infinity | cjwatson: I vaguely recall it being discussed a week or two back, but did we decide if we were going to hard freeze with queue reviews, or just do some sort of massive britney block? | 00:06 |
---|---|---|
infinity | cjwatson: (I actually prefer the hard freeze, honestly, the reviews are helpful, and a britney block could just lead to a ton of crap in proposed that we need to dump at release...) | 00:07 |
=== mcasadevall__ is now known as NCommander | ||
micahg | can someone please help chromium-browser ubuntu2 to the release pocket? | 01:34 |
vibhav | Sorry for cross posting, but the topic needs to be updates here too :) | 04:23 |
vibhav | (feature freeze->final beta freeze) | 04:23 |
=== infinity changed the topic of #ubuntu-release to: Ubuntu 12.10 and 12.04.2 released | Archive: Final Beta Freeze | Raring Ringtail Release Coordination. Please don't upload things during freezes where you shouldn't, or be prepared to apologise to the release team | we accept payment in cash, check or beer | melior malum quod cognoscis | ||
=== mmrazik is now known as mmrazik|lunch | ||
ogra_ | cjwatson, hmm, something is wrong with the cdimage publisher code ... the nexus7 desktop image /current link is stuck on an image from the 15th | 11:42 |
=== mmrazik|lunch is now known as mmrazik | ||
ogra_ | cjwatson, hmm, seems to be even worse ... http://cdimage.ubuntu.com/daily-preinstalled/ has subdirs for the 15th, 26th and 27th but only the 15th is populated | 11:48 |
ogra_ | the others are empoty dirs | 11:48 |
ogra_ | weird, in the log i see it downloading the ext4 file ... but it doesnt seem to find it (and in fact its not in the preinstalled subdir on nusakan) | 12:08 |
ogra_ | cjwatson, aha, seems the preinstalled filesystems end up in /srv/cdimage.ubuntu.com/scratch/ubuntu/raring/daily-preinstalled/live/ instead of /srv/cdimage.ubuntu.com/scratch/ubuntu/raring/daily-preinstalled/preinstalled/ | 12:24 |
ogra_ | ah, line 509 in livefs.py | 12:31 |
ogra_ | (or rather 510 since it is wrapped) | 12:33 |
cjwatson | infinity: there was a thread on -release about that, I think - I haven't really been following | 12:38 |
cjwatson | ogra_: ok, give me a minute to have a look | 12:38 |
cjwatson | (holiday today so may get interrupted) | 12:38 |
ogra_ | cjwatson, ah, i didnt even expect to see you around ... was just starting to quickly hack a "default_dir" var in to switch it for preinstalled :) | 12:39 |
cjwatson | I'd rather you didn't | 12:39 |
ogra_ | yeah, quick hack ... | 12:39 |
cjwatson | There's really no reason to have a different path there | 12:39 |
cjwatson | I was deliberately trying to consolidate live and preinstalled, since the bulk of the differences there were entirely artificial | 12:40 |
ogra_ | ah, so you fix it on the publisher side ... k | 12:40 |
ogra_ | thats the part i hadn't looked at yet | 12:40 |
cjwatson | I hadn't realised that debian-cd cared | 12:41 |
cjwatson | It's like a single giant layering violation | 12:41 |
cjwatson | The reason it's safe to have it be .../live/ is that the image_type is always different | 12:42 |
cjwatson | So I'm just going to s/PREINSTALLEDIMAGES/LIVEIMAGES/g, basically | 12:43 |
ogra_ | yeah | 12:43 |
cjwatson | Sorry for not noticing the failures | 12:44 |
ogra_ | well, it was me not noticing it | 12:44 |
ogra_ | and i still wouldnt have ... if there wasnt a user in #ubuntu-touch complaining about the n7 desktop images | 12:44 |
cjwatson | Just trying a fresh build now | 12:45 |
ogra_ | well, essentially complaining about compiz still crashing in ubiquity-dm :) | 12:45 |
cjwatson | Urgh, what | 12:45 |
* cjwatson has broken something else, hang on a minute | 12:45 | |
cjwatson | all hail pdb | 12:46 |
cjwatson | And it's broken really early which is why we haven't even been getting failure logs | 12:47 |
ogra_ | the livefs builds look all fine, i dont think you need to re-do them ... unless you want to test buildlive | 12:47 |
* ogra_ sees proper rootfs.tar.gz for all of them | 12:48 | |
cjwatson | No, I hadn't been planning to | 12:48 |
cjwatson | Ah, missing test | 12:51 |
