[00:08] <leonardr> edwingrubbs: fwiw, yes
[00:08]  * leonardr heading out
[08:55] <jtv> hi al-maisan!  Thanks.
[08:56] <jtv> al-maisan: want me to do adi's branch?
[08:56] <al-maisan> Good morning jtv 
[08:56] <al-maisan> jtv: if you like, I guess noodles775 will start reviewing at some point as well..
[08:56] <noodles775> yep, I'll be there soon.
[08:58]  * jtv backs off
[09:42] <al-maisan> hello jtv, errm, just wanted to let you know that a branch of mine (currently reviewed by allenap) does away with `IBuildFarmJobDispatchEstimation` altogether
[09:42] <al-maisan> i.e. your branch that adds an implementation of that interface is not needed any more
[09:42]  * jtv is silent as face progresses through a wild mix of moods and expressions
[09:43] <al-maisan> sorry for the extra work you have put in but ..
[09:43] <jtv> oh well :)
[09:43] <al-maisan> .. a solution that allows us to remove `IBuildFarmJobDispatchEstimation` became obvious only recently
[09:44] <jtv> That is good news though.  Did you go with the simulation as Tim suggested?
[09:45] <al-maisan> jtv: not really but we now have enough data in the BuildQueue table to estimate dispatch times w/o having to look beyond that
[09:45] <jtv> that's great... do you think you'll be needing me to provide more info for that data?
[09:46] <al-maisan> jtv: not really as long as you set BuildQueue.processor, BuildQueue.virtualized and BuildQueue.estimated_duration
[09:47] <jtv> al-maisan: I'm not sure I do though
[09:47] <al-maisan> well, the defaults for BuildQueue.processor and BuildQueue.virtualized are NULL, so that should be fine
[09:48] <al-maisan> However, the BuildQueue.estimated_duration property is required. Wasn't henninge working on that?
[09:48] <jtv> No, I think I implemented that one as a hard-coded constant.
[09:48] <jtv> I think he looked at it and found that the actual job there took essentially zero time.
[09:49] <al-maisan> jtv: as long as the constant is close enough to reality that would be fine as well :)
[09:49] <al-maisan> hmm .. so what is the constant he chose? Zero?
[09:50] <jtv> I think I picked one minute...
[09:51] <jtv> ...which is probably high.
[09:51] <al-maisan> jtv: hmm .. that's a pretty conservative guesstimate
[09:52] <al-maisan> jtv: are you storing the actual running time after the translation import job completes?
[09:52] <jtv> I was thinking we could leave it at that unless and until we find that it's really affecting the reliability of estimates.
[09:52] <al-maisan> jtv: I have no problem with that .. let's keep it for the time being.
[09:52] <jtv> Isn't storing the running time something that's done in the generic buildmaster code?
[09:52]  * al-maisan looks
[09:55] <allenap> al-maisan: What does EXTRACT(EPOCH FROM ...) do?
[09:56] <al-maisan> allenap: it converts the interval to seconds
[09:56] <allenap> al-maisan: Cool.
[09:57] <al-maisan> allenap: http://www.postgresql.org/docs/8.1/static/functions-datetime.html#FUNCTIONS-DATETIME-EXTRACT
[09:58] <allenap> al-maisan: Awesome, thank you.
[09:58] <al-maisan> you are welcome :)
[09:59] <al-maisan> jtv: you were right, the actual running time of the job is stored in BuildBase.storeBuildInfo()
[10:00] <jtv> al-maisan: common sense, design-wise, to keep that generic.  :)
[10:01] <al-maisan> jtv: yup, but common sense is not at all that common ;)
[10:02] <jtv> true, alas
[10:06] <al-maisan> adiroiban: your mp is reviewed & approved already (https://code.edge.launchpad.net/~adiroiban/launchpad/bug-340662-take-2/+merge/17598)
[10:06] <al-maisan> is that correct?
[10:06] <al-maisan> adiroiban: can I remove you from the queue?
[10:06] <adiroiban> hm
[10:06] <adiroiban> let me check
[10:07] <adiroiban> it is for bug 509252
[10:07] <mup> Bug #509252: Remove AdminPoTemplateSubset from security.py <cleanup> <Launchpad Translations:New for adiroiban> <https://launchpad.net/bugs/509252>
[10:08] <al-maisan> adiroiban: sorry .. I clicked on the wrong link
[10:09] <adiroiban> np
[10:33] <al-maisan> adiroiban: looks good, r=me
[10:35] <adiroiban> al-maisan: thanks! Do you have time to land it or should I ask someone else?
