[12:49] <dilys> Merge to thelove@canonical.com/dists--bazaar--1.5: new build (patch-61)
[12:50] <dilys> Merge to thelove@canonical.com/bazaar--devo--1.5: Added a --versions to commands displaying revision lists. (patch-47: Matthieu.Moy@imag.fr)
[09:26] <dilys> Merge to thelove@canonical.com/dists--bazaar--1.5: new build (patch-62)
[09:28] <dilys> Merge to thelove@canonical.com/bazaar--devo--1.5: --from and multiple versions support for "missing" (patch-48: Matthieu.Moy@imag.fr)
[09:29] <bob2> daf: why does dilys sometimes not have a name?
[09:29] <bob2> are those pqm requests really sent without one?
[09:56] <ddaa> The name is the name of the archives of the merged patchlogs (when != from the destination archive)
[09:56] <ddaa> you can have several names when doing a roll-up merge
[09:57] <bob2> hm, I guess I assumed it just took the From: name of whoever sent the request
[09:57] <ddaa> and none in the case of an automatic commit, or same archive merge
[09:57] <ddaa> it _is_ confusing
[09:57] <bob2> you're home again now?
[09:57] <ddaa> I only understood it after going WTF at lifeless
[09:57] <bob2> hah
[09:58] <ddaa> FSVO again, I guess
[09:58] <bob2> hah
[09:58] <bob2> got time for a python q then?
[09:58] <ddaa> sure
[09:58] <bob2> so, I have a moderately complicated class
[09:59] <bob2> with lots of neatly factored-out little methods
[09:59] <bob2> testing the little methods is non-trivial, tho, since they depend on object state
[09:59] <bob2> would you go with "setting state manually in the test suite" or "making the methods take state as arguments"?
[10:00] <ddaa> That's not a python specific question, rather a TDD best practices question
[10:00] <bob2> yeah, true
[10:01] <ddaa> What I would do, would be doing feature tests, i.e. testing whole stories setting up the state running the method being tested, and checkinh the output and new state, without breaking encapsulation.
[10:01] <bob2> hmmm
[10:01] <ddaa> Such tests are useful to have, but that's not the end all...
[10:02] <bob2> I have that too, but I feel safer having low-level "this method does what it's supposed to" tests, too
[10:02] <ddaa> Right, one of my problem is that I have trouble writing unit tests.
[10:02] <bob2> the other issue is that some methofs change external state (e.g. write out files, send email), which makes it a pain to test things higher up the stack
[10:03] <bob2> does it seem like a terrible idea to test a subclass that replaces the externally-modifying tests with stubs?
[10:03] <ddaa> I'd rather set the state manually in your case. Because if your API is  message-passing style, designed around accumulating state in the object, it would not make sense to graft a functional API on top of it for testing.
[10:04] <bob2> that's what I was thinking
[10:05] <ddaa> I would stub out the externally modifying classes too. Actually, google for "Mock Objects", that's a kind of stub with testing smarties.
[10:05] <bob2> ah, I forgot about them
[10:05] <bob2> well, in this context, anyway; I do use them for testing inter-object stuff
[10:05] <ddaa> Basically, a mock a stub that provides machinery to test interaction with other objects.
[10:06] <ddaa> While a mere stub is just something that provides enough API to be able to run the code being tested.
[10:07] <ddaa> Nice of you to ask me about that, but I'm really no authority in OOP...
[10:07] <ddaa> re: subclassing
[10:08] <ddaa> lifeless wolud tell you to use compositon :P
[10:08] <bob2> haha
[10:08] <ddaa> Which might well make sense in that case.
[10:09] <bob2> that's a good point, actually
[10:09] <ddaa> to have a separate object to handle communication with the outside world.
[10:09] <bob2> thanks for your thoughts :)
[10:09] <ddaa> my pleasure
[11:40] <mvo> could someone of the python people review a interface for me? python-apt has a "pure" python interface now that hides the details from apt_pkg. a quick look if it looks pythonic enough would be really cool
[12:21] <ddaa> I do not think we have any python-dev here :P
[12:22] <ddaa> mvo: where can I find the code you want reviewed? From the ubuntu archive?
[12:23] <mvo> ddaa: http://people.ubuntu.com/~mvo/arch/ubuntu, michael.vogt@ubuntu.com--2005/python-apt--mvo--0, directory "apt"
[12:23] <mvo> ddaa: thanks :)
[12:23] <ddaa> did not say I would do it :P
[12:23] <mvo> ddaa: aha :)
[12:24] <ddaa> Thinking of it, I probably focus on my assigned tasks... but I'm curious :)
[12:24] <ddaa> * I should probably
[12:25] <mvo> ddaa: well, if you look at it, it would be very kind if you could /msg me your comments!