ogra_ | oh, the path was even hardcoded in CONF.sh | 12:52 |
cjwatson | Right, cdimage.config fixed (and test cases added); cron.daily-preinstalled running now | 12:53 |
cjwatson | I'm not likely to be around much more for several hours - if anyone wants to start builds of other things affected by the same problem (which won't have been everything - it was any project/image_type combination for which there was an open-ended series range in etc/default-arches), feel free, otherwise I guess cron will sort it out | 12:55 |
ogra_ | yeah, lets leave it to cron | 12:57 |
* ogra_ didnt plan to work during his vacation either ... i would have added a hack to be ripped out on tuesday | 12:57 | |
cjwatson | ogra_: that seems to have worked | 13:14 |
ogra_ | yep, and i see a /pending dir too now :) | 13:14 |
* ogra_ would dance ... but its good friday and i cant break the law :P (dance prohibition in germany) | 13:18 | |
ogra_ | (its that one law that makes germans feel like americans :) ) | 13:19 |
* skaet figures that as long as germans can drink beer on good friday, they can survive without the dancing ;) | 13:29 | |
ogra_ | skaet, haha, yeah ... | 13:36 |
skaet | :) | 13:37 |
ogra_ | (some here call it footloose friday here :) ) | 13:37 |
ogra_ | -here | 13:37 |
ogra_ | clubs make a lot of money though ... people tend to gather at 11 and the music starts at 12 | 13:38 |
skaet | dancing starting at midnight hmm... in the US the equivalent is probably holiday shopping starting on the Friday after Thanksgiving. | 13:44 |
Trevinho | Hello, do we really need a FFe for this: https://bugs.launchpad.net/unity/+bug/1152477 ? | 14:14 |
ubot2` | Launchpad bug 1152477 in unity "[FFe] Window list in right click context menu should indicate the focused window" [Medium,In progress] | 14:14 |
Trevinho | it's a really minor change, involving no API change or string change | 14:14 |
ogra_ | its a new feature ... | 14:15 |
ogra_ | and the doc team will need to take it into accound in screenshots for example ... | 14:15 |
didrocks | ogra_: we do agree :) | 14:15 |
ogra_ | its definitely a "just nod it off" FFe | 14:16 |
ogra_ | buut nonthelsee an FFe | 14:16 |
didrocks | yep | 14:16 |
ogra_ | *nontheless | 14:16 |
xnox | Trevinho: doesn't it now need a UserInterfaceFreezeException as well now?! | 14:21 |
Trevinho | xnox: it does not affect any screenshot or big ui change. | 14:24 |
ogra_ | how do you know the docteam didnt make a screenshot of the right click menu ? | 14:24 |
Trevinho | ogra_: the right-click menu could not been affected in all the cases by this change | 14:25 |
Trevinho | ogra_: you need that the doc team has two-windows of the same app opened and one of them focused to notice a change | 14:25 |
ogra_ | well, the process is to get the FFe nodded off and i think xnox is right that you also need the docteam to agree | 14:25 |
ogra_ | its not like thats hard or anything | 14:25 |
Trevinho | ogra_: and the change would be trival anyway (not seeing a * in an item is really a small thing) | 14:25 |
ogra_ | sure, still we're past freezes and there is a process for this ... if the change is trivial it will go very fast ... | 14:26 |
Trevinho | ogra_: ok... | 14:27 |
ogra_ | (the time you spent discussing it here might have been sufficient to mail the docteam ;) ) | 14:27 |
Trevinho | ogra_: I only was asking this because in the past we never neded it for such trivial changes. | 14:27 |
=== mmrazik is now known as mmrazik|afk | ||
infinity | cjwatson: Ahh, indeed, it was on the list. I've followed up. Thanks for having a slightly less swiss-cheesy brain. | 15:32 |
ogra_ | infinity, so you got some sleep ? | 15:33 |
infinity | ogra_: I did! | 15:33 |
infinity | ogra_: It does happen once in a while. Just not every night. :P | 15:33 |
ogra_ | yay ! | 15:33 |
infinity | (But last night, I actually slept, uhm... 5 hours?) | 15:33 |
infinity | Which, for me, isn't a bad stretch. | 15:34 |
ogra_ | athst something at least | 15:34 |
ogra_ | *that's | 15:34 |
=== mmrazik|afk is now known as mmrazik | ||
