=== thumper [n=tim@166.179.24.250] has joined #launchpad-meeting === thumper [n=tim@222.152.34.146] has joined #launchpad-meeting === thumper [n=tim@166-179-19-162.jamamobile.co.nz] has joined #launchpad-meeting === _thumper_ [n=tim@166-179-28-124.jamamobile.co.nz] has joined #launchpad-meeting === ddaa [n=ddaa@nor75-18-82-241-238-155.fbx.proxad.net] has joined #launchpad-meeting [10:00] MEETING STARTS [10:00] here :-) [10:00] == Agenda == [10:00] Next meeting Monday 18 December, 09:00-09:45 UTC. [10:00] * roll call [10:00] * production status [10:00] * status reports [10:00] == Roll call == [10:00] No apology. [10:00] lifeless: ping [10:00] jamesh: ping [10:00] lifeless is on leave [10:00] SteveA: ping [10:00] pong [10:00] right [10:01] good turnout tonight === spiv [n=andrew@218-214-66-203.people.net.au] has joined #launchpad-meeting [10:01] Hello. [10:02] hi [10:02] spiv: welcome [10:02] so, lifeless on vac [10:02] Shouldn't this be on the internal irc server? [10:02] poolie late [10:02] hu? [10:02] I guess it would make more sense [10:02] spiv: it has always been on the public one before, but I don't mind [10:02] spiv: it is notionally public. [10:03] yeah [10:03] jamesh: oh, ok, I assumed otherwise. [10:03] spiv: the #launchpad-meeting channel did not move [10:03] In that case, carry on! [10:03] anything secret squirrel and we can go sideways === ddaa carries on [10:03] == Production status == [10:03] New rollouts or production problems. [10:03] That I know, at least. [10:04] Unless anybody has something to report here [10:04] not from me [10:04] like, e.g. the deployment of the supermirror-smart-server :) [10:04] I'll carry on [10:04] == Status reports == [10:04] * everybody: read https://blueprints.launchpad.net/distros/ubuntu/+spec/code-review [10:04] Did everybody read this spec? [10:04] readit === spiv looks [10:05] No, I haven't yet. [10:05] I've decided to grab some of the issues hilighted in it [10:06] spiv: jamesh: please make sure to read this, it's quite important for the near future of the launchpad-bazaar stuff [10:06] I looked over it a bit last week. [10:06] bugs 58889, 71303 and 73975 [10:06] * spiv: supermirror-smart-server. [10:06] * ddaa, _thumper_: look at spec-branches. [10:06] I did not look at the spec-branch work. [10:06] ddaa: note name change [10:06] I wonder if we need a search capability for branch listing [10:06] in addition to status filtering [10:07] thumper: name change? [10:07] jamesh: searching on what? [10:07] ddaa: from _thumper_ [10:07] thumper: branch names and descriptions [10:07] jamesh: that would certainly be good [10:07] sounds like a plan [10:07] but not right now [10:07] braindump it [10:07] even without a wiki page, just to catch the idea [10:08] jamesh: want to look at thumper's work on branch filtering and post a braindump to the ML on how to add fti search? [10:08] No news from me, as I was at a conference last week, but https://launchpad.canonical.com/SupermirrorSmartServerPlan now exists (as requested at a previous meeting) [10:08] jamesh: if you don't I will, more karma to me [10:09] ddaa: okay. I had a very brief look at the branch already. I'll have a more detailed look. [10:09] ddaa: jamesh has commented already... [10:09] I just suggested having a look so the branch searching thing could be thought out in this context. [10:09] I was more commenting on issues that had already been brought up [10:10] spiv: are you going to have time to focus on it this week? [10:10] oh, ok [10:10] * ddaa: pyrex. [10:10] ddaa: yes, I think so. [10:10] Got a quick review from spiv, did not have the time to take any action. Actually, did not have the time to act on any review this week. [10:11] spiv: looking forward to having some good news to tell in the next launchpad meeting :) [10:11] ddaa: me too! :) [10:11] at least, some more details in the SupermirrorSmartServerPlan, it's a bit sketchy now [10:12] but better if it's more concrete stuff like "wsgi integration tested at spiv's", you see what I mean :) [10:12] * ddaa: progress on motu code-review [10:12] * poolie: bzr-lp features. [10:12] * jamesh: branch browsing, can help? [10:13] ddaa: still finishing off the bug import stuff. Can probably help later on this week. [10:13] on motu code-review, commented on the spec, did not get a lot of feedback. Grilled dholbach to get his requirement on branch filtering. Then thumper started implementing it. [10:13] :) [10:14] Also had a round of feedback on the email notification side with thumper. Will try to push the spec forward this week. [10:14] jamesh: I'd like if you could tell me when you get started, and if you need anything. [10:14] Not that I think you're going to need anything, but it's nice to propose :) [10:15] poolie is not here today, so no bzr-lp update [10:15] I