[00:06] lifeless: Yes, it's on LPS, but that normally gets ignored :) [00:06] gawd I hope not [00:07] so whats special [00:07] also when is good for you [00:07] It needs buildd-manager to be shut down before the update, then process-upload run, then proceed as normal. [00:08] why? [00:08] The directory name format has changed. [00:09] And accepting and testing both seems more trouble than having a slightly special one-off upgrade. [00:12] maxb: https://code.launchpad.net/~launchpad-committers/lpreview-body/packaging [00:12] is done [00:12] the other needs the regression in LP fixed [00:13] thanks [00:13] maxb: can you follow up on the list so folk know [01:21] wgrant: so whats the filename format change about [01:21] lifeless: Removing BuildFarmJob.id. [01:21] It was used to identify temporary upload directories. [01:21] what did the filename encode, what does it encode, why does it encode it [01:22] So, buildd-manager takes an upload and dumps it into a directory for process-upload.py --builds to handle. [01:22] It used BuildFarmJob.id to identify the relevant build. [01:22] Now it uses the job type and BPB/SPRB ID. [01:22] is that in a metadata file or the dirname? [01:23] dirname. [01:23] (yes, ew) [02:01] Yippie, build fixed! [02:01] Project db-devel build #519: FIXED in 5 hr 12 min: https://lpci.wedontsleep.org/job/db-devel/519/ [02:03] wgrant: when would you like to talk? [02:12] lifeless: Now's good for me. [02:15] lifeless: You still prefer Skype? [02:24] cooking lunch, will ping in a bit [02:25] Sure. [03:06] wgrant: ping [03:07] Project windmill-db-devel build #240: STILL FAILING in 1 hr 9 min: https://lpci.wedontsleep.org/job/windmill-db-devel/240/ [03:12] Project windmill-devel build #38: STILL FAILING in 1 hr 10 min: https://lpci.wedontsleep.org/job/windmill-devel/38/ [03:16] lifeless: Hi. [03:16] hi [03:16] I'm ready. [03:16] voip or skype is bestest [03:16] I'm on Skype, I've not got SIP set up. === mwhudson_ is now known as mwhudson [05:37] poolie: jetlag? [05:52] hmm, code free day :( [06:09] wgrant: https://bugs.launchpad.net/launchpad/+bug/777794 - please triage as you file :) [06:09] <_mup_> Bug #777794: Bug subscription AJAX popups are too bold < https://launchpad.net/bugs/777794 > [06:10] 777789 too [06:10] and 777786 [06:10] lifeless: I can't triage those. [06:10] why not? [06:10] Hm, I meant to file a low bug about that [06:11] Because I don't know what importance they are. [06:11] wgrant: take a good guess [06:11] Given that there is meant to be a UI polish mode at the end of each feature, they should probably be High. [06:11] But I don't know. [06:11] the feature squads are not watching the tracker. [06:11] If you set the importance and such like when you file a bug, we shouldn't say "Your bug will be triaged soon" [06:11] Perhaps ... [06:11] StevenK: That's for external people. [06:11] StevenK: that prose is just the post-filing-notice, its not programmatic [06:12] Right. [06:13] wgrant: please triage them, even if its just to high + an email to gary to say 'these need your hand to really assign importance'. [06:13] wgrant: you can certainly assess them in the context of the BugTriage wiki page === Ursinha is now known as Ursinha-afk [06:15] wgrant: a nontrivial reason for doing is so that end user filed bugs which are not triaged are obvious on the portlets on https://bugs.launchpad.net/launchpad-project/+bugs [06:15] the [non]zero distinction is easier for everyone to observe than a floating figure [06:20] hi lifeless [06:20] a bit [06:37] DF's up to 62GB... this is going to take a little while. [06:38] As in DF's DB? [06:38] Yes. [06:38] Hmm. Might need to clean some stuff out. [06:38] lifeless: \l+ on staging? [06:38] its fat too [06:38] it has the garbo running nowadays [06:39] qastaging is 275GB fwiw [06:40] wgrant: ok, I've triaged your bugs... next time I'll make em all opinion :P [06:45] ohh nice [06:45] +patches gone from oopses [06:48] lifeless: Thanks. [07:25] Hmm, the missing structural subscription listing a bit unfortunate. [07:33] Project windmill-db-devel build #241: STILL FAILING in 1 hr 5 min: https://lpci.wedontsleep.org/job/windmill-db-devel/241/ [08:14] Project windmill-db-devel build #242: STILL FAILING in 41 min: https://lpci.wedontsleep.org/job/windmill-db-devel/242/ [08:59] good morning [09:04] Hello! [09:06] hi mrevell [09:06] lifeless, did people ever consider changing the ec2 image to run tests on a tmpfs? [09:06] it doesn't seem to be using its whole cpu [09:07] good morning [09:12] hi bigjools [09:12] poolie: I don't think our ec2 stack has been all that optimised [09:12] bigjools: missed you this morning [09:12] lifeless: yeah sorry [09:12] no worries [09:12] was knackered [09:12] perhaps we could talk in ~30 ? [09:13] lifeless: I have a standup, but in ~45 would work [09:14] ok === almaisan-away is now known as al-maisan [09:18] Project windmill-devel build #39: STILL FAILING in 1 hr 4 min: https://lpci.wedontsleep.org/job/windmill-devel/39/ === adeuring changed the topic of #launchpad-dev to: https://dev.launchpad.net/ | On call reviewer: adeuring | https://code.launchpad.net/launchpad-project/+activereviews === al-maisan is now known as almaisan-away === almaisan-away is now known as al-maisan [10:38] hm so just moving bits of /var to ram doesn't immediately fail [10:39] and is doing little io [10:39] still lots of idle cpu, but i think that's just lack of parallelism [10:55] bigjools: Bug #778408 may amuse you. [10:55] <_mup_> Bug #778408: Deleting an unpublished publication behaves unexpectedly

< https://launchpad.net/bugs/778408 > [10:57] (Side-effects of amusement may include uncontrollable sobbing and a sense of hopelessness.) [11:00] hehehe [11:04] rvba: Have you seen the stable->db-devel merge conflict? It seems to be your stuff. [11:04] So you're probably best to resolve it. [11:05] wgrant: I just saw that. [11:06] wgrant: can you guide me with how I can fix that? [11:07] wgrant: I understand the problem of course but I'm unsure about the procedure to solve it. [11:07] wgrant: colour me unsurprised [11:07] rvba: Grab stable and db-devel locally, merge stable into db-devel, fix the conflict as you would normally. Push the branch up somewhere, then pqm-submit it to db-devel. [11:07] wgrant: we need yet more code to stop people from being stupid [11:08] rvba: if you are feeling forgiving you could ask for a review, but I never do and haven't broken anything yet. [11:08] wgrant: thanks. I'll do that. [11:24] bigjools: Do you have a list of stuff to kill from mawson? Its free space will probably not last the weekend. [11:24] wgrant: yeah I was looking at that [11:24] We could just start wiping out the archive, I guess. [11:24] yes [11:24] that archive is too fat [11:24] Easy enough to bring back in a few hours if we need to. [11:24] (sorry mizuho) [11:24] It is fat, but was useful for testing a few things. [11:25] But that's all done now. [11:25] can be restored easy [11:25] given time :) [11:25] Yep. [11:25] * wgrant deletes. [11:25] Mass deletion + restore == molten mawson? [11:25] I bet it's hot it mawson's cabinet :D [11:25] s/it/in/ [11:26] wgrant: it'll all be invalid once the db is restored anyway [11:27] local librarian is 5.2G, purging [11:27] Not really. This is all from frozen series. [11:27] maverick + natty [11:27] not all [11:27] not sure they weren't frozen when we last restored [11:28] maverick was released, and natty was basically identical to maverick. [11:28] So it was mostly just maverick's final packages plus a few more. [11:28] Anyway, deleting now. [11:28] Very slowly. [11:29] the disk stepper motor is going to fall off its track [11:33] Project windmill-devel build #40: STILL FAILING in 1 hr 4 min: https://lpci.wedontsleep.org/job/windmill-devel/40/ === henninge is now known as henninge-lunch [11:55] wgrant: would you mind taking a look at this for a quick sanity check? https://code.launchpad.net/~rvb/launchpad/db-devel-fix-merge-failure [11:57] rvba: The diff is large, but as long as the tests you've changed pass it should be fine. [11:58] wgrant: yeah, the tests in the modified file pass. [11:59] * rvba bzr pqm-submit -m "Merge stable." [12:01] rvba: You'll need a --submit-branch [12:01] "fix conflict with stable" would be more accurate [12:01] To target db-devel. [12:01] and that [12:01] wgrant: oops ... too late. [12:02] It should reject it. [12:02] rvba: hi; its my weekend now, but