[10:33] <JamesTait> Morning all! :)
[10:44] <karni> Morning guys :)
[11:04] <gatox> good morning!
[11:09] <mandel> gatox, morning!
[11:10] <gatox> mandel, how are you?? we have u1-cp fixed in jenkins! :D
[11:11] <mandel> gatox, hurray!
[11:11] <mandel> gatox, here everything ok, although I need some reviews from you for the fsevents daemon
[11:11] <mandel> gatox, if you can read c++ you can read objective-c :)
[11:12] <gatox> mandel, ah yes.....  i'll do it right now
[11:13] <mandel> gatox, thx!
[11:13] <mandel> gatox, read the code and ask any question I'd like you to also be able to at least understand it
[11:14] <gatox> mandel, ack!
[11:45] <gatox> mandel, i'll fix the headers of the files, it says: //  Copyright (c) 2012 __MyCompanyName__. All rights reserved.
[11:45] <mandel> gatox, there is another branch for that with no code changes :)
[11:46] <gatox> mandel, ahhhh ok jeje
[11:46] <mandel> gatox, i don't want to add noise to a code review
[11:46] <gatox> of course
[12:17] <gatox> mandel, the modified files looks fine for me..... i'm reviewing the new files now, i found a little typo at this moment: // assert that the correct metho type is received
[12:17] <gatox> in: objc/FsEventsTests/FsEvents/FSEventsDispatcherTests.m
[12:17] <mandel> gatox, cool, can you add the error to the mp please?
[12:17] <gatox> ok
[12:18] <mandel> gatox, I'm got a feeling you will find more than one
[12:18] <gatox> mandel, it's silly..... but better to fix it now that the file is new, that see it appear mix with some code in the future
[12:19] <mandel> gatox, 100000 time agreed :)
[12:34] <mandel> ok, lunch!
[12:44] <gatox> mandel, message added to the MP
[12:44] <gatox> mandel, i added a trace of the compilation failing too
[12:56] <ralsina> morning!
[13:05] <dobey> hi
[13:06] <gatox> ralsina, dobey hi
[13:09] <alecu> mandel, https://code.launchpad.net/~mandel/ubuntuone-fsevents-daemon/add-run-tests/+merge/106227
[13:14] <ralsina> mandel, alecu: do you guys remember why ussoc on windows doesn't die of timeout like on linux?
[13:14] <ralsina> mandel, alecu: I know we had a good reason, but can't remember it
[13:14] <ralsina> mandel, alecu: and it may be time to reconsider it
[13:15] <alecu> ralsina, I think the reason was "we can make it behave properly later"
[13:15] <ralsina> alecu: because it was doing things like starting twice, right?
[13:16] <alecu> ralsina, can't recall. nessita is the one with the prodigal memory for this kind of things.
[13:16]  * alecu can hardly recall what he ate yesterday.
[13:16] <alecu> oh, fish it was.
[13:16]  * alecu smells t-shirt
[13:17] <alecu> fish it was.
[13:43]  * mandel back
[13:44] <mandel> gatox, i have not looked at the trace, but did you add OCMock to the path?
[13:44] <mandel> ralsina, alecu, AFAIK the process will start, die and will never start again
[13:45] <gatox> mandel, my bad..... forget to read the description of the change
[13:45] <gatox> trying that now
[13:45] <mandel> gatox,  :)
[13:45] <mandel> gatox, no worries, if I know the issue without looking it means that I expected that to happen hehehe
[13:46] <mandel> alecu, interesting the segfault, which revno is that?
[13:46] <briancurtin> elopio: on your machine that has this testability stuff, where is testability.dll and did it come with PyQt or did you install it separately?
[13:48] <gatox> mandel, can you show me your path please..... so i can find it easier
[13:49] <mandel> gatox, with xcode 4 it should be in /Applications/Xcode.app/Contents/Developer/Library/Frameworks
[13:49] <mandel> gatox, or you can configure the env var :)
[13:50] <gatox> mandel, thx
[13:50] <alecu> mandel, I got the segfault with trunk + your branch merged in.
[13:50] <mandel> gatox, I placed it there for convenience
[13:50] <mandel> alecu, that probably means that the run shell got merged wrong, can you pastebin it?
[13:50] <mandel> alecu, I might need to fix the way it merges agains trunk
[13:51] <ralsina> briancurtin: I don't know where it comes from but I know we will have trouble with it because it's a plugin
[13:51] <ralsina> briancurtin: and plugins don't work well with py2exe
[13:51] <alecu> mandel, bzr says that run-tests.sh was freshly created, not merged.
[13:52] <mandel> alecu, hm.. weird, let me check
[13:53] <mandel> alecu, oh, is the run-tests branch, I know what that happens it uses otest directly, give me 2 mins and is fixed :)
[13:53] <alecu> ralsina, briancurtin: do we really want testability.dll in our installer? Though I suppose it might make sense for the nightlies.
