| StevenK | Bleh, webservice.patch is giving me 401 :-( | 00:46 |
|---|---|---|
| StevenK | How am getting ValueError: (<Bug at 0x112af190>, 'subscribe', 'launchpad.Edit'), the webservice is for the owner of the bug :-( | 01:08 |
| wgrant | StevenK: The owner doesn't always have access if it's private | 01:10 |
| StevenK | Which it isn't. | 01:11 |
| StevenK | wgrant: Diff is at http://pastebin.ubuntu.com/1195794/ , test output at http://pastebin.ubuntu.com/1195796/ | 01:14 |
| wgrant | StevenK: Have you checked that webservice_for is not defaulting to an RO token? | 01:16 |
| lifeless | or anonymous ? | 01:17 |
| wgrant | It's not anonymous, since a person is being passed in | 01:17 |
| StevenK | Oh, bleh, I bet that's it. | 01:19 |
| StevenK | permission=OAuthPermission.READ_PUBLIC | 01:20 |
| StevenK | Yeah | 01:20 |
| wgrant | :D | 01:20 |
| StevenK | lifeless: You'll never guess what the fix is. | 01:35 |
| wgrant | Is it just a matter of checking in the subscriber if the modification event actually has any modifications? | 01:39 |
| lifeless | StevenK: don't use the API ? | 01:40 |
| StevenK | if event.edited_fields is None: return | 01:40 |
| StevenK | wallyworld, wgrant: https://code.launchpad.net/~stevenk/launchpad/subscription-no-longer-updates/+merge/123467 | 01:46 |
| wallyworld | StevenK: busy, will look soon | 01:46 |
| StevenK | wgrant: Do you think 1021773 and 1035338 could be dupes? I can not load the oops to compare them. | 02:09 |
| wgrant | StevenK: Yes, they're probably dupes | 02:11 |
| wgrant | And they're probably related to a lot of the scanner timeouts too | 02:12 |
| wgrant | Some thing or things are holding locks on branch for many seconds | 02:12 |
| wgrant | StevenK: What if edited_fields is an empty list or so? | 02:13 |
| wgrant | We probably just care if it evaluates to boolean false | 02:14 |
| StevenK | wgrant: I can change that bit | 02:15 |
| StevenK | wgrant: http://pastebin.ubuntu.com/1195864/ Is that your only concern? | 02:17 |
| wgrant | StevenK: I think so | 02:20 |
| wgrant | StevenK: Are there similar subscribers on other objects? | 02:20 |
| StevenK | wgrant: I have no idea -- I didn't think anything but bugs used Zope subscribers. | 02:34 |
| StevenK | And I didn't know bugs used them until Friday afternoon. | 02:34 |
| wgrant | It's largely localised to Bugs, but I'd look around just in case | 02:34 |
| lifeless | bugs | 02:34 |
| lifeless | answers | 02:34 |
| lifeless | soyuz has some bits I think | 02:34 |
| lifeless | code | 02:34 |
| lifeless | localised, no, centre of gravity, yes. | 02:35 |
| wgrant | Soyuz doesn't really use them | 02:35 |
| wgrant | Certainly not for ObjectModifiedEvents | 02:35 |
| StevenK | Soyuz has one. | 02:35 |
| wgrant | Oh? | 02:35 |
| StevenK | <subscriber | 02:35 |
| wgrant | In other news, where are my OOPS reports | 02:35 |
| StevenK | for="lp.soyuz.interfaces.archivesubscriber.IArchiveSubscriber | 02:35 |
| StevenK | lazr.lifecycle.interfaces.IObjectCreatedEvent" | 02:35 |
| StevenK | handler="lp.soyuz.mail.notifications.notify_new_ppa_subscription"/> | 02:35 |
| wgrant | Ah | 02:35 |
| StevenK | wgrant: I can't think of a similar pattern off the top of my head, anyway. We call a method on X, don't modify X and return Y | 02:37 |
| wgrant | I'd look around for the other ObjectModifiedEvent subscribers and see what they do | 02:37 |
| wgrant | 24460 0 Archive:EntryResource:getPublishedBinaries | 02:38 |
| wgrant | ouch | 02:38 |
| lifeless | not deployed yet ? | 02:40 |
| wgrant | A couple of hours ago | 02:42 |
| wgrant | So tomorrow's report should be clean | 02:42 |
| wgrant | hopefully | 02:42 |
| StevenK | wgrant: Everything looks fine -- I'm not sure if subscribing to a branch over the API causes the same issue. | 02:46 |
| StevenK | wgrant: Which is immune, so everything is fine. | 02:57 |