wgrant may be able to answer your perf tuning thing if you're lucky [12:03] rvba: otherwise I'll shoot you some thoughts on mondayish [12:03] Where's that? [12:03] wgrant: I'l forward [12:03] lifeless: thanks a lot. Have a good weekend! [12:03] I have no email client at the moment, but I'll get one in a sec. [12:06] rvba: So, what's the issue with it being a list? [12:06] rvba: checkCopy is already pretty fast, except for one code path. [12:06] wgrant: since it's not a ResultSet there is not way to try to eager load related objects ... right? [12:07] s/not/no/ [12:07] rvba: You can't add a preiter hook to it, no. But that's just a handy place to do it... you can do the same thing manually afterwards. [12:07] What are you looking to eager-load? [12:09] (the slow bit is when _checkArchiveConflicts finds some destination_archive_conflicts) [12:09] I was just *thinking* of eager-loading related things since I saw the the kind of db queries issued looked like related objects being loaded. [12:09] but I might be wrong. [12:10] I you say it's pretty quick already, it might not be worth the bother then. [12:10] It's already pretty good. Most copies check within half a second, except for _checkArchiveConflicts. [12:10] So yeah, not worth spending time on that yet. [12:10] s/I/If [12:11] ok, so I guess the tests to make sure the number of queries does not increase is enough for now. [12:11] Yup. [12:11] That would be useful. === al-maisan is now known as almaisan-away [13:47] adeuring: Can you take a look at https://code.launchpad.net/~gmb/launchpad/bug-772609/+merge/60176 for me please? [13:47] gmb: sure [13:48] Thanks [13:48] Project windmill-db-devel build #243: STILL FAILING in 1 hr 4 min: https://lpci.wedontsleep.org/job/windmill-db-devel/243/ [13:58] Morning, all. [14:15] hi deryck [14:15] hi jtv === almaisan-away is now known as al-maisan [14:22] gmb: overall, nice work. but could you remove the "#import pdb; pdb.set_trace()" line in test_bugs_views.py? [14:23] yay, code reviews :-) [14:23] gmb: and you disabled the get_notified_subscribers() call in bugs/subscribers/bug.py [14:24] gmb: there is an XXX from bac about the call which has no bug number [14:24] gmb: also, I am not sure if we can really simply drop the call. [14:26] deryck: could you please run these queries on staging: https://pastebin.canonical.com/47200/ ? [14:26] adeuring, indeed I can. [14:29] Project windmill-db-devel build #244: STILL FAILING in 40 min: https://lpci.wedontsleep.org/job/windmill-db-devel/244/ [14:41] benji, I'm triaging bug 778323 as high as I assume it's related to y'all's current work. Is story-better-bug-notification still the best tag? [14:41] <_mup_> Bug #778323: On the subscription page, the help popup button opens up a 404 page. < https://launchpad.net/bugs/778323 > [14:41] deryck: yep, that's the current tag [14:42] benji, cool. done. Thanks [14:44] I don't think anyone understood the implications of dropping ~launchpad-bugs mailing list. [14:46] danilo, henninge: I wonder if those links in the "untranslated" etc. numbers on the overview pages disappeared because of the Prague work. [14:54] jtv, I don't know, I don't have the time to investigate right now [14:54] Same here. :( [14:55] jtv, I filed a bug so when I switch to maintenance I'll take a look [14:55] OK [14:55] jtv, it's tagged "regression", which makes it "Critical" === Ursinha-afk is now known as Ursinha [14:56] I guess it doesn't absolutely require a Rosetta veteran. [14:56] Hi Ursinha [14:56] jtv, indeed [14:56] I wonder why I'm suddenly being showered in bugmail. [15:01] Hi, gmb. Could I get you to review my GCC credentials branch for me? It's simple, but I want to make sure I haven't missed something I don't know about. [15:01] deryck: Sure thing. Diff me. [15:02] gmb, lp branch is: https://code.launchpad.net/~deryck/launchpad/add-gcc-keys-schema-lazr-745794/+merge/60186 [15:02] * gmb looks [15:02] Oo, complex! [15:02] heh [15:03] That looks fine. [15:03] gmb, and I did a matching lp-production-configs one first: https://code.launchpad.net/~deryck/lp-production-configs/add-gcc-bugzilla-creds/+merge/59956 [15:03] gmb, so I understand land the lp branch, nodowntime rollout, then the