should talk with poolie about what he thinks it should look like. [10:15] jamesh: the emails? [10:16] the branch browser, I think [10:16] thumper: branch browser [10:16] jamesh: I think we agreed on "just webserve" [10:16] ddaa, jamesh: initially, adding features as we want them if they are not there [10:16] as a stopgap, we could even run it on the bare id-based filesystem and keep it private [10:17] then add the mapping to the virtual filesystem [10:17] ddaa: last meeting he mentioned having a talk about it this week [10:17] ddaa: also suggested by poolie, he is very much in favour of just getting something up quick [10:17] which is why I mentioned it. [10:17] hi [10:17] jamesh: I think we're all in agreement on this. Get something up quick even if it's ugly. [10:17] although if ugly, keep private [10:18] initially [10:18] remember "pretty is a feature" [10:18] I think, if it's got the virtual filesystem, it would be pretty enough for the public. [10:18] If people want prettier, they would be welcome to hack webserve. [10:19] not having a _working_ web UI to showcase is bad for bzr [10:19] and geoffredo's site is unusably slow [10:20] So, that finishes the agenda. [10:20] == Any other business? == [10:20] SteveA: hello, you're late [10:20] SteveA: getting access to vostok for me [10:20] can we swing that? [10:21] I might not be able to work all day. The heating is off in my appt building, and my fingers are getting cold numb. [10:21] yes [10:21] I just read the scrollback [10:21] ddaa: is there a warm internet cafe near you? [10:21] not that I know [10:21] or a shop that sells portable heaters? [10:21] ddaa: why is the heating off? [10:22] thumper: no idea [10:22] should be back today though [10:22] ddaa: not even Starbucks? === thumper ducks [10:22] thumper: this is Paris, there are like 3 starbucks in the entire town. [10:22] ddaa: give it time [10:22] there are zero starbucks in Perth [10:23] http://www.world66.com/europe/france/paris/cybercafes [10:23] spiv: have you found out anything that would update the supermirror SS plan? === ddaa knew he had some other business to discuss [10:25] SteveA: not yet. [10:25] spiv: finding out whether apache has the hook you need sounds like it needs a small experiment [10:25] so describe the experiment and its result in the wiki page [10:26] (I wouldn't expect the docs to say this explicitly, and I've found apache docs a bit patchy in describing the interaction of different directives) [10:26] Yeah, I expect it to involve a hour or two of mucking about. [10:27] SteveA: re portable heaters, my appt is too small and crowded to get yet another piece of equipment [10:28] SteveA: you read the mail about "software you can get using bzr"? [10:28] I was wondering about the title [10:28] isn't it more about bazaar activity? [10:28] showing what is recent? or explicitly targetting "software you can get with bzr"? [10:29] ddaa: no [10:29] it's a feature request from sabfdl [10:29] ddaa: what's the subject of the email? [10:29] I spoke with mark about this on the phone at the end of last week [10:29] he specifically wants a page listing product names [10:29] Subject: Braindump: Software you can get using bzr [10:30] SteveA: ok [10:30] ddaa: I don't have that email [10:30] Date: Fri, 08 Dec 2006 21:10:35 +0100 [10:30] launchpad list [10:32] for reasons unknown, I don't have it, and I don't recall having seen it [10:32] SteveA: forwarded to you now [10:33] https://lists.ubuntu.com/mailman/private/launchpad/2006-December/012603.html [10:33] ta [10:33] ddaa: ok, reading it [10:34] Importd does not currently set Productseries.datelastsynced, so we do [10:34] not know when was the last successful import. This can be fixed quickly [10:34] by a small patch to importd. But this would be a good opportunity to do [10:34] more work, and start separating the status reporting from the job runner. [10:34] [10:34] this kind of thing worries me [10:34] I'll try to explain why [10:34] ddaa: weird. I didn't get that email either === thumper thinks perhaps it looks like spam [10:35] you're appear to be saying "I could make this quick and simple fix, but I won't because I'm worried if I do that, I'll feel less need to do the larger refactoring" [10:35] meaning I can hack it in less than an afternoon (tests included), but I'd prefer to start tackling the larger job of separating the job runner (now buildbot) from the status reporting to go incrementally in the directory of improtd-ng. The Plan is to have job-runner -> job-reporter -> cscvs. [10:36] ddaa: as far as I can see from reading your mail, that's the only feature we're missing to make the page Mark asked for work [10:36] is that correct? [10:36] no [10:36] "If the last sync was successful, but was not mirrored yet, we do not know how old