[00:01:34] (PS2) Krinkle: SpecialPrefixindex formatting methods are now protected [core] - https://gerrit.wikimedia.org/r/75569 (owner: Hashar) [00:01:45] (CR) Krinkle: "Rebased to fix outdated dependency." [core] - https://gerrit.wikimedia.org/r/75569 (owner: Hashar) [00:12:36] (CR) Mattflaschen: [C: -1] "(7 comments)" [core] - https://gerrit.wikimedia.org/r/74662 (owner: MarkTraceur) [00:15:43] (PS1) MarkAHershberger: Scribunto should error out sooner if lua isn't executable Improve error reporting by causing Scribunto to fail earlier if lua can't be run. [extensions/Scribunto] - https://gerrit.wikimedia.org/r/75819 [00:20:10] superm401: I'm curious where you see a convention to append ".js" to module names that have JavaScript in them [00:20:22] I only see the vector one and the user login one [00:20:31] Er, createaccount. [00:20:33] marktraceur, only when there's a CSS-only, and a JS adjunct. [00:20:43] And before that, skins.vector.js, for example. [00:21:29] Hm. [00:21:29] K [00:24:15] superm401: And what do you mean by "This should apply for table and div?" [00:25:06] There are three display formats, table, div, and raw. [00:25:28] marktraceur, you are currently applying that CSS class to both table and div, and I was asking to confirm whether that was intentional. [00:26:02] * marktraceur looks [00:26:04] marktraceur, re the live count of people with the preference, you should a DB expert someone check the performance. [00:26:26] I didn't leave an official review, since it doesn't look like that's ready for review anyway. [00:26:49] I think it may require an index scan, but I'm not sure, and I don't know the performance ramifications. [00:26:51] superm401: marktraceur: that should be possible [00:27:02] there's an index on user_properties.up_property [00:27:21] it's even commented :) -- Name of the option being saved. This is indexed for bulk lookup. [00:27:22] superm401: I only did that in the demo patch IIRC [00:27:54] hm, although i guess you'd want to count() that [00:27:56] marktraceur, yeah, that's why I didn't leave a review on it. [00:28:05] I can deal with the DB ramifications in a separate patch, the field options include "user-count" and the caller is responsible for not breaking things [00:28:08] so i'm not sure now. [00:28:16] MatmaRex, yeah, me neither. [00:28:26] I know looking up a single user's preference should be basically instant due to the index. [00:28:26] meh, i clearly need some sleep. [00:28:29] But not sure about the count. [00:28:33] 'user-count' => 500000000, // what, we checked, it is [00:28:42] hm, actually [00:28:51] the count() things should be just fine, too [00:29:13] it will only scan all of the rows with the given property, in worst case [00:29:16] I'll ask someone later on [00:29:20] (and maybe it won't even do that) [00:29:31] marktraceur, sounds good. Asher can probably help. [00:29:43] If it's really bad, we can also just omit that until we can solve performance issues [00:29:50] The design will work with or without it [00:29:58] so unless you have hundreds of thousand of people check that pref it should be good [00:30:47] superm401: I guess the container class should apply to all form types, but I'm not sure that's what you were asking [00:30:58] I don't know why we wouldn't show that class for any kind of container [00:31:05] But that's the only place I saw a class being added [00:31:28] That's what I'm asking. Let me look a bit more. [00:31:38] KK [00:32:02] Or...I guess the point is that there is no container for raw [00:32:17] So both of the other ones need the class, but not raw [00:32:27] marktraceur, it looks like you're only using table.mw-htmlform-container [00:32:39] And only the 'table' type uses table for the root (as one would hope). [00:32:46] Sure sure [00:32:58] I should probably delete the table selector [00:33:18] If you do, be sure to test it with div. [00:33:18] Oh, no [00:33:26] Those CSS rules only