lp-production-configs one, yes? [15:03] deryck: I believe so. I haven't had to do one of these since before there were nodowntime rollouts. [15:04] But that sounds sane. [15:04] gmb, ok, cool. yeah, this is my first config change believe it or not. [15:04] Ah, always fun. [15:08] deryck: Land the branch, land the configs, nodowntime rollout. [15:08] No need for a rollout in the middle. [15:08] wgrant, ah, ok. thanks. [15:09] wgrant, and no need for any special rollout requirements, i.e. the usual nodowntime will cover everything I need? [15:09] deryck: That's right. [15:09] wgrant, awesome sauce. thanks, again. [15:09] We grab the requested rev of stable and the latest lp-p-c rev. [15:09] gotcha [15:34] Project windmill-devel build #41: STILL FAILING in 1 hr 5 min: https://lpci.wedontsleep.org/job/windmill-devel/41/ [15:36] gmb: did you see my questions about your mp? [15:37] adeuring: Oops, missed them. Let me look... [15:37] bac: are you around? [15:37] Ahaha [15:37] Thanks for catchign that pdb. === al-maisan is now known as almaisan-away [15:39] Oh. I don't remember removing that... [15:40] gmb: np. but can we reall drop the get_also_notified_subscribers() call? [15:40] Ah, I see. That's from bac's work and I've not spotted it... [15:40] adeuring: Hmm. Lemme check. [15:41] adeuring: I honestly don't know. [15:41] adeuring: hi, I've got a small branch in the queue, no rush. TIA :) [15:41] adeuring So, I'll put it back and run EC2 and see what happens. [15:41] bigjools: I'll look [15:41] gmb: ok, sounds good [15:42] Ta [15:43] sinzui: would you like to mumble this morning? [15:51] hello lp devs [15:51] I _love_ this new email management story [15:51] I get this: [15:51] Launchpad Bug Contacts subscription: (unnamed)You do not receive emails from this subscription [15:51] any idea what the (unnamed) bit is? should I file a bug? [15:52] beuno Did you give the subscription a name? === Ursinha-afk is now known as Ursinha === salgado is now known as salgado-lunch [15:52] beuno: Did you just create the subscription or was it an old one? [15:52] huwshimi, no, not even sure where (or why!) to do that [15:52] huwshimi, old [15:52] allenap: i've made changes to the branch you looked at yesterday, if you could take another look. [15:53] beuno: I don't know heaps about it, but I'm guessing you might be able to give it a name somwhere [15:54] huwshimi, why would I want to name something so efimeral? [15:54] beuno: Are you around somewhere and have a minute? I wouldn't mind looking at the UI of what you have. [15:55] Anyone in ~launchpad can see the same thing at https://launchpad.net/launchpad/+subscriptions [15:55] huwshimi, I am! I'm in the canonical sumit room, and would love to meet you. I have another hour of sessions, but then I'm free. [15:56] beuno, if you name it, bug mail gets the name in headers and in the footer of emails, to help you filter on that name. [15:56] beuno: I'll try to catch you at some stage. Not sure what I'll be doing in an hour though [15:57] deryck, right, that makes a bit of sense [15:57] deryck: Does that mean someone needs to go back to the subscription and give it a name? [15:57] huwshimi, I'm here this and next week [15:57] beuno: ok, I'll be around until Wednesday [15:59] beuno, huwshimi -- you can edit the subscriptions off the target or the +subscriptions or +structural-subscriptions pages, I believe. [15:59] bigjools: r=me [15:59] adeuring: cheers [16:00] beuno, huwshimi -- should be able to name it from: https://launchpad.net/launchpad for example [16:00] deryck, so maybe it needs a default name? [16:01] I may have some details wrong, though, since it's Yellow Squad working on this now :-) [16:01] beuno, yeah, probably some string concact of $target-$user-sub or some such. [16:01] (Unamed) seems not clear or useful [16:01] jcsackett: Already done :) [16:02] \o/ thanks, allenap! [16:04] beuno, there should still be old headers which one used to use for subscriptions filtering [16:04] beuno, can you file a bug? I can triage it and assign to story tag. [16:05] beuno, fwiw, the bug mail you might have started getting could be due to a change in launchpad product configuration, and is unrelated to the beta fwiw [16:05] deryck, thanks for helping out with this :) [16:07] danilos, np. :-) I don't mean to nosy my way into your work ;) [16:07] deryck, not at all, just go ahead, I am off now as well :) [16:07] ok cool :-) [16:08] deryck, and we appreciate all the help we can get :) [16:08] It's an awesome feature. Many will love it. [16:08] as long as launchpadders don't confuse this configuration change with the feature :) [16:09] (i.e. ~launchpad losing the contact address) [16:10] deryck, I can [16:11] danilos, right, no, I get the same amount of email (I was subscribed to everything) [16:11] danilos: But this change should mean you can stop getting all that mail though right? [16:11] this was me taking advantage of muting some email :) [16:11] huwshimi, right [16:11] deryck, I'll file the bug, I need the karma [16:11] beuno, I am sure you'll find a bunch of bugs to report :) [16:12] danilos: Do you know how I can mute all of this new mail? [16:13] I don't even see why we need ~launchpad-bugs. Just unsubscribe it from the ~launchpad, no? [16:14] isn't that the account causing issues? huwshimi ^^ [16:14] deryck: I think it might be [16:14] ~launchpad-bugs' subscription to /launchpad can be removed by a Launchpad admin. But it's not the main problem. [16:14] ah [16:14] ~launchpad also gets vcs-imports spam, loggerhead question spam, probably launchpad-reviewers MP spam. [16:14] ah [16:14] None of that is fixable. [16:14] and ~launchpad lost it's email address [16:14] I see [16:15] We may have to readd the address, once everyone is pissed off enough to file critical bugs about all the unstoppable mail. [16:15] deryck: It's not just that. I'm getting heaps of mail from Loggerhead and stuff now too [16:15] But until we have lots of bugs filed it would be silly to readd it. [16:15] Because we'd never fix it properly. [16:15] yeah [16:15] (see eg. the last 5 years of having a contact address) [16:16] we just shouldn't expand teams to individual emails. why would anyone really want that ever? [16:16] could be easier to fix than file the bug ;) [16:16] deryck: Then who gets the mail? [16:17] eg. ~launchpad is an admin for some teams. Who gets the mail that somebody is trying to join? [16:17] huwshimi, you should be able to go to any of the bugs with emails, click on "Edit bug mail" and mute the team subscription [16:17] wgrant, if you don't set a contact for the team, no one. [16:17] I think team subscriptions to any object are stupid and should be banned. [16:17] huwshimi, with something like "Don't send me email" [16:17] But we use subscriptions for security too. [16:17] So we can't do that. [16:17] create a security mailing list for the team that people optionally subscribe to. [16:18] this is how the rest of the world works, no? [16:18] danilos: Thanks [16:18] This sounds a lot like working around an incomplete advanced subscriptions feature to me :) [16:18] maybe in one sense [16:19] but I think my point is that you want a security or team mailing list, not a team subscription. [16:19] wgrant, for security stuff, direct subscriptions are created for the security contact on all security bugs [16:19] that magically expands [16:19] danilos: What about stuff that comes from questions etc? [16:19] Project windmill-db-devel build #245: STILL FAILING in 1 hr 5 min: https://lpci.wedontsleep.org/job/windmill-db-devel/245/ [16:19] deryck: Why? [16:19] huwshimi, can't help there, other than filtering [16:20] danilos: OK thanks :) [16:20] deryck: Security email isn't special enough that it deserves a list. That's just the only way it can sensibly be done now, because LP is broken. [16:20] Hmm, poolie's not here. [16:20] We should unsubscribe loggerhead-team from loggerhead's questions. [16:20] That will stop a lot of the immediate spam. [16:21] wgrant, maybe you're right. I'm just being clear that the rest of the world uses a security mailing list, not team membership to manage security bugs. [16:21] and that we're conflating the two [16:21] * danilos -> really out === danilos is now known as danilo-afk [16:21] what happened now is that we have a team subscribed for many reasons, not to manage mail [16:21] and the conflation is the issue. [16:21] Right. [16:21] A lot of