[11:40] hey luca. I suddenly have 20 minutes free. :-) I have a few follow-up questions about your reply (whicvh I really appreciate). Do you have a moment, or are you lunching? [11:41] sure, guicaht? [11:41] guichat^ [11:45] luca thanks yes [11:45] sorry didn't see [12:25] luca: got a sec? [12:30] guihelp: I need to retrieve the env object from within a viewlet. Is there a way for viewlets to access their own viewContainer? [12:31] frankban: shouldn't the view container give the viewlet anything it needs? Pass down cfg or helpers? [12:31] frankban: i need to do the same thing and have modified the call to render to accept it. hold on and i'll paste my changes [12:33] frankban: http://paste.ubuntu.com/5804591/ [12:34] bac: why not just pass in the db? Seems like this connects the two classes a lot now for testing and such? [12:34] bac: very nice, thank you [12:37] rick_h: how to retrieve, e.g., the env from the db then? However, perhaps we could just pass ViewContainer.options? [12:37] frankban: yea, I was seeing that. This is added to the root class vs a class that's later inheriting I guess. [12:38] frankban: I like the idea of an optional cfg second param instead [12:38] render(model, options) [12:39] frankban: but i'm passing the viewcontainer. can't you get options and env from it? [12:39] sounds good, also extended to the other viewlet methods I guess. what is options? viewContainer.options? [12:39] rick_h: ^^^ [12:39] frankban: yeah [12:40] bac: render(model, {db: ciewContainer.get('db')}) [12:40] rick_h: sorry i didn't see your comment [12:40] frankban: options is just a arg name. [12:40] frankban: call it cfg, foobar, whatever. Basically a space for passing in any optional data? [12:40] frankban: I guess I'm just looking at this diff and not the big picture of how it's working. So ignore me if it doesn't apply I guess. [12:40] rick_h: bcsaller, hatch and i discussed this yesterday and thought it may need to be broader, but i see your point [12:41] I just get twitchy when I see something getting the full instance of something else as those two classes are now tied at the hip vs something more generic [12:41] rick_h: AFAICT this is a framework, our render is called for us, so I guess we must decide what to pass as second arg [12:42] frankban: right, then nvm I guess. I've not used it yet so was looking at things like a Y.View calling another Y.View where it has control over how render is called [12:42] frankban: is the init custom then? [12:42] frankban: e.g. could the viewlet be given optional init config and then it's already there for this.render()? [12:45] rick_h: you can pass options to the viewContainer initializer (an Y.View). a viewlet is more a configuration object, with hooks/methods called by the viewContainer. It seems nice to me if the viewlet can receive (or store, as you suggested) those option values. I am also fine with Brad;s solution of passing the whole viewContainer, even if I guess we will be mostly interested in its options [13:01] rick_h: I'm in a meeting now for an hour but can we catch-up after? [13:02] luca: sent an email on your way. Feel free to reply to that or we can chat after your meeting. [13:02] rick_h: ah brilliant, cheers :) [13:02] BradCrittenden, rick_h: what I described: http://pastebin.ubuntu.com/5804671/ [13:04] frankban: cool. I like that way it's limited to bits of data vs the whole class and the temtation to try to share methods/etc across. Seems less 'dirty'. [13:07] frankban: looks good === BradCrittenden is now known as bac [13:16] anyone up for a second review of https://codereview.appspot.com/10677043/ ? [13:28] teknico, any takers? if not, please invoke j u j u g u i :-) [13:29] luca, I don't understand this line from inspector behavior: [13:29] "You can scroll the inspector but the scroll goes underneath the dark grey divider line" [13:29] gary_poster: nope, proceeding with deployment of Secret Weapon [13:30] cool teknico :-) [13:30] luca, and that line seems key :-) [13:30] * gary_poster goes to another call, but more explanation would be great [13:30] gary_poster: we up? === BradCrittenden is now known as bac [13:31] brb [13:31] we are bac [13:31] cool [13:31] `juju deploy jujugui`: anyone up for a second review of https://codereview.appspot.com/10677043/ ? [13:31] heh === benji__ is now known as benji [13:31] teknico: hi, I'll take it === benji is now known as Guest32120 [13:32] gary_poster: am i in the wrong place? [13:32] benji: I mean, thank you ;-) [13:32] whoops [13:32] bac, maybe. 