[08:57:56] thedj: Is that JS i18n stuff in 1.16? [09:07:54] JeroenDeDauw: don't think so. it's in the js2 branch only i suspect. [09:09:32] thedj: ok thanks, I'll have to check if I can use it then. [09:10:19] JeroenDeDauw: otherwise, you will probably have to use the same method that the ajax (un)watch uses atm. [09:10:23] but that's ugly :D [10:51:25] Ryan_Lane: When you're around, could you please set the timezone on the usability prototype to UTC? It was in EST last week and is in EDT now, which messes up wiki timestamps, IRC log timestamps and causes "timestamp in the future" warnings when extracting tarballs [11:17:30] RoanKattouw: can i get your opinion on something? [11:17:37] Sure [11:18:12] so when saving snapshots of the iframes html in history states, i'm doing this currently: context.$contents.html() [11:18:49] which does not return the html as it exists. specifically, templates aren't wrapped [11:19:23] do you think i should change that to something like context.$iframe.find('body').html() to ensure we're storing the markup and don't have to reparse everything when we do a undo/redo? [11:21:29] context.$content.html() should Just Work [11:21:36] It should include template wrappers [11:21:52] In fact, context.$content IS the [11:22:31] and we're not doing any trickery to strip out .wikieditor-noincludes? [11:23:32] hmm maybes something else is going on here, now i'm seeing those both being identical [11:23:38] forget i said anything for now :) [11:25:14] I think those wrappers are being stripped/unwrapped in some post-processing step somewhere [11:40:49] adam_miller: Do you have Opera handy? [11:40:58] yes sir [11:41:14] There's this bug that makes me wonder whether Math.min() is broken in Opera [11:41:14] opera 10.something [11:41:19] Awesome [11:41:24] i'll check [11:41:32] how do you suspect it's broken? [11:41:36] Could you answer the questions in https://bugzilla.wikimedia.org/show_bug.cgi?id=22841#c1 first [11:41:42] Well question 2 [11:41:55] (Question 1 requires testing in another browser, don't worry about that for now) [11:42:27] If I was wrong about question 2, could you fire up a debugger and tell me what Math.min( 1, 0 ) is? [11:42:41] Ah [11:42:50] And Math.min( 1, 1 ) too [11:43:17] it definitely does not insert the header rows [11:44:01] But does it eat the top row when you tell it not to create a header row? [11:44:54] RoanKattouw: no i dont think so. the number of rows and cols is always what I ask for [11:45:02] Hm [11:45:29] Are Math.min( 1, 0 ) and Math.min( 1, 1 ) correct? (0 and 1 respectively) [11:45:52] >>> Math.min(1,0) [11:45:52] 0 [11:45:53] >>> Math.min(1,1) [11:45:53] 1 [11:46:00] yesir [11:46:09] Hmph [11:46:19] OK open the table dialog and make sure the headers checkbox is checked [11:46:38] Then enter $j( '#wikieditor-toolbar-table-dimensions-header:checked' ).size() into the console [11:46:50] 1 [11:47:35] Hmm [11:48:43] I seriously can't see what's going wrong here [11:49:09] I suggest you try breaking at Toolbar.js:1514 and stepping through the loop that generates the table code [11:49:21] *RoanKattouw apologizes for monopolizing adam_miller 's time [11:49:45] RoanKattouw: if i uncheck and recheck the Add Header Row option, it will then insert it [11:49:54] ^ [11:50:01] OK so it's got to do with the initial state [11:50:30] Could you check what $j( '#wikieditor-toolbar-table-dimensions-header:checked' ).size() is in this initial state? [11:50:32] $j( '#wikieditor-toolbar-table-dimensions-header:checked' ).size() returns 0 until i do that [11:50:39] Aaah [11:50:55] What about $j('#wikieditor-toolbar-table-dimensions-header').is(':checked') [11:51:07] or $j('#wikieditor-toolbar-table-dimensions-header').attr('checked') [11:52:30] true and true [11:53:35] OK [11:53:39] Thanks for the research [11:53:46] I'll be sure to credit you in my commit