[01:48] <dilys> Merge to devel/launchpad/: [trivial]  Update analyze-error-reports: display more/all exceptions, and cosmetic makeover on titles and headings (r1841: kiko)
[01:52] <kiko-afk> y a y
[01:57] <sfllaw> kiko-afk: How do you type when you're AFK?
[01:58] <mdke> sfllaw: you should see him sleep-type
[02:00] <mdke> jordi: not much chance of you being around I guess, but ping?
[02:02] <kiko-afk> naked waitresses type for me
[02:02] <lifeless> gee
[02:02] <lifeless> you need more imagination
[02:03] <mdke> kiko-afk: are you available for some "argh wtf has happened to all these translations" help?
[04:15] <bddebian> cprov: ping?
[04:16] <cprov> bddebian: hi
[04:17] <bddebian> Hello.  Sorry to bother you but I was told you were the person to ask about @ubuntu.com e-mail address not working?
[04:18] <cprov> bddebian: no problem, unfortunatelly I'm not in touch with the @ubuntu redirection, but tell me what is wrong ?  have you set your preferred email properly ?
[04:19] <bddebian> cprov: As far as I know, though I guess it wouldn't hurt to check again
[04:20] <bddebian> Hmm, looks correct
[04:22] <cprov> bddebian: I can't do anything right now, but I think you can file a bug on Soyuz, I will redirect it to elmo asap
[04:22] <bddebian> cprov: Oh, I can bug elmo, thx
[04:23] <cprov> bddebian: right, better. He's the person behind redirect.
[04:23] <bddebian> OK, thanks for your time
[04:26] <cprov> bddebian: anytime
[04:31] <mpt> Goooooooooooooooooooood afternoon Launchpadders!
[04:33] <bddebian> Hmmm
[04:33] <bddebian> Hello mpt
[04:39] <poningru> hi question: I wanted to request/work on a feature for the edgy espresso partitioning tool, a dual boot 'option' in the partitioning tool, should I put that in as feature request in malone or should I add it as a braindump on the wiki?
[04:39] <poningru> or should I ask this in another channel? like ubuntu-dev?
[04:41] <cprov> poningru, ubundu-dev, it seems to require a specification (use blueprint), coordination with developers and then malone bugs  
[04:42] <cprov> poningru: anyway, it's just a hint, ask Colin for espresso roadmap
[04:42] <poningru> oh true
[05:09] <poningru> I was looking through the specs: https://launchpad.net/distros/ubuntu/+specs and there seem to be 3 migration specs
[05:10] <poningru> two of them have the same title
[05:10] <poningru> just do a search for 'migrat' on the page
[05:59] <mpt> poningru, the specification already includes a dual-booting option
[05:59] <mpt> the existing specification, I mean
[06:01] <mpt> poningru, https://wiki.ubuntu.com/UbuntuExpress/PartitioningTool
[06:01] <poningru> mpt: hehe yeah burgundiva pointed that out to ignorant ole me
[06:02] <poningru> err burgundavia
[09:44] <sivang> re all
[09:44] <ivoks> hi
[09:45] <sivang> hey ivoks , sup?
[09:45] <ivoks> still waking up :)
[09:53] <ivoks> hm... why are there three lines when english has singluar and plural (two lines) in rosetta? :)
[10:08] <jamesh> ivoks: have the URL?
[10:09] <ivoks> second...
[10:09] <ivoks> https://launchpad.net/distros/ubuntu/dapper/+source/rhythmbox/+pots/rhythmbox/hr/+translate?offset=10&show=untranslated
[10:10] <ivoks> i saw that on many locations...
[10:10] <ivoks> i also found one funny word in rhythmbox; undownloaded :)
[10:17] <SteveA> ivoks: that page is to translate into croatian
[10:18] <SteveA> croatian has three plural forms
[10:18] <ivoks> SteveA: right
[10:18] <ivoks> we do?
[10:18] <ivoks> hm...
[10:18] <SteveA> the expression is in a box on the left
[10:18] <SteveA> n%10==1 && n%100!=11 ? 0 : n%10>=2 && n%10<=4 && (n%100<10 || n%100>=20) ? 1 : 2
[10:18] <ivoks> ah...