| wgrant | OK | 02:59 |
| wgrant | r=me, thanks | 02:59 |
| lifeless | review plox: https://code.launchpad.net/~lifeless/js-oopsd/post/+merge/123474 | 04:24 |
| lifeless | wallyworld: or StevenK: ^ wgrant, is obsessing on performance. | 04:25 |
| wgrant | Correctly obsessing; this missing index is at the root of all but one known disclosure performance regression :) | 04:26 |
| wallyworld | otp, just a sec | 04:26 |
| wallyworld | wgrant: what missing index is that? | 04:36 |
| * StevenK stabs bugactivity | 04:38 | |
| lifeless | wallyworld: the inverse index on a M:M join table | 04:38 |
| wgrant | wallyworld: See #launchpad-ops. accesspolicygrant(grantee) isn't indexed, so finding a list of policies that a person can see uses accesspolicygrant(policy, grantee) | 04:39 |
| wallyworld | which M:M table? | 04:39 |
| wgrant | And takes 100ms rather than 1ms | 04:39 |
| wallyworld | ok | 04:39 |
| wgrant | This fixes all the branch access checks | 04:39 |
| wallyworld | excellent | 04:39 |
| wgrant | It possibly won't fix code.launchpad.net/ubuntu, as the MP query there is pretty obscene, but I have a simple query workaround for that | 04:39 |
| StevenK | Sigh, it's been so long since I've had do preloading, I've forgotten what to look for. :-( | 05:09 |
| * lifeless begs for a review | 05:11 | |
| jam | lifeless: in your overview, you say any browser to do "POST" or "GET", but in your "Allow-Methods" header, you have "PUT", "DELETE", am I missing something there? | 05:14 |
| jam | and your test seems a little bit circular (rather than asserting what the actual content is, it asserts that the content matches the list that you say it should be) | 05:18 |
| jam | so if I put whatever I want it cors_headers, the test still passes. | 05:18 |
| jam | (vs saying "GET" is in the list of Access-Control-Allow-Methods) | 05:18 |
| jam | so it depends what you want to assert | 05:19 |
| lifeless | thats true | 05:19 |
| jam | if you are asserting that passing the cors_headers is hooked up == good | 05:19 |
| lifeless | I had a more specific test but perhaps over enthusiastically refactored | 05:19 |
| jam | asserting that it conforms to the spec == better. | 05:19 |
| lifeless | the it should be GET+POST | 05:19 |
| lifeless | jam: I'll a) fix the static set of headers and b) convert the preflight check to be manual rather than circular. | 05:20 |
| jam | lifeless: aside from that sanity check, I don't know that it is worth having *me* fully understand CORS, so I'll trust that we can iterate until everything works. | 05:21 |
| jam | (If it doesn't work properly in browsers, then we update the test suite until it does) | 05:21 |
| lifeless | I'm not sure anyone really understands CORS. | 05:22 |
| lifeless | I've not read a more confusing spec. | 05:22 |
| jam | lifeless: which is why I think iterating towards 'working' is good. | 05:22 |
| jam | As browsers may not conform at all to the spec either. | 05:22 |
| lifeless | that too | 05:22 |
| jam | confusing specs tend not to be followed by anyone very well :) | 05:22 |
| jam | lifeless: review posted | 05:27 |
| lifeless | jam: I've pushed a revised version | 05:28 |
| lifeless | jam: if there was nothing more in the review, it should be good | 05:28 |
| jam | lifeless: the only other thing I thought about, do we want to have a wide-open access policy? | 05:29 |
| === Ursinha is now known as Ursinha-afk | ||
| lifeless | also need a review for https://code.launchpad.net/~lifeless/python-oops-tools/bug-1048470/+merge/123480 please :) | 05:30 |
| lifeless | jam: yes. | 05:30 |
| jam | lifeless: mostly just making sure that was true. It might be good to have a code-level comment about it. | 05:30 |
| lifeless | jam: Folk that want to attack us can just ignore the policy entirely. | 05:30 |
