=== zumbi is now known as Guest18413 [08:55] morning all! [09:14] Happy Monday, people! :D [11:05] good morning! [11:18] mandel, how was your python+android talk? [11:18] gatox, it went well.. sorry I did not ping you, it turned out I though it was on the wrong day hehe [11:18] I'm not good with dates :P [11:19] mandel, jejjee no problem! [11:19] mandel, why you didn't make a post about python+android?? [11:19] you always make post about the new things you learn [11:20] * gatox wants to read it from mandel's blog..... is too lazy to google jeje [11:20] gatox, will do, I have to do a decent post hehe [11:20] jeje [11:21] gatox, to much work to do.. atm doing a nice twisted test case that can be used to lauch a twisted service and client and close them correctly :) [11:22] gatox, I also have to write one about greenlets and force you use them instead of threads :) [11:22] mandel, great [11:22] mandel, me? [11:23] gatox, well, you because you are here hehe [11:23] jejeeje ahhhhh okok [11:23] * gatox reads about greenlet [11:26] gatox, coroutines for the win! [11:27] mandel, jeje yep, i've been reading..... sounds good [11:48] hello, all! [11:48] ralsina, dobey: ping [11:54] alecu, hi! [11:55] hola gatox! how was your weekend? [11:56] alecu, awesome!! I rested a lot!! i feel like a new person!! :P [11:56] alecu, yours? [11:56] gatox, I felt like crap thru most of the weekend, so not so good :P [11:57] gatox, I'm still feeling kind of dizzy... not sure if I'll take the day today. [11:57] *the day off [11:57] alecu, ohhhh.... [11:58] alecu, get some rest [11:58] gatox, yup, I should. But I need to discuss some matters with ralsina first...! === Guest18413 is now known as zumbi [12:15] alecu, gatox, I had dst in spain, when do we have standup from now? [12:16] mandel, dst? [12:16] mandel, three hours from now. [12:16] mandel, 2.45hs [12:16] gatox, http://en.wikipedia.org/wiki/Daylight_saving_time :) [12:16] ahhhhhhhhhh [12:17] i didn't know what dst stands for [12:17] gatox, acronyms are like that, or they give you the full info or you are wtf?! [12:17] mandel, yep..... i'm not particularly good with acronyms [12:18] gatox, is one of those that you learn after woking with dates hehe [12:18] jeje [12:30] alecu, i need your dbus powers for a quick questionn [12:30] gatox, here, have them. All yours! [12:31] alecu, ..... not funny :P [12:31] alecu, if i modified the returned values from dbus_interface and put that instance of u1-client to run...... when i check the methods with d-feet it should return the new value i set??... or i have to modified this somewhere else in the code? [12:32] gatox, can you point me at the code? [12:32] alecu, /ubuntuone-client/ubuntuone/platform/linux/dbus_interface.py [12:34] any of the methods that are returning a value...... just need to know, if i modified something that return a string, to return now: "return 'my_string'"...... with d-feet when calling that method i'm going to get "my_string" [12:34] (because i'm trying to do that...... but it seems that is not working..... so i want to know if i'm doing something wrong) [12:36] gatox, the method you changed, is it decorated with @dbus.service.method? [12:36] gatox, (btw: which method have you changed????) [12:36] alecu, yes, it has that decorator: bandwidth_throttling_enabled [12:36] gatox, and... in case it is decorated... what's the "out_signature"? [12:36] alecu, don't panic.... i'm just testing something! [12:37] gatox, (I'm asking because it's not good to change our current APIs) [12:37] alecu, the out_signature was "b".... and i was returning False.... then i changed to "s"..... and i was returning a string.... but nothing works [12:38] gatox, what does "nothing works" mean in this context? [12:38] alecu, nono..... i'm not changing anything, don't worry,... i'm just trying to debug something to know where the bug come from [12:38] gatox, do you get a stacktrace from SD? [12:38] good morning! [12:38] alecu, that no matter what i change d-feet always return 1 for that method [12:38] hola ralsina! [12:38] ralsina, hi there! o/ [12:38] gatox, then it's very likely that the SD you are running is not the one from that tree [12:39] anyone needs my morning reviews slots? [12:39] gatox, how are you starting SD? [12:39] ralsina, absolutely yes [12:39] alecu: I saw some strange comments from you on friday after EOW, what happened? [12:39] ralsina, if you insist :P https://code.launchpad.net/~diegosarmentero/ubuntu-sso-client/reset-error/+merge/99039 [12:39] gatox: thanks! [12:40] alecu, ahhhh ok..... so the answer is: "if i change the returned values in those methods..... i should be able to see the new value with d-feet" [12:42] ralsina, I'm asking gatox to review that delicate branch too. [12:42] alecu: ack [12:44] alecu: aproved *that* branch [12:44] ralsina, thanks. [12:46] gatox, *how* are you starting sd from that folder? [12:47] ralsina, and after that branch, I have this other one: https://code.launchpad.net/~alecu/ubuntuone-client/stop-proxy-tunnel/+merge/99134 [12:47] alecu, first: u1sdtool -q (to shutdown the actual), then: PYTHONPATH=. bin/ubuntuone-syncdaemon [12:48] alecu: ack [12:48] alecu: could you (if you have a slot) check why syncdaemon doesn't stop on windows? I am at a loss there. [12:48] ralsina, yes: I've got that bug as the first thing for this morning [12:49] bug #963404, right? [12:49] Launchpad bug 963404 in Ubuntu One Client "On windows, syncdaemon can't be stopped." [High,Triaged] https://launchpad.net/bugs/963404 [12:49] alecu, +1 [12:49] alecu: thanks [12:49] gatox, great [12:49] alecu: right [12:49] gatox, that seems to be right. In any case, to check that sd is working, try running it with debug on: [12:50] alecu: you may have to fix the "not starting proxytunnel causes syncdaemon to go autist" bug first [12:50] U1_DEBUG=True PYTHONPATH=. bin/ubuntuone-syncdaemon --debug [12:50] alecu, true.... thanks [12:50] ralsina, lol. [12:50] alecu, ralsina how do you feel about this: http://paste.ubuntu.com/900387/ [12:50] ? [12:50] alecu: my current guess is we are doing a defertothread somewhere and it's getting hung, but I have no first clue on how to debug that [12:51] mandel: how should I feel, and what's that? [12:51] mandel: oh, the test case! [12:51] ralsina, example usage: http://paste.ubuntu.com/900389/ [12:51] mandel: I am the wrong person to review that [12:52] ralsina, more the API than anything else :) [12:52] mandel, I like the general shape. But what happens if one test needs more than one server? [12:53] mandel, aggregation beats inheritance. [12:53] mandel, also: it would be right if you can include a small test that uses this, to check on the API. [12:53] alecu, I used that wen doing the IPC the first time and it ended being a mess hehe (the aggregation part I mean) [12:54] +1 on aggregation being a mess that time [12:54] alecu, example usage: http://paste.ubuntu.com/900389/ [12:54] tests were absolutely incomprehensible (to me, bah) [12:54] mandel, that's the same url, isn't it? [12:55] alecu, I have http://paste.ubuntu.com/900387/ and http://paste.ubuntu.com/900389/ [12:55] oh, doh. [12:55] alecu, very very similar ones hehe [12:55] alecu: nessita is on the mechanic getting her mirror fixed (apparently it says she's not the