ReleaseEngineering/Meetings/2007-oct-29

From FedoraProject

Jump to: navigation, search

Fedora Release Engineering Meeting :: Monday 2007-10-29

IRC Transcript

-!- f13 changed the topic of #fedora-meeting to: Fedora Release Engineering - Fedora 8 release meeting13:06
* wwoods on the scene.. like a test machine13:07
jeremypoelcat: ping13:07
f13wwoods: how do we stand qa wise?13:07
poelcatpresent13:07
* poelcat just found a bug in slocate13:07
poelcatit crashes13:07
poelcatjeremy: pong13:08
f13(I'm going to try and make this quick since I"m fighting compose issues)13:08
jwbmy limited QA is pretty good13:09
jwbbut i rarely test installs13:09
jeremyas far as NetworkManager -- I talked with dcbw about half an hour ago, and the conclusion we came to is that unless we find something horrific, sticking with the current NM seems like the best approach.  he's working on getting gconf migration working, but that kind of scares me13:09
wwoodsf13: stuff is mostly solid13:09
jeremytoo many cases to easily test, if something goes wrong, it's likely to be a crasher, etc13:09
warrenjeremy, without gconf migration, what happens to a user upgrading from FC6 or F7?13:09
wwoodsother than the fact that we don't have an installable tree ATM13:10
jeremywarren: they have to enter the settings for their network again.  which isn't great, but I'd rather that than it crashes on them13:10
nottingof course, if on install they have to enter their settings again, that sort of defeats the point of adding migration later13:10
jeremynotting: I'm not following you....13:11
nottingjeremy: well, if you're going to make everyone who uses NM re-enter their config, adding migration of their old settings later isn't really going to help anyone13:11
jeremynotting: helps release skippers!   also anyone that respins with updates13:12
jeremynotting: honestly, if we'd had it two weeks I'd be jumping all over it.  but it's going to be a not small amount of code with lots of string munging written in C. sooo :)13:13
jwbew13:13
wwoodsso yeah, scanning the blocker list real quick-like13:15
wwoodsso there's this standing bug where Turkish installs die13:15
jeremywwoods: I did a turkish install successfully yesterday13:15
wwoodswhich totally sucks but I don't think we can fix that in time... unless13:16
wwoodsunless jeremy pulls a miracle out of his ass, I was about to say13:16
wwoodset voila13:16
warrenIt seems that the change in -37 made a HUGE difference for my iwl3945 on x86_6413:16
wwoodsif I'm reading it right, we might have a fix for the AGP boot blocker13:16
jeremywwoods: yep, there's a patch there.  the other two kernel bugs have patches too... I was going to track down davej/cebbert after this13:17
* davej is invoked13:17
wwoodsSELinux stuff seems mostly sorted, except the ps3 thing that magically appeared today13:17
jeremywwoods: crontab is still broken by selinux13:17
jwbmy biggest REGRESSION that i've found is the backlight stuff.  but like i said last week, it's not a blocker13:17
warrenjwb, bug #?13:17
jeremywwoods: trying to track down dwalsh (he's not in the office today), and also getting other people to check13:18
jwbwarren, no bug yet.  i'm actually trying things myself13:18
nottingi'm attempting to fix ppp. *sigh*13:18
wwoodsjeremy: really? I thought that was fixed. bah and humbug13:18
jwbwarren, like building a kernel without jeremy's thinkpad patch to see if it magically fixes it13:18
poelcatcrontab doesn't work13:18
jeremynotting: *sigh*13:18
nottingjeremy: hey, it's beem broken since the end of june.13:18
wwoodsoh, and are we gonna build crc32c into the kernel so iSCSI works?13:18
jeremywwoods: I agree :/13:18
jeremywwoods: that's one of the kernel bugs13:18
davejI'll build it in, jcm seems to be on the hook for fixing it in mod-init-tools later.13:19
wwoodsit's a grotesque hack but as long as we have a plan for getting rid of it, I'm OK13:20
jeremythe blocker list is looking shockingly good right now13:21
wwoodsis there a plan for the qla2xxx thing?13:21
wwoodsjeremy: yeah, I think a couple of those are actually closeable too13:21
wwoodsgnokii, f'rinstance, and the Turkish installs13:21