| jam | as it looks like a security hole | 05:30 |
| jam | I thought it is probably ok | 05:30 |
| lifeless | jam: we can't ever trust the data we get for this service at all. | 05:30 |
| jam | lifeless: aiui, this is about the fact that launchpad.net may send out a javascript header, but we want to post the OOPS to oops.canonical.com, right? | 05:31 |
| jam | and anyone can just post to oops.canonical.com | 05:31 |
| jam | and this is just about telling javascript that it is ok that even though launchpad.net gave you the JS file, it can send data to oops. | 05:32 |
| lifeless | yes, or u1, or ... | 05:32 |
| jam | lifeless: so you've gotten an approve* from me | 05:33 |
| lifeless | tight rules are used to prevent query methods that return sensitive data, and are secured by cookies or whatever, from being fooled via CRSF attacks and the like | 05:33 |
| jam | not sure if you need someone official on it :) | 05:33 |
| lifeless | jam: you shold be official | 05:33 |
| lifeless | wallyworld: thanks! I totally don't know why I didn't thinkyou'd reviewed. | 05:34 |
| lifeless | wallyworld: slow mail day or something. | 05:34 |
| wallyworld | np | 05:35 |
| jam | lifeless: I've never been officially given the 'launchpad reviewer' title, at least that I can remember. | 05:35 |
| lifeless | jam: you're in the group; do you have a mentor reviewer to give you the LP magic sauce ? | 05:36 |
| StevenK | TAL, please die in a large chemical fire. http://pastebin.ubuntu.com/1196061/ | 05:37 |
| lifeless | wallyworld: if I can trouble you again .. https://code.launchpad.net/~lifeless/python-oops-tools/bug-1048470/+merge/123480 | 05:38 |
| jam | lifeless: https://code.launchpad.net/~lifeless/python-oops-tools/bug-1048470/+merge/123480 just to validate assumptions, the decoding of 'data' already maps things to Unicode, so that we don't have to worry that the data might be a UTF8 string that we are then doing .encode('utf8') on (which will break with the magic decode step). | 05:38 |
| lifeless | wallyworld: its about 4 lines of code | 05:38 |
| wallyworld | sure | 05:38 |
| lifeless | wallyworld: or... jam has it :P | 05:38 |
| lifeless | wallyworld: so no worries | 05:38 |
| lifeless | jam: right, its coming out of the ORM | 05:38 |
| lifeless | jam: so its always unicode. | 05:38 |
| lifeless | jam: its assigned slightly higher in the file | 05:38 |
| jam | lifeless: so the thing you changed was 'pageid' related, but the test is testing 'topic'. | 05:38 |
| lifeless | jam: topic in the wire protocol is mapped to pageid in the schema | 05:39 |
| lifeless | the schema is old and rather more LP centric than the wire protocol. | 05:39 |
| lifeless | creaky even | 05:39 |
| jam | lifeless: I don't think I've ever had an official mentor for obtaining LP secret sauce. | 05:39 |
| lifeless | jam: so, lets get you one. I know, gmb, he has nothing to do :P | 05:40 |
| lifeless | jam: you clearly know python and how to do code review; the only thing you need is to convince a mentor reviewer than you can tell the difference between code you grok and code that will contain surprising gems. | 05:40 |
| lifeless | lets see what tarmac thinks of that one | 05:41 |
| wallyworld | lifeless: should there also be a test for the utf8 encoded oopsid? | 05:42 |
| StevenK | wgrant: http://pastebin.ubuntu.com/1196061/ | 05:47 |
| lifeless | wallyworld: mmm, I only changed that because the test I wrote caught it :) | 05:47 |
| lifeless | wallyworld: so, I don't think an explicit one is called for, no. This whole code could be radically improved by using a template engine. | 05:47 |
| wallyworld | ah, ok. gotta love tests :-) | 05:47 |
| wallyworld | +1 to the template engine | 05:47 |
| lifeless | yeah, I added the test to validate my assumption about the cause of the problem, and then fixed everything it uncovered. | 05:47 |
