=== SteveA_ [n=steve@costello.z3u.com] has joined #launchpad-meeting === mpool [n=mbp@ozlabs.org] has joined #launchpad-meeting === SteveA_ [n=steve@costello.z3u.com] has joined #launchpad-meeting === jamesh [n=james@203-59-20-109.dyn.iinet.net.au] has joined #launchpad-meeting === SteveA [n=steve@costello.z3u.com] has joined #launchpad-meeting [12:05] lifeless: I guess ddaa is on holiday. Was there any decision on who would chair? === lifeless votes jamesh [12:06] SteveA, mpool, lifeless, spiv, jamesh: is everyone here? [12:06] I'm here. [12:08] no SteveA or mpool? [12:09] I'll nag [12:10] should we start without them? [12:10] ok, mpool nagged succesfully [12:10] I haven't nagged stevea [12:13] well, we're almost a third of the way through the timeslot, so we should probably start. [12:13] * roll call [12:13] * production status [12:13] * Smart server [12:13] * SFTP advertising [12:13] * vcs-import knits [12:13] * tarball-spider [12:13] * Python import [12:13] * critical bugs [12:13] * pending sysadmin tasks [12:13] * any other business [12:13] hi [12:13] hi [12:14] first up is production status. Does anyone have anything to report? [12:15] will the supermirror be down for writing when launchpad is down tomorrow? [12:15] if so, do the right people know about it? [12:15] Yes. [12:16] we've still got bug #53825 (input validation problems in the branch puller), but I don't think there is any new issues. [12:16] will it fail elegantly, or inelegantly? [12:16] It will refuse authentication. [12:16] as a future thing, can we make it refuse with a message [12:16] like "launchpad is down for maintenance" ? [12:16] SFTP doesn't give us a nice way to say "sorry, we're down for maintenance" that I know of. [12:16] then at some point, we should do authserver cacheing with it [12:16] or some such [12:16] Well, http://bazaar.launchpad.net/ will still be serving read-only branches ok. [12:17] maybe that's fixed when we go to two databases [12:17] mpool: what do you think about announcing it to the bzr list? [12:17] (unless the cronscript generating the branch name mapping goes silly when the db is down?) [12:17] announcing the launchpad downtime, that is [12:17] I think 2 dbs is the right fix for it [12:17] i think spiv should send a message to the list [12:17] or someone [12:17] it would be nice if it could continue in r/o mode [12:18] spiv: looks like the rewritemap script opens the output file for writing after connecting to the DB, so assuming that fails it should be fine. [12:19] I suggested to stub at the launchpad meeting that we should point the authserver at a r/o copy of the database -- that would still allow writes to existing branches, at least, but I don't know if he's had time to do that, it may be too much to ask at short notice. [12:19] next on the agenda is the smart server status [12:19] we kept working on it [12:20] spiv: a read only connection to staging might be an option for that (or maybe for future downtime like this) [12:20] so nothing new to report for smart server? [12:20] some parts merged [12:20] more in progress [12:20] all conformance test passing, which means we're now in conversion mode rather than model overhaul [12:21] cool. [12:21] next up is SFTP advertising [12:21] I published the team branches article last week: http://blogs.gnome.org/view/jamesh/2006/08/17/1 [12:22] there is a note in the agenda about spiv blogging about similarities between SVN and bzr usage with checkouts [12:22] it was well received outside of canonical, jamesh [12:22] mgedmin read it and commented on irc about bzr looking better and better all the time [12:23] SteveA: cool. We should check if it has a visible effect in the cricket statistics [12:23] since it is tracking the number of team owned branches [12:23] if the sftp server gets improved to not need --make-parents (or whatever the option is) from bzr [12:23] we may have got another user today, olive was the linchpin for him [12:23] then you could update the blog with that [12:23] what does "olive was the linchpin" mean? [12:24] sounds like secret agent codes [12:24] as his users include only-comfortable -with-guis folk [12:24] olive is our gui [12:24] written as part of SoC [12:24] i see [12:24] KDE and GTK, known to work on windows [12:24] it's a bit of an opaque codename [12:24] spiv: do you think getting rid of the --make-parents requirement would be easy for someone else to do, or would it require more twisted knowledge? [12:25] jamesh: I think your analysis is probably correct -- it's probably not too hard. [12:25] (I mean --create-prefix rather than --make-parents) [12:25] The twisted knowledge necessary should be cargo-cultable [12:26] spiv: okay. I'll have a go at it tomorrow then. [12:26] if I get in over my head I'll ping you. [12:27] lifeless: imo, don't