[11:31] * gmb takes an early lunch [12:06] * danilos -> lunch as well === benji_ is now known as benji [13:05] danilos. my email replies about the bug notification stuff were OK by you? [13:05] gary_poster: Howdy. So, I'm still working on the devpad cronjob. Running into issues with buildbot since buildbot can't connect out. This might need a) Someone who knows more about buildbot than I to look at it or b) Same person as in (a) to answer questions when I have them. [13:05] gmb, buildbot? [13:06] buildout. [13:06] heh, ok [13:06] Two letters, dammit. [13:06] :-) [13:07] gmb, cheap hack might work. Thinking. Lemme get the branch to see what the buildout looks like these days [13:13] gary_poster: It's now at lp:~launchpad/lp2kanban/trunk [13:13] Rather than under ~yellow/+junk [13:13] gmb, yeah, yay :-) already got it there. almost ready with idea. [13:13] I actually have two [13:13] Cool\ [13:13] trying for the less annoying one :-P [13:15] gary_poster, oh, I haven't looked at the email at all :/ sorry, will do so now [13:15] gmb, ok I'll list both ideas. First one is to do what LP does. I consider this the more annoying approach, but I remember better how to do it since that's what we do now [13:16] For that one, in buildout.cfg, add this [13:16] (in [buildout] [13:16] ) [13:16] eggs-directory = eggs [13:16] download-cache = download-cache [13:16] install-from-cache = true [13:16] Then make the download-cache and eggs directories [13:17] and copy those up wholesale from local bits [13:17] That's kinda gross [13:17] Ahaha. [13:17] but we know it works [13:17] I like a good kludge. [13:17] :-) [13:17] The less annoying approach is this one [13:18] You upload the necessary eggs (as found in your download-cache, or whatever) to a directory on lp [13:18] ideally this would be (1) owned by a group [13:18] and (2) already a published directory [13:19] (like ~/public) [13:19] or ~/html or whatever we call it [13:19] then in buildout.cfg [13:19] you say index=http://...the address to that directory... [13:20] assuming that devpad can access itself over the net via the same path we can externally, that means we all use the same eggs for this [13:20] and it is nice and easy [13:20] Okay. [13:20] however, that might be wrong [13:20] I'll do the first to make sure that the script actually works (I'm wondering about devpad connecting out to LeanKitKanban) [13:20] And then try the second. [13:21] ok sounds good [13:21] makes sense [13:21] the docs for the second are as usual hard to make practical use of [13:21] but they are here [13:21] http://pypi.python.org/pypi/zc.buildout/1.5.2#finding-distributions [13:21] it may be that we have to do other random hacks to make the second work [13:21] which I won't bother with till later [13:22] Okiedoke. [13:23] (sadly, I used to like buildout! :-P You can decide for yourself if the sadness is that I no longer do, or that I did initially, because I'm not sure :-) ) [13:23] danilos, understood, np. [13:24] Heh. [13:30] bac benji danilos gmb, call in 3 (I must "prepare") [13:32] bac and benji sent their apologies [13:33] * gary_poster tries to build the new lp2kanban quickly... [13:53] gary_poster, emailed jtv, looked through your emails, they are all good, I'll just take a look at the race-cond bug description [13:54] cool danilos, thanks [13:56] gary_poster, btw, nobody complained or noticed the missing "subscribed by" titles [13:56] danilos, heh [13:56] I wouldn't have [15:22] gary_poster: Is there a good way to populate my local download-cache so that I can rsync it up to devpad? download-cache/dist is empty for my lp2kanban and the versions specificed in buildout.cfg are later than the ones in the LP download-cache, so using that doesn't work. [15:23] gmb, you tried rerunning buildout locally already, after making the same changes? [15:23] gary_poster: I have. Actually, let me try this in a fresh branch... [15:23] fresh shouldn't make a diff [15:24] gary_poster: No, but at least I have a better idea what's a result of buildbot and what's a result of me mucking about. [15:25] :-) if you get me a copy of the chanes, I can try running it locally [15:31] gary_poster: Oh, d'oh. PEBKAC. [15:31] gmb, cool :-) [15:31] If you set install-from-cache = true and there's no cache, it's not gonna work. [15:34] :-) [15:39] gary_poster: So, computer says "no" as far as connecting out to LKK. Going to talk to the LOSAs now. The error is spectacularly non-obvious, so I'm going to file a bug for that and fix it later. http://pastebin.ubuntu.com/630416/ [15:43] hi gary_poster, i've made some good progress but i'm a bit stuck atm on what i suspect is some missing zcml fu. got a sec? [15:44] gmb, ok, progress at least [15:44] bac, ok, sure [15:44] Yes indeed. [15:44] gmb: firewall issue? [15:44] bac: Yus. [15:44] yeah [15:45] gary_poster: so two odd things. one, i've specified a plural name for IProcessorFamily as 'processor_families' but it shows up in lplib as 'processor_familys'. [15:45] heh [15:45] gary_poster: and the more important, i get error trying to get a canonical_url for IProcessor. that's the one where i'm missing zcml i'm quite sure [15:46] gary_poster: the diff is at http://pastebin.ubuntu.com/630418/ [15:46] * gary_poster wonders if benji would have a better chance of helping quickly with the plural thing. [15:46] looking [15:46] gary_poster: these classes have never been exposed to the UI or API before. only way to tweak them is through SQL [15:49] bac, did you expect the canonical_url for IProcessor to work, or you didn't expect it because you didn't know what to add? I don't obviously see any current zcml that would expose it [15:49] I mean, in the diff [15:50] gary_poster: no, i'm not surprised it doesn't work. [15:50] ok cool [15:50] gary_poster: i'm looking for direction on the proper way to hook it up [15:50] gotcha. Looking. [15:50] I can look at the plural thing if needed. [15:52] bac, I have no idea, but am doing greppy things. My current lead is "canonical.launchpad.webapp.interfaces.ICanonicalUrlData" [15:52] benji, do you have any idea? [15:52] I mean about the plural thing [15:52] not off the top of my head, but I'd be glad to look into it [15:53] benji, k. ...yeah, why don't you look into the plural thing bac mentioned while I look into the canonical url thing [15:53] k [15:54] bac: I assume this is the same branch as yesterday [15:54] benji: yes [15:55] benji: let me re-push [15:55] done [15:56] k [15:59] bac, are you already familiar with browser:url? I wasn't but it appears to be common, and appears to be the way LP expects these things to be done [16:00] gary_poster: do i need to expose a top level stepto for +process too [16:00] gary_poster: yeah, i was just looking at that [16:00] gary_poster: something like: [16:00] for="lp.soyuz.interfaces.processor.IProcessor" [16:00] path_expression="string:+processor/${id}" [16:00] parent_utility="canonical.launchpad.webapp.interfaces.ILaunchpadRoot"/> [16:00] bac, yeah, sounds like a good start to me [16:00] so, in addition to +processor-family i need to add +processor [16:01] bac, it sounds like you are on the right track to me, so I'm gonna do CHR, but feel free to ping again [16:02] gary_poster: but i wonder if that is too inflationary wrt to top-level entry points? i suspect francis may have strong opinions [16:03] bac, does hwdb have a parent url? [16:03] bac: it appears that you have to provide both a plural and singular in order for either of them to be used [16:03] like /hwdb/... [16:03] +hwdb, yes [16:03] bac, I have no idea, but would that be reasonable to use then? [16:03] but it is an IApplication [16:03] (if you experimentally verify that as the case, I will file a bug about generating an error if you only provide one) [16:03] why is that bad bac? [16:04] gary_poster: oh, you mean hang the processor off +hwdb or use it as an example? [16:06] bac, hang it off +hwdb, possibly as +hwdb/processor/ or something. That may be horrible--I know nothing about hwdb so I don't know what their URL patterns looks like, and where conflicts might arise. +hwdb/+processor? [16:06] gary_poster: actually a processor belongs to only one ProcessorFamily [16:06] so i can do +processor_family/processor/${id} [16:06] bac, +1 [16:07] assuming that means that this will not conflict with a processor_family named "processor," or that we consciously don't care and have handled the potential OOPS if it does [16:07] bac ^^ [16:08] gary_poster: will keep in mind [16:08] cool [16:10] benji: we have plenty of examples of export_as_webservice_entry with only a plural_name. i assume they work... [16:11] bac: maybe I'm reading the code wrong. How do I reproduce the problem? [16:11] benji: with my branch, run the server and then connect to it using lplib is one way to see [16:12] dir(lp) shows the top level collections and you'll see process_familys [16:12] or, i'm sure it is visible in wadl file too [16:12] benji: do you use lp-shell? man, it rocks [16:13] gmb, could you take a look at https://answers.launchpad.net/launchpad/+question/161972 and see if you have any wisdom to impart? [16:13] nope, I don't think I've heard of it [16:13] gary_poster: Sure [16:13] thank you [16:13] * gmb sees HWDB, gets scared. [16:13] benji: http://pastebin.ubuntu.com/630428/plain/ [16:13] gary_poster: I think this is one that abel would be best for. [16:13] I'll ping him. [16:14] thanks gmb [16:14] benji: http://bradcrittenden.net/post/1173745688/interacting-with-launchpadlib-using-lp-shell [16:20] yay danilos :-) [16:20] yaay :) [16:21] gary_poster, and jtv was quick to react, 29mins ago marked as qa-ok on https://bugs.launchpad.net/launchpad/+bug/394645 :) [16:21] <_mup_> Bug #394645: IDistroSeries.getQueueItems deprecated by IPackageUploadSet.getPackageUploads