prettiest in the kingdom anymore) [12:55] * ralsina is going to go to hell for that joke and he knows it, and he can't resist. [12:55] alecu, I like the idea of supporting several servers.. let me think about it since is certainly possible to do [12:56] ralsina, lol yet bad joke hehe [12:56] alecu: so, I don't think we are going to get a fixed package before release [12:56] mandel: compared to yours, that joke is a classic ;-) [12:56] ralsina, jejejje [12:57] ralsina, just because you are classic buahahaah [12:57] :P [12:57] alecu: we should be able to get one that people will get right after installing, though [12:58] alecu, getting more than one server implies that the developer should pass the client factory to use and agains which server factory to connect, I'll see how well that looks like [12:58] mandel, let me write a small pastebin on how I picture the api. [12:59] alecu, please do! [12:59] alecu: nessita says she needs *that* fix proposed against stable-3-0 [13:00] alecu, also, what do you think about the context manager? it avoids cases in which the server disconnected deferred is called do to calls in the setUp [13:01] ralsina, do I have an hour til the standup? [13:01] mandel: 2 [13:01] * ralsina needs to fix that damn thing [13:02] ralsina, cool, them I'm off to have lunch [13:02] * mandel gets screwed up by dst.. [13:02] mandel, why "yield self.client_connected" in the super in the test? [13:03] argh, everyone, sorry about the mail you are going to get about standup moving, it's wrong and will be undone in a minute [13:03] alecu, ensure we connected.. recommended by http://mumak.net/stuff/twisted-disconnect.html [13:03] mandel, http://pastebin.ubuntu.com/900404/ [13:04] mandel: standup is at the right time in the team's calendar, use it ;-) [13:05] alecu, then in the tests you have to make you connected to the client [13:06] mandel, I think it should be the most common case, where the test yield on some deferred that's only fired after the client has done some work. [13:07] mandel, but I agree that exposing that option may make sense for some corner cases. [13:07] mandel, so I think that bit should be optional [13:07] mandel, ... [13:07] mandel, in any case, I still don't believe we can come up with a perfect solution if we don't give this a few iterations. [13:08] mandel, by replacing some bits of the tests that we already have with this [13:09] alecu, yes, I think we can do what you showed with the code present in the test rather than in the setUp with no problems, running more than one servrice is harder (in terms of api) yet doable [13:09] alecu, but I don't think we have that use case [13:10] mandel, agreed [13:12] mandel, I didn't see the bit with the context manager... [13:12] alecu, I did in a test for the ipc in sso, but do to the way the building of the client is used [13:13] alecu, I'm going to fix the sso tests on windows using this to see how it brakes 'cause it will brake [13:14] mandel, and you still think that refactoring tests is a good thing to do now.... [13:14] mandel, I don't like the solution with the context manager too much. [13:15] mandel, if we do it with aggregation, then we can have two independent servers, and we don't need that thing with the context manager, right? [13:15] alecu, me neither, is all I could come up with to fix those tests I've seen, or fully refactor it which is not something I want to do [13:16] alecu, yes.. since is not the same guy.. I'll play with that approach after lunch [13:17] * mandel lunch [13:21] alecu: hi [13:35] nessita: I think I addressed your remaining concerns on https://code.launchpad.net/~brian.curtin/ubuntuone-control-panel/autostart-clean/+merge/99089 [13:37] briancurtin: nessita is not around yet, and she has no logs, so you will have to repeat that after she arrives [13:37] briancurtin: and good morning! [13:37] ralsina: good morning, and thanks === yofel_ is now known as yofel [13:48] hello all [13:48] nessita, hi [13:50] ralsina, alecu: I guess we can make a new release for ussoc (re logging proxy credentials), but since we're in beta freeze we can't upload to ubuntu [13:51] ralsina: perhaps you can ask in #ubuntu-release about this? they may give us "permission" to upload given the security issue [13:51] nessita, it would be for u1-client [13:51] nessita: ok, asking [13:51] ralsina, ^ [13:51] alecu: ah, ok, sorry. u1-client then [13:51] alecu: are you 100% sure that's the only place? [13:52] ralsina, good point. I'm checking sso too. [13:53] nessita: i think i addressed your concerns on both of these: https://code.launchpad.net/~brian.curtin/ubuntuone-control-panel/no-symlinks-on-windows/+merge/98736 and https://code.launchpad.net/~brian.curtin/ubuntuone-control-panel/autostart-clean/+merge/99089 [13:54] briancurtin: no-symlink in windows was merged on friday... no? [13:54] nessita: oops, im looking at the old proposal [13:54] i thought something felt weird about that [13:54] hola nessita [13:55] hola dobey [13:56] nessita: you forgot to add python-dirspec as a dep, to python-ubuntuone-control-panel i think [13:57] ralsina, now I'm more sure that it's the only place. [13:57] dobey: you migrated u1cp to dirspec? [13:57] or, who did? [13:58] nessita: no. you used it in one branch as per my suggestion, instead of hard-coding "~/.local/share/" [13:58] dobey: oh, right [13:58] dobey: would you please file me a bug in u1cp? [13:58] there was one filed. let me find it [13:58] nessita: answer in -release: "you can in any case upload to the queue, letting the package in is a manual process during beta freeze" [13:59] nessita: there was one about a missing dirspec dependency this weekend [14:00] ralsina: haven't looked email yet... will do soon, and search for that [14:00] hmm, i can't find it [14:01] oh fun [14:01] nessita, i think you will enjoy this: http://pypi.python.org/pypi/pep8nazi/0.1 [14:01] :P [14:01] searching for people/teams in the "asign a bug" dialog thing on lp isn't working [14:02] gatox: LOL [14:02] briancurtin: you may wanna reject this proposal https://code.launchpad.net/~brian.curtin/ubuntuone-control-panel/no-symlinks-on-windows/+merge/98736 [14:03] nessita: its rejected and deleted [14:03] briancurtin: ah, ack. Ideally, do not delete branches that have votes in it, just reject them [14:04] nessita: even after the branch/proposal which replaces them is landed? [14:04] * mandel back [14:05] briancurtin: yes [14:05] nessita, FYI example usage for the tests case: http://paste.ubuntu.com/900389/ [14:06] nessita, there is some more work needed to ensure that that is how we want it to work, I'm testing that api with the broken tests on windows [14:06] grr, launchpad [14:06] i give up [14:07] mandel: after q very quick look, that code does not look like the code you use to add... wanna give me a brief summary of what's the goal of it? [14:08] nessita, that code is using: http://paste.ubuntu.com/900387/ [14:09] nessita, the idea is that you pass the factories you want to use for the service client tests and the test case will use such factories with a protocol that has the required deferreds [14:10] nessita, so, in general, you tell the testcase, listen to a port using MyServerFactory and use MyClientFactory to create the clients, then I'll use yield