[02:45:42] <MichaelShavlovsk>	 hi, I cannot understand behavior of $dbw->selectRow, everything should be simple, but in it returns empty row while there is some data in the DB to return.
[02:46:42] <marktraceur>	 MichaelShavlovsk: Can you be more verbose? Pastebin?
[02:46:45] <marktraceur>	 !pastebin | MichaelShavlovsk
[02:46:45] <wm-bot>	 MichaelShavlovsk: Please do not paste more than 2-3 lines of text into the channel as it disrupts the flow of conversation. When sharing multiple lines of code, please use a pastebin such as <http://dpaste.com/> or <http://p.defau.lt/> and post a link to your paste in the channel.
[02:47:13] <MichaelShavlovsk>	 sure, one sec
[02:49:52] <MichaelShavlovsk>	 marktraceur: http://dpaste.com/806618/
[02:50:04] <MichaelShavlovsk>	 maybe I should paste more
[02:50:47] <MichaelShavlovsk>	 If I do select by my hands then there is data in the table
[02:53:57] <MichaelShavlovsk>	 marktraceur: it is very similar to Job.php class http://svn.wikimedia.org/doc/Job_8php_source.html
[02:54:15] <MichaelShavlovsk>	 lines 164 - 166
[02:55:52] <MichaelShavlovsk>	 marktraceur:why on line 166 they check whether $row->minjob is null or not, if there no data then $row should be false, is not it?
[03:01:13] <MichaelShavlovsk>	 marktraceur: in description of selectRow: "If the query returns no rows, false is returned", but in my case $row is not a false and $row->minid is null ( in Job.php they even check for this condition, why? ).
[03:05:14] <marktraceur>	 MichaelShavlovsk: Wrong order
[03:05:19] <marktraceur>	 I think
[03:05:27] <marktraceur>	 No, it's not
[03:05:28] <marktraceur>	 Sorry
[03:05:54] <marktraceur>	 MichaelShavlovsk: Yes, $row is false and $row->(member) is null
[03:05:56] <marktraceur>	 Blame PHP
[03:06:10] <MichaelShavlovsk>	 why, I don't understand
[03:06:16] <marktraceur>	 MichaelShavlovsk: Try #php
[03:06:42] <MichaelShavlovsk>	 where? in cli
[03:07:13] <MichaelShavlovsk>	 no no, $row is not false
[03:07:17] <MichaelShavlovsk>	 this is the problem
[03:08:33] <marktraceur>	 MichaelShavlovsk: I can see, that would be a problem
[03:12:22] <marktraceur>	 MichaelShavlovsk: We're worried MIN is doing something funny, can you maybe try without that? (it won't do what you want, but may confirm a bug)
[03:13:13] <MichaelShavlovsk>	 marktraceur: okay
[03:17:25] <marktraceur>	 MichaelShavlovsk: The other thing you could try is sele
[03:17:31] <marktraceur>	 Er, sorry
[03:18:04] <marktraceur>	 MichaelShavlovsk: You could try 'minid' => '4', instead of 'minid' => 'MIN(jq_article_id)' , so that you get back a value for that parameter
[03:18:19] <MichaelShavlovsk>	 okay
[03:18:23] * marktraceur  is actually not marktraceur at the moment (since the typo thing) but instead a friend of his hanging out at his place

[03:18:32] <marktraceur>	 But hence the 'we'
[03:31:13] <lexein>	 Psst, hey, wanna get marktraceur blocked or banned?  Do a bunch of edits while he's not at the keyboard!  We won't tell!
[03:31:26] <lexein>	 Funny!
[03:32:18] <MichaelShavlovsk>	 marktraceur: well, if I do 'minid' => '4' I have the same result
[03:35:46] <marktraceur>	 <friend>What SQL is it executing? Is there a way to log that? That's the real question. (Then Mark says...</friend>
[03:35:49] <marktraceur>	 !sqllog
[03:35:49] <wm-bot>	 To log all SQL queries to a file, set $wgDebugDumpSql = true; and set $wgDebugLogFile to a filename
[03:36:04] <marktraceur>	 <friend>Yeah, do that, and see what the SQL it's actually executing is.</friend>
[03:37:45] <MichaelShavlovsk>	 marktraceur: yes I am loggin it, what exactly you are interested in? I got white spaces on this line 		wfDebugLog( 'blamemaps', strval(($row->minid)) );
[03:38:08] <marktraceur>	 "white spaces on this line"?