< https://launchpad.net/bugs/394645 > [16:21] cool :-) [16:26] gary_poster, while waiting for henning to become available for a preimp, I've closed one critical bug (OOPS) which doesn't seem to happen anymore (it was happening every day until March 1st then suddenly stopped), and looked into another timeout bug and identified the core problem and potential solution (a solution that seems to work for that case, but since it's "remove an index that's in the way", I need to ensure it's not harmful [16:26] elsewhere) [16:27] gary_poster, do you perhaps want these bugs on the kanban board? [16:27] gary_poster, (one was basically half an hour with matsubara querying the OOPS DB, other some playtime on lpmain_staging DB) [16:28] danilos, sounds great. I haven't been putting the "close by declaring dead" bugs on the board myself (I had one yesterday too), but I'm happy to see them. If you intend to continue on the timeout bug, yeah, def put it on the board. [16:28] gary_poster, yeah, I do for the timeout bug, so I'll put it up, thanks [16:29] cool [16:30] I am off now, tomorrow I am OCR and a help contact :) [16:31] bac: I'm having trouble reproducing the problem: http://pastebin.ubuntu.com/630436/ [16:32] benji: so you have nothing wrt processor family [16:32] bac: not that I can see [16:32] I've verified that I'm running a branch with your most recent branch merged in and there are no other servers running, and I see the requests in the access log of the expected server [16:32] benji: i see the same thing if i connect using 'devel' version. if i use '1.0' version i see processor_familys [16:33] oddities abound [16:33] hmm, let me look at that [16:36] bac: yeah, I don't see anything about families with lp = Launchpad.login_with('test', 'dev', version='1.0') or lp = Launchpad.login_with('test', 'dev', version='devel') [16:36] benji: that's very odd. you did a 'make clean build'? [16:37] bac: ooh, I think you diagnosed it [16:37] bac, fwiw, mrevell has not replied to https://support.one.ubuntu.com/Ticket/Display.html?id=2081 in a long time so maybe worth some commercial love when you get a chance. [16:37] gary_poster: ok [16:37] benji: or is the wadl being cached? [16:37] i think it may be! [16:38] that has bitten me in the past [16:41] bac: yep, the WADL was being cached by launchpadlib, once I removed that I began seeing the correctly named attribute: "processor_families" [16:42] benji: me too. sorry for the goose chase [16:42] I needed some excersize. [16:49] gary_poster: when do you fly out to dublin? what's your routing? i'm on DL via ATL [16:50] bac, same here. I leave 6:30 PM Sat [16:50] return flight was weirder [16:50] Delta leaving 10:30 AM Sat, through Kennedy [16:51] Don't remember why [16:53] yeah, i'm not returning until the following saturday. [16:55] gary_poster: it's odd on that commercial ticket matthew seems to have led the guy on about private hosting. now i've got to be the bad guy. [16:55] bac, uh-oh :-( [16:55] maybe that's why he didn't reply :-P [16:57] :) [17:49] lunch/getting last-minute birthday presents for younger son [19:43] bac, gmb, did you already countersign your allhands stuff? I meant to ask this morning on the call [19:43] gary_poster: nope. will do now. [19:43] cool [19:45] gary_poster: i have countersigned, validated, and confirmed [19:45] cool thanks bac. I'll take you off the card [19:45] is that it for a while, i hope? [19:46] here's my favorite part: The objective sheet can't be anymore updated, unless it is reopened by your manager. [19:46] bac, y, afaik [19:46] umm, bad english [19:46] yeah :-/ [19:46] i filed a bug on that, oh, i guess a year ago [19:46] heh [19:46] yeah [19:47] improvements on that app never happen afaik [20:00] It looks like it's written in Domino (Lotus Notes). [20:00] * benji prepares to do CHR. [20:10] * gary_poster was only so-so on CHR today, and fears benji's wrath [20:27] gary_poster: heh; no worries, there were lots of projects to review, but now that I have the disable_projects.py script, it's quite a bit easier [20:27] benji, where is it? [20:27] I did have one almost coherent question that I had to answer. [20:27] heh [20:27] lp:lp-dev-utils [20:28] I thought it might be something like that. Thanks [20:49] benji: were there still lots of projects to review? i did a bunch yesterday [20:50] bac: yeah, but they were all registered in the last 24 hours; aparently it's a popular site ;) [20:50] gary_poster, benji: is there a trick in the API to setting a property to a list of things? i want to do: archive.enabled_restricted_families = [arm] but i get a NoBoundRepresentationError [20:51] I don't know of one. As a matter of fact, I don't know much about how collections work. [20:51] you might try archive.enabled_restricted_families[:] = [arm] [20:54] bac, I'm afraid I don't know either :-/ [20:54] TypeError: 'Collection' object does not support item assignment [20:54] benji: ^^ [20:55] [rhetorically] well, how are you supposed to mutate it then? [20:55] yeah, there must be a way [20:57] from looking at the code, it looks like Collections are read-only; still looking [20:57] let's see what francis says... [20:57] Good idea [20:57] I was looking at https://help.launchpad.net/API/launchpadlib and getting nowhere [20:59] well, it seems it is undoable [20:59] not doable [20:59] impossible [21:03] who would want to change a list, really? [21:43] The list we should check frequently: https://bugs.launchpad.net/launchpad-project/+bugs?field.tag=escalated [21:44] bugs danilos should mentor us on: bug 734765 and bug 775691 [21:44] <_mup_> Bug #734765: Show linked upstream projects to the translatable templates in the +templates page

< https://launchpad.net/bugs/734765 > [21:44] <_mup_> Bug #775691: Empty translations on one side do not get translated by the other side

< https://launchpad.net/bugs/775691 >