summary :) [11:54:07] how lame is it that opera named their js debug tool Dragonfly? [11:55:02] What's its usefulness on a scale of 1 (IE's Firebug ripoff) to 10 (real Firebug)? [11:55:21] 6.5 [11:55:36] what's really lame is that they don't let you use console.log [11:55:53] you gotta use some proprietary nonsense [11:56:24] window.opera.postError [11:56:39] Yeah at least IE supports console.log [12:00:10] RoanKattouw: would template parsing still be happening after the 'ready' event is triggered in jquery.wikieditor.js? [12:00:41] Offhand I don't think so [12:01:16] the initial history state is always saved without the template markup [12:01:18] can't figure out why [12:01:34] Hm [12:01:53] // Let modules know we're ready to start working with the content [12:01:55] context.fn.trigger( 'ready' ); [12:01:56] // Only save HTML now: ready handlers may have modified it [12:01:58] context.oldHTML = context.oldDelayedHTML = context.$content.html(); [12:02:03] That suggests it should be with template markup [12:03:08] This is very weird, from reading the code I'm almost certain wrapping will have happened by the time trigger() returns [12:03:43] Maybe use a debugger and put a breakpoint in highlight.mark(), then step through the code I quoted above to see when it gets called [12:03:51] it totally doesn't though. as soon as you click on the iframe contents, another history state is saved [12:04:04] because it thinks the html has changed from the initial history state [12:07:10] RoanKattouw: do you use the same test page that Trevor sent me? has three templates in it and four headings [12:07:35] I don't use that exact page, no, but I should probably start [12:07:45] I can haz pastie? [12:09:33] http://pastebin.com/zQxMM2i5 [12:10:08] adam_miller: Committed fix for the Opera bug [12:10:22] *RoanKattouw slams head into table for forgetting to credit someone for the second time today [12:10:26] Thanks [12:10:39] i just noticed that the inline template on that test page is not wrapped until i make a change to the text. [12:10:55] you bet. no need to credit [12:13:36] RoanKattouw: figured out the history issue. the ready event is fired before the mark event is [12:14:08] The mark event is fired as a response to the ready event [12:14:11] So that makes total sense [12:14:27] Or does trigger('ready') finish before 'mark' fires? That'd be very weird [12:15:33] "This '''paragraph''' has some crazy whitespace in it. Like this for instance" [12:15:43] adam_miller: I don't see any crazy whitespace in that paragraph, copypaste bug? xD [12:16:10] possibly, but i just sent you the pastbin link trevor sent me last week [12:16:34] Hm [12:16:42] I'll ask him when he gets in, shouldn't be long [12:16:49] so the ready event handler on wikieditor.js finishes before the mark event handler on templateEditor.js is called [12:17:05] That's weird [12:17:17] Could you inspect the call stack for the latter call? [12:17:35] whats the easiest way to do that? [12:18:26] Firebug :) [12:18:39] Place a break point in templateEditor.mark() [12:18:59] And once you've gotten execution to pause there, go to the "Stack" tab on the right [12:19:07] yeah got it [12:19:16] mark() [12:19:16] trigger() [12:19:16] mark() [12:19:16] It'll list a function stack and you can click on functions to see exactly where in the function the call is [12:19:17] ready() [12:19:17] trigger() [12:19:17] (?)() [12:19:17] handle() [12:19:18] (?)() [12:19:49] Could you click on the first (?)() and see which line that's on? [12:21:37] 250 of jquery.min.js [12:22:58] And the second (?)