[02:29] hello, i went into the web U1 interface and selected a folder to "stop syncing". while i do understand the message that it does *not* delete my local folder, do i understand correctly that it deleted the server copy? === m_conley is now known as m_conley_away [08:32] morning! [08:38] achiang: yes, you are right, the server-side copy will be removed and UDF will no longer be subscribed on the client [08:49] Good morning, all! *8O) [11:13] good morning [11:40] gatox, morning! [11:40] mandel, hi [11:43] gatox, I finally managed to fix the tests for domain sockets! [11:43] gatox, I need to ensure that I did not brake any on your windows vm ;) [11:44] mandel, great...... i have the run-tests script working on mac, with a script to setup the env... only need to fix u1lint and i'll ready to propose the branches [11:44] gatox, superb! we start moving, great news \o/ [11:45] mandel, give your branches i'll test them [11:45] gatox, we have to be very very careful windows tests are kept like the are, linux is not a problem we have tarmac, but we are the windows tarmac ;) === yofel_ is now known as yofel [12:23] EHLO one.ubuntu.com [12:25] alecu, buenas [12:25] alecu, hello! [12:25] alecu, logging ssl branch updating with the message you proposed [12:25] alecu, and I'm off to have lunch :) [12:48] good morning! [12:49] ralsina, good morning [13:02] mandel: hi, about the SSL-moar-info-please thing, poking you :) [13:06] hmm [13:09] hola ralsina. Hola dobey. [13:09] ralsina: is the installer briancurtin emailed to us yesterday the one that we should have some users try that are impacted by the 3.0 auth issues? [13:10] rye, mandel is off for lunch. Got any more info on the issue? [13:10] joshuahoover: exactly [13:10] joshuahoover, yes [13:10] joshuahoover: should give us the actual SSL error [13:10] rye: brian uploaded an installer yesterday with mandel's branch in it, I can give you the link [13:10] ralsina: very good, i'll see who we can get to test that out for us [13:11] rye: http://u1.to/~brian.curtin/g/3.0.0-windows-installer [13:12] hola alecu [13:31] ralsina: oh, that's what is it, ok [13:40] * mandel back! [13:44] alecu, so in https://code.launchpad.net/~mandel/ubuntu-sso-client/fix-ssl/+merge/103147 I would add the logger, I know is not used, but ensures that the weclient api is consistent over all implementations [13:44] alecu, if we remove the skip we will have failures such because of the lack of logger [14:06] ralsina: lucio and I have been discussing whether or not to ask for a bit of design help on the u1todo example app, to make it more compelling. We decided to ask for your opinion in the matter... :) [14:06] thisfred: please do, I have seen how that looks and it needs it ;-) [14:07] thisfred: although it's an example app so the theming code may be distracting [14:07] thisfred: but just a qss can do wonders for it [14:07] thisfred: and layout [14:08] ralsina: yeah I want to keep the layout code to a minimum, but since the ui is already separate, we should be able to spruce it up a little [14:08] ralsina: ok, so do I ask lisettte directly? (hi! ;) or do I need to go through channels, and if so what are they? [14:09] let's just ask lisette, but it all depends on how much work she has on her plate [14:10] thisfred: we could also just STEAL [14:10] It's not super urgent. I'm trying to keep the code and ui as independent as possible, so adding it later should still be easy enough [14:10] thisfred: yes, the major functions are already in place [14:10] hi ralsina, thisfred, what's up? [14:10] ralsina: from what? u1cp? [14:10] thisfred: and minimalistic is nice [14:10] thisfred: from the world. Find a todo app people like :-) [14:10] lisettte: hi, I'm developing a little example app on top of u1db [14:11] lisettte: mostly to show developers how to do stuff [14:11] thisfred: sounds cool [14:12] ugh. not feeling the greatest today [14:12] lisettte: we'd like it to look a little less ugly than I made it so far, but it doesn't have to be on brand or anything, since it's just example code. [14:13] thisfred: aha [14:17] ok, ubuntuone-scripts are unmanageable now. Starting to manage [14:18] dobey: you *are* the greatest. There, better? [14:19] no :) [14:20] hey, I tried! [14:20] but i did just troll Neil