[00:28] <chaselivingston> ping ralsina: any update on a new version of the mac app?
[00:28] <chaselivingston> ping mmcc: any update on a new version of the mac app?
[00:29] <ralsina> chaselivingston: not today, early tomorrow
[00:30] <chaselivingston> ralsina: awesome, will it be sent out via email like the first one?
[00:31] <ralsina> chaselivingston: yes
[00:31] <chaselivingston> ralsina: sweet. thanks for all you guys' hard work on this!
[00:31] <ralsina> chaselivingston: it's our job :-)
[00:32] <chaselivingston> ralsina: haha, well you're doing a darn good job, can't wait to see the app released to the public
[00:32] <ralsina> chaselivingston: me too
[00:33] <ralsina> chaselivingston: the problem with ports like this is that we get little feedback until it's like 95% done because it just doesn't work at all before that
[00:33] <chaselivingston> ralsina: yeah, that makes sense. trying to give you guys as much as I can :)
[00:33] <ralsina> chaselivingston: and it's much appreciated!
[00:33] <chaselivingston> ralsina: good, sometimes I feel like I'm a little annoying with it ;)
[00:33] <ralsina> chaselivingston: not at all!
[00:34] <ralsina> mmcc: you don't by chance have a bundle to give chaselivingston?
[00:34] <ralsina> mmcc: unofficially :-)
[00:35] <ralsina> damned, 9:30PM and pizza is here! Gotta run ;-)
[00:36] <chaselivingston> ralsina: enjoy!
[06:08] <mmcc> welp, not a simple retain / release bug.
[07:46] <mandel> morning all!
[07:47] <mandel> ralsina, mmcc, you should not be using the root logger from syncdaemon.. agh! there was a reason I did not do it that way!
[07:47] <mmcc> mandel: oops, so why didn't it work the way you did it? where were those messages going?
[07:48] <mandel> mmcc, mainly because is using the wrong namespace, had I used ubuntuone.syncdaemon it would have gone to the correct file, not big deal, is an easy fix :)
[07:49] <mmcc> mandel: it didn't go to any file that I could see…
[07:49] <mandel> mmcc, yes, that is because there is a general config that will pick the loggers with the namespaces, because I forgot to use the correct namespace it was going no where
[07:50] <mmcc> aha.
[07:50] <mandel> mmcc, I good example is in ubuntuone/platform/filesystem_notifications/monitor/windows.py
[07:50] <mmcc> what problem is caused by using the root logger as it is in trunk now?
[07:50] <mandel> mmcc, anyways, not a big problem and I know you were in a hurry so it makes sense you did a quick fix
[07:51] <mandel> mmcc, the main concern is that platform is imported from other projects and there is no guarantee that the syncdaemon code wont install a reactor
[07:51] <mandel> mmcc, is better to keep things clean in that way
[07:52] <mmcc> oh, but platform/filesystem_notifications will only ever be imported from syncdaemon, right? I just wrote a test script that fails if it detects that…
[07:52] <mmcc> for that same reaosn
[07:52] <mmcc> reason
[07:53] <mandel> mmcc, it should only be imported from there yes, but we don't know, also, all other packages use the same style I used
[07:53] <mandel> mmcc, anyway, is a stupid style problem, nothing else and it requires like a 2 line diff
[07:54] <mandel> mmcc, ignore it, specially because is late there, did you manage to build the bundle with the stale broker branch?
[07:54] <mmcc> mandel: no, not yet. will do after I sleep… I never reproduced that issue, btw
[07:55] <mmcc> real quick - in fsevents_daemon.py you have logging.getLogger('ubuntuone.SyncDaemon.FSMonitor') -- should that be going to syncdaemon.log?
[07:56] <mmcc> I can't confirm that add_watch is getting called in the daemon implementation of FilesystemMonitor
[07:57] <mmcc> I finally got the launchd-launched daemon up and running, had a crash that took forever to figure out, and now it's running and getting events but apparently the daemon client code in syncdaemon isn't registering any watch paths
[07:57] <mandel> mmcc, yes, that is the trick
[07:57] <mmcc> you mean, yes it should be going to syncdaemon.log? because then something's crazy, since a log in add_watch is not showing up
[07:58] <mandel> mmcc,  you should be getting the logs, can you send me the code you are working on and I'll debug that for you?
[07:58] <mandel> mmcc, what was the crash problem?
[07:59] <mmcc> you were sending 'self' in the CFSocketContext inside a class method, where self is the class, not the instance you just created…
[07:59] <mandel> ouch, stupid me
[08:00] <mmcc> it was hard to figure out because the code was duplicated in CommSocketServer and FSEventsCommSocketServer, and fixing it in CommSocketServer didn't fix it, but I couldn't figure out why
[08:00] <mandel> mmcc, do you have a branch for that so I can fully understand it?