is the currently published code." [10:37] ok [10:37] so that too [10:37] I don't understand what the distinction is between "last sync" and "mirrored" [10:37] so, making the import system actually use one field we already have in the database [10:37] can you enlighten me? [10:37] given those two, yes, we can do the page Mark asked for [10:37] and adding a new field, and making the import system do that [10:38] ddaa: how long do you think doing both those things would take? [10:38] thumper: sync == code converted from CVS/Subversion to Bazaar [10:38] thumper: syncs don't happen directly on the mirror, there's an intermediate holding place. [10:38] thumper: importd published stuff to an internal http server, then the branch puller puts this data in the published location [10:38] thumper: mirror == code mirrored to http://bazaar.launchpad.net [10:38] ok, ta [10:39] SteveA: All included (coding, review, deployment) it should be about 8 hours of work for me [10:39] ok [10:39] then there the db patch issue, but it should be cherrypickable [10:40] and how long do you think "tackling the larger job of separating the job runner from the status reporting to go incrementally in the directory of improtd-ng" [10:40] will take? [10:40] not sure... double it at least [10:40] double what? [10:40] 8 more hours [10:41] it's hard to say, it's not complicated work, but it's a bit of new infrastructure, so there's the usual overhead [10:41] new script, new test cases, etc. [10:41] ok [10:41] here's my frustration with what you wrote [10:42] 3 mins to go [10:42] you wrote that email in response to a request to write a particular feature [10:42] and, for very tenuous reasons, you're trying to combine writing that feature [10:42] which has a clear bounds [10:42] with doing some more open-ended refactoring [10:42] sure, I do not have a clear reason for this [10:43] then WHY did you do it? [10:43] to annoy you? :) [10:43] well, because I'm just trying to push this refactoring any way I can [10:43] ever heard the phrase "career limiting move" ? [10:43] frustrated of not getting around to doing it [10:43] we're in a distributed company [10:43] communication has an overhead [10:44] and a larger overhead than in an "all in one office" company [10:44] and I don't like it when you make communication muddy to further a separate agenda than I've asked you about [10:44] please think upon this [10:45] next time, for example, send me a separate email saying how important you think the refactoring is [10:45] but keep the other email to the point [10:45] I think you know this issue. I understand it's not smart doing it. [10:45] so: stop doing it immedately [10:45] that's all. [10:46] Anything else you are unhappy about the recent communication spec work I've been doing recently? [10:46] non [10:46] other than that, the email you wrote about this "bzr products listed on a page" is clear [10:46] I've been doing a lot of it, so it feels a bit frustrating that the only feedback I get from you is about this single sentence. [10:46] I didn't see that email until today [10:46] Oh yes other business: [10:47] cscvs has been released, in case people did not notice [10:47] cool. that must feel good, as you've been wanting to do that for a while [10:47] the stuff is still lacking some code from hct, I plan to clean that up soon (it's part of the larger "remove dep on pybaz" problem) [10:47] and the cmd line is still quite broken [10:48] so for the time being I want the release to be low-key [10:48] if people express an interest in the code, direct them to it [10:48] ddaa: where was the release announced? [10:48] It was not announced. [10:48] not surprising that it passed me by then... [10:48] I want to wait until we have ironed out some process to get the code in [10:48] I mean code contribution [10:49] and until the most blatant breakage in the cmd line is fixed [10:49] then we can announce it [10:49] would it be worth asking the marketing guys for input on a release announcement? [10:49] I expliticitly want this release to stay discreet for now [10:49] the tool sucks as a standalone conversion tool as it is [10:49] perhaps time it with the 1.0 bzr release? [10:49] there is a low, but steady, demand for access to this code [10:50] I already notified sivang, and he'll talk to janimo about it [10:50] okay. It wouldn't hurt to see if they have any suggestions for the announcement before hand. [10:51] so, when we have a patch submission process, and when the most blatant brokeness is fixed, we can announce it. ATM it's just "a dump of broken corporate code" [10:51] so, feel free to tell people about it, but I'd rather avoid publicity [10:52] the product on launchpad is launchpad-cscvs by the way [10:53] ok [10:53] BTW, meeting closed [10:53] yay, sleep [10:53] SteveA: I did more than just this email in the past couple of weeks [10:53] spiv: I'd like to get some pings from you during the week about progress looking into the smartserver + supermirror [10:54] specs about deleted branches, email notification, requirements on branch status filtering [10:54] spiv: because I'm interested in seeing how it will come out [10:54] SteveA: sure. [10:54] thanks [10:54] actually, I spent pretty much all my time doing help-text, cscvs release, and spec/guidance work. [10:55] after the infrastructure for this "page of products" [10:55] what would you like to do next? [10:55] after? === ddaa calms down the passive-agressive [10:57] continue spec work around launcphad-bazaar, clear out backlog of review replies, spend a little time fixing cscvs, spend a little time improving bzr-svn, and finish the pyrex work. [10:58] also, keep spending about one hour a day on small UI improvements to launchpad (currently, spent on help-text) [10:59] I'd like to have a call with you later so we can talk over the relative priorities of these things [11:00] also, Kiko and I will be introducing a new concept into Launchpad work, starting for each team within Launchpad as soon as the 1.0 work for that team in complete [11:00] that is, Fridays will be "work on whatever Launchpad stuff you think needs it" days [11:00] mh [11:01] too bad there's only one friday a week :) [11:01] we're going to start planning Launchpad work based on four day weeks of scheduled work [11:01] for completing specs etc. [11:02] and on Fridays, you get to set the priorities and activities for that day [11:02] formal announcement on Thursday [11:02] sounds like a good idea [11:02] so we can get to do all those important things that we never get around to [11:04] workrave in 2 mins [12:04] http://svn.collab.net/repos/svn/trunk/notes/svnsync.txt [12:07] lifeless: dude, enjoy your holiday! :) [12:08] What are you doing in a work channel? :) [12:13] " Q: How does svnsync deal with parts of the master repository that I'm not authorized to read? " [12:13] mh, that's another cscvs bug in the waiting === carlos [n=carlos@195.Red-88-0-156.dynamicIP.rima-tde.net] has joined #launchpad-meeting === danilos [n=danilo@adsl-233-130.eunet.yu] has joined #launchpad-meeting === kiko [n=kiko@200-171-140-32.dsl.telesp.net.br] has joined #launchpad-meeting [04:11] hi [04:11] good morning guys [04:11] how are you [04:11] I am fine, thanks [04:11] fine too, thanks [04:11] and you? [04:12] pretty good [04:12] so [04:12] carlos, what happened last monday? [04:13] I sent TranslationReview review answer and the modification to allow anonymous browsing was implemented and sent to you [04:13] for review [04:13] I saw that [04:13] but you had said that you would get everything reviewed and landed by monday [04:13] and you said that in fact you were going to come back in the afternoon [04:13] so danilo and I waited for you [04:14] hmm [04:14] I was, but I didn't connect... [04:14] was I supposed to connect to IRC? [04:14] you were supposed to at least tell us the status of your work by end of day monday [04:14] sorry, I was not aware of that... [04:14] huh? [04:15] you said very explicitly that you would be back, and we had spoken on friday about us ensuring that we got things done by monday [04:15] I said that I will work and handle the merge, but I was not thinking on connecting to IRC, sorry... [04:15] I sent you a brief report, with copy to danilo [04:15] but the merge wasn't done either, was it? [04:15] with the diff I sent to yoiu [04:16] you [04:16] no, because the review required more work [04:16] and you didn't tell us that either [04:16] so... hard for us to understand what's going on in valencia. [04:17] I would like that to sort of thing to be solved [04:17] ok, sorry about that. I was not as communicative as I should be... [04:18] so how are things this week [04:18] what's snagged on review? [04:19] kiko: about IRC connection, I usually don't connect when I'm supposed to be on vacations, to be focus on the thing that needs to be done, even on vacations. I will try to connect at least to #canonical next time [04:19] you've got my weekly reports, and the last thing I did last week was fighting with zipfile module [04:20] basically, I've done the exports, need to test them thoroughoutly, and add a test (as discussed on Friday) [04:20] carlos, I don't care. if you say you will merge something and then disappear, I will get upset. [04:20] so yes, I expect better communication, and this isn't the first time. [04:20] about the review: I think the review from Salgado should come today and then merged, about TranslationReview, I need to move pagetests from using database code to use BeautifulSoup. I had a meeting today with Bjorn about it, he helped me