1c5347001459ec985658a81e5c3a019abc715e05 ? [13:33] bac from calendar [13:33] no, did it change with the new calendar invite? [13:33] benji: thank you [13:44] rick_h: have a minute for some browser wierdness discussion? [13:46] benji: sec, finishing up standup [13:46] l [13:46] k [13:46] now everyone knows I'm using a QWERTY keyboard [13:46] doh! [14:00] bac, you still here? [14:00] is anyone able to "make code-doc" successfully in trunk? [14:01] teknico: nope [14:02] benji: free now, guichat? [14:02] rick_h: I figured it out. Your contribution will be noted for posterity. [14:03] woot! I love being awesomely helpful by not attending! [14:03] I guess 'by absence' would fit better [14:05] rick_h, hatch, either of you have idea on what to grep for for that yuidoc failure (http://pastebin.ubuntu.com/5804801/) [14:05] ? [14:07] gary_poster: looking [14:08] thx [14:16] gary_poster: try subapp [14:16] er [14:16] subclass [14:16] just thinking [14:17] hatch you mean grep for subclass? [14:17] yeah sorry [14:17] it's possible that the class isn't defined if ther is a subclass of juju.something [14:18] brb [14:24] gary_poster: any luck? [14:24] hatch, if you comment this out then yuidoc stops complaining [14:24] http://pastebin.ubuntu.com/5804847/ [14:24] but [14:24] then there is nothing in the yuidoc directory (if you move any old ones aside first) [14:25] yeah that's definitely not right [14:25] umm can you bisect to find the failure [14:25] er... [14:25] check out the last revno [14:25] I can't--more mtgs--but maybe teknico? :-) [14:26] I'm looking at where the error is thrown in node_modules/yuidocjs/lib/docparser.js and it makes no sense :/ [14:26] console.log(modules) [get a nice fancy list] console.log error'ing module [it's not in the previous list...] [14:28] for some reason it's trying to find a 'juju' module but there isn't a 'juju' module file. [14:28] hatch: when can we have a one-on-one about viewlet/databinding? [14:28] nevvvaaaa [14:28] nice :-) [14:28] ok but really pretty soon [14:28] just waiting for the morning rush to slow down around here :) [14:29] rick_h: the first doc comment gary_post3r pointed out, in apps/subapps/browser/browser.js , says "@module juju" [14:31] bac: are you able to make a branch and commit with that view-container update? I can if you arent' able to [14:31] hatch: frankban has come up with an alternative that looks better [14:32] oh? I thought that was pretty clean :) what's the alternative? [14:32] http://pastebin.ubuntu.com/5804671/ [14:32] i haven't integrated his change yet [14:32] teknico: gotcha. Yea, guess there's not really a juju module to be a part of. Ends up just a namespace. [14:34] teknico: do the docs go anywhere? [14:34] I must be blind today [14:34] bac: frankban hmm I don't really like that as much because then you don't have access to the viewlet for anything [14:35] rick_h: they should go in the top level yuidoc/ dir [14:35] rick_h: but it does not get created anymore [14:35] sinzui: Can we do a mid-imp review of https://code.launchpad.net/~abentley/charmworld/config-storage/+merge/171801 ? [14:35] teknico: so I changed it to https://pastebin.canonical.com/93510/ and no error [14:35] teknico: but no docs [14:36] hatch: other way around? [14:36] hatch: isn't this the viewlet getting access to the viewContainer? [14:36] hatch: it gives what we need with less coupling. [14:36] yeah.....ok you're right I read this wrong [14:36] the context will still be the viewlet [14:37] * hatch goes to grab his coffee [14:37] rick_h: yeah, something else is amiss [14:37] gary_poster: the divider line is the dark grey line under the "upgrade" available notification. [14:42] http://gizmodo.com/how-a-fridge-full-of-beer-that-only-unlocks-for-candian-596858353 [14:42] rofl [14:43] teknico: ok want to chat now? [14:44] I have another call in 15 but we can probably get it done in time [14:44] hatch: yes please [14:44] guichat is open [14:47] sinzui, abentley, jcsackett, rick_h could one of you review these two small MPs: https://code.launchpad.net/~adeuring/charmworld/1192559-forgotten-jenkins-results/+merge/171825 https://code.launchpad.net/~adeuring/charmworld/1192544-dont-hide-charms-with-processing-errors/+merge/171756 [14:47] adeuring: looking. [14:47] thanks [14:49] adeuring: could FailingStoreProviderResults be a function? [14:50] adeuring: You're replacing a function, so that seems intuitive to