[03:38:27] <marktraceur>	 What I'm interested in knowing first is: what is the SQL that MediaWiki is executing corresponding to the query you wrote up at http://dpaste.com/806618/ ?
[03:38:30] <marktraceur>	 Is it the SQL you expect?
[03:38:50] <marktraceur>	 If not, then the problem has to do with how we're expressing the query.
[03:39:08] <marktraceur>	 If so, then it has something to do with how MW is returning the results through its internal database API.
[03:39:29] <marktraceur>	 Either way, I'd be interested in seeing the results of the same SQL query, if you execute it using the 'mysql' command line tool.
[03:39:47] <MichaelShavlovsk>	 what this line should return ? $row = $dbr->selectRow( self::jobQueueTable, array( 'minid' => '2','maxid' => '2h' ),'1=1',__METHOD__ );
[03:40:25] <MichaelShavlovsk>	 log shows that $row->minid is empty
[03:41:12] <marktraceur>	 Since I'm still the pseudo-random friend, I don't know; but I would find out the raw SQL it's executing, and see what 'mysql' (the CLI tool) thinks when you execute the same SQL.
[03:41:47] <marktraceur>	 unlike marktraceur (who is busy at the moment) I might not have quick answers, but I can try to be helpful and direct us both toward more understanding.
[03:42:12] <MichaelShavlovsk>	 marktraceur: I don't know what exactly sql querry correspond to that
[03:42:37] <marktraceur>	 I think it should be findable in the log
[03:43:22] <MichaelShavlovsk>	 yes, I will log just before executing querry, one sec
[03:46:45] <MichaelShavlovsk>	 marktraceur: yes sql querry works fine it is "SELECT  2,2  FROM `jobqueue`  WHERE 1 = 1  LIMIT 1"
[03:46:57] <MichaelShavlovsk>	 and it returns 2, 2
[13:52:54] <gerrit-wm>	 Change merged: Diederik; [analytics/webstatscollector] (master) - https://gerrit.wikimedia.org/r/24551
[14:39:35] <gerrit-wm>	 New patchset: Diederik; "Added automagically build numbering using git" [analytics/webstatscollector] (master) - https://gerrit.wikimedia.org/r/25502
[14:40:38] <gerrit-wm>	 New review: Diederik; "Ok." [analytics/webstatscollector] (master); V: 1 C: 2;  - https://gerrit.wikimedia.org/r/25502
[14:40:39] <gerrit-wm>	 Change merged: Diederik; [analytics/webstatscollector] (master) - https://gerrit.wikimedia.org/r/25502
[15:44:45] <chrismcmahon>	 hi zeljkof!
[15:44:55] <zeljkof>	 hi chrismcmahon
[16:12:38] <marktraceur>	 zeljkof: Oh hi there! How's it going?
[16:15:03] <zeljkof>	 hi marktraceur
[16:15:31] <chrismcmahon>	 zeljkof: marktraceur did some work on a watir test I had for our UploadWizard feature
[16:17:23] <marktraceur>	 I'm also getting pretty happy with phpunit and qunit from other projects
[16:18:16] <chrismcmahon>	 \0/ marktraceur!  I'll be asking about that stuff pretty soon now.
[16:18:28] <^demon>	 our unit tests are so much faster on a ssd.
[16:18:29] <^demon>	 :p
[16:19:10] <Reedy>	 lols
[16:20:02] <marktraceur>	 chrismcmahon: hashar knows a lot about the phpunit side, Krinkle can help more with qunit....if they're not around I can do my best to help :)
[17:14:49] <brion>	 any thoughts or objections on my responsive images code so far? https://gerrit.wikimedia.org/r/#/c/24115/
[17:15:12] <brion>	 mainly targeted at mobile (another change set actually enables it on MobileFrontend)
[17:15:17] <brion>	 but also at retina display users ;)
[17:16:44] <brion>	 there's a bit of a hack in how the srcset entries get set from Linker::makeImageLink and Linker::makeThumbLink but that can be pushed down to lower levels later, as long as the technique is sound
[17:16:58] <^demon>	 brion: I haven't tried them yet. I should.
