=== spm is now known as stevemci === stevemci is now known as spm === spm is now known as stevemci === stevemci is now known as spm === spm is now known as stevemci === stevemci is now known as spm === zyga-afk is now known as zyga [07:28] morning all! [07:32] hi mandel :) [07:33] mmcc, what are you doing here! [07:33] mandel, I left you some nice long reviews, feel free to send me angry emails while I go sleep [07:33] mmcc, you could have sent me an email instead of waiting, you know that right? [07:33] hehehe [07:34] it's only just past midnight here, so I'm just up late catching up. don't get used to it :) [07:34] heh [07:34] ok, actually going now. cheers [07:35] mmcc, good night! [08:41] Morning all! :) [09:19] tea break :) [09:52] * mandel back === yofel_ is now known as yofel [11:42] good morning! [12:33] ralsina, welcome back! [12:33] ralsina, I'm off to have lunch, wanna catch up later? [12:33] mandel: yes please [12:33] mandel: buen provecho [12:34] ralsina, thx, I'll be fast = heat food, eat and let the dog pee :) [12:34] mandel: no rush, don't fall on things, take care ;-) [12:42] mandel: take your time, I forgot I have a call now for about 30 minutes [13:21] ralsina, I'm back let me know whenever you are free [13:21] mandel: now is good [13:21] mandel: mumble? [13:21] ralsina, sure, launching the thing [13:22] ralsina, can you hear me? [13:22] mandel: no, and no red lips [13:23] ralsina, wierd, I hear you, I see my red lips, see yours.. etc.. [13:23] ralsina, ok [13:28] hmm [13:30] dobey, hi! [13:31] hi [13:33] hello all! [13:47] hola alecu! [13:47] alecu: thought you were on vacations! [13:48] ralsina: no, I only took yesterday off since Amelia is on winter vacations and nobody could babysit her. [13:49] alecu: cool! [14:00] alecu: I am in catchup mode, can we have our 1-1 in about 30'? [14:01] ralsina: sure [14:57] are we doing mumble or hangout? [14:58] I think I may have to skipit [14:58] * alecu readies his pirate captain hat [14:58] oh :-( [14:58] let's schedule it for tomorrow same bat-time, same bat-channel [14:58] sounds good. text standup for now? [14:59] oh, ok then [15:03] if not i'm just going to lunch then :) [15:03] me [15:03] Can't connect to mumble here, is it going on? [15:04] mmcc: doing it tomorrow [15:04] mmcc: so you have the next 24 hours to figure it out...or else [15:04] briancurtin: Aha, thx. [15:04] i guess that means we're skipping standup too then [15:05] since nobody else is saying anything? [15:08] dobey: EOM before it even started, i guess. enjoy lunch [15:08] :) [15:12] erm [15:12] ok, lunch time [15:15] briancurtin, ralsina, alecu, dobey, stop porting things to python3 and port the to ChinesePython: http://www.chinesepython.org/home.html [15:15] if you are lazy, someone is translating python in chinese, as 'import sys' is '載入 系統' [15:19] WTF (it would be a WTF in any language other than english) [15:20] mandel, this reminds me of a portuguese version I got for Martin Fowler's "Refactoring" book [15:20] instead of "getPrice", for example, it was "lerPreço". Pretty crappy. [15:20] i like that they're still using the super old python.org website look [15:21] diogobaeder, yes, they should not translate IT books.. ins spain is very common, horrible, horrible.. [15:22] mandel, +1 [15:24] mandel, everytime someone asks me if there's a portuguese version for an IT book originally in english: http://is.gd/MB9tic [15:25] lunch time === diogobaeder is now known as diogobaeder-lunc [15:38] mandel: the "latests news" are from Feb 2004 [15:38] mandel: and the latest release is based on python 2.1.3... [15:39] mandel: so, we should keep porting to python3, that already supports all kinds of unicode identifiers [15:40] alecu, but but.. we could get so many community contributions from china! I know is the only reason why they are holding back.. [15:40] (and for mandel and diogo's happiness, non-english books with translated identifiers won't be exclusive to Java)! [15:42] "cinco de mayo".