jeremywwoods: we have the firmware packaged now.  not really handled properly by anaconda, but it was way too late hwen we found out about it to be able to fix it sanely :(13:22
jeremywwoods: yeah, a number are in MODIFIED because I figure me investigating, fixing and closing is a tad too much ;-)13:22
wwoodslemme know which ones you think can be retested / closed and I'll do so13:22
wwoodsand/or throw them into NEEDINFO reporter13:23
jeremywwoods: ones in MODIFIED should be pretty testable.  I don't think I fixed anything that wasn't trivially testable13:23
wwoodsgotcha13:23
wwoodsI'll check 'em out. so crontab/selinux sounds like it's the big thing we're waiting for13:24
wwoods(and fixing rawhide to have images again)13:24
jeremyyep13:24
jeremyso, fingers crossed, tomorrow's rawhide should hopefully be able to be it13:24
wwoodsa couple of people are pissy about firewire not working, but.. not a regression from F7, small number of users, planned fix for F9 (maybe an F8 update), and available workarounds13:25
wwoodserr, amend that: firewire not working with iso devices (cameras etc) and OHCI 1.0 controllers13:25
jeremy*nod*13:25
wwoodsrahul added the NV nosuspend change to the common bug page - we might want a whitelist or something in the future13:26
wwoodsbut I assume we can push out future pm-utils updates to handle something along those lines once we figure out what works13:26
jeremywell, I'd rather that we fix the driver and then just pull the hack rather than trying to do a whitelist ;)13:26
wwoodsone thing that worries me is longstanding bugs with the updates to libata etc13:26
wwoodspeople who were not updating to F7 because it didn't work on their hardware13:27
wwoodsif it still doesn't work with F8, they're kind of in trouble13:27
warrenhow widespread?13:27
warrenwwoods, do we have any such boxes, or did somebody test it?13:28
wwoodsspotty - mostly seems to affect subsets of people with old hardware + wonky drives13:28
jeremywwoods: I suspect there's not much we can do there13:28
wwoodsor things we don't have hardware to test (intel dmraid)13:28
jeremypeter is looking at the dmraid one now13:28
wwoodsI agree that there's not much we can do for the old hardware case but dmraid worries me more13:28
jeremy(he wasn't able to reproduce when using intel dmraid last week, but is trying again)13:29
warrendmraid bug #?13:29
wwoodsalan pointed out some raid+libata bugs but those turned out to mostly be mdraid13:29
jeremywarren: 34916113:29
wwoodswarren: https://bugzilla.redhat.com/show_bug.cgi?id=349161 is the blocker one13:29
wwoodsthere's some other similar things in need of investigation at the bottom of http://fedoraproject.org/wiki/Bugs/F8Common13:30
wwoodsoh, if there's anything on that page that's fixed now, lemme know13:30
* wwoods eyes "NetworkManager Segmentation Faults" and "kernel does not boot on SK8V motherboards" suspiciously13:31
nottingisn't sk8v the same agp bug?13:32
jeremythe former I haven't seen in a long while... the latter is the agp kernel bug and davej says he's going to add the patch13:32
wwoodsnotting: yes13:32
wwoodsoh, I've been testing yum upgrades from f8t3 to rawhide (we said we'd provide help/docs for people to do that) and adding stuff to http://fedoraproject.org/wiki/Tools/yum/YumUpgradeFaq13:33
wwoodsso if anyone asks, that's where the information lives13:33
* wwoods done13:34
davejnotting: not entirely sure the sk8v bug is the same cause as the agp bug.13:34
warrenSwitching from wired to wireless nm-applet crash was fixed?13:35
jeremywwoods: I'll take a look at the raid/lvm bugs at the bottom of F8Common as soon as I get a livecd-tools with all the currnet configs built13:35
jeremywarren: I can't reproduce a crash trying with today's rawhide13:35
jeremyI switched back and forth and back and forth and ...13:35
wwoodswarren: should be fixed, I'll be testing it in a sec13:35
wwoods(waiting for powerbook to reboot)13:36
jeremyanyway, do we have anything else?  or just keep trucking and fixing stuff13:37
jeremyand when do we want to reconvene?13:37
warrensame time tomorrow?13:38
nottinggoing to conflict with board .maybe slightly later?13:38
* jeremy is flexible...13:40
