[09:52] <gmb> Hurrah for /++form++ links. Sadly, in this case, I can't use one.
[09:52] <wgrant> Why does DistroSeries.lucilleconfig still exist? Shouldn't ComponentSelection cover that?
[10:00] <bigjools> wgrant: legacy.  it needs to die
[10:00] <wgrant> bigjools: Ah, good.
[10:01] <gmb> Argh. The bloody branch picker is horrible, horrible, horrible.
[11:02] <deryck> Good morning, all.
[11:09] <mrevell> yo deryck
[12:33] <beuno> good morning
[12:33] <noodles775> hey beuno !
[12:34] <beuno> I slept today!  :)
[12:34] <noodles775> Good to hear!
[12:39] <gmb> deryck: Damn, I can see why firebug's annoying you now.
[12:39] <wgrant> It's not just me, then?
[12:40] <deryck> gmb, yeah, it's beyond painful to me now.  I can't even use it without popping it out into a new window now.
[12:40] <gmb> wgrant: I don't think you've done all that much to annoy deryck. Keep trying, though.
[12:40] <gmb> Oh, wait.
[12:40] <deryck> heh
[12:41] <wgrant> Heh.
[12:41] <deryck> I'm actually not easily annoyed... but firebug and I are on not-so-friendly terms these days.
[12:45] <gmb> deryck: Do you know if there are any good notes about using the JS webservice client? It's frustrating me to death.
[12:46] <deryck> gmb, I don't know of any notes, sorry.  I just grep around the js files we have for prior art.
[12:46] <gmb> Rats.
[12:46] <gmb> deryck: Oh well, thanks anyway. Maybe I need food before continuing. Yes, that seems likely.
[12:46]  * gmb -> lunch
[12:51] <mrevell> I've just posted an interview with the people from Open Mind, a MIT project to build a database of common sense concepts. They're using LP. Some interesting stuff -- http://blog.launchpad.net/projects/open-mind-and-launchpad
[13:07] <deryck> oh happy freakin' day I fixed firebug for me.
[13:08] <deryck> gmb, uninstall firebug, close FF, grep your profile directory for firebug entries (mostly locatestore.rdf and prefs.js), delete those entries and then reinstall firebug.  Firebug behaves again!
[13:50] <bac> mrevell-lunch: ping me when you're back from lunch
[13:51] <beuno> mrevell-lunch, are we on for our call in 10?
[13:54] <deryck> leonardr, ping.
[13:54] <leonardr> hi deryck
[13:55] <deryck> leonardr, hi.  I have a question about the js web service client and trying to get a result back in html...
[13:55] <leonardr> ok
[13:55] <deryck> leonardr, I'm trying something like: https://pastebin.canonical.com/20439/
[13:55] <deryck> leonardr, am I on track here, or is there a better way?
[13:57] <leonardr> deryck, what is the PATCHplugin? i've never heard of it
[13:57] <leonardr> the code looks fine but i just never heard of that PATCHplugin
[13:57] <leonardr> the way you get an xhtml representation is to specify 'accept', as you did
[13:58] <deryck> leonardr, PATCHplugin is in lib/canonical/launchpad/javascript/client/client.js and comment says: * This plugin overrides the widget _saveData method to update the * underlying model object using a PATCH call.
[13:59] <leonardr> oh, ok. so it's a way of making the web service the destination of widget actions
[14:00] <deryck> leonardr, right.  So this looks like the right path I'm headed down then?
[14:00] <leonardr> deryck, yes, it looks like it
[14:00] <deryck> leonardr, excellent.  thanks for the help!
[14:01] <leonardr> np
[14:10] <gmb> deryck: Ah! That's... annoying but useful. Thanks.
[14:16] <deryck> gmb, yeah, you would think it could blow away previous defaults or somehow be backwards compatible.
[14:31] <sinzui> bac, Edwin, salgado: standup in 2 minutes
[14:34] <intellectronica> leonardr: i'm having an interesting problem using the webservice. it seems it can't mix between api.launchpad.net/beta and launchpad.net/api/beta url. if i post to and using the other in a value it complains
[14:35] <intellectronica> leonardr: i've got a workaround, but i think it would make sense to allow both, no?
[14:35] <leonardr> intellectronica, are you using an ajax client?
[14:35] <intellectronica> leonardr: indeed
[14:36] <intellectronica> oh, of course, domain restrictions!