me. [14:50] abentley: no, when you pass a function to patch.object(), it is called during the call of patch.object() [14:50] adeuring: Not in my experience. [14:51] abentley: well, I tried it... [14:51] adeuring: See test_store.py for some examples. [14:51] * adeuring is looking [14:55] bcsaller, when you are here, could you take bug 1195354? made card in miscellaneous [14:55] <_mup_> Bug #1195354: Export exports relations incorrectly [14:55] teknico, is "Broken tests for pyJuju sandbox" landed yet [14:56] gary_poster: it is, just yet [14:56] moved the card [14:56] abentley: got it. I misunderstood the parameter "new" [14:57] gary_poster: kapil's email said relations: [14:57] - [blog, [db, memcached]] [14:58] one line change either way I think [14:58] adeuring: The changes you've made to the tests in 1192544-dont-hide-charms-with-processing-errors look wrong to me. You should be testing that charms with errors are not hidden, not changing them to refer to store errors. I believe the store error stuff already has tests. [14:58] bcsaller, xchat might have eaten what you said, I am afraid, but I think you are saying that this will be easy :-) . Maybe also that you need to clarify something with Kapil? [14:59] both those things. My orig email suggested what the bug says but he offered back that its a list rather than ':' delimited [14:59] abentley: OK, having tests that charms with errors are shown make seinse. But I could not find tests for the store_error case. Maybe I missed them though... [14:59] I can do either though, like I said, one line change :) [15:02] adeuring: in search.py: test_api_search_charm_with_store_error, test_search_charm_with_store_error. In models.py, test_find_charms_default [15:03] hey hazmat, could you connect with bcsaller on list vs bug above? [15:03] abentley: ok, so I'll just change the originals tests that charms with processing errors are not hidden. [15:03] adeuring: +1 [15:04] gary_poster: canonicaladmin for you [15:04] adeuring: r=me on 1192559-forgotten-jenkins-results [15:04] abentley: thanks [15:05] sinzui, am I right that manage.jujucharms.com will integrate with SSO? [15:05] gary_poster: it does already. The login function should work for anything in ~charmers [15:05] sinzui, and are there any others on the charm side of the world that you know of? [15:06] rick_h, when I click on login, I get a 503 Service Unavailable [15:06] No server is available to handle this request. [15:06] gary_poster: hmmm wfm [15:06] rick_h, ok worked that time [15:06] gary_poster: looks like new SSO layout/etc [15:06] yeah [15:06] gary_poster, I am logged in [15:07] bac, accepted :-) [15:07] gary_poster: oh, and another. my sick day from a while bak [15:07] back [15:07] gary_poster, I just logged in [15:08] gary_poster, I think the bug here is that we did not test what someone who has not role to login sees [15:08] sinzui, ack works for me now as I said. could you please add manage.jujucharms.com to bug 1184961 [15:08] so it gets an icon [15:08] doe SSO [15:08] for [15:08] gary_poster, ~charmers and the m.jc.com developers can loggin [15:09] gary_poster, logging in lets you do charm QA. [15:09] sinzui, ack cool thanks. my main point is that icon bug. could you add yourself to the bug? [15:09] I mean, mj.c [15:10] rick_h: free to talk? [15:10] luca: sure thing [15:10] luca: taking over guichat [15:11] gary_poster, yellow/ui group appear to be the only staff that are NOT in https://launchpad.net/~juju-jitsu/+members [15:11] sinzui, lol [15:12] sinzui, you saw the icon bug request, right? :-D [15:12] gary_poster, I think your people would be helped by membership...and I will look into what happens when you are not permitted to sign in [15:12] bcsaller, re emai. those are separate services.. - [blog, [db, memcached]] is two relations in one line.. no qualifiers.. qualifiers are done as 'service:rel_name' [15:12] gary_poster, No I didn't see it [15:12] number? [15:12] hazmat: ah, must have misunderstood, thanks [15:12] sinzui, bug 1184961 [15:13] thank you. I will look into this. [15:14] thanks sinzui [15:15] ha ha, the deploy was so fast it happened as I was testing login/logout [15:16] abentley, your login/logout link staste bug still exists and somewhat confusing when a deploy happens [15:16] abentley: i updated the branch [15:17] adeuring: r=me then. [15:17] abentley: thanks [15:23] adeuring: I had already approved it, conditional on fixing the tests. [15:25] sinzui: That's weird. I