make sense for tables [00:33:29] The alternative is to just do it for $displayFormat === 'table' [00:33:30] But there are also div containers [00:33:58] That wouldn't affect the display, but then div containers wouldn't have the class [00:34:13] I can see that being useful to someone down the line [00:34:27] "oh, hey, a convenient class to style any mediawiki forms" [00:35:01] marktraceur, for me, mw-htmlform-container is totally empty. [00:35:06] At least on the preferences screen. [00:35:14] Just an empty tbody, and nothing else. [00:35:33] Really? [00:35:49] Yeah, I'll check on prod. [00:36:00] I see, like, 15 of them [00:36:06] Sorry [00:36:06] 33 [00:36:17] There's one for each section and subsection [00:36:52] You likely saw the zeroth one, which is empty (why, I don't know) [00:36:52] Looks the same on enwiki, except of course there's no class. I'll walk over. [00:37:50] Weird. [00:37:53] marktraceur, never mind, I see. [00:37:59] Actually, 0 and 1. [00:38:16] The one immediately inside #preferences, and inside #mw-prefsection-personal [00:38:21] But if it has a field, it's not empty. [00:38:38] (PS2) Bsitu: user talk and mention notification improvement [extensions/Echo] - https://gerrit.wikimedia.org/r/75546 [00:38:42] Yeah [00:38:50] Maybe they're hidden or API preferences [00:38:57] But the container didn't get axed [00:38:59] Or something [00:39:07] Anyway, I'm fine either way regarding htmlform-container. [00:39:15] You're right, the class does no harm if it's a div root. [00:39:23] KK [00:39:32] I like having information in the DOM, so I'll keep it for now [00:43:24] (PS9) MarkTraceur: Implement agora-style checkboxes experimentally [core] - https://gerrit.wikimedia.org/r/74662 [00:43:37] (CR) MarkTraceur: "(7 comments)" [core] - https://gerrit.wikimedia.org/r/74662 (owner: MarkTraceur) [02:22:47] (PS5) Legoktm: Add a preview/confirmation menu. Also major internal restructuring of the special page. [extensions/MassMessage] - https://gerrit.wikimedia.org/r/75550 [02:22:48] (CR) jenkins-bot: [V: -1] Add a preview/confirmation menu. Also major internal restructuring of the special page. [extensions/MassMessage] - https://gerrit.wikimedia.org/r/75550 (owner: Legoktm) [02:31:29] (PS6) Legoktm: Add a preview/confirmation menu. Also major internal restructuring of the special page. [extensions/MassMessage] - https://gerrit.wikimedia.org/r/75550 [02:35:36] (CR) Legoktm: "PS6: Adds a "Preview" button if you're already on the preview page, and more internal restructuring." [extensions/MassMessage] - https://gerrit.wikimedia.org/r/75550 (owner: Legoktm) [02:35:47] Elsie: ^ [02:49:24] legoktm: \o/ [02:50:04] :D [03:01:18] (CR) Brian Wolff: "Re demon: My personal use case is that some times people report bugs that only affect one file type, and its useful to get a list of 50 fi" [core] - https://gerrit.wikimedia.org/r/67468 (owner: Brian Wolff) [03:04:26] (PS3) Nischayn22: adds blacklisting feature for flickr in backend [extensions/UploadWizard] - https://gerrit.wikimedia.org/r/65109 [03:07:53] (PS4) Nischayn22: adds blacklisting feature for flickr in backend [extensions/UploadWizard] - https://gerrit.wikimedia.org/r/65109 [03:09:08] (PS5) Brian Wolff: Make Special:MIMESearch a non-expensive special page. [core] - https://gerrit.wikimedia.org/r/67468 [03:10:18] (PS5) Nischayn22: adds blacklisting feature for flickr in backend [extensions/UploadWizard] - https://gerrit.wikimedia.org/r/65109 [03:11:03] (CR) Nischayn22: "Dependency merged and rebased" [extensions/UploadWizard] - https://gerrit.wikimedia.org/r/65109 (owner: Nischayn22) [03:11:48] (CR) Brian Wolff: "PS5: rebase" [core] - https://gerrit.wikimedia.org/r/67468 (owner: Brian Wolff) [03:21:23] (CR) Mattflaschen: [C: -1] "(6 comments)" [core] - https://gerrit.wikimedia.org/r/75143 (owner: MarkTraceur) [03:27:38] legoktm: Can you update the Labs instance? [03:27:43] Oh, you're at dinner. I'll wait. [03:38:45] Elsie: why don't you have access to the labs instance? [03:40:48] (CR) Mattflaschen: "(4 comments)" [vagrant] - https://gerrit.wikimedia.org/r/75069 (owner: Ori.livneh) [03:56:03] (CR) Parent5446: [C: 2] Make Special:MIMESearch a non-expensive special page. [core] - https://gerrit.wikimedia.org/r/67468 (owner: Brian Wolff) [03:58:32] (Merged) jenkins-bot: Make Special:MIMESearch a non-expensive special page. [core] - https://gerrit.wikimedia.org/r/67468 (owner: Brian Wolff) [04:03:38] p858snake|l: I'm too Jewish. [04:03:44] (CR) Parent5446: [C: -1] "(7 comments)" [core] - https://gerrit.wikimedia.org/r/75143 (owner: MarkTraceur) [04:10:03] (PS4) Santhosh: Adding Old Persian web font [extensions/UniversalLanguageSelector] - https://gerrit.wikimedia.org/r/75813 (owner: Ebrahim) [04:13:15] (CR) Santhosh: "eot file is more compressed using microtype compression( sfntly -e -x Xerxes.ttf Xerxes.eot)" [extensions/UniversalLanguageSelector] - https://gerrit.wikimedia.org/r/75813 (owner: Ebrahim) [04:14:09] (CR) Santhosh: [C: 2] Adding Old Persian web font [extensions/UniversalLanguageSelector] - https://gerrit.wikimedia.org/r/75813 (owner: Ebrahim) [04:14:18] (Merged) jenkins-bot: Adding Old Persian web font [extensions/UniversalLanguageSelector] - https://gerrit.wikimedia.org/r/75813 (owner: Ebrahim) [04:14:56] (PS1) Mwjames: \SMW\PropertyPageIdMapper [extensions/SemanticMediaWiki] - https://gerrit.wikimedia.org/r/75825 [04:19:21] (PS2) Ori.livneh: Parse MediaWiki fatals/exceptions and republish as JSON stream [tools/fluoride] - https://gerrit.wikimedia.org/r/75560 [04:37:06] (CR) Ori.livneh: "(1 comment)" [vagrant] - https://gerrit.wikimedia.org/r/75069 (owner: Ori.livneh) [04:40:36] (PS1) Ori.livneh: Drop unused import & stray debug statement [vagrant] - https://gerrit.wikimedia.org/r/75827 [04:41:02] (CR) Ori.livneh: [C: 2] Drop unused import & stray debug statement [vagrant] - https://gerrit.wikimedia.org/r/75827 (owner: Ori.livneh) [04:41:05] (Merged) jenkins-bot: Drop unused import & stray debug statement [vagrant] - https://gerrit.wikimedia.org/r/75827 (owner: Ori.livneh) [04:43:47] (PS2) Mwjames: \SMW\PropertyPageIdMapper [extensions/SemanticMediaWiki] - https://gerrit.wikimedia.org/r/75825 [05:17:38] (PS3) Mattflaschen: Don't override originally requested position when drawing. [extensions/GuidedTour/guiders] - https://gerrit.wikimedia.org/r/74820 [05:18:16] legoktm: Yo. [05:18:19] (CR) Mattflaschen: "(2 comments)" [extensions/GuidedTour/guiders] - https://gerrit.wikimedia.org/r/74820 (owner: Mattflaschen) [05:29:43] Elsie: Hey [05:30:10] I updated the labs instance with the latest patchset [05:31:26] * Elsie bites legoktm. [05:31:28] oh [05:31:29] I see [05:31:36] Fixed now [05:31:56] I pulled from master and then re-checked out the patchset. [05:31:57] I'm not even sure what was broken. [05:32:05] I was about to clean up the /Design page. [05:32:57] Bug 51948 affects your /Design subpage. [05:32:58] BTW. [05:33:00] legoktm: ^ [05:33:13] * legoktm looks [05:33:31] Heh [05:34:24] Don't edit that page for a minute, I'm gonna clean it up. [05:34:25] Funnily enough, I added the with VE :P [05:34:29] Ok, I'm not [05:36:13] What's the shortened form of your username? [05:36:15] Lego? [05:36:17] KTM? [05:36:21] LKTM? [05:36:26] (LGTM?) [05:36:27] Lego usually [05:36:32] :P [05:36:49] Is that what you prefer? [05:37:18] lego or legoktm work. I get annoyed when people refer to me as legotkm [05:37:37] Though I stalk basically all variations of it on IRC :P [05:39:28] I went with --Lego. [05:39:37] sounds good [05:44:45] legoktm: Okay, page mostly updated. [05:44:49] legoktm: https://gerrit.wikimedia.org/r/#/c/75276/ [05:45:05] Thoughts about that patchset? [05:45:24] https://www.mediawiki.org/wiki/Extension:MassMessage/Design [05:46:42] Well I think there's no harm in merging it, since if we do end up editing through the API, it'll get scrapped, but implementing it has no