[10:19] <ivoks> thanks
[10:19] <SteveA> no problem
[10:33] <ivoks> SteveA: if i understant that right, this means that if n=1 and n=11, singular will be used?
[10:33] <ivoks> that is; first line?
[10:37] <SteveA> if n is 11, i think form 2 will be used
[10:37] <SteveA> >>> def whatform(n):
[10:37] <SteveA> ...   if n%10==1 and n%100!=11:
[10:37] <SteveA> ...     return 0
[10:37] <SteveA> ...   if  n%10>=2 and n%10<=4 and (n%100<10 or n%100>=20):
[10:37] <SteveA> ...     return 1
[10:37] <SteveA> ...   return 2
[10:38] <SteveA> 
[10:38] <SteveA> >>> for number in range(150):
[10:38] <SteveA> ...   print number, whatform(number)
[10:38] <SteveA> 
[10:38] <SteveA> try that in python
[10:40] <ivoks> right... that's ok :)
[10:40] <ivoks> thank you very much
[10:41] <ivoks> i missed that ! in n%100! :)
[10:42] <SteveA> rosetta should have a feature to show you the plural form expression expanded like that
[10:43] <SteveA> but, if you speak croatian
[10:43] <jamesh> alternatively we could provide a label for each plural form for a language
[10:43] <SteveA> you should be able to just think how it should sound
[10:43] <jamesh> if there are standard names for each form
[10:43] <ivoks> SteveA: i do
[10:43] <SteveA> like in lithuanian, if i'm thinking of the coins...
[10:43] <ivoks> but the thing is that in 99% cases 1 and 2 are same
[10:43] <SteveA> i have 0 lit
[10:43] <SteveA> i have 1 litas
[10:44] <SteveA> i have 2 litai
[10:44] <SteveA> i have 9 litai
[10:44] <SteveA> i have 10 lit
[10:44] <ivoks> i understand
[10:44] <SteveA> ... 11 lit ... 20 lit ... 21 litai ... 22 lit ... 100 lit ... 101 litas
[10:45] <ivoks> that's same for croatian
[10:45] <ivoks> but in this case, when i have All %d artists, 1 and 2 are same
[10:46] <SteveA> because it is genitive ? 
[10:46] <ivoks> yeah
[10:46] <SteveA> i see
[10:46] <ivoks> thanks for your help
[10:46] <jamesh> then you might have to repeat the translation for 2 plural forms
[10:47] <SteveA> looking at the translatable text in english, saying "All 1 file" is odd
[10:47] <ivoks> i know, that's not the problem
[10:47] <ivoks> most important thing is that i now understand how it works :)
[10:47] <ivoks> SteveA: right, i removed All from translation
[10:47] <ivoks> for singular
[11:53] <SteveA> mpt: hello
[11:53] <SteveA> mpt: thanks for the menus data info.  a question: what do the /!\ symbols mean?
[11:56] <mpt> SteveA, it means those items were in the diagram but there don't seem to be special pages for them
[11:56] <mpt> or at least I couldn't find them
[11:56] <SteveA> ok
[11:56] <mpt> so they might not be necessary
[11:56] <SteveA> or we might need a page for them
[11:56] <mpt> (for the special breadcrumb-menu definition of "necessary")
[11:56] <mpt> yes
[11:56] <SteveA> anyway, the page is fine for me.  i can create stuff on the server for this now
[11:57] <mpt> cool
[11:57] <SteveA> although, i probably won't do a lot today.  it is a public holiday
[12:09] <mpt> Anyone got time for a five-minute TAL review?
[12:33] <lifeless> mpt: sure
[12:34] <lifeless> SteveA: you had something I needed to put in this weeks review meeting
[12:34] <lifeless> review meeting in 25 
[01:11] <mpt> https://chinstrap.warthogs.hbd.com/~dsilvers/paste/fileIFhKX4.html if anyone's got five minutes spare for a review
[01:16] <lifeless> meep, got distraced
[01:16] <lifeless> who is here for the review meeting
[01:16] <lifeless>  ?