[13:54] <briancurtin> if i can find out how to make it work, i'd only package it on a test build
[13:54] <briancurtin> so add some type of flag, or change the SHOW_CMD option to TEST_BUILD and key it off that
[13:55] <mandel> alecu, fix pushed, you should not longer have that segfault
[13:58] <ralsina> alecu: just for nightlies yes
[13:59] <gatox_mac> mandel, i'm getting this:
[13:59] <gatox_mac> ** BUILD SUCCEEDED **
[13:59] <gatox_mac> Running tests
[13:59] <gatox_mac> xcodebuild: error: The project 'FsEvents' does not contain a scheme named 'FsEvents'.
[13:59] <gatox_mac> any clues?
[13:59] <mandel> gatox_mac, uhm.. weird, there should be an fsevents scheme, can you please open the project in xcode?
[14:00] <mandel> gatox_mac, the scheme is that thing next to the stop button, can you let me know what it has?
[14:01] <gatox_mac> opening...
[14:02] <gatox_mac> mandel, wait… it's installing some components before continuing
[14:04] <mandel> gatox_mac, hm.. interesting
[14:07] <gatox_mac> mandel, so… i double click the .xcodeproj file Xcode opens and says: workspace integrity: couldn't load fsevens.xcodeproj
[14:08] <mandel> gatox, wtf?! that is new to me..
[14:08] <dobey> hrmm
[14:08] <mandel> gatox_mac, does it offer any fixing?
[14:09] <gatox_mac> mandel, i don't see anything to fix it
[14:09] <gatox_mac> mandel, mmm it seems to be fix if i reopen
[14:10] <gatox_mac> mandel, trying to run the tests again
[14:10] <mandel> gatox_mac, if that works, can you do a bzr diff and tell me what changed?
[14:10] <gatox_mac> mandel, now: xcodebuild: error: Failed to build project FsEvents with scheme FsEvents.
[14:10] <gatox_mac> 	Reason: Scheme "FsEvents" is not testable.
[14:11] <mandel> gatox_mac, give me a sec, I'll test it using my test account
[14:11] <gatox_mac> ok
[14:11] <dobey> everyone!
[14:11] <gatox_mac> dobey, yes?
[14:12] <mandel> hello!
[14:12] <dobey> please don't set branches to "Approved" for merging into trunk for now. there is a problem with tarmac, which needs to be resolved first
[14:12] <mandel> ack
[14:12] <dobey> am trying to get it fixed
[14:12] <gatox_mac> dobey, ack
[14:18] <dobey> wtf
[14:19] <dobey> "Lote de 100 fontes de chocolate a menos de metade do valor de mercado!"
[14:19] <dobey> con churros!
[14:22] <mandel> gatox_mac, I managed to reproduce the issue, xcode is stupid..
[14:22] <gatox> mandel, it seems that the rule is: "if you want to know if something really works.... it has to go through any of my machines" jeje
[14:24] <mandel> gatox,  is more of, IDE usually are stupid :)
[14:25] <gatox> mandel, i have nothing to say jeje
[14:26] <mandel> gatox, I'm about to push the change.. only if people that wrote IDEs could make it simple to know which files are per user and which not..
[14:27] <gatox> mandel, well... i know of an ide that is really cool and create only one file in the project folder and it's for the user jejeejej
[14:28] <ralsina> dobey: that's portuguese not spanish!
[14:28] <mandel> ralsina, or spanish with a churro in your mouth hehehe
[14:29] <gatox> mandel, jejejjee
[14:29] <gatox> stupid bug!!
[14:29]  * gatox gets angry
[14:30] <dobey> heh
[14:31] <mandel> gatox, I fix the issue pushing now
[14:32] <gatox> mandel, ok..... running the tests
[14:32] <mandel> gatox, revno 25
[14:32] <mmcc> morning folks!
[14:33] <mandel> gatox, turns out that you have to explicitly tell xcode that a scheme is shared or it will placed in the userdata
[14:33] <mmcc> my connection died 7 hours ago, so I'm already caught up!
[14:35] <gatox_mac> mandel, http://paste.ubuntu.com/1014864/
[14:35] <mandel> mmcc, did you hear anything from the ocmock guys?
[14:36] <mandel> gatox_mac, revno?
[14:36] <mmcc> mandel: just going thru email now
[14:36] <dobey> oh, pylint. grmbl
[14:36] <dobey> must get rid of that junk
[14:36] <mandel> dobey, yes
[14:38] <gatox> well..... maybe in half an hour after running ubuntuone-client tests i'll be able to propose the branch... :@
[14:38] <mmcc> mandel, I did get a response from Erik, the OCMock guy:
[14:38] <mmcc> he said: "let me state first that I am not a lawyer. That said, my intent and interpretation of this part of the license is that the attribution is required for the actual use of OCMock. I don't think your file sync client uses OCMock. I assume OCMock is only used during the development of the file sync client. Again, I am not a lawyer, but to me this sounds like attribution would only be required when you talk about the developmen
[14:38] <mmcc> your software."
[14:38] <mandel> gatox, so which revno is giving you that error?
[14:39] <gatox_mac> mandel, 25
[14:39] <mmcc> also ralsina, see my ircspam above ^
[14:39] <gatox_mac> is that ok?