[08:01] <mmcc> not yet
[08:01] <mmcc> need to look over the xcodeproj and make sure I haven't introduced weird unnecessary changes
[08:02] <mmcc> trying to figure out how to send you the code so you can debug
[08:02] <mmcc> it takes some explanation - had to make sbjson a static lib, and put it in the objc/ directory, along with its headers
[08:03] <mmcc> and you'll need to re-sign everything every time you change anything, which means you need a new cert
[08:03] <mmcc> let me try to clean things up and write an email, but I'm fading here …
[08:04] <mandel> mmcc, sure, either way, do not work til to late, is not worth it
[08:04] <mmcc> did you have a chance to try building the app bundle?
[08:04] <mandel> mmcc, yes but had some small problem with my env so I'm fixing that before
[08:06] <mandel> mmcc, I'll be trying to exaplin the reconnects but that is from the serve side of the ipc so the stale broker code is needed just in case we ever have a reconnect, the fact that we get the exception means that client and server are connected but the remote objects where gc in the server side
[08:09] <mandel> mmcc, if you still have energy this should work for you: https://code.launchpad.net/~mandel/ubuntuone-client/correct-logger/+merge/121553
[08:09] <mmcc> mandel: that's gonna wait until after sleep :)
[08:09] <mandel> mmcc, no problem :)
[08:09] <mandel> mmcc, have a good night!
[08:10] <mmcc> mandel: not yet, I'm sending you an email that'll let you debug this daemon interaction problem I'm seeing - just as long as you don't need to touch the daemon itself
[08:10] <mmcc> because that's the hard part to build
[08:11] <mandel> mmcc, sure, I can play around with it without changing it.. I have an idea of where the problem might be
[08:11] <mandel> mmcc, are you getting the logs stating that the fs monitor used is the daemon one?
[08:11] <mmcc> mandel: yes, that's working
[08:12] <mmcc> what I'm seeing is that when I have monitor set to 'daemon' in syncdaemon.conf , it appears to connect (the daemon prints something about a new user to the console)
[08:12] <mmcc> but I get no further action past the SYS_USER_CONNECT event in the syncdaemon logs
[08:12] <mmcc> no local rescan either
[08:13] <mmcc> if I use 'default' in the .conf, that still works fine
[08:13] <mmcc> without the daemon, obviously
[08:13] <mandel> mmcc, ok, if you do not get to sys_user_connect it means that we might be getting an exception, don't worry I'll take it from here I have seen this happening before
[08:16] <mmcc> it does get to there, but I don't see any exceptions anywhere
[08:16] <mandel> mmcc, probably twisted is hiding it in a deferred
[08:18] <mmcc> whee
[08:18] <mandel> mmcc, yes, so go to bed, I'll deal with it :)
[08:19] <mmcc> ok, ok - look in your email for info on how to tweak the bundle I put in the u1 mac client builds share just now
[08:20] <mmcc> (and before I go to sleep, do you see that bundle?)
[08:21] <mandel> mmcc, was it attach to the email?
[08:21] <mandel> mmcc, I have not file attached in it
[08:22] <mmcc> mandel: no, it's in the u1 mac client builds shared folder on u1
[08:22] <mandel> mmcc, let me check
[08:22] <mandel> mmcc, is it u1-mac-client-sign.. etc?
[08:22] <mmcc> yes
[08:23]  * mmcc out
[08:23] <mandel> mmcc, ok, rest well o/
[08:48] <JamesTait> Good morning all! :)
[08:58] <ralsina> good morning!
[09:01] <lifeless> require_gather_details = skipIf(gather_details is None,
[09:01] <lifeless>         "gather_details() is not available.")
[09:01] <lifeless> bah
[09:02] <lifeless> sorry
[09:08] <mandel> ralsina, good very early morning!
[09:08] <mandel> ralsina, review for https://code.launchpad.net/~mandel/ubuntuone-client/correct-logger/+merge/121553
[09:08] <ralsina> hi mandel. I see I just missed mmcc :)
[09:08] <ralsina> mandel: sure!
[09:08] <mandel> ralsina, we should not use the root_logger
[09:09] <ralsina> mandel: I never quite understood our logging
[09:09] <ralsina> mandel: not that I tried very hard
[09:09] <mandel> ralsina, also, we do the following http://paste.ubuntu.com/1171403/ in our code, which means that the app goes nuts in debug mode when you try to sync the share from lissette
[09:09] <mandel> ralsina, is hard.. and a PITA
[09:10] <ralsina> mandel: I found out os_helper.walk fails on invalid utf8 filenames in ubuntu, so those things may be more broken than you think
[09:10] <ralsina> mandel: oh, it breaks because of the %
[09:11] <mandel> ralsina, yes.. I need to look for all the bloody log messages and fix it
[09:11] <ralsina> mandel: ack
[09:11] <mandel> ralsina, or at least for the one I know it breaks
[09:11] <ralsina> mandel: debugging log messages is hell
[09:12] <mandel> ralsina, just found because I was running the app with U1_DEBUG..