[01:17] <lifeless> mpt: is the sortable flag undesirable ?
[01:17] <SteveA> me
[01:17] <lifeless> woo
[01:18] <lifeless> lets have a meeting
[01:18] <spiv> I'm here.
[01:18] <lifeless> SteveA: what was the thing from the weekly lp meeting that was meant to go on the review meeting agenda ?
[01:18] <SteveA> it's a crap day for a meeting though
[01:18] <SteveA> hmm, don't know
[01:18] <SteveA> mpt: have you looked at doing the summary yet?
[01:20] <SteveA> Apr 27 15:37:20 <SteveA>	MeetingAction: lifeless to put this on the agenda for the reviewers' meeting
[01:20] <lifeless> yes, thats what I remember seeing in my scrollback.
[01:20] <lifeless> but what does it mean!
[01:20] <SteveA> the issue is about what level of quality is needed to get code merged into RF
[01:20] <SteveA> my view, which mark supports is the following
[01:21] <SteveA>  - database code and tests must be upto scratch
[01:21] <SteveA>  - python view code must also be up to scratch (but can be slightly more flexible)
[01:21] <SteveA>  - templates, in terms of usability and UI review and perfection, whatever, just land it
[01:21] <SteveA> so, basically, code must be good
[01:22] <SteveA> its functionality -- whether it does exactly the right thing or looks perfect to the user
[01:22] <SteveA> well, it is better to get it merged into the mainline so that it can be improved in that regard by everyone
[01:22] <SteveA> and not have the expense of maintaining forked code
[01:22] <SteveA> i guess i'm trying to make a distinction between code quality
[01:22] <lifeless> so. Why is core code special in that regard?
[01:22] <SteveA> and functional quality
[01:23] <SteveA> so, the code must be of a good quality, but what it does / how it *renders* to users can be imperfect
[01:23] <SteveA> the specific issue in the weekly development meeting was ddaa had been reluctant to merge some work
[01:23] <lifeless> what I am missing is how quality is defined: from a User Story design approach, doing the wrong thing is clearly of low quality
[01:24] <SteveA> because although the code was appropriately tested etc
[01:24] <SteveA> mpt had various outstanding review points about the UI being sufficiently usable
[01:24] <SteveA> so, my rule is: UI review can occur after landing in RF
[01:25] <SteveA> to answer your various questions
[01:25] <SteveA>  - i've tried to explain what the issue was
[01:25] <jordi> mdke: I wasn't quite here, no :)
[01:26] <SteveA>  - core code is special because it is maintained and relied upon by more people than the code to one particular page template.  a breakage or lack of maintainability in a database class will cause a big problem.  a breakage or lack of maintainability in a template affects just that one template.
[01:27] <jordi> hello all
[01:27] <lifeless> SteveA: aren't templates reused as components in rendered pages ?
[01:27] <SteveA>  - "doing the wrong thing" may well be of low quality, but it depends exactly how far off the mark it is.  "wrong" isn't black and white.
[01:27] <SteveA> very few templates are reused as fragments of other pages.
[01:27] <SteveA> those ones need more care.
[01:27] <lifeless> ok
[01:27] <spiv> SteveA: i.e. issues with core code are likely to impede other developers, but issues with UI code/templates usually won't, so are tolerable?
[01:28] <lifeless> I'm concerned that we're going to make the reviewers life hard with a too intangible requirement. But we'll try
[01:28] <SteveA> spiv: yes.  we must have very high standards for core code.  we can *choose* on particular occassions to relax requirements for ui code or templates. but...
[01:28] <SteveA> the main thrust of my argument is not that we should allow poor quality code
[01:28] <mpt> lifeless, it is nonsensical for any listing that's batched, because it sorts only the batch
[01:28] <SteveA> in terms of well formedness or tests
[01:28] <lifeless> SteveA: does this mean we can start requiring a matching balance of test types ?
[01:29] <SteveA> let me finish
[01:29] <lifeless> mpt: lets come back to that after the meeting. 
[01:29] <lifeless> SteveA: doing so, sorry.