| lifeless | Stuff trying to understand this stuff :> | 05:47 |
| wgrant | StevenK: It's probably raising an AttributeError because your code is terrible | 05:48 |
| wgrant | Anywhere inside the attribute it complains about | 05:48 |
| wgrant | An AttributeError will just be squashed into that useless LocationError | 05:48 |
| StevenK | wgrant: I didn't think http://pastebin.ubuntu.com/1196075/ was terrible :-( | 05:49 |
| wgrant | StevenK: Possibly personID isn't on IBugActivity | 05:50 |
| wgrant | So it's not accessible | 05:50 |
| StevenK | ForbiddenAttribute: ('personID', <BugActivity at 0xc852410>) | 05:52 |
| StevenK | But it's there if you dropkick the security proxy | 05:52 |
| lifeless | that would be the attribute not being on the interface | 05:53 |
| StevenK | It's SQLBase, so I have no idea why it isn't. | 05:54 |
| lifeless | because they aren't magic. | 05:54 |
| lifeless | They have to be explicitly listed like all other attributes. | 05:55 |
| wgrant | The interface is not SQLBase | 06:08 |
| wgrant | SQLBase only affects the model | 06:08 |
| StevenK | Yeah, I realized after lifeless' comment | 06:08 |
| StevenK | wallyworld, wgrant: https://code.launchpad.net/~stevenk/launchpad/bugtask-activity-preload/+merge/123484 | 06:09 |
| wgrant | StevenK: Do you really care about just ValidPersonCache? | 06:10 |
| wgrant | You probably want the Person too | 06:10 |
| wgrant | Which means you want PersonSet.getPrecachedPersonsFromIDs | 06:10 |
| StevenK | wgrant: It's strange, if I did load_related(Person, ...) the query count went to 15. | 06:11 |
| wgrant | wallyworld: I was able to reproduce the request a review bug on prod a couple of hours ago | 06:22 |
| wgrant | wallyworld: It works fine until you click to submit | 06:22 |
| wgrant | Then it just disappears and does nothing | 06:22 |
| wallyworld | wgrant: worked for me locally | 06:22 |
| wallyworld | i didn't hit submit on prod, i'll try on qas | 06:22 |
| wgrant | wallyworld: With the yui 3.5 FF set? | 06:22 |
| wallyworld | i think so, unless i've remade my schema, which i may have done | 06:23 |
| wallyworld | the bug said the link wasn't even clickable | 06:23 |
| wallyworld | which is not the case anymore | 06:23 |
| wgrant | It doesn't say that the link was unclickable | 06:24 |
| wgrant | It even says "I was trying to request lifeless review a db patch branch.", which suggests it died post-input | 06:24 |
| wallyworld | it says the console shows an error when it is clicked | 06:24 |
| wallyworld | which implies the link didn't work | 06:24 |
| wgrant | It doesn't say anything about clicking, just that the green link was being used | 06:24 |
| wallyworld | how do you ue a link without clicking it? | 06:25 |
| wgrant | You don't, but it doesn't say it failed during the clicking of the link | 06:25 |
| wgrant | Just that it failed some time while using the functionality behind that link | 06:25 |
| StevenK | wgrant: http://pastebin.ubuntu.com/1196108/ when preloading Person only | 06:26 |
| wallyworld | sure, but i seem to have a recollection i saw the same issue and the link itself didn;t work | 06:26 |
| wgrant | wallyworld: Hm, I've never seen that | 06:26 |
| wgrant | StevenK: But it does reduce the query count somewhat? | 06:26 |
| StevenK | wgrant: It does not. It goes from 13 to 15. | 06:27 |
| wgrant | Oh | 06:27 |
| wgrant | I wonder if you're failing to invalidate the cache properly | 06:27 |
| StevenK | 97+ Store.of(bug).flush() | 06:27 |
| StevenK | ^ Not sufficent? | 06:27 |
| wgrant | flush is not invalidate | 06:27 |
| wgrant | flush flushes | 06:27 |
| wgrant | invalidate invalidates :) | 06:27 |
| wallyworld | wgrant: submitting a review request works fine on qas too | 06:29 |
| wgrant | wallyworld: Ah, qastaging only has 3.5.1 enabled for rick | 06:30 |
