
ack
<@adamwill:fedora.im>
17:26:18
!agreed 2320158 - AcceptedFreezeException (Final) - this is accepted as we generally accept FTI fixes in the interest of keeping the frozen final repo as consistent as possible
<@adamwill:fedora.im>
17:26:27
ok, let's do a quick review of:
<@adamwill:fedora.im>
17:26:32
!topic Accepted Final blockers
<@adamwill:fedora.im>
17:26:43
!info Accepted Blocker, distribution, ON_QA
<@adamwill:fedora.im>
17:26:43
<@adamwill:fedora.im>
17:26:43
<@adamwill:fedora.im>
17:26:43
!topic (2320081) osbuild tasks are consistently failing and breaking composes, but building ARM minimal with ImageFactory instead would compromise device support
<@adamwill:fedora.im>
17:27:03
!info this is marked ON_QA as osbuild team believes they have resolved the issue. we will fire the next compose when we're ready, and find out. :D no action required ATM
<@adamwill:fedora.im>
17:27:11
!topic (2318710) updating F41 Beta in offline mode reboots into maintenance mode due to "ext4 bad orphan inode" errors
<@adamwill:fedora.im>
17:27:11
<@adamwill:fedora.im>
17:27:11
!info Accepted Blocker, kernel, ON_QA
<@adamwill:fedora.im>
17:27:11
<@adamwill:fedora.im>
17:27:28
this one is back because kparal tested and did not find it's fixed for him. though i tested twice and it was definitely fixed for me
<@adamwill:fedora.im>
17:27:46
kparal has been testing update of F41 beta to final, I have been testing upgrade of F40 to F41, so that may be the difference. we will both test some more
<@nirik:matrix.scrye.com>
17:28:31
Wonder if it also happens on 39 > 41
<@adamwill:fedora.im>
17:28:46
the bug did happen, i verified that. i didn't test the fix in that case yet
<@decathorpe:fedora.im>
17:28:55
I wonder why "orphan inode repairs are necessary" even happens ...
<@adamwill:fedora.im>
17:29:05
yes, we all wonder that, but didn't get to figuring it out yet.
<@adamwill:fedora.im>
17:29:36
possibly the filesystem isn't unmounted properly on shutdown from the special offline upgrade environment
<@nirik:matrix.scrye.com>
17:30:45
it's probibly cups.
<@nirik:matrix.scrye.com>
17:30:53
(I kid)
<@kparal:matrix.org>
17:33:49
I also have lvm, not sure if that could matter
<@adamwill:fedora.im>
17:33:50
!info there's some uncertainty about whether this is fully fixed yet, kparal and adamw are seeing differing results in testing. we will continue to test, if others want to help out that'd be great
<@adamwill:fedora.im>
17:34:11
ah, i was testing just on direct ext4 root, no LVM
<@adamwill:fedora.im>
17:34:18
let's mix up the variables a bit...
<@kparal:matrix.org>
17:34:25
I'll run more tests tomorrow
<@adamwill:fedora.im>
17:34:46
thanks, i'll try different scenarios today and see if i can pin anything down
<@kparal:matrix.org>
17:35:28
I double checked, the resulted system had the correct version of e2fsprogs
<@adamwill:fedora.im>
17:35:37
yeah i saw that in the rpm -qa output :/
<@adamwill:fedora.im>
17:35:46
the initramfs thing seems like a potential theory at least, i'll work on that first
<@kparal:matrix.org>
17:36:22
not sure how to check e2fsck version in initramfs env
<@kparal:matrix.org>
17:36:26
it doesn't have rpm
<@kparal:matrix.org>
17:36:41
I guess I could just cmp on the binaries...
<@adamwill:fedora.im>
17:36:46
oh haha
<@adamwill:fedora.im>
17:36:49
i think i can guess the problem
<@adamwill:fedora.im>
17:37:03
it's that old chestnut: "didn't apply the patch"
<@adamwill:fedora.im>
17:37:06
https://src.fedoraproject.org/rpms/e2fsprogs/c/ddab4e6f808d38e822f3e5083383ba6434c61006?branch=rawhide
<@conan_kudo:matrix.org>
17:37:10
lol
<@adamwill:fedora.im>
17:37:12
the spec doesn't use autosetup
<@adamwill:fedora.im>
17:37:25
i was testing with my own scratch build, where I *did* apply the patch. :D good thing you re-tested!
<@kparal:matrix.org>
17:37:40
heh
<@conan_kudo:matrix.org>
17:37:42
you can put spaces between `Patch` and the number?!
