2012/09/21/#launchpad-dev.txt

[02:07] <wallyworld_> wgrant: i thought unclosed <br>tags were valid html5?
[02:07] <sinzui> no
[02:07] <wgrant> wallyworld_: They're valid HTML5, but not valid XHTML
[02:07] <wgrant> And we are a polyglot
[02:07] <wgrant> Because people parse us with XML parsers
[02:07] <wgrant> Because SGML is awful, and HTML5 is only mildly better
[02:07] <sinzui> we will not permit them because we cannot check our own work...and we do not know what the browser will do
[02:08] <wallyworld_> ok, i'll lp-land a quick fix, didn't see the mp comment before i landed
[02:08] <wgrant> Formally we treat our markup as an HTML5/XHTML polyglot, and I've fixed most historical issues that prevent parsing as either.
[02:09] <sinzui> I patch pocket-lint a few months back because it chocked on html5 with html4 entities
[02:09] <wallyworld_> sinzui: if you're stil lawake, i'd love a quick chat about 1052639
[02:09] <sinzui> they are the same of course, but old tools do not know that
[02:10] <sinzui> wallyworld_, okay
[02:11] <wallyworld_> sinzui: just need to reboot - mumble hates me
=== bigjools-afk is now known as bigjools
[02:38] <wallyworld_> wgrant: i just created a branch to fix those couple of <br> tags i left in and doing a search finds lots more. i'll fix those while i'm there
[02:38] <wallyworld_> well, a few anyway
[02:39] <wallyworld_> many are in comments
[02:39] <james_w> lifeless, is there a way to refer to the root context in handlebars?
[02:39] <james_w> I'm looking for a way to do conditional logic based in a partial not on the thing being rendered, but the context in which it is being rendered
[02:40] <wgrant> wallyworld_: I know there are some in structsub JS
[02:40] <wgrant> But what others?
[02:40] <wgrant> Ah, there's one on an edit form
[02:40] <james_w> and I might be any number of layers deep, and don't want to duplicate that config at all layers
[02:40] <wgrant> wallyworld_: Beware about fixing the ones in JS
[02:40] <lifeless> james_w: I don't think so. I mean, we could add / vs ../
[02:40] <wgrant> wallyworld_: Some of our UI relies on the markup being misparsed
[02:40] <wallyworld_> wgrant: yes, they are mainly js ones
[02:41] <wallyworld_> ok, will leave alone then
[02:41] <lifeless> james_w: I'm not sure its a good idea though, as the compiler is pretty unaware of how deep it is today.
[02:41] <wgrant> wallyworld_: eg. the structsub thing relied on accidental empty table cells due to invalid markup
[02:41] <wgrant> Fixing the <br> there might change the nesting
[02:41] <james_w> lifeless, yeah, it feels wrong, I'm just struggling to find a solution that doesn't multiply data
[02:42] <james_w> {{#with foo=bar}} would work
[02:42] <james_w> as would {{>partial foo=bar}}
[02:43] <lifeless> james_w: passing it down in userspace, as it were?
[02:43] <lifeless> james_w: that seems like a solid approach to me
[02:43] <james_w> lifeless, those aren't supported syntaxes currently?
[02:46] <lifeless> so, remember that ../ is template nesting, not context nesting
[02:46] <lifeless> james_w: does that help ?
[02:47] <james_w> lifeless, what do you mean?
[02:49] <lifeless> you aren't an arbitrary depth down
[02:50] <lifeless> james_w: anyhow, handlebars does support keyword parameters
[02:51] <james_w> http://paste.ubuntu.com/1217974/
[02:52] <lifeless> hmmm, No OOPS Tools Production OOPSes found for 2012-09-20 <- not sure I believe that.
[02:52] <wgrant> lifeless: Wasn't it upgraded just after midnight?
[02:54] <lifeless> could be
[02:56] <lifeless> james_w: so, I'm suggesting
[02:57] <lifeless> james_w: http://paste.ubuntu.com/1217979/
[02:57] <lifeless> I don't know if that will work OOTB
[02:57] <lifeless> but it may
[02:58] <james_w> it doesn't
[02:58] <lifeless> you tried ?