()? [12:23:46] 1098 of jquery.wikiEditor.js which is the trigger('ready') line [12:24:30] And handle()? [12:29:05] adam_miller: What *exactly* is on line 1098? Is it context.fn.trigger( 'ready' ); or something else? [12:29:16] context.fn.trigger('ready') [12:29:35] handle() is line 277 of jquery.min [12:29:56] *RoanKattouw WTFs [12:30:01] I'll look into this after dinner [13:28:37] nkomura: you there? [13:28:45] hi [13:28:50] hi :) [13:31:05] hi peoples [13:35:27] Morning guys [13:39:53] Ryan_Lane_: Poke [13:39:58] Ryan_Lane: When you're around, could you please set the timezone on the usability prototype to UTC? It was in EST last week and is in EDT now, which messes up wiki timestamps, IRC log timestamps and causes "timestamp in the future" warnings when extracting tarballs [13:40:32] RoanKattouw: ugh. i'll take care of that when I get home (6 hours from now) [13:40:39] I should clarify that I'm not complaining about it changing from EST to EDT specifically, but about the fact that it's not UTC whereas all other hosts within the cluster use that [13:40:46] RoanKattouw: I should probably fix that for all the systems. [13:40:52] Yeah [13:40:54] I don't see how it changed from EST to EDT [13:41:36] Probably locale-based or something [13:41:39] so all wikimedia systems are set GMT? [13:41:46] It knows which country it's in and when DST takes effect there [13:41:48] Yes [13:41:58] Well, UTC [13:42:09] ah, right daylight savings time [13:42:17] stupid DST [13:42:32] Although the UK is in BST as opposed to GMT now [13:42:49] So I'm not sure whether GMT does or doesn't include DST, or whether the definition varies [13:43:02] At least UTC is clearly defined [13:43:19] UTC it is then :) [13:43:55] I believe GMT includes DST. My timezone is either -5 or -6 depending on DST [13:47:24] RoanKattouw: so we're going to put the ethnio scrip live again [13:47:32] should be fun [13:47:35] Ethnio? [13:47:42] the recruiting thing [13:47:52] Ah [13:47:59] http://meta.wikimedia.org/w/index.php?title=Special:NoticeTemplate/view&template=ethnio_recruitment_pitch [13:49:00] Am I supposed to see anything beyond textboxes and code there? [13:49:07] Or do I get filtered because I'm not in the Bay Area? [13:50:42] aha [13:55:08] no - you don't see it because you have a cookie on your computer [13:55:16] delete the cookies for that page and it will appear [13:55:17] Ah I'm not an anon [13:55:21] "like magic!" [13:58:44] Hm the JS only seems to check for the ethnio_displayed cookie [13:58:46] Which I don't have [14:00:52] Oh [14:00:53] &&wgUserName===null [14:02:18] TrevorParscal: When I click the button, nothing happens [14:02:30] Oh wait now it does ^^ [14:03:09] oh - so I think that's the "the script hasn't loaded yet so the binding isn't set yet issue" [14:03:19] that must be what nimish was seeing the other day [14:04:18] When I mouseover it does show "javascript:void( Ethnio.show() )" [14:07:40] *RoanKattouw glares at wgThrottle=0 [14:08:06] There's a wgThrottle>0 clause in the onload if [14:08:08] *RoanKattouw removes it [14:08:46] um [14:08:50] don't do that [14:08:59] that's the default [14:09:23] which gets overwritten by /mnt/upload6/centralnotice/enwiki/ethnio/throttle.js [14:09:29] I commented out the wgThrottle>0 clause [14:09:51] Oh right [14:10:00] we need that clause [14:10:14] or it won't turn off when we want it to [14:10:18] I'll put it back [14:10:24] thanks :) [14:10:25] Hm the issue persists [14:10:30] ? [14:10:33] There's clearly href="javascript:..." on the [14:10:47] And the javascript: URL shows up in my status bar [14:10:54] avascript:void( Ethnio.show() ); yeah - that's normal [14:10:57] But clicking the doesn't do anything [14:11:36] hmm [14:11:53] well, I'm interested in fixing it [14:12:07] but I can't recreate the issues on my computer [14:12:52] I was hoping we could have a quick meeting [14:13:05] just to run down the priority list / status of things [14:13:13] adam_miller: you around? [14:13:14] OK [14:13:31] TrevorParscal: i'm here [14:13:38] ahoy! [14:13:52] maybe nimish and I could grab a room with a phone or something... [14:14:29] I'm thinking 