[14:39] <mandel> gatox_mac, yes, can you open xcode, add edit the scheme
[14:39] <gatox_mac> mandel, yes…
[14:39] <mandel> gatox_mac, in the name next to the stop, there is a context menu
[14:39] <ralsina> mmcc: the number of people who is not a lawyer and writes licenses never ceases to amaze
[14:40] <gatox_mac> mandel, yes
[14:40] <ralsina> mmcc: short answer, I don't know
[14:40] <mmcc> so it sounds like he is OK with how we want to use it, but I'm not sure if we're exactly on the same page
[14:40] <gatox_mac> mandel, i'm in edit scheme
[14:40] <ralsina> mmcc: longer answer, let's move on with it
[14:40] <mandel> gatox_mac, there in the test tab can you open it?
[14:40] <ralsina> mmcc: you mentioned chrome used it? If it did and is gpl, then we are in good shape
[14:41] <mmcc> ralsina: move on with using OCMock?
[14:41]  * ralsina assumes google has lawyers
[14:41] <ralsina> mmcc: yes
[14:41] <dobey> haha
[14:41] <gatox_mac> mandel, yes… already there
[14:41] <dobey> chrome or chromium?
[14:41] <mandel> gatox_mac, in the panel of the right, is there a bundle called FsEventsTests?
[14:41] <mmcc> ralsina: chrome does use it but I need to check their license. and the commit comment thread I found discussing it had a lot of "uh, I'm not a lawyer but this seems fine"
[14:42] <dobey> chrome is not GPL
[14:42] <mandel> gatox_mac, or a screenshot please :)
[14:42] <gatox_mac> mandel, yes
[14:42] <gatox_mac> mandel, ok… screenshot
[14:42] <dobey> and people probably misread the license
[14:42] <mmcc> chromium:  http://code.google.com/p/chromium/issues/detail?id=28291
[14:42] <dobey> does chrome include a statement advertising that it's using OCMock created by whoever that guy is that created it?
[14:43] <gatox_mac> mandel, do you know what would be awesome?? if u1 was working in mac, it would be easier to share the screenshots… jejejejejee recursion!
[14:43] <ralsina> chrome is BSD
[14:44] <dobey> chrome is proprietary
[14:44] <ralsina> chromium
[14:44] <dobey> chromium is a lot of mess
[14:44] <dobey> it's not all bsd
[14:44] <ralsina> damn, chromium licensing is a collage
[14:44] <dobey> yes
[14:44] <mandel> gatox_mac, while you upload the image, can you branch the code in a completely diff location and try to run the tests
[14:44] <ralsina> it has MPL/GPL/LGPL/MIT/Ms-PL/BSD licenses
[14:45] <gatox_mac> mandel, http://ubuntuone.com/7MCq3uU9iVvdygM89LzWYj
[14:45] <mandel> gatox_mac, maybe you have a little mess in the xcode files
[14:45] <dobey> it is a royal clusterfrack
[14:45] <gatox_mac> mandel, ack
[14:45] <ralsina> mmcc: there is no alternative to ocmock?
[14:45] <ralsina> mmcc: I would rather spend a day looking for one than a week removing it
[14:45] <mandel> gatox, funny, because it seems ok, so if you do command+u it should run the tests inside xcode..
[14:45] <mandel> gatox_mac, ^
[14:45] <dobey> i would avoid ocmock unless the license gets fixed
[14:46] <gatox_mac> mandel, checking
[14:46] <mmcc> ralsina: I'm not aware of anything that nice, but we can keep looking. We could also probably add some patch-style features to something else if we needed to
[14:46] <ralsina> mmcc: in parallel, we can ask our lawyer
[14:47] <ralsina> mmcc: but I don't expect that to be useful because "use" is so fuzzy here
[14:47] <mmcc> ralsina: ok. I get the impression that this guy would be willing to let us use it how we want, I feel like we just talked past each other this first time around
[14:48] <gatox_mac> mandel, finally!!!!!!! tests succeeded!!!!!
[14:48] <gatox_mac> 47 tests executed
[14:48] <gatox_mac> the Xcode files were dirty!
[14:48] <mmcc> I'll forward you the email I sent - I should've CCd you first
[14:48] <gatox_mac> dirty files
[14:48] <dobey> mmcc: what was the conversation? did you ask him to remove the advertising clause and make it plain MIT/X licensed?
[14:49] <mandel> gatox_mac, ok, so a clean branch works, right?
[14:49] <gatox_mac> mandel, yap!
[14:49] <mandel> gatox_mac, cool :)
[14:49] <gatox_mac> mandel, so.. if you fix the typos. it's a +1 from me
[14:49] <mandel> gatox_mac, ok, on it right now before I forget!