deGrasse Tyson again, so it's a start. [14:20] dobey, ping... question: in mac we are using the buildout to configure the development environment, the thing is, im trying to keep the things as independent as possible, but u1lint is starting pylint as a process and this process is tryinng to use the python from the system, not the one from the buildout... it is reasonable to modify u1lint to start the python from the buildout instead of the one from the system for MAC?? (just wanted to kn [14:20] ow if i can do that... to avoid adding a lot of folders from the buildout to the pythonpath of the system) [14:20] you mentioned that on "Iron Sky" the moon nazis are not astronomically correct? [14:21] gatox: why not just manipulate PATH? [14:21] ralsina: no. much much better than that: https://twitter.com/#!/dohbee/status/194792155791892480 [14:22] he [14:22] ralsina, path? i'm addinng a python_u1... to avoid crashing and don't affect the system just to run tests, etc [14:22] gatox: i have no idea how the buildout works. why does pylint not "just work" ? [14:22] gatox: let's talk about this on mumble in 10'. You are probably making it too hard for yourself :-) [14:23] gatox: and how do we solve this for other places where we do subprocess.Popen()? [14:23] gatox: for example: u1lint probably has system python on its shebang, and shouldn't [14:23] dobey, inn other places we have all the dependencies installed in the system...... that is not what the buiildout do [14:23] gatox: or pylint does [14:23] ralsina, yes, i know [14:24] ralsina: conversly, i recently saw a report about us using "env python" on shebang, and saying we shouldn't [14:24] ralsina, ok...... let me know when you can have a quick mumble, and i'll tell you what is going on [14:24] gatox: in both cases, that's because they were installed using the system python, since setup.py is supposed to put the "right" python there [14:24] dobey: yes env python is bad for things that are installed system wide [14:24] gatox: subprocesses don't really start correctly in the buildout - i have the same problem in Windows that i've temporarily fixed by finding the "bin\python-script.py" and inserting it as the first argument to sys.executable [14:25] hmm [14:25] briancurtin, i add the bin folder from the buildout to the PATH...... but i rename the python in there to python_u1, and use that to run tests, etc [14:25] so if you want to run Popen([sys.executable, "yourscript.py", "your arg"]) it really has to be [sys.executable, "path\to\bin\python-script.py", "yourscript.py"...] [14:26] so, let's call the custom python "python" and add that to the PATH [14:26] why would that not work? [14:26] except of course for things that have the python hardcoded [14:27] briancurtin: hrmm, nobody should ever be passing sys.executable to Popen() [14:27] dobey: any reason? [14:27] lisettte: to give you an idea of how much we need your help, this is what it looks like now: http://ubuntuone.com/1aHSGoXtfWsHs91TmR6CKt Note that this is not super urgent, so it needs to be planned in at some point, and again, the aim is not to make it look as good as possible, but more like someone has actually given any thought to it. [14:27] ralsina, well... in this case doing that i'll have the same problem, because pylint has that..... [14:28] gatox: unless we install pylint using that python [14:28] also: one feature that isn't shown yet is "ta [14:28] or we also do a virtualenv [14:28] briancurtin: well, it's python, so any random piece of code you import could change it out form under you. and there should never be any need to use it. if you find yourself using it, you're grasping straws to work around a larger problem, it would seem [14:28] "tags", which have to be attachable to any item, and searchable/filterable [14:29] ralsina, actually now the ONLY problem remaining is for pylint... if it is possible to modify u1lint to add 1 line...... problem solved [14:29] dobey: so how would you start a script using the same python interpreter you used? i get that someone could change it, but we're running tests with it (the common usage of Popen([sys.executable...])), so...don't change it [14:30] thisfred: cool [14:31] briancurtin: why should you start a script with the same interpreter? [14:31] dobey: because