[14:36] <leonardr> actually, it might not be because of domain restrictions, now that i think about it
[14:36] <intellectronica> still, i'm not trying to access a different domain. i just pass a uri as a value
[14:37] <leonardr> i think the parser assumes incoming urls use the same host as the host of the current request
[14:37] <leonardr> that'd be pretty hard to change; how annoying is this for you?
[14:38] <intellectronica> leonardr: it's annoying in that i have to write code especially to convert the result of canonical_url (which uses the api domain) to the format which uses the /api root. it seems wrong to have that hard coded
[14:38] <intellectronica> as i said, it works now, so i can leave it with an XXX, but i think that it would be good if in the future we handled that
[14:38] <leonardr> intellectronica: i think bjornt has fixed the underlying problem
[14:39] <intellectronica> leonardr: even in the absence of a webservice request?
[14:40] <leonardr> intellectronica: ok, let's back up
[14:40] <leonardr> where are you getting the data that you have to convert?
[14:40] <intellectronica> the problem i have is that i do the formatting in a normal web request
[14:41] <intellectronica> leonardr: while processing a normal browser request, i take an object (from a vocabulary) and i want to get the api url for it, which i will use when issuing a PATCH request from the AJAX client
[14:41] <leonardr> ok, you are writing code in launchpad
[14:41] <leonardr> i thought you were writing client code
[14:41] <intellectronica> yup
[14:42] <intellectronica> leonardr: well, i'm doing both. the client code wants an api url in the same domain as the page
[14:42] <intellectronica> but i'm preparing this value in advance when rendering the page
[14:43] <leonardr> intellectronica, what bjornt did was change canonical_url so that if you did not pass in a request object, it would give you the object's website url, not the web service url
[14:44] <intellectronica> leonardr: right, so that won't solve my problem
[14:44] <leonardr> but that won't help you. you need to get the object's web service url for the web service as accessed through the website
[14:44] <intellectronica> leonardr: what i have to do now is get the website url, parse it, inject '/api/beta' and unparse
[14:45] <intellectronica> leonardr: maybe all i need to do is extract this to a function we can use everywhere. i don't know if it makes sense to overload canonical_url with more stuff
[14:46] <leonardr> intellectronica: in theory there's some request object you could create and pass into canonical_url that would make canonical_url do what you want
[14:46] <leonardr> canonical_url already generates those urls under some circumstances
[14:46] <intellectronica> leonardr: ok, that may be a better way to implement that, but i still don't think i should have to do that every time i need to generate a url
[14:47] <intellectronica> leonardr: also, it seems a bit wrong to create a request object just for that purpose
[14:47] <leonardr> intellectronica, we create request objects all the time for similar purposes
[14:47] <leonardr> changing canonical_url is a huge task
[14:48] <intellectronica> fair enough. do you maybe remember a place where we do that?
[14:48] <intellectronica> leonardr: i agree. i think i'll just create a new helper function, webservice_canonical_url() or something like that
[14:48] <leonardr> intellectronica: check out bjornt's code. you can do the same thing in your helper function, but also check canonical_url to see what the request has to look like for that kind of url to be generated
[14:49] <intellectronica> leonardr: cool, thanks
[14:51] <BjornT> intellectronica: it might work if you simply do: canonical_url(ob, request=IWebServiceClientRequest(request))
[14:51] <BjornT> intellectronica: but hard to say without looking at your code
[14:54] <intellectronica> BjornT: yes, that seems to work just fine. thanks!
[15:11] <gmb> Gah. Did I mention that I hate Javascript?
[15:11] <mars> :)
[15:17] <maxb> Does anyone like Javascript? :-)
[15:31] <kfogel> mrevell: ready to skype?
[15:32] <mrevell> Hi kfogel, almost, just making a cup of tea
[15:33] <kfogel> mrevell: no rush, I'm on skype now
[15:34] <kfogel> mrevell: meaning, available when you're ready, but take your time
[15:35] <gmb> Does anyone have any idea under what circumstances a named_post would just sort of vanish into the ether when called from Javascript? Neither success nor failure callbacks are called.
[15:36] <gmb> Ah, wait.
[15:36] <gmb> If I actually read the output from the server, that would help.
[15:53] <rockstar> morning
[15:54] <noodles775_> beuno: It'd be great to be able to find out exactly what the purpose and use-cases of a page is on LP - documenting that info in some permanent location.
[15:54] <noodles775_> beuno: I think mars mentioned something along those lines the other week...