[12:25] <ddaa> mvo: baz remark, you can just say "baz get http://people.ubuntu.com/~mvo/arch/ubuntu/python-apt--mvo--0/"
[12:26] <mvo> ddaa: that's a nice feature. since what version of baz does it exist?
[12:26] <ddaa> 1.3
[12:26] <ddaa> or 1.4...
[12:28] <ddaa> actually, without the trailing slash...
[03:26] <dilys> New Malone bug 1609 filed on Bazaar by David Allouche: Possible arch-cache race condition leading to data corruption
[03:26] <dilys> https://launchpad.ubuntu.com/malone/bugs/1609
[04:49] <SteveA> silbs: hello
[04:51] <silbs> SteveA: hi. welcome back.
[04:54] <kiko> ahoy
[04:54] <kiko> carlos!
[04:54] <kiko> daf!
[04:55] <carlos> kiko, !
[04:55] <lifeless> ddaa , bob2 - jsuch long running tests are doctests in nasture. that said, you will have many repeats to handle different expected conditions, which is where I'd use unit tests with setup and tteardown that give appropriate eady to test objects
[04:55] <carlos> so Brazil was connected again?
[04:55] <kiko> carlos, apparently. fuck fuck fuck.
[04:55] <ddaa> doctests?
[04:55] <ddaa> long-running?
[04:56] <ddaa> -ECONTEXT
[04:56] <carlos> kiko, that's because they know I'm going there and know I need network connection :-D
[04:56] <kiko> yeah
[04:56] <kiko> so what's up with daf?
[04:56] <morgs> Coming soon, wireless AP with built in satphone...
[04:56] <carlos> kiko, waiting for me
[04:56] <carlos> kiko, at London
[04:59] <lifeless> ddaa: do a lastlog lifeless ;-)
[05:00] <ddaa> Doesn't fit.
[05:00] <ddaa> What do you mean by "lastlog"?
[05:02] <lifeless> '/lastlog lifeless'
[05:02] <ddaa> my irc client does not know about this command
[05:02] <lifeless> you and bob2 invoked my name
[05:03] <carlos> mpt, around?
[05:03] <ddaa> Yes, I remember this discussion. But still, I do not see the connection with "doctests in nature" and "long running".
[05:03] <carlos> mpt, I got your camera, but I was not able to get the memory card
[05:04] <kiko> carlos, what sort of memory card?
[05:04] <lifeless> oh, if you need to setup a object and then call method X method Y beother bar on it and they are idempotent methods, a doctest will  fit that quite well IMO
[05:04] <lifeless> though not in th emethod docstring 
[05:04] <carlos> kiko, sd
[05:05] <kiko> carlos, we can buy them cheap here, no worries
[05:05] <carlos> ok
[05:05] <carlos> cool
[05:13] <mpt> carlos: excellent, thanks
[05:14] <carlos> mpt, also, they gave me a couple of rechargable AA batteries and the plug to recharge them for free
[05:16] <jamesh> carlos: what sort of camera?
[05:17] <carlos> mpt, ? I don't remember the model
[05:18] <mpt> carlos: Let me guess, it's a Spanish battery charger?
[05:18] <carlos> mpt, yeah ;-)
[05:18] <carlos> European
[05:18] <mpt> jamesh: Pentax S-50
[05:18] <mpt> carlos: Yeah, my last one came with a US battery charger that was similarly useless
[05:18] <mpt> carlos: you can keep it :-)
[05:20] <jamesh> mpt: cool.  that's the updated version of my camera (I have the S40)
[05:20] <carlos> well, It's with my luggage and I already checkin so...
[05:22] <carlos> lifeless, any chance to process my cherrypick request?
[05:23] <lifeless> carlos: dude, we've been offline for 2 days
[05:23] <lifeless> no internet at the sprint or the hotel.
[05:23] <carlos> lifeless, I know, that's why I'm asking :-)
[05:23] <carlos> if there is a chance
[05:24] <lifeless> carlos - it conflicts
[05:24] <carlos> fuck
[05:24] <lifeless> stub tried 
[05:24] <carlos> ok
[05:24] <carlos> then will need to wait until next week
[05:24] <carlos> I don't have time to fix it
[05:24] <carlos> before the flight
[05:24] <lifeless> can it wait till the amin rollout on monday/tuestday ?
[05:25] <carlos> yeah, because I cannot fix it now, it should wait anyway
[05:41] <carlos> see you!
[06:19] <ddaa> morgs: ping
[06:20] <morgs> ddaa: gnip
[06:20] <ddaa> is your change for branch display currently on staging?
[06:20] <morgs> ddaa: I'll take a quick look...
[06:21] <morgs> ddaa: yes, here's an example: https://staging.ubuntu.com/products/gstreamer
[06:22] <ddaa> Is there a policy for using tables for everything?