[10:35] <al-maisan> adiroiban: I believe we are in RC mode .. but I can land it once 10.02 opens up
[10:37] <adiroiban> al-maisan: sure. I was refering to land it then the queue will be open... probably it would be best to ask you when the queue is open
[10:38] <al-maisan> adiroiban: thanks!
[10:43] <allenap> al-maisan: I've finally finished your review. Lots of minor suggestions, nothing major. Looks good :)
[10:43] <al-maisan> allenap: great, thanks!
[10:59] <noodles775> adiroiban: I can take a look at that branch, or did you want someone specific to look at it?
[11:01] <adiroiban> noodles775: hi. nope. you can review it
[11:29] <noodles775> adiroiban: hi, your new test seems to pass even when I revert the change to product.py?
[11:31] <noodles775> (BTW: thanks for adding extra tests too!)
[11:32] <adiroiban> noodles775: uh... yes. I forget to test „obsolete is _not_ in extracted_text”
[11:44] <adiroiban> noodles775: I pushed the missing test
[11:45] <noodles775> adiroiban: thanks... I'm actually looking at the pagetest, and thinking that most of it is really testing the *view* functionality and should really be in a view/browser test...
[11:46] <noodles775> I mean, you're really testing the untranslatable_series property... what do you think?
[11:46] <adiroiban> noodles775: yes. I will move them in browser/tests
[11:47] <noodles775> adiroiban: So perhaps just add one series in the pagetest and verify that it appears, then yes, the rest can go in the view test.
[11:48] <adiroiban> noodles775: or 2...one active and one obsolete
[11:48] <noodles775> adiroiban: also, I (personally) think it would be worth sorting the results of untranslatable_series - random order is not good. What do you think?
[11:48] <noodles775> adiroiban: well, the fact that the obsolete series won't be included should be in the view test (as it's view code).
[11:49] <adiroiban> noodles775: ok. alphabetically sorted?
[11:50] <noodles775> adiroiban: if that make the most sense, yes (another option would be by the status, but you'll know more about which makes most sense :)
[11:51] <noodles775> adiroiban: So I'll make this as needs-fixing for the moment. Just ping me when you're done and I'll go through the code in detail. Thanks!
[11:52] <adiroiban> noodles775: ok. thanks
[14:36] <beuno> kfogel, any magic trickery available to load the sample data?
[14:36] <beuno> ro should I just clean the db and run the SQL?
[14:37] <kfogel> beuno: in standup, will answer after (inability to multitask, is all)
[14:37] <beuno> thanks
[14:40] <kfogel> beuno: most ppl in #launchpad-dev should know, though; I recall it's pretty easy to load -- abel has shown me how before
[14:40] <kfogel> beuno: of course, abel is in this standup with me :-)
[14:50] <kfogel> beuno: okay, so I *think* you just do 'cd database/schema; cp current-dev.sql MY_SAFE_current-dev.sql; cp newsampledata-dev.sql current-dev.sql'
[14:50] <kfogel> beuno: and then a 'make schema' may be necessary
[14:51] <beuno> kfogel, sounds like what I want, thanks, I'll give it a go
[14:51] <kfogel> beuno: might want to check with someone that that last step is both necessary and safe
[14:51] <beuno> kfogel, it's ok, I have a helmet on
[14:51] <kfogel> beuno: that is, there's also the "current.sql" file; I'm not sure what it is or whether 'make schema' blows it away, though I doubt 'make schema' is destructive to any file on disk (its whole point is to be destructive to the db)
[14:51] <kfogel> ok
[15:01] <beuno> kfogel, no padding between the borders and the content of the tooltips!  :)
[15:03] <adiroiban> noodles775: I'we pushed the changes. Hope you got the latests diff in your mailbox :)
[15:04] <noodles775> Thanks adiroiban, I'll take a look shortly.
[15:04] <kfogel> beuno: need padding?
[15:04] <beuno> kfogel, the content is squished against the border
[15:05] <intellectronica> beuno, kfogel: b.t.w don't forget that there's a lazr widget wanting to happen there. we didn't take this route to being with because we thought it's better to first have a concrete implementation, but that's already the second implementation in LP and i'm sure other apps would benefit from this too
[15:05] <intellectronica> i'm talking about the fancy tooltip b.t.w
[15:06] <kfogel> intellectronica: I *had* forgotten, actually, thank you for reminding me.  IOW, we want to abstract out the tooltip functionality?