[02:58] <james_w> yeah
[02:59] <james_w> let me get the error
[03:02] <james_w> pybars/frompymeta.py in moduleFromSource
[03:02] <james_w>         code = compile(source, filename, "exec")
[03:02] <james_w> non-keyword arg after keyword arg (pymeta_grammar__render__195.py, line 8)
[03:02] <lifeless> ok
[03:02] <lifeless> so, there is a protocol for calling helpers with kw arguments
[03:03] <lifeless> it makes sense to me that a template called with extra kw arguments would shove them in the context
[03:03] <lifeless> I suggest either getting handlebars.js or making it require an option to enable or something
[03:03] <lifeless> so we don't go spuriously incompatible indefinitely.
[03:03] <lifeless> james_w: if you think its a sensible approach
[03:04]  * lifeless grabs a drink before fixing https://bugs.launchpad.net/python-timeline-django/+bug/1050852
[03:04] <_mup_> Bug #1050852: Error in action handling <python-timeline-django:New> < https://launchpad.net/bugs/1050852 >
[03:04] <james_w> lifeless, ok
[03:05] <james_w> lifeless, I won't do that now as I'm just trying things out, but if the experiment is a success then I'll still do it if I haven't found a better way to address this
[03:07] <lifeless> kk
=== almaisan-away is now known as al-maisan
=== al-maisan is now known as almaisan-away
[06:10] <wgrant> StevenK: What do you think of the restriction that non-virtual main archives can't be arch-restricted?
[06:10] <wgrant> That is:
[06:10] <wgrant>         if self.is_main and not self.require_virtualized:
[06:10] <wgrant>             return getUtility(IProcessorFamilySet).getRestricted()
[06:10] <wgrant> It seems like a pretty pointless special case
[06:10] <wgrant> So I'm going to give Ubuntu primary/partner all restricted archs and remove it, unless you can think of a reason not to
[06:11] <StevenK> main archives being purpose PRIMARY and PARTNER?
[06:11] <wgrant> And DEBUG, but yes
[06:11] <StevenK> Right
[06:11] <StevenK> Sounds like a good plan
[06:11] <StevenK> Remove PARTNER while you're at it
[06:11] <wgrant> Heh
[07:05] <wgrant> StevenK: https://code.launchpad.net/~wgrant/launchpad/bug-897999/+merge/125640
[07:09] <StevenK> wgrant: I was about to comment on your crimes against indenting when I realised it was code you were deleting.
[07:10] <StevenK> wgrant: r=me
[07:33] <adeuring> good morning
=== bigjools-afk is now known as bigjools
[08:35] <cjwatson> StevenK: totally QAable on mawson - it just got trumped by secure boot stuff yesterday
=== almaisan-away is now known as al-maisan
[08:36] <cjwatson> StevenK: I'll do it this morning
=== adeuring changed the topic of #launchpad-dev to: http://dev.launchpad.net/ | On call reviewer: adeuring | Firefighting: - | Critical bugs: ∞
[10:47] <czajkowski> jelmer: can this be closed https://answers.launchpad.net/launchpad/+question/206501
[10:57] <cjwatson> wow, that rules file is insane
[10:57] <cjwatson> I'll follow up
[11:04] <czajkowski> cjwatson: thank you
=== al-maisan is now known as almaisan-away
[12:14] <cjwatson> wgrant: Do you have any remaining concerns about https://code.launchpad.net/~cjwatson/launchpad/remove-queue-tool/+merge/114464, once process-accepted-bugs-job is deployed and the associated ops work done?  I'm just merging in current devel now
[12:15] <cjwatson> Because, -1956
[12:15] <wgrant> cjwatson: Well, we should probably live through a bit of freeze first. It's very likely that the bugs thing will be sufficient to get rid of all the timeouts, but we're pretty screwed if it doesn't.
[12:15] <cjwatson> Yeah - fortunately we have a freeze now
[12:15] <wgrant> Right :)
[12:15] <cjwatson> And if we have this deployed and enabled on say Monday, we'll have that freeze period plus the flush at the end
[12:16] <cjwatson> Which I think will be enough
[12:16] <wgrant> Indeed
[12:28] <vila> hi guys, I found a socks_proxy=socks://lpdev.local:8110/ in my env, any idea where it's coming from ?