client_connected to know I'm ready to start testing [14:10] nessita, tearing the server and client correctly is done by a Cleanup that is added by the parent [14:11] mandel: we usually do not have a custom factory (nor for servers or clients).... why not accepting that as parameter, but defaulting to the twisted ones? [14:12] alecu: let me know when you have the patch ready [14:12] nessita, you don't have a custom factory in the production code but for the tests you need one since you are changing the protocol class attribute [14:12] mandel: right, but we always use the sme one, why not provide that in that patch? [14:12] nessita, I will. [14:12] nessita, passing the factory will allow this code not only to work with Pb but with any service provided over a Tcp socket [14:12] nessita, I'm getting this, btw: http://pastebin.ubuntu.com/900503/ [14:13] mandel: atm we're just trying to solve the PB case... [14:13] mandel: so, is good to have a parameter for that, but also to have a default [14:13] mandel: so, callers do not need to build any factory at all [14:13] alecu: smells like an outdated system [14:13] nessita, sure, I can set the default one to be Pb that is not an issue at all [14:14] alecu: in particular, old ussoc [14:14] nessita, yes: I'm getting an error installing the newer: http://pastebin.ubuntu.com/900507/ [14:14] mandel: I was hoping we could have the thing you added to ussoc, in the last branch that started this conversation, to be moved to devtools [14:15] alecu: looks like you have some ensalada between nightlies and not nightlies? [14:15] alecu: is this a new install? [14:15] nessita, and that is exactly what that code is doing but in a more generic way, PbServerFactory and PbClientFactory are small extensions of the default, for example: http://twistedmatrix.com/trac/browser/tags/releases/twisted-8.2.0/twisted/spread/pb.py#L1160 [14:15] alecu: do you have nightlies available and up to date? [14:16] nessita, no: it's my desktop, that I've updated from O to P during the weekend. [14:16] alecu: right, so nightlies are disabled [14:16] nessita, tis approach is more generic and will allow people to write better tests using a server and a client in twisted [14:16] alecu: can you please enabled them back again and update? [14:16] nessita, doh. thanks! [14:16] mandel: right, but our agreement was "think" general, but implement specific for PB [14:17] mandel: so, ideally, I would like that code in sso will never have to create a custom factory [14:17] and have that provided by devtools [14:17] nessita, sure I'll do that, is just hanging a few lines [14:18] mandel: that's great: the gain is that otherwise we will adding the same few lines in different projects [14:18] nessita, I know :) [14:20] mandel, nessita: I want this for other tests, besides PB [14:20] alecu: for the package, we need a patch of level 1 (can be generated with bzr diff -p 1) [14:20] alecu: yes, that's what we talked [14:20] alecu: we also said do not go general just now, think general and implement PB, right? [14:20] nessita, I wholly agree with mandel's idea of having to explicitly pass the factory. [14:21] alecu: I think we should have (perhaps) a general test case that accepts the factory, and then specializations of testcases for specific factories. Again, that looks a next step (and not this one) [14:22] alecu: I thought you agreed to "do not solve general now"... I'm getting confused [14:26] gatox: what's this line for in the reset-error MP? + exc = failure [14:26] nessita, I agreed to fix only PB for now, yes. But I think we should make the code that we use for that be as general as possible. [14:26] nessita, :P was testing something and forgot to change that [14:27] nessita, and having that general bit of code creating the factories specific to this test does not make sense to me. [14:27] alecu: can you explain a bit more? I'm starting to think we're talking about different things [14:28] hmm, i wonder why people removed python-dirspec from the install [14:28] or maybe they're all on kubuntu [14:28] nessita, fixed [14:30] morning, all [14:31] nessita, the code that mandel is working on starts a listenTCP with a given server factory with some protocol on a random port. Then it starts a connectTCP to that port with a given client factory with some other protocol. [14:31] nessita, in all cases it's the factory the one responsible for creating instances of the right protocols [14:32] nessita, but in many cases we are using custom factories, since they hold instance variables that are used for all protocols created by it. [14:33] nessita, so, in my opinion it makes a lot of sense to be able to "pass" factories into these utility functions that mandel is creating. [14:33] hey there, urbanape! [14:33] nessita, and it makes a lot less sense to pass just instances of the protocols. Or of the classes of the protocols. [14:34] alecu: yes, agreed so far except the last sentence. My point is most of the cases we're using the *same* factories, but we write the class twice or more. An example is the ServerProtocol, SaveProtocolServerFactory, SaveClientFactory: I don't want that we need to create those every time, I think that those should be provided in devtools [14:34] alecu: those classes are from https://code.launchpad.net/~mandel/ubuntu-sso-client/fix-broken-tests/+merge/98868 [14:35] alecu: so, we can talk about how to provide that API the best, but I don't think those classes should be built every time we want to use the base test case mandel is working on [14:36] nessita, I agree on that [14:38] nessita, so: I agree on the need to have helpers or mixins for server factories when what we are testing are clients. [14:38] alecu: so, perhaps you have a better suggestion, but I was thinking we should have some API like this: http://pastebin.ubuntu.com/900545/ [14:40] nessita, I requested mandel that we should have aggregation, not inheritance for this. Something like: http://pastebin.ubuntu.com/900404/ [14:40] alecu: why? [14:40] I mean, what do we gain with that? [14:41] nessita, so a given test may start more than one instance of this pattern. And so I can inherit from that and change the listenTCP to listenUNIX or something else, and I can do it in a general way. [14:42] nessita, if I add this to a given testcase instance it would be a mess to change. [14:43] alecu: ok, but, in the 90% of the cases the use case is much simpler (start a server, start a client, use the client), and I would like that we provide a testcase (again, may be a specialization or similar) to trivially use for this case [14:43] alecu: so, as a immediate goal, I would like that mandel's branch provides a test case that can be use in the ussoc tests by just changing the inheritance class in the ussoc end [14:47] nessita, I agree on having a specialization that can be used as a mixin in any test that needs the simpler way. I'm just stressing that the lower level bits should be separated, because I have some use cases for that too. [14:47] alecu: +1 to consider that. I was just asking to solve the specific PB situation now, but of course considering the larger picture [14:48] nessita, so, the paste I put above can be seen as what's inside the "simplified" testcase you mention on your paste. [14:48] nessita, and 