[15:54] <beuno> noodles775_, yes
[15:54] <noodles775_> beuno: but I've created https://dev.launchpad.net/SoyuzBuildersIndexPage
[15:55] <noodles775_> as an example, and will fill it out as I get info, but is there anything you'd add (I'm about to create a second one for another page)
[15:56] <beuno> noodles775_, on calls, but will look at it afer them
[15:56] <noodles775_> beuno: yep, no rush. Thanks.
[16:36]  * rockstar is glad to have 4 cores, since make schema is chewing on an entire core
[17:50] <sinzui> beuno: I would like your opinion of the screencaps for bug 399010.
[17:50] <ubot3> Malone bug 399010 in launchpad-registry "Update simple edit pages to base-layout" [High,In progress] https://launchpad.net/bugs/399010
[17:50] <mup> Bug #399010: Update simple edit pages to base-layout <story-ui-3> <Launchpad Registry:In Progress by sinzui> <https://launchpad.net/bugs/399010>
[17:50] <mup> Bug #399010: Update simple edit pages to base-layout <story-ui-3> <Launchpad Registry:In Progress by sinzui> <https://launchpad.net/bugs/399010>
[17:50] <mup> Bug #399010: Update simple edit pages to base-layout <story-ui-3> <Launchpad Registry:In Progress by sinzui> <https://launchpad.net/bugs/399010>
[17:50]  * sinzui now hacks on project page
[17:51] <rockstar> Looks like we need to either kill ubot3 or mup...
[17:53] <gmb> rockstar: Better still, make them fight until one dies.
[17:53] <sinzui> two bots enter, one bot leaves
[17:54] <sinzui> I suspect mup is cheating and may in fact be 3 bots
[18:07]  * gmb EODs
[18:12] <beuno> sinzui, finishing lunch, and will look
[18:24] <abentley> rockstar: chat?
[18:25] <rockstar> abentley, sure.  One sec.
[18:32] <abentley> rockstar: https://launchpad.canonical.com/DeveloperDocumentation/TemplateCodeReuse
[18:48] <rockstar> abentley, actually, the more I think about it, the more it would seem to follow current conventions if it was a link to a single submit button form value.
[18:48] <rockstar> abentley, although I am not qualified to make UI decisions.
[18:48]  * rockstar hates the look of HTML buttons
[18:50] <abentley> rockstar: I think following conventions is good, up to a point.  I think that having people go to another page just to click a submit button is past that point.
[18:50] <rockstar> Yes, this is true.  I'm pretty torn.
[18:51] <abentley> You could almost justify it if you were asking for confirmation, because claiming a review can't be undone.
[18:51] <beuno> sinzui,
[18:52] <beuno> are you up for a massive challenge?
[18:52] <rockstar> abentley, ah, yes, that would actually make sense.
[18:52] <sinzui> Dude. I just discovered that the new heading and the title edit-widget are incompatible. I am already in a challenge
[18:53] <beuno> sinzui, heh, ok
[18:53] <beuno> I'll try and tackle it then
[18:53] <beuno> we need to fix our freakin forms
[18:53] <sinzui> beuno: but I am available
[18:53] <beuno> this right-aligned bullshit has to stop
[18:53] <rockstar> beuno, AMEN!
[18:54] <beuno> I've tried to train my eyes to ignore it, but I can't
[18:54] <sinzui> beuno. Yeah. There are some near infinite length fields too
[18:54] <beuno> sinzui, I'll address
[18:54] <beuno> I'll do it NOW
[18:55] <sinzui> beuno: a subtle change to launchpad-form-macros will change everything, as too might the css rule
[19:03] <beuno> sinzui, more from my ourage soon
[19:03] <beuno> I was looking through your screencaps
[19:03] <beuno> but that issue lept coming up
[19:04] <beuno> sinzui, the template changes look
[19:04] <beuno> fine
[19:04] <beuno> there are a lot of peripheral issues to solve
[19:04] <sinzui> beuno: I understand. That is why the screencaps are repetative...so it is obvious there is something else we need to do to land my branch
[19:04] <beuno> :)
[19:05] <beuno> sinzui, there's something else
[19:05] <beuno> about a change we made to edit/add/remove icons
[19:05] <beuno> the rule should be: if you can edit/remove an object, the icon should be on it's right
[19:05] <beuno> if it's an action, on the left
[19:05] <beuno> is there any way we can change that?
[19:06] <beuno> I don't like the result of having the add/remove on the right of action links
[19:07] <sinzui> beuno: How do we know the link is an action?
[19:07] <beuno> sinzui, my guess is that the only way is to specify it?