[01:29] <mpt> As for the topic of the meeting, http://www.ok-cancel.com/comic/46.html
[01:29] <SteveA> but, that we may defer UI concerns such as "is this the best UI" or "does it allow all the UI things we want" or "is this somewhat misleading as a UI"
[01:30] <SteveA> until after the code has landed in RF, because these UI things can be incrementally improved later on
[01:30] <jordi> is labour day celebrated in your respective countries?
[01:30] <mpt> SteveA, I'm halfway through the meeting minutes (the page is up, not fully sorted yet)
[01:30] <SteveA> mpt's reviews are highly valuable
[01:30] <mpt> I'll do the rest when I wake up
[01:30] <SteveA> but they shoudn't delay getting stuff into RF in the same way as "no tests for database classes" should delay stuff getting into RF
[01:30] <SteveA> (done)
[01:31] <lifeless> so I like having a policy of 'only let good core code in'
[01:32] <sivang> mpt: hehe
[01:32] <lifeless> as I say though, I'm concerned that the line is going to be tricky to draw well, and separately it seems to me that the real distinction isn't around the place the code is used, its around the type of evaluation being made of it
[01:32] <sivang> mpt: (re the link)
[01:33] <lifeless> that is, in a template, a broken template is poor quality and shouldn't be let in. A template that doesn't meet the aesthetic for the UI however, can still be a good quality zpt program.
[01:34] <lifeless> ...
[01:34] <SteveA> yes
[01:34] <SteveA> no broken templates
[01:34] <lifeless> my point is that this separation of quality applies to the core code as well.
[01:34] <SteveA> but we can allow ugly templates
[01:34] <lifeless> because the core code has to evolve to support the UI as the UI's aesthetic is improved.
[01:34] <SteveA> i don't want to allow poor designs in core code to land in RF
[01:34] <SteveA> but, a poor UI design can land
[01:36] <lifeless> sure, I get that.
[01:36] <mpt> sivang, of course, we don't even have usability testing yet :-)
[01:36] <lifeless> and please dont think I'm trying to be difficult here. I want to ensure we have a predictable guideline for the reviewers
[01:37] <lifeless> and for the developers asking for review
[01:37] <lifeless> so one way to do this is to say:
[01:37] <sivang> mpt: :-)
[01:38] <lifeless> 'the pre-commit reviews do not review UI, they review database, python code and zpt/zcml syntax'
[01:38] <lifeless> UI reviews can be requested as a developer wishes, and will typically occur after merging to the mainline
[01:39] <lifeless> I think this will meet your objectives, and my desire for keeping the process clean and understandable
[01:39] <SteveA> i don't like that
[01:39] <lifeless> ok. Why not?
[01:39] <SteveA> i would like reviewers to actually use the code, where it changes UI
[01:39] <SteveA> to run launchpad with the new code, and try it out, when templates have been added
[01:39] <SteveA> so they can really understand what is being done
[01:40] <spiv> I think understanding the UI changes is relevant to understanding the code when reviewing.
[01:40] <sivang> that could be sort of usaility testing. 
[01:40] <SteveA> an important part of the review process is asking "what is this patch for?"
[01:40] <SteveA> then seeing if there are tests that test what it is essentially for
[01:41] <mpt> sivang, nah, LP developers know tens of times more about Launchpad than others do
[01:41] <mpt> spiv, that's why I added a "Demo URL:" line to the review template
[01:41] <SteveA> and, often, trying the code out, particularly for larger code landings
[01:42] <lifeless> I think changes that affect the ui definately need a willingness to run it, though it shouldn't be mandatory, to see what it delivers
[01:43] <lifeless> theres often no way for a reviewer to 'get it' without running it 
[01:43] <lifeless> but its also extremely time consuming to review multi thousand line patches
[01:43] <SteveA> i think it is useful for the reviewers to review the UI, but reviewers should often say "the UI design is sucky, but the correct elements are there.  land the code, and then look at the UI on a new branch"
[01:43] <spiv> It's easy to think "well, I'm not really sure what this bit of code is for, but I'm sure they have a good reason" if you don't understand the high-level aim of the patch.  Reading sufficiently clear tests can help a lot here.
