/srv/irclogs.ubuntu.com/2015/02/25/#ubuntu-release.txt

=== pleia2_ is now known as pleia2
=== rcj is now known as Guest72514
=== Odd_Blok1 is now known as Odd_Bloke
flexiondotorgLaney, would it be possible to sync ubuntu-mate-artwork 0.4.3ubuntu1 from proposed so it can be used to spin Beta1?09:36
Laneyflexiondotorg: OK (FWIW the request is for an 'unblock', it's being held because of the beta 1 freeze)09:37
Laneyhttp://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#ubuntu-mate-artwork09:37
LaneyDone - once rmadison shows the new version is in vivid you can start your rebuild09:39
flexiondotorgLaney, Thanks.10:11
oSoMoNwhat’s the freeze currently in application that prevents packages from migrating from proposed to release?10:22
cjwatsonoSoMoN: beta 110:26
oSoMoNcjwatson, pardon my ignorance, when will it be lifted?10:28
cjwatsonoSoMoN: after beta 1 is released.  see the release schedule10:28
cjwatsonhttps://wiki.ubuntu.com/VividVervet/ReleaseSchedule10:28
oSoMoNtomorrow if I read correctly then. thanks10:29
=== Symmetria is now known as Vortexia
=== Vortexia is now known as Symmetria
cjwatsonflexiondotorg: ^- powerpc appears to have succeeded too, it's just only on the tracker for daily and not beta 1 at the moment13:51
cjwatsonLaney: ^- doesn't appear to have download information set up on the tracker13:51
cjwatsonhttp://iso.qa.ubuntu.com/qatracker/milestones/326/builds/89692/downloads13:52
flexiondotorgcjwatson, OK. Is that something I can fix?13:52
cjwatsonflexiondotorg: I think it needs ubuntu-release, which is why I asked Laney13:54
cjwatsonoh, sorry, you mean beta 113:54
cjwatsonI don't remember, maybe Laney can13:55
flexiondotorgcjwatson, What is most important is it produce a powerpc iso I can get the ppc guys playing with.13:56
cjwatsonflexiondotorg: Well, you have that13:58
cjwatsonflexiondotorg: http://cdimage.ubuntu.com/ubuntu-mate/daily-live/current/13:59
cjwatsonDon't commit the error of thinking that the tracker is everything :)13:59
flexiondotorgcjwatson, 😃13:59
=== Guest72514 is now known as rcj
=== rcj is now known as Guest22529
Laneyflexiondotorg: cjwatson: Done, added to Beta 1 milestone also14:22
flexiondotorgLaney, Thanks!14:28
cjwatsonLaney: ta14:41
flexiondotorgLaney, I rebuild the Ubuntu MATE isos a little while ago.14:59
flexiondotorgLaney, This spin is looking good for Beta 1.14:59
flexiondotorgLaney, Do I have to do anything to mark it as Beta 1 candidate or will it just automatically get ellected?14:59
Laneyflexiondotorg: Use "Mark as ready" on the tracker to indicate so15:00
flexiondotorgLaney, does that mean "i've tested it and it is ready"?15:01
LaneyIt means that it's ready to be released with the milestone15:01
flexiondotorgOK15:02
LaneyWhether it's tested or not is a matter for the person pressing the button. :P15:02
bdmurraycjwatson: the autopackage stuff is ready for review https://code.launchpad.net/~brian-murray/ubuntu-archive-tools/autopkg-excuses/+merge/25079515:10
bdmurrayand there is another ubuntu-archive-tools mp - https://code.launchpad.net/~brian-murray/ubuntu-archive-tools/no-deleted/+merge/25081115:12
cjwatsonoh, right, yeah, those were on my to-do for today, let me look now15:13
cjwatsonbdmurray: both merged, thanks!15:19
bdmurraycjwatson: great, thank you15:21
infinitybdmurray: Say, if you're fiddling in sru-report, I have a bug for you that I was just looking at how best to fix.15:24
infinitybdmurray: In get_srus, we skip packages we don't want showing up in the lists due to noise (langpacks, kernels, etc), but that's the wrong place to skip them, as it means they also don't show up on the removals list.15:25
infinitybdmurray: Which has never been a problem until we did britney integration, now we really do want the proposed versions deleted if they're hanging around.15:26
infinitycjwatson: Though, maybe this is a self-solving problem if the ultimate goal is for migration to be unblocks, and britney does an atomic copy/delete for me, like it does in the devel series?15:26
infinityWould still be nice to have the report tell me about removals when that screws up, though.15:27
bdmurrayinfinity: noted15:31
infinitybdmurray: The naive fix would have just been to move the filter into the HTML output section instead, but if we're not displaying all the info for those, it's pretty inefficient (especially for kernels) to go bug-scraping, etc.  So, perhaps something more clever like having them still be in the srus list, but skip scraping for extra info, and add a display=bool property set to false?  I dunno.15:34
cjwatsoninfinity: Yeah, I think this is ultimately why I just had to do a big pile of binary removals from -proposed for old kernels.  I was thinking about the same thing today.15:34
infinitycjwatson: Yeah, it's definitely the cause of your pain.  We used to remove old kernels from proposed, and then I started filtering them out of the list, which meant they hung around forever.15:35