[19:07] <beuno> by default it's an action, but a parameter modifies that?
[19:07] <sinzui> Well in th screencaps, what is wrong with the links?
[19:07] <sinzui> which ones are wrong
[19:08] <beuno> Related actions
[19:08] <sinzui> beuno: you are editing the project in all cases
[19:08] <beuno> sinzui, right, but they are action links
[19:08] <beuno> what I mean by object, is the actual's object name
[19:08] <beuno> Project name (i)
[19:09] <beuno> (I) Edit something something
[19:09] <beuno> in the first case, the important information is the object's name, and secondary is the fact you can edit it
[19:09] <beuno> and in the second case, the important information is that you can edit something
[19:10] <sinzui> beuno: If the icon must follow something else in the page, then we use fmt:icon, and we make fmt:icon always render the with the icon on the left
[19:11] <beuno> sinzui, and for editable objects?
[19:11] <sinzui> So the rule is: developers may only use fmt:icon  when they are placing a link to edit the "thing" that just rendered
[19:12] <sinzui> is editable objects the same as "thing" that just rendered?
[19:12] <beuno> yes, but that means it should be on the right
[19:12] <beuno> I think add will *always* be on the left, for example
[19:13] <beuno> because it has to tell you what it adds
[19:13] <sinzui> beuno: I think the rule then is the link text can never be rendered before the icon. The icon may be rendered after the object's text/representation
[19:13] <beuno> but edit and remove placed next to something is clear enough
[19:13] <beuno> yes to that second statement
[19:14] <beuno> rockstar, have you heard about BjornT's super inline bug commenting branch?
[19:14] <sinzui> beuno: in https://bugs.edge.launchpad.net/launchpad-registry/+bug/399010, I see "Unsubscribe (-)".
[19:14] <mup> Bug #399010: Update simple edit pages to base-layout <story-ui-3> <Launchpad Registry:In Progress by sinzui> <https://launchpad.net/bugs/399010>
[19:14] <ubot3> Malone bug 399010 in launchpad-registry "Update simple edit pages to base-layout" [High,In progress]
[19:14] <mup> Bug #399010: Update simple edit pages to base-layout <story-ui-3> <Launchpad Registry:In Progress by sinzui> <https://launchpad.net/bugs/399010>
[19:14] <beuno> rockstar, I think you should steal it for reviews as soon as it lands
[19:14] <beuno> sinzui, right. I think that looks wrong.
[19:15] <beuno> it should be "(-) Unsubscribe"
[19:15] <sinzui> beuno: I think by your rule (fuck you mup) the link should be...okay we agree
[19:15] <beuno> :)
[19:15] <beuno> Martin Albisetti (-)
[19:15] <sinzui> beuno: I think I can fix this in a few deft lines of code.
[19:15] <beuno> unsubscribes me
[19:15] <beuno> sinzui, you will elevate your status from personal heroe to personal god
[19:16] <sinzui> If only I had physique of a code instead of an deskbound hacker
[19:16] <sinzui> wow, how did "code" get into that sentence
[19:17] <sinzui> oh well.
[19:17]  * beuno wonders what he meant
[19:17] <sinzui> s/code/god/
[19:17] <beuno> :)
[19:17] <beuno> BIG typo there
[19:19] <EdwinGrubbs> sinzui: the SourcePackageNameVocabulary is in registry/model/sourcepackagename.py instead of registry/vocabularies.py. Is this part of the reorg or just wrong?
[19:20] <sinzui> I think it was wrong and was not automatically fixed by the migration script
[19:21] <rockstar> beuno, sounds good.  I have a backlog of branches that need to land first, but I'll keep it in mind.
[19:24] <beuno> bigjools, mockup is looking better
[19:24] <beuno> although seeing it with data it makes me realize some of my thoughts where bogus
[19:26] <beuno> sinzui, I went through an intensive review of navigation with the UX team and Mark today
[19:26] <beuno> and we're looking good
[19:26] <beuno> I should get a final-ish mockup back tomorrow
[19:27] <sinzui> rock
[19:27] <beuno> which is *very* similar to the last one you saw
[19:35] <rockstar> Would someone like to do a JS review for me?