15 minute meeting, run down the list of things we are trying to get done / need to get done by tomorow afternoon [14:14:31] works for me...RoanKattouw feel like calling in? [14:14:55] AHA [14:15:04] I found the issue with the Ethnio button [14:15:17] Yeah I can call in [14:15:35] brb [14:15:38] Ethnio.should_display = function() { [14:15:39] return ( ( Math.floor( Math.random() * wgThrottle ) == 0 ) && more crap her [14:15:47] The return value of that function changes EACH TIME IT'S CALLED [14:16:03] Ethnio.should_display = function() { [14:16:05] return ( ( Math.floor( Math.random() * wgThrottle ) == 0 ) && more crap her [14:16:07] The return value of that function changes EACH TIME IT'S CALLED [14:16:16] And each time it's called is on every click [14:16:19] RoanKattouw: yeah it's supposed to randomly show up [14:16:26] oh it's called on every CLICK?? [14:16:36] Yes, *and* on deciding whether or not to show up [14:16:40] *RoanKattouw fixes [14:16:42] hahahaha [14:16:47] thanks [14:19:02] Yay works now [14:19:31] RoanKattouw: ext 655 [14:19:38] TrevorParscal__: did you find a room? [14:19:46] OK, lemme grab my headphones [14:19:47] board room [14:19:53] come on over [14:19:57] I booked it any everything! [14:20:41] adam_miller: so, are you thinking you'll use skype? [14:21:06] TrevorParscal: yes, but i can dial into the conference line if you want [14:21:10] hmm [14:21:25] you could call the office and get ext. 655 maybe? [14:21:47] +1 415 839 6885 [14:21:50] yes [14:22:20] ok, i can do that. now? 3 minutes? 5 minutes? 15 minutes? [14:22:24] Now [14:24:33] roan are you on hold? [14:24:42] "Far end released call" [14:24:47] Lemme call back [14:25:07] It's ringing now [14:25:28] adam call in [14:25:31] TrevorParscal: now you hung up on me to let roan on [14:25:36] then I will conference with roan's extension [14:25:38] sorry abou tthat [14:25:43] figuring the phones out still [14:25:47] He just hung up on me too [14:25:50] roan's ext: 677? [14:25:51] OK that's cool [14:25:53] Yes [14:26:07] TrevorParscal: Did you ask RobH about getting adam_miller an ext or should I? [14:26:16] I will ask about it [14:26:46] ... [14:29:04] Bug #68282 - Cannot Edit from Wrist Watch [14:32:57] xD [14:33:58] WOO HOO [14:34:00] we are live [14:34:08] currently accepting recruits [14:35:26] ^^ [14:35:41] parutron: Did you notice I fixed a ridiculous bug in it 20 mins ago? [14:35:55] the double click!! [14:36:09] i did - thanks RoanKattouw nimish_g and TrevorParscal [14:36:16] most radical central notice team ever. [14:37:04] we have our first recruit! [14:37:07] never gets old! [14:37:32] i cleared my cookie, but i don't see the banner [14:38:14] nkomura: Doesn't show if you're logged in, and it randomly decides whether or not to show up [14:38:25] i'm not logged in [14:38:36] but i haven't tried that many click yet [14:39:05] what's the throttling level? [14:39:31] 40% of the central notice 20% [14:39:35] so 8/100 [14:39:42] they're coming in pretty fast. [14:39:47] i encountered one! [14:39:48] parutron: I think it was 20% [14:40:01] i'm gonna gauge where we are at in a few hours to see if we need to change the throttle or not [14:40:06] for now we are running it for three days [14:41:14] nice, scrolling is working [14:42:01] parutron: pls delete my submission [14:50:03] http://p2p.wrox.com/content/articles/identifying-iphone-safari-user-agent [15:07:49] Mozilla /4.0 BlackBerry8700/4.2.0 Profile/MIDP-2.0 [15:07:49] Configuration/CLDC-1.1 VendorID/-1 [15:07:54] http://na.blackberry.com/eng/devjournals/resources/journals/mar_2007/bdjs.jsp [15:11:18] *RoanKattouw wonders whether anyone got the Sunday @ 2:30 am thing [15:50:53] Sunday at 2:30am ... [15:51:03] I saw it - what was that referencing? [15:55:48] Well, that time never existed [15:56:23] See also http://identi.ca/notice/24706274 [15:57:39] AH! [15:57:46] dang