90% can inherit from that testcase, while the 10% can use the MandelClass directly. [14:48] nessita, does that make sense? [14:48] alecu: exactly [14:48] yes [14:51] brb! need to restart [14:55] nessita, is control panel with the sso login working for you? [14:56] gatox: looking [14:56] gatox: apparently not :-/ [14:57] nessita, i keep getting "There was a problem trying to setup the page" [14:57] nessita, that is blocking me for the other issue i'm trying to fix... can i create a bug for that and start working on this? [14:58] gaaaaah I know what happened [14:58] gatox: your main-moved branch broke it [14:58] gatox: can you please file the bug, and propose a branch where the main module is listed in setup.py? [14:58] nessita, it seems...... but it was working when i tested [14:59] nessita, yes [14:59] gatox: the ubuntu_sso/qt/main/ is not being installed [14:59] nessita, ahhhh ok [14:59] gatox: thanks! and great catch [15:00] standup people! [15:00] me [15:00] me [15:01] me [15:01] me [15:02] me [15:02] gatox, dobey, thisfred, urbanape: standup [15:02] already me [15:02] meh === h00k_ is now known as h00k [15:03] mandel: go? [15:03] DONE: First implementation of a test cases that allows to start twisted services, connect to the via a client and close correctly. [15:03] TODO: more work with the above to include comments from alecu and nessita. [15:03] BLOCKED: no [15:03] next: briancurtin [15:03] alecu: me? [15:03] DONE: complete autostart and symlink branches, IRL testing [15:03] TODO: get approval on autostart, anything else needed for windows, get on with the mac work [15:03] BLOCKED: need coffee [15:03] NEXT: nessita [15:04] DONE: tried to improve the notification test suite in u1client (could not do it), reviews, advices [15:04] TODO: wasted 3 hours of my life in the car repair shop (they can't tell left from right :-/), more reviews, start with bug #945078 [15:04] BLOCKED: no [15:04] NEXT: ralsina [15:04] me [15:04] Launchpad bug 945078 in Ubuntu One Control Panel trunk "Qt UI: without credentials, if there is an error in the first screen, it freezes and nothing happens" [High,Triaged] https://launchpad.net/bugs/945078 [15:04] DONE: lots of small windows bugfixes, administrivia, reviews. TODO: more of the same, tech leads call, move forward with windows release (maybe) BLOCKED: I won't release with syncdaemon being unstoppable :-( NEXT: gatox [15:04] DONE: [15:04] Proposed the Reset password branch (there was a problem with the handling of WebClientError to create the error message in the ui), worked on the limit bandwidth issue, found the sso login not working issue (proposing a branch for this) [15:04] TODO: [15:04] Finish with the limit bandwidth issue as soon as i fix the sso problem., then keep fixing ui stuff. [15:04] BLOCKED: [15:04] No [15:04] dobey, go [15:04] λ DONE: bug triage, bug #963323, partial bug #961342 [15:04] λ TODO: perf review stuff, code reviews, finish #961342 [15:04] λ BLCK: none. [15:04] Launchpad bug 963323 in rhythmbox-ubuntuone (Ubuntu Precise) "Multiple empty Ubuntu One libraries" [High,Triaged] https://launchpad.net/bugs/963323 [15:04] Launchpad bug 961342 in Ubuntu One Client "GI error while trying to show notifications" [High,Confirmed] https://launchpad.net/bugs/961342 [15:04] alecu [15:04] oh, performance review!!! [15:04] * nessita adds to tomorrow's TODO [15:05] also, I need to upload new packages for u1cp and u1client today [15:05] DONE: some proxy bugfixing [15:05] TODO: allhands reviews, more bugfixing [15:05] BLOCKED: no [15:05] nessita: por que client? [15:05] urbanape, thisfred: fight for who's next [15:05] dobey: proxy credentials are being logged :-/ alecu will provide the level 1 patch [15:06] my bad. I left a debugging log :-( [15:06] * alecu blushes [15:06] nessita: ah ok. archive is frozen, so you might need to poke release team [15:06] dobey: yes, ralsina already gave the heads up [15:06] ok [15:07] oops ME [15:07] alecu: anyways, u1client by default logs in INFO [15:07] alecu: but we need to fix anyways ASAP [15:07] alecu: is that why you pinged me earlier? [15:07] DONE: Explore Savannah, GA TODO: functions and mappin [15:07] nessita: no it doesn't. it logs in DEBUG by default for development releases [15:08] gs in u1db indexes [15:08] BLOCKED: no [15:08] dobey: ah, I thought it was DEBUG only for nightlies [15:08] dobey, yup. I pinged you as soon as I started today, and ne-ssita and ra-lsina too. [15:09] nessita, should I include the diff for the testcase too in the diff -p 1 ? [15:10] nessita: nope, any release for X.Y.Z where Y is an odd number [15:10] alecu: yes please [15:10] dobey: good to know, thanks [15:11] but the proxy package is not installed by default, and not getting installed by ubuntuone-installer, so affected installs is probably very small if any, right now [15:12] nessita: also, need to teach you how to move bugs to a different project/package, instead of marking invalid and adding a new package with "also affects distribution" :) [15:13] nessita, I've attached the patch to the bug: https://bugs.launchpad.net/ubuntuone-client/+bug/963568 [15:13] Error: ubuntu bug 963568 not found [15:13] dobey: how can they be moved? [15:13] alecu: thanks [15:13] brb, getting nagged to reboot for updates [15:13] dobey: I know I can edit the project [15:13] dobey: but I've tried to move the source package, and I couldn't find how [15:13] although, it's not really doable for projects that are targetted to a series, though for packages it is [15:13] nessita, the fix also landed in trunk [15:14] nessita: if you click the expander arrow next to the package name (or series for ubuntu bugs), it should be obvious [15:14] dobey: how? :-) [15:14] alecu: ack [15:14] dobey: looking [15:15] dobey: ah! I think I never have expanded that arrow :-D [15:15] ok, duly note dnow [15:15] hehe [15:16] nessita: for packages/projects which aren't targeted to a specific series, you can even move a bug from affecting ubuntu, over to just the upstream project, or from the upstream project, over to ubuntu. [15:17] and for project bugs that are targeted to a series, you can't move them it seems. :-/ [15:18] nessita, fixed: https://code.launchpad.net/~diegosarmentero/ubuntu-sso-client/install-main-module/+merge/99352 and i tested this installing this branch on my machine [15:19] gatox: awesome! [15:19] hey, all: anybody wants to do a review? https://code.launchpad.net/~alecu/ubuntuone-client/stop-proxy-tunnel/+merge/99134 [15:20] gatox: reset-error looks great! [15:21] nessita, great [15:21] gatox: looking the main-module now [15:23] oh, whoops. [15:24] DONE: ubuntu-sso-client mac progress [15:24] TODO: finish him [15:24] BLOCK: None [15:27] gatox: bug #965400 [15:28] Launchpad bug 965400 in Ubuntu One Control Panel "UnicodeDecodeError on portuguese windows XP" [Undecided,New] https://launchpad.net/bugs/965400 [15:28] ralsina, mmmmm i'm going to test that..... [15:29] gatox: ack [15:30] gatox: can you please remove the 'ubuntu_sso.qt.main.tests', line? we're not installing tests for the UI [15:30] nessita, ah ok.... i get confuse with the other lines installing tests [15:31] gatox: yeap, sorry [15:31] nessita, done [15:31] gatox: otherwise, I'm approving and globally approving when this is done [15:31] gatox: nice! [15:40] * gatox lunch === gatox is now known as gatox_lunch [15:41] briancurtin: quick question re: addtoautostart branch [15:41] nessita: what's up [15:42] briancurtin: given that you're patching it only in AutoStartTestCase, does that mean that the real thing is called in the MainWindowTestCase? [15:42] briancurtin: so, when running the test suite in windows, will u1 be really added to the autostart? [15:42] nessita: the function only has effect when run from frozen binaries [15:43] nessita: so in the test suite it's really a noop [15:43] briancurtin: ah, ok, thanks [15:46] * alecu reboots === gatox_lunch is now known as gatox [15:51] ralsina: should i work on the syncdaemon shutdown issue or is that covered? [15:55] briancurtin: alecu has that one [15:55] briancurtin: so, do you have ny pending branches? [15:56] ralsina: the only one we need to release is the autostart branch which nessita was just asking about, just waiting on approval [15:57] ralsina: anything else blocking the release? === gatox is now known as gatox_lunch [15:57] ralsina: about to approve [15:57] tetsing IRL now [15:57] but was blocked by the non working sso :-), which I'm updating now [15:58] briancurtin: how about this: TunnelRunner failing to spawn should make syncdaemon start as no-proxy, not go insane :-) [15:58] briancurtin: alecu may have some idea about how to fix that, so talk to him [15:59] ralsina: sounds good, looking into it [15:59] briancurtin: the unstoppable syncdaemon is the only blocker I have noted [15:59] briancurtin, that's what SD does while running on linux [16:00] briancurtin, the ubuntuone.proxy python package can be completely gone on linux, since it's a different binary package than SD. (ubuntuone-client-proxy vs. ubuntuone-client) [16:01] alecu: ack, thanks [16:02] * alecu needs to restart X, again. [16:07] briancurtin: approved [16:08] ralsina: you will also have to review that (you're requested as reviewer) [16:08] nessita: ack [16:09] dobey: did you file me the bug for u1cp packaging? :-) [16:11] briancurtin, are you running a windows box atm? [16:11] mandel: always [16:11] nessita: i couldn't find the bug i was thinking of. launchpad search is being very unhelpful, sorry. i'm not filing a new bug, there is already a bug filed which i was thinking of [16:12] briancurtin, hehe can you help me running some tests in it? I want to make sure everything works in all stations [16:12] mandel: yep, let me know what you want to run [16:13] briancurtin, can you add lp:~mandel/ubuntuone-dev-tools/tcp-testcases in your python path and run the tests on windows for lp:~mandel/ubuntu-sso-client/fix-broken-tests [16:14] briancurtin, if everything is correct I should also be fixing the mysterious failing tests on mac os x :) [16:14] ah of course [16:14] someone filed it in the wrong palce [16:15] nessita: is bug #963538 [16:15] Launchpad bug 963538 in ubuntuone-control-panel (Ubuntu) "ubuntuone-control-panel-qt consistently fails to start" [High,Confirmed] https://launchpad.net/bugs/963538 [16:16] dobey: thanks [16:16] nessita: i presume it also needs fixing in nightlies as well. [16:16] dobey: yes === gatox_lunch is now known as gatox [16:21] ok [16:21] i need to get some lunch [16:21] bbiab [16:25] nessita, in the ubuntu_sso/main/tests/test_clients.py do the linux tests use dbus? [16:26] mandel: yes (and please please pretty please keep it like that :-)) [16:26] nessita, I don't want to remove it, that is for sure, I'm trying to understand the tests correctly to do the right thing [16:27] mandel: thanks! :-) [16:27] nessita, so, the dbsu address, which one is it? [16:27] mandel: may I ask why you need that? (so I can give an accurate answer) [16:28] nessita, in those tests you are starting a UbuntuSSOProxy which on windows has two problems to run the tests, first it looks for a precise key in the registry of windows which does not have to be present and two it starts a PBServer which I need to make sure that does not leave a dirty reactor [16:29] mandel: the registry in the server is mandatory [16:29] mandel: everyone should have the key in the registry, even if it's pointing to a dummy localtion [16:29] nessita, I don't and I get lots of errors due to that [16:29] mandel: right, so you should have a key [16:30] mandel: there is script that adds them [16:30] mandel: you just have to run trunk/ubuntu_sso/main/tests/ubuntuone.reg [16:30] Can I request my Ubuntu One account upgraded as I am a new Canonical Employee (or where do I request this?) [16:31] ralsina, i'm testing the unicode issue...... that seems to be some kind of regression or something.... that was fixed long time ago :S [16:31] nessita, but doesn't it smell a little that if that the tests depend on the enviroment and assuming that is ok, should they be creating it? [16:31] gatox: ack [16:31] ralsina, i'll start working on that now [16:31] gatox: cool, god luck! [16:32] mandel: we had to make a decision back then, and we decided that having the key in place (even if it points to a dummy place) was the best option [16:32] mandel: handling not having the key is too much work compared to what we gain [16:33] mandel: https://pastebin.canonical.com/63042/ [16:33] nessita, ack, then lets move to the second question, in windows we have to make sure that the reactor is left clean, that implies patching the start of the proxy (or doing something with it, I'm not yet sure) [16:33] briancurtin, thx! just like I suspected, I should not trust my vm :( [16:34] mandel: yes, I can mumble about it when this meeting I'm in finishes [16:34] nessita, sure, ping me whenever you are ready :) [16:34] mandel: yes! [16:36] ralsina, mmmm this is weird, trying the projects from sources i don't have that bug... is there any py2exe thing that you know that i should have in mind? [16:36] gatox: none comes to mind [16:36] ralsina, i'll try to create the .exes in a weird-name env.... [16:38] briancurtin: how's your day looking? [16:40] urbanape: i'm past my Windows blockers, but i do have one more thing to do that i'm working on right now with TunnelRunner. seems like i should be done early afternoon [16:40] great, thanks [16:41] briancurtin, urbanape I should be able to pass you a branch which fixes the sso tests on windows and mac after I have a chat with nessita [16:42] thanks [16:59] dobey: I'm getting these lintian warnings and errors in u1client: http://pastebin.ubuntu.com/900763/ [16:59] mandel: ready in mumble [17:01] mandel: ready when you are [17:02] nessita, rebooting the machine to get on mac os and we mumble [17:02] mandel: ack [17:02] nessita, sorry but it does not work on linux :( [17:05] nessita, ready :) [17:10] nessita, lp:~mandel/ubuntu-sso-client/fix-broken-tests [17:10] branching [17:25] nessita: you only need to fix the ubuntuone-client-proxy one there [17:26] dobey: ok, so these are present in ubuntu now, no? [17:27] nessita: the other two warnings are fine to ignore. i didn't see any about ubuntuone-proxy-tunnel when i did the upload to ubuntu [17:27] dobey: ack, will fix [17:27] nessita: but if it is there, please fix it [17:27] sure [17:28] thanks [17:36] ok, EOD for me, briancurtin urbanape I'll update that branch for you my morning so that you have it ready when you arrive [17:36] nessita, thx again for the chat! [17:37] all, laters! [17:37] mandel: my pleasure!!! [17:37] mandel: thanks, take it easy [17:39] thanks, mandel [18:01] nessita: btw, did you see my comment on your add-recorder branch? [18:02] dobey: yes, will reply when I finish with packages.... any suggestion for the new module name? [18:02] nessita: recorder i think would be fine, unless you have a better idea. [18:03] though Recorder vs. Recorded is a 1 char difference, and that can be troublesome sometimes :) [18:03] dobey: right... something more generic like... [18:03] hum, food [18:03] but i don't have any better suggestions for that at the moment [18:03] will reply after having lunch, I'm starving :-) [18:03] * nessita -> brb food real quick [18:04] (inser that argentine saying that == bon appetit, but which i can't remember how to spell) [18:04] buen provecho! [18:04] that one [18:10] I will have lunch too, since it seems it's the IN thing to do === ralsina is now known as ralsina_lunch [18:24] ok, i'm back [18:24] uploading u1client to ubuntu now... [18:31] briancurtin, I'm trying to debug the "sd not stopping" problem on windows. [18:32] ralsina_lunch: fabulous models all over the world do it, even [18:32] briancurtin, I'm having a hard time getting this branch to run. Probably my deps are outdated [18:33] briancurtin, is there a wiki with instructions on how to run the windows dev env? [18:33] alecu: none of them have changed since i started. there this wiki page (https://wiki.ubuntu.com/UbuntuOne/Contribute/WindowsTesting) but it doesn't list external deps [18:33] alecu, ralsina_lunch: new u1client (with alecu's patch) is now released to Ubuntu, so is the new controlpanel with the missing dep [18:33] dobey: ^ [18:34] dobey: so, is helpers a too generic module name for this? [18:34] nessita: awesome, great work [18:34] alecu: i can give you some tips to setup via the buildout environment that i'm currently using [18:35] nessita: i think so. most of the things in devtools are helpers. and we probably want to put others into their own modules as well, depending on what they do, and how much code they are [18:36] nessita, awesome. [18:36] dobey: what about ubuntuone.devtools.recording ? [18:36] briancurtin, trying the buildout sounds interesting. [18:36] nessita: vs ubuntuone.devtools.testing.recording? === ralsina_lunch is now known as ralsina [18:37] alecu: it's a bit messy since it involves a number of branches, but let me get the directions sorted out. i'll let you know in a few minutes [18:38] nessita: can you explain a bit more what it does exactly, beyond what the docstrings say? i'm not entirely sure i fully understand what it's supposed to do. maybe if i understand it better, i'll have a better suggestion :) [18:42] dobey: ack. Before I meant: ubuntuone.devtools.testing.recording. The Recorder class is a class that have mainly 2 uses: [18:42] * to build fakes, so you can just inherit from it and build fake objects easily, that will know which methods were called on it [18:42] * to be used with the class decorator, add_recorder, so you can grab your existent class Foo, decorate it, and then every instance of Foo will behave like before (with the same side effects and all), but every call (with its params) made to it will be recorder (so you can assert over that in a test for example) [18:43] nessita: ok, that helps. let me think about it for a bit, and i'll get back to you. thanks. :) [18:44] dobey: thank you [19:06] ralsina: would you have any qt-magic slot left? [19:06] alecu: try this https://pastebin.canonical.com/63065/ [19:06] nessita: sure, but I also have a killer headache, so magic may be low [19:07] nessita: but let's try :-) [19:07] ralsina: you in linux? [19:07] nessita: yes, currently [19:07] i need some fresh air before i burn this apartment down...the building is now for sale so i'll have to move soon :/ [19:07] ralsina: pushing a branch then [19:07] be back in a bit [19:08] ralsina: so, remember the issue that mandel has regarding dialog.exec_() freezing? [19:08] nessita: yes [19:08] briancurtin, I'll take a look, thanks. [19:08] briancurtin, that sounds like really bad news, right? [19:08] ralsina: so, can you branch lp:~nataliabidart/ubuntuone-control-panel/handle-credentials-error [19:09] nessita: branching... [19:09] alecu: its not like we have to leave immediately, but soon they'll kick us and the other tenants out [19:09] ralsina: after branching, please remove your current device, and please make the /usr/lib/ubuntu-sso-client/ubuntu-sso-login-qt fail to start (quickest way if editing the __main__ and add a failure) [19:09] nessita: ack [19:10] ralsina: so, after that, open controlpanel from my branch and click on signin [19:10] briancurtin, that sucks! :-( [19:10] nessita: the one in usr/lib? Doesn't u1cp try to open it from its own bin/ first? [19:11] ralsina: nopes for sso [19:11] oh, it's sso [19:11] ralsina: sso itself will, but u1cp no [19:11] nessita: ooook, getting a bit dzzy. Ok, trying it now [19:12] ralsina: so, after clicking on signin, since the sso service can not open the qt UI, it will fail, and my branch is supposed to fix that. The problem is that when running exec_ on a dialog that shows the error, I get bad things happening [19:12] ralsina: using exec_ will freeze the app; using .show() + connecting to finished will trigger a seg fault [19:12] nessita: ok, give me 1' [19:12] bad news: this is happening for every error that is raised and catched by the handle_errors decorator [19:12] :-/ [19:13] nessita: ok, got the segfault [19:13] nessita: give me 15' to see what's going on, then [19:13] ralsina: yes, because the exec_ is commented out [19:13] nessita: where is u1cp is this code? [19:13] ralsina: you can dig into ubuntuone/controlpanel/gui/qt/__init__.py to play with the exec_ bits [19:13] nessita: ack [19:14] nessita, btw: I've taken a look at the "set_transient_for" issue from Qt that you mentioned last week. I have not found a clean way to do it in Qt, other than using some 3rd party library like python-xlib, which I'm not sure if we have available on the CD. [19:14] alecu: we can probably use xcb and ctypes [19:15] nessita: box is deleted before the deferred is triggered [19:15] nessita, ralsina: since this is a hint for the wm we might try some of those other ways, and if it's not available we can safely ignore it. [19:15] ok, I need to pick amelia at kinder [19:15] see you in a while! === alecu is now known as alecu_kinderrun [19:16] nessita: so, try usual hack for thing getting dropped :-) [19:16] ralsina: why the box is deleted? [19:16] ralsina: and why exec_ does not work? [19:16] nessita: because we reach the end of the function where it's defied [19:16] ralsina: there is a yield before that [19:16] ralsina: the end is not reached, see the print with the '3' in it [19:17] nessita: I do get the 3 printed before the segfault [19:17] ralsina: the '3' is printed before the seg fault [19:17] ralsina: exactly, so the method finishes when the box was closed [19:17] nessita: keep in mind that signals will be triggered by entering the event loop, which we don't do until after this function ends [19:17] ralsina: I think that's not correct... no? [19:17] nessita: easy to check! [19:17] ralsina: why