[19:37] <sinzui> beuno: This is the fix for the modify icons: https://pastebin.canonical.com/20476/ I replaces "sprite-after" with "sprint modify". We do not have a .modify class. If you need to act just on those links, you now have a hook
[19:38] <sinzui> s/"sprint modify"/"sprite modify"/
[19:38] <beuno> sinzui, rock
[19:38] <beuno> thanks
[19:38] <beuno> deryck may go crazy
[19:38] <sinzui> This will land when pqm opens
[19:38] <beuno> thanks sinzui
[19:39]  * deryck looks at scrollback, smelling icon changes again
[19:42] <deryck> beuno, why will I go crazy?  Because of changing the icon positioning again? :)
[19:43] <sinzui> deryck: only if you are not using the approved way of making links. fmt:link JDRT
[19:43]  * beuno hides
[19:43] <deryck> sinzui, fmt:link doesn't exist in js. :)
[19:43] <sinzui> deryck: fmt:link-icon and fmt:icon-link will also work because They are just aliases for fmt:link
[19:44] <sinzui> deryck: if you used sprite-after, you will need to change it to sprite modify.
[19:44] <deryck> sinzui, yeah, I don't think it's going to be a pain.  Was just going to give beuno a hard time about it. :)
[19:45]  * beuno unhides!
[19:45] <beuno> aha!
[19:45] <beuno> :)
[19:45] <beuno> I promise, I'm standarizing on this rule
[19:46] <beuno> sinzui, how crazy do you think I am if I drop the use of tables from forms?
[19:46] <rockstar> deryck, I think you might like the branch I just sent you.  I'm working on devising a way of pulling markup out of js and back to tal again.
[19:46] <beuno> I will have to spend a year of my life fixing tests, right?
[19:46] <deryck> rockstar, oh cool
[19:46] <sinzui> beuno: you will cry. the problem is larger than sprites
[19:46] <deryck> rockstar, did you ever branch subscriptions?
[19:46] <beuno> crap
[19:46] <deryck> ever do, that is.
[19:46] <sinzui> beuno: there are a lot of widgets that render as tr and td combos because they know all forms are tables
[19:47] <rockstar> deryck, I haven't finished it, but I will need to make LOTS of changes to it.  It's going to be more complicated than bug subscriptions, because our subscriptions have state.
[19:47] <deryck> rockstar, right.
[19:47] <beuno> ok, well, I'm going to ask to get this for my birthday then
[19:47] <sinzui> beuno: they also assume the table is 2 column. I have seen BAD thing happen when they do not mix well
[19:48] <deryck> rockstar, I do have nicer Subscriber and Subscription objects in js now, which you may be better able to re-use.  If you haven't seen those yet.
[19:48] <rockstar> deryck, I saw that.
[19:48] <deryck> doesn't help the state difference, but handling users is slightly better.
[19:49] <beuno> sinzui, I could cheat and make everything colspan=2  :)
[19:49] <beuno> especially now that flacoste is on vacations and can't see me
[19:49] <sinzui> beuno: hey! that is my cheat. I used it last release
[19:49] <beuno> fantastic, I know have someone to blame
[19:50] <sinzui> beuno: use the change password view to verify our hacks work
[19:52]  * beuno stabs stuff
[20:05] <beuno> sinzui, the password change form seems to do something funky
[20:05] <sinzui> beuno: it does!
[20:06] <beuno> why?!
[20:06] <sinzui> beuno: solve that form and most of the widget weirdness will also be solved
[20:06] <beuno> and where do I stab it?
[20:06] <beuno> I mean
[20:06] <beuno> I changed the forms macros
[20:06] <beuno> and the first field is fixed
[20:06] <beuno> but the next two remain the same
[20:06] <beuno> they have th's out of nowhere!
[20:07] <sinzui> beuno: The password widget is inserting it's own markup. many of our widgets do
[20:07]  * beuno bangs head against table
[20:07] <beuno> sinzui, so I should dive into python?
[20:07] <beuno> why does it *just* do that for 2 fields?  when all 3 of them are exactly the same?
[20:08] <sinzui> beuno: EdwinGrubbs discovered that the widgets in that form never align. We decided that setting the width of one column will give the illusion on good design
[20:09] <sinzui> beuno: I think you need to look at the HTML templates used by each widget.
[20:09] <beuno> *sigh*
[20:10]  * beuno removes infinite length for input boxes
[20:10]  * sinzui high five's beuno
[20:11] <beuno> sinzui, I will want you to review this branch
[20:11] <sinzui> beuno: I think these are the problematic controls: https://pastebin.canonical.com/20478/
[20:11] <sinzui> sure
[20:11] <beuno> I want to make non-optional labels bold
[20:11] <beuno> thoughts?
