[02:53:12] where does jquery.mwCoolCats.js live? (i have this feeling that there's more than one copy out there) [02:55:27] who knows how UploadWizard works? guillom, neilk, raindrift is all? [03:00:31] I know a little bit about it, although now is not a good time for the names you mentioned (they are either asleep or at home) [03:00:46] there's a new script in core that does ajaxCategories for normal wiki pages [03:04:11] hrmmm [03:05:16] Krinkle: so, one of the many questions i'm tackling is why i see cats suggested which don't exist on commons. i'm thinking it's coming from lucene and it's using the wrong index? [03:05:30] dunno [03:05:58] http://svn.wikimedia.org/viewvc/mediawiki/branches/wmf/1.17wmf1/extensions/UploadWizard/ [03:06:04] that's what runs on Commons [03:06:09] oh, i was using trunk... [03:06:20] The cluster doesn't run trunk :P [03:06:40] i know, but there can't be that many people hacking that extension on trunk besides those 3 :) [03:06:43] UploadWizard has been deployed not long ago though, so it's should be similar. I don't remember changes to CoolCats [03:06:49] right [03:07:15] Krinkle: have time to query DB? toolserver or elsewhere [03:07:21] sure [03:07:31] *jeremyb put a request in #-toolserver [03:50:57] Krinkle: Myra: moving back here? [03:51:14] What. [03:51:14] so, there's at least 3 tracking bugs for UploadWizard for different versions [03:51:21] how do i know which one to use? [03:51:32] Krinkle already filed a bug. [03:51:45] for, e.g. blocking for 26411 [03:51:50] orly [03:52:01] I don't care which UW version, we don't have to care. Just file it and do it in component UploadWizard [03:52:10] https://bugzilla.wikimedia.org/30718 [03:52:20] manager will decide priority and milestone for which it should be ready [03:52:31] I didn't even know UW had versions. [03:52:34] i mean which to block against like 27260 [03:52:43] !b 27260 [03:52:43] --elephant-- https://bugzilla.wikimedia.org/show_bug.cgi?id=27260 [03:52:59] Oh. [03:53:04] yeah, it's a little weird :P [03:53:08] I know you meant that, it's fine without a blocker [03:53:08] I don't think there are any UW deployments planned. [03:53:18] 27801 has no dependencies [03:53:19] Or UW releases or whatever. [03:53:22] it's deployed all the time for the last few weeks [03:53:46] and !b 27800 has some [03:53:49] !b 27800 [03:53:49] --elephant-- https://bugzilla.wikimedia.org/show_bug.cgi?id=27800 [15:43:02] Strategy wiki. Where the language-selector of the Welcome message is bigger than the welcome text itself. [15:43:31] http://strategy.wikimedia.org/wiki/Thread:User_talk:Krinkle/Welcome! [15:45:13] http://strategy.wikimedia.org/w/index.php?title=Thread:User_talk:Krinkle/Welcome%21 [15:58:57] RoanKattouw: https://commons.wikimedia.org/x (https), but the 404 error thinks it is on http and redirect to that as such. Should use current protocol instead and upload.wikimedia links protocol-relative instead of hardcoded http [15:59:04] Not sure where that 404 error is kept though [15:59:27] Oh, right [15:59:31] File a bug and assign to me [15:59:33] ok [15:59:37] I just found two more https bugs on Commons [15:59:51] And one (WikiMiniatlas from meta) can't be fixed till we deploy https on meta [15:59:56] Which will happen on Tuesda [16:00:14] well,it can be fixed. And afaik is fixed now. It dimly loads it from secure/meta [16:00:20] simply* [16:00:27] Yes, but only if you access commons via secure [16:00:32] Not for https://commons [16:00:41] it should check https not https://secure [16:00:48] oh right, wgServer doesn't have https on https://commonswiki [16:00:52] Once meta has HTTPS, that whole if-else can be killed and we can just set var metaBase = '//meta.wikimedia.org'; [16:00:55] That too [16:00:57] But it's Ok [16:01:01] yeah [16:01:08] We can leave it broken over the weekend and fix it on Tuesday [16:01:21] it just gets it from non-secure, not that broken. [16:01:31] It's mixed content [16:01:35] So it needs to be fixed [16:01:41] The other mixed thing is geoiplookup [16:01:45] I'm filing 3 bugs now [16:02:01] Your 404 issue, WikiMiniatlas, and geoiplookup [16:02:24] I'm filing 404 [16:03:37] OK [16:04:13] lol, bugzilla's URL field doesn't linkify https links [16:04:20] https://bugzilla.wikimedia.org/show_bug.cgi?id=30733 [16:04:37] bug bugzilla [16:06:58] Krinkle: Re https://bugzilla.wikimedia.org/show_bug.cgi?id=30734 I may need your help to identify other instances [16:10:19] RoanKattouw: Yeah [16:11:39] Ah nice. https://bugzilla.mozilla.org/show_bug.cgi?id=120603#c1 They have a thing for "Summary: x -> y" [16:12:48] ? I don't see it [16:23:30] RoanKattouw: A little bar directly below the first comment. [16:23:36] Summary: [RFE URL field should be clickable ??? [RFE] URL field should be intuitively clickable [16:24:16] Doesn't display for me [17:30:18] Krinkle: Did I ever send you the pictures you took in SF with my camera? [17:30:26] yes, on the plain [17:30:31] plane* [17:30:40] Oh ,right [17:30:54] You copied it straight off the SD card on the IST-AMS flight [17:45:45] RoanKattouw: btw, any idea what's causing this ? http://www.mediawiki.org/wiki/Special:Code/MediaWiki/96013#code-comments [17:45:53] One of the gadgets isn't showing up in the prefs [17:46:01] They're not hidden [17:46:07] (!isHidden()) [17:46:24] Strange [17:46:50] Can you make an XML dump (Special:Export) of the three Gadget_definition: pages and send it to me? [17:47:47] ok [17:51:02] RoanKattouw: http://toolserver.org/~krinkle/tmp/KrinkleDev-20110903175025.xml [17:51:13] https too :) [17:51:13] https://toolserver.org/~krinkle/tmp/KrinkleDev-20110903175025.xml [17:51:24] experimental on ts [17:51:28] yay [17:51:50] ERROR: certificate common name `*.toolserver.org' doesn't match requested host name `toolserver.org'. [17:51:59] wget is stupid [17:52:16] just save as in browser and Special:Import ? [17:52:22] I downloaded it [17:52:26] I'll import it later [17:57:06] RoanKattouw: Where should I note down about 1) exception throwing on NS_GADGET_DEFINITION edits that are not valid, 2) normalization/validation/prettify ? [17:57:33] Backend section of the task management [17:58:53] btw, I got an idea. In one of those hooks you can deny an edit, right ? Just like edit conflict triggers such event (or permission denied). I was thinking of throwing up an error if the page content is not valid JSON. (ie. missing } or unclosed " ). [17:59:04] If we'd normalize to {} (since it's invalid) the user would have to revert that and fix it [17:59:39] and simply show the edit page again (not save the edit) [18:00:52] Yes, that was my idea all along [18:00:56] ok [18:02:01] If we do it in PST, I don't have to do any normalization in the front-end. I just get the metadata as json object from API, show the form, apply modifications to object and stringily back to ApiEdit (which then normalizes and prettifies it) [18:02:45] Yeah [18:02:52] I'm not yet fully sure how I'll do normalization [18:03:09] normalization would include things like removing duplicates and sorting of array values ? [18:03:23] ie. what resource loader does as well internally. [18:03:44] Normalization is essentially json_encode(json_decode($str)) [18:04:06] Hmm, yeah we could do those things too [18:04:23] encode/decode just undoes prettification afaik [18:05:09] so normalization first, then to database and prettified normalized version to page text ? That's going to be interesting [18:05:35] Well, FormatJson::encode(FormatJson::decode($str), WITH_PRETTIFICATION); [18:05:37] or something [18:05:54] Although, hmm [18:06:01] database wouldn't be prettifyd right ? [18:06:16] gadgets table that is, text table would be [18:06:27] Yeah, the gadgets table would probably need a non-prettified version [18:24:49] Anyone got Windows/IE at hand ? [18:46:07] Krinkle: I have a VM [18:46:12] IE8 on XP [18:46:30] http://jsbin.com/ahaxe [18:46:39] I wonder if that plays well in IE [18:47:08] ie. enter a bunch of text and it expands with a little head room (but the head room should always be the same, not grow exponentially) [18:47:23] and decrease as well when removing 1 char (backspace) or multiple (select and backspace) [19:04:01] *RoanKattouw tries that in IE [19:05:46] Works perfectly [19:06:01] In IE8 on XP [19:26:10] RoanKattouw: Thanks [21:06:35] hi Ryan_Lane! emailed the toolserver & wiki-research-l lists, about to hit a few more (re hackathon) [21:06:55] awesome :) [21:11:11] I'll hit up gnocode, nolaphp and net2no [21:12:32] Ryan_Lane: sweet [21:12:50] *sumanah sends to mediawiki-l [21:14:30] sumanah: working on labor day weekend eh? :D [21:15:12] Ryan_Lane: yeah. Just feel an urgency to get these invites out, because it's 6 weeks away and people's calendars close up [21:16:11] *Ryan_Lane nods [21:16:14] I'll send some emails out today too :) [21:31:00] Ryan_Lane: ok, hit wikitech-l. [21:31:47] as in, I just hit wikitech-l