[00:03:31] Krenair: I am doing a useless code review, mostly small doc stuff [00:03:47] ok [00:04:53] (03CR) 10Tim Starling: [C: 04-1] "I filed bug 52253 for the issue with protocol-relative URLs. I assume removing tags would cause the same kinds of problems. Many HT" [core] - 10https://gerrit.wikimedia.org/r/70036 (owner: 10Krinkle) [00:04:54] (03PS1) 10MarkAHershberger: Install without XML support [core] - 10https://gerrit.wikimedia.org/r/76663 [00:06:16] (03PS11) 10Kaldari: Adding hook handler for creating human-friendly relative timestamp [extensions/cldr] - 10https://gerrit.wikimedia.org/r/45499 [00:08:05] (03CR) 10MZMcBride: "(1 comment)" [extensions/MassMessage] - 10https://gerrit.wikimedia.org/r/76647 (owner: 10Legoktm) [00:12:21] (03CR) 10MZMcBride: "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/76643 (owner: 10Ori.livneh) [00:12:22] (03CR) 10Legoktm: "(1 comment)" [extensions/MassMessage] - 10https://gerrit.wikimedia.org/r/76647 (owner: 10Legoktm) [00:13:59] (03CR) 10MZMcBride: "The README at is no longer using MediaWiki markup." [core] - 10https://gerrit.wikimedia.org/r/76643 (owner: 10Ori.livneh) [00:15:02] (03PS2) 10MarkAHershberger: Install without XML support [core] - 10https://gerrit.wikimedia.org/r/76663 [00:16:09] (03CR) 10jenkins-bot: [V: 04-1] Install without XML support [core] - 10https://gerrit.wikimedia.org/r/76663 (owner: 10MarkAHershberger) [00:16:25] (03CR) 10Matmarex: "I think a convention to use .ser extension exists, we have a bunch of such files in ./serialized/." [core] - 10https://gerrit.wikimedia.org/r/76663 (owner: 10MarkAHershberger) [00:21:43] (03CR) 10MarkAHershberger: "thanks. I didn't realize that." [core] - 10https://gerrit.wikimedia.org/r/76663 (owner: 10MarkAHershberger) [00:24:20] (03CR) 10Mwalker: [C: 031] "(1 comment)" [extensions/BookManagerv2] - 10https://gerrit.wikimedia.org/r/76644 (owner: 10Mollywhite) [00:24:40] Elsie: The table should have the id on the ? Or one of the s? [00:24:51] The tr. [00:24:53] legoktm: ^ [00:25:04] Ok, submitting a core patch [00:26:09] (03CR) 10Tim Starling: [C: 04-2] "-2 per https://bugzilla.wikimedia.org/show_bug.cgi?id=50040#c4" [core] - 10https://gerrit.wikimedia.org/r/70036 (owner: 10Krinkle) [00:35:42] (03CR) 10Hashar: [C: 04-1] "(11 comments)" [core] - 10https://gerrit.wikimedia.org/r/52922 (owner: 10Victor Vasiliev) [00:35:52] Krenair: review posted :-) [00:35:59] ty [00:36:06] Krenair: I like how you have split everything [00:36:11] that makes code WAYYYY easier to follow [00:36:14] I don't [00:36:15] and hopefully testable [00:36:25] :p [00:36:53] and i forgot something but I can ask there [00:37:14] Krenair: would it make sense to reach out with the bot authors to find out whether the proposed json output would be fine to them ? [00:37:17] I forgot to change that name for wfErrorLog [00:38:02] Yes, probably. [00:38:17] (03PS1) 10Bsitu: Update email text color and add a period to notification text [extensions/Echo] - 10https://gerrit.wikimedia.org/r/76667 [00:38:35] (03CR) 10Krinkle: [C: 04-1] "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/52922 (owner: 10Victor Vasiliev) [00:38:55] Particularly whether there's too much information and I'd like to know if there's any important bit of info we missed [00:39:29] (03PS2) 10Bsitu: Update email text color and add a period to notification text [extensions/Echo] - 10https://gerrit.wikimedia.org/r/76667 [00:42:08] hashar, re "Whenever the 'engine' is missing from the feed definition, we would get a fatal since it is the class name. I am wondering whether we should discard that engine and log or throw an exception." [00:42:24] (03PS3) 10MarkAHershberger: Install without XML support [core] - 10https://gerrit.wikimedia.org/r/76663 [00:42:43] Krenair: I have no idea how to reach out to bot authors. Krinkle might know. [00:42:44] I don't think so [00:43:24] (03CR) 10jenkins-bot: [V: 04-1] Install without XML support [core] - 10https://gerrit.wikimedia.org/r/76663 (owner: 10MarkAHershberger) [00:43:48] hashar: Krenair: what kind of bots, regarding what? [00:44:15] hashar was thinking about contacting people who run bots which parse the current IRC RC feeds [00:44:25] about giving them the option to receive json instead [00:44:46] I'd recommend not spending any time on this. It will likely be wrongly received and not put to full use. [00:44:59] come on timo :-D [00:45:13] I'm speaking as staff of CVN here, not a developer [00:45:24] This is in need of greater refactor, and parsing JSON over IRC would still be crappy [00:45:30] and in fact has a ton of issues [00:45:32] yeah I understood :) [00:45:39] such as UDP's behaviour of cropping the packet [00:45:42] leaving it unusable [00:45:59] Greater refactor? [00:46:06] krenair has been working on slightly refactoring it [00:46:16] what do you have in mind? [00:46:17] basically I'd like migrate once (not for eternity, but at least a slightly bigger refactor) [00:46:26] giving us a way to implement different transport such as a TCP to a queue system ( hint hint zmq) [00:46:29] eg. from color parser irc to a proper socket for json [00:46:30] not irc [00:46:50] (03PS1) 10Bsitu: Add period to the end of email notification text [extensions/PageTriage] - 10https://gerrit.wikimedia.org/r/76670 [00:46:50] and set up a program that feeds irc from the json feed. [00:47:00] definitely not having mediawiki do both directly [00:47:28] Also, various implementation details might have to change again when I start working on the new countervandalism and patrolling tools. [00:47:34] (as a wmf project probably) [00:47:56] so I don't think it makes sense right now to do this, I certainly don't plan no spending any time refactoring the cvn bots for it. [00:48:08] well there is no idea to refactor it [00:48:16] because it works right now, and I'm not refactoring a parser with the only gain of it not being hacky. it works as hacky as it is. [00:49:00] we talked extensively about it with yuvi /ori earlier this morning, the rough idea is to have the IRC Feed send over TCP to a zero mqueue [00:49:09] and that's all I'm going to say about it for now. Working on other thing snow. [00:49:20] Talked about it why? Who's asking for this? [00:49:21] and then get subscriber to publish that in redis queues for some labs bot to consume [00:49:32] eventually move the IRC stuff to be a subscriber of the zero mqueue [00:49:44] I don't have time for this now. [00:49:56] yeah yeah [00:49:59] contact me elsehow if you want to contact cvn or bots or anything else from me. sorry. [00:50:04] we were merely wondering how to reach out to some bots :-] [00:50:32] it depends a lot since differnet bots have different abstraction levels. [00:50:42] very few will actually need to be engaged in the change you're looking for [00:50:53] to find out whether they could use some more fields. I guess as a first step it could just have the same amount of data being provided by the current irc feed [00:51:17] api recentchanges packet is what I think would be most useful [00:51:26] not invent another one [00:51:57] last year when I had some breathing room I worked in a simple proof-of-concept D-JSON [00:52:24] to allow splitting up packets between mediawiki emitting UDP and the server that receives them for further distribution (the subscription socket service0 [00:52:53] https://github.com/Krinkle/D-JSON there may be other projects that do the same, i just wrote it up as something I used to run this service at the TOolserver for a while [00:53:04] (based on polling the API and IRC) [00:53:29] subscribable from WebSockets in the browser and in tools on the server-side [00:53:36] :) [00:53:50] https://github.com/Krinkle/D-JSON/blob/master/test/D-JSON.test.js#L56-L69 [00:53:59] (03PS6) 10Alex Monk: Replace remaining sajax use [core] - 10https://gerrit.wikimedia.org/r/65641 [00:54:06] [00:54:13] thanks timo :) [00:54:36] I can act as proxy if you have a message to send (draft on email or etherpad and send me alink) [00:54:44] Krenair: so at least one point from above is using the same format as the API for recentchanges [00:55:01] Krinkle: that would be kind from you! thanks for the offering :) [00:59:53] (03PS1) 10Legoktm: For rows added by the SpecialStatsAddExtra hook, add an HTML id equal to the message key [core] - 10https://gerrit.wikimedia.org/r/76671 [00:59:53] Elsie: ^ [01:01:05] RoanKattouw: do you know of a method in mediawiki javascript to detect what CSS media mode I'm in? [01:01:10] specifically if I'm in print mode [01:01:25] Hmmm [01:01:39] I don't, actually [01:01:45] But why do you need this? [01:01:55] In theory you should be able to achieve what you want using media styles [01:01:56] to prevent CentralNotice banners from being shown in print view [01:02:23] and we already use JS to modify the centralNotice div visability [01:02:42] so I was just going to have it not even bother to request a banner in print view [01:02:53] mwalker: use classes instead of inline styles to toggle visibility, then hide that class in print mode and/or only show in media=screen [01:03:45] as for not making the request, when people press print, most browsers print the current active DOM with the printable styles applied, so it is already requested [01:03:48] Krinkle: I would normally agree -- but there is far too custom user code for that to work reliably [01:04:13] mwalker: some people click 'Printable version' that one has printable=1 in the url query, so that'd be your queue [01:04:41] that's sort of what I was looking at; but merp [01:07:20] (03CR) 10Hashar: "And I found out that ApiQueryRecentChanges::extractRowInfo( $sql_row ) had duplicated the logic from RecentChanges() :/ I really hate ou" [core] - 10https://gerrit.wikimedia.org/r/52922 (owner: 10Victor Vasiliev) [01:07:50] (03PS11) 10Bsitu: user talk and mention notification improvement [extensions/Echo] - 10https://gerrit.wikimedia.org/r/75546 [01:08:07] (03CR) 10Legoktm: "(1 comment)" [extensions/MassMessage] - 10https://gerrit.wikimedia.org/r/76647 (owner: 10Legoktm) [01:15:43] and 3am off to bed. Have a good afternoon [01:17:24] (03CR) 10Mattflaschen: [C: 04-1] "(6 comments)" [extensions/PronunciationRecording] - 10https://gerrit.wikimedia.org/r/74413 (owner: 10Mdale) [01:17:54] (03CR) 10Catrope: [C: 031] Use wfResetOutputBuffers in ResourceLoader [core] - 10https://gerrit.wikimedia.org/r/74570 (owner: 10Hoo man) [01:25:02] (03CR) 10Hashar: [C: 031] Remove UploadBlacklist [tools/release] - 10https://gerrit.wikimedia.org/r/76230 (owner: 10Reedy) [01:30:23] (03CR) 10Hashar: "(1 comment)" [qa/browsertests] - 10https://gerrit.wikimedia.org/r/75793 (owner: 10Manybubbles) [01:58:44] (03CR) 10MZMcBride: "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/76671 (owner: 10Legoktm) [02:00:50] (03CR) 10Tim Starling: [C: 032] "No comments in 2 weeks. I'm pretty sure this is harmless." [extensions/Score] - 10https://gerrit.wikimedia.org/r/73720 (owner: 10Tim Starling) [02:01:07] (03Merged) 10jenkins-bot: Reduce Lilypond memory usage [extensions/Score] - 10https://gerrit.wikimedia.org/r/73720 (owner: 10Tim Starling) [02:01:46] (03CR) 10Tim Starling: [C: 032 V: 032] fss.ini used '#' for comments, which is deprecated [php/FastStringSearch] - 10https://gerrit.wikimedia.org/r/76511 (owner: 10Hashar) [02:06:08] (03CR) 10Mattflaschen: [C: 04-1] "(1 comment)" [extensions/Annotator] - 10https://gerrit.wikimedia.org/r/75106 (owner: 10Rjain) [02:11:55] (03PS5) 10Tim Starling: Return messages in a consistent form from Status objects [core] - 10https://gerrit.wikimedia.org/r/74176 (owner: 10Daniel Kinzler) [02:11:56] (03CR) 10Tim Starling: [C: 032] Return messages in a consistent form from Status objects [core] - 10https://gerrit.wikimedia.org/r/74176 (owner: 10Daniel Kinzler) [02:16:05] (03Merged) 10jenkins-bot: Return messages in a consistent form from Status objects [core] - 10https://gerrit.wikimedia.org/r/74176 (owner: 10Daniel Kinzler) [02:18:17] (03CR) 10Mwalker: "(2 comments)" [extensions/BookManagerv2] - 10https://gerrit.wikimedia.org/r/76463 (owner: 10Mollywhite) [03:25:02] How can I add code blocks in gerrit comments? [03:28:55] leading whitespace. at least two spaces, IIRC. [03:35:33] thanks [03:40:27] One space is fine, I think. [03:41:11] <^demon> I think you have to have at least on \n before the [ ]+ [03:53:35] ^demon is apparently waiting in the shadows for the upload blacklist to be disabled so that he can Goatse every wiki. [03:53:50] (Did I use that verb correctly?) [03:53:56] <^demon> Sssh, don't spoil my plan. [04:03:08] Krinkle: there's a newish dev tools feature in Chromium, "Show potential scroll bottlenecks", that highlights DOM elements that could slow down scroll smoothness. Interestingly it flags the wikitext editor on MW.org (haven't tried elsewhere). https://code.google.com/p/chromium/issues/detail?id=253552 suggests it's caused by elements that have touch or wheel event listeners. [04:03:13] http://i.imgur.com/Kt4BtcK.png [04:04:13] I wonder if it's just a fact of life with that have content that flows past the edge of the element, or something that we could potentially improve. [04:05:15]