[14:49] <mmcc> dobey: check your email
[14:49] <gatox_mac> mandel, cool! let me know
[14:50] <gatox_mac> mandel, btw… great job!! :D
[14:50] <mmcc> dobey, ralsina I didn't ask him to change the license, I asked him for a clarification
[14:50] <mandel> alecu, there was an issue with the xcode config files, please when testing add-run-tests do a clean branch and not a pull
[14:50] <mandel> gatox_mac, meh, just luck :)
[14:50] <gatox> mandel, i mean about the branch :P
[14:50] <alecu> mandel, ack
[14:51] <mandel> gatox, oh, hehe
[14:51] <gatox> mandel, and your objective-c lessons
[14:51] <gatox> jejej
[14:51]  * mmcc believes XCode developers do not share .xcodeproj files
[14:51] <mmcc> I would bet money that XCode itself is built using makefiles and awk
[14:52] <gatox> mmcc, they are really selfish! jeje
[14:52] <alecu> mandel, all tests pass that way
[14:52] <mandel> mmcc, I'll move to make later, no a priority atm
[14:52] <mandel> alecu, superb!
[14:52] <mmcc> I suspect they also build all their UI in code, and don't use Interface Builder. Or else only one developer ever touches those things...
[14:52]  * dobey "fixes" the pylilnt problem
[14:53] <alecu> mandel, I tried merging with trunk again, and I did not get the segfault either :P
[14:53] <mmcc> mandel: no hurry on that :) I like to complain, but it does sort of work...
[14:53] <mandel> mmcc, is good to have someone complaining with good reasons
[14:54] <mandel> mmcc, I just did not know that schemes are not share by default on xcode, what a stupid feature
[14:54]  * dobey really needs to do expenses
[14:55] <mmcc> mandel: I didn't know that either. XCode has changed a lot in the last version or so.
[14:55] <dobey> before that deadline passes in a few days
[15:00] <gatox> me!
[15:00] <gatox> alecu, thisfred briancurtin mmcc ralsina dobey standup?
[15:00] <briancurtin> me
[15:01] <ralsina> oh right
[15:01] <ralsina> me
[15:01] <thisfred> me
[15:01] <dobey> meh
[15:02] <mmcc> me
[15:02] <dobey> gatox: don't forget mandel
[15:02] <gatox> oh!
[15:02] <gatox> mandel, sorry
[15:03] <gatox> mandel, alecu ?
[15:03] <mandel> me
[15:03] <ralsina> alecu is last, go gatox!
[15:04] <gatox> DONE:
[15:04] <gatox> Review an objective-c branch and fight with some xcode complications when running the tests. Fixed Bug #998079.
[15:04] <gatox> TODO:
[15:04] <gatox> Bug #1006329 to close my unicode issues. Add my findings about macfsevents in the wiki, go back to fsevents and start looking how to add a wait condition to the watcher when a new folder is added so we don't lose events. 1-1 with ralsina.
[15:04] <gatox> BLOCKED:
[15:04] <gatox> No
[15:04] <gatox> briancurtin, go
[15:04] <briancurtin> DONE: lazr removal (one more pending)
[15:04] <briancurtin> TODO: get testability built and included in the automated builds
[15:04] <briancurtin> BLOCKED: nope
[15:04] <briancurtin> NEXT: ralsina
[15:04] <ralsina> DONE: mgmt call, some 1-1s, reviews, reading about dash development, trying to understand the mac universe and failing at it TODO: understand what I did not, more reviews BLOCKED: no, NEXT, thisfred
[15:04] <thisfred> DONE: bug #1004673 TODO: bug #1004673 , estimate the rest of the work for this phase of the u1db client BLOCKED: no NEXT: dobeyh
[15:04] <dobey> λ DONE: set up milestones, fix to tarmac config
[15:04] <dobey> λ TODO: expenses, unblock tarmac, triage
[15:04] <dobey> λ BLCK: none.
[15:04] <dobey> mmc
[15:04] <mmcc> DONE: setup-mac.py, objc test frameworks
[15:04] <mmcc> TODO: more -^--------^
[15:04] <mmcc> BLCK: NO
[15:04] <mmcc> NEXT: mandel
[15:04] <dobey> c
[15:04] <mmcc> dobey: si
[15:04] <mandel> DONE: Worked in adding tests to the FSEvent class (I little stuck there).
[15:04] <mandel> TODO: More tests, add domain sockets server tests.
[15:04] <mandel> BLOCKED: no
[15:04] <mandel> alecu, please
[15:04] <alecu> DONE: learned about objective c blocks; reviews; following a SD bug that rye found
[15:04] <alecu> TODO: more deep SD debugging
[15:04] <alecu> BLOCKED: no
[15:04] <thisfred> fg
[15:05] <mandel> gatox, fixed (I removed the comment is was idiotic) and pushed
[15:05] <gatox> alecu, question: did you fix the other unicode issue i had from the other day? because i don't see it in my queue
[15:05] <gatox> mandel, ack!
[15:05] <mmcc> ralsina: which parts of the mac universe?