the event loop will not be re-entered? [19:18] nessita: it only is entered after you leave this function, or call processEvents() [19:18] ralsina: why? [19:18] nessita: because that's how it works [19:18] nessita: how would it get started while you are in this function? [19:18] ralsina: the event loop is already running [19:18] no? [19:19] nessita: no, it's waiting for this to end [19:19] nessita: it's in this thread, so it can't be running [19:19] nessita: if I make box "global" it doesn't segfault [19:19] ralsina: but I'm yielding, so the main loop can keep working [19:20] nessita: but we are not using qtreactor [19:20] the event loop will wait until you return, I think [19:20] ralsina: but deferred integrates with any mainloop (that's my understanding) [19:20] nessita: no idea there [19:21] ralsina: otherwise all our calls to the backend would block the UI, and we're not blocking the UI [19:21] nessita: good point [19:21] ralsina: I'm 99.99% sure the deferred mechanism integrates with gtk and qt mainloops [19:21] even if there is no reactor running [19:21] nessita: could you check if the callback is getting called before this function ends? [19:21] if it's not, then that's why it segfaults [19:21] if it is, then I don't know [19:22] ralsina: the callback is being called befire this function ends, but the control is being released to Qt main loop on each yield [19:22] the function inner is a chain of callbacks... so that can be confusing [19:23] nessita: there is *some* (no idea what) event that is being called after the function ends that has a reference to box [19:23] I think I understand what's happening now [19:23] nessita: that's why it segfaults. I know no more than that from this debugging [19:23] ack, thanks!! [19:23] nessita: pleasure! [19:23] ralsina: can I disconnect a callback somehow, from a qt signal? [19:23] nessita: sure [19:23] nessita: whatever.signal.disconnect(callback) [19:24] nessita: keep in mind that *may* give an exception if it's not connected [19:24] ack [19:29] ralsina: how did you make the box global? I'm doing exactly so, and I'm getting: [19:29] QWidget: Must construct a QApplication before a QPaintDevice [19:29] Aborted (core dumped) [19:30] nessita: I did a globals=[] and then globals.append(box) [19:30] aaahhhh [19:30] the globals, of course, in global scope [19:31] ralsina: on the other hand, did you have any progress on getting the Hide details button not growing like crazy? [19:31] nessita: no, got sidetracked [19:31] ack [19:31] nessita: and now I am behind on perf. reviews, so... [19:32] nessita: should be just adding a qss rule for that button that doesn't have the hack in it [19:33] ralsina: any idea how to identify that button? [19:33] ralsina: the button is not "ours" [19:34] nessita: is it in a QMessageBox ? [19:34] ralsina: yes [19:34] nessita: then using a nesting selector, I think [19:34] QMessageBox > QPushButton is, IMHO, too general, no? [19:34] * briancurtin back [19:34] nessita: no, not really [19:34] I would like to say 'the show/hide details button' [19:35] nessita: hmmmm would have to check the sources to see if it has a name :-) [19:35] right [19:35] nessita: looking... [19:36] nessita: it's QMessageBox.detailsButton but doesn't seem to have an objectName :-( [19:36] ralsina: detailsButton is the class name? [19:36] nessita: no, the instance name [19:36] ah... [19:39] nessita: you could do box.detailsButton.setObjectName("something") but not sure that would work [19:40] nessita: it's a private attribute, so it's just QMessageBox > QPushButton. Or, if you name box, QMessageBox#thename > QPushButton === alecu_kinderrun is now known as alecu [19:47] nessita, ralsina: deferreds work right with any kind of mainloop; they don't need any reactor. [19:47] mixing a reactor with a mainloop is what makes things explode usually. [19:48] right [19:58] ralsina, the unicode issues seems to be REALLY difficult to reproduce...... if i run the installed backend, but executing cp from sources it works.... i'm trying to reproduce it creating the .exes because the problem seems to be related somehow to that [19:59] gatox: ack [19:59] gatox: good luck! [19:59] ralsina, jeje thanks [19:59] ralsina, it was just to let you know [20:03] ralsina: I'm defining a style for QMessageBox > QPushButton and as far as I debug is not applied... I confirmed I have no typos. I'm using this: [20:03] 370 QMessageBox > QPushButton { [20:03] 371 background: red; [20:03] 372 border: 2px solid green; [20:03] 373 } [20:03] ralsina: the same settings applied to a QMessageBox work [20:04] nessita: set it for ::focus [20:04] nessita: the ::focus selector is probably more specific and winning [20:04] ralsina: but the stylesheet is never applied [20:04] how come it will work on focus? [20:04] nessita: the more specific selector is always applied [20:05] ralsina: is this correct? [20:05] 370 QMessageBox > QPushButton:focus { [20:05] (is not working either) [20:05] ralsina: my guess is that the pushbutton is not a child of the messagebox [20:05] not sure if ">" means child [20:05] or can be deeper in the hierarchy [20:06] right [20:06] > means "direct children" [20:06] nessita: let me check the sources, then [20:07] nessita: insert a QWidget in between those [20:07] * nessita throws up and inserts a qwidget [20:08] nessita: it should be a QButtonBox but I am not sure [20:08] ok, this (kinda) works 370 QMessageBox > QWidget > QPushButton { [20:08] will keep playing [20:08] ralsina: thanks! === jml_ is now known as jml [20:09] nessita: it's a QDialogButtonBox if you wat to be specific [20:09] nessita: I think :-) [20:09] ralsina: may I adopt the saying "con más hacks terribles que ralsina"? :) [20:09] briancurtin, alecu: I am about to slip the windows release [20:09] nessita: no, because I do them on request ;-) [20:09] lol [20:10] alecu: have you been able to take a look at the evil undead syncdaemon? [20:10] briancurtin: have you had any luck with the evil poisonous TunnelRunner ? [20:10] alecu, briancurtin: I depend on your ETAs to decide how much I slip [20:10] ralsina: i have an idea but i'm working through the tests now [20:10] briancurtin: ack [20:17] aquarius: ping [20:17] pong [20:19] aquarius: would you or someone be interested in coming to ireland for a conference? speaking on U1? [20:19] aquarius: same conference a before only this time we're 5 years older? [20:19] czajkowski, aha, yes, I have a bit set to respond to your tweet about this :) [20:20] when is it>? [20:20] aquarius: october date to be finalised once we have key speaker [20:20] czajkowski, I can do it [20:20] hopefully early october and not towards the end [20:20] lovely [20:20] czajkowski, do you have some sense of attendee count? [20:20] aquarius: I'm in planning mode [20:20] so no idea as of now [20:20] but I do have UUPC coming :) [20:21] and am now trying to help the new guy sort out funding [20:21] same campus, same saturday night dinner, and hopefully same hotel [20:21] people, EOD here..... see you tomorrow...... i hope with some idea of how to reproduce this unicode issue :P bye! [20:22] czajkowski, pencil me in [20:22] sweet [20:22] aquarius: know any arm folks in UK/EU ? [20:22] folks who'd enjoy my kinda conference? [20:22] czajkowski, I do not, but the