21:00:55 <otavio> #startmeeting
21:00:55 <MeetBot> Meeting started Mon Nov 23 21:00:55 2009 UTC.  The chair is otavio. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:00:55 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
21:01:06 <otavio> Could people wave please?
21:01:29 * fezie waves
21:02:13 <otavio> fezie: be welcome
21:02:38 * fil waves
21:02:42 <otavio> cjwatson: ?
21:02:45 <otavio> tbm: ?
21:02:46 <otavio> bubulle: ?
21:02:59 <otavio> luk_: ?
21:03:09 <otavio> maks: ?
21:03:26 <maks> pong
21:03:36 <otavio> Womble2: ?
21:03:39 <otavio> waldi: ?
21:03:44 <waldi> yes
21:03:57 <otavio> Sledge: ?
21:04:01 <otavio> dannf: ?
21:04:06 <otavio> Well
21:04:19 <otavio> #topic General status update
21:04:42 * cjwatson sort of here but not really
21:04:54 <CIA-3> debian-installer: 03cjwatson * r61452 10packages/arch/i386/grub-installer/ (debian/changelog grub-installer): Use GRUB Legacy on multipath disks for now (works around #442382 et al).
21:05:05 <otavio> We have two really serious issues to be fixed before we can release Alpha 1
21:05:36 <otavio> First is the g-i installer breakage due the GTK changes done during 2.1[78] development.
21:06:21 <cjwatson> from a wider standpoint, the problem is that nobody maintains the gdk directfb backend properly
21:06:27 <otavio> From what I've discussed with GNOME team please it looks to be due some deep changes to reduce the amount of code that diverges between X11 and DFB backend however I didn't look at the code.
21:06:44 <otavio> cjwatson: this is indeed a problem for us
21:07:07 <otavio> cjwatson: specially since people at GNOME team (in Debian) don't bother about it as well
21:07:26 <cjwatson> it's hard to ask them to do that when upstream is largely uninterested, although Sven Neumann has helped in the past
21:08:08 <otavio> I've talked to Jeremy - Lunar^ - and he said he can take a look at it but not before began of December
21:08:25 <otavio> cjwatson: is you able to take a look at it?
21:08:41 <cjwatson> I can't promise success, but I'll have a look this week
21:08:52 <otavio> cjwatson: good; thanks in advance by doing it.
21:09:18 * fil ponders if threatening to drop the graphical installer might provoke someone to care about it -- probably not ;-)
21:09:24 <otavio> So next issue. We have a very weird issue with udev that erase CD-RW medias
21:09:41 <otavio> It has been properly identified but it is still not fixed.
21:10:39 <otavio> It looks to be a conflict about SATA ATA and ATAPI CMDs and then the used CMD erases it. udev maintainer is aware of the issue and I hope it will be fixed soon
21:11:04 <otavio> cjwatson: btw, please upload grub-installer ;-)
21:11:52 <cjwatson> will do
21:12:01 <otavio> I've uploaded mipsel kernel udebs on Saturday to allow partman-ext3 to migrate to testing. I hope this is the last blocker (from packages POV) to installer upload.
21:12:20 <otavio> Well ...
21:13:09 <otavio> With this background in mind it looks we still have some time until we have installer in a releasable state so I'd like to ask about the possibility to us to move to 2.6.31 kernel udebs now.
21:13:17 <otavio> I see few pros doing it
21:13:37 <otavio> - Squeeze looks to target 2.6.32 as kernel and 2.6.31 is our current
21:13:41 <otavio> kernel so we'd be near of the target one in the release;
21:13:46 <otavio> 
21:13:50 <otavio> - We use this period of time until g-i is working again to stabilize
21:13:55 <otavio> the kernel and other stuff;
21:14:00 <otavio> 
21:14:03 <otavio> - We help kernel team to get more test on 2.6.31 and this help us to
21:14:08 <otavio> reduce distance from kernel team when possible;
21:14:10 <CIA-3> debian-installer: 03cjwatson * r61453 10packages/arch/i386/grub-installer/debian/changelog: releasing version 1.48
21:14:27 <otavio> Squeeze kernel looks to going to be 2.6.32 or 2.6.33
21:14:59 <CIA-3> debian-installer: 03cjwatson 07packages * r61454 10/arch/i386/grub-installer/1.48/: tagging version 1.48
21:15:00 <otavio> So I'd like to ask what the opinion of you guys about it?