[12:29] <vila> as in: I setup an lpdev instance long ago, but this line is nowhere to be seen in my .bashrc/.profile
[12:29] <vila> I discover it recently as it breaks a badly isolated test in another project
=== skaet_ is now known as skaet
[14:00] <wgrant> vila: That's not a standard Launchpad thing.
[14:00] <wgrant> vila: No tool I know of uses lpdev.local
[14:00] <wgrant> Or socks_proxy
[14:06] <vila> wgrant: there was a wiki page mentioning socks at some point (I can't find it anymore), the .local is from avahi and comes from me. But I don't remember ever putting that in any .rc file :-/
[14:06] <vila> (and all the significant ones are under version control anyway)
[14:07] <vila> wgrant: note that (from memory) the wiki page was referring to some FF setting, not an env var... but since I *never* used socks myself...
[14:38] <cr3> hi folks, I'd like to have a project and team deleted from launchpad. should I ask a question to launchpad (itself)?
[14:40] <czajkowski> cr3: yes plase
[14:40] <cr3> czajkowski: cheers!
[14:54] <czajkowski> cr3: need the list deleted so just waiting on that to happen
[14:55] <cr3> czajkowski: thanks, no rush
[14:56] <cr3> czajkowski: this is mostly about keeping Launchpad tidy :)
=== adeuring changed the topic of #launchpad-dev to:  http://dev.launchpad.net/ | On call reviewer: - | Firefighting: - | Critical bugs: ∞
=== matsubara__ is now known as matsubara
[18:10] <sinzui> jcsackett, how goes to demolition of the suggest ubuntu packages portlet?
[18:11] <jcsackett> sinzui: slowly, though not because of the task itself. i have had a rocky-ish day so far in terms of other things needing doing.
[18:12] <sinzui> I keep getting locked out of my computer bug the screen-lock
[18:12] <jcsackett> sinzui: that sounds like a pain.
[18:12] <sinzui> I just disabled it so I can talk without needing to kill procs or reboot
[18:13] <jcsackett> sinzui: dig.
[18:15] <sinzui> jcsackett, did you see Bug #105376 ? There is a db column that can be removed too I think
[18:15] <_mup_> Bug #105376: Kubuntu 7.04beta Alt. CD install not create a new user. <debian-installer (Ubuntu):Fix Released by brian-murray> < https://launchpad.net/bugs/105376 >
[18:15] <sinzui> jcsackett, did you see Bug #1053765 ? There is a db column that can be removed too I think
[18:15] <_mup_> Bug #1053765: Suggest ubuntu packages portlet has nothing to do <package-link> <projects> <tech-debt> <Launchpad itself:Triaged by jcsackett> < https://launchpad.net/bugs/1053765 >
[18:15] <sinzui> I missed a digit
[18:15] <jcsackett> sinzui: i didn't see that bug. thanks.
[18:16] <jcsackett> ah, i see you updated my card as well.
[18:32] <jcsackett> sinzui: do we want to obliterate the portlet, period? or just the part that's trying to encourage linking if not known?
[18:32] <sinzui> just the suggestions
[18:32] <jcsackett> ok; that's what i was assuming, but i realized it could be meant to remove all of it.
[18:32] <sinzui> we still want to show the 5 most recent packages built from the project's releases
[18:33] <jcsackett> sinzui: dig.
=== almaisan-away is now known as al-maisan
=== al-maisan is now known as almaisan-away
[22:03] <wallyworld_> sinzui: hi, with that new getSHaredProducts branch, i did not return distros because of the lazr restful issues with heterogeneous results. i was thinking about a separate api for distros
[22:03] <wallyworld_> also, there are some comments in the bug about teams vs users
[22:04] <wallyworld_> the idea is that user is removed from teams first, so we only need the api to check for direct shring for the user
[22:05] <wallyworld_> ah hang on, you are talking about teams for maintainer and driver
[22:05] <wallyworld_> i do need to fix this then
[22:05] <wallyworld_> and you also want commercial admins to get back all results, not just admins