we don't want to use the broken system python [14:32] why wouldn't you? i dont get it. problem: i want to run another script - solution: run another script [14:32] yes, why do you *care* what interpreter it runs with, or that it's python? [14:33] dobey: because pylint *loads* the module it's parsing. If the module uses things that are not installed on the python pylint is using, it breaks horribly. [14:35] how is that a problem with Popen() though? [14:38] dobey: it's a problem with pylint using the wrong interpreter. If when calling popen you call "therightpython pylint" it works [14:38] why is pylint using the wrong interpreter? PATH is wrong? [14:39] dobey, pylint is using the interpreter from the system, and the buildout generates his own python with all the libs [14:40] we want to use the one from the buildout [14:40] but *WHY* is it using the system one? [14:40] but without breaking the one from the system [14:40] dobey: because pylint's first line says "#!/usr/bin/python" [14:40] so do all of our scripts [14:41] dobey: not i they are installed via setup.py with the python you want to use. The shebang is corrected on install. [14:41] Which is why I believe that pylint is not installed correctly [14:42] ralsina, ( briancurtin correct if i'm wrong ) we don't have pylint installed, we have the eggs inside the eggs folder from the buildout and we are taking them from there [14:43] gatox: correct === Guest70849 is now known as zumbi [14:44] ralsina, so that's why we need to call it in this case like: python_u1 pylint *args [14:44] Dr_Who: thanks for all the input on the Files app. I'll be incorporating a lot of that over the next few days. === zumbi is now known as Guest86751 [14:44] gatox: if pylint is not installed, where did that pylint came from/ [14:44] ? [14:45] ralsina, fromm the buildout, the buildout download a bunch of stuff...... but don't install those libs, instead creeate a custom python that use them, so your system remains clean from the development environment [14:45] gatox, FYI in order to ensure that we do not brake lots of things with the tcpactivation I first have to make some changes in devtool [14:45] gatox: and it installs a pylint script with the wrong shebang? [14:45] mandel, ack [14:46] you're very welcome urbanape :-) [14:46] briancurtin, can you explain that? ^^ [14:46] urbanape, will certainly do some more [14:46] ralsina: maybe it got built with system python? [14:47] dobey: that's my guess [14:47] a full log of the buildout build would be helpful i guess :) [14:47] * ralsina suggests editig the freaking pylint script [14:47] gatox: i have no idea. on windows theres no shebang. we have an env.bat script which sets up paths and copies pylint (and a few others) out of their deep egg structures and puts them in the bin\ folder to be run [14:47] briancurtin: and those bats use the system python instead of the custom one? [14:47] i'm doing something similar in mac [14:48] ralsina: everything is started from bin\python.exe which is actually the system python but called with bin\python-script.py to setup "the buildout" (modifies sys.path to point to all of the eggs and whatnot") [14:49] briancurtin: ok, I am lost now. But that's normal. [14:49] * ralsina suggests using a virtualenv [14:49] but probably just more hassle [14:49] ralsina, we are going to have some problems with PyQt with virtualenv i think [14:50] gatox: I just copy it [14:50] gatox: since it doesn't install in virtualenv [14:50] gatox: or buildout for that matter [14:50] bin\python.exe actually executes """C:\Python27\python.exe bin\python-script.py""" to give your bin\python.exe the right environment [14:50] briancurtin: ahhhhh [14:50] ok, so if it's just pylint, and we use pylint for nothing else, let's change pylint.bat (and the pylint shebang on mac) so it does the right thing/ [14:51] or let's not run pylint on windows where it never wors anyway [14:53] https://code.launchpad.net/~dobey/ubuntuone-client/queue-limit-3-0/+merge/103174 [14:54] https://code.launchpad.net/~dobey/ubuntuone-client/fix-981255-3-0/+merge/103179 [14:54] thisfred, urbanape: ^^ if you could please review those. they're simple backports from trunk to stable-3-0 [14:54] kk [14:55] ralsina: btw, what about switch to pyflakes/pep8 or flake8 