[00:23:04] MaxSem: just sent you the patch [13:26:53] MaxSem: Fatal error: Call to a member function getText() on a non-object in /usr/local/apache/common-local/php-1.20wmf7/extensions/MobileFrontend/api/ApiMobileView.php [13:26:53] on line 160 [13:27:08] Didn't you fix that one already? :/ [13:27:24] you -1'd it;) [13:28:18] Reedy, https://gerrit.wikimedia.org/r/#/c/15559/1/api/ApiMobileView.php [13:28:42] Oh, I see [13:28:43] $pool = new PoolWorkArticleView( $this, $parserOptions, $oldid, $useParserCache ); [13:28:43] $pool->execute(); [13:29:19] Change merged: Reedy; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/15559 [16:46:16] awjr, around? [16:46:25] MaxSem: moostly [16:46:27] *mostly [16:46:28] what's up? [16:47:04] wanted to discuss your patch [16:47:23] MaxSem would you mind waitint til after standup? [16:47:28] sure [16:54:25] wheee https://gerrit.wikimedia.org/r/15905 [17:00:45] hi preilly - feeling better? [17:00:57] standup, anyone? [17:01:17] i'll start a hangout [17:02:04] MaxSem, preilly, philinje, jdlrobson: https://plus.google.com/hangouts/_/3a6168b14ab6de17b261bb5b542b804c7a714eaf [17:08:14] MaxSem im ready to chat about the patch whenever you're ready [17:08:17] btw is your uncle ok? [17:08:58] in hospital [17:09:15] :( im sorry dude [17:09:45] awjr, regarding the patch: [17:10:45] 1) it doesn't contain the schema change [17:11:15] that's true [17:11:31] 2) and the admin_tree populating script [17:11:52] ok i can add those [17:13:17] i can't beileve they still havent made my TS account [17:13:23] :( [17:13:42] MaxSem do you know if there someone i can harass about it [17:13:44] 3) formatter gets passed a query-continue parameter, however this action has no pagination [17:13:56] #wikimedia-toolserver [17:14:02] ty [17:14:14] yeah i wasnt sure about pagination [17:14:20] the admin is DaB. [17:14:28] cool thanks [17:17:38] awjr: more reports of broken sticky cookies - this results in most emails to mobile feedback at the moment [17:17:54] latest one is on android 4.1.1 [17:17:59] ugh [17:18:14] jdlrobson when you get those can you forward them to me? [17:18:23] they are all the same - don't add much [17:18:25] just say it's broken [17:18:27] ok [17:18:37] but they include device info? [17:19:06] 4) use quote() instead of sanitize(), the latter doesn't enclose the statement in quotes [17:19:57] 5) however, in this case you're overescaping as constructs like 'field' => $value are already escaped by select() [17:20:30] (but not $where[] = "field >= $value") [17:20:36] jdlrobson: alternatively, can i somehow receive the feedback emails? [17:20:45] MaxSem doh [17:20:46] talk to philinje [17:20:51] k [17:21:14] philinje: would it be possible for me to receive mobile feedback emails for MobileFrontend? [17:21:16] on a side note somehow my development instance of mediawiki has gonna hella slow [17:21:29] even with memcached it's taking 10s to load a page [17:22:03] jdlrobson, also, after I wrote all the code I sent you, Yuvi said that `country` is not needed due to `adm0` [17:22:17] brr, s/jdlrobson/awjr/ [17:23:00] oh - adm0 = country then? [17:23:20] jdlrobson have you tried turning it off and on again? [17:23:29] yes awjr multiple times [17:23:32] adm0 means what? [17:23:53] highest level administrative zone [17:24:00] ~country [17:24:01] which i guess maps to country [17:24:03] yeah [17:25:41] jdlrobson, from my experience, dev wikis run faster on SQLite;) [17:25:46] blargh ive been getting dns errors from google [17:25:58] i just don't understand how it got so slow :( [17:27:13] $wgProfiler['class'] = 'Profiler'; [17:27:57] (assuming you have debug log enabled) [17:30:36] [18-Jul-2012 10:30:21] PHP Fatal error: Allowed memory size of 52428800 bytes exhausted (tried to allocate 523800 bytes) in /Users/jrobson/Sites/w/includes/profiler/Profiler.php on line 175 mmm [17:32:04] 50 megs? oh come on, it's 2012! :P [17:32:37] lol [17:32:47] is that defined in php.ini? [17:34:15] yes [17:34:23] memory_limit, i think [17:37:14] [18-Jul-2012 10:36:55] PHP Fatal error: Maximum execution time of 30 seconds exceeded in /Users/jrobson/Sites/w/includes/parser/Preprocessor_DOM.php on line 1084 [17:37:22] interesting.. [17:37:58] I think this is a mediawiki problem… are you guys running an up to date mediawiki dev version? [17:40:09] always [17:42:10] mm [17:42:23] is it possible I imported a large article and it's slowed down the whole thing? [17:42:28] I don't know