21:15:13 <cjwatson> otavio: if it were me, I think I'd prefer the devil I know over the devil I don't, but I can understand the other position as well, so I think it's up to whoever's doing the work
21:16:09 <otavio> cjwatson: yes; I see this point as well however this period of time until we have g-i and udev fixed might be enough (or mostly) to us to have kernel done
21:16:43 <otavio> Specially if kernel team could help us to solve issues fast if we found them. (like speeding up a upload or so)
21:16:56 <otavio> maks, waldi, Womble2 : ^
21:16:59 <cjwatson> it is not clear to me that the "udev bug" is actually in uev
21:17:01 <cjwatson> udev
21:17:16 <cjwatson> all ata_id does is ioctl(HDIO_GET_IDENTITY)
21:17:23 <otavio> cjwatson: take a look at the bug; it is at kernel or udev
21:17:33 <otavio> cjwatson: yes; it can be at kernel ...
21:17:34 <waldi> otavio: you are already 4 weeks behind
21:18:03 <cjwatson> otavio: yes, I read the bug
21:18:05 <otavio> waldi: let's skip this discussion; I'm just asking if we can count on you (as team) on it
21:18:21 <cjwatson> ah, I'm looking at a pre-release of 147
21:18:37 <otavio> cjwatson: it looks like it changed the quering schema
21:18:42 <otavio> cjwatson: and this introduced it
21:18:46 <cjwatson> thank you, I can read the bug ;-)
21:18:55 <otavio> cjwatson: it changed to "advanced" ATA cmds
21:18:56 <cjwatson> I'm trying to follow through to actual code
21:19:47 <otavio> cjwatson: so my idea was to reduce the perid of time we'd be out of sync with the kernel.
21:20:00 <otavio> cjwatson: but this depends on the support we'll have from kernel team
21:20:08 <otavio> cjwatson: otherwise we stick with current one for a1
21:20:25 <cjwatson> I understand; as I say, I think it's up to whoever's doing the work to sync up d-i and get it working again (there's usually some breakage somewhere)
21:21:28 <cjwatson> anyway, the udev bug is fixed in udev git
21:21:41 <cjwatson> I'll forward a reference to the bug
21:21:45 <otavio> cjwatson: good; thx
21:22:09 <otavio> cjwatson: I fear Ubuntu is also affected by this bug as well?
21:22:47 <cjwatson> no
21:23:01 <cjwatson> not 9.10, anyway. we shipped a pre-release of udev 147 before this change
21:23:04 <fil> otavio: it seems like a good idea to me (the jump to 2.6.31) -- but my opinion's not really worth much in this particular case
21:23:17 <cjwatson> and 10.04 will ship with something rather newer I expect
21:23:26 <otavio> cjwatson: I see
21:24:00 <otavio> So; I'll post this idea for further discussion in mailing list and start a local branch for 2.6.31 work here locally
21:24:42 <otavio> fezie: from grub side; I'd like to know about the issue that Frans and Robert has been discussing lately
21:24:57 * otavio goes to fetch the bug number
21:25:05 <cjwatson> if the udev fix is uploaded reasonably swiftly, and we find a fix for the g-i bug, though, I think we should release with what we have if possible rather than introducing further indeterminate delays
21:25:22 <cjwatson> i.e. if that all happens before a kernel change is ready
21:25:25 <cjwatson> anyway, gone
21:25:26 <fezie> otavio: the thing is that the grub-pc package now asks for a device to install to
21:25:26 <otavio> fezie: #556738
21:25:32 <otavio> fezie: yes
21:25:52 <otavio> cjwatson: yes; I fully agree
21:26:03 <fezie> I think the best would be to just preseed this from grub-installer
21:26:14 <cjwatson> fezie: that's what I do in Ubuntu, I thought I'd committed that to Debian ...