[06:23] <ddaa> I think that branch listing would be better served, semantically, by a dt/dd list
[06:23] <ddaa> and that would mitigate the ridiculously-wide-table problem
[06:24] <morgs> ddaa: I'm fine with that, I'll give it a try on Monday
[06:24] <ddaa> thus helping avoid the annoyingly-small-font-size problem...
[06:24] <ddaa> sabdfl had me refocused on all-things-branches-and-bazaar-in-launchpad-and-friends recently
[06:25] <morgs> Do you think the title and URL are the right things to display?
[06:25] <ddaa> so, time permitting, I would be able to contribute more code to this kind of things
[06:25] <morgs> Great
[06:25] <ddaa> But I'm so happy with your work, that I really see little reason for that :)
[06:30] <morgs> ddaa: If you think there's anything a user would benefit from, let me know. AFAIK the url is the most useful thing?
[06:30] <ddaa> URL is the only useful thing ATM
[06:30] <morgs> OK
[06:31] <ddaa> There's something in my pipe coming right from HQ which I suspect is related to branch display. I think in the short term you can just get away with listing branches in a <ul>
[06:32] <ddaa> I mean, listing urls
[06:32] <morgs> Makes sense. OK.
[06:32] <ddaa> In the medium term I gather I'm expected to write something that will make the rest of the webapp world pale in humility ;)
[06:33] <morgs> Sort of google maps meets the bazaar? :-P
[06:34] <morgs> Anyway, I'm out of here.
[06:35] <ddaa> Yeah, sort of something like that.
[06:36] <ddaa> OMG, this satellite view thing is painfully cool
[06:54] <dilys> Merge to thelove@canonical.com/dists--bazaar--1.5: new build (patch-63)
[06:55] <dilys> Merge to thelove@canonical.com/bazaar--devo--1.5: Implicit ^...$ around the regexps for rbrowse (patch-49: Matthieu.Moy@imag.fr)
[07:34] <jordi> hmm
[07:34] <jordi> it's tricky to find the url that generates statistics for all the breezy modulesi n Catalan
[07:35] <jordi> ah, got it
[07:42] <dilys> New Malone bug 1612 filed on The Launchpad by Scott James Remnant: After logging in, launchpad leaves me up a blind alley without my clothes
[07:42] <dilys> https://launchpad.ubuntu.com/malone/bugs/1612
[07:44] <dilys> New Malone bug 1613 filed on The Launchpad by Robert Collins: cannot sign coc due to launchpad ignoring signing keys
[07:44] <dilys> https://launchpad.ubuntu.com/malone/bugs/1613
[07:53] <dilys> New Malone bug 1614 filed on The Launchpad by Robert Collins: bugs related people portal should uniqify the maintainers list
[07:53] <dilys> https://launchpad.ubuntu.com/malone/bugs/1614
[07:57] <bob2> lifeless: yeah, I have both doctests and unittests
[08:02] <cprov> lifeless: I really didn't get your point on bug # 1613, want to talk about .
[08:14] <bob2> (for that reason)
[08:44] <lilo> hi all
[08:45] <Kinnison> Hello lilo
[08:45] <lilo> sorry about the kline earlier, one of your IP's hit a user limit.... we have to kline on those to catch bot attacks quickly enough
[08:45] <lilo> I added new I: lines, so the problem shouldn't recur anytime soon
[08:45] <lilo> (hey Kinnison)
[08:47] <Keybuk> lilo: are bots typically registered with NickServ?
[08:47] <Kinnison> Thanks for that lilo
[08:48] <Kinnison> It's a conference IP :-)
[08:48] <lilo> oic! yah, we need to figure out how to get people to check with us before setting those up
[08:48] <Keybuk> if not, that could be a sweet optimisation for the auto-killer
[08:48] <lilo> more likely we just need some way to look for conferences in process and plug a calendar into the kline process
[08:48] <lilo> I've been thinking about that, but we don't have the facilities yet
[08:49] <lilo> anyway, let us know if you need anything at all
[08:54] <SteveA> cheers lilo
[09:05] <dilys> Merge to thelove@canonical.com/dists--bazaar--1.5: new build (patch-64)
[09:06] <dilys> Merge to thelove@canonical.com/bazaar--devo--1.5: baz lock-revision without argument, informative message at the end (patch-50: Matthieu.Moy@imag.fr)
[10:11] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [trivial, partial rs=SteveA]  nuke the old MaloneApplicationView, fix a typo and deal with null SourcePackage.currentrelease -- fixing bug 1549 in the process. Also fixes (once and for all) the handling of the bugtrackertype enum in checkwatches (patch-2190: christian.reis@canonical.com)
[10:37] <kiko> otherwise pqm might go rusty!
[11:32] <dilys> Merge to rocketfuel@canonical.com/launchpad--devel--0: [trivial]  fix bugs link in facet menu for person (patch-2191: james.henstridge@canonical.com)