=== ddaa [n=ddaa@nor75-18-82-241-238-155.fbx.proxad.net] has joined #launchpad-meeting === jamesh [n=james@203-59-20-109.dyn.iinet.net.au] has joined #launchpad-meeting [12:01] Still missing mpool and Steve === mpool [n=mbp@ozlabs.tip.net.au] has joined #launchpad-meeting === SteveA [n=steve@costello.z3u.com] has joined #launchpad-meeting [12:02] hi [12:03] Hello [12:03] Before the meeting starts [12:03] I will have a short week === irvin [n=ipp@ubuntu/member/irvin] has joined #launchpad-meeting [12:03] since I'm on leave starting thursday [12:03] I'll not be paying much attention [12:03] I'm in UI meetings here today [12:03] until the 28th [12:04] SteveA: there's one or two points where I think I'll need your attention, but okay [12:04] MEETING STARTS [12:04] == Agenda == [12:04] * roll call [12:04] * production status [12:04] * Smart server [12:04] * SFTP advertising [12:04] * vcs-import knits [12:04] * URTF [12:04] * Python import [12:04] * bzr lp:// [12:04] * important bugs according to bzr community [12:04] hi [12:04] * 1.0 targets, what, who when? [12:04] * critical bugs [12:04] * pending sysadmin tasks [12:04] * any other business [12:04] == Roll call == [12:04] ddaa is here [12:04] So is Steve [12:04] mpool said hi [12:04] so did lifeless [12:04] I'm here [12:04] I'm here [12:05] Wonderful. [12:05] == Production status == [12:05] Nothing new. https://launchpad.net/products/launchpad-bazaar/+bug/53825 [12:05] jamesh: so, will we get oops from script or do we need a different plan to deal with those errors? [12:05] last week you said you'll have a meeting with salgado about that [12:06] Mh, apparently not... [12:06] == Smart server == [12:06] mpool, spiv: status? [12:06] ddaa: I talked with matsubara about it last week. I don't think we'll have polished infrastructure for it soon, but I could set things up to get simple traceback OOPS reports [12:07] ddaa: getting SQL statement logs is a fair bit more work due to differences between webapp and zopeless environments [12:07] We paired a couple of days on it last week, and have scheduled another day on Wednesday. [12:07] ddaa: continuing to work with spiv on it; pairing on wednesday this week [12:08] spiv: mpool: any delivery estimate, for the non-integrated bit at least? [12:08] There are a fair few tests that make assumptions about the transport that need tweaking. [12:08] jamesh: basically, kiko is your client for that stuff, since he's the one I expect to find the error spam most annnoying first. [12:09] The plan is to get the last ~30 tests to pass, then start implementing actual smart operations on the new transport. [12:09] spiv: think you'll get it up and running on launchpad for 1.0? [12:10] I think there's probably only a day or so left on getting the tests working, but I'm not sure how quickly it'll progress after that. I am getting more fluent with the bzrlib code, though. [12:10] or is that a post 1.0 feature? [12:10] When's launchpad 1.0? [12:10] end of september === spiv hmms [12:12] spiv: you'll have time to answer that at the end of the meeting [12:12] moving on [12:12] I think it's possible, but I'm not going to commit to that right now. [12:12] == SFTP advertising == [12:12] jamesh posted a second draft at the end of the week. I intended to review it friday but got -ENOROUNDTUIT. [12:12] Action: ddaa to review jamesh second draft real soon. [12:12] spiv: last week you said you'd have a look, did you? [12:12] I did too. I haven't looked yet, I'll do that immediately after this meeting and reply. [12:12] spiv: two things for the smartserver: running it on the supermirror so that bzr users at large get good network performance with the SM. [12:12] I posted the second draft as a followup to the first thread [12:12] also, running it on devpad for launchpad use [12:12] Would be nice to get that blog posted at the end of the week. [12:14] SteveA: if we make the smart server post 1.0, it would certainly be nice to have run on devpad for a while, so it can mature, before putting it on launchpad. [12:14] == vcs-imports knits == [12:14] As far as I can tell, it's done. All vcs-imports branches are know knits. But that consumed all my