[15:06] <ralsina> mmcc: where things go, how things install, basic stuff
[15:06] <ralsina> mmcc: I have installed 6 or 7 things and some evenhave real installers after you click the dmg
[15:07] <ralsina> mmcc: which I thought was a nono, but then again, they are crap like silverlight and flash
[15:07] <mmcc> ralsina: yeah, it's a bit of a mess. there's best practices, then there's other stuff
[15:07] <ralsina> mmcc: haha, yeah, like everywhere else :-)
[15:07] <mandel> ralsina, anything done by flash is usually wrong in any platform
[15:08] <ralsina> OTOH, I am starting to understand xcode somewhat
[15:08] <mmcc> ralsina: yep. except the App Store enforces strict UX for installers (ie, no installers, period)
[15:08] <ralsina> mmcc: should be a short guideline
[15:08] <mmcc> ralsina: what should?
[15:08] <ralsina> mmcc: the how to do installers guideline. "STOP"
[15:09] <dobey> "The road to hell is paved with good intentions."
[15:09] <mmcc> ralsina: heh, yeah. it's about that short.
[15:09] <ralsina> anyway, everything is just a bit foreign so I am forcing myself to use it half a day until I understand it somewhat
[15:09] <ralsina> and it's being unpleasant, but it's mostly my fault for never using a mac before, I guess
[15:10] <mmcc> ralsina: don't blame the user
[15:10] <dobey> mmcc: oh, you also e-mailed the wrong guy i guess?
[15:10] <mmcc> dobey: huh?
[15:10] <ralsina> mmcc: I am not a power user, I am a problem user
[15:11] <dobey> mmcc: how is "Erik Eoernenburg" related to "Mulle Kybernetik" ?
[15:11] <mmcc> dobey: he's erikdoe on github, it's his project
[15:12] <mmcc> his face is the rightmost picture on http://www.mulle-kybernetik.com/en/index.html
[15:12] <dobey> yeah i see that now
[15:12] <dobey> but what exactly is mulle kybernetik
[15:15] <mmcc> dobey: I'm not really sure about that. I think it's sort of a side-project software company that those three guys work on, or at least identify with
[15:15] <mandel> dobey, is a cybernetic mule in a funny lang?
[15:16] <ralsina> Kybernetyk is cybernetic in german
[15:16] <mmcc> they do have an about page... but it's not super enlightening
[15:16] <mmcc> http://www.mulle-kybernetik.com/en/about.html
[15:16] <ralsina> and mulle is "me" in estonian ;-)
[15:16] <dobey> mmcc: i'm not sure it's actually a company. i think it's just a "hey we're some guys that like some things"
[15:17] <dobey> mandel: given the logo, i suspect mulle is mole, not mule :P
[15:17] <mmcc> dobey: yeah, I guess it's mainly just a website that they all contribute to, only mostly it's 'Nat!', who I can't find a real name for
[15:18] <mandel> dobey, do cybernetic moles dream about electric sheep?
[15:18] <dobey> i think they dream about electrict dirt
[15:19] <dobey> mmcc: basically, that reply is not comforting at all. it basically says "i have no idea what my own license means exactly"
[15:19] <dobey> and "oh, also, i am not a lawyer"
[15:21] <mmcc> dobey: agreed. I couldn't figure out why he chose that license, if he's not really clear on what it means. I mean, he wrote the code, so ...?
[15:21] <dobey> mmcc: and presumably wrote the license
[15:22] <dobey> it's not even a good advertising clause
[15:22] <mmcc> dobey: exactly, since I've never seen that particular license anywhere else. Maybe one of the other guys wrote the license for M-K, and he just used it?
[15:23] <dobey> certainly possible
[15:23] <dobey> i could probably get them to change it
[15:24]  * mandel imagines dobey chopping off a horse head and sending it via cheap post delivery to germany 
[15:25] <dobey> heh
[15:32] <dobey> ok everyone
[15:32] <dobey> trunk tarmac should be usable again now.
[15:38] <mmcc> ralsina dobey: so the 3-clause BSD is also a no-no for us? I'm looking at Kiwi again, which does let you stub parts of existing objects, basically the same as twisted's patch.
[15:39] <dobey> mmcc: right, bsd + advertising is not gpl compat either
[15:40] <mmcc> urgh.
[15:40] <gatox> ralsina, mandel can i have a really small review from you? https://code.launchpad.net/~diegosarmentero/ubuntuone-client/publish-file/+merge/107996
[15:42] <elopio> briancurtin: sorry. testability.dll comes with the qtsdk. I found it at c:/QtSDK/Desktop/Qt/4.8.1/mingw/plugins/testability
[15:42] <briancurtin> elopio: cool, I'll see about getting it in
[15:42] <elopio> briancurtin: and I didn't install pyqt.
[15:42] <ralsina> gatox: sure
[15:43] <gatox> thx
[15:43] <ralsina> elopio: if the SDK is built using mingw, that's yet another problem
[15:43] <ralsina> elopio: because we are using a set of DLLs built using VS and it's not compatible
[15:43] <elopio> ralsina: there's an sdk for vs.
[15:44] <ralsina> briancurtin: here's some info about Qt plugins and py2exe http://www.py2exe.org/index.cgi/Py2exeAndPyQt
[15:44] <ralsina> elopio: ok, we can try to get it from there and just copy it insto the installation folder, really
[15:44] <ralsina> elopio: requires no code changes
[15:44] <elopio> I can try with it instead.