[01:44] <lifeless> spiv: if you don't understand the reason, 'ASK'
[01:44] <spiv> lifeless: Random idea -- can we get jamesh to sort diffs to put tests at the top?
[01:44] <lifeless> spiv: sure. add it to the feature requests list please
[01:44] <spiv> lifeless: Sure, I make myself do that (or dig further), and almost always find it valuable :)
[01:45] <SteveA> lifeless: can we have a product called launchpad-development-infrastructure (or something), and file bugs against it?
[01:45] <lifeless> SteveA: ok. I think then that this policy will do what you want.. 
[01:46] <SteveA> lifeless: would you state the policy?
[01:46] <lifeless> "Reviewers please do not seek perfection on the UI. Its ok for the UI to need polish as long as the patch delivers the key goals it set out to."
[01:46] <lifeless> SteveA: I was just typing it :)
[01:47] <lifeless> SteveA: (product) that sounds fine. I'll set one up
[01:47] <SteveA> ah, i saw ".." and thought it was a typo.  "..." would have made me wait
[01:47] <lifeless> it was a typo.
[01:47] <lifeless> it was missing a .
[01:47] <SteveA> the policy looks good to me
[01:47] <lifeless> ok.
[01:48] <lifeless> I'd like to raise a different but related problem
[01:48] <SteveA> i think this would have addressed the issue with ddaa wanting to land the branches stuff only after getting the UI passed a full UI review
[01:48] <lifeless> I need to do some hard checks on this, but we seem to have a unbalanced code-area/test-volume ratio
[01:48] <lifeless> SteveA: ddaa asked for the ui review and chose to wait.
[01:49] <SteveA> yes, and had he been aware of this policy
[01:49] <lifeless> but its possible.
[01:49] <SteveA> he could have chosen to go ahead and merge, in accordance with the policy
[01:49] <lifeless> anyway, I'm supportive of this policy tweak
[01:50] <SteveA> what is "code-area" ?
[01:50] <lifeless> do you know the food group triangle ?
[01:50] <lifeless> at the top there are sweets and fats, at the bottom fruit and vegesm and in the middle carbs and protein
[01:50] <lifeless> it has a wide base and the top is a point
[01:51] <SteveA> i've seen a picture
[01:51] <lifeless> there is a similar triangle for testing: 
[01:52] <lifeless> small numbers of GUI tests, moderate numbers of integration tests that test connecting or composing groups of functionality, and lots of unit tests testing specifics
[01:52] <SteveA> the particular food triangle you mentioned doesn't apply to hunter-gatherer societies, who have to catch and kill their own protein, and would need to start farming from scratch to raise grains.
[01:52] <sivang> lifeless: this is sorted calorie wise?
[01:53] <sivang> (given sweets are at the top)
[01:53] <SteveA> i think we should have comprehensive acceptance / workflow tests, that do test the GUI in a way that is comparable to users actually using the GUI
[01:53] <lifeless> SteveA: until I've done some actual hard stats I'm hesitant to assert this, but I have a strong feeling that we have this inverted for much of launchpad.
[01:54] <SteveA> we should have unit tests of coarse-grained units -- the same units that we think about and work upon
[01:54] <SteveA> we have quite large classes in launchpad typically, and i support having tests of those classes
[01:55] <SteveA> by experts-in-OO terms, our classes are far too large, and have too many purposes for each class
[01:55] <spiv> lifeless is referring to something like this, I think: http://photos1.blogger.com/blogger/7546/664/1600/testing-pyramid.jpg
[01:55] <SteveA> but, this pattern works quite well for launchpad development, given the specific nature of our development
[01:55] <SteveA> but, i think we need to take this into account when deciding what "unit test" means for launchpad
[01:56] <lifeless> SteveA: sure. 
[01:56] <SteveA> i support the idea of more unit tests, provided we choose the units well
[01:57] <lifeless> my concern about this is not driven by OO-size considerations
[01:57] <SteveA> we will see individual pagetests disappear, and be replaced by user-interaction-story tests
[01:57] <lifeless> its that its easy to generate untested code 
[01:57] <lifeless> when a UI changes
[01:58] <SteveA> i agree with the goal of having tests that still give good coverage in the absence of UI
[01:59] <lifeless> if the sole test for a code path is by a UI level test - user-story or other - and the UI changes, then the code becomes untested
[01:59] <lifeless> SteveA: the corollary to that goal is that the number of tests that do not test the UI should be >> than the number of UI tests.