to figure the best way to check translation form in that way [04:21] carlos, hmmm. I see. what will beautifulsoup help there? [04:21] ah, you are querying for results? [04:21] and bjorn wants you to test the page output? [04:21] kiko: the fact that active translations are not inside a textarea or input fields, we cannot query them from pagetests, it's just plain html text [04:22] well, I test the content, but based on what we have in the database [04:22] right, you'd need to parse with BS indeed [04:22] yeah, yeah. [04:22] right [04:22] how hard is that change? [04:22] seems much more easy than I expected [04:22] iterating over html tags [04:22] okay, good news [04:22] will you get it done by tomorrrow? [04:23] yes [04:23] I hope to get the final review from Bjorn today too [04:23] ok. [04:23] but the only remaining issue I think there will be is the BeautifulSoup chnage [04:23] ok, good news [04:24] danilos, did you get over the zipfile issue? [04:24] kiko: well, worked around it as I initially planned: recreated new zipfile in memory instead of trying to modify existing one [04:25] right. [04:25] that's the better plan I think too [04:25] at the moment, I'm trying to get export to work properly [04:25] danilos, what's stuck there? [04:25] well, my code is simply not getting called when I run the export cronscript [04:26] so, I probably missed something, and need to look into it better [04:26] the thing is that regular po export is using a caching mechanism, which I am not at the moment [04:26] so, there's probably something missing there [04:26] danilos: forget about caching at this time [04:27] it should work without caching support [04:27] carlos: right [04:28] basically, this was the last thing I did on Friday, and haven't had a chance to look at it properly [04:28] the thing is that all the classes are there, but it simply finishes really quickly without exporting anything [04:28] danilos, add some tracing code [04:28] to see what its doing [04:28] or pdb it [04:28] kiko: right, I'll do that [04:29] danilos, how's it gone today? [04:29] kiko: my plan was to actually find out where it's getting first (using some tracing code), and see where is it failing [04:29] right, that's pretty sane to me. [04:29] oh, I've started really late today, so I've got like 6 more hours of work [04:29] basically, haven't done much [04:30] but this is the main thing for the rest of the day, anyway [04:30] okay [04:30] so, I'll let you know how it goes, and will ping you and others on #launchpad-code if I get stuck [04:30] carlos, what else came up in your review with carlos? [04:30] yeah, danilos, please do -- I'm at your service [04:30] ? [04:30] err [04:30] with bjorn [04:31] my left wrist hurts today, fucks me over [04:31] kiko, talk with kiko about it :P [04:31] well [04:31] most of the changes were related with tests [04:31] and the way to document it [04:33] carlos, was he happy with the general design? [04:33] and also, I did a mistake and with the hurry to deliver the first review, I forgot to review the diff and notice that I miss that I added some kind of documentation that Bjorn told me to change later in the review [04:33] but that's already solved (or should, I still need his reply) [04:33] okay. [04:33] the only complain about the design was related with a behaviour I did on javascript [04:34] carlos, danilos: what's the plan for tomorrow? [04:34] carlos, what JS? the copy stuff? [04:34] well, I'd rather talk about tommorow later in the day: I want to see how will these things go today first [04:34] no, the way to handle the 'New suggestion' flag [04:34] and when you can type in new info to the textarea [04:34] if I manage to get out of it today, it's up for review tommorow [04:35] I just followed his advice and implemented it the same way Malone handles it [04:35] but that's an "if" (no ETA ;) [04:35] danilos, okay. it sounds like you will still lack tests though, after this. [04:35] carlos, to auto-select the radiobutton? [04:35] right, but we've already discussed how I am going to do them [04:36] kiko: checkbuttons [04:36] and I'll test such programmed result using local copy of firefox [04:36] kiko: lack tests? [04:36] kiko: I don't think so [04:36] last review was just a review to improve current tests [04:36] danilos, okay, but likely your reviewer will want tests that are more unit-like [04:36] carlos: that was about me [04:36] no code changes (other than renames) [04:36] oh [04:36] kiko: ah, so you think more of doctests vs. pagetests, or? [04:37] carlos, okay. do you need help with the BS changes? [04:37] ok ;-) [04:37] danilos, well, you'll need doctests for API you add/modify [04:37] kiko: or simply that I should test the new classes I've added as well? [04:37] pagetests for pages you change [04:37]