<@nirik:matrix.scrye.com>
17:37:44
it seems to have -V
<@frantisekz:fedora.im>
17:37:51
where went the proposal to remove the setup macro? ! :D
<@adamwill:fedora.im>
17:37:58
!info we now have a BETTER suspect for the discrepancy here: the official update build did not apply the patch that fixes the bug
<@adamwill:fedora.im>
17:38:04
!action adamw to fix that up, kparal to re-test
<@adamwill:fedora.im>
17:38:15
Conan Kudo: inorite?! rpm is weird
<@decathorpe:fedora.im>
17:38:24
yes - it's cursed! :D
<@conan_kudo:matrix.org>
17:38:30
just switch it to autosetup
<@adamwill:fedora.im>
17:39:47
i don't like doing that to other people's packags
<@adamwill:fedora.im>
17:39:56
i'm pinging eric to see if he wants to fix this himself before i stomp on it
<@adamwill:fedora.im>
17:39:59
anyhoo, moving along
<@adamwill:fedora.im>
17:42:19
!topic (2318559) Fedora media writer downloads wrong iso when kde spin is selected
<@adamwill:fedora.im>
17:42:19
!info Accepted 0-day Blocker, mediawriter, ASSIGNED
<@adamwill:fedora.im>
17:42:19
<@adamwill:fedora.im>
17:42:19
<@nirik:matrix.scrye.com>
17:43:11
So this is fixed, but we need websites to point to their newer builds I guess?
<@adamwill:fedora.im>
17:43:36
yeah, um, i haven't really been tracking this one, let's see
<@nirik:matrix.scrye.com>
17:44:00
I thought there was a websites ticket, but I can't find it.
<@adamwill:fedora.im>
17:45:04
hum...can you file one?
<@nirik:matrix.scrye.com>
17:45:55
I could have sworn there was one, but yeah, I can look more and file if not.
<@adamwill:fedora.im>
17:47:10
thanks a lot
<@adamwill:fedora.im>
17:47:48
!info this is more or less dealt with, though we need the websites updated to point to the new release of fmw, and there's some oddness to how the metadata got to be wrong that we should work out
<@nirik:matrix.scrye.com>
17:49:39
ok. I see what happened. It was filed as a releng ticket, told to refile as websites, and that never happened for whatever reason. Will file.
<@adamwill:fedora.im>
17:49:48
thanks
<@adamwill:fedora.im>
17:49:52
ok, that's the crop
<@adamwill:fedora.im>
17:49:54
!topic Open floor
<@adamwill:fedora.im>
17:49:59
any other business, folks?
<@nielsenb:fedora.im>
17:50:34
Not from me
<@frantisekz:fedora.im>
17:51:11
just a q, adamw would you be asking for the compose now and shall I secretary it right away, or it'd be fine ~45 minutes?
<@nirik:matrix.scrye.com>
17:51:11
oh, no, it is filed: and closed yesterday
<@nirik:matrix.scrye.com>
17:51:16
https://gitlab.com/fedora/websites-apps/fedora-websites/fedora-websites-3.0/-/issues/314
<@adamwill:fedora.im>
17:54:33
oohkay
<@adamwill:fedora.im>
17:55:00
František Zatloukal: secretary isn't urgent i don't think as everything we need to pull in was already at least accepted FE
<@adamwill:fedora.im>
17:55:08
oh, except maybe the qt one, if you can do that it'd be great
<@adamwill:fedora.im>
17:55:43
nirik: https://fedoraproject.org/workstation/download does seem to point to https://flathub.org/apps/details/org.fedoraproject.MediaWriter
<@frantisekz:fedora.im>
17:55:54
will handle that!
<@adamwill:fedora.im>
17:56:00
er, that's not hte link i meant
<@adamwill:fedora.im>
17:56:08
oh hum
<@nirik:matrix.scrye.com>
17:56:09
if you are on linux probibly
<@frantisekz:fedora.im>
17:56:09
the orca/qt one
<@adamwill:fedora.im>
17:56:11
windows points to https://github.com/FedoraQt/MediaWriter/releases/latest
<@nirik:matrix.scrye.com>
17:56:17
yeah
<@adamwill:fedora.im>
17:56:17
linux points to https://flathub.org/apps/details/org.fedoraproject.MediaWriter
<@adamwill:fedora.im>
17:56:21
so we should make sure flathub has the latest...
<@adamwill:fedora.im>
17:56:30
nope, it's still 5.1.2
<@nirik:matrix.scrye.com>
17:57:34
ok, we should get them to update that and/or point to the fedora flatpak
<@adamwill:fedora.im>
17:57:46
i've commented in the bg
<@adamwill:fedora.im>
17:57:50
i've commented in the bug
<@adamwill:fedora.im>
17:57:54
not sure if we need to flag it anywhere else
<@nirik:matrix.scrye.com>
17:58:39
we can circle back later I think?
<@adamwill:fedora.im>
18:00:14
sure
<@adamwill:fedora.im>
18:00:21
we're at time, so i guess that's all
<@adamwill:fedora.im>
18:00:40
thanks for coming folks! looks like we might be able to make it to a compose later today once we recheck a few things and line them all upo
<@nirik:matrix.scrye.com>
18:00:51
sure hope so
<@adamwill:fedora.im>
18:01:09
!endmeeting