| wallyworld | :-( i assumed it was on for everyone in lp | 06:31 |
| wgrant | Likewise, but I just checked and it's not :( | 06:31 |
| wallyworld | bollocks | 06:32 |
| wgrant | Heh, we just passed MP 123456 a few hours ago | 06:32 |
| wgrant | wallyworld: Try rerequesting a review from lifeless on https://code.launchpad.net/~wgrant/launchpad/i-am-stupid/+merge/123478 | 06:33 |
| wallyworld | yep, so that failed | 06:34 |
| StevenK | wgrant: So my numbers were off. Slightly | 06:37 |
| StevenK | wgrant: Putting in list(getUtility(IPersonSet).getPrecachedPersonsFromIDs( | 06:37 |
| StevenK | ids, need_validity=True)) | 06:37 |
| StevenK | Still has a bunch of queries against ValidPersonCache | 06:38 |
| wgrant | That's very odd | 06:38 |
| wgrant | The same number of queries? | 06:38 |
| StevenK | So with no preloading, 50 bugactivity rows is 109 queries. With that getPrecachedPersonsFromIDs, it drops to 33. | 06:39 |
| wgrant | Oh, 50? | 06:39 |
| wgrant | You can't test with 50 | 06:39 |
| wgrant | The dev storm cache size is only like 100 | 06:39 |
| StevenK | Why not? | 06:40 |
| wgrant | Also, that's going to be slow even if the cache is huge :) | 06:40 |
| wgrant | No point going above, say, 5 | 06:40 |
| StevenK | My original test had 10 | 06:40 |
| StevenK | 10 bugactivity rows drops from 29 queries to 7. | 06:42 |
| wgrant | That sounds roughly correct, right? | 06:42 |
| StevenK | Yeah | 06:42 |
| StevenK | wgrant: Well, 5 rows is 7 queries, as is 10. Which is good. | 06:43 |
| wgrant | Yes :) | 06:43 |
| StevenK | wgrant: http://pastebin.ubuntu.com/1196125/ | 06:43 |
| wgrant | StevenK: I'd inline the list comp, but otherwise good | 06:44 |
| StevenK | Ah yes, that would work. | 06:44 |
| StevenK | Haha, wallyworld approved it 32 minutes ago, I just didn't notice. | 06:46 |
| wgrant | Heh, so he did | 06:46 |
| wgrant | I was right, though :) | 06:47 |
| StevenK | Not disputing that one bit. | 06:47 |
| adeuring | good morning | 07:40 |
| StevenK | wgrant: Heh, if not event.edited_fields is not quite the solution -- the 313 failures from ec2 attests to that. | 07:50 |
| === almaisan-away is now known as al-maisan | ||
| === al-maisan is now known as almaisan-away | ||
| wgrant | stub: Morning | 08:28 |
| === almaisan-away is now known as al-maisan | ||
| wgrant | stub: Could you review https://code.launchpad.net/~wgrant/launchpad/i-am-stupid/+merge/123478? | 09:44 |
| lifeless | wgrant: love the name | 09:44 |
| wgrant | I try. | 09:45 |
| mgz | such a let down, just missing indicies | 09:45 |
| nigelb | lol | 09:45 |
| nigelb | Shouldn't it go into db-devel or something? | 09:45 |
| wgrant | Nah, nowadays indices go to devel and get applied live, without downtime | 09:46 |
| nigelb | Ooooh. Exciting :D | 09:46 |
| stub | k | 09:47 |
| stub | wgrant: Not stupid, I was hoping the primary key indexes would make these unnecessary | 09:48 |
| wgrant | The only really important one is apg(grantee, policy), and it was pretty obvious that that was necessary from the start, but I apparently forgot to create it | 09:49 |
| stub | well... grantor is needed or person merge explodes | 09:49 |
| stub | yeah, but the (policy, grantee) would often be usable instead of the inverse | 09:49 |
| stub | These can go live now? | 09:50 |
| wgrant | Please | 09:50 |
| wgrant | I will be watching the DBR clear up immediately :) | 09:50 |
| wgrant | Well | 09:51 |
| wgrant | I would be if my Yubikey wasn't flashing angrily at me | 09:51 |
| czajkowski | hah | 09:51 |
| wgrant | This is inconvenient | 09:53 |
| czajkowski | usually I have to go find my phone and hope it's charged | 09:54 |
| stub | wgrant: Might be worth making the compound indexes UNIQUE? | 09:54 |
| stub | wgrant: Don't think it makes any difference now, but a future planner might notice. | 09:54 |
