[00:00] Bert_2: It's the table at the top of the bug page that shows status/importance/etc. [00:00] wgrant: mmm, on submitting a filled in bug my instance now errors on that [00:01] What's the error? [00:01] Could be another missing celebrity breaking a permission check [00:02] here's the full oopsinfo: http://pastebin.com/dzEKFBYd [00:03] it seems to be the case I cannot view bugdetails, only the table of bugs [00:03] the oops I gave you is shown on displaying a specific bug [00:03] Ah, I think I know what that is. [00:03] Give me a sec [00:04] take your time ;) [00:10] Bert_2: On https://launchpad.dev/ubuntu/+edit (or the equivalent for your domain), select 'Bugs in this project are tracked in Launchpad' [00:12] wgrant: did that, and it fixed it :O [00:12] so, why was it not working and why is it now ? [00:13] If you click on the expander icon at the left of the task table, you'll see that the Distribution widget defaults to Ubuntu [00:13] That's a bit of a hack -- it's hardcoded [00:14] The widget is configured to only show distros that track their bugs in Launchpad, so the default value of Ubuntu was illegal. [00:14] the Target:Distribution:Ubuntu thing you mean ? [00:14] Yeah [00:14] oh, I see [00:15] thanks a lot man, I'm pretty sure I couldn't have found that by myself :p [00:17] Yeah, it was pretty obscure, but I was able to guess it right away as I've had the misfortune of running into that code before. [00:18] I'm fixing the script to add those essential celebrities and set that Ubuntu bit [00:18] wgrant: awesome, that'll help a lot of people out in the future [00:19] now as we'll be using bugs, answers and bluepints I'll let you know if I run into any more issues so the script is fully up to date ;) [00:19] Yup [00:19] awesome :D [00:33] wgrant: thanks again for your help, now I'm off to bed, bye [00:33] Bert_2: Night. [00:47] wgrant: http://pastebin.ubuntu.com/1109290/ [00:48] wgrant: I wouldn't expect subscriber to be there for the new tests -- so I'm guessing I have something backward [00:49] StevenK: Will look soon [01:10] wgrant: Is it soon yet? :-) [01:11] StevenK: it is "Too Sooooon!" [01:21] StevenK: + subscribers = BugSubscriberSet().union( [01:21] + self.structural_subscribers, [01:21] + self.all_pillar_owners_without_bug_supervisors, [01:21] + self.all_assignees) [01:21] StevenK: None of those seem to be filtered [01:37] wgrant: I thought self.structural_subscribers would be? Although the structuralsubscription model code is messy at best [01:38] StevenK: Right, but the others aren't... [01:40] self.all_pillar_owners_without_bug_supervisors probably does not require filtering [01:41] self.all_assignees is currently causing me grief [01:41] It requires filtering [01:41] To /dev/null [01:41] That's the thing we said we were going to delete [01:41] On the basis that it doesn't make sense [01:41] Ah [01:41] And things that don't make any senses probably shouldn't exists [01:41] * StevenK deletes it [01:41] -plurals [01:41] * wgrant stabs the branch scanner [01:41] StevenK: Separate branch [01:41] :-( [01:42] Probably the same branch as the bug supervisor default sub removal [01:42] Which is *after* this branch. Bah. [01:42] Nah [01:42] Can be before this branch [01:42] Bug supervisor default structsub removal, that is [01:42] Not the default private_bugs direct subscription [02:27] wgrant: http://pastebin.ubuntu.com/1109408/ Ignoring the printf debugging, it seems like forbidden_recipients_filter and forbidden_subscribers are backwards [02:29] StevenK: What suggests that? [02:30] wgrant: My printf debugging does [02:30] Specific examples? [02:31] Ah [02:31] Well, yes, forbidden_subscribers is backwards [02:31] So I'd suggest making it not backwards [02:31] So forbidden_recipients_filter is backward? [02:31] forbidden_recipients_filter returns Storm conditions that restrict the returned subscriptions to those which have access [02:34] Oh, damn it, that is still keyed on BugSubscription [02:34] I need both BugSubscription and StructuralSubscription [02:38] So no wonder it returns [] :-( [03:12] wgrant: http://pastebin.ubuntu.com/1109443/ is the StructuralSubscription query [03:15] StevenK: Does it work? [03:15] lifeless: Is bzrlib's Graph.find_difference meant to be slow? [03:16] lifeless: It's taking upwards of 5 minutes to calculate for one of my LP branches. It was about 5000 mainline revs behind devel, and it chokes on the merge I did yesterday. [03:21] it uses _find_border_ancestors [03:21] which notes: This will scale with the number of