[15:44] <elopio> ralsina: that sounds cool.
[15:46] <dobey> yay
[15:47] <dobey> briancurtin: ./ubuntu_sso/networkstate/tests/test_linux.py:58:42: E271 multiple spaces after keyword
[15:47] <briancurtin> dobey: where is that from?
[15:47] <dobey> briancurtin: in your remove-lazr-from-setup branch
[15:48] <dobey> briancurtin: could be a new complaint from pep8 on quantal, as it has version 1.1 while precise has 0.6.1
[15:49] <mmcc> well, http://oops.se/objcunit looks promising, but it's 10 years since the last release, and the link is broken. I sent an email... :\
[15:53] <ralsina> gatox: +1
[15:53] <gatox> ralsina, cool! thanks
[15:53] <gatox> mmcc, could you take a look at this: https://code.launchpad.net/~diegosarmentero/ubuntuone-client/publish-file/+merge/107996 ?
[15:53] <mmcc> gatox: sure
[15:54] <gatox> mmcc, thx, it's a really small branch
[15:58] <dobey> ok, need to get lunch. bbiab
[15:59] <gatox> alecu, ping
[16:01] <gatox> ok..... lunch for me too........ brb
[16:04] <alecu> gatox_lunch, pong
[16:04] <alecu> ok, I'll get lunch too
[16:07] <mandel> and EOD for me, laters!
[16:45] <gatox_lunch> alecu, ping..... for when your lunch end :P
[16:46] <alecu> gatox_lunch, my lunch got delayed...
[16:46] <alecu> gatox_lunch, so, pong
[16:47] <gatox> alecu, quick question..... do you know if get_abspath in filesystem_manager (u1-client) should handle (returning) only bytes or unicode?
[16:47] <alecu> gatox, no, I don't know.
[16:47] <gatox> alecu, ok then! :P just wanted to know if it was one of those cases
[16:48] <mmcc> gatox, your branch looks fine to me. I had some test failures that seem unrelated - dirty reactors in test_perspective_broker. I think my branches might be out of sync.
[16:49] <gatox> mmcc, in windows/linux/mac?
[16:49] <mmcc> gatox: linux
[16:49] <briancurtin> googling failed me. anyone know of a way to branch just a subdirectory of a repository? i really just want scripts/devsetup out of ubuntuone-windows-installer
[16:50] <gatox> mmcc, weird..... maybe you have sometihng out of date
[16:50] <briancurtin> oh, hmm, maybe bzr export will do it. checking
[16:50] <gatox> mmcc, are you running the tests with: make check?
[16:50] <mmcc> gatox: yes
[16:51] <gatox> mmcc, and do you have clientdefs.py and logging.conf in the proper places?
[16:52] <gatox> maybe u1-dev-tools is out of date
[16:57] <mmcc> gatox: I do have those files, yes.
[16:58] <gatox> mmcc, sudo apt-get upgrade?? to update u1-dev-tools maybe?
[17:00] <mmcc> gatox: I ran 'sudo apt-get install ubuntu-sso-client' before running the tests, since my installe version of the sso client didn't have mandel's recent changes (missing get_activation_cmdline)
[17:01] <gatox> mmcc, sudo apt-get install upgrade..... will upgrade all the system..... also, this might be failing in your machine because you don't have ubuntuone-dev-tools up to date
[17:01] <dobey> hmm
[17:01] <mmcc> so I have recent nightlies, and 'apt-get build-dep ubuntuone-client' claims I have no updates to add
[17:01] <mmcc> brb
[17:01] <dobey> what is failing?
[17:01] <gatox> dobey, mmcc is having some dirty reactor failures trying to run the tests from my branch
[17:01] <dobey> briancurtin: i don't think you can branch/checkout only a subdir in bzr
[17:02] <gatox> i think that he might  not have the last version of u1-dev-tools
[17:02] <dobey> mmcc: what version of ubuntuone-dev-tools do you have?
[17:03] <mmcc> dobey: how do i check?
[17:04] <dobey> mmcc: dpkg -l ubuntuone-dev-tools
[17:04] <mmcc> 3.1+r71-19~precise1
[17:05] <dobey> you don't have the latest version
[17:05] <dobey> build-dep doesn't upgrade already installed packages
[17:05] <dobey> does apt-get upgrade not want to upgrade stuff?
[17:05] <briancurtin> dobey: yeah what i had to do was branch and then "bzr export dest my/branch/subdir" to get subdir as dest (i wanted C:\u1 to be ubuntuone-windows-installer/scripts/devsetup as the home of my dev folder)
[17:06] <mmcc> dobey - it wants to upgrade a lot of stuff :) I will get in the habit of upgrading everything before tests, I guess
[17:06] <dobey> :)
[17:07] <mmcc> thanks dobey, gatox. re-running tests over lunch, I'll approve that branch when I get back
[17:07] <dobey> briancurtin: you can do "bzr export -d lp:project dest subdir" to avoid pulling a local branch first
[17:08] <gatox> mmcc, ack
[17:09]  * mmcc lunches
[17:17] <ralsina> dobey: any strng feelings about not generating clientdefs.py.in with autotools?