round tuits at the end the week. [12:14] lifeless: mpool: I posted two patches to the bzr mailing list, that were needed for that. I would like to have them in rocketfuel shortly after I come back from leave (on the 28th). They are currently a divergence from rocketfuel in the production code. [12:15] lifeless: apologies for blowing up on you on #bzr saturday morning [12:16] Is that an action for anybody? [12:16] blow up at me ? [12:17] well, blaming you for turning that rollout into a 12 hours marathon [12:17] ddaa: I need to focus on this UI discussion. I'll read the email summary after this meeting. Please ping me directly on this channel if there is something I need to contribute to. [12:17] SteveA: ack [12:17] ta [12:18] Okay, moving on. I'd just like some quick feedback on those patch, although I'll probably won't be able to act on it before leave. [12:18] == URTF == [12:18] lifeless: jamesh: last week there was some talk about unblocking jamesh with urtf, how did that go? [12:18] URTF ? [12:18] Upstream Release Tarball Finder, you know... [12:18] previously known as Dyson [12:19] I'll look at implementing what was discussed this week [12:19] jamesh: where can we see the output of that discussion? [12:19] I was blocked last week due to sprints, rather than on input from other people [12:20] I'd be interested in knowing what's going on with that thing, since I spend so many hours stuffing the db with data for that. [12:20] s/spend/spent/ [12:20] its in the bug [12:20] ddaa: https://launchpad.net/products/launchpad/+bug/53698 [12:20] Thanks [12:20] == Python import == [12:20] Last week I was asked to look into the problem with the Python import, that manifest itself as a DNS failure at random points in the middle of the import. [12:20] Planned to do so friday. Did not. Bot noticed that other large SVN imports fail in the same way. That's probably our most serious cscvs bug ATM. [12:20] Action: ddaa will file bug about that failure, and maybe look at it before going to leave. [12:21] Though it's down the list of priorities, so would not like to promise looking at it soon. [12:22] == bzr lp:// == [12:22] sabdfl made some noises about that recently, which resulted in some activity. [12:22] mpool wrote a spec about the bzr-launchpad integration aspect, I will review it this week. [12:22] I put a braindump on the mailing list, but had no feedback. In short it boils down to: [12:22] * one attribute: Use ProductSeries.branch both for the vcs-imports branch and for the user specified branch. That leads to model constraints, restrictions for the user, and complication. [12:22] * two attributes: Use different attributes, say import_branch and explicit_trunk_branch. That essentially acknowledges that import != series, even though it's stuffed in the same table, and makes things much more flexible. [12:22] ddaa: yes, would appreciate that [12:23] mpool: it's one of the important things for this week, so you can make progress on that while I'm away. [12:23] what is the URL of the spec? [12:23] speaking of specs [12:23] https://launchpad.canonical.com/PrivateBranches is nearly reviewed [12:23] mpool and ddaa favor the "two attributes" approach. But that is a sensitive topic, so I would like consensus or word from sabdfl before putting out a proper spec. [12:23] ddaa: i'm not really qualified to give feedback on the db changes [12:23] who will be? [12:23] SteveA: your attention would be welcome about the db change issue, since it's a loaded issue involving sabdfl. [12:23] jamesh: https://launchpad.canonical.com/BranchIndirection [12:24] who else should review either the db or ui specs? [12:24] lifeless: talk about that end of meeting [12:24] mpool: stub gets to review all db changes [12:25] I'm not sure that mpt has much to say about the UI in that case. [12:25] me either [12:25] spiv, can you please read it [12:25] then let's JFDI [12:25] JFDI? [12:25] JustFuckingDoIt [12:25] ddaa: hi [12:25] mpool: ok [12:26] ddaa: what's the question? [12:26] SteveA: backlog, just after the == bzr lp:// == [12:26] discussion about changing the db schema to have two branch attributes in ProductSeries [12:27] Since I do not