[09:16] <ralsina> mandel: global+1 on the logger branch
[09:16] <mandel> thx
[09:16] <mandel> ralsina, fixing the logging problem and will push it
[09:17] <mandel> ralsina, regarding the stale broker, do I really need to find out why twisted reconnects?
[09:18] <pedronis> rye: maybe you can with a web-op revert the code in one of the ec2 slaves and see if it makes a difference, but it seems the conditions of that bug are complicated...
[09:20] <ralsina> mandel: it would be a nice to have, but not for today
[09:21] <ralsina> it would be nice since reconnects probably hurt ur UI responsiveness
[09:21] <ralsina> but not today because we are in a hurry
[09:23] <mandel> ralsina, then here you go: https://code.launchpad.net/~mandel/ubuntuone-client/stale-broker/+merge/121456
[09:23] <mandel> ralsina, I realized that the same has to be done in sso for the windows port, will do that asap
[09:24] <ralsina> mandel: really no rush about fixing widows bugs, we are not doing windows releases
[09:25] <mandel> ralsina, I know, but else I will forget and will never be fixed :)
[09:31] <ralsina> mandel: basically, we should not have any " % " in any line that also has "log" in it, right?
[09:31] <ralsina> mandel: "else I will forget" is why we have a bug tracker ;-)
[09:32] <mandel> ralsina, yes, never ever have that, yet there are some parts where the message is generated using another message
[09:32] <mandel> plus a funny log decorator..
[09:32] <mandel> ralsina, and yes, I'll file the bung in lp
[09:54] <mandel> ralsina, somehow I managed to see sd crash which explains the stale brokers..
[09:54] <mandel> ralsina, now I just need to understand paste.ubuntu.com/1171475
[09:57] <ralsina> mandel: I don't think that contains any useful information
[09:57] <mandel> ralsina, nop, nor our logs..
[09:58] <mandel> ralsina, but you know I have a webcam hehe
[09:58] <ralsina> mandel: eh?
[09:59] <mandel> ralsina, the the bottom of the pastebin, their report include all the hardware
[09:59] <mandel> very useful info when u1 crashes
[09:59] <ralsina> he
[10:03] <mandel> ralsina, I think I found the culprit fsevents: Fatal Python error: GC object already tracked
[10:03] <mandel> cute..
[10:13] <ralsina> mandel: +1
[10:23] <ralsina> breakfast!
[11:45] <Marcinn> +c
[12:12] <alecu> hello, all!
[12:39] <mandel> lunch time for me
[12:40] <mandel> alecu, hile I'm away, can you do some testing for me?
[12:41] <mandel> alecu, can you run sd from trunk and accept lissetes design share, the app should crash quite early
[12:41] <ralsina> hello again!
[12:42] <alecu> mandel: on osx or on q or how?
[12:42] <alecu> hello back, ralsina
[12:42] <mandel> alecu, os x AFAIK
[12:43] <mandel> alecu, but make sure you set U1_DEBUG=1
[12:43] <mandel> ralsina, same if you can please ^
[12:43] <ralsina> mandel: that bug doesn't sound mac-specific at all
[12:43] <mandel> ralsina, I have been all morning on the mac, feel free to test it on other platforms
[12:43] <ralsina> mandel: will do it on precise
[12:44] <mandel> it has to be in debug mode, is very important
[12:44] <ralsina> mandel: sure
[12:44] <ralsina> mandel: what will crash u1cp or sd?
[12:44] <mandel> ralsina, sd
[12:44] <ralsina> mandel: ok
[12:44] <mandel> ralsina, with that GC object already tracked error..
[12:44] <ralsina> mandel: ok
[12:45] <mandel> ralsina, alecu, besides the logging failing because it has a % in the path (I think I found all occurrences of that) you should also get that or similar
[12:45] <mandel> I'm off to get sushi and I'll be back with more energy :)
[12:58] <ralsina> I am getting weird IPC errors. I am rebooting just in case.
[13:12] <ralsina> grmbl unity-3d froze my mouse :-(
[13:12] <ralsina> restarting the session
[13:13] <ralsina> then again, how dows one logout without a mouse?
[13:14] <ralsina> ah, unlock
[13:18] <ralsina> alecu: look at this error I am getting: (process:8202): libindicate-ERROR **: Unable to get session bus: Error spawning command line `dbus-launch --autolaunch=3ca1568960a3745bf6fc17df00000018 --binary-syntax --close-stderr': Failed to fork (Cannot allocate memory)
[13:30] <alecu> ralsina: that looks... weird
[13:30] <ralsina> alecu: only happens on unity, too
[13:30] <alecu> ralsina: who's giving that error?
[13:30] <ralsina> alecu: syncdaemon
[13:30] <alecu> ralsina: under virtualbox? on q?
[13:30] <ralsina> alecu: on P, real box
[13:31] <alecu> ralsina: u1 nightlies?
