[00:10:28] *hexmode looks at email [00:11:06] is the PostgreSQL issue the only blocker now? [00:11:11] TimStarling: ^? [00:14:25] yes [00:14:53] I'm working on it now [00:15:20] TimStarling: "it" == tarball or Pg? [00:15:30] we've had postgres support since 1.7, there would have to be a really really good reason to drop it [00:15:34] pg [00:15:41] k [00:16:10] that's why I asked what your reasons were [00:17:10] sure... I'm trying to put together something now [00:17:28] I think I have good diffs, they just uncovered new holes [00:22:02] why did you make getConnection() use wgDBuser in r81439? [00:22:29] PostgreSQL's user model is pretty much identical to Oracle's, but we didn't need to do that there [00:23:41] as an aside, we really need a better way to track CR blockers to a release....are there any other unreviewed changes like this? [00:26:33] assuming Roan reviewed the ones that he backported, then I think we're ok [00:26:52] *hexmode looks at 81439 [00:27:11] hexmode and freakolowsky committed changes directly to REL1_17, which is how I ended up looking at them [00:27:36] if they had committed to trunk and added a 1.17 tag then Roan would have done them [00:28:28] so, 81439 introduces the $database param, right? [00:30:03] *robla disappears into meeting [00:30:03] in oracle, a seperate method were created, but a different connection was needed [00:30:34] *robla has a couple more minutes before he disappears [00:30:46] after you mentioned it [00:30:54] I looked at how oracle handled it [00:31:02] no, in oracle, the same user was used, just with a DBO_SYSDBA flag [00:31:03] and saw the seperate method [00:31:14] so I'm trying to duplicate that [00:31:27] postgresql doesn't have a DBO_SYSDBA flag [00:31:30] no such flag exists in PG AFAIK [00:31:32] right [00:31:59] so the installation user (the bug that started this all) has some weird permissions [00:32:05] it isn't a super user [00:32:12] but it can create users [00:32:16] and databases [00:32:44] the second connection is needed b/c of the way permissions are handled [00:33:13] lemme see if I can show you what I came up with [00:33:33] *hexmode goes to test his code [00:39:26] hrm.... maybe I just did something that worked. [00:45:16] *hexmode pasted "timstarling: almost works, still get wfGetDB() upon reinstall" at http://paste2.org/p/1461340 [00:46:06] *jorm yawns. [00:46:12] werdna: get ahold of alolita? [00:46:44] yep [14:57:22] The Association for Psychological Science's Wikipedia initiative http://www.psychologicalscience.org/index.php/members/aps-wikipedia-initiative is interested in getting more psychologists & students editing our wikis. Great! They say it would be great if a prof could batch-create 50 accounts so the students don't have to sign up themselves. Possible? [14:58:31] the ip limit will be the issue [14:58:50] one could write a little script and feed it account names and email addresses I suppose [15:03:47] apergos: who could run that? [15:04:24] It's enough to find a local sysop actually [15:05:07] my notes from the discussion: maybe could create a script, such that you could do it if you have the account_creator permission.... do it via current Account Creator sys. When someone wants to create a batch of accounts, they temp req those permissions. or Rosta from APS gets agreement from community to let her have that perm for this purpose. [15:05:36] their largest course so far was 85 [15:05:38] students [15:05:41] I think accountcreator flag is no big deal [15:05:58] we can create em via the api I bet [15:06:12] hence it's only a matter of an account running it that has privs [15:06:47] how many sysops are there? [15:06:52] 1800 [15:07:57] and you can ask each of them to make someone an accountcreator [15:11:48] ok, I gotta run [15:12:01] public forum about the debt [15:12:07] see folks later tonight [15:12:31] night apergos [15:12:41] later [15:13:17] Nemo_bis: so you're suggesting the workflow would be: I have a list of 50 usernames, I contact a sysop in IRC and ask them to create them for me? [15:13:53] sumanah, an accountcreator it's probably easier but yes [15:14:15] ok, thanks! [15:14:17] Or just ask a sysop to add you or someone else to accountcreator group :) [15:14:26] Sweet. I'll tell Rosta. [15:15:00] I assumed we didn't allow batch creation, because a quick Google didn't find me any directions, and because it seems like a gateway to spam/vandalism [15:20:08] *sumanah emails APS developers [15:20:14] er, just the one. [15:22:16] In case you create acounts by yourself, remember that you need to fill in the username and email and then click "by email" (nothing else), so that password is automatically generated and sent [15:23:22] thanks Nemo_bis - will append! [15:24:15] so the other issue APS is dealing with: their interface involves frames, and it seems like you cannot edit a Wikipedia page inside a frame. Is that the desired behavior? [17:03:22] !pad WMF-TOC [17:03:22] --elephant-- http://etherpad.wikimedia.org/WMF-TOC [17:10:05] alolita1: Which room is the scrum ? [17:13:22] alolita1: $operation_map = array( [17:13:23] 'highslows' => 'populateHighsLows', [17:13:23] 'problems' => 'populateProblems', [17:13:23] ); [17:13:24] [17:13:36] only those two are currently in the database/extension code base. [17:56:01] an issue Association for Psychological Sciences is dealing with: their interface involves frames, and it seems like you cannot edit a Wikipedia page inside a frame. Is that the desired behavior? [18:00:39] <^demon|away> sumanah: Intended behavior. "Desirable" is debatable :) [18:01:42] lol [18:03:17] ^demon|away: ok. what would you recommend as a workaround or a substitute? the point is to let people log into APS's site and edit from there - http://hciresearch2.hcii.cs.cmu.edu/~rfarzan/wikipedia/tool/wsignup.php [18:03:42] <^demon|away> Their site could do it via the API, rather than iframing it. [18:03:57] I'll suggest that to her, thanks [18:05:18] done [18:06:06] <^demon|away> sumanah: The iframe restrictions are for two reasons: one so people don't hotlink our content inside of their sites, and also so you can't be tricked into thinking you're on wp but really on the third-party site who then steals your credentials. [18:06:28] makes sense, ^demon|away [18:10:57] I don't think they should edit via the API [18:11:04] they would all appear to be one user [18:11:20] They do ask for our WP credentials, iirc [18:11:22] or they would need to provide their password to the intermediate site [18:14:00] *sumanah sends that as appendix [19:13:24] what does isPingLimitable() do? [19:13:53] Rate limiting I think [19:14:52] rate limiting means the interval restriction for carrying out an activity? [19:15:52] Yes [19:15:59] Like at most N actions in M seconds [19:18:42] got it, thanks :) [19:54:17] !swarm [19:54:17] --elephant-- I don't know anything about "swarm". [19:54:19] !testswarm [19:54:19] --elephant-- I don't know anything about "testswarm". [20:02:14] Ryan_Lane: Could you look at the vms ? [20:14:48] Krinkle: sure. gimme a sec [20:16:00] reconnected [20:16:17] after I get done with a few high priority things I'll work on a script to restart them [22:02:47] tomaszf, other dr. who nerds: http://www.ho-lo.co.il/archives/587 [22:31:51] I just posted to wikitech-l.