have my hands free to with that table as I wish, and since sabdfl seems to feel strongly about it (see SourceSourceRefactoring, last year) [12:28] the two attribute solution sounds like the simpler option. [12:28] jamesh: certainly, but it's acking that vcs-import != series, and for some reason that appears to be taboo. [12:29] No sound technical reason AFAICT [12:29] that's why I'm incomfortable about it [12:29] ddaa: really? [12:29] ddaa: you are uncomfortable because you got what you wanted [12:29] ? [12:29] I did not get that yet. [12:30] But if we have a consensus on using two attributes, great. [12:30] I'm happy. [12:30] SteveA: thank you [12:30] == important bugs according to bzr community == [12:30] mpool: when we had that phone call a couple of weeks ago, I suggested you get feedback from the bzr community about launchpad integration issues. [12:30] That would be useful data to have at the beginning of september, can you commit to delivering that? [12:31] ddaa: on consideration, i think the problem is that integration is just not very visible yet [12:31] Mh [12:31] Actually, I mean more than just integration of the client with the server. [12:31] so "what are the bugs in it" may not get good results [12:32] I also mean things like the UI to the branch registry, bug branches, sftp server, tec. [12:32] ddaa: what question are we trying to answer? [12:32] s/tec./etc./ [12:32] network problems [12:32] back [12:32] The question is: "what are the bugs or features that the bzr community as a group appears to desire the most" [12:33] ok [12:33] ddaa: on the issue you asked me about [12:33] As an help to prioritizing tasks. [12:33] ddaa: if we go for 2 table fields now, can it be collapsed into 1 later with a simple database update? [12:33] SteveA: not really [12:34] it's adding degrees of freedom to the system [12:34] ddaa: I'd think it would be pretty easy: "update productseries set user_branch = import_branch where user_branch is NULL" then remove the import_branch [12:35] jamesh: it's not enough [12:35] jamesh: let's talk about that after meeting [12:35] okay [12:35] ACTION: jamesh and ddaa to come agreement about db change for ProductSeries === irvin [n=ipp@ubuntu/member/irvin] has left #launchpad-meeting ["Leaving"] [12:36] ddaa: i think we should just decide [12:36] and then perhaps ask for suggestions [12:36] well, I'd certainly not let user whims be the only factor in deciding [12:36] i'll try to give you a list before you leave, and then we can continue [12:37] if we start showing that we're shipping features there, people will suggest more [12:37] but since I'm not going to decide on anything before september, asking the community might be a good use of the time until then. [12:37] why? [12:37] because I'll be only leave [12:37] oh, from this weekend for two weeks? [12:37] ok [12:37] and because I have my hands very full for this week w/o dowing planning [12:37] s/dowing/doing/ [12:38] moving on [12:38] btw it looks likely that someone new will be starting here in september, also working on lp bzr features [12:38] ddaa: ok, in that case i'll work on it but less urgently [12:38] and talk to spiv, steve, mark, robertc, etc [12:39] that's not what I consider the "bzr community", but okay [12:39] Let's move on. [12:39] == 1.0 targets == [12:39] I'm a bit unclear on what is everybody's target for Launchpad 1.0 (end of september). [12:39] For me: I think it was bzr transition. But I'm thinking of delaying the Arch support removal until we have the new hire. That would be a good exercice to get him acquainted with a lot of the code. [12:39] Everybody else: What are your Launchpad 1.0 targets? [12:39] i know it's not [12:39] hm, let's talk about that afterwards [12:39] My Launchpad 1.0 target is the smart server on the supermirror. [12:40] ACTION: ddaa mpool talk about gathering planning input [12:40] spiv: sounds like a challenge [12:40] Yeah. Achievable, I think. [12:41] mpool: bzr-roundtrip-svn is marked 1.0 too [12:41] The integration with the supermirror will be a couple of days effort, I think, not much. The main issue is having a smart server in the first place :) [12:41] what's the status