what else I could have done.. [17:46:09] jdlrobson have you made code changes? [17:46:18] no :( [17:46:20] :_( [17:48:13] jdlrobson, have you trie runJobs.php? [17:48:26] it's just the mobile site and larger articles that seems slow [17:49:47] MaxSem: runJobs.php yup and update.ph [18:35:24] New patchset: Kghbln; "Kill de-formal message superseded by it's de fallback message" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/15949 [18:41:58] MaxSem: can you hit this url http://mobile-testing.wmflabs.org/carrier-test/?cid=1010 with Opera Mini [18:47:00] MaxSem|android, http://mobile-testing.wmflabs.org/carrier-test/?cid=1010 [18:48:33] preilly, done [18:48:39] MaxSem: thanks [18:49:07] MaxSem: was this it 1010:213.87.130.126:android [18:49:18] MaxSem: and that was from Opera Mini? [18:49:41] yeah, android browser first, then opera [18:51:13] MaxSem: change the cid for the Opera request [18:52:12] done, 101055 [18:52:25] MaxSem: 101055:213.87.130.126, 217.212.231.216:operamini [18:52:32] MaxSem: okay thanks [19:03:12] MaxSem: im not sure where the best place is to put the sql for creating the admin_tree table nor a script to populate it. the only sql i see in the checkout is in erfgoedbot - is that where i should be doing this? [19:03:24] yup [19:17:48] MaxSem: got a second? [19:17:54] sure [19:18:05] MaxSem: what is the current state of GeoData? [19:19:11] as I said, I'm playing with Sphinx support for it, guess it will make ops happer [19:19:26] *happier [19:22:25] MaxSem: because it's too slow [19:22:36] MaxSem: in MySQL land? [19:22:44] ?? [19:23:43] MaxSem: is it because it's too slow in MySQL directly? [19:24:09] well, at least for me on Windows Sphinx is slower [19:24:26] currently trying on Linux [19:25:50] MaxSem: have you tried it on production with a fuller dataset? [19:26:27] yes, I tested with a real dataset of 700k+ points [19:26:41] MaxSem: but on what box? [19:26:56] but not on production [19:27:11] MaxSem: so on your local machine with 700k+ points [19:27:17] yup [19:27:19] MaxSem: and what was the issue? [19:28:45] with what? [19:29:02] MaxSem: let me try to ask this a different way [19:29:16] MaxSem: what is keeping us from deployment GeoData live to the cluster? [19:29:28] s/deployment/deploying [19:30:40] nothing? but I still would like to complete my today's experiments - in such case, Asher will not need to worry if we will melt his servers [19:31:25] and we could just create the tables and start collecting the data while other prerequisites are being installed [19:31:43] MaxSem: what are other prerequisites? [19:32:53] can we talk on Skype or something? [19:33:15] MaxSem: sure let's talk after lunch [19:33:39] IRC is not good for serious talks, unfortunately [19:50:55] What is the status of https://gerrit.wikimedia.org/r/#/c/14419/ [20:17:03] preilly, I'd say that commit is deadlocked. I see Krinkle's point, but his example of breakage is pretty broken by itself [20:17:31] we served MW:Common.css for a while, and we saw no problems with it [20:35:45] MaxSem should i be concerned right now about bad data in the monuments_all table? [20:36:21] what kind of bad data? [20:36:31] as im working on the queries to populate the admin_tree table im finding inconsistent data for some monuments. for instance, im finding a bunch of monuments that have adm2 data but nothing in the adm1 field [20:36:50] also [20:37:00] semi-unrelated, adm1 does not always = country [20:37:23] er wait sorry nevermind that last point [20:37:49] yeah. there are such funky countries as it-88 [20:38:31] more concerning is lack of adm1 data when there is adm2 data [20:39:14] adm2 but no adm1 is a bug. per multichill, you should simply ignore such rows by stopping tree creation on NUI [20:39:21] NULL values [20:41:05] same appears true for adm3 - where adm3 will have data but not neessarily have adm2 data [20:41:06] :( [20:41:14] these are nasty queries. [21:24:50] preilly, ready to continue our talk? [21:36:17] New review: preilly; "This doesn't seem right.... why are you removing the key?" [mediawiki/extensions/MobileFrontend] (master); V: 0 C: -1; - https://gerrit.wikimedia.org/r/15949 [21:42:48] New review: Kghbln; "de has been changed in a way that it is now usable for de-formal, too. Thus de fits for both cases d..." [mediawiki/extensions/MobileFrontend] (master) C: 1; - https://gerrit.wikimedia.org/r/15949 [21:49:00] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/15949