it, I should have known that one [16:01:18] *TrevorParscal goes to verify the blackberry block [16:01:50] nkomura: I sent out an e-mail to you guys and Calcey about the fix for the one Opera bug as well as two blacklisting actions. I'll leave it up to you to instruct them to test again [16:02:10] k [16:02:24] how about the bug reports from last week? [16:02:34] i see it [16:02:34] i haven't gone through each case yet, [16:02:40] I'll look into those [16:02:49] I'll just skim through the open bug list again [16:02:51] thanks [16:02:52] I'm going to verify the blackberry, iphone and ipod blocks against sandbox.6 [16:02:53] In a minute [16:03:05] TrevorParscal: Hold on, lemme actually update s-6 first [16:03:09] k [16:03:12] Or test against a normal or deployment prototype [16:03:27] lets do the sandbox [16:03:37] let me know when it's ready [16:04:03] s-6 is ready [16:06:21] awesomeness [16:11:28] hmm - i got full on iframe an all on an iPhone [16:11:41] Weird [16:12:03] *RoanKattouw sets up a script to alert() the user agent [16:12:08] :) [16:12:56] iPhone produces clean diffs [16:12:57] :) [16:13:07] the iframe is not ediable however :( [16:13:32] http://toolserver.org/~catrope/ua.html [16:16:34] Mozilla/5.0 (iPhone; U; CPU iPhone OS 3_0 like Mac OS X; en-us) AppleWebKit/528.18 (KHTML, like Geko) Version/4.0 Mobile/7A341 Safari/528.16 [16:16:50] and, hello back [16:18:06] Thanks, I'll debug that AU [16:18:08] *UA [16:19:06] Hm $j.browser.name == 'iphone' [16:19:12] RoanKattouw, TrevorParscal; why don't you guys make ethnio recruitment a geonotice? (i.e. it only shows up for IPs within xx of SF) [16:19:36] *RoanKattouw looks away, knows nothing [16:19:57] I say that notice a few hours ago and fixed a glaring error in its randomization thingy just before it went live [16:20:02] That's all I know :) [16:20:10] enwiki already has a lot of the code already written, it would just need to be ported to work here. [16:20:10] Heh. [16:21:45] casey: we will be doing that before our next study for sure - but we're super tight on dev resources right now [16:22:07] Okay, for reference, see the first few lines on http://en.wikipedia.org/wiki/MediaWiki:Common.js/watchlist.js [16:22:11] TrevorParscal: If I want ['&&', false] to work, I have to add && as a supported operator I think [16:22:20] ah [16:22:54] Oh wait that's not it [16:23:00] val can only be a string or a number [16:23:13] ah [16:23:18] dang it [16:23:27] && 0 should work [16:23:32] what if we added a way to do this [16:23:52] so, if typeof val == 'bool' && val == false [16:23:55] or something [16:24:00] so blocking things out looks sane [16:24:58] ideally the notation would be 'ipod': false [16:25:23] so, we could check if browser is false [16:25:27] on line 94 [16:25:35] then we can use nice notation [16:25:39] what say ye? [16:25:53] *TrevorParscal checks blackberry [16:25:54] brb [16:26:01] *RoanKattouw already committed false -> 0 [16:26:14] Yeah ideally it'd just be false, agreed [16:26:19] Lemme do that [16:32:05] TrevorParscal: Done [16:35:51] TrevorParscal: Prototypes and sandboxes 6 and 8 updated with new blacklisting code, please test [16:38:33] k [16:38:38] blackberry blocking is workin [16:38:59] had to try 2 diff blackberry devices, one was the old one that has the plain old mobile versions of things [16:39:05] and the other has the "real" internet [16:39:13] good news, vector worked well on that [16:39:33] Cool [16:39:46] brb - iPod won't get on the network [16:41:55] *TrevorParscal had the URL wrong [16:47:56] RoanKattouw: when highlight marks templates, should it be placing an empty span in front of the template, or actually wrapping the span around the template? [16:48:19] The latter [16:48:27] It shouldn't add empty spans any more [16:48:39] For headers it should addClass() the