of that, last we talked about it, you were about to have a call with mark about using jelmer's stuff without supporting it explicitly in launchpad IIRC. [12:42] spiv: +1 [12:42] lifeless: from your interest, I also gather that tracking-versions is 1.0, is it? [12:43] it's not marked as such in the spec tracker [12:43] its as soon as it can be done [12:43] I dont know what makes something a 1.0 feature [12:44] AFAIK, sabdfl makes things 1.0 features [12:44] so, then its probably not a 1.0 feature [12:44] but its still as soon as possible ;) [12:44] mpool: about bzr-roundtrip-svn? [12:45] ddaa: i had a call and i sent a mail reporting the result [12:45] i'm sure you were cc'd [12:45] mh? [12:45] did you not see it, or was it lost? [12:45] I do not remember seeing it [12:46] i remember writing it [12:46] ACTION: ddaa and mpool to sort out the lost mail issue [12:46] maybe there was a mail problem somewhere [12:46] no, it's worth announcing here [12:46] so the decision was that we will take no action on it [12:46] CANCEL THAT ACTION [12:46] it's not a priority for canonical [12:47] cool [12:47] jelmer can keep going with his work; we can support it and give advice on e.g. how to make id assignment better [12:47] but the main things for us to focus on are [12:47] - improving bzr performance [12:47] - getting the other import priorities [12:47] - exposing more feautres on launchpad [12:47] cool? [12:48] Sounds quite sensible a plan. [12:48] If you have not sent it to the ML, please do so. [12:48] :( [12:48] i hate losing mail [12:48] == Pending sysadmin tasks == [12:49] Nothing I guess. [12:49] Admins are perfect as usual those days. [12:49] == Critical bugs == [12:49] Skipping that [12:49] In short, everything is in progress [12:49] I have partly implemented fix for two of them. [12:49] Plan to get the code up for review this week. [12:49] That will be the bulk of my time. [12:50] == Any other business? == [12:51] yes [12:51] https://launchpad.canonical.com/PrivateBranches#preview [12:51] almost too late [12:51] I've replied to the reviews from spiv and ddaa [12:51] please go through them and either: [12:51] - delete the review question [you are happy] [12:51] - identify what you'd like me to do to make you happy on that point [12:51] Ok. [12:52] ACTION: spiv and ddaa to review PrivateBranches again [12:52] will try to do it this week [12:52] lifeless: is that very important? [12:53] its a blocker for using lp with lp [12:53] bit of a lynchpin [12:53] lynchin-r-us dear friend [12:53] so, its important, but long term. Dont panic on it, but dont defer indefinately. [12:53] Okay, but I'd rather get outstanding code finished before leaving. [12:54] At the latest, I'll review quickly when I'm back. [12:54] ddaa: ok, mail is sent [12:54] mpool: thank you [12:54] I think that's all sorted out. [12:55] MEETING CLOSED [12:55] Reviewers have a break before the next meeting. [12:55] I'll have a restbreak too. [12:55] mpool: how much longer are you around today? [12:57] i can be around for a while [12:57] mpool: I think we have one or two items to discuss post meeting [12:58] Let's have it at 1130 UTC === ddaa runs out [12:58] mpool: re. the client side of BranchIndirection, it sounds like it would need the same infrastructure as needed to handle HTTP redirects [12:59] jamesh: yes, it's essentially a different front-end to the same functionality [12:59] jamesh: yes, i think so - could you add some details on that? [12:59] and also on the server side === ddaa runs out really [01:00] mpool: e.g. another use for cross-scheme redirects would be if I set up my web server to redirect with "Location: bzr-smart-server://..." [01:01] right [01:01] and you could argue that rather than xmlrpc, we should really just do the query over http [01:01] and expect a redirection back [01:01] which would be a pretty cool thing to do for http://bazaar.launchpad.net/ -- redirect the user to smart server if their user-agent indicates they have a new enough bzr [01:06] yep === ddaa [n=ddaa@nor75-18-82-241-238-155.fbx.proxad.net] has joined #launchpad-meeting === jelmer [n=jelmer@a62-251-123-16.adsl.xs4all.nl] has joined #launchpad-meeting