[01:59] <SteveA> i don't think that follows
[01:59] <SteveA> consider a UI that is complex and has many paths through it
[02:00] <SteveA> i think the corollary is that the coverage of non-ui tests should exceed the coverage of UI tests
[02:00] <SteveA> this coverage may be achieved efficiently or inefficiently in terms of number of tests
[02:00] <lifeless> ok, sure.
[02:01] <mpt> If Launchpad becomes more like an application, and less like a series of forms, it will become less testable at the pagetest level anyway (JavaScript, XmlHttp, etc)
[02:01] <lifeless> so I'm raising a request that the review team be allowed to start pushing back on patches where the non UI test coverage does not meet or exceed the UI test coverage
[02:02] <SteveA> spiv: i've seen that pyramid before.  i don't think it is a helpful diagram.  it presents something that is often true of well-tested projects, without examining the underlying reasons that such a distribution of types of test leads to well-testedness
[02:02] <SteveA> lifeless: agreed.
[02:02] <lifeless> SteveA: thank you.
[02:02] <SteveA> i like that you have phrased that without using "unit test" or "integration test"
[02:03] <lifeless> SteveA: softly softly catchee mousey
[02:03] <SteveA> and focused on UI test coverage vs non-UI test coverage
[02:03] <SteveA> i will ask that reviewers apply a similar focus in their reviews
[02:04] <SteveA> bbiam
[02:04] <lifeless> SteveA: classifying the nature of the written tests is secondary to classifying the risks inherent in how they achieve their testing
[02:05] <lifeless> My perspective on the most efficient way to write tests for the code paths in a system is irrelevant to addressing the specific risk I raised here
[02:05] <SteveA> good.  we agree on what this is all about.
[02:06] <lifeless> ok. 
[02:06] <lifeless>  * review queue status
[02:07] <lifeless> kiko-afk still has cprovs branches to review, and as cprov is well and truely back, they are now getting quite old.  But I understand staff reviews affecting that
[02:07] <lifeless> jamesh has an old bzrtools branch - two weeks now
[02:07] <lifeless> and SteveA - you have a branch of brads to review
[02:08] <lifeless> other than that its under control
[02:08] <SteveA> noted
[02:08] <lifeless> I'd like to change our process to differentiate between 'days in current state' and 'days since put up for review'
[02:09] <lifeless> can you allocate some time to jamesh to do that ?
[02:09] <lifeless> it should be trackable quite easily
[02:10] <SteveA> i propose you file a bug on that, and james can add it to his queue, to be started after he's landed the latest work for david
[02:10] <lifeless> thanks
[02:10] <lifeless> ok, next meeting, May 8th, same time ?
[02:11] <SteveA> +1
[02:11] <lifeless> +1
[02:12] <spiv> +1
[02:12] <lifeless> ok
[02:12] <lifeless> any last requests ?
[02:14] <lifeless> 111
[02:14] <lifeless> 110
[02:14] <lifeless> 101
[02:14] <lifeless> 100
[02:14] <lifeless> 011
[02:14] <lifeless> 010
[02:14] <lifeless> 001
[02:14] <lifeless> 000
[02:16] <richips> Hi; I'm traslating using Rosetta, and I've been asking if there isn't a dictionary for the 'labeled expressions' such as the menu names or the gui buttons
[02:17] <richips> If there's not, that would be an improvement for faster and standared translations
[02:19] <lifeless> SteveA: did you have a look at my talk ?
[02:20] <richips> no one works in rosetta?
[02:20] <lifeless> richips: carlos does, but hes on leave today
[02:20] <lifeless> its a public holiday in much of the world
[02:20] <richips> ya, here (in spain) it's too
[02:21] <richips> :P
[02:27] <spiv> richips: I don't think we have that yet, but there are some plans in that direction: https://wiki.launchpad.canonical.com/LanguageGlossaryLinks
[02:33] <richips> spiv, that sounds cool
[02:34] <richips> I have see an application, too, named KBabel... It's wright to use it in ubuntu translation proyects?