, and templates should be wrapped in a [16:49:37] RoanKattouw: ipod blocking is working [16:49:42] testing iphone again [16:49:44] Good [16:50:51] RoanKattouw: i think i'm seeing a spot where it's still placing an empty span before the text [16:51:47] iphone blocking works nwo [16:52:34] using that test article i sent you the pastebin for earlier, the inline template does not get wrapped until you make an edit to the iframe's contents manually. I was poking at it a little and it looks like the marker node is just an empty span for that one the first time mark is called. [16:55:48] Is this inline template inside another template? [16:57:39] no, it's in a paragraph [16:58:11] lorem ipsum dolor sit blah blah{{testing|this=that}} labore et ... [17:00:07] Hm weird [17:00:18] I'll poke at it later [17:00:25] I'm taking a mental break reading the news :) [17:01:00] When will the JS2 i18n stuff [ http://www.mediawiki.org/wiki/JS2_Overview#How_to_include_localized_messages_in_your_JavaScript ] be available on trunk? [17:01:35] RoanKattouw: loving the new 'ipod': false syntax! [17:01:52] Whee [17:01:54] JeroenDeDauw: it's going to take a while [17:02:54] JeroenDeDauw: I'm advocating and facilitating that as hard as I can, but there's a gap to be bridged between Michael Dale and Tim Starling in coding style, priorities and motivations [17:03:21] plus it just takes allot time to get things done in MediaWiki (perhaps it should not, but it does) [17:04:04] We are going to do a branch of 1.17 which integrates just the script loader and basic JS support for it [17:04:08] soon [17:04:17] as in next month [17:13:59] TrevorParscal: is event.jQueryNode something we invented? [17:14:12] Yes [17:14:26] This is a hell of a break you're taking RoanKattouw [17:14:54] I know [17:15:09] I need to have i18n stuff in the JS of an extension. What's the recommended approach for now? [17:29:17] JeroenDeDauw: did you take a look at how the UsabilityInitaitve extension works? [17:29:44] That's gonna take him all night, we abstracted that into obvilion :) [17:29:48] *oblivion [17:30:34] JeroenDeDauw: in short, we load the messages into the page using an inline script [17:30:53] we use a function called loadGM to set the messages in the inline script [17:31:08] and another function gM to get them by key [17:31:23] *TrevorParscal wants to lead the way to gettext [17:32:07] No gM() is no more [17:32:12] It's called mw.usability.getMsg() now [17:32:14] oh [17:32:16] right [17:32:21] sorry [17:33:22] RoanKattouw: actually if you take the SF article and remove the infobox settlement you see the reshuffling text [17:33:29] Thanks [17:33:47] so it's not nested templates...I un-nested all of them in the article and still happens [17:40:04] RoanKattouw: and to make things even more fun, the shift is definitely different depending on which browser I use [17:40:13] *sigh* [17:40:18] Which browsers does it happen on? [17:40:21] even tho we're just supporting FF [17:40:32] FF and Safari, those are the only ones I've checked so far [17:40:56] FF 3.5 [17:42:37] Ryan_Lane , if you're around, it looks like thumbnails are not generated on the Commons prototype [17:42:41] or, at least, not displayed [17:43:18] RoanKattouw did you write getElementAtCursor? [17:43:26] No wasn't me [17:43:33] TrevorParscal? [17:44:37] I think nimish_g wrote it [17:45:10] it was nimish, I sort of got him started on that path, but he implimented it [17:45:35] yeah I wrote it [17:45:35] nimish_g: getElementAtCursor -> why are you always returning the parentNode of the start container? [17:45:37] wassup? [17:45:45] because the start container is a textnode [17:45:53] it's *always* a textnode [17:46:34] well, not always... it's a textnode inside the element I want for all the template stuff [17:46:46] for empty lines, it's a