[13:31] <dobey> brb
[13:31] <ralsina> alecu: trunk
[13:33] <alecu> ralsina: I don't get that on trunk on a fully up to date P...
[13:34] <ralsina> alecu: I may have more files than you. sd goes up to 1GB memory usage before failing
[13:34] <alecu> ralsina: does SD start at all, or does it break there?
[13:34] <ralsina> alecu: it starts, then does that and dies
[13:34] <ralsina> alecu: may be because of sni-qt, also
[13:35] <alecu> ralsina: ah... I don't have sni-qt, I think.
[13:35] <alecu> ralsina: top says my SD uses 567m 125m  20m
[13:35] <ralsina> mine is 1g 231m
[13:35] <ralsina> but I have many dozen thousand files
[13:36] <dobey> not quite yet
[13:37] <ralsina> alecu: now I am not on unity, and things work much more smoothly
[13:38] <ralsina> although all gtk apps look like gnome 1.0
[13:39] <alecu> ralsina: try starting gnome-settings-daemon!
[13:39] <ralsina> alecu: oh, thanks!
[13:45] <dobey> now, brb
[13:55] <mmccphone> Hi team
[13:55] <mmccphone> mandel: Any luck with the daemon?
[13:58] <mmccphone> I'm out strolling the baby, who woke up when I went to sleep, and tortured mom until 6, so I'm afk for another 20 min or so
[13:58] <mmccphone> But I will get the reconnect fix into a build as soon as I get home
[14:15] <mandel> mmccphone, I have been stuck with a crash of sd..
[14:15] <mandel> alecu, ralsina did you manage to reproduce the crash?
[14:16] <mmccphone> When is it crashing? I saw you mention the bad logging format string caused by lisettes share
[14:16] <mmccphone> I noticed that once but couldn't reproduce it myself
[14:18] <mandel> mmccphone, I've noticed it crashing when you accept a share that has % paths in it and U1_DEBUG is set.. but I don't get why it should crash
[14:18] <mandel> mmccphone, it also returns a nice cpython error regarding memory management, so I gave priority to this than to the daemon, sorry
[14:19] <dobey> oh, fml
[14:20] <alecu> mandel: does it look like this? 2012-08-28 11:20:36,437 - twisted - INFO - pause reads for: %s 0.0285714285714
[14:21] <alecu> mandel: it's not crashing, though
[14:21] <mmcc> mandel: is the cpython error connected to the share path?
[14:22] <mmcc> when I noticed the share path issue, it was just an exception, not  a crash (although it did block progress)
[14:22] <mandel> mmcc, it just happens in random places.. the exact error is  Fatal Python error: GC object already tracked
[14:23] <mandel> mmcc, it comes after a number or logging errors, in my system is easy to reproduce
[14:23] <mmcc> hmm, well let me know if I can help
[14:23] <mandel> alecu, did you use the this version of macfsevents: https://github.com/mandel-macaque/macfsevents
[14:23] <alecu> mandel: oh, I got a different error..
[14:24] <alecu> mandel: this is on Q
[14:24] <mandel> mmcc, can you try and reproduce the error, mainly run u1sd with U1_DEBUG and accept/download lissettes design share
[14:24] <mandel> alecu, ok, do you have the chance to try on a darwin box?
[14:25] <mmcc> mandel: ok
[14:25] <mandel> mmcc, awesome, thx!
[14:26] <alecu> mandel: I can try, but I need to fix dependencies on it first. That's why I'm trying to avoid it!
[14:26] <mandel> alecu, lol, ok
[14:26] <alecu> mandel: http://pastebin.ubuntu.com/1171881/
[14:29] <ralsina> mandel: no
[14:29] <mmcc> mandel: yes: http://paste.ubuntu.com/1171890/
[14:29] <mmcc> same as alecu
[14:30] <ralsina> mandel: I have been to busy trying to make gatox' branch not crash on my data
[14:30] <mandel> ralsina, ok, nop
[14:31] <mandel> mmcc, alecu, that bug is due to mklog being stupid and doing logger.debug('my_path%30.png some other things %s %s', 'blah', 'blah') plus another logging method in sync..
[14:31] <mandel> let me push a fixed version
[14:31] <mmcc> mandel: great
[14:31] <dobey> yay, i love dbus :(
[14:31] <dobey> hrmm
[14:32] <dobey> thisfred_, mmcc: care to review https://code.launchpad.net/~dobey/rhythmbox-ubuntuone/fix-1042769/+merge/121627 ?
[14:32] <mandel> dobey, I dare you to use the twisted ipc we have in all other platforms!
[14:33] <dobey> mandel: that doesn't help with talking to anything else; like aptdaemon
[14:33] <thisfred_> would love to
[14:33] <thisfred_> but, ehhh,
[14:33] <mandel> dobey, even if it did, is a lot harder to work with :)
[14:33] <thisfred_> ok, I will
[14:34] <thisfred_> dobey str('plugin') doesn't do anything, or am I crazy?