wwoodsoh hey, I have a question / discussion topic: should we remove old test release torrents after a certain time?13:43
wwoodsgetting tired of people installing f8t3 and reporting bugs against that13:43
jeremywe should do some torrent cleanup after the release13:43
wwoodsboth bugs in outdated packages and filing them against the version 'f8test3'13:43
jeremyalthough also, we're going to start hopefully having people hit 'torrent.fedoraproject.org' directly a little less too13:43
* mmcgrath agrees, possibly looking at a redesign of the torrent page would be good and finding a good home for the videos.13:44
jeremymmcgrath: fedoraproject.org/Videos with nicer links ?13:44
jeremybut neither here nor there13:44
mmcgrath<nod> something.13:45
warrenI do not like green eggs and ham.13:45
nottingwould you like them with a mouse?13:46
wwoodssomething to discuss in-depth at a later time, but yeah. really tired of people installing 4-week-old bits and telling me they're broken.13:46
wwoodswell, yeah dude, we spent the past 4 weeks fixing those bits. except it's harder for you to get those bits. anyway. I'm flexible for meeting time tomorrow13:46
warrenMeeting tomorrow at undisclosed time, at undisclosed location13:47
jeremyhow about 2?13:47
warrensounds good13:47
nottingwfm13:47
jeremy2 pm tomorrow it is13:48
poelcatjeremy: what is ck-list ?13:50
jeremypoelcat: lists sessions according to console kit13:51
poelcatjeremy: and why severity "low" ? sound is completely broken13:51
poelcatjeremy: i removed it!13:51
poelcati don't want it :)13:51
jeremyif you remove console kit, then you don't get console permissions!13:51
jeremyduh!13:51
wwoodswe ignore priority and severity on bug reports13:52
wwoodsset them to whatever the fuck you like13:52
poelcati thought consolekit was all about changing users?13:52
poelcati'm the only user; I don't want it :)13:52
jeremyno, it's not13:52
f13oh wow, forgot this was still going :/13:52
f13poelcat: there is root.13:53
f13poelcat: that's another user.13:53
* jeremy takes away poelcat's blocker filing privileges ;-)13:53
poelcathey what about crontab ?13:54
wwoodsyou wilfully removed random system packages that you didn't understand, and then filed a bug when stuff broke, and *then* added that bug to the blocker list?13:54
wwoodsyou now owe me a beer13:54
jeremypoelcat: crontab is still being looked into13:54
poelcatwwoods: or you owe me a better page explaiing why I need consolekit13:54
poelcati only disabled the service :)13:55
wwoodsno13:55
wwoodsheh13:55
f13ConsoleKit is a system daemon for tracking what users are logged13:55
f13into the system and how they interact with the computer (e.g.13:55
f13which keyboard and mouse they use).13:55
poelcatf13: okay :) it was advertised in our last release as a snazzy way to change users on the fly13:57
poelcati thought that was its only benefit13:57
f13no no, it just /supports/ the bits that allow you to choose users13:57
wwoodsit handles the backend bits, like changing the owner of the console devices (sound, keyboard, mouse, etc) to the person who is currently logged in13:57
f13without ConcoleKit you can't do fast user switching, but it doesn't do the switching for you thats fast user switcher.13:57
f13ConsoleKit and PolicyKit just take care of assigning ACLs where they need to be for access to hardware.13:57
jeremypoelcat: that's fast-user-switching.  not consolekit13:57
* poelcat now understands why yum remove ConsoleKit looked so ugly (I didn't do it btw)13:57
* poelcat stops being confused :)13:58
poelcatso.... is updatedb failing a blocker?13:58
* poelcat has caused enough trouble today... why not14:00
mmcgrathpoelcat: is it selinux?14:00
jeremypoelcat: updatedb just worked for me14:00
poelcatjeremy: does it need ConsoleKit too? ;-)14:00
poelcati'm trying to run a slimmed down system here... 1G of ram isn't what it used to be :)14:01
* wwoods running on 640MB14:04
jeremypoelcat: what's the bug #?14:09
poelcatjeremy: bug 35706114:14
jeremypoelcat: based on the strace, looks like something specific in your home directory (since that's where it's scanning at that point)14:21
f13hrm.14:45

Generated by irclog2html.py 2.3 by Marius Gedminas - find it at mg.pov.lt!