[17:17:22] <brion>	 yes, mr retina macbook pro
[17:17:27] <brion>	 you should! :)
[17:22:12] <^demon>	 Seems to work fine. Are there any areas I should be looking for *really* noticeable improvement?
[17:23:03] <brion>	 photos should look sharper, svg & png diagrams cleaner
[17:23:09] <brion>	 in the current version not everything will improve though
[17:23:29] <brion>	 <gallery> items, and the large-size zoom on the image page don't get the extra srcset values right now
[17:23:42] <^demon>	 Gotcha. Pictures do seem a bit sharper.
[17:23:52] <^demon>	 Lemme upload a higher-def photo to my local wiki and compare.
[17:23:53] <brion>	 i could special-case them, or i could go whole-hog and retool MediaHandlers
[17:24:14] <brion>	 another question is how best to handle InstantCommons
[17:24:37] <brion>	 right now it does three separate API requests to Commons to fetch the 1.0, 1.5, and 2.0x image thumb URLs
[17:24:40] <brion>	 which is a bit slow
[17:25:02] <apergos>	 poor swift
[17:25:10] <apergos>	 swift isn't...
[17:25:16] <brion>	 if we built multi-resolution handling in at a lower level, we could perhaps expose a 'thumbnailsrcset' in the imageinfo so we need only make one request
[17:25:52] <^demon>	 InstantCommons was never more than a dirty hack ;-)
[17:26:00] <brion>	 it's a beautiful hack :)
[17:26:03] <^demon>	 Indeed.
[17:26:04] <apergos>	 and now we'll never be rid of it
[17:26:04] <brion>	 but far from perfect
[17:26:16] <brion>	 if we could consolidate the api requests that'd be a big help
[17:26:30] <^demon>	 apergos: Early versions contained the comment "Hacky and inefficient, do not use except for testing" :)
[17:26:35] <apergos>	 hahaha
[17:26:45] <brion>	 something about a warning that it may kill your dog
[17:26:47] <apergos>	 yeah well you know how that goes... about as well as "temporary until X"
[17:28:34] <^demon>	 brion: It's been awhile since I played with the ForeignApiRepo code, but it was dead simple so would be easy to review fixes for :)
[17:28:44] <brion>	 whee
[17:30:09] <apergos>	 so if one api request gets all three thumb sizes, we have 1/3 as less ulimit time (if there are limits set for api requests) for each of those thumbs to get done...
[17:31:27] <apergos>	 wonder what a good workaround will be
[17:31:42] <^demon>	 Back to the topic of "stop providing all thumbs in all sizes all the time?"
[17:31:53] <^demon>	 And preferring "standard" sizes.
[17:32:17] <apergos>	 yeah so
[17:32:39] <apergos>	 I haven't got around to writing up my summary of that thread yet (I shall soon though) but
[17:32:53] <brion>	 apergos: well for sites like ours, the actual rendering is delayed until use
[17:32:55] <apergos>	 in an ideal world we don't care about cost of space so we just store everything forever
[17:33:14] <brion>	 but it does potentially increase our thumb storage...
[17:33:33] <^demon>	 ForeignApiRepo could maybe normalize sizes a bit before requesting them? So if you request a 198px image, the repo returns a 200px one and your browser can scale the rest.
[17:33:44] <apergos>	 also in an ideal world we could decide we don't care about cpu cost so instead of storing anything we have piles and piles of scalers and plenty of swift backend
[17:33:45] <^demon>	 That would be pretty easy to do within the ForeignApiRepo itself and not impact anything else.