[14:34] <thisfred_> yes is an acceptable answer
[14:35] <ralsina> alecu: so, what do we do with diego's branch? Are there any needsfixings left? I have one but it's rather minor (you can't type spaces in the search box)
[14:35] <dobey> thisfred_: it does with unicode_literals
[14:35] <dobey> thisfred_: converts unicode object to a str object
[14:35] <thisfred_> dobey, if it can
[14:35] <thisfred_> but this is already a string literal
[14:35] <dobey> thisfred_: so in py3 it's a no-op, but in py2 it makes stuff which sucks, not crash
[14:36] <dobey> thisfred_: no, it's a unicode literal
[14:36] <thisfred_> I doubt it for string literals
[14:36] <thisfred_> dobey, in py 2?
[14:36] <thisfred_> it does not have a u
[14:36] <dobey> thisfred_: from __future__ import unicode_literals
[14:36] <thisfred_> ah
[14:36] <thisfred_> so yes, I'm crazy :)
[14:36] <mandel> mmcc, can you try with lp:~mandel/ubuntuone-client/fix-logging
[14:36] <alecu> ralsina: I don't have any other concern with it. I left the tests running this morning... let me check that console!
[14:36] <mandel> mmcc, that one fixes the issue of the logging, might bring with it something funnier..
[14:37] <thisfred_> dobey, so python 3 still has string objects?
[14:37] <thisfred_> or str there just returns unicode
[14:38] <thisfred_> dobey, also, for the non-literals, str will fail if they contain non-ascii, right
[14:38] <thisfred_> ah, but we only use it on literals here
[14:38] <thisfred_> nm
[14:38] <thisfred_> +1
[14:39] <mmcc> ok dobey
[14:39] <dobey> well if someone wants to make a glib signal name with non-ascii, may they FOADIAF
[14:40] <mmcc> +1 dobey
[14:41] <mandel> dobey, lol
[14:45] <ralsina> mmcc: can you create a bundle for vila and jfunk to test?
[14:46] <mmcc> ralsina: in process. just updating so I have all the fixes + mandel's reconnect patch
[14:46] <ralsina> mmcc: could you add gatox's shares branch?
[14:47] <mmcc> ralsina: if you point me to it, sure
[14:47] <ralsina> sure
[14:47] <ralsina> mmcc: as soon as I uncrash chromium that is :-/
[14:49] <ralsina> mmcc: https://code.launchpad.net/~diegosarmentero/ubuntuone-control-panel/tab-shares-functions/+merge/121283
[14:50] <ralsina> alecu: if that's the case with gatox's branch, when you change your vote I will turn my isues into bugs I will fix after that merges
[14:54] <mandel> I think I'm going crazy.. mmcc any luck crashing sd?
[14:54] <mmcc> mandel, it didn't crash for me, but I have mostly been pulling things together for a build so didn't try it too much
[14:54] <mandel> mmcc, ok, I'll keep looking, not to worry
[14:57] <ralsina> mandel, alecu, dobey, thisfred_, briancurtin, mmcc: standup in 3' ends in 5'!
[14:57] <thisfred_> me
[14:57] <thisfred_> there :P
[14:57] <ralsina> thisfred_: premature me-ing, that's a paddling
[14:58] <alecu> ralsina: +1d gatox's branch.
[14:58] <ralsina> alecu: awesome, I'll sort out my stuff on it in a bit :-)
[14:58] <thisfred_> corporal punishment eh
[14:58] <ralsina> thisfred_: no, I use a wii and a robot
[14:59] <thisfred_> uncorporeal punishment?
[14:59] <alecu> me
[15:00] <alecu> ouch
[15:00] <thisfred_> another paddling
[15:00] <thisfred_> me
[15:00] <mandel> me
[15:00] <briancurtin> me
[15:00] <ralsina> me
[15:00] <ralsina> mmcc, dobey?
[15:00] <mmcc> me
[15:00] <dobey> me
[15:01] <ralsina> alecu, go!
[15:01] <alecu> DONE: long review for gatox, more vala and libsoup
[15:01] <alecu> TODO: c++ with nux on the plate today (yuck!)
[15:01] <alecu> BLOCKED: no
[15:01] <alecu> NEXT: thisfred_
[15:02] <ralsina> go briancurtin
[15:02] <briancurtin> DONE: a few small branches to support working on dev-tools on 3
[15:02] <briancurtin> TODO: isolate this bytes formatting bug from the rest, should make it easier to solve the rest of the puzzle
[15:02] <briancurtin> NEXT: mandel
[15:02] <mandel> DONE:Fixed logging issues related to paths that contain % in their name. Found out that sd crashes in my system badly.
[15:02] <mandel> TODO: Find why I get a mem management issue on mac. FIx the bloody thing.
[15:02] <mandel> BLOCKED: no idea why sd crashes.