Mirv | FYI I'm not sure what's the story behind qtpim upload at https://launchpad.net/ubuntu/raring/+queue?queue_state=0&queue_text= , but from my point of view it shouldn't have been uploaded (packaging is not complete, the organizer doesn't work at all without some patches I and a community member have submitted upstream etc) | 17:21 |
=== pgraner-afk is now known as pgraner | ||
pgraner | infinity, can you trigger a rebuild for https://bugs.launchpad.net/ubuntu/+bug/1129949 | 17:25 |
ubot2` | Launchpad bug 1129949 in ubuntu "corrupt package libgrail5_3.0.6-0ubuntu0.12.04.01_i386.deb on archive.ubuntu.com" [Undecided,New] | 17:25 |
ScottK | Mirv: Rejected. | 17:27 |
=== mmrazik is now known as mmrazik|afk | ||
infinity | pgraner: Erm, wait. What? | 17:27 |
pgraner | infinity, tiaz pinged me and asked that I do it, but I don't have the voodoo to make it happen | 17:28 |
infinity | pgraner: If the actual file is corrupt, we should just fix it. | 17:28 |
Mirv | ScottK: thanks | 17:28 |
infinity | (And if it is, that's worrying, to say the least) | 17:28 |
ScottK | Easy decision since there's no FFe. | 17:28 |
pgraner | infinity, I asked tiaz to join so you can talk to him :) | 17:30 |
infinity | tiaz: Yo. | 17:30 |
infinity | Why are we rebuilding packages to fix filesystem corruption? | 17:31 |
infinity | (And why do we have corruption at all?) | 17:31 |
tiaz | infinity: sorry, it occurred to me that I hadn't checked that we're still serving the bad package. I did confirm the one provided was bad | 17:32 |
tiaz | looking for it now, just a sec | 17:32 |
infinity | So, the one on pepo matches the librarian. | 17:33 |
infinity | Is it broken on syncproxy, or the archive frontends? | 17:33 |
tiaz | infinity: I think it was broken on the frontends and is no longer. the one I just fetched from a.u.c is correct, but I don't know which mirror I got it from. I'm going to pull it from all of them to make sure, but this may be a noop | 17:34 |
infinity | If you just make sure all the archive frontends (and syncproxy) match pepo, and artificially touch the package to futz the timestamp, it should cascase and fix all the mirrors. | 17:34 |
infinity | The timestamp/touch thing is essential, since most people don't rsync with sums (cause, ow, painful on a 0.5TB mirror), so if the file name/length/time don't change, they won't fix the file locally. | 17:35 |
infinity | I'll touch it on pepo now, so it doesn't revert on mirror pushes. | 17:35 |
infinity | All I can assume here is that one archive frontend broke, and it happened to be a parent to some other ccTLD mirror. | 17:36 |
infinity | Would be nice if we knew which frontend that was, since it could point at something much worse. | 17:36 |
tiaz | infinity: all the current mirrors are correct. I've touched them just in case, I'm now looking at if we had a different set when this was opened, just in case | 17:38 |
infinity | Oh, I didn't notice the bug was a month old. | 17:39 |
infinity | That makes it a bit harder to trace. | 17:39 |
infinity | Bother. | 17:39 |
infinity | tiaz: I assume you double-checked syncproxy too? | 17:39 |
tiaz | infinity: yeah | 17:40 |
infinity | Kay. Well, no one should be rsyncing (or, in fact, allowed to rsync) from anything that isn't a child of one of the archive frontends. So, if they're all happy, it should resolve itself. | 17:41 |
infinity | And, like I said, I futzed the timestamp on pepo, and that should bubble out to the world and force every mirror to re-sync the package for no reason. :P | 17:41 |
infinity | But that's still better than a whole new SRU for nothing. | 17:42 |
infinity | I do wish the original reporter had been a bit more verbose. | 17:42 |
tiaz | it would be really nice if they'd specified what mirror they got it from, yeah. | 17:43 |
cjwatson | For all we know it was from an apt proxy of some kind. | 17:43 |
cjwatson | And they only think it was corrupt on a mirror. | 17:43 |
infinity | Yep. | 17:43 |
infinity | But this quick touch faff isn't harming anyone. | 17:43 |
infinity | So, whatever. | 17:43 |
infinity | Belt and bracers. | 17:44 |
infinity | And such. | 17:44 |
cjwatson | Some versions of apt-cacher-ng, say, are REALLY good at corrupting packages | 17:44 |