infinitycjwatson: I never cared, since it has zero impact on the size of /pool/ and minimal impact on proposed/Packages, but it suddenly matters.15:35
cjwatsoninfinity: If I get round to finishing my copyPackage(move=True) patch, then you could just use that in sru-release, of course, and the removals section no longer matters.15:36
cjwatson(Or movePackage.  Whatever.)15:36
cjwatsonproposed-migration would use that too under the hood, of course, but it could equally be used directly.15:37
infinityAnd, of course, my filtering was just a cargo-cult of the existing langpack filtering, but langpacks don't suffer the same issue, since they don't introduce NEW binaries on every upload.15:37
cjwatsonOr, even now, sru-release could do the same thing that promote-to-release does and do the delete.15:37
infinitycjwatson: Yeah, sru-release could do the copy-and-delete together, but we still know that (very rarely) fails to delete sometimes, hence the removal list is still helpful.15:38
infinityWe still catch the odd manual removal needed in vivid here and there.15:38
infinityUnless that was finally hunted down and stamped out?15:38
cjwatsonRight.  This is ultimately why I want an atomic move operation.15:38
cjwatsonIt's entirely bogus to have to do those separately.15:38
infinityYep.15:38
infinityAn actual atomic move would be lovely.15:38
cjwatsonThe case where copyPackage succeeds, the actual copy job fails, and the delete succeeds is much worse, and does happen once in a blue moon.15:39
infinityI'd also like some atomic batch operations... I wonder if I ever files bugs.15:39
cjwatsonBatching can be harder due to timeouts, but atomic move of a single package is quite easy.15:39
infinity(ie: change-override should be using an atomic batch interface, and an atomic batch for a set of moves would also be awesome, so I don't get hit by a random delay making linux-meta release before linux)15:40
cjwatsonI had most of the patch for that but I keep getting distracted ...15:40
infinitycjwatson: I assume the API is flexible enough to be able to crank up timeouts for specific operations?15:42
infinitycjwatson: Seems a waste to back this whole thing on a database with transactional support and not leverage it more.15:42
cjwatsonThat's controllable by feature flags per-operation if necessary, yes.15:43
infinityAnd, actually, a batch change-overrides would probably be much faster than what we currently do, too.  So long as it doesn't lock an entire table when it happens. :P15:43
infinityOr worse, several.15:43
cjwatsonThat's probably the issue.  I don't understand the data model quite well enough for that.15:43
cjwatsonAlso we'd need to put the operation on some other base object.15:44
infinitycjwatson: Well, I can hold off on these fancy requests, since I know that both overrides and copies are on wgrant's hitlist for his feature redesign.15:45
cjwatsonAnd I suspect we'd end up encoding a bit more policy into the Launchpad side, because it would need to have things like checking for which architectures still have active publications, that are currently checked on the change-override side.15:45
cjwatsonYeah.15:45
infinitycjwatson: So, a fresh data model might make these things easier.  Though, "will this allow for sane batching and locking" should probably be a design goal.15:45
cjwatsonI need to go off and understand the current bits of that redesign at some point.15:46
cjwatsonBut slightly full right now. :)15:46
=== alex_abreu is now known as alex-abreu
elfyI suspect that I've missed Laney by a country mile ...18:26
Laneyelfy: suspect no more18:57
elfy:)18:58
elfyso  - it looks like everyone (except us) are all looking good where testing18:58
elfyI'll chase studio up shortly18:59
elfyxubuntu has a contingency plan - doesn't look like we'll get our issue sorted in time to test18:59
elfynot sure what other things you need from me tbh18:59
elfyatm anyway19:00
elfyI'll sort out the release announcment tomorrow19:00
Laneyjust herd the cats into being ready in time, and send the announcement tomorrow19:00
elfyyea19:00
elfyloads of fun, I'm not internet friendly at work :)19:01
wxlhey infinity you ever figure out what's wrong with our alternate? it looks like the issue may extend beyond just that last point release :( http://pastebin.com/VKECPUCK19:51
wxlinfinity: nevermind. bad tester.19:53
infinitywxl: I'm not sure what you mean?  You mean trusty dailies still show it?  Of course they do, nothing has changed to fix it since 14.04.2 was released.19:53
wxli meant in vivid infinity but i see that this is based off of 14.04.2 :/19:54
elfyinfinity: someone a lot cleverer than me has worked out why the install windows are off -> somewhere :)20:53
=== Guest22529 is now known as rcj
=== rcj is now known as Guest62484
infinityelfy: Oh?  Have a pointer to the analysis/fix?  I'm curious.23:08
=== s8321414_ is now known as s8321414

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!