índice(3) [16:05] mmcc, I updated https://code.launchpad.net/~mandel/ubuntuone-fsevents-daemon/remove-arc-dispatcher/+merge/116434 and https://code.launchpad.net/~mandel/ubuntuone-fsevents-daemon/remove-arc-commsocket/+merge/116439 [16:06] mmcc, will get https://code.launchpad.net/~mandel/ubuntuone-fsevents-daemon/remove-arc-commsocket-server/+merge/116440 done later [16:06] ralsina, alecu, did we completely forget about the stand up? [16:07] mandel: ok, looking at dispatcher and commsocket now [16:08] herb, i just read the backlog, no standup :P [16:11] mandel: not important but for future reference, -retain returns self, so you can do foo = [inFoo retain]; [16:12] and interestingly, they've now marked all those ref counting methods as "Obsolete" [16:13] mmcc, since when? [16:13] mmcc, oh, and one question, you still have the 10.6 machine around, right? [16:14] mandel yes I have it around, why? [16:14] mandel: I'm not sure when they changed the docs. Might even have been yesterday [16:14] mmcc, because is the only one we have to test it work in 10,6 :) [16:15] anyway, obsolete != deprecated, just discouraged [16:15] crap. LP is still showing prerequisite changes in this diff. that's helpful... [16:15] mandel: right. for arc removal, I'll have to test with the last branch though since the test bundle doesn't link unless everything is non-arc [16:15] mmcc, yeah, lets keep moving this way but not in a rush, ralsina will talk in the techleads meeting about if we should ignore 10.6 or not [16:16] mmcc, yes, I know, but I though doing all these mini branches was better for the reviewers :) [16:16] mandel: no, it's good - I'm not complaining [16:17] :) [16:17] herb, we are old schoole, look what twisted has now: from twisted.internet import inotify [16:18] sorry herb, ignore the last message [16:18] * mandel curses at his tab [16:19] mandel: do you just hit tab after every word? [16:20] mmcc, no, but in this case I did for some reason, I wanted just to do 'he' [16:20] I don't understand how I work sometimes.. [16:20] heh. [16:21] mandel, mmcc: I'm thinking of updating to 10.8. Any reasons I shouldn't? [16:22] alecu, bugs, but it would be nice to have an intrepid soul so that we run our tests there :) [16:22] alecu, I might wait 'til 10.8.1 [16:22] alecu: fear of the unknown? [16:22] but mandel pointed out you can do dual boot if you want [16:23] so far I've been hearing that 10.8 is reasonably stable, compared to 10.7.0 [16:23] ok, here I go. [16:23] btw: this is just awesome: [16:23] Twitter is currently down for <%= reason %>. [16:23] We expect to be back in <%= deadline %> [16:23] https://twitter.com [16:23] heh, I saw that too. [16:24] alecu, maybe the use gtalk as a backend.. [16:24] mandel: lolz [16:27] if youtube and reddit go down the internet might be as productive as it has never been before.. [16:35] mandel: just read your comment in remove-arc-commsocket about self.sockRef being a property. You're right of course, sorry about the confusion. Another reason I don't like properties and dot syntax in ObjC - it confuses old timers like me :) [16:35] mmcc, no problem, I had to think several times about it [16:36] mmcc, lets remove them as a design decision in a later step :) [16:36] yeah, my complaint is that there's no way to know when you're looking at an assignment if it's going to call a method somewhere or not. You have to go poke at the header file to see if it's a property and what kind of property it is [16:36] the code looks nice like python but you can't assume it does what you think it should:\ [16:37] mmcc, well in python you could also have a property, but having properties with side effects are not good and with arc they do [16:38] mmcc, as in, var is retained etc.. [16:39] right. in python you don't have to worry about what a property call is doing to memory management [16:41] yep yep [16:42] ok, EOD for me I have started rugby training already :-/ [16:42] mmcc, see you tonight if you are still up :) [16:43] heh alright mandel, don't count on it :\ [17:01] hmm, no way to just poke at what certs pyOpenSSL sees by default? I can get_cert_store but the X509Store object it gives me only lets me add [17:05] context - we are currently just shipping PEM files and loading them from somewhere (/etc/ssl on linux, Common AppData on windows). dobey says I shouldn't just put them in XDG_CONFIG_HOME, which makes sense since we're putting them in a global place on other platforms, but we can't write to the osx global config location without escalating privileges, which isn't OK in a sandbox… [17:06] we can add certs programmatically, but I wanted to have a way to see which ones we need to install that way [17:07] in the above, adding programmatically means saving them to the system keychain. [17:09] can we