uncommon ancestors. [03:21] wgrant: what is it doing? Can you lsprof it ? [03:27] lifeless: My smaller test cases are 5x faster after a pack. Trying the original case again now [03:30] 1039155 0 2.3077 2.3077 [03:30] blink [03:31] That takes 11s unprofiled, and it's b.repository.get_graph().find_difference('william.grant@canonical.com-20120717002424-zaimmc2rnew8s0qp', 'launchpad@pqm.canonical.com-20120724215633-t9eizm9m7wlb8nwo') -- the former rev being a merge from devel a week ago, so it's only a week of revisions. [03:32] After packing the full 14-month diff takes 150s. [03:34] I'm not quite sure why calculating a 277-node graph difference requires more than a million list extends... [03:35] I don't, at this remove, remember which apis are fast and which are not. [03:35] I *think* this is meant to be fast [03:45] wgrant: No, which is why I'm pasting it [03:45] And 476s for the full diff on an unpacked repo. [03:45] StevenK: What doesn't it do? [03:45] StevenK: Note that that is finding forbidden subscribers [03:46] wgrant: I thought Not(Or(*get_bug_privacy...) would filter out forbiddens [03:48] StevenK: get_bug_privacy_filter filters bugs based on privacy [03:49] You use it when you want bugs that are visible to someone [03:49] Read it :) [03:49] The first clause it returns is information_type.is_in(PUBLIC_INFORMATION_TYPES) [03:50] wgrant: Right, so it looks like the filtering in the depths of structural subscriptions was backward [04:01] IntegrityError: duplicate key value violates unique constraint "accesspolicy__product__type__key" [04:01] DETAIL: Key (product, type)=(27, 4) already exists. [04:01] :-( [04:48] wgrant, wallyworld: https://code.launchpad.net/~stevenk/launchpad/one-linkcve/+merge/116582 [04:50] StevenK: so you no longer need @export_operation_as('linkCVE') [04:51] Oh, duh [04:52] other than that looks ok i think [04:52] wallyworld: And it's -15 now :-) [04:53] wallyworld: That change has pushed [04:53] done [04:54] wallyworld: Thanks [04:54] np [05:13] wgrant: Do you have a suggestion how to squeeze filtering into BugSubsciptionInfo.duplicate_subscriptions ? I've been thinking about using filter_forbidden_subscriptions, but the multiple dupes thing is tripping me up. [07:52] good morning === frankban changed the topic of #launchpad-dev to: http://dev.launchpad.net/ | On call reviewer: frankban* (rvba) | Firefighting: - | Critical bugs: 4.0*10^2 [08:16] has the font rendering on Launchpad changed recently? [08:16] jml: yes I think huw as working on that [08:17] dont know if it landed though [08:17] czajkowski: as in, "it's a bug and huw is fixing it", or as in, "yes, and it's a work in progress", or as in, "yes. tremble at our newly attained beauty?" [08:18] certainly looks nicer to me :) [08:19] he was working on it [08:19] it's not a bug [08:19] and I think the code was landed and rolled out [08:20] so yes praise our new beauty! [08:21] czajkowski: :) === almaisan-away is now known as al-maisan [09:06] stub: The sampledata changes are not noise. There are two new product columns that I added last week, and one removed from bugsummary. [09:07] But they were all changes made by patches, so nothing to review. [09:25] wgrant: cool. [10:49] morning [10:49] gmb, ping [10:50] Ergo^: as I said yesterday, already, he's offline for 2 weeks. [10:50] yes, jsut noticed he showed online now ;-) [10:51] yes a lot of us run screens and while we're connected we may not always be here [10:51] czajkowski: to be fair, some launchpadders seem to have trouble with the concept of "holidays" :P [10:51] jelmer: yes you and wgrant do, gmb doesn't :) [10:52] seems my xchat client shows "away" statuses, with delay [10:54] mpt: I have a bug for you! [10:54] payback! [10:54] czajkowski, mmmm, delicious [10:55] mpt: mind if I show it to you [10:55] much easier than me attempting to explain [10:55] .c [10:55] I am waiting with bated breath === matsubara is now known as matsubara-afk === al-maisan is now known as almaisan-away [11:53] wallyworld, rick_h_: I think your JS changes may have conflicted, as I have a failure on ec2. [11:53] actual = ["Failure in /var/launchpad/test/lib/lp/bugs/javascript/tests/test_duplicates.html.lp.bugs.duplicates_tests.test_duplicate_form_submission_success: Unexpected error: Result of expression 'uri' [undefined] is not an object."] [11:54] hmm. let me check [11:54] That's the only failure [11:56] buildbot seems happy so far but i think the yui tests are run near the end [11:56] Yeah, this was in the last 5 minutes of the test suite [11:56] bollocks [11:56] I have another instance 15 minutes behind that hasn't failed yet [11:56] Ah [11:56] Now it has [11:56] In the last two minutes [11:57] i'll have to pull devel and take a look [11:57] but i have something i need to attend to first [11:59] wgrant: looking, yea bet it's something with the lp.client changes [11:59] I renamed lp_original_uri to just uri looking at wallyworld's change to see if something there didn't agree since tests passed on my end. [12:06] i didn't make any lp client changes [12:06] wallyworld: right I did, and wonder if you used it in any way that broke [12:06] I'm pulling and updating, will run the tests and see [12:06] ok, thanks [12:07] i used a couple of 'standard' methods [12:07] right, nothing jumped out at me when I reviewed it yesterday, but will look [12:09] http://www.happyprog.com/tdgotchi/?repost=true [12:14] wallyworld: wgrant one word fix, going to just submit straight to pqm to keep from killing buildbot all day if that's cool [12:14] rick_h_: go for it, thanks! [12:15] rick_h_: the current build will fail to you will need to force it when it happens [12:15] wallyworld: will do, thanks for the reminder. Will keep an eye on it [12:15] np, thanks for fixing === abentley changed the topic of #launchpad-dev to: http://dev.launchpad.net/ | On call reviewer: frankban* (rvba), abentley | Firefighting: - | Critical bugs: 4.0*10^2 [13:17] deryck: ping, got a quick sec? [13:17] rick_h_, sure. what's up? [13:17] deryck: helping popey help someone with an issue using ubuntu-bug and sent me a screenshot: http://uploads.mitechie.com/lp/lp_bug_screenshot.png anything look strange that would toss an error you know of off the top of your head? [13:18] deryck: wondering if () in the summary or something? [13:19] rick_h_: we give rather usless error help don't we [13:19] czajkowski: :) could be a bit better I suppose [13:19] rick_h_, no, it shouldn't choke over the parens. [13:20] rick_h_, there's some required field or something not displayed. I wonder if something in extra options was set incorrectly somehow? [13:20] rick_h_, or some field with bad data. [13:22] deryck: ok, I'll see if I can get some more info and maybe talk him into submitting it manually to make sure that works [13:22] rick_h_, or maybe he already filed a bug against the token and it's no longer valid? The random number thing in the url is what links to the apport filed data. Maybe something there. [13:22] deryck: yea, this was an ubuntu-bug ... filing [13:25] rick_h_, change that message to Error: Success, that will teach them [13:25] deryck: ok cool, looks like a bad tag name: rc-6.0.0-0ubuntu5~test1 [13:33] oh dang, standup time. [13:33] rick_h_, abentley, adeuring -- coming, sorry. forgot time. [13:34] deryck: rgr [13:36] adeuring, https://plus.google.com/hangouts/_/6e08b518c64d571c487a6e9da883b81733756c2d?authuser=0&hl=en [14:18] adeuring: Why does test_run_missing_ready_does_not_return_results call a script instead of running the code in-process? === Guest69114 is now known as Ursinha [14:19] abentley: I can't remember exactly... Might be related to config dteails. [14:35] abentley: can you peek at https://code.launchpad.net/~rharding/launchpad/no_more_tmp/+merge/116663 when you get a sec, short one [14:37] rick_h_: You're creating a directory named ${buildout:yui-directory}/yui-${:yui_version}.zip ? [14:38] abentley: yea, because I copy out to yui-${:yui_version} [14:38] it was either that or create a tmp directory in there to extract/copy out of [14:38] and that just seemed like creating a dir for no good reason [14:39] Does this avoid creating a directory? It sounds like it would just be a different name if you created a tmp directory. [14:40] abentley: right, this avoids creating the directory and extracting to it vs just extracting straight to yui-version.zip as a dir [14:40] it's extracting the yui-version.zip from source-deps into build/js/ and then pulling only the build (no tests, docs, etc) from that zip into build/js/yui-3.3.0 [14:41] rick_h_: I find extensions in directory names surprising. Would ${buildout:yui-directory}/yui-${:yui_version}-extracted or ${buildout:yui-directory}/yui-${:yui_version}-tmp or something work for you? [14:41] abentley: sure thing, all the same to me. [14:41] will update, thanks [14:41] rick_h_: r=me. [14:48] mrevell: do you have time to talk [14:49] sinzui, I think I'm just about to going into a meeting . I'm in the new London office. [14:50] okay. If you have time today. ping me for a short talk about my