[17:18] <ralsina> dobey: alternatively, I can do a script that does the same thing autotools does with it, so we can generate it on windows
[17:19] <gatox> fraking unicode bug!
[17:24] <dobey> ralsina: i have plans to fix that
[17:25] <ralsina> dobey: ok, trust you with it then.
[17:37] <gatox> ralsina, oops....... 1-1
[17:37] <ralsina> gatox: vamos
[17:46] <dobey> hrmm
[17:46] <dobey> i guess we need to copy the relevant pylint bits and such, to the stable/beta PPAs as well
[17:50] <briancurtin> quick MP if anyone has a minute, just bumps pep8 version in buildout - https://code.launchpad.net/~brian.curtin/ubuntuone-windows-installer/pep8-version-bump/+merge/108024
[17:50] <rye> mandel: ralsina anybody knows how to build some script to archive the ubuntuone logs on windows? bat/cscript/powershell ?
[17:51] <ralsina> rye: if you do it with python we can make it an exe
[17:52] <briancurtin> rye: you mean to have a client run it to grab what you need?
[17:52] <ralsina> rye: as a plus, python has its own zip implementation, so you don't require crap
[17:54] <briancurtin> my last company had a similar script (little C# app) that you just double click and it zips the right directory and drops it on the desktop (easiest way)
[18:05]  * dobey wonders what to do about this SRU stuff
[18:10] <gatox> ok..... this is not reproducible
[18:15] <rye> briancurtin: that's what I am after too.
[18:16] <briancurtin> rye: i can write you up a pretty simple python script that'd do this, but having to ship a zip file of everything py2exe/python needs to run would be a bit heavy. could just write a little C# app for you that'll do it and just distribute the EXE
[18:16] <ralsina> dobey: I am worried about it. The requirements are hard to meet without just making stuff up.
[18:17] <ralsina> briancurtin: if you do a python script we can ship it with the software
[18:17] <ralsina> although of course we can ship exes too
[18:17] <rye> briancurtin: yep. I was thinking about employing other python compiled thing to be later included into u1 and which compresses everything in a little tiny exe
[18:17] <dobey> ralsina: as it turns out, making shit up is my forte
[18:17] <rye> if these exist these days
[18:17] <briancurtin> ralsina, rye: true. something like ubuntuone-diagnostic.exe
[18:19] <ralsina> dobey: haha
[18:38]  * dobey mails the tech board
[18:39] <briancurtin> elopio: how did you get this testability stuff to work in the first place? its not included in the SDK i downloaded, and i cant compile it because the repository is empty...so im not really sure where to go from here
[18:41] <briancurtin> i'll be back in a bit. going to spend my lunch time waiting in line to send a package. fun times.
[18:42] <gatox> briancurtin, been there
[18:46] <elopio> briancurtin: I followed this
[18:46] <elopio> http://projects.developer.nokia.com/Testabilitydriver/wiki/WindowsHowToStart
[18:47] <elopio> and then found the missing parts here:
[18:47] <elopio> http://projects.developer.nokia.com/Testabilitydriver/wiki/WindowsInstallation
[18:47] <elopio> it was not funny, btw.
[18:54]  * alecu is wasted, and will take the evening off.
[18:54] <dobey> whoot
[18:54] <alecu> please send me an sms or a gtalk to reach me.
[18:56] <dobey> ralsina: well, just mailed tech board to request micro release exception for u1 :)
[18:59] <ralsina> dobey: awesome
[19:16] <dobey> i hope my new t-shirt arrives this week
[19:16]  * dobey wonders how long it takes for a t-shirt to get here, from Paris
[19:16] <ralsina> 10 days
[19:17]  * ralsina is practicing estimates
[19:17] <dobey> heh
[19:17] <dobey> postal planning poker
[19:18] <gatox> alecu, not now..... but when you have some time, you can take a look at the wiki with the macfsevents section updated with the results i get
[19:19] <ralsina> gatox: alecu is out for the day
[19:19] <dobey> hrmm, maybe i should make a new ppa
[19:19] <gatox> ralsina, yes, i know..... just leaving the log here for him to see it later :P
[19:19] <dobey> which has only things which we don't maintain ourselves, but which we need to run tests and such
[19:20] <dobey> for the tarmac instances to use
[19:20] <gatox> ok..... going back to mac fsevents now!