[15:02] <mandel> ralsina, please
[15:02] <thisfred_> DONE: noodling around with u1db and web music API. Found we need server side u1db modifications for it to work. Tried to guilt webm0nkey into making those changes. TODO: more attempted guilting | adding u1db to web music api BLOCKED: mildly, unless I do the server work myself, which may have to happen NEXT: mandel
[15:02] <ralsina> DONE: mgmt call, other calls, fixed speed saving bug, helped here & there woke up at 5AM, did more reviews TODO: more calls, help around, reviews, help QA if needed, fix a couple of smallish bugs, BLOCKED: no NEXT: mmcc
[15:02] <mmcc> DONE: daemon launches via launchd
[15:02] <mmcc> TODO: QA release & daemon doesn't do anything after launching
[15:02] <mmcc> BLCK: no
[15:02] <mmcc> NEXT: dobey
[15:02] <dobey> DONE: releases
[15:02] <dobey> TODO: releases, fix nautilus plug-in for Q, icon generating magic, music store work
[15:02] <dobey> BLCK: None.
[15:03] <ralsina> dobey: "fix nautilus plug-in" is not affected by the UI Freeze right? And icons are?
[15:03] <ralsina> dobey: just trying to know the order of things :-)
[15:03] <ralsina> TODO: paddle thisfred_
[15:04] <ralsina> https://twitter.com/ralsina/status/240464955755859968
[15:05] <ralsina> This standup was ALMOST perfect. thisfred_ you got the tweet but I am probably being unfair :-)
[15:05] <dobey> ralsina: i suppose they both are. nautilus plug-in issue is the evolution API change and thus necessary UI change
[15:05] <ralsina> EOM
[15:05] <thisfred_> ralsina, what?
[15:05] <ralsina> dobey: ugh :-)
[15:05] <ralsina> that was a :-(
[15:06] <ralsina> thisfred_: it took you 1:50 from the "next thisfred" until you reported
[15:07] <thisfred_> ralsina, shows less than a minute here
[15:07] <ralsina> thisfred_: ok, maybe it 's lag
[15:07] <thisfred_> but fair enough
[15:07] <ralsina> thisfred_: it's like a reality show
[15:07] <ralsina> thisfred_: someone loses every time
[15:08] <ralsina> thisfred_: ;-)
[15:09] <thisfred_> I got distracted. Which is my whole problem with the: get a calendar reminder and/or a ping some time before, then remember to say me, then keep watching the channel, and ignore everyone else talking at you until some time after. But yeah, I've not thought of a better solution either
[15:09] <ralsina> thisfred_: it's 2 minutes
[15:09] <ralsina> we did end it in two minutes, BTW
[15:10] <thisfred_> of waiting. I hate waiting. I can't do waiting, I get distracted :)
[15:10] <ralsina> thisfred_: hehe
[15:10] <thisfred_> that's what every single of my school report cards said: easily distracted
[15:10]  * ralsina goes get the laser pointer
[15:10] <vila> ralsina, mmcc : is the bundle 1) ready but I don't know where to get it OR 2) being built, just be patient ?
[15:11] <ralsina> vila: 2
[15:11] <ralsina> vila: sorry about the delay
[15:11]  * vila watches the blinken lights
[15:11] <alecu> thisfred_: what? who gets distracted?
[15:11] <thisfred_> hehe
[15:11] <ralsina> thisfred_, alecu, vila: look, shiny!
[15:11] <vila> :)
[15:12] <alecu> briancurtin: can I help you somehow with the "bytes formatting bug"?
[15:12] <mandel> thisfred_, I can show you distracted: http://www.urbandictionary.com/define.php?term=pineapples
[15:12] <mandel> thisfred_, first definition
[15:13] <briancurtin> alecu: i think i have it figured out, i just need to separate it out from some other things that are going on in order to get it 100%. some of these issues pop up, then you fix them, then more tests can run but later on what you fixed fails
[15:14] <alecu> briancurtin: sounds familiar! just like ipc on sd! ;-)
[15:16] <alecu> briancurtin: just let me know if you have something for review, or if I can help discussing any unicode issue.
[15:16] <briancurtin> alecu: will do, i'll ping you
[15:23] <dobey> alright, lunch time. bbiab
[15:28] <ralsina> there, gatox's tab-shares is now up for merging :-)
[15:30] <mmcc> build is uploading
[15:30] <mmcc> it has gatox' share links branch, and mandel's reconnect branch.
[15:31] <mmcc> it also loads the root daemon and connects to it but (on my system) does not ever add paths
[15:31] <mmcc> I'd like to see if anyone else sees that issue, but if they do, there's not much to test afterward
[15:32] <ralsina> mmcc: it doesn't fallback to the other fsm?
[15:33] <mmcc> ralsina, no - we don't fallback in case of error - we probably should - but worse, it doesn't get any errors to trigger falling back
[15:33] <mandel> ralsina, mmcc, so it connects and does nothing..