21:26:19 <fezie> oh
21:26:22 <cjwatson> if I forgot, I'll deal with that
21:26:30 <otavio> fezie: could you and cjwatson take a look at that?
21:26:38 <otavio> fezie: good
21:26:39 <otavio> cjwatson: good
21:26:43 <fezie> cjwatson: ah right I remember some talk about that with you
21:27:08 <cjwatson> yeah, I see a delta there
21:27:09 <otavio> fezie: so please you or cjwatson upload grub-installer (again) once it is done?
21:27:26 <cjwatson> anyway, baby's crying, really need to go deal with her now
21:27:37 <otavio> cjwatson: no problem; you helped a lot :-)
21:27:44 <fezie> ok I can do the upload and even try to merge it if cjwatson wants me to do it
21:27:45 <otavio> cjwatson: go there and thanks a lot
21:28:04 <otavio> fezie: I'm sure cjwatson is OK with that
21:28:13 <fezie> :)
21:28:18 <otavio> fezie: so please go ahead
21:28:37 <otavio> So ...
21:28:54 <otavio> My current planned timeline for this all is:
21:28:59 * otavio goes to grab it
21:30:03 <otavio> The timeline I have in mind would be:
21:30:04 <otavio> 27 Nov 2009: have udebs of kernel uploaded 30 Nov 2009: meeting and review current status 04 Dec 2009: have g-i working again (if possible) 07 Dec 2009: meeting and review current status
21:30:07 <otavio> (the remaining assume g-i is working again at 07 Dec 2009)
21:30:10 <otavio> 11 Dec 2009: debian-installer is uploaded 15 Dec 2009: daily images are changed to use squeeze installer 17 Dec 2009: image test starts 21 Dec 2009: release
21:30:13 <otavio> oops
21:30:16 <otavio> crap
21:31:21 <otavio> http://pastebin.com/f4f6246b3
21:31:22 <otavio> ok
21:31:29 <otavio> I'll post it in ml anyway
21:31:53 <otavio> If we can fix g-i and udev fast we can speed it up a lot
21:33:03 <otavio> So far this is what I have as a status update
21:33:39 <otavio> A lot of new things has been commited and uploaded for tzsetup and localechooser and we will see how it behave in production in next days ...
21:33:58 <otavio> but besides those two major issues we're more or less ok
21:34:09 <otavio> #topic Another topic?
21:34:19 <otavio> Someone has another topic for discussion?
21:34:49 <fezie> me not
21:35:26 <fil> hopefully I'll have time to finally get round to doing the auto-install stuff we talkied about at DC9 -- would now be a good time to do that?  (sounds like it if I'm quick)
21:36:44 <otavio> fil: I doubt it will be at a1 but please go and run
21:36:53 <otavio> take 1
21:36:57 <otavio> take 2
21:37:03 <otavio> no other topic?
21:37:04 <fil> (mostly reinstating the stuff that was yanked out at the last minute in the lenny release, but then needing things like signed preseed/run etc.)
21:37:25 <otavio> fil: yes; I sort of remember that ...
21:37:35 <otavio> fil: it would be nice if you could prepare a patch for it
21:37:44 <otavio> fil: and then we discuss it on ml as usual :-)
21:38:18 <otavio> fil: but IIRC it is not a big change anymore since most part of it is in now. Am I wrong?
21:39:53 <fil> otavio: I think that's right -- it's mostly a case of fixing the bug that we didn't have time to fix before -- to do a decent patch will take perhaps more time than doing something that reintroduces that bug, and then fixing the bug -- I'll have a look
21:40:11 <otavio> fil: good
21:40:18 <otavio> so people I think it is done.
21:40:27 <otavio> I'll summarise it in a mail to ml
21:41:01 <fil> it's all confined to preseeding anyway -- the issue that kept it out was that preseeding is used by CDD's on their CDs
21:41:56 <otavio> #endmeeting