and just killing pylint? [14:55] urbanape, say did you have any thoughts / feedback on the little design tweek I had sent you ? [14:55] dobey: pyflakes/pep8 doesn't notice missing docstrings [14:56] dobey: and that was just the first time we tried to use it on something [14:56] Dr_Who: we actually had some mockups like that, and have some UI work on the roadmap that will enter that territory. [14:57] o cool ? is that up on a wiki somewhere or akin ? [14:57] mmmmm it's really easy to check for missing strings with ast.... [14:57] * gatox is going to send a patch to pyflakes [14:57] missing docstrings [14:58] in any case, even a first quick test did not fill me with confidence :-( [14:58] boo [15:00] me [15:00] does PEP 8 require docstrings? or is that all PEP 257? [15:01] 257, 8 just references it [15:02] ralsina, dobey thisfred mandel alecu briancurtin mmcc standup? [15:02] me [15:02] meh [15:02] dobey, pep8 does requires some docstrings; pep257 expands on it. http://www.python.org/dev/peps/pep-0008/#documentation-strings [15:02] me [15:03] but... "Docstrings are not necessary for non-public methods" [15:03] me [15:03] * mmcc doesn't know what to say here. me? [15:03] me [15:03] mmcc: yes :-) [15:03] me [15:03] ups [15:03] DONE: [15:03] Improve MAC OS scripts to configure the development environment. Almost done, just need to fix some issues running pylint. [15:03] TODO: [15:03] Finish with the mac os scripts and propose the branches for all U1. [15:03] BLOCKED: [15:03] No [15:03] thisfred, go [15:04] DONE: found and fixed indexing bug in u1db, added real persistence to u1todo TODO: bug #987414 BLOCKED: no NEXT: dobey [15:04] Launchpad bug 987414 in U1DB "Implement get_index_keys_from_index" [Wishlist,Triaged] https://launchpad.net/bugs/987414 [15:04] λ DONE: backports, expense, review [15:04] λ TODO: backports, SRUs, u1db packaging/buildsys [15:04] λ BLCK: none. [15:04] ralsina [15:04] DONE: tech leads call, mgmt call, helped around, some reviews, management stuff, etc. TODO: help around, reviews, etc. BLOCKED: no [15:04] alecu? [15:04] DONE: some research on SSL handshake error, team leads meeting, back to security issues [15:04] TODO: roadmap call, code some more [15:04] BLOCKED: no [15:04] NEXT: briancurtin [15:04] DONE: fought with Qt/PyQt version matching, got environment re-setup after that to do a one-off release and test it [15:04] TODO: i think i have Qt/PyQt figured out, have to rebuild it. fix up installer issues [15:04] BLOCKED: None [15:04] NEXT: mmcc [15:04] DONE: email, wiki setup; code and dependencies downloaded [15:04] BLOCKED: None [15:04] TODO: run tests, ask lots of questions [15:05] ok, I go [15:05] DONE: Update ssl logging branch according to reviews. Some reviews. Work on tcpactivation got to the point of having tests running via tcp and domain sockets. Work on getting squid tests running on windows (will propose in a little) [15:05] TODO: add a bug so that tcp test cases from ubuntuone devtools can use endpoints (allow domain sockets to be used) Propose squid tests on windows. [15:05] BLOCKED: no [15:05] dobey: you're good to go [15:05] * urbanape needs to remove himself from review days [15:06] oh right, guess we should put mmcc there instead :) [15:06] urbanape: yes, and you can probably stop showing up on standups, although you are welcome :-) [15:06] dobey, mmcc: yes, mmcc you are going to have a code review day. Take friday so you are not alone :-) [15:07] mmcc: at least for the 1st month [15:07] mmcc: the review day schedule should be in your google calendar [15:07] ralsina, ok, I'll check [15:08] make sure you log in with the right account to google docs :) [15:08] ralsina: I have nothing on there right now. [15:08] mmcc: hmmm [15:08] dobey: yep :) I use a separate browser for my personal account, for years now [15:08] mmcc: do you have any calendars other than your own? [15:08] ralsina: no - just mine and "Tasks" [15:09] mmcc: ok, sending you a few [15:09] mmcc: your account name? [15:09] mike.mccracken@canonical.com [15:09] ralsina, any idea why would I have a TCP miss in squid when trying to get a resource form localhost? [15:10] mandel: TCP_MISS is normal, it means "getting from server" [15:11] mmcc: there, you should now have 3 more calendars [15:11] ralsina, I get