| wgrant | stub: Right, it doesn't make any difference now. | 09:54 |
| wgrant | But I could... | 09:54 |
| wgrant | Possibly a good idea | 09:54 |
| stub | yeah | 09:54 |
| * wgrant does so | 09:57 | |
| wgrant | stub: That's done | 09:59 |
| wgrant | Diff not updated yet, though | 09:59 |
| wgrant | Hm | 10:01 |
| wgrant | Do we still call unique indices __key? | 10:01 |
| wgrant | Rather than __idx? | 10:01 |
| stub | yeah, __key | 10:04 |
| wgrant | stub: Done | 10:07 |
| wgrant | I think that's it... | 10:07 |
| jelmer | mgz: your summon worked! | 10:09 |
| mgz | magic :) | 10:09 |
| stub | wgrant: All done on production | 10:14 |
| stub | and qastaging | 10:14 |
| wgrant | stub: Thanks | 10:16 |
| wgrant | branch and APG will hopefully drop off the top of the DBR's Most Read Tables list... | 10:17 |
| wgrant | Well, APG is dropping to nothing, but branch is still terrible :( | 10:29 |
| wgrant | stub: So, I see 9.2 has been released... just as we are almost completely migrated to 9.1 :) | 11:58 |
| rick_h_ | chaining repliate ftw! | 11:59 |
| === gary_poster|away is now known as gary_poster | ||
| === benji changed the topic of #launchpad-dev to: http://dev.launchpad.net/ | On call reviewer: benji | Firefighting: - | Critical bugs: ∞ | ||
| jam | benji: poke about lpsetup. It appears that my patch cannot land because the tarmac process is broken? | 12:59 |
| jam | (someone said that if a patch breaks, it doesn't teardown the lxc instance, so the next submission will fail because the lxc is still there) | 12:59 |
| jam | I'm not entirely sure who to talk to, but I figured starting with someone on yellow might be reasonable. | 13:00 |
| benji | jam: hmm, I hadn't heard of that; I'll see what I can do about it | 13:00 |
| jam | benji: https://code.launchpad.net/~jameinel/lpsetup/missing_lxc_1045728/+merge/122663 is the merge request | 13:01 |
| benji | bac: are you aware of such a problem? ^^^ | 13:01 |
| jam | the failure from tarmac was 'juju', 'bootstrap', '-e', 'lpsetup-testing-lxc' was failing. | 13:01 |
| bac | benji, jam: the first thing to do is to ssh into the machine and have a look around. | 13:04 |
| bac | you should be able to 'ssh tarmac@lptarmac.no-ip.org'. you'll need your .ssh/config configured like https://pastebin.canonical.com/74130/ | 13:04 |
| benji | ok, I'll take a look | 13:05 |
| jam | bac: as user 'ubuntu' or as user 'tarmac' ? | 13:05 |
| bac | jam: tarmac is easiest. either works and both have sudo. | 13:08 |
| bac | jam: did you get access to the machine? | 13:12 |
| jam | bac: otp now, will try later. | 13:12 |
| benji | bac: I did. No lxc containers are running, no juju environment was bootstrapped. I bootstrapped the test env and it seemed to work fine. (and then I destroyed it) | 13:13 |
| benji | is there some way I can re-run the merge request? | 13:14 |
| bac | benji: yes. the log file is in ~/.config/tarmac/tarmac.log. grep for Merging and it'll show you the branch name | 13:15 |
| bac | cd ~/repos/lpsetup/trunk | 13:15 |
| bac | hey benji, wanna use tb? | 13:15 |
| benji | bac: termbeamer is up and sending you my session | 13:17 |
| benji | bac: juju appears to be hung waiting for the unit to come up | 13:42 |
| benji | juju status times out | 13:43 |
| czajkowski | bac: ello when do you guys have your stand up? | 13:44 |
| bac | czajkowski: 1600Z | 13:46 |
| bac | why? | 13:46 |
| bac | benji: that's ungood. | 13:47 |
| bac | benji: we could bounce the instance but that doesn't solve the problem. | 13:48 |
| benji | bac: yeah; I was hoping youhad seen this before; let me investigate some more and see if I can get closer to a root cause | 13:50 |
| czajkowski | bac: might be good to join a few in case we have over lap on stuff | 13:52 |
| czajkowski | given you're now on maintenance no ? | 13:52 |
| stub | wgrant: Yup, but still waiting for a point release or two. | 13:54 |