cjwatson | Like the one in precise | 17:44 |
infinity | My apt-cacher-ng behaves quite well. I must be lucky. | 17:44 |
cjwatson | (Yes, there's an SRU in train for this, I believe) | 17:44 |
infinity | Oh, or I've never run the precise version. :P | 17:44 |
cjwatson | The one in precise occasionally thinks headers are data. | 17:45 |
cjwatson | Comedy results. | 17:45 |
infinity | I still need to sit down with a large stick and beat the notion of combined mirrors into its tiny brain. | 17:45 |
tiaz | infinity: I was going to close it as we checked and it's not a problem, but I see adconrad just commented. should I leave it? | 17:45 |
cjwatson | tiaz: /whois | 17:46 |
infinity | You can sort of fool it into putting archive.u.c and ports.u.c into the same cache (so you don't have to download _all.deb twice), which is shiny and all, but then the cron job explodes when it tries to verify things. | 17:46 |
cjwatson | :) | 17:46 |
infinity | Silly. | 17:46 |
infinity | tiaz: ... srsly? | 17:46 |
infinity | tiaz: How long have we worked together, dude? :) | 17:46 |
ScottK | Not long enough apparently. | 17:47 |
infinity | Or too long. I've been known to drive people insane. | 17:47 |
tiaz | infinity: ha! I'm terrible with names and tend to remember irc nicks :) | 17:48 |
xnox | infinity: who can accept stuff if you freeze the archive. Release team + archive admins? | 20:27 |
ScottK | xnox: Technically yes, but by policy the release team. | 20:31 |
xnox | ScottK: infinity: will there still be britney block for everything on CDs? | 20:32 |
ScottK | We've yet to actually write the britney block that would manage that. | 20:33 |
ScottK | Although we got close. | 20:33 |
* ScottK doesn't see much point. | 20:33 | |
infinity | I see no point in a britney block if we freeze the queue. | 20:40 |
xnox | Ok. Also archive rebuild in progress.... | 20:45 |
xnox | infinity: AAs can't twiddle archive block as easily? | 20:48 |
xnox | s/archive/britney/ | 20:49 |
ScottK | The ones that aren't also on the release team should know better. | 20:52 |
xnox | ScottK: I'm confused about bug 1077624. | 21:02 |
ubot2` | Launchpad bug 1077624 in simgear (Ubuntu) "Raring: Update Flightgear to version 2.10.0" [Wishlist,Confirmed] https://launchpad.net/bugs/1077624 | 21:02 |
xnox | it doesn't look like FFe has been granted yet, but it shows up in the sponsorship queue. | 21:02 |
ScottK | xnox: OK. It does now. | 21:14 |
xnox | ScottK: about bug 1159832, am I ok to sync the new bitcoin from experimental? | 21:15 |
ubot2` | Launchpad bug 1159832 in bitcoin (Ubuntu) "[FFE] bitcoin: Mandatory upgrade on May 15" [High,New] https://launchpad.net/bugs/1159832 | 21:15 |
ScottK | xnox: Yes. | 21:15 |
xnox | ScottK: thanks. | 21:15 |
* xnox testing. | 21:15 | |
ScottK | No problem. | 21:15 |
ScottK | Thanks for checking. | 21:15 |
phillw1 | cjwatson: if you're still about, I have a failure on a mini-iso who should be informed? | 23:39 |
phillw1 | cjwatson: for your, and others information https://bugs.launchpad.net/ubuntu/+source/debian-installer/+bug/1161898 | 23:42 |
ubot2` | Launchpad bug 1161898 in debian-installer "Raring 1386 mini.iso complains about mismatched kernel" [Undecided,New] | 23:42 |
cjwatson | phillw1: It's invalid. I'll reply on the bug. | 23:47 |
cjwatson | stgraber: ^- You might want to look at that, though, to see if there's some reason the tracker hadn't been updated. I don't know whether that's usually automatic or manual. | 23:51 |
cjwatson | stgraber: The tracker said 20101020ubuntu225 for all the netboot images, despite the last d-i update being a day or two ago. | 23:52 |
cjwatson | (20101020ubuntu226) | 23:52 |
phillw1 | cjwatson: thanks, it's one of the lubuntu testers trying to get some pre-testing in for the mini iso's that we do use as 'worst case' smallest installs :D | 23:52 |
cjwatson | stgraber: I don't know whether this is related to the upgrade testcases not being updated, as mentioned on ubuntu-release@; I don't know whether that's normally automatic or manual | 23:52 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!