[17:46:50] yeah i think there are occasions where it's not a textnode [17:47:24] yeah, the only time I saw that was when there were empty lines [17:47:25] i'm going to try changing that to only return the parent if it's a textnode [17:47:29] ok [17:47:46] a TextNode, actually [17:47:57] .nodeName == '#text' [17:48:17] the code you wrote to fix enter keypress events in the template text was breaking because it doesn't always return the element you were assuming it would [17:48:42] aaaah [17:48:57] i changed teh second conditional in the keydown handler to this: [17:48:58] $evtElem.hasClass( 'wikiEditor-template-text' ) || $evtElem.children( '.wikiEditor-template-text' ).length == 1 [17:49:00] and it worked [17:49:32] but i'm thinking maybe thats a hack and we should just fix the getElementAtCursor function [17:50:00] yeah, probably. that second one would capture enters in places we don't want [17:56:58] k [17:57:06] fixed the label update issue [17:57:14] beware, capsule markup changed slightly [17:57:35] also there's a bit of a performance improvement because of the way I'm building out capsules now [17:59:00] Yeah I think that's where most of the remaining slowness was [18:05:45] there were like 7 or 8 dom manipulations including text. append, prepend and addClass, now there's just 1 prepend [18:06:38] So bascially you're building it first, then throwing the whole thing in the DOM? [18:07:36] nimish_g: committed that fix, now just need to deal with the cursor being repositioned at the top of the text whenever you press enter [18:12:57] adam_miller: are you using OSX? [18:13:08] but, of course [18:13:27] hm...do you see the cursor moving to the front of the template whenever you enter *any* text? [18:13:32] me and TrevorParscal have that [18:14:15] oh yeah, forgot it was broader than the enter key [18:15:51] yeah. if you remember I had a commit a while back where I bailed out of the 'mark' function if we'd already marked that node [18:16:07] it broke the hell out of the ntoc, but solved the cursor jumping issue [18:16:57] my proposed solution, if we can't find one soon is to just not run the mark code *at all* for templates except for on ready [18:17:25] Which means template wrappers don't update at all, ever [18:21:43] RoanKattouw: do you mean? when I did that, the only thing that was broken was new templates didn't wrap [18:21:48] *what do you mean? [18:22:17] Well when you change the name in a template call, there'll be no update [18:22:30] I think the template dialog code still checks and updates the model if needed [18:23:16] yeah, the template name would only get updated if you expand and contract the arrow [18:23:54] It sounds like an acceptable workaround to me, if we wanna reduce the feature set in that way [18:24:14] RoanKattouw: what triggers the mark event? i can't find it [18:24:43] adam_miller: highlight.mark() [18:25:03] a "realchange" event triggers it [18:25:08] or a "ready" event [18:25:23] I was thinking we could do the same thing we're doing for keypress events in the templates - check the current element the caret is in, and skip the mark event for the templateEditor if it's inside a template [18:26:12] Problem is you're dealing with delayed change events, which are triggered by a set of change events [18:26:27] that's lots of extra processing...there should be a way to just use the 'realchange' event and just not do stuff if the marker is a template type [18:28:10] ha, and when i attempted to implement this, it just completely stripped the markup from around the template when i edited it :) [18:28:40] haha [18:29:11] comment out line 42 and 43 from highlight, and don't use NTOC [18:29:28] and things work fine [18:31:11] rather, don't pay attention to NTOC and it works fine [18:32:43] hey! i put a lot of blood sweat and tears into that NTOC [18:32:55] i'm not disabling it [18:33:36] haha [18:36:16] alright dinner time. i may be