couldn't reproduce it. [15:26] abentley, It happened during the rollout. It was fast. I cannot make it happen now [15:27] sinzui: Can we do a mid-imp review of https://code.launchpad.net/~abentley/charmworld/config-storage/+merge/171801 ? [15:27] oh, and it just did... abentley. I just logged out. I had to force reload to see login again [15:27] * sinzui reads [15:28] gary_poster: jenkins failed but it seems unrelated to my branch, can we retry? [15:29] teknico I will retry if you look at video and see if you learn anything, and maybe doc in fole. :-) [15:29] file [15:30] gary_poster: oh, the video, I keep forgetting that. what do you mean by "doc in file"? [15:33] I guess "document in a file anything I learn" [15:36] teknico, I was thinking like your FIXMEs [15:36] Makyo, coming to meeting sorry [15:41] rick_h: got a minute for a chat? [15:44] hatch: rgr [15:44] in guichat [15:45] hatch: trying [15:50] jujugui call in 10 kanban now [15:56] sinzui: how's that review going? [15:58] abentley, In the last in your branch (# This is not acceptable to Elasticsearch, because config is a list.) you demonstrate that 'e' was not found. I think it was deleted by the exception block. Should it be? If the config options is already in the right format, I would have indexed it after a no-op [15:59] Oh [15:59] it was options that was a list [15:59] abentley, we have configs that that a list? [15:59] do we also have configs that are a string? [15:59] hatch: do you have a minute for a chat after the call? [16:00] sinzui: can i grab you to chat right after the juju gui standup? probably 15 min? [16:00] jcsackett, yes [16:00] cool, thanks. [16:02] gary_poster: the underlying problem with the Jenkins failure is a timeout with "Charm API server did not respond", so maybe a network hiccup [16:03] abentley, I think my questions are moot. We don't support config being anything other than a dict with a single item of 'options': {}. Anything else is ignored now. your branch continues to ignore it, but we also get an exception in ingest to document the charm is bad [16:03] teknico, ah ok. I think they had a deployment [16:03] jujugui sorry got lost in irc and mail. coming [16:03] sinzui: I forget what the precise mapping issue I ran into yesterday was. But we had something that is supposed to be a map in our format, and instead it was a list. [16:04] sinzui: This originally just aborted the migration, which is unacceptable. So I handled it by swallowing the exception. [16:05] sinzui: But since this is a migration, any existing document would be in the old format. So we should not retain it. We should delete it. [16:05] abentley, understood and agreed. We let users craft the yaml by hand. I suppose proof should/does callout this issue [16:07] sinzui: Right. [16:08] abentley, the migration script does a print. Shouldn't it be a log because for webops to investigate. [16:08] or maybe you expect webops to check the juju-log [16:09] sinzui: I don't think that migrate has logging enabled. I should check that. [16:09] sinzui: yeah, no references to 'log' at all. [16:10] abentley, I think all stdout is captured by the juju-log. I think I saw the issue in the log when we had migration 007 bow up [16:10] blow up [16:10] sinzui: Right. I meant that I don't think the migrate script has python logging enabled. [16:11] okay. I follow. That is a nice to have. I think your branch is in a good state. and +1 for documenting what a DISTRO_CHARM_DATA looks like [16:12] * sinzui thinks bzr_branch leaked through, but is nice ti see as an addition [16:12] sinzui: Cool. Yes, DISTRO_CHARM_DATA allows me to justify the changes to the old JSON API. [16:13] sinzui: Okay, can we talk about how we would re-implement migration? [16:14] yeah [16:20] frankban: i'll be right back then we can chat [16:20] hatch: thanks [16:20] thanks hatch. i had not merged from trunk and didn't have ghost-inspector yet [16:20] to look at event handling [16:20] sinzui: want to join me in tinyurl.com/guichat? [16:21] jcsackett, in a few minutes, I am in another hangout [16:21] ok. [16:23] bac: you'll also want to create your own object similar to the one on line 126 of ghost-inspector.js and then include it into the controller [16:24] you do that by adding it to the array on views/service.js:1176 [16:24] I can outline in more detail if needed [16:24] frankban: guichat? [16:24] hatch: sure [16:25] sinzui: when you're free ping me and i'll send an invite; guichat was