[15:33] <mmcc> it thinks it's working fine but add_watch never gets called as far as I can tell
[15:34] <ralsina> mmcc: ok, that's quite a step backwards...
[15:34] <mmcc> mandel: the daemon gets and processes events but since nobody sent it an add watch message, it doesn't dispatch any
[15:34] <mandel> ralsina, mmcc I think we should give qa one based on fsevents and let me some time to debug that..
[15:34] <mmcc> ralsina: the default in the bundle is still not to use the root daemon
[15:34] <mmcc> (because it doesn't work :) )
[15:35] <mandel> mmcc, so, it will work then? I mean the sync
[15:35] <mmcc> mandel: if you have syncdaemon.conf's fs_monitor setting set to 'default' then it'll work the same as before
[15:35] <mmcc> and that's the default
[15:35] <mandel> ok
[15:36] <mmcc> the only visible difference will be that it'll ask you to install the helper, but it won't matter what you say
[15:36] <mmcc> I mean, it'll matter - - it'll get installed if you install it, but it won't get used
[15:36] <mandel> mmcc, well, then is a step closer :)
[15:36] <mmcc> btw, here is the .app: http://ubuntuone.com/2C1raiI65BlgnS7JOsyOjh
[15:36] <mmcc> should I send an email? to whom?
[15:37] <mmcc> also, the app is signed, so I'm curious for those on 10.8 what it says
[15:37] <mmcc> when you double click it
[15:37] <mandel> mmcc, is that bundle in the share?
[15:37] <ralsina> mmcc: vila, jfunk, elopio, me
[15:38] <mandel> and me :)
[15:38] <ralsina> mmcc: 33MB, that's *reasonable*!
[15:38] <ralsina> mandel: well, you are here
[15:38] <vila> ralsina: thanks
[15:38] <mandel> ralsina, records :)
[15:39] <mmcc> mandel - you should see it in the share
[15:39] <mandel> mmcc, awesome!
[15:47] <mmcc> brb, way past coffee time
[15:56] <ralsina> lunch is looking like a good idea, my belly says
[15:57] <ralsina> will be back in a bit
[15:57] <mandel> mmcc, double clicking in 10.8 does not work yet out of the box
[16:02] <mmcc> mandel: ok, not surprised. thanks
[16:02] <mandel> ralsina, mmcc, officially something in my machine is out, I managed to crash sd from the bundle
[16:02] <mmcc> mandel: crash how?
[16:03] <mandel> mmcc, just using the thing, nothing else
[16:03] <mmcc> mandel: I mean, what happens, how does it crash?
[16:04] <mmcc> and are you using the daemon?
[16:04] <mandel> mmcc, the sd process dies and the control panel never gets the info
[16:06] <mandel> mmcc, no, I'm trying to debug a little
[16:06] <mandel> mmcc, yet I have been happing this with trunk while other have not
[16:07] <mmcc> mandel: is your sd.conf different from ours?
[16:07] <mandel> mmcc, no, I have not touch anything in that area
[16:08] <mmcc> also - to all mac users / testers - remember that you can log in as a guest user to get a clean env that will be deleted when you log them out: http://support.apple.com/kb/PH11321
[16:09] <chaselivingston> mmcc: any news on when an update will be out?
[16:09] <mmcc> chaselivingston: now, see backlog
[16:10] <chaselivingston> just saw it, thanks. mandel just pm'ed me the link as well
[16:11] <mandel> mmcc, I think I really broke my machine at some point.. :*(
[16:12] <mmcc> mandel: try from a guest account?
[16:12] <mandel> mmcc, good point
[16:22] <chaselivingston> mandel, mmcc, ralsina: switching tabs I now get an endless "getting info, please wait" dialog
[16:24] <mandel> chaselivingston, awesome, in a way
[16:24] <mandel> chaselivingston, can you open the activity monitor and look for ubuntuone-syncdaemon?
[16:25] <mandel> chaselivingston, I've been having that bug but I was the only one
[16:25] <chaselivingston> mandel: I see syncdefaultsd
[16:25] <mandel> chaselivingston, that from os x, so no ubuntuone-syncdaemon, right?
[16:26] <chaselivingston> mandel: right
[16:26] <mandel> chaselivingston, close and reopen and check that the guy appears, he should be able to stay there a few mins
[16:26] <mandel> chaselivingston, or enough to catch him :)
[16:26] <mmcc> chaselivingston: is there anything interesting in ~/Library/Caches/ubuntuone/log/syncdaemon.log ?
[16:27] <chaselivingston> mandel: I do see it when I reopen the cp
[16:27] <chaselivingston> mmcc: let me look...
[16:27] <mmcc> like an error or a backtrace or something? (note that that file rotates - it'll only show output from the currently running SD process)
[16:27] <mandel> mmcc, interesting thing is that with a new user I have no problems
[16:27] <mandel> mmcc, could this be the metadata corruption thing?
