[00:05:34] <^demon>	zakg: But we do optimistically release every quarter :)
[00:05:45] <^demon>	Practically, it just doesn't happen ;-)
[00:06:17] <zakg>	^demon: RIght, which is why I removed the note. :)
[00:06:46] <^demon>	Actually, I don't think we ever got a quarterly release out.
[00:07:02] <^demon>	I think we just had like 3 decently timed releases in a row, and somehow that turned into "we release quarterly"
[00:07:16] <RoanKattouw>	Like I said
[00:07:24] <RoanKattouw>	There was never a calendar year with 4 releases
[00:07:29] <RoanKattouw>	Highest was 3
[00:07:44] <Shirley>	Are the mailing lists being slow?
[00:07:57] <^demon>	Been fine for me.
[00:08:06] <Shirley>	Hmm, posted to wikitech-l.
[00:08:07] <zakg>	Perhaps we could just redefine the length of the year... a kilo-day year would help a lot. ;-)
[00:08:22] <Shirley>	Don't see it in the archives or my inbox though. Oh well.
[00:09:06] <RoanKattouw>	I didn't get it either
[00:09:27] <RoanKattouw>	zakg: That'd mean I'm a 6yo, no thanks ;)
[00:09:39] <^demon>	Well seeing as it's already been decided we're branching in 4 days, has anyone picked a branch point yet?
[00:10:06] <RoanKattouw>	I was hoping you and I could work that out
[00:10:14] <RoanKattouw>	Something close to where HEAD's gonna be on Tuesday I guess
[00:10:27] <RoanKattouw>	I see no reason to branch significantly behind HEAD
[00:10:51] <^demon>	Hmm, let's list all the problems with HEAD.
[00:11:11] <Shirley>	That would be the fixme list, right?
[00:11:18] <^demon>	+a billion other unreviewed things.
[00:11:22] <Shirley>	There's all sorts of weird shit in trunk.
[00:11:24] <RoanKattouw>	We have some of that on that Etherpad-turned-wikipage
[00:11:43] <^demon>	I saw, and it's incomplete.
[00:11:47] <RoanKattouw>	Of coures
[00:12:06] <^demon>	 /maintenance/tests/phpunit is a disaster.
[00:12:10] <Reedy>	haha
[00:12:22] <RoanKattouw>	^demon: Yeah and that really matters for deploying ^^
[00:13:00] <Reedy>	"OMG THE UNIT TESTS DISTRIBUTED WITH 1.17 DON'T PASS" BLOCKER HIGHEST
[00:13:01] <RoanKattouw>	But we should look at where the unreviewed revs for WMF code (trunk + deployed exts) are and see if they're concentrated anywhere
[00:13:25] <^demon>	I would like to see http://www.mediawiki.org/wiki/Special:Code/MediaWiki/status/new?path=/trunk/phase3 much shorter.
[00:13:44] <^demon>	Granted I've been busy as crap with school the last week and a half or so, so my time's been limited at best.
[00:13:46] <Reedy>	Roans reviewing quite a lot...
[00:14:04] <RoanKattouw>	I just caught up, yay
[00:14:20] <^demon>	Nobody reviews my code.
[00:14:22] <RoanKattouw>	^demon: Yeah we need to get as many people on that as we can
[00:14:29] <TrevorParscal>	^demon: hi
[00:14:39] <RoanKattouw>	And I'll be happy to prioritize reviewing your code because it's very likely few others will
[00:15:09] <TrevorParscal>	^demon: did you weigh in already on preference to branch the 1.17 release or do a feature freeze on trunk? Roan preferred the branch approach
[00:15:58] <^demon>	I've advocated feature freezes in the past, but Tim and Roan have convinced me otherwise over time. I'd prefer to branch like we usually do.