[15:06] <intellectronica> yes, we want to abstract the tooltip functionality using a YUI widget, and put that in lazr-js
[15:07] <intellectronica> it will be hard to find resources to take on this task, but just recording it for now will already be a step forward
[15:07] <kfogel> intellectronica: would an XXX comment with a bug be appropriate?  (I could put the XXX comment both in lib/lp/bugs/templates/bugtarget-patches.pt and in the appropriate place for the other spot we're using a fancy tooltip right now).
[15:08] <beuno> kfogel, change div.popupTitle{padding:0 1em;} to padding:0.5em 1em
[15:08] <intellectronica> kfogel: yes, i think an XXX with a bug would be appropriate. the bug should probably be on lazr-js (rather than launchpad)
[15:08] <kfogel> beuno: I was just about to ask you how to do that, thank you :-)
[15:09] <kfogel> intellectronica: *nod*
[15:09] <kfogel> beuno: trying it now
[15:11] <kfogel> beuno: http://paste.ubuntu.com/364621/
[15:11] <kfogel> beuno: something like that?
[15:11] <kfogel> beuno: (it appears to have the desired effect in my launchpad.dev instance)
[15:11] <beuno> kfogel, yes, but it's better if you do that on the css file
[15:12] <kfogel> beuno: oh, right -- popupTitle is a shared class
[15:12] <kfogel> one sec
[15:15] <kfogel> beuno: I did this http://paste.ubuntu.com/364623/, but is there some way to do a page reload that will reload the CSS file?  I'm not seeing the new padding now.
[15:15] <beuno> kfogel, IIRC, you need to run make again (yuk!)
[15:16] <kfogel> beuno: oh, to make combo.css?
[15:16] <beuno> yeah
[15:16] <kfogel> right, forgot about that
[15:16] <kfogel> cool, doing now
[15:16] <kfogel> beuno: while that's going on, any other UI tweaks?
[15:16] <beuno> kfogel, triple tasking, so I can't say for sure
[15:17] <beuno> the drop down looks odd
[15:18] <kfogel> beuno: ...can you be more specific?  Should we get rid of the little two-headed lizard icon? :-)
[15:19] <beuno> kfogel, I haven't gotten to the suggestion part
[15:19] <beuno> it just floats there
[15:19] <beuno> I'm trying to think what to do with it while I explain to the team I'm leaving what I did this morning  :)
[15:20] <kfogel> beuno: no rush -- come back to this later if triple-casking is killing you.  Excuse me, I mean "triple-tasking" (cough, cough).
[15:21] <beuno> :)
[15:21] <beuno> sinzui, how's the TL call?  I miss everyone
[15:21] <sinzui> beuno:  join in 
[15:22] <beuno> about to!
[16:04] <beuno> kfogel, bug icons don't match the importance
[16:05] <beuno> and lastly, I'd add an "Order: " label to the ordering drop down
[16:07] <beuno> I'll add all these things to the MP
[16:07] <kfogel> beuno: oh, that's right, bug icons get colored according to importance, I totally forgot.
[16:07] <kfogel> Thank you.
[16:07] <kfogel> I'm not actually color-blind, but I might as well be -- I never notice that stuff.
[16:09] <beuno> heh, it gets better when looking at other peoples' stuff
[17:59] <leonardr> rockstar, i have a fun branch for you
[17:59] <leonardr> https://code.edge.launchpad.net/~leonardr/lazr.restful/enforce-version-order/+merge/18223
[18:00] <leonardr> ("if it was fun to write, it should be fun to understand")
[18:01] <leonardr> i'm taking my wife out to lunch, so if you run into problems and need my input, just kick the branch back to me. but it should be pretty self-explanatory
[18:01] <rockstar> leonardr, okay.  I'm also heading out to run an errand.  I'll get to it when I get back.
[18:01] <leonardr> all right
[18:17] <henninge> Hey jtv, I hope you didn't find my review too harsh ... ;)
[18:17] <henninge> jtv: Also, what are you doing up this time of night?
[18:18] <jtv> henninge: wrapping up a few things I couldn't do in my own timezone.  :-)
[18:20] <jtv> henninge: no worries about the review... I would say the reason this is (slightly) translations-specific is that we specifically check for read-only mode and make decisions based on it, in this case leading to an assertion failure.  Other requests will mostly just fail with a "transaction is read-only" database error, which is easy enough to filter out.
[20:03] <leonardr> rockstar, fwiw, i'm back
[20:03] <rockstar> leonardr, cool.  Looking at your branch now.