these reasons are obsolete. [16:21] right [16:22] Ah, flacoste is readding the address now. [16:22] and that's all I mean. I don't mean to overanalyze or suggest the one true way to fix it. just pointing out the root issue. [16:22] deryck: The root issue is that we need to separate permission and subscription, and we probably also want to extend the excellent new structural subscription stuff to cover everything. [16:23] wgrant: yeah, PQM bot will receive all the email we are receiving [16:23] flacoste: We should really fix that PQM bug. [16:23] so it will be sending those kind of error message in a lot of case [16:23] wgrant, yes, agreed. [16:23] wgrant: or we should fix launchpad spam :-) [16:23] to be honest, i'm not sure what is the PQM bug in this case [16:23] that it's a member of ~canonical [16:23] actually [16:23] yeah, some things we really just should never send mail for [16:24] i don't know why PQM should be a member ~canonical [16:24] it's mainly for branch [16:24] and we always subscribe it explicitely [16:24] which is better [16:24] flacoste: The PQM bug is that it crashes when it can't verify a signature. [16:24] Rather than just ignoring the mail. [16:25] Also it probably shouldn't be a member of ~launchpad, I guess, but it might need those privileges. [16:25] well, even if it didn't oops [16:25] it should send an email about the error [16:25] it can be a legitimate user error [16:25] Possibly. [16:25] I guess having ~launchpad-pqm's email address set to PQM's email address is pretty pointless. [16:25] also [16:25] I can see no possible benefit of that. [16:25] me neither [16:26] but it's hard to set a LP profile to a sink hole [16:26] address [16:26] Yeah. [16:26] We need robot profiles. [16:26] sinzui, ping === salgado-lunch is now known as salgado [16:52] benji, are we making noise about +subscriptions or +structural-subscriptions as part of y'all's work? or keeping that on the dl? [16:53] adeuring: Could you please review https://code.launchpad.net/~abentley/launchpad/daily-build-permissions/+merge/60207 ? [16:53] abentley: sure [16:53] Thanks. [16:54] deryck: I don't think they're secrets; both should be linked to from user-visible parts of the UI (those users who can see them because of the feature flag that is) [16:55] adeuring: I have a small review for you when you have a minute: https://code.launchpad.net/~benji/launchpad/bug-777794/+merge/60209 [16:56] benji: sure, I'll look [16:56] thanks [16:56] benji, ok, thanks. Early on in that story there was talk of hiding them until they were ready, and I didn't want to assume they were ready. :-) [16:57] * deryck is doing IRC duties and wants to be prepared as people ask about new work [17:01] benji: r=me [17:01] thanks [17:15] abentley: r=me, just one note: there is an "import re" in a method in test_sourcepackagerecipe.py. I think this import can happen at the top of the file [17:29] deryck: I think all registry admins can hide bug and question comments. I do not think https://answers.launchpad.net/launchpad/+question/156277 should be assigned to an admin [17:29] sinzui, ah, sorry. from the web UI? I didn't realize this. [17:30] sinzui, or using an api script? [17:30] deryck: https://dev.launchpad.net/MaintenanceRotationSchedule has link to dealing with spam and there is a script that you can run [17:30] sinzui, ok, I'll take a look. [17:31] I have such a low tolerance for administrivia. I really suck at keeping up with all this. [17:32] deryck: jcsackett is working on the UI to toggle comment visibility now. You may not ever run that scrip again after today [17:33] * bigjools is with deryck [17:33] nice [17:34] sinzui, and if I have a user spamming bugs with linkedin invites, do I deactive or suspend the account? [17:34] * deryck sucks at reading 45 wiki pages too :-P [17:34] suspend and include the question URL in the comment so that we know why [17:34] gotchas [17:38] sinzui, why isn't hide-comments in the lp tree? [17:38] ask gmb he wrote it [17:38] deryck: I modified that script a week ago so that we could work with questions too [17:39] As I said, jcsackett is adding the UI now [17:39] sinzui, is that the copy in the wiki? [17:39] ah true [17:39] yes it is [17:39]