[00:16:29] <TrevorParscal>	that was my pref as well - but it's more important that you and roan (and tim) are happy with the approach than anyone else at this point
[00:16:32] <^demon>	RoanKattouw: select count(*) from code_rev where cr_author = 'demon' and cr_status = 'new';
[00:16:33] <TrevorParscal>	since you guys are doing the work
[00:16:35] <^demon>	125.
[00:16:42] <RoanKattouw>	Eek
[00:16:51] <RoanKattouw>	OK well I reviewed almost a hundred revisions today
[00:16:53] <TrevorParscal>	^demon: I will review you CSS/JS and front-end PHP code :)
[00:16:58] <^demon>	Yeah, I'm approx 1/7th of the unreviewed total :)
[00:17:06] <RoanKattouw>	TrevorParscal: Yay
[00:17:06] <^demon>	To trunk
[00:17:09] <TrevorParscal>	all 3 revs of it... (random guess)
[00:20:01] <^demon>	RoanKattouw: Can we review + sync r75939?
[00:20:14] <^demon>	It'd make reviewing by author easier :)
[00:20:24] <RoanKattouw>	Sure gimme a minute
[00:21:19] <^demon>	That whole query UI needs redoing, but I don't mind using query params in the meantime
[00:29:47] <TrevorParscal>	$.params is cool
[00:29:55] <TrevorParscal>	but it doesn't comprehend anchors
[00:30:07] <TrevorParscal>	er.. well it just genreates stuff
[00:30:11] <TrevorParscal>	I'm thinking about something else
[00:30:15] <TrevorParscal>	it's cool
[01:00:29] <Reedy>	lol TrevorParscal "Mailing list? I want to join!"
[01:00:37] <TrevorParscal>	ha ha
[01:00:43] *TrevorParscal 	stalks Roan
[01:00:59] <RoanKattouw>	You made that post when it was still leaking IPs
[01:01:07] <RoanKattouw>	Fortunately you were at the office, judging from the IP it leaked
[01:01:20] <TrevorParscal>	I can haz stalkerz?
[01:03:37] <Reedy>	TrevorParscal, don't your kids stalk you?
[01:03:42] <Reedy>	New - Total number of results: 1655
[01:04:28] *TrevorParscal 	plays Nirvana in the office
[01:07:30] <TrevorParscal>	Reedy: I think so, yes
[01:40:01] <werdna>	omg wikileaks!@
[01:41:32] <Reedy>	1599 new
[01:41:33] <Reedy>	:)
[01:42:32] <Reedy>	Bedtime
[01:42:34] <Reedy>	Night all
[17:33:47] <vvv>	May anyone check if lookup of revs by specific tag is broken in trunk?
[17:34:29] <Reedy>	sure
[17:34:44] <Reedy>	vvv, you mean see if going to /tag, and clicking a tag finds anything?
[17:34:56] <Reedy>	A database query syntax error has occurred. This may indicate a bug in the software. The last attempted database query was:
[17:34:57] <Reedy>	(SQL query hidden)
[17:34:57] <Reedy>	from within function "IndexPager::reallyDoQuery (SvnRevTagTablePager)". Database returned error "1054: Unknown column 'cr_repo_id' in 'on clause' (10.0.6.49)".
[17:35:04] <Reedy>	They're broken on 1.16wmf4 :/
[17:36:03] <Reedy>	vvv, same error
[17:37:58] <Reedy>	SvnRevTagTablePager hasn't been changed recently
[17:38:15] <vvv>	I have the same
[17:38:30] <Reedy>	It's a change in SvnRevTablePager
[17:38:35] <Reedy>	But AFAIK, roan didn't merge much..