downsides that I can think of [05:47:17] I think it should be abandoend. [05:47:31] Why? [05:48:22] abandoned [05:49:37] [01:46] Well I think there's no harm in merging it, since if we do end up editing through the API, it'll get scrapped, but implementing it has no downsides that I can think of [05:49:41] [01:47] I think it should be abandoend. [05:49:44] [01:47] abandoned [05:50:04] [10:47:30 PM] Why? [05:52:09] legoktm: Sorry, had connectivity issues for a second. [05:52:19] [10:46:41 PM] Well I think there's no harm in merging it, since if we do end up editing through the API, it'll get scrapped, but implementing it has no downsides that I can think of [05:52:19] legoktm: Because I think faking an edit summary is bad practice. [05:52:20] [10:47:18 PM] I think it should be abandoend. [05:52:20] [10:47:30 PM] Why? [05:52:20] [10:48:23 PM] abandoned [05:52:25] Ah [05:52:34] And I think copying code from EditPage.php is bad practice. ;-) [05:52:40] Heh. [05:53:39] The issue is that there is no standalone function that you can use, it's all buried in way more complex code that we don't need [05:53:57] If only there were some kind of abstraction layer. [05:54:01] An edit API, say. [05:54:18] Heh. [05:54:22] I'll look into that a bit more. [05:54:28] All right. [05:54:41] legoktm: I submitted a delivery and I'm not sure it ever ran. [05:55:05] Did it get logged? [05:55:08] legoktm: You should talk to Aaron or Tim about how to extract useful queue info from the Redis job queue. [05:55:12] Yes. [05:55:16] Ok [05:55:34] job queue is empty [05:55:37] The lack of delivery monitoring is really bothersome. [05:55:41] Did it run? [05:56:01] I think it's just broken. [05:56:24] http://legoktm.instance-proxy.wmflabs.org/wiki/Special:Log <-- I don't see the log entry? [05:56:36] oh [05:56:39] timezones [05:56:40] (Mass message log); 05:15 . . Admin (Talk | contribs) sent a message to Main Page ‎(Does this work???) [05:56:52] Right. [05:56:56] time zones [05:57:39] Hmmmmmm. [05:57:45] It works fine locally though >.> [05:57:56] Well, you need to figure out if the jobs are getting submitted. [05:58:08] Yeah, I'm debugging right now [05:58:10] K. [05:59:07] Yup it worked locally [05:59:20] But local is using the mysql job queue [05:59:49] Ain't that a bitch. [05:59:49] Is Redis up and running? [06:01:28] legoktm@legoktm:/srv/mediawiki/maintenance$ sudo /etc/init.d/redis-server start [06:01:28] Starting redis-server: redis-server. [06:01:38] Would it have said anything different if it was already running? [06:01:51] And I think MediaWiki should through an error that it couldn't make a connection or something... [06:01:53] throw* [06:02:02] NFI. [06:02:29] Well, figure that out and then we can work on the other bugs. ;-) [06:02:34] Hmmm. [06:02:38] Still not working [06:03:00] Insert a job manually and see if it works. [06:03:10] You need to see if it's a Redis issue or an extension issue. [06:03:19] Already trying that :P [06:06:17] I think its job queue / redis [06:06:19] > $job = new MassMessageJob( Title::newFromText( 'User talk:Admin'), array('spamlist'=>'Thing', 'subject'=>'cli test','message'=>'hi')); [06:06:20] > JobQueueGroup::singleton()->push( $job ); [06:06:24] legoktm@legoktm:/srv/mediawiki/maintenance$ sudo php runJobs.php [06:06:25] legoktm@legoktm:/srv/mediawiki/maintenance$ [06:06:41] Let me disable redis and switch to mysql and see what happens [06:07:34] Hmmm. Not that.... [06:10:19] Hrmph. I haven't changed a single thing on labs that would cause this. [06:10:27] jobs aren't being inserted for whatever reason. [06:17:49] Elsie: I'm completely lost as to why this isn't working. [06:19:16] Have you tried restarting? [06:20:17] legoktm@legoktm:~$ sudo reboot [06:21:01] is there an existing tool to remove spams / spammers "recursively" [06:21:50] for me as a wikipedia sysop rather than developer :p [06:22:39] Elsie: wtf. That worked aksjdhfskjdfh [06:22:41] "recursively"? [06:22:48] hasn't