[02:40] <lifeless> SteveA: https://launchpad.net/products/launchpad-development-infrastructure if you would care to review the product ?
[02:54] <salgado> hey lifeless. do you have a few minutes to disccus the remaining points of that branch-puller changes, so I can land it? (kiko's been complaining a lot lately, because he get spammed with errors all the time)
[02:56] <lifeless> salgado: sure, a few minutes
[02:59] <salgado> lifeless, so, have you seen my latest reply?
[03:00] <lifeless> checking
[03:02] <lifeless> found it
[03:04] <lifeless> replying
[03:06] <bradb> spiv: ping
[03:06] <spiv> bradb: pong
[03:06] <bradb> spiv: The problem with syncUpdate is that I need to add that method to the IBug interface.
[03:06] <salgado> lifeless, there's a few other exceptions that kiko reported: https://chinstrap.ubuntu.com/~dsilvers/paste/filecIy7g3.html
[03:07] <spiv> bradb: Ah, good point.  Damn.
[03:07] <bradb> don't I?
[03:07] <lifeless> salgado: lets get a patch in for the current ones we've debugged so far
[03:07] <lifeless> salgado: the perfect is the enemy of the good
[03:07] <salgado> indeed
[03:07] <bradb> maybe adding it only to ZCML will work, but hm
[03:07] <spiv> Well, if the test in question returns a security proxied object...
[03:08] <bradb> right, so adding it just in ZCML works
[03:08] <bradb> i.e. as part of allow attributes="..."
[03:09] <bradb> spiv: Is it okay to have to allow that method in ZCML?
[03:09] <spiv> Cool  Well, I'm happy either way now.  All things being equal, I'd prefer using syncUpdate, but things aren't quite equal here :)
[03:09] <bradb> i don't mind the ZCML hack.
[03:09] <spiv> I'm now wavering, so pick whichever you prefer.
[03:10] <bradb> Presumably most objects don't have this problem, but bugs have special security requirements, hence the ZCML change needed here.
[03:12] <spiv> bradb: Btw, when you merge this, don't forget to clean up the second, older entry for it on PendingReviews in salgado's queue.
[03:13] <lifeless> salgado: replied
[03:13] <bradb> spiv: right, i just removed both old requests of mine from salgado's queue, thanks
[03:14] <spiv> lifeless: is there a way to specify a revno in a url?
[03:14] <lifeless> mpt: I'm crashing, I presume you have. Ping me tomorrow.
[03:15] <lifeless> spiv: for pending-reviews yes, but pqm doesn't grok it yet.
[03:15] <lifeless> http://...../;revno=xxxx
[03:15] <spiv> lifeless: Cool.  Thanks.
[03:16] <lifeless> salgado: re new exceptions, the branch client error - file a bug on the authserver
[03:16] <lifeless> salgado: we cant catch that one, because we cant report status from there on in
[03:17] <lifeless> (because we dont know the cause, transient or not)
[03:17] <lifeless> salgado: the   File "/srv/sm-ng/production/launchpad/cronscripts/../lib/bzrlib/transport/http.py", line+182, in has mutter('url error code: %s for has url: %r', e.code, path)
[03:17] <lifeless> AttributeError: URLError instance has no attribute 'code'
[03:18] <lifeless> exception, please check the upstream bzrlib - it should not access e.code in UrlError, only in HttpError, and catch URLError in a separate block
[03:18] <lifeless> if upstream if faulty, file a bug there. If not, hassle spiv to finish merging upstream into rocketfuel
[03:18] <lifeless> finally,   File "/usr/lib/python2.4/urllib.py", line 813, in __init__ self.read = self.fp.read
[03:18] <lifeless> AttributeError: 'NoneType' object has no attribute 'read'
[03:19] <lifeless> looks like a failure to read the branch list. So it was probably during a rollout.