[17:38:56] <Reedy>	http://www.mediawiki.org/w/index.php?title=Special:Code/MediaWiki/author/catrope&path=/branches/wmf/1.16wmf4/extensions/CodeReview
[17:39:23] <Reedy>	I'm sure they worked recently
[17:39:27] <vvv>	They are not broken in live version
[17:39:57] <Reedy>	http://www.mediawiki.org/wiki/Special:Code/MediaWiki/tag/ok
[17:39:59] <Reedy>	That errors for me
[17:40:11] <vvv>	`code_rev`,`code_tags` LEFT JOIN `code_comment` ON ((cc_repo_id = cr_repo_id
[17:40:24] <vvv>	The problem seems to be here
[17:41:22] <Reedy>	Hmm
[17:41:30] <Reedy>	It seems to intermittantly work on MW.org :/
[17:41:39] <vvv>	Does the same for me
[17:41:50] <Reedy>	Seems like one of the DB's is out of sync
[17:42:15] <vvv>	srv196 is one
[17:42:25] <Reedy>	10.0.6.49
[17:42:40] <vvv>	Or we should understand whether it is DB or MW
[17:42:50] <Reedy>	yeah
[17:42:57] <Reedy>	though, it being intermittant is strange
[17:57:28] <Reedy>	vvv, I'm wondering if there's some caching
[17:57:33] <Reedy>	And that's why it's intermittant on mw.org
[17:57:39] <vvv>	Of what?
[17:58:01] <Reedy>	the tag pages
[17:58:03] *vvv 	is trying to find the revision where it was introduced using svn up
[17:58:51] <Reedy>	I'm guessing it's against 				'join_conds' => array(
[17:58:51] <Reedy>						'code_comment' => array( 'LEFT JOIN', 'cc_repo_id = cr_repo_id AND cc_rev_id = cr_id' ),
[17:58:51] <Reedy>					)
[17:58:55] <Reedy>	but that's not changed recently
[17:59:01] <vvv>	Yeah
[17:59:09] <vvv>	It's broken for me on r60000
[17:59:35] <Reedy>	hmm
[18:00:06] <Reedy>	That seems weird
[18:00:39] <Reedy>	We're synced upto ~75000 on most revisions
[18:01:30] <vvv>	And on r50000
[18:02:46] <Reedy>	:/
[18:02:51] <Reedy>	I don't believe this
[18:02:56] <Reedy>	vvv, using trunk MW?
[18:03:03] <vvv>	Yep
[18:03:08] <vvv>	Means that something in core changed
[18:03:13] <Reedy>	Yeah
[18:03:13] <Reedy>	But
[18:03:21] <Reedy>	it surely wouldn't affect 1.16wmf4, surely?
[18:03:59] <vvv>	As we see, it affects 1.16wmf4
[18:04:21] <Reedy>	Yeah :/
[18:04:31] <Reedy>	Has anything else been done to both recently..
[18:13:56] <vvv>	Reedy: seems like 10.0.6.49 is the only DB to trigger that error
[18:19:21] <Reedy>	vvv, poking Ariel, seems to be the only tech staff about
[18:21:29] <vvv>	Reedy: hm, what about MySQL versions?
[18:22:29] <Reedy>	reedy@ubuntu64-esxi:~/mediawiki/trunk$ mysql --version
[18:22:29] <Reedy>	mysql  Ver 14.14 Distrib 5.1.49, for debian-linux-gnu (x86_64) using readline 6.1
[18:23:06] <vvv>	Yeah
[18:23:26] <vvv>	Admin log shows that s3 was switched to a new 5.1 host yesterday
[18:23:57] <Reedy>	<apergos> db39  this one is now 5.1, I think it was not earlier
[18:25:14] <Reedy>	<apergos> yes all the errors seem to be coming via db39 so that looks like the issue#
[18:25:42] <vvv>	May he check what was previous server and MySQL version?
[18:27:13] <Reedy>	vvv, AHA!