the following: paste.ubuntu.com/944210 [15:11] ralsina, the deny is expected, the miss are not happening on linux [15:12] ralsina: yes. what is the the core hours schedule about? it's busy… [15:12] mmcc: I usually have that one disabled, but is to show you when each person in online services is supposed to be working [15:12] mmcc: just in case you want to see if you can meet with someone [15:12] ralsina: ah, ok. I get it [15:12] mandel: as I said, TCP_MISS is perfectly normal [15:13] mandel: TCP_MISS or TCP_HIT are "good" results [15:15] ralsina, hm.. so when that happens, what is the client mean to do? [15:15] mandel: makes no diference at all to the client [15:15] mandel: the client still gets a page [15:15] mandel: TCP_MISS => gets it from the server TCP_HIT => gets it from cache [15:16] ralsina, I'm getting a 504 Gateway timeout [15:16] mandel: hmmm [15:16] mandel: that's something else [15:16] mandel: it could be it's not able to connect to the server [15:17] ralsina, to localhost with no firewall, weird.. [15:17] damn sinuses [15:17] mandel: squid should never "create" a 504 [15:18] * dobey has now deleted all the currently existing maverick packages from nightlies [15:18] EOL FTW [15:19] mandel: btw, it seems squid is just slow to start on the lucid ppa build. :( https://launchpadlibrarian.net/102707979/buildlog_ubuntu-lucid-i386.ubuntuone-dev-tools_3.1%2Br66-19~lucid1_BUILDING.txt.gz [15:20] dobey, looking [15:21] dobey, shall we wait longer? for the rest of the platforms there is no problem [15:21] mandel: how long are we waiting now exactly? [15:22] mmcc: if you want, you can just swap review days with me (Tuesdays) [15:23] urbanape: I don't know enough to have a preference :) whatever's easiest, I guess… [15:23] well, I just switched it to eric/?? for today. [15:24] urbanape: but today might be a bad day to start ;) (or a good day?) [15:24] Like the Klingons say, "Today is a good day to review merge proposals." [15:24] I just switched it to ralsina/ralsina [15:24] urbanape: those are some nerdy Klingons [15:24] (it's a lot like dying) [15:24] thisfred: hmmm [15:25] thisfred: what's that, "anything can happen tuesdays"? [15:25] oh wait, did you complete my evaluation yet? :D [15:25] urbanape: reviews are not at all like dying. Dying you only do once. Reviews, we do twice. [15:25] hahahahaha [15:26] thisfred: yes, it says, "ask jdo when he's back from the dentist" [15:26] dobey, look at like 222 in ubuntuone/devtools/services/squid.py [15:26] dobey, we look and try several times to know the state of the proxy, we can either increase the numbers or just add 10, 15 to it, longer might be too much [15:27] ralsina: truly anything can happen in that case :D [15:28] "I'll get me coat then..." [15:29] thisfred: he [15:30] Perhaps today *is* a good day to dye. [15:34] punning clans are here again [15:35] thisfred, what mouse walks using just two legs? [15:35] * thisfred puts fingers in ears and hums loudly [15:36] thisfred, Mickey Mouse! [15:36] thisfred, and the duck? [15:37] * thisfred hums louder [15:39] thisfred, then I wont tell you :P [15:41] windows es un hijo de puta que merece una muerte lenta [15:42] ralsina, if you use locahols on windows + squid it does not know how to resolve it to 127.0.0.1 [15:42] mandel: yuck [15:43] ping localhols fails on linux too ;-) [15:43] ralsina, double yuck, good news, ubuntuone-dev-tools squid tests now run on windows, which means I'll propose the branch and we can start running proxy tests on jenkins [15:43] mandel: nice! [15:48] briancurtin, ralsina, question, I'm using Popen to launch squid, if I use pipes on widows it always crashes, any idea? [15:48] mandel: do you have any example code? [15:48] hmm [15:49] briancurtin, sure, give me a sec and I'll push the branch [15:49] mandel: it should just work, code please [15:50] lolcaholics unanimous [15:53] * gatox lunch === gatox is now known as gatox_lunch [15:53] briancurtin, ralsina, if I do not pass -X to squid (which means be very verbose in stdout) it does not crash [15:54] briancurtin, ralsina, you can find the code in lp:~mandel/ubuntuone-dev-tools/fix-squid-tests [15:54] mandel: are you capturing stdout/stderr? If you are not, things may crash on windows [15:54] cautionary tale: I used the string "lolcathost" once in a functional test to see what happened with illegal urls. Then I forgot all about that, and later when something changed and it failed, neither I now someone else who was looking at it actually saw it wasn't localhost, so we lost about an hour wondering why the test was asserting that it *shouldn't* work against localhost. [15:54] ralsina, I do and when I don't it works [15:55] There's a lesson in there somewhere [15:55] thisfred: next time use something that doesn't look like a typo :-) [15:55] thisfred, on windows we have been seeing dns lookups for localhost.. is weird [15:55] mandel: then don't :-) [15:55] ralsina: that could well be it :) [15:55] ralsina, I want to do things right :) [15:55] mandel: localhost may or may not be in your hosts [15:55] mandel: that ship has sailed [15:55] ralsina, I could remove the -X will get less info but will work [15:56] mandel: ok, no idea, really [15:57] ralsina, then I'll remove the -X and screw windows since we have access to the jenkins machine to see wtf is happening === Guest86751 is now known as zumbi [16:03] mandel: sounds like a plan [16:03] mandel: let's pick our battles === gatox_lunch is now known as gatox [16:51] mandel, since this will run inside trial, have you considered reactor.spawnProcess for squid? [16:53] mandel: capturing stdout/stderr on windows, crashes? [16:54] alecu, nop, but this is an update to present code, it sounds like a good idea for improvement [16:54] dobey, capturing A LOT of stdout makes squid on windows crash, not python [16:54] oh [16:54] well [16:55] something inappropriate about squid's mother, then [16:56] dobey, I did mention that too :) [16:56] heh [16:56] i would expect no less of you, mandel [16:56] thisfred, please review: https://code.launchpad.net/~mandel/ubuntuone-dev-tools/fix-squid-tests/+merge/103325 [16:57] dobey, hehe I know [16:57] dobey, feel free to review ^ [16:57] ralsina: 1-1 in a few minutes? [16:57] ralsina, briancurtin, I'd appreciate if you can test https://code.launchpad.net/~mandel/ubuntuone-dev-tools/fix-squid-tests/+merge/103325 on your windows boxes [16:57] briancurtin: let's do it [16:58] dobey, also, squid fails to parse C:\path\in\windows you have to add it to the config as C:\\squid\\has\\mummy\\issues [16:59] of course [16:59] typo intended :) [16:59] dobey, and fails to pick up localhost.. [17:00] * thisfred reviews [17:00] alecu, ralsina, FYI I'll be working a little tom morning on ubuntuone-dev-tools to add support for domain sockets in the tcp tests cases I added last time, that way I wont leave dirty reactor on linux and we will be able to run the tcpactivation and ipc tests both on domain sockets and tcp sockets [17:01] uh [17:02] mandel, great. [17:03] ok, with that EOD from here, laters o/ [17:07] and i need to get lunch [17:14] alecu, I'm off to walk the dog and rugby, check https://code.launchpad.net/~mandel/ubuntu-sso-client/fix-ssl/+merge/103147 if you want I can remove logger, no problema what so ever, I'll me back at my 1:00 am or something like that [17:14] mandel, great, thanks. [17:19] be back shortly, running to the bank and to grab a sandwich [17:25] that nice feeling when the windows implementation works on linux === ralsina is now known as ralsina_lunch [17:43] * briancurtin back [17:43] ok, now i have EVERYTHING working on mac [17:44] gatox: I'm listening, what wasn't working before? I'm still digging around getting familiar with the code I've grabbed… [17:45] (so that'd be helpful info - knowing what shouldn't work :) ) [17:46] mmcc, i was just creating a couple of scripts so when you finish with the buildout, you can run another script to set some env-vars..... and then you will be able to run tests and lint checks, etc..... without a lot fo paths configuration and stuff [17:46] gatox: aha, so that's very relevant to my interests, as you might imagine [17:46] mmcc, i'll update the document as soon as the branches start landing [17:46] mmcc, yes :P [17:49] gatox: ok, good - for now I will just keep reading code/readmes/pyqt docs… [17:53] urbanape, ping [17:58] urbanape, hi, i'm going to create a branch from your mac-branch-port to fix the remaining things, so we can speed up that to make it land and we don't have to bother you with all the changes now that