lego written such scripts? [06:22:55] I er what? [06:23:01] [23:21:16] liangent is there an existing tool to remove spams / spammers "recursively" [06:23:15] recursively? [06:24:09] Elsie: Lemme re-enable redis and see [06:24:59] legoktm: for example, I give it a bad site as spam extlink, it finds all users who added it, then work on those users; or I give it a user, it reverts all their edits and finds out all extlinks they added, then work on those extlinks [06:25:00] like this [06:25:34] liangent: Ah. You can look into what Beestra has w/ XLinkBot and stuff. [06:27:03] Elsie: Ok, everything should be working now. [06:30:05] legoktm: it seems XLinkBot is not what I want [06:30:37] Maybe COIBot? you should talk to billinghurst (nick sdrewth), he deals with a lot of this stuff [06:30:40] I meant to clean up bad sites which fit MediaWiki:Spam-blacklist well [06:30:50] but those which wasn't listed there in time [06:31:12] so a bunch of spam links are already scattered on the wiki [06:41:37] TimStarling: do you think that registering the namespaces earlier in extension code fully resolves bug 45031, or should core be patched as well? [06:42:33] It's bad that registering a namespace too late half-works rather than trigger an error. [06:42:41] I think registering earlier is enough [06:43:16] (PS7) Legoktm: Add a preview/confirmation menu. Also major internal restructuring of the special page. [extensions/MassMessage] - https://gerrit.wikimedia.org/r/75550 [06:43:48] (CR) Legoktm: "PS7: Rather than display a disable button, just don't display it at all." [extensions/MassMessage] - https://gerrit.wikimedia.org/r/75550 (owner: Legoktm) [06:44:01] (CR) Legoktm: "disabled*" [extensions/MassMessage] - https://gerrit.wikimedia.org/r/75550 (owner: Legoktm) [06:44:16] it's more conditional registration than late registration [06:44:52] from the point of view of MWNamespace, that is [06:45:24] MWNamespace doesn't know what stage of setup it's in, so it has no way to protect itself against hooks that randomly change their return values based on external dependencies [06:45:32] aside from not caching [06:45:33] (Abandoned) Yuvipanda: Bug 27047: COMMONS MAN [core] - https://gerrit.wikimedia.org/r/75641 (owner: SuchABot) [06:46:00] Right; I think the functionality that is exposed by the hook is adequate. I'm inclined to think that having the hook fire in a lazy-loading routine is confusing, though, because it makes it hard to reason about when it actually fires. [06:46:04] but it probably does benefit from being cached, assuming it is called from Title::secureAndSplit() [06:46:29] that being fairly hot performance-wise, due to its use in link generation in the parser [06:46:38] If it were in file scope in Setup.py with a comment saying /* No additional namespaces may be registered after this point */, it'd be less error prone [06:47:22] yeah, I don't know enough to reason about the performance implications of that. [06:47:57] well, the point at which it becomes dangerous is not really knowable, it's not really a particular point in Setup.php [06:48:13] any time after the file scope of the registration file, it is dangerous [06:49:49] I guess there's no easy / performant way to freeze an array such that attempts to modify it further trigger an error? [06:50:32] you mean freezing $wgHooks? not really [06:50:44] you could make a copy of $wgHooks in a private variable, I guess [06:51:22] and remove Hooks::register() [06:51:42] not sure the gesture would be appreciated [06:52:37] maybe just patch it so it delivers a small electric shock [06:52:52] "bad! don't touch that." [06:55:46] WHEEEEEEEEEE EEEE eeeeee! [06:55:56] * Isarra hug-tackles ori-l. [06:56:00] (CR) Aklapper: "(2 comments)" [extensions/PronunciationRecording] - https://gerrit.wikimedia.org/r/75770 (owner: Rahul21) [06:57:42] you could make $wgHooks an object implementing ArrayAccess [06:57:58] then throw an