[18:27:15] <Reedy>	https://bugzilla.wikimedia.org/show_bug.cgi?id=24166
[18:27:19] <Reedy>	http://www.mediawiki.org/wiki/Special:Code/MediaWiki/72116
[18:27:25] <Reedy>	I fixed something similar in the core API a while back
[18:28:50] <Reedy>	vvv, probably v4
[18:28:56] <vvv>	Huh, that means I was right about the reason
[18:29:24] <Reedy>	SELECT cr_id,cr_repo_id,cr_status,COUNT(DISTINCT cc_id) AS comments,cr_path,cr_message,cr_author,cr_timestamp FROM `code_tags`,`code_rev` LEFT JOIN `code_comment` ON ((cc_repo_id = cr_repo_id AND cc_rev_id = cr_id)) WHERE cr_repo_id = '3' AND (cr_repo_id=ct_repo_id) AND (cr_id=ct_rev_id) AND ct_tag = 'api' GROUP BY cr_id ORDER BY cr_id DESC LIMIT 51
[18:29:28] <Reedy>	that works
[18:29:32] <Reedy>	Switch the 2 from tables
[18:29:34] <Reedy>	Job done
[18:30:10] <Reedy>			$info['tables'][] = 'code_tags';
[18:30:21] <Reedy>	vvv, so, we need to get it to insert that earlier in the array
[18:31:18] <Reedy>	$info['tables'] = array_merge(  array( 'code_tags' ), $info['tables'] );
[18:31:23] <vvv>	Heh
[18:31:23] <vvv>	Reedy: that would be very fragile
[18:31:31] <vvv>	Hm, that would not be
[18:32:25] <Reedy>	Yup, fixed
[18:32:52] <Reedy>	vvv, mind if I commit?
[18:32:58] <vvv>	I don't
[18:33:01] <vvv>	Commit it
[18:33:12] <vvv>	Yet those MySQL regressions make me sad
[18:33:45] <Reedy>	:(
[18:37:11] <vvv>	Reedy: you should add commit about why it should not be changed back in some "code cleanup"
[18:37:15] <vvv>	*comment
[18:37:26] <Reedy>	That's a good point
[18:37:33] <Reedy>	I did for http://www.mediawiki.org/wiki/Special:Code/MediaWiki/72116
[18:37:38] <Reedy>	OH
[18:37:40] <Reedy>	^demon, :D
[18:37:49] <Reedy>	Fancy scapping me a revision out please?
[18:37:53] *^demon 	yawns
[18:38:02] <Reedy>	http://www.mediawiki.org/wiki/Special:Code/MediaWiki/72116
[18:38:24] <^demon>	Can you at least give me 5 minutes to finish waking up?
[18:38:24] <^demon>	:p
[18:38:29] <Reedy>	Pfft
[18:38:34] <Reedy>	It's after lunchtime for you ;)
[18:38:36] <^demon>	Eyes blurry, haven't had a cigarette, etc etc.
[18:38:49] <Reedy>	Sure, it's not overly urgent :)
[18:38:59] <^demon>	Yes, it's after lunch. It's also a saturday and the first time I've gotten to sleep in in weeks :p
[18:39:05] <Reedy>	haha
[18:45:26] <^demon>	Whoops, timed out fenari connection won't do me any good :p
[18:47:33] <Reedy>	Seemingly Total number of results: 77,378
[18:47:34] <Reedy>	isn't right for number of things tagged "lol" on my local install :D
[19:23:10] <vvv>	Reedy: oh, and that query filesorts
[19:23:34] <Reedy>	aha, cheers :)
[19:24:36] <Reedy>	vvv, logged on the bug :)
[19:27:02] <vvv>	Reedy: what you gonna do about them?
[19:27:31] <Reedy>	vvv, the intention is when Roan comes over to mine on the w/c 20th december, we've got the querypages API bug to do
[19:27:51] <Reedy>	Then we're gonna look at seeing if we can put a few extra indexes in to improve the situtation for the CR queries
[20:18:12] <vvv>	Reedy: the generic problem is: there is code_something with cs_repo_id, cs_rev_id and cs_value; the only index is primary key on all three columns in the order listed. There is code_rev which is joined with code_something by cr_repo_id and cr_id.
[20:18:21] <vvv>	The query on code_something filesorts