[17:34:24] <apergos>	 since we don't live in either of those worlds we are going to have to compromise, which prolly means some on demand and some fixed sizes in an lru cache
[17:34:41] <apergos>	 but that won't happen over night
[17:34:56] <apergos>	 however, normalizing sizes is a good approach, bear in mind that...
[17:35:27] <apergos>	 a reason people request the weird sizes is that the one they have at the standard size has something wrong with it and for whatever reason they can't get the purge working
[17:36:35] <^demon>	 So fix purge ;-)
[17:37:19] <apergos>	 yeahright
[17:37:57] <apergos>	 but a lot of sizes are just "I can get yet another weird size, oh let's do"
[17:38:31] <^demon>	 brion: As a first stab at it, I think it's good. Definitely some visible improvements.
[17:38:37] <^demon>	 No errors/regressions that I see.
[17:39:13] <brion>	 yay!
[17:39:17] <brion>	 thanks for checking it over
[17:39:48] <^demon>	 yw
[17:57:33] <Reedy>	 reedy@fenari:~$ mwscript cleanupUploadStash.php commonswiki
[17:57:33] <Reedy>	 Getting list of files to clean up...
[17:57:33] <Reedy>	 Removing 52945 file(s)...
[17:57:33] <Reedy>	 Set $wgShowExceptionDetails = true; in LocalSettings.php to show detailed debugging information.
[17:57:34] <Reedy>	 Fuckyeah
[17:59:27] <Nemo_bis>	 O_o
[18:02:39] <gerrit-wm>	 New patchset: Stefan.petrea; "enhancements to debianize.sh" [analytics/webstatscollector] (master) - https://gerrit.wikimedia.org/r/25522
[18:02:42] <Reedy>	 Oh FFS
[18:04:07] <Reedy>	 No aaron
[18:04:26] <csteipp>	 Anyone know how to push a patch into gerrit, that relies on another patch? I've seen it done, but not sure how to tell gerrit that's what I want..
[18:07:46] <gerrit-wm>	 New review: Diederik; "Ok." [analytics/webstatscollector] (master); V: 1 C: 2;  - https://gerrit.wikimedia.org/r/25522
[18:07:46] <gerrit-wm>	 Change merged: Diederik; [analytics/webstatscollector] (master) - https://gerrit.wikimedia.org/r/25522
[18:19:11] <Krenair>	 csteipp, I thought you download the other patch and commited after it
[18:19:47] <csteipp>	 Sortof... I found some directions on the git workflow that I just followed... and it seems to have worked
[18:21:22] <Reedy>	 If you commit them both to the same feature branch, IIRC gerrit makes them dependant
[18:24:00] <csteipp>	 http://www.mediawiki.org/wiki/Git/Workflow#Create_a_dependency seems to have worked ok... Hopefully gerrit plays nice when everything merges in.
[18:27:53] <Reedy>	 hmm, looks like it just does it in 2 steps rebasing onto the branch
[18:28:00] <Reedy>	 if it shows up in gerrit, it worked
[18:33:04] <gerrit-wm>	 New patchset: Diederik; "Fix typo" [analytics/webstatscollector] (master) - https://gerrit.wikimedia.org/r/25528
[18:33:23] <gerrit-wm>	 Change merged: Diederik; [analytics/webstatscollector] (master) - https://gerrit.wikimedia.org/r/25528
[19:23:15] <Nemo_bis>	 https://gerrit.wikimedia.org/r/#/c/25527/ : can the unit test output to check include HTML entities? or are special whitespaces ignored?
[19:25:02] <^demon>	 It's just doing a direct string comparison, does $in == $out.
[19:25:18] <^demon>	 Just update the tests for the expected output.
[19:25:33] <Nemo_bis>	 ^demon: what if it includes &nbsp; aka &#160; ?
[19:26:22] <^demon>	 The test doesn't know any better...
[19:26:43] <^demon>	 You could maybe add some normalization to the test case, str_replace() everything to ' ' before comparison.