anxieties. [14:52] * czajkowski hugs sinzui [14:52] sinzui: read the warthogs mail [14:52] and did laugh [14:52] worst holiday from hell? [14:54] My in-laws invited me to the same beach for another week of near-death-experiences at the end of August [14:54] sinzui: will you be going ? [14:54] I am uncertain. My wife and kids will [14:57] sinzui: avoid the beach bugs. [14:57] czajkowski: : this is the non-published link of what I did yesterday. I gave up trying to get the flashes out of the video since it is a known openshot tansition bug: http://youtu.be/Vq6VsKuFY_o [14:58] jcsackett: since Anne did not select the house, there is no guarantee there will be internet. [14:58] sinzui: cheers [14:58] I guess I could take the car and drive 2.5 hours to get to that Chinese restaurant [14:58] sinzui: or... you could tkae a holiday [14:59] dear purple squad you need to learn how to take off your holidays properly [14:59] czajkowski: in addition to his many allergies, sinzui has a fatal allergy to not working on launchpad. [14:59] jcsackett: so does wgrant [15:00] My meditation is TDD. Zen and the Art of Test Driven Development. I don't have to work on Lp, but I need to do something. [15:00] sinzui, Hey, I have 15 minutes now or longer than that in 90 minutes. Which works best for you? [15:00] sinzui: arduinos ? [15:01] oh, yeah. we should get him started on that. i need to know more people doing arduino hacking. [15:02] sinzui: can i grab you after you're done talking to mrevell? [15:02] mrevell: now would be fine === almaisan-away is now known as al-maisan [15:02] sinzui, I'll grab a room [15:02] mrevell: https://plus.google.com/hangouts/_/410f6d1a701063146f8bb2bc46e26dfa04cd2539?authuser=0&hl=en when you are ready [15:05] rick_h_, hey, not sure if it's inviting you: https://plus.google.com/hangouts/_/73a8173804e27710a1abf877e26aefe3848f70d3?authuser=0&hl=en [15:17] frankban: Could you please review https://code.launchpad.net/~abentley/launchpad/enable-tests/+merge/116678 ? [15:17] sure abentley [15:19] mrevell: http://youtu.be/Vq6VsKuFY_o [15:33] rick_h_, abentley, adeuring -- I'm playing with G+ events to get a named hangout for our standup, so forgive if you get event invite spammed now. [15:33] deryck: np [15:40] sinzui: do you have a minute to talk about https://bugs.launchpad.net/launchpad/+bug/918745 ? [15:40] <_mup_> Bug #918745: PersonMergeJob sends emails to people who do not care

< https://launchpad.net/bugs/918745 > [15:40] I do [15:41] sinzui: I don't understand your comment. Why would the target of the merge care if the job failed, unless the target of the merge requested the job? [15:42] abentley: this happens when the user deletes the team, they do not care about it, and there is nothing they can do to fix it. [15:43] abentley: we can send the email to the person who requested the deletion. It will delay us from fixing the data in the team [15:44] abentley: we fixed the most common cause last week. We do not allow people to delete teams with private branches because they cannot be given to any other team [15:44] sinzui: If this error is not meant to be sent to users, I think we should let the normal oops mechanism handle it. [15:45] abentley: that is fine. the user will just create a question for us to look up and then fix the data as we do when ~registry gets the email [15:46] Your change is good because it tells the user the delete failed [15:49] sinzui: If we want to send email to a celebrity, that could make sense. But is the target guaranteed to be ~registry, or could it be a normal user/team? [15:52] abentley: The email went tot the target of the merge. We get the oopses where ~registry is missing the proper BVP. Other teams get them when the change is a merge. Remember that delete is always a merge into ~registry because there are undeletable artrefacts. [15:53] sinzui: Okay, I think I'll go ahead with the change. [15:54] okay [15:54] abentley: you want to send the email to the user that scheduled the job? [15:55] sinzui: Yes. [15:55] fab [15:57] rvba: I don't know what value is best, either. In most cases, we don't hit the sleep at all. [15:57] rvba: So maybe it doesn't matter. [15:58] abentley: ok then. [15:58] rvba: But in cases where we race, I guess it doesn't hurt to do a fraction of a second. The processor has nothing better to do :-) [15:58] abentley: that was my point :) [15:59] rvba: Okay, I'll change the sleep to (0.1) and the range to 100. Cool? [15:59] abentley: cool. [16:00] rvba: (while I didn't write sleep(1) it in *this* branch, I wrote it