call it "olive" in public. if you have $$$$ to spend on marketing like apple, then you can call your GUI "aqua" or other cute names. [12:27] but it's cliquey and offputting for a small project to obfuscate names, where more obvious (but boring) ones will work [12:27] next up on the agenda is "vcs-import knits", but last week's minutes says that it was completed so I'll skip it [12:28] SteveA: fair enough. I've had almost no involvement with it though, so I'll redirect your feedback to mpool [12:28] thanks [12:28] in line with the other naming thread, we should just move to calling them 'bazaar gui' [12:28] they need to all technically merge together [12:28] ah, about knits [12:28] james troup (or another sysadmin) commented to me [12:28] that there was a lot, a LOT, less space being used [12:29] on the SM machine [12:29] that's good to hear [12:29] he was concerned that some evil data loss had occured [12:29] so, in future, maybe warn the admins of such improvements ;-) [12:29] heh, we did [12:29] when we rolled it out [12:29] next up is the script formerly known as dyson [12:30] I landed my fixes for the last round of bugs today [12:30] that's the... tarball non-spider? [12:30] and renamed it to "product-release-finder" [12:30] lifeless has asked stub to do another test run of it on staging, so we'll have more to report next week [12:32] next is Python import. Given ddaa isn't here, I'll skip it [12:32] unless anyone else knows the status [12:32] noidea [12:32] okay. On to critical bugs. [12:32] bug 31308: Cannot set branch associated to a product series. [12:33] Mark mailed some of us with some concerns about having two branch references in the product series, but decided that it sounded okay after I explained what we were planning [12:33] lifeless: are you still responsible for specing this? [12:34] nope [12:34] last meeting ddaa said to not spec it [12:34] or around that time [12:34] anyhow, its now just down to implementation of the extra attribute [12:34] is david going to do that? [12:35] last meeting's minutes says it has been assigned back to him [12:35] The bug is assigned to him. [12:35] bug 37897: renaming project, product or series breaks vcs imports [12:36] looks like this one is now in my review queue [12:36] using the design discussed at the sprint [12:36] bug 51130: cannot use +admin on a branch I own [12:37] the branch mentioned in last minutes for this one is in BjornT's review queue [12:37] so I guess things are looking pretty good. [12:37] next up is pending sysadmin tasks [12:38] is anyone blocked on sysadmin stuff? === jamesh takes that as a no [12:39] next then? [12:39] there are a few items listed under proposed items, but I think they are unchanged from last week. Should I go through them? [12:40] the first is lp: URLs for bzr. [12:40] you're the matre'd [12:40] there have been some comments added to the spec which probably need integrating. I don't think there is much more to add. [12:40] jamesh, spiv: have you read the spec? [12:41] mpool: yes. I added some of the comments. [12:41] mpool: not yet, I'll do that. [12:42] mpool: it looked pretty good. The client side would be pretty easy to implement with better redirect handling in bzr :) [12:43] ok, so: ACTION: spiv, read the spec [12:43] ACTION: mbp read/integrate the comments [12:43] next up is "important bugs according to bzr community" [12:44] i spoke to david and mark [12:44] i'll update the spec list instead [12:44] so that item is closed [12:44] okay. [12:44] next is "Move vcs-import data out of ProductSeries" [12:45] I don't think there is much to add here. It probably needs a bit of specing [12:45] next is "1.0 targets, what, who, when?" [12:46] we've already covered the smart server work, so I assume that's on target [12:47] jamesh: do you have any 1.0 bazaar-related targets? [12:47] does that spec about lp: urls take into account our canonical pillar names? [12:47] spiv: do you have any other ditto? [12:47] SteveA: yes [12:47] great [12:47] https://launchpad.canonical.com/BranchIndirection [12:47] if you want to read it [12:47] thanks [12:48] mpool: other ditto? [12:48] mpool: I'm not working on any of the 1.0 target bazaar specs at the moment. [12:48] any lp 1.0 targets related to bazaar? [12:48] that spec should be registered in LP ... [12:49] Oh, right. Just the smart server. [12:50] One last thing before we finish. There was an action item from the last meeting that we haven't covered [12:50] ACTION: spiv and ddaa to review PrivateBranches again [12:50] spiv: is that still a todo? [12:51] Yes, but I've skimmed and it won't take long for me to do. [12:51] okay. I think that's it. [12:51] Meeting ends. [12:51] thanks for chairing [12:51] thanks jamesh === spiv [n=andrew@218-214-66-203.people.net.au] has joined #launchpad-meeting