[20:12] <beuno> oh, THANK YOU for that list
[20:12] <sinzui> beuno: I support every effort to fix the madness. I looked at the issue last year and walked away
[20:47] <beuno> does anyone remember the large ec2 instance's name?
[20:47] <beuno> -i x1.large?
[20:48] <beuno> gary_poster, ^
[20:48] <beuno> (and, can we make that the default?)
[20:48] <gary_poster> beuno do ec2test --help
[20:48] <beuno> ah  :)
[20:48] <beuno> I did, but didn't read
[20:48] <gary_poster> :-)
[20:48] <beuno> thanks gary_poster
[20:48] <gary_poster> np
[20:48] <beuno> any reason why that's not the default?
[20:49] <beuno> oh, it says it does now
[21:02] <maxb> gary_poster: Hi - Your name has been mentioned in connection with Launchpad + Python 2.5 a bit - are there any bugs / branches that I, being interested in that, might want to keep an eye on?
[21:04] <gary_poster> maxb: hi.  When I get back to this branch and get it landed, we'll have the infrastructure for others to start helping (by moving to newer versions of dependencies, for instance): https://edge.launchpad.net/~gary/launchpad/zbuildout
[21:04] <maxb> Can you un-private that branch, if you're at liberty to do so?
[21:07] <gary_poster> maxb: oh, sure, sorry.  haven't done that.  beuno (since you tagged me ;-) ), I either ask a losa to do that, or subscribe the launchpad public team, or just push another new branch up, right?
[21:09] <maxb> There is no edit icon for you next to "This branch is private" on https://edge.launchpad.net/~gary/launchpad/zbuildout ?
[21:09] <beuno> yes
[21:09] <maxb> (I'm randomly speculating that it should work the same way as private bugs, since I've never seen a private branch)
[21:09] <beuno> subscribe people
[21:10] <gary_poster> maxb: no
[21:10] <gary_poster> beuno ack thanks
[21:11] <gary_poster> maxb: kinda busy, I'll just add you for now.  what's your account?
[21:11] <maxb> Just maxb, but no rush, it can wait until you can do it properly
[21:11] <maxb> sidnei already pushed a public copy a couple of days ago
[21:12] <gary_poster> maxb: ok cool.  I plan to do it before EOD.  thanks for understanding.  ping me anytime you like
[21:15]  * gmb wishes for a pqm-submit --do-what-i-mean option.
[21:16] <beuno> fsck
[21:16] <beuno> I forgot to use --headless
[21:17] <beuno> gary_poster, do I have any other options than killing it?
[21:17] <gary_poster> beuno: no, sorry
[21:17] <beuno> no problem, stupidity has a price
[21:18] <beuno> in this case, probably about 2usd
[21:42] <gmb> \0/ wgrant becomes the first community contributor to have code landed on devel. Hurrah, etc.
[21:42] <gmb> (at least I think he does)
[21:50] <kfogel> maxb: hey, I noticed you edited the dev.launchpad.net wiki (thanks).  Was it open when you first tried, or did you have to get someone to remove the ACL?
[22:30] <thumper> morning
[22:30] <beuno> hey thumper
[22:30] <mars> morning thumper
[22:30] <beuno> I want to talk to you when you're caffeinated  :)
[22:30] <thumper> I see that pqm is open again
[22:30]  * thumper submits some branches
[22:30]  * mars was just being social
[22:33] <thumper> beuno: what was your bug number for your branch listing oopses?
[22:33] <thumper> hi mars
[22:34] <thumper> beuno: nm
[22:36] <beuno> kai
[22:36] <thumper> beuno: LP looks very sucky in FF 2 on Windows :(
[22:36] <beuno> does it?
[22:36] <beuno> I'll have to take a look and see why
[22:36] <thumper> yep
[22:37] <thumper> sprits :(
[22:37] <thumper> sprites
[22:37] <beuno> really?  they should work fine on FF2, there's no black magic in them
[22:38] <thumper> beuno: we can talk now if you like
[22:38] <thumper> beuno: I have 20 minutes before the standup
[22:38] <beuno> thumper, sounds good
[22:38]  * beuno headphoneses
[22:39] <beuno> thumper, ready
[22:40] <thumper> beuno: my skype is having trouble finding you
[22:40] <thumper> beuno: try to call me
[22:44] <maxb> kfogel: I waited until I noticed an Edit link. I understand it's generally open now.