needed. :-) [16:26] hatch: i'm working on the service inspector so i think your point about adding to the array is not right [16:26] s/right/required [16:37] bcsaller: are you around? [16:37] can you join guichat with frankban and I [16:37] yes [16:39] sinzui: i've got to run out. i'll ping you when i'm back, see if you're free then. [16:39] okay. sorry jcsackett [16:50] sinzui: ping, got a sec to chat? [16:56] sinzui: jcsackett can I get reviews on this short branch please? https://codereview.appspot.com/10691043/ [17:02] hatch: that did fix it :-/ [17:04] bcsaller: how to add that formatting function for each value in constraints? I'd like to use that to convert undefined to an empty string [17:05] frankban: did you try using constraints as the name, I'm not sure that would work but maybe I can make that happen [17:06] because you're right, we don't want to have to repeat that everywhere and we don't always know all the key names in the code anyway [17:07] rick_h: I have a charm-token-related question for you when you have a minute [17:08] bcsaller: it does not work with just 'constraints', yeah it would be a nice functionality [17:08] sinzui: Notes on our chat: https://docs.google.com/a/canonical.com/document/d/1hYCZWzTdtyv9IfZ1wdx82COrYSMwRRfeIRkVTfNMQvw/edit [17:08] frankban: ok, I'll look at hacking something up. [17:09] bcsaller: great [17:12] sinzui: i am back, are you free? [17:15] benji: hey, sorry I missed your ping [17:16] benji: what's up? [17:17] bcsaller: darn....oh well those things happen :) [17:18] jcsackett: hatch can you guys peek at this short one please? https://codereview.appspot.com/10694043 fixes a bug that annoys hatch so I can earn brownie points to disagree with something else later on :P [17:18] hatch: adding support for hierarchical bindings to deal with nested key paths. So methods on 'constraints' will be used for 'constraints.foo' if that doesn't define the method in question. [17:19] rick_h: I want to talk about svg icons and .small .medium and .large [17:19] guichat? [17:19] benji: sure thing [17:31] bcsaller: I'm not sure what you're saying [17:31] hatch: I can explain it in chat or show you in a bit in code [17:32] ok ummm i'll make a hangout [17:32] calling [17:32] not sure if I picked the right account :) [17:36] rick_h, LGTM [17:38] gary_poster: ty much [17:40] jcsackett: thanks as well. You caught me trying ot make sure each test failed/passed lol [17:41] I need to do better at that [17:52] oh man there is a lot of drama around X and MIR [17:52] hatch: welcome to canonical [17:52] why are people so resistant to change haha [17:53] ok fellas [17:53] "RAWR don't event attempt improve my linux RAWR" [17:53] so the drag and drop is pretty much awesome [17:53] (thats'a how I read these comment threads) [17:54] jcastro: benji is working on some browser tweaks and such to it now [17:54] nod [17:54] hatch. All developers believe it is fine for themselves to introduce change, but distrust anyone else proposing a change that might topple their changes [17:54] jcastro: yes that is awesome :) [17:54] jcastro: so send beer his way and hold off if something's not quite right yet [17:54] I was like "ok if this doesn't land it's no big deal" [17:54] now after using it it's like, why would I ever click the add button ever? [17:54] sinzui: lol [17:54] these developers need to work on a webapp haha [17:55] hatch, indeed. [17:57] howdy sinzui can you peek at https://codereview.appspot.com/10691043/ real quick and I wanted to see if you had time to chat on priority/next [17:57] * sinzui looks [17:57] I do have chat time. I am all talk and no action [17:57] woot [17:58] 17px? are you mad. Why choose a prime number of a dimension (other than 2) [17:58] it's what the ratio worked out to be going off the other images. 16.8px actually [17:59] .35 * charm icon size ish [17:59] 17 px is a poor input when scaling images designed for 64/128/256 [17:59] * sinzui looks for crack pipes [17:59] .35 * 48px [17:59] the large one is 57px :P [18:00] and that came from UX [18:02] rick_h: it seems that after('render') is indeed called after render() is called, but not after it is really, really rendered by the browser. [18:02] "17px? are you mad" ha! [18:02] benji: oh shoot me [18:02] however, if instead of after('render', myFunc) I use setTimeout(myFunc, 1000) it works [18:02] :/ [18:03] rick_h, LGTM. [18:04] rick_h, my daughter only sets the volume of the TV to increments of 5. She sites a