| bac | benji: ok. if you ever want to bounce it, just use 'poweroff' in the instance and then the startup script in ./bin on your local machine | 13:55 |
| benji | bac: where is the startup script? | 13:57 |
| bac | in the config branch in bin | 13:58 |
| bac | benji: lp:lp-tarmac-configs | 13:58 |
| benji | cool, thanks | 13:58 |
| rick_h_ | deryck: ping, when you get a sec then I want to sanity check the +newproduct ProjectGroup vs +new and Product | 14:16 |
| rick_h_ | deryck: http://paste.mitechie.com/show/782/ for the quick rundown on the bits fitting together | 14:17 |
| deryck | rick_h_, looking | 14:26 |
| deryck | rick_h_, so looking at that, it seems to me you only care about getting allowed types in the second step if the context is IProduct. | 14:28 |
| deryck | rick_h_, or else you could have the same method on IProject that only returns Public. | 14:29 |
| rick_h_ | deryck: right, ok that sounds good | 14:29 |
| === al-maisan is now known as almaisan-away | ||
| === matsubara is now known as matsubara-lunch | ||
| === deryck is now known as deryck[lunch] | ||
| benji | bac: how long should the inttance take to start? it has been around 30 minutes of "[localhost] local: scp -r -o StrictHostKeyChecking=no * 10.55.60.129:" | 16:28 |
| * benji gets some lunch. | 16:30 | |
| === Ursinha` is now known as Ursinha | ||
| === Ursinha is now known as Guest37586 | ||
| === deryck[lunch] is now known as deryck | ||
| === Ursinha-afk is now known as Ursinha | ||
| deryck | lifeless, ping. wondering if my 10.4 second db patch, according to dbupragde.log is okay for deployment. | 17:18 |
| rick_h_ | benji: mp for you if you get a sec: https://code.launchpad.net/~rharding/launchpad/pp_register/+merge/122666 | 17:22 |
| rick_h_ | deryck: I left the fake attribute in there with an XXX for the db side since I figured it was more ugly to try to dynamically add in the @property conditionally | 17:23 |
| deryck | rick_h_, it's all behind a feature flag, too, right? | 17:24 |
| rick_h_ | deryck: right | 17:24 |
| deryck | ok | 17:24 |
| benji | rick_h_: sure; it will be a while though | 17:24 |
| rick_h_ | benji no hurry, thanks | 17:25 |
| rick_h_ | deryck: and since I'm using an @property and you'll have a db attribute there shouldn't be much to resolve no matter who gets in first for conflict stuff. | 17:25 |
| deryck | rick_h_, right. sounds fine to me, but I'll defer to benji for the real review. | 17:26 |
| rick_h_ | deryck: right, just fyi your way | 17:26 |
| deryck | gotchas | 17:26 |
| benji | bac: I can't get the tarmac container to start; do you mind trying for me? | 17:40 |
| bac | benji: can you ssh to it? | 17:43 |
| benji | bac: nope | 17:47 |
| bac | benji: you did try by ip addr? | 17:47 |
| benji | bac: nope; let me find it (I don't have it on hand) | 17:48 |
| benji | found it | 17:49 |
| benji | bac: ssh tarmac@10.55.60.123 fails with "Permission denied (publickey)." | 17:50 |
| bac | benji: yep, don't use tarmac@. | 17:51 |
| === matsubara-lunch is now known as matsubara | ||
| bac | he's probably not created yet. if you don't specify you log in as ubuntu | 17:51 |
| benji | bac: "ssh 10.55.60.123" still gives me "Permission denied (publickey)." | 18:04 |
| bac | benji: uhg | 18:07 |
| lifeless | deryck: definitely not | 18:39 |
| lifeless | deryck: we're looking for 4-5 seconds max IIRC the math correctly. | 18:40 |
| lifeless | deryck: which patch was it | 18:40 |
| lifeless | ? | 18:40 |
| abentley | lifeless: I believe deryck meant https://code.launchpad.net/~deryck/launchpad/product-db-changes-for-privacy/+merge/122941 | 18:59 |
| deryck | lifeless, yeah, it's 2209-33-1, the specification_sharing_policy column. | 18:59 |
| deryck | lifeless, I have no idea how to get a single column add any faster. :) | 18:59 |