back later for some more caret jumping fun and dialog fixing [18:36:18] just don't pay attention when it updates wrong =P I mean we could hardcode "if marker.type == 'template' return;" or whatever we named those things, but I'd rather not [18:36:21] cool [18:47:04] RoanKattouw, I tried to purge at least 3 different image pages :) [18:47:20] And none of the thumbs show up? [18:47:25] Maybe they're misconfigured [18:47:34] Thumb generation that is [18:47:58] nothing shows [18:48:08] http://208.80.152.249/w/index.php?title=Special%3ASearch&search=paris&fulltext=Search [18:48:38] Yeah the search index needs rebuilding too [19:02:54] flipzagging, hi; are you all set up to make updates on the Commons prototype, or do you need anything else? [19:05:41] i don't believe I have any instructions about that [19:05:54] I figured I would deal with it after tuesday [19:07:16] guillom: I guess Roan is the person to ask? [19:07:36] If you need stuff, yes [19:07:51] guillom: Could you e-mail me about the thumb issue and the search issue so I get both done tomorrow? [19:07:54] RoanKattouw: so can I just ssh into this prototype? [19:08:02] sure [19:32:06] Yes [19:32:18] If your user/pass don't work or you don't have them, I can add them for you [19:32:28] Also, if you're not a sudoer, poke someone who is [19:34:10] grr - spellcheck="false" on a contentEditable span SHOULD work, but it's not working [19:36:16] RoanKattouw: hostname? [19:36:32] 208.80.152.249 [19:36:37] heh [19:36:40] I don't think it has a DNS yet [19:37:16] Oh wait commonsprototype.tesla.usability.wikimedia.org ^^ [19:37:26] We probably want a shorter alias for that [19:38:17] k, that's fine for now [19:38:21] who is the DNS person here [19:38:25] RobH [19:39:18] commons-proto.wikimedia.org ? [19:39:30] commons.usability.wikimedia.org is not much better [19:40:19] hey mdale. Where are you, sxsw still? [19:40:38] I just got back will be in the office tomorrow. [19:40:57] i thought you were going to Mexico [19:47:49] in a week or so [19:55:53] hi howief [19:56:02] hey werdna [19:56:19] may not be super responsive, i'm on a train. [19:56:23] no worries [19:56:29] thanks for rolling out the changes to LT [19:56:36] i haven't had a chance to look at it yet [19:56:44] yvw [19:56:45] but i'm looking forward to seeing the changes [19:57:02] yay [20:03:53] werdna: [20:03:56] i think it looks great [20:04:07] the interaction seems pretty intuitive [20:04:48] not for this round of changes, but next round of changes i'm thinking one indicator that replies are related could be by having the horizonal line either lighter or even absent [20:04:55] i'll send you a mockup if that doesn't make sense [20:05:15] but i'd like to get feedback from eugene and philippe [20:45:07] howief: nod, I wan to wait just a little longer [20:45:30] there are a few bugs reported by our favourite troll Natalie, and I'd like to get those dusted off before expanding the deployment [20:45:47] sounds good [20:48:16] maybe wait another day or two for anything else to get shaken out [20:48:32] good idea [20:48:33] there was some fedback that the grey boxes might look nicer if they had softer lines, and were blue instead of grey. [20:48:36] thoughts? [20:48:51] cool be interesting to try [20:48:57] hard to say without looking at a mockup [20:49:00] nod [20:49:06] I'll have a play with it and send you a screenshot [20:49:21] cool - thanks [20:49:30] have a computing lab in 10 minutes [20:49:36] maybe I'll learn about if statements! [20:49:40] hahaa [20:49:50] :) [20:50:36] better get over to that [20:50:47] have fun [20:50:50] talk to you later [21:16:29] werdna: If you stopped adding/removing features, there'd be fewer bugs. :P [21:20:35] werdna: Are there any plans for deployment on actual wikis? [23:36:40] Natalie: yes, meta in the next few weeks [23:36:43] hopefully before april