study that 57% of the population are bothered by displays that accept odd numbers. I now set the TV to prime numbers only. I like to watch her twitch. === matsubara is now known as matsubara-lunch [18:07] prime numbers only....so the tv is either very loud or very quiet? haha [18:07] benji: ah ffs, it's a race since the actual rendering is done off the same event. [18:07] hatch: go smack someone on YUI around for the widget renderer crap [18:07] what widget renderer crap? [18:08] * hatch gets his pimp hand ready [18:08] http://yuilibrary.com/yui/docs/api/files/widget_js_Widget.js.html#l579 [18:08] ^^ wtf, why is the actual DOM rendering done off the event RENDER? [18:08] I guess so they can enforce the fireOnce-ness of it...ugh. [18:08] * benji hands hatch an animal print fedora. [18:09] benji: ok, hack time. [18:09] yay! I get to add a setTimeout to the code base [18:09] * benji greps [18:09] benji: no no no [18:09] darn! I'm not the first one [18:09] :P [18:09] * hatch puts my pimp hand away [18:09] actually that's how I'd write it too [18:09] :) [18:10] well not exactly [18:10] but close haha [18:10] hatch: I think what we can do is change the event to after(renderedChange) [18:10] I don't know what the problem is so I can't say what to do :) [18:10] benji: http://yuilibrary.com/yui/docs/api/files/widget_js_Widget.js.html#l583 is after the actual rendering, the ATTR for RENDERED ('rendered') is set to TRUE [18:11] benji: so we can add an event watch for that, since we KNOW it's after the REAL rendering. [18:11] benji: and then do your work [18:11] ooh, I like it [18:11] that'll be this.on('renderedChanged', myFunc), right? [18:11] benji: so to watch for ATTR changes you'll have to use the event syntax renderedChange I believe [18:11] * rick_h goes to dbl check attr event watch syntax [18:12] * benji wonders where those docs are. [18:12] use after [18:12] yep, after would be better [18:12] well, I guess it doesn't matter in this case, but generally "after" is what I want for attribute changes [18:12] benji: yea, so per http://yuilibrary.com/yui/docs/attribute/#events looks right [18:12] bac, do you have a few minutes to review these my text changes to the juju-gui charm? https://codereview.appspot.com/10698043/ [18:13] sinzui: ok, have tmie to chat now, sorry. benji is distracting me :P [18:13] :) [18:13] sinzui: y [18:13] sinzui: invite on the way [18:14] benji: 'on' will allow you to prevent the attribute change value, 'after' you can be sure that it hasn't been prevented by anything else [18:14] mind if I ask what the issue is? [18:14] ah, that's a good point [18:14] sinzui: in HACKING you are telling them to get the released version of the charm to start work, not the devel branch. [18:14] hatch: this doesn't seem to work for me: [18:14] this.addEvent(this.after('renderedChanged', Y.bind(this._addDraggability, this, container))) [18:15] ~juju-gui-charmers is release, ~juju-gui is devel, sinzui [18:15] Change [18:15] no d [18:15] I need to inspect some element dimentions after they are rendered [18:15] ah! [18:15] benji: this.after('render', function() {}) doesn't work? [18:16] that's kind of the purpose of the after render :) [18:16] event* [18:16] hatch: nope because that event is (apparently) fired after all the render() methods are called, not when the nodes have actually been rendered by the browser [18:17] bac, I think I am asking them to hack on the dev version [18:18] sinzui: line 22 of HACKING refers to lp:~juju-gui-charmers [18:18] bac: I think that is wrong then [18:18] yeah, i think so [18:18] benji: I don't think so, reading the code here [18:18] are you sure you used 'after' and not 'on' [18:19] after fires /after/ the defaultFN is run [18:19] whic is aftter the elements are in the dom [18:19] if I put a debugger call in the event handler I can't the elements on the screen [18:19] and they have no computed style [18:19] (which is what I really want) [18:20] oh, I htink that's because the renderUI isn't attaching the element to the contentBox [18:20] can you see if the content/boundingbox is in the dom? [18:21] * benji looks [18:21] sorry for the q's I just want to know where the real issue lies [18:21] so, this.after('render', function() { debugger; }); [18:22] then there should be a yui3-charmtoken element in the dom [18:22] basically (there is more in there, than just the debugger call [18:22] here is the renderUI funciton: [18:22] http://paste.ubuntu.com/5805499/ [18:22]