[22:45] <maxb> I see the front page is locked - presumably to discourage stupidity
[22:48]  * ajmitch wonders why rocketfuel-get is failing now as it tries to get a bzr lock
[22:49] <ajmitch> more specifically, it's only ~/launchpad/lp-sourcedeps//download-cache that's failing to update, saying that it's a readonly transport
[22:51] <wgrant> gmb: Thanks!
[22:54] <EdwinGrubbs> sinzui: we kinda forgot about the weekly meeting.
[22:54] <sinzui> EdwinGrubbs: yeah. We can talk sometime tomorrow about the pickker work
[22:58] <ajmitch> wgrant: as a quick check, are you able to update ~/launchpad/lp-sourcedeps/download-cache ? It doesn't want to update with bzr+ssh for me
[22:59] <wgrant> ajmitch: WFM
[22:59] <ajmitch> bzr+ssh for you?
[23:00] <wgrant> Yep.
[23:00] <ajmitch> interesting, I just get:
[23:00] <ajmitch> bzr: ERROR: Cannot lock LockDir(lp-45197200:///~launchpad/lp-source-dependencies/trunk/.bzr/branchlock): Transport operation not possible: readonly transport
[23:01] <wgrant> What are you trying to do?
[23:01] <ajmitch> Run rocketfuel-get & not have it fail
[23:02] <wgrant> What bzr action gives that error?
[23:03] <ajmitch> bzr up "$LP_DOWNLOAD_CACHE_PATH"
[23:03] <ajmitch> which is ~/launchpad/lp-sourcedeps/download-cache
[23:04] <ajmitch> nothing out of the ordinary there
[23:04] <wgrant> Very odd.
[23:04] <ajmitch> The only thing I can see is that it's a checkout, not a branch
[23:14] <beuno> thumper, I'm filing the "bzr push lp:project" bug anyway, since I can't find it
[23:14] <beuno> unless you can  :)
[23:14] <thumper> beuno: ok
[23:15] <thumper> beuno: we could always dupe it anyway
[23:15] <beuno> yes, free kama
[23:15] <beuno> karma as well
[23:16] <beuno> aha
[23:16] <beuno> bug 181401
[23:16] <ubot3> Malone bug 181401 in launchpad-code "should be able to push to lp:project or lp:project/series to make a new branch" [Medium,Triaged] https://launchpad.net/bugs/181401
[23:16] <mup> Bug #181401: should be able to push to lp:project or lp:project/series to make a new branch <Launchpad Bazaar Integration:Triaged> <https://launchpad.net/bugs/181401>
[23:16] <mup> Bug #181401: should be able to push to lp:project or lp:project/series to make a new branch <Launchpad Bazaar Integration:Triaged> <https://launchpad.net/bugs/181401>
[23:16] <mup> Bug #181401: should be able to push to lp:project or lp:project/series to make a new branch <Launchpad Bazaar Integration:Triaged> <https://launchpad.net/bugs/181401>
[23:17] <beuno> thank you mup
[23:17] <rockstar> Who is in charge of ubot3?  Can we get it disabled?
[23:17] <beuno> done  :)
[23:17] <rockstar> Thank you beuno
[23:18]  * wgrant had better get going on another branch before bigjools clears the queue tonight.
[23:18] <maxb> mup would seem to have a bug too, since it's replying three times when only a person and ubot3 said anything
[23:18] <wgrant> maxb: The third was probably from the URL in ubot3's reply.
[23:19] <maxb> hmm
[23:19] <wgrant> It's reasonable to look for multiple bugs, so I don't think that's a bug.
[23:19] <beuno> it's mentioned 3 times
[23:19] <beuno> once by me
[23:19] <beuno> and twice by ubot3
[23:21] <wgrant> At least ubottu and derivatives have a limit on how often they'll detect a mention of a bug.
[23:25] <wgrant> I can't see how badly I killed buildbot :(
[23:27] <ajmitch> Do you suspect that you may have murdered it?
[23:27] <wgrant> It's the first community contribution, and it's by me, so something has to go wrong.
[23:28] <thumper> wgrant: what do you think has happened?
[23:28] <thumper> wgrant: I don't see anything wrong
[23:29] <wgrant> thumper: Ah, good.
[23:29] <beuno> thumper, any idea how to conditionally display a specific CSS class in TAL?
[23:30] <beuno> as in, always show the tag
[23:30] <thumper> beuno: yes
[23:30] <beuno> but on a condition, display a class
[23:30] <mwhudson> is there a reason the buildbot waterfall is still private btw?