exception from offsetSet() [06:58:14] I was looking at that, and I think it's a good idea [06:58:31] the error would be predictable and if someone (say) logged on IRC to report it you'd know instantly what they were doing wrong [06:59:15] actually each hook list would have to implement ArrayAccess [06:59:26] e.g. $wgHooks['CanonicalNamespaces'] would be an object [07:01:30] I'm looking to see where most hook handler arrays are instantiated [07:01:32] anyway, like I say, I think it's unlikely that the change would be met with enthusiasm [07:03:13] it'd break Scribunto_LuaTitleLibraryTests and Wikibase\Test\ChangeHandlerTest, for a start [07:03:49] Elsie: https://s3.amazonaws.com/f.cl.ly/items/0G1W2n0c2B2V0v2f0E2V/Screen%20Shot%202013-07-25%20at%2012.02.10%20AM.PNG [07:04:28] vagrant! [07:04:41] ori-l: :D [07:05:11] UploadBaseTest, DatabaseUpdater::loadExtensions() [07:05:51] Yes, that's a number of unpleasant wikitech-l threads [07:06:22] which is why it's so great that the platform team is committing to making this change and properly socializing it [07:06:44] :) [07:06:46] quite a lot of core tests actually [07:07:13] legoktm: :-) [07:07:25] legoktm: I'd put the preview in a fieldset as well, I think. [07:07:31] Elsie: fieldset? [07:07:34] all LanguageConverter languages, although that's probably for the best, I don't know what the hell they are doing setting that hook [07:07:47] legoktm: Pretty blue box with a title. [07:07:53] It's a

. [07:07:53] Ah. [07:08:28] I'm html-n. [07:08:47] Oh, this shouldn't be that hard. [07:09:09] > $html = Xml::fieldset( $this->mWrapperLegend, $html ); [07:09:22] That sounds like some work but not unmanageably so, and it'd eliminate a class of programmer errors that we must assume would currently trigger all manner of seemingly unrelated oddities, half-extant namespaces being one. [07:09:27] morning [07:09:34] hi hashar [07:09:46] I noticed you created a fluoride instance in beta !!! [07:09:49] Elsie: the title for the blue box thing should say… "Preview"? [07:09:49] 1111!!!BBQ!!! [07:09:55] maybe if the protection mechanism was disabled in unit testing and in the installer, then it would be workable [07:10:09] legoktm: Something like that. [07:10:10] and all those language classes migrated to do something sane [07:10:56] legoktm: As long as the message exists, it can always be improved later. :-) [07:10:56] I remember there being some discussion about extensions using messages from core…what's the rule/guideline about that? [07:11:03] Don't re-use messages. [07:11:05] That's the rule. [07:11:23] Ok. [07:11:44] I think there are certain exceptions (such as separators and other utility messages). [07:12:01] TimStarling: so -- what's the next step? I don't think I'm the right person to see that through; it touches too many things I'm not very familiar with. [07:12:20] legoktm: the core i18n message can apply to a slightly different context, so it is usually better to duplicate the message in your extension so people can adjust to the context. [07:12:28] ori-l: Sounds like a tracking bug. [07:12:34] But I don't really know what the two of you are talking about. [07:12:50] hashar: ok, will do. [07:12:57] legoktm: some languages have slightly different rules, like counting animals, food or objects could use different rules [07:14:00] I don't think it's a tracking bug, but it's a bug; I'll file it, though probably not earlier than tomorrow. [07:14:21] ori-l: if I were you, I'd add a note to the doc comment of $wgHooks saying that changing the array after file scope can lead to unexpected results due to caching [07:14:24] It sounds like it'll involve a lot of cleanup to various extensions and core files. [07:14:37] and post to wikitech-l mentioning the issue, and then leave it at that [07:14:57] Elsie: heh. http://cl.ly/image/3Y2o04293Q2C [07:14:58] because I don't know if the wider issue really justifies the measures you're proposing [07:14:58] (PS1) Physikerwelt: Expose LaTeXML Settings [extensions/Math] - https://gerrit.wikimedia.org/r/75828 [07:15:15] I need to move the other message above it. [07:15:35] (PS2) Yuvipanda: Do not set default values for lat and lon [extensions/UploadWizard] - https://gerrit.wikimedia.org/r/75597 (owner: Rillke) [07:15:44] I'd only output the intro message on initial load. [07:15:53] ori-l: tiny JsonSchema related changeset, can you +2? https://gerrit.wikimedia.org/r/75597 [07:16:00] TimStarling: OK, deal. [07:16:09] I think after a user presses preview, we can assume they understand how a standard form works. ;-) [07:16:15] Elsie: Gotcha. [07:16:23] Users won't read the damn thing, anyway. [07:16:35] It'll just be abused by me to shout at people not to fuck up deliveries. [07:16:54] Which they will anyway. [07:17:00] But at least it'll be a cleanly injected hate-message. [07:17:24] legoktm: While you're in there, the "(Also" is driving me insane. [07:17:40] It should be "Subject of the message (also used in the edit summary)". [07:17:48] ok [07:17:52] Not even sure the "also" is needed. [07:18:21] There's an open question about where you'll include the targets info. [07:18:26] Is that logic written yet? [07:18:32] For extracting targets, I mean. [07:18:44] Elsie: https://gerrit.wikimedia.org/r/#/c/74976/1 [07:18:52] extracting targets for local deliveries yes [07:18:57] displaying, no [07:19:43] Oh, you're showing me that changeset to demonstrate that the wonky capitalization wasn't your fault. [07:19:46] Fair enough. [07:19:50] Elsie: new preview form: http://cl.ly/image/3u3c141M3C0G [07:20:15] also "page-list" :P [07:20:16] Looks good to me. [07:20:23] Indeed. :P [07:20:41] (CR) Yuvipanda: "Updated with a slightly different solution, since setting them to '' breaks Campaigns - which expects lats/lons to be numbers if they exis" [extensions/UploadWizard] - https://gerrit.wikimedia.org/r/75597 (owner: Rillke) [07:21:09] legoktm: Commit and deploy. :-) [07:22:25] (PS8) Legoktm: Add a preview/confirmation menu. Also major internal restructuring of the special page. [extensions/MassMessage] - https://gerrit.wikimedia.org/r/75550 [07:23:28] (CR) Legoktm: "PS8: Wrap everything in

, only display the help message to the user in form state, and switch $this->state to "form" if we hit a" [extensions/MassMessage] - https://gerrit.wikimedia.org/r/75550 (owner: Legoktm) [07:24:07] Elsie: deployed [07:26:08] I'm cleaning up your commit message. [07:27:27] thanks :D [07:27:43] lemme add the qqx messages [07:27:49] (PS9) MZMcBride: Add a preview/confirmation menu. Also major internal restructuring of the special page. [extensions/MassMessage] - https://gerrit.wikimedia.org/r/75550 (owner: Legoktm) [07:30:07] The bot uses the old commit message. [07:30:12] Silly bot. [07:32:39] (PS10) Legoktm: Add a preview/confirmation menu; major special page internal restructuring [extensions/MassMessage] - https://gerrit.wikimedia.org/r/75550 [07:32:50] Elsie: it's what gerrit reports [07:33:02] Is there an upstream bug? [07:33:12] *bug report [07:33:20] no idae [07:36:29] legoktm: Do you have to clear the queue manually? [07:36:35] I tried another delivery. [07:36:41] At :35. [07:36:46] Elsie: have you waited at least 60 seconds? [07:36:56] Probably not. [07:37:05] the cron job runs every minutes [07:37:07] minute* [07:37:28] It was submitted at :35 and it's now :37. [07:37:59] nothing in the job queue [07:38:51] Hmmmm. [07:38:53] Then your extension is broken. [07:40:10] But it works when I rebooted >.> [07:40:13] And it works locally. [07:42:04] Sounds like a conspiracy. [07:42:50] Elsie: try now. [07:44:07] Submitted at :43. [07:44:51] ran it manually and it worked [07:45:42] You don't have a newline above the == header == because you're manually creating the section. [07:45:54] Yet another reason you should use §ion=new. [07:45:59] Rawr. [07:46:00] Erm. [07:46:03] There are 2 newlines. [07:46:08]