not just load them from the .app internally? [17:11] dobey: hmm. yes, I guess so. Why don't we do that on Windows? [17:13] mmcc: on windows we put them inside our app, more or less [17:14] ralsina: ok, I guess that's what Common AppData means, even though it's not actually with the other files? Definitely still catching up here on windows distribution :) === diogobaeder-lunc is now known as diogobaeder [17:15] mmcc: I suspect we were following some convention of other. That part was not done by a real windows developer :-) [17:15] (meaning, I did it) [17:16] ralsina: heh, ok. So if just loading the cert files from within the app's Resources file is OK, that's a nice easy fix [17:16] mmcc: yes, it's not meant for general usage anyway, so whatever works [17:21] it would be nice to use the standard system conventions though, and avoid shipping any certs at all [17:21] but, yay cross-platform is pain [17:21] ralsina: not meant for general usage? [17:21] mmcc: not for other apps [17:21] ralsina: oh, ok [17:41] * briancurtin pharmacy/lunch === mmcc_ is now known as mmcc [18:22] * briancurtin back [19:29] bbiab, gotta run an errand [19:30] just noticed that the recent filesystem_notifications refactoring is installing the default reactor too soon again. so I'm going for lunch now… [19:31] going to try to make it a working lunch [19:31] but will be ask for a bit getting there [19:31] s/ask/afk [19:58] mmcc argh [19:58] that bug will never die [20:10] ralsina: yep.. the design decision that keeps on givin' [20:59] ralsina: if only we could just drop twisted. et voila, no more default reactor :) [20:59] * dobey is having plenty of not-fun at the moment with twisted :-/ [21:01] * mmcc is right there with you [21:01] thanks to the fun of __import__ [21:01] dobey: if wishes were candy and nuts, we'd all have a merry christmas! [21:01] * ralsina knows that's not how that goes [21:02] if wishes were candy and nuts, there wouldn't be a christmas [21:02] dobey: it seems like you need to get in touch with a guy in korea: http://www.codinghorror.com/blog/2012/07/the-ips-lcd-revolution.html [21:03] yeah i know, right [21:03] oh wait [21:03] also, looks like I need to go parenting [21:03] those are 27" [21:03] dobey: but this is what you really really want: http://news.ycombinator.com/item?id=4298415 [21:05] well, on something that doesn't suck so much power, weigh 35 lbs, or be so thick [21:05] i want the new panasonic display, really [21:05] I knew someone who had one of those IBM suckers. He said it was hard to use anything else after he got used to it [21:06] it is [21:06] same reason i kept my SGI 1600SW for so long [21:06] and it's still sitting in the closet. totally need to sell it [21:10] give me OLED [21:12] anyone know how to make __import__ not break the world? [21:20] dobey: what's the symptom? [21:20] * mmcc promises nothing [21:21] http://pastebin.ubuntu.com/1112654/ [21:21] twisted imports not working from something which i'm __import__()ing [21:24] dobey: am I missing something or should that be globals(), locals() ? [21:24] mmcc: i don't know; but same error happens either way, or if they are None [21:25] hrm ok… [21:27] wild guess: is it a relative import name clash? is it looking for twisted.scripts in devtools.runners.twisted.py? [21:27] oh, maybe; but i changed level to 0 and still failed :( [21:28] hmm [21:29] indeed, renaming got past that issue [21:31] mmcc: thanks [21:31] dobey: hey, sure. always glad to help with some wild guessing [21:36] :) [21:37] alright, off for now. have a good evening all [21:45] for the u1client tests, does trial just run each test in a thread? is there a thread pool with a set size? [21:47] I ask because I'm running the full test suite and it's dying with 'too many open files' right around when it's launched about 250 threads, which makes me very suspicious of a number of things [21:49] that is, many hundreds of the tests fail because of errno 24, too many open files. [21:55] in other news this confused me: from ubuntuone.platform.filesystem_notifications.monitor import ( [21:55] linux as filesystem_notifications [21:55] ) [21:55] because I first saw filesystem_notifications.FilesystemMonitor in the code and couldn't figure out where that was coming from. [22:05] ok, I've fixed that reactor issue, need to test it on windows to be sure, then I'll propose it [22:05] time to go for now