yesterday in another branch :-) === al-maisan is now known as almaisan-away === matsubara-afk is now known as matsubara [16:35] Hi, blueprintslists in my instance of launchpad come in pages of 5, launchpad.net has pages of 20, I seem to be unable to find any mentioning of a limit in the templates, does anyone know where this number is set ? === deryck is now known as deryck[lunch] [17:03] development launchpad defaults to 5 items per list. This allows us to see batching work with our small sampledata size [17:06] sinzui: and how can it be changed, cause I can't really seem to find a place to change the 5 to 20 or more [17:06] it is defined in the schema. [17:06] schema-lazr.conf ? [17:07] It can be redefined in configs/development/launchpad-lazr.conf [17:08] what is the option called, cause I don't see any option with a 5 in schema-lazr.conf that have to do with it, I think ? === frankban changed the topic of #launchpad-dev to: http://dev.launchpad.net/ | On call reviewer: abentley | Firefighting: - | Critical bugs: 4.0*10^2 [17:09] I don't know the name of the key, but it is defined in schema-lazr.conf. Search for 5 [17:09] I just did [17:10] the only ones with just five are announcement_batch_size and bugnotification_interval [17:10] and consecutive_failure_limit [17:10] those don't seem to have to do with blueprints :S [17:11] As I said, this is about batching, not applications [17:11] Change the value in configs/development/launchpad-lazr.conf as I suggested [17:12] so announcement_batch_size is the one ? [17:14] sinzui: I'm not using development [17:14] I have my own configdir [17:15] Bert_2: This channel is for launchpad development. There no channel for support, the code is unsupported [17:17] sinzui: I have supplied multiple additions to the bootstrap-db-from-scratch script, I'm not just here to get support [17:22] got it, thx anyway sinzui === deryck[lunch] is now known as deryck === salgado is now known as salgado-lunch [19:03] abentley: celery test failed on buildbot can you dbl check please? [19:03] rick_h_: sure. [19:08] rick_h_: The fix in r15684 was broken. It's corrected in r15688. [19:09] is there an issue with LP code hosting - I am trying to set trunk to my branch and it is rejecting it [19:09] abentley ok, will force another round [19:09] the branch is lp:~jonobacon/ubuntu-accomplishments-system/validation-service [19:10] rick_h_: thanks for letting me know, though. [19:10] abentley: started up aonther build then, look slike 15688 was merged so here we go again wheee :) === salgado-lunch is now known as salgado [21:42] lifeless: do you have a few minutes to discuss/listen to a story about the monsters in the forest of sharing [21:49] I do [21:49] sinzui: skype? G+ ? [21:49] skype I think [21:50] signed in .... now [21:59] abentley: I'm rolling back the celery changes, as in 15688 both tests have failed [22:35] sinzui: we finished the standup. all good, no-one is blocked etc [22:42] hello [22:42] sinzui: can you talk to cjohnston he has some questions over your recent blog post about project maintainers and private buts [22:42] *bugs [22:42] man I am knackered [22:42] * cjohnston thinks its bed time ;-) [22:43] and that may go for me too [22:43] lol [22:59] wgrant: StevenK: look at this [22:59] https://bugs.launchpad.net/launchpad/+bug/1029093 [22:59] <_mup_> Bug #1029093: Can't change the information type of a bug that has dupes filed on deactivated project < https://launchpad.net/bugs/1029093 > [23:01] cjohnston: I suggest asking the question :) [23:02] I was waiting to see if he responded (or atleast now that I see that he is here) ;-) [23:03] sinzui: I'm a little confused about marking bugs private.. From the help pages for LP, it says that projects can set all bugs by default to be private. I tried implementing this setting and wasn't allowed to because I don't have a commercial licence.. if nothing else, the text is confusing [23:04] https://help.launchpad.net/Bugs/PrivateBugs#Private_bugs [23:07] cjohnston: Yeah, all the documentation around privacy is pretty terrible. [23:07] We're almost done completely reworking it (including replacing the default private bugs checkbox), after which the docs can be rewritten [23:08] wgrant: quick pm? [23:08] cjohnston: Sure [23:09] wgrant: http://pastebin.ubuntu.com/1111033/ is the dupe query [23:13] wgrant: cjohnston is an unsual area, the project is for canonical but he's a contractor [23:13] czajkowski: Yeah [23:14] cjohnston: Might be best to discuss this in #launchpad-ops on irc.c.c [23:15] indeed