you went back to your team, if there is any problem with this please let me know :D [17:59] and i'll propose the new branch [18:00] facundobatista: so, last night i decided that u1sd simply couldn't handle a directory of 100GB+ so i did a workaround and just sync subdirs directly [18:00] achiang, probably, I don't think anybody tried with throwing 100GB at once... [18:02] gatox: still figuring things out - so urbanape is on a different team and was just helping get the OSX port started? how long has the effort been going on and how far is it along? [18:02] mmcc: not long [18:02] and yeah, I'm on the Web & Mobile team [18:03] gatox: pong [18:03] urbanape, hi... did you read my message? [18:03] urbanape, it's ok with you? [18:03] nope, no problem [18:05] urbanape: is there any overlap between the OSX code and your iOS code? [18:05] gatox, dobey: reviews please https://code.launchpad.net/~ralsina/ubuntuone-control-panel/unique_in_ubuntu/+merge/103336 [18:05] ralsina_lunch, on it [18:05] dobey: that one should go in the P SRU so be harsh with it please ;-) === ralsina_lunch is now known as ralsina [18:05] mmcc: not really. We could ostensibly re-use the SSO code, but right now, we're hewing more closely to the linux and mac ports. [18:06] urbanape: s/mac/win ? [18:06] er, yeah [18:07] yeah so I was just looking at the SSO stuff. I'm curious how it's done on iOS. I've used the OSX keychain, but I don't know what the done thing on iOS is… [18:15] ralsina, +1 [18:16] gatox: thanks! [18:21] gatox: does one of your pending doc changes involve installing dbus? either I have a path wrong or I need to install… [18:22] mmcc, no, you are probably trying to run the tests in the wrong branch or not ignoring the proper files [18:22] mmcc, you don't need dbus [18:22] mmcc, are you using this branch? https://code.launchpad.net/~urbanape/ubuntu-sso-client/initial-darwin-port/+merge/101112 [18:24] gatox: sorry, bzr newbie. what's the fastest way to answer that question? :) [18:25] bzr info didn't seem helpful there [18:25] mmcc, are you trying to run the tests for sso? [18:25] yes. [18:27] for completeness' sake, here's what I'm doing: [18:27] mmcc, show me the result of "bzr info" inside the sso foldeer [18:27] Users/mmccrack/Documents/Canonical/Source/buildout-env/scripts/devsetup/parts/ubuntu-sso-client [18:27] % bzr info [18:27] Standalone tree (format: 2a) [18:27] Location: [18:27] branch root: . [18:27] Related branches: [18:27] parent branch: bzr+ssh://bazaar.launchpad.net/+branch/ubuntu-sso-client/ [18:28] oof, that was ugly. sorry for the multiline spam there [18:28] mmcc, right...... so, there do: bzr merge lp:~urbanape/ubuntu-sso-client/initial-darwin-port [18:28] mmcc, and you will merge your local copy of sso with the content of that branch [18:29] mmcc, you can also branch locally if you want to leave that one clean, like this: [18:29] bzr branch ubuntu-sso-client new-sso-branch...... [18:29] and merge in the new branch..... so you will keep a clean copy of sso just in case [18:30] gatox: ok, well for now I just merged in place. I can always grab a clean copy later I guess [18:30] thanks [18:30] mmcc, yes..... [18:31] mmcc, now, try to run the tests in that branch as the docs says, and let me know if that works [18:33] ok, so the doc has a PYTHONPATH override that I need to translate to my filesystem again. just a minute [18:35] mmcc, yes, the script that i made is going to avoid doing that..... but is not in trunk yet [18:37] gatox: that's good news :) [18:38] mmcc, yep! :D [18:39] bah, where are the free downloads on google music now [18:41] gatox: so now it's dying when it tries to import resources_rc from ubuntu_sso.qt.ui - I note the import line has been wrapped with pylint comments… [18:41] mmcc, yes...... sorry about that...... as you are running the test manually, you need to compile everything by your own..... [18:41] mmcc, execute: python setup.py build [18:42] aha, ok [18:42] mmcc, later when the ./run-mac-tests will be available, you wont need to do that (among other things) [18:45] gatox: are there other parts I need to build? [18:45] from ubuntuone.devtools.testing.txwebserver import HTTPWebServer [18:45] ImportError: No module named txwebserver [18:45] yay buildout using the old devtools [18:46]