arm team at canonical will :) [20:23] where does one find them ? [20:27] czajkowski: you might want to ask about arm people in #linaro perhaps :) [20:27] I asked in canonical channel fell on deaf ears [20:27] but will try thanks [20:27] ask on freenode #linaro [20:28] and probably during normal hours if you're looking for people already in uk/ie :) [20:29] dobey: thanks [20:29] though yeah, timing might be bad [20:35] dobey: you know if you were closer I'd invite you :) [20:36] czajkowski: what converence is it? [20:36] although, october is just bad timing, too :) [20:37] Skycon it's a conference to mark my unis computer society birthday :) skycon.skynet.ie/2007 was the last one I was involved in [20:37] which is how I got to meet the charming aquarius [20:37] czajkowski, amber's a linaro person now [20:37] nods [20:39] czajkowski: has there been one since 2009? [20:40] nope 2007 was our 15th and then one of the guys thought he should run one to keep the name alive which was kinda hit and miss but was a good idea [20:40] so now 2012 is the year we're 20 :) [20:42] ah [20:42] might want to get a basic web site up that doesn't point at 2009 :) [20:43] that'll be changing when 2012 goes live [20:43] kinda annoying alright [20:52] ralsina, hola. I've been banging my head at making my branches work right. [20:52] alecu: ack [20:53] ralsina, I can run all projects on my dev environ, but SD does not start listening for ipc [20:54] ralsina, and sso seems to not be listening to ipc either. [20:54] ralsina, so I'm trying with the buildout instructions that briancurtin gave me. [20:56] alecu: you trying to start syncdaemon how? [20:56] alecu: same question for ussoc [20:56] nessita: [20:56] set PYTHONPATH=.;e:\ubuntu-sso-client\trunk;e:\ubuntuone-client\trunk;e:\ubuntuone-control-panel\trunk [20:56] set TRIAL_TEMP_DIR=c:\temp\tritritridir [20:56] set DEBUG=True [20:56] set U1_DEBUG=True [20:56] python bin\ubuntu-sso-login [20:56] and in the other folder [20:57] python bin\ubuntuone-syncdaemon --debug [20:57] alecu: have any trace? [20:57] nessita, no stacktrace in the logs nor on the console. [20:57] nessita, it just blocks. [20:57] let me try here [20:57] nessita, let me find the exact log [20:57] alecu: after this snapshot finishes merging with another one :-/ [20:58] doesn't PYTHONPATH have to use : instead of ; on windows, because of some silliness? [20:59] dobey: is the other way around [20:59] it uses ; [20:59] : on Linux, ; on windows [20:59] nessita@dali:~$ echo $PATH [20:59] /home/software/tarmac/bin:/home/software/mc/bin:/home/nessita/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games [20:59] i know ; is the standard on windows [20:59] but i recall you guys discussing PYTHONPATH or something have a problem where ; wasn't working right and you had to use : instead [21:00] dobey: that was "just" for pylint [21:00] ah ok [21:00] dobey: pylint expects : pythonpath separator on windows [21:00] which is so buggy [21:00] right [21:01] so it was pythonpath; i just forgot the details [21:01] thanks [21:06] ralsina, briancurtin: I now know why my SD is not listening for IPC. [21:07] it's blocked while trying to start the tunnelrunner :P [21:07] dobey: anytime! [21:08] alecu: what about the blocked ussoc? [21:09] nessita, my guess is that -login is trying to open -login-qt, but there's no .exe [21:10] alecu: very likely [21:10] alecu: I know, I told you that on friday [21:11] ralsina, you should have remind me today! [21:11] alecu: I did, I think (looking for logs) [21:11] ralsina, briancurtin: is there any way to have windows be smart and read the scripts hashbang? [21:11] nessita: o [21:11] nessita: No [21:11] nessita: https://bitbucket.org/vinay.sajip/pylauncher - we're including this in 3.3 [21:11] So, no ;-) [21:11] * nessita just realizes how naive she is when she asked "is there any way to have windows be smart" [21:12] ralsina, I know you mentioned that was broken too. But I did not realize it was blocking *this*. [21:12] * alecu brbs [21:12] alecu: "alecu: could you (if you have a slot) check why syncdaemon doesn't stop on windows? I am at a loss there." followed by "alecu: you may have to fix the "not starting proxytunnel causes syncdaemon to go autist" bug first" [21:12] alecu: but yes, I should have been more specific [21:13] i'm making TunnelRunner be smarter about failing on startup, i have a few cases figured out [21:13] ralsina, briancurtin: can we use pylauncher in our dev envs to make a python script be also an executable? [21:14] nessita: pylauncher is effectiely a shebang processor. it'd be nifty but i dont know that it solves any of our problems around exes [21:14] ralsina, someone else: may I have reviews https://code.launchpad.net/~nataliabidart/ubuntuone-control-panel/handle-credentials-error/+merge/99420 [21:14] briancurtin: I would like to spwan a process in windows using a python script (foo.py) and have that being the same as calling "python foo.py" [21:16] nessita: gotcha, but probably won't finish before EOD [21:16] ralsina: is ok, tomorrow is just fine [21:16] nessita: it works if we call them .pyw I think [21:16] nessita: i think you could do that but i havent dug deep enough into pylauncher - i know it launches another thing itself, can't remember if its a process or a job object or something [21:17] briancurtin: ack [21:31] EOD for me, will probably do some extra tonight, so mail me if you need anythiing [21:31] ralsina: ack, have a good evening [21:37] briancurtin, you are being kicked out? wtf? the states is worse than I though.. [21:37] I'm gone to pilates class! [21:38] see ya crowd tomorrow [21:39] mandel: at some point i'll probably be kicked out, but it's a few months away so i'll try to leave before then. the guy who owns the building sucks at paying the bills so the bank is taking it away [21:40] briancurtin, the world is going down the drain.. the good thing is that you might find a better place.. or at least a better land lord :) [21:41] mandel: yeah i should be able to find a better place. this place is nice and big but also a bit pricey for being kind of crappy and old [21:42] i'm kind of hoping the land lord just runs away. i had a friend in the same situation who ended up living in his place for several months with no one to collect rent, then the bank paid him a few grand to move out so they could tear it down [21:43] later all [21:47] briancurtin, ha! that sounds like a bad movie hehehe In spains is similar, they cannot kick you out or at least is very difficult so people stop paying the rent until they are kick out which usually takes several years [21:50] mandel: i'm calling a friend who does bank/mortgage stuff to ask him if i can just stop paying my rent right now. i figure i'll move in 2 months, so keeping that 2 months rent would be nice...it's not like the owner is paying the bills anymore anyway [21:50] the guy already lost the building so if i pay him it's right into his pocket [21:51] briancurtin, as long as you don't get in trouble it does sound legit, I would do the same [21:52] rather than pay his beer fund, i'd rather pay my own [22:02] briancurtin, or mine for that matter :) [22:21] i'm out of here, have a good evening mandel [22:21] briancurtin, you too! good luck with the house!