[03:19] <lifeless> or perhaps when gangotri:1 fell over
[03:19] <lifeless> gnight, HTH
[03:20] <salgado> yeah, thanks lifeless.
[05:33] <avoine> hi, i just want to know if launchpad is written in python?
[05:57] <dsas> avoine, yes it is.
[06:45] <WaterSevenUb> Something seems wrong with this https://launchpad.net/distros/ubuntu/dapper/+source/ubuntu-docs/+pots/website-index/ doc side or rosetta side?
[07:54] <dilys> Merge to devel/launchpad/: [r=spiv]  Fix bug 29752 (If a bug is marked as a duplicate, its subscribers should be notified when the duplicate bug changes) (r1842: Brad Bollenbach)
[08:00] <lamont> specifically, what's the best module to use for parsing xml returns from web pages in python...
[08:00] <lamont> (granted that it's obviously ECHAN)
[08:14] <mdke> jordi: np, got your mail, thanks.
[09:52] <nictuku> is the c
[09:53] <nictuku> is the CoC signing process working, as far as you all know?
[09:55] <salgado> nictuku, it's working for the 1.0 version, but not for the 1.0.1
[09:56] <nictuku> is there a bug for that, so I can know when it's fixed?
[09:56] <salgado> nictuku, bug 39547
[09:56] <Ubugtu> Malone bug 39547 in launchpad "Code of Conduct 1.0.1 signatures not accepted" [Critical,Confirmed]  http://launchpad.net/bugs/39547
[09:57] <nictuku> thank you!
[09:58] <salgado> you're welcome. :)
[10:09] <seaLne> hi is there a way to remove an open poll?
[10:10] <Ju> Hi all !
[10:10] <seaLne> just asked
[10:10] <Ju> ;-)
[10:10] <Ju> ok...
[10:11] <Ju> seaLne: https://launchpad.net/products/launchpad/+bug/28670
[10:11] <Ubugtu> Malone bug 28670 in launchpad "Shouldn't be able to create a poll with zero options" [Normal,Confirmed]  
[10:15] <salgado> seaLne, no, it's not possible to remove polls at all.  why do you want to remove it, though?
[10:15] <seaLne> it has no voting options
[10:16] <seaLne> https://launchpad.net/people/behindubuntu/+poll/frequence
[10:18] <salgado> I see. I need to change that poll-creation form so it's possible to create the options at the same time the poll is created. 
[10:18] <seaLne> it would also be usefull to be able to modify polls after they are open
[10:21] <salgado> you can do some changes, but most of them are not allowed, and that's intentional
[10:21] <salgado> for instance, adding a new option after the poll is open and somebody already voted will not be fair with these people who already voted
[10:22] <salgado> in this case it should be allowed, though, as nobody voted yet
[10:24] <Ju> salgado: totally agree.
[10:25] <Ju> but an open poll should not determined only by the start date, I thought (i've just created the poll) it should consider a "I'm done" settings
[10:26] <Ju> why not use a input for each questions instead of large textarea ?
[10:29] <salgado> Ju, you mean, the "Proposition" text area?
[10:30] <Ju> yep
[10:30] <salgado> the proposition need to be kept on that page, that's not the same thing as the options that are going to be voted
[10:30] <Ju> because I don't see differences between proposition and options
[10:30] <salgado> we need to keep the proposition and maybe add another textarea where you can enter one option on each line
[10:30] <Ju> or maybe I'm all wrong ;-/ https://launchpad.net/people/ubuntu-fr/+poll/test
[10:32] <salgado> Ju, I think https://launchpad.net/people/ubuntu-dev/+poll/tb-nomination-mjg59-2005 is a good example
[10:34] <Ju> oh ok... I see ;-)
[10:34] <salgado> but the UI doesn't make this very clear, so if you file a bug about it I promise I'll improve the UI when I get to fix all these poll-related bugs
[10:34] <Ju> ok ;-)
[10:35] <salgado> I haven't done so yet because there's other higher priority bugs/specs that I need to fix/implement
[10:35] <Ju> I can understand...
[10:36] <Ju> the main thing is : "The proposition that is going to be voted :" become when done : "Poll overview" it's confusing, I'll fill a bug