[23:30]  * beuno waits to be enlightened
[23:30] <mars> beuno, eeew, not fun to do in TAL
[23:30] <thumper> mwhudson: not any more
[23:31] <thumper> beuno: well...
[23:31] <thumper> beuno: it depends
[23:31] <wgrant> Particularly in Python 2.4...
[23:31] <beuno> I know it's something like  $condition|something/something
[23:31] <thumper> beuno: you want to use tal:attributes="class ..."
[23:31] <beuno> yes, I got that far
[23:31] <thumper> beuno: got a concrete example for me?
[23:31] <beuno> but blah|blah only returns True/False
[23:32] <beuno> yes
[23:32] <beuno> one sec
[23:32] <mwhudson> beuno: put it on the view
[23:32] <mwhudson> i think returning None will mean not including the attribute at all
[23:32] <beuno>           <label tal:attributes="for widget/name;
[23:32] <beuno>                                  class show_optional|I-want-a-class-here"
[23:32] <beuno>                  tal:content="string:${widget/label}:">Label</label>
[23:33] <beuno> the thing is
[23:33] <mwhudson> or a python: expression, but ewww
[23:33] <beuno> this is the forms macros
[23:33] <beuno> so I could put it in it's view
[23:33] <beuno> but I ahve this right above it:
[23:33] <beuno>         <tal:block tal:condition="display_label|widget/display_label|python:True">
[23:34] <thumper> as mwhudson said, a python tal expression is one way
[23:34] <thumper> or put the classname onto the view itself
[23:35] <mwhudson> beuno: how about you tell us what you're actually trying to do?
[23:35] <mwhudson> (a line that never gets old!)
[23:35] <beuno> ok
[23:35] <beuno> I want for labels on form elements to be bold if they're mandatory
[23:36] <beuno> I somehow feel explaining that won't help
[23:36] <beuno> but there you have it   :)
[23:36] <thumper> tal:define="classname python: "" and show_optional or "someclass""
[23:36] <thumper> no
[23:36] <thumper> actually that is exactly wrong
[23:36] <thumper> :(
[23:36] <wgrant> Sounds like you need Python 2.5.
[23:37] <beuno> I know I've done this before
[23:37] <thumper> switch "someclass" and ""
[23:37] <mwhudson> tal:attributes="class python:[None, 'show_optional'][bool(show_optional)]"
[23:37] <mwhudson> it's horrific, but it'll work
[23:38] <thumper> ick
[23:38] <thumper> that'd need a comment for sure
[23:38] <beuno> UI is fun
[23:38] <mars> tbh, I'd bounce it
[23:38] <mars> put that in the view
[23:38] <mwhudson> beuno: nicer would be widget/fmt:label-css
[23:38] <mwhudson> and some magic in tales.py
[23:38] <beuno> yes, nicer is more expensive than I can afford
[23:39] <thumper> why?
[23:39] <beuno> so I think I'm just not doing it
[23:39] <mwhudson> how is that expensive?
[23:39] <thumper> beuno: get wgrant to write it :)
[23:39] <beuno> I need to start poking in more places, test it, etc
[23:39]  * wgrant fortunately has lectures today.
[23:40] <mwhudson> beuno: you're not very receptive (and this is a good thing!) to "too much work" when you do ui reviews :)
[23:40] <beuno> mwhudson, heh
[23:40] <beuno> which is why I try and do as little code as possible in LP  :)
[23:40] <beuno> I'm a terrible judge for myself
[23:43] <beuno> but somebody has to make forms stop sucking
[23:45] <thumper> beuno: so just DTRT and do it properly :)
[23:46] <mwhudson> beuno: i'm happy to help with coding suggestions, etc
[23:46] <mwhudson> but this is not the place for hacks
[23:47] <beuno> I'm going to wait and see if changing this in forms breaks dozens of tests or not
[23:48] <beuno> and decide based on that if I don't implement that UI feature at all
[23:48] <beuno> I won't hack it in, it's not worth it
[23:48] <beuno> so either DTRT or not doing that change
[23:57] <thumper> mwhudson: how do I ec2test jml's branches without grabbing and pushing them myself?
[23:58] <thumper> mwhudson: is the a convenient param?
[23:58] <mwhudson> thumper: ec2test.py lp:~whatever
[23:58] <thumper> ah fantastic
[23:58] <thumper> mwhudson: I'll test submit jml's branches
[23:58] <mwhudson> thumper: you'll need to supply a submit address somehow if you want to use -s
[23:59] <mwhudson> and submit branch
[23:59] <thumper> mwhudson: ok