| deryck | lifeless, sorry for the delay, but I've got pool people here to give and estimate on my pool servicing. I need to step away again, but won't be too long. | 19:00 |
| lifeless | deryck: you're fine, a close inspection shows we have a time reporting bug. | 19:11 |
| lifeless | 2012-09-10 00:40:09 INFO Applying /srv/staging.launchpad.net/staging/launchpad/database/schema/patch-2209-23-1.sql | 19:12 |
| lifeless | 2012-09-10 00:40:20 INFO Applying /srv/staging.launchpad.net/staging/launchpad/database/schema/patch-2209-31-1.sql | 19:12 |
| lifeless | 23-1 took 11 seconds | 19:12 |
| lifeless | I need to see what that is. | 19:12 |
| lifeless | it might be wgrants index, in which case it went on concurrently on prod and we can ignore it. | 19:12 |
| lifeless | deryck: which is is, so your patch is fine, it isn't 11 seconds long. | 19:14 |
| lifeless | deryck: could you please file a bug about this? It will confuse others.. | 19:15 |
| lifeless | probably fallout from slony | 19:15 |
| bac | benji: i was able to restart the c'stack instance with no problem. did you update your env per the instructions on the internal wiki? | 19:29 |
| benji | bac: by sourcing ~/.canonistack/novarc? if so, yes | 19:29 |
| bac | benji: in .canonistack do you have a benji_lcy01.key or just benji.key? | 19:30 |
| benji | bac: benji_lcy01.key | 19:30 |
| bac | benji: ok, then, i'm not sure what the problem is | 19:30 |
| bac | maybe later we can use a termbeamer to see what is going on | 19:30 |
| benji | bac: that'd be great; ping me when you have a moment | 19:31 |
| abentley | deryck: Can we chat? | 19:32 |
| deryck | lifeless, sure, I can file a bug. | 19:41 |
| deryck | abentley, sure. | 19:41 |
| abentley | deryck: Okay, g+ hangout. | 19:42 |
| deryck | ok, coming now.... | 19:43 |
| lifeless | gary_poster: I've fixed (just pushed to trunk now) a bug in testrepository causing test times to be badly undercounted when the stream isn't timed. I think this won't affect paralleltest due to the subunit stream from zope being timed, but if it isn't timed, then I'm wrong, and you would want the fix. | 19:44 |
| gary_poster | lifeless, cool! it is timed for us because we pass -vvv (or something silly like that) to bin/test, but might be helpful for us even in other situations. | 19:45 |
| lifeless | https://bugs.launchpad.net/testrepository/+bug/1048126 for reference | 19:47 |
| _mup_ | Bug #1048126: Test times of untimestamped streams are bogus <Testrepository:Triaged by lifeless> < https://launchpad.net/bugs/1048126 > | 19:47 |
| lifeless | flacoste: hi ? | 20:04 |
| flacoste | hi lifeless | 20:04 |
| lifeless | we on ? | 20:04 |
| flacoste | we are | 20:04 |
| lifeless | hah | 20:05 |
| lifeless | thats something skype handles better | 20:06 |
| lifeless | when both sides call | 20:06 |
| lifeless | I'll start a fresh fresh one | 20:06 |
| flacoste | lifeless: hang on, i'm with google on the phone (search broken) | 20:06 |
| lifeless | ok | 20:06 |
| lifeless | grab me whenever | 20:07 |
| flacoste | lifeless: can you send me a new invite? | 20:15 |
| lifeless | sure | 20:16 |
| lifeless | flacoste: you should have it... | 20:18 |
| bac | benji: still around? | 21:04 |
| benji | bac: just :) | 21:04 |
| === benji changed the topic of #launchpad-dev to: http://dev.launchpad.net/ | On call reviewer: - | Firefighting: - | Critical bugs: ∞ | ||
| bac | benji: https://plus.google.com/hangouts/_/ef48f894900cf80079fb2996e64e6b0ae17df36d?authuser=1&hl=en-US | 21:04 |
| benji | bac: I'm afraid I'm away from my hangout setup; mind if we do it in the morning? | 21:05 |
| bac | benji: nope. have a good evening. | 21:05 |
| benji | thanks; you too! | 21:05 |
| === spm is now known as stevemci | ||
| === stevemci is now known as spm | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!