[19:20]  * dobey wonders what all changes are needed exactly, for us to be able to switch to pyflakes for everything
[19:22] <gatox> dobey, one of the problems with pyflakes is that you don't have a way to ignore part of the code as we do with pylint
[19:22] <dobey> gatox: i don't think that's a problem really
[19:23] <dobey> gatox: at least, it's not a problem in all the places where we currently use pyflakes :)
[19:24] <gatox> dobey, no..... but in some situations we need to write (for u1-client) not really pretty code, just to force pyflakes to pass..... don't get me wrong..... i prefer pyflakes.... i forked it and added some functionality to pyflakes and fixed some stuff for personal use.... i should propose that
[19:25] <dobey> gatox: there is only one situation where that is the case, and it's because of a bug in pylint which needs to be fixed. the solution for that isn't to add # ignore comments everywhere
[19:25] <dobey> also, i think it was said that flake8 has a way to do inline ignores
[19:28] <gatox> that would be nice....... i REALLY prefer pyflakes... just wasn't sure about those weird cases
[19:29] <dobey> gatox: the only weird case i know aobut is the "try: import foo except ImportError: import bar as foo" case where it complains about overwriting an undefined variable
[19:29] <dobey> or undeclared, whatever
[19:30] <gatox> dobey, yep..... that and one you import something that you don't use, but in some cases we do that, because we want that inside that namespace
[19:30] <gatox> those are the only thing that come to my mind
[19:31] <gatox> dobey, but yes..... in favor to change to pyflakes jejeje i wasn't arguing, just saying that there are some things that we have to keep in mind
[19:31] <dobey> gatox: that latter case we're probably doing wrong anyway, if pyflakes complains about it. the right solution for that case i think, is to use __all__, which we have generally failed at doing
[19:31] <gatox> yap
[19:31] <dobey> gatox: yes. i just want an exhaustive list, so we can get them all fixed, and switch
[19:32] <dobey> also, pyflakes supports python3 i think. while pylint is insanely complex and tedious
[19:33] <gatox> dobey, pyflakes support python3?? i would need to check that.... because i was having some problems between pyflakes and python3 code
[19:33] <dobey> i think so, but could be wrong
[19:34] <dobey> ah no, i am wrong
[19:35] <dobey> but it shouldn't be hard to port it to py3
[19:38] <gatox> dobey, no..... i'm starting to do that btw....
[19:38]  * mmcc was wondering why we weren't using __all__
[19:38] <gatox> not now..... but it's in my after-work tasks :P
[19:39] <dobey> heh
[19:43] <mmcc> seriously, is there a tradeoff wrt. using __all__ that I don't know about? I was going to ask back when we were doing that platform refactoring but forgot to
[19:44] <dobey> mmcc: well if __all__ breaks imports for things that aren't in __all__, then __all__ makes it hard to import those other things for tests or internal usage, no? :)
[19:46] <mmcc> dobey: tests! that makes sense.
[19:47] <dobey> or maybe i just don't fully understand how it works
[19:50] <ralsina> dobey: tests should not import the things that the module doesn't export, because then you are testing the implementation details
[19:50] <ralsina> dobey: but that's mostly a theoretical argument I am not going to argue very hard ;-)
[19:50] <ralsina> plus it's trivial to disable __all__ for testing
[19:51] <mmcc> hrm, it seems __all__ only affects "from foo import *" ?
[19:51] <dobey> yeah, i don't think __all__ works quite right anyway
[19:51] <ralsina> mmcc: heh, you are right. So it only matters when you do it wrong.
[19:52] <mmcc> and you can manipulate __path__ to point to platform-specific implementations
[19:52] <mmcc> that is, in __init__.py, you can manipulate __path__ to point to a platform-specific implementation in a submodule
[19:53] <mmcc> this is from an old doc, that the 2.7 docs say is still mostly up to date, but I don't know how py3 changes this, esp. wrt pep 420 (right number? the no-__init__ implicit module thing?)
[19:53] <mmcc> http://www.python.org/doc/essays/packages.html
[19:55] <dobey> i just can't take pep 420 seriously
[19:55] <mmcc> that might be a nice way to get rid of all those "import darwin as source, func1 = source.func1, func2 = source.func2, etc etc" things
[19:55] <mmcc> dobey: because they should've reserved the number for an april fools joke pep?
[19:56] <dobey> or something
[19:56] <dobey> it just has the wrong number to be a serious pep
[19:57] <mmcc>  not that anything called "pep" sounds that serious to begin with
[19:57] <dobey> true
[19:58] <dobey> and on that particular example, it doesn't help that i've been rewatching the complete series of Daria
[20:00] <dobey> though all this SRU stuff is putting me in the mood to watch Brazil
[20:01] <gatox> saying: pep8 in spanish is really funny :P
[20:02] <dobey> heh
[20:02] <dobey> pep ocho?
[20:02] <gatox> yes..... but it's sounds like "precioso" but talking weird
[20:03] <gatox> like a baby or sometihng like that
[20:03] <dobey> gatox: tomar el churro la boca.
[20:04] <gatox> hehehehehehe
[20:07] <gatox> ok..... eod for me people!! enjoy and see you tomorrow!
[20:07] <mmcc> bye gatox
[20:07] <ralsina> bye gatox
[20:07] <gatox> bye!
[20:29] <ralsina> EOD for me, will do reviews & stuff at night, so mail me if anyone needs anything
[20:29] <ralsina> bye!
[20:30] <mmcc> bye ralsina
[21:52] <dobey> later all
[22:44] <mmcc> ok, I'm heading out. feels like I'm getting closer to a working app with py2app now. up next: some issues with not seeing the Qt image plugins at packaging time
[22:45]  * mmcc leaves