[19:27:11] <Nemo_bis>	 oh well, I don't need people to think I'm a proper coder, I'll just let jenkins yell at me and try again
[19:49:32] <gerrit-wm>	 New patchset: Stefan.petrea; "arguments, and fixing gcc warnings" [analytics/webstatscollector] (master) - https://gerrit.wikimedia.org/r/25542
[19:53:25] <gerrit-wm>	 New review: Diederik; "Ok." [analytics/webstatscollector] (master); V: 1 C: 2;  - https://gerrit.wikimedia.org/r/25542
[19:53:25] <gerrit-wm>	 Change merged: Diederik; [analytics/webstatscollector] (master) - https://gerrit.wikimedia.org/r/25542
[20:40:56] <brion>	 http://www.mediawiki.org/wiki/Git seems to be missing info on how to check out the extensions.
[20:42:48] <brion>	 fine i'll do this the hard way :P
[20:42:58] <Reedy>	 I reported this a while ago ;) https://bugzilla.wikimedia.org/show_bug.cgi?id=37977
[20:43:01] <brion>	  git remote show origin
[20:43:07] <brion>	   Fetch URL: https://gerrit.wikimedia.org/r/p/mediawiki/extensions.git
[20:43:09] <brion>	 ta-daa
[20:45:54] <brion>	 https://www.mediawiki.org/w/index.php?title=Git&diff=588079&oldid=586833
[20:46:23] <Reedy>	 git submodule update --init works as one command :D
[20:47:21] <brion>	 \o/
[21:03:40] <gerrit-wm>	 New patchset: Stefan.petrea; "filter bot switch, test switch renamed to debug" [analytics/webstatscollector] (master) - https://gerrit.wikimedia.org/r/25585
[21:05:54] <gerrit-wm>	 New review: Diederik; "Ok." [analytics/webstatscollector] (master); V: 1 C: 2;  - https://gerrit.wikimedia.org/r/25585
[21:05:55] <gerrit-wm>	 Change merged: Diederik; [analytics/webstatscollector] (master) - https://gerrit.wikimedia.org/r/25585
[21:08:09] <Nemo_bis>	 Reedy: so is that consistent with https://www.mediawiki.org/wiki/Download_from_Git#Download_all_extensions now? :)
[21:11:28] <brion>	 i'll be honest
[21:11:34] <brion>	 i always find the directions by googling 'mediawiki git'
[21:12:13] <brion>	 first hit is the 'git' page, so it should probably more prominently link to the actual download directions if for some reason they'
[21:12:15] <Reedy>	 Who needs mediawiki search?
[21:12:17] <brion>	 re on a separate page
[21:12:58] <apergos1>	 :-D  I also find the stuff I need by goggling (usually mediawiki gerrit workflow)
[21:13:05] <apergos1>	 er
[21:13:07] <apergos1>	 googling!!
[21:14:08] <brion>	 bah Fatal error: Maximum function nesting level of '100' reached, aborting! in /srv/com.leuksman/mw/includes/parser/Preprocessor_DOM.php on line 985
[21:14:15] <brion>	 rendering copy of 'San Francisco'
[21:14:18] <brion>	 sighhhhh
[21:50:20] <gwicke>	 brion: ahhh- we already opened a bug about that article in Parsoid: https://bugzilla.wikimedia.org/show_bug.cgi?id=40577
[21:56:26] <brion>	 fun :D
[23:14:41] <awjr>	 ori-l: ping
[23:14:55] <ori-l>	 awjr: hey
[23:14:57] <awjr>	 yo!
[23:15:17] <awjr>	 maryana told me i should talk to you about how you guys are gathering stats about events on the e3 team
[23:15:45] <awjr>	 i know we chatted some about this kind of thing in regards to the WLM mobile app a few weeks back - is that the same thing?
[23:16:09] <jorm>	 when did the color change for the "updated since last visit" text?
[23:16:18] <jorm>	 that needs to be reverted.  it's not vivid enough, the green now.
[23:16:23] <awjr>	 we need a way to get stats about various events in MobileFrontend but we can't use ClickTracking
[23:16:33] <kaldari>	 rmoen: got time for a question about the new edit interface?
[23:16:42] <awjr>	 ori-l ^^
[23:16:50] <ori-l>	 awjr: it's the same thing.. i'll pm