[16:27] <mmcc> mandel: SOLVED! let's go get a drink
[16:28] <mandel> mmcc, is a shot in the dark.. but since it works in my guest account which was clean..
[16:28] <mmcc> mandel: what metadata issue? not sure I follow
[16:28] <chaselivingston> mandel: and now it's gone in activity monitor… cp is still running
[16:28] <mandel> chaselivingston, do you know how to delete the metadata? do not delete the files, just metadata
[16:28] <chaselivingston> mandel: I don't….
[16:29] <mandel> mmcc, there is a metadata bug where some people get corrupted data and things go bad
[16:29] <mmcc> I don't either :) mandel, please elaborate :)
[16:29] <mmcc> do we have any backtrace or anything from syncdaemon?
[16:29] <mmcc> anything in the Console.app ?
[16:30] <chaselivingston> mmcc: just submitted https://bugs.launchpad.net/ubuntuone-control-panel/+bug/1042834
[16:31] <mandel> mmcc, where is the metadata stored in mac os x?
[16:32] <mandel> mmcc, I know logs are in ~/Library/Caches/ubuntuone but the rest?
[16:33] <mmcc> mandel: looking
[16:33] <mandel> chaselivingston, found it, close the app and rm ~/Library/Application\ Support/syncdaemon
[16:33] <mandel> mmcc, got it ^
[16:34] <mandel> chaselivingston, do not delete shares so that you don't redownload them
[16:34] <mmcc> mandel: I don't have anything there
[16:34] <mmcc> is that really a path on your system?
[16:34] <mandel> mmcc, yes
[16:34] <mmcc> or did you mean application support/ubuntuone/syncdaemon
[16:34] <mmcc> which is where it is on mine
[16:35] <mandel> mmcc, I guess you are not in 10.8
[16:35] <chaselivingston> mandel: ok deleted that, reopening cp now
[16:36] <mmcc> mandel: that shouldn't matter
[16:36] <mandel> mmcc, I do have that path, and chaselivingston found it too
[16:37] <mandel> chaselivingston, will take a little longer because it has to re-create all metadata
[16:37] <chaselivingston> mandel, mmcc: still running after quitting cp after deleting that metadata
[16:37] <mandel> chaselivingston, so it stays there, right?
[16:37] <chaselivingston> mandel: yes, so far
[16:38] <mandel> chaselivingston, semi good news then :)
[16:38] <mandel> chaselivingston, same here :)
[16:38] <mmcc> I know, but the path is set in syncdaemon/config.py line 168 (or 86) and that is it
[16:39] <chaselivingston> mandel: cool
[16:39] <mmcc> I have no idea how you guys got those directories there, or why it would fix anything
[16:39] <mandel> mmcc, deleting the metadata fixes it but because it was corrupted
[16:39] <chaselivingston> mine was under ~/Users/chaselivingston/Library/Application Support/ etc....
[16:39] <chaselivingston> mmcc: ^
[16:39] <mandel> mmcc, could me in ~/my_photos/syncdaemon for all we care
[16:41] <mandel> mmcc, sd has a metadata stored in tritcask that is used to compare the local state with the server state
[16:41] <mmcc> chaselivingston: that initial ~ is a typo right?
[16:41] <mandel> mmcc, I have it in the same place, with ~ aka $HOME
[16:41] <mmcc> ? you have a folder $HOME/Users/mandel/Library/etcetc?
[16:41] <chaselivingston> mmcc: … I don't think so
[16:42] <mmcc> if so, something went very wrong there
[16:42] <mmcc> and mandel, I understand the bug, but why oh why is it looking for the files in the wrong place?
[16:42] <mandel> mmcc, yes, I have a /Users/mandel/Library
[16:42] <mmcc> do you guys also have anything at $HOME/Library/Application\ Support/ubuntuone?
[16:43] <mmcc> no mandel, I'm asking if you have /Users/mandel/Users/mandel/Library/
[16:43] <mmcc> and then I'm asking if you have anything at the right place - /Users/mandel/Library/Application\ Support/ubuntuone
[16:43] <chaselivingston> mmcc: here's the path: http://cl.ly/J3ln
[16:43] <mandel> mmcc, no, no, no second username
[16:44] <mmcc> ok good. false alarm
[16:44] <mmcc> but you *do* have application support/syncdaemon instead of application support/ubuntuone/syncdaemon?
[16:45] <chaselivingston> mmcc: negative, appsupport/u1/sd
[16:45] <mandel> mmcc, I have everything in the right place, I promise, under ~/Library but in the right palce
[16:45] <mandel> s/palce/place
[16:45] <mandel> chaselivingston, is the thing working ok for you?
[16:46] <chaselivingston> mandel: seems to be, sd is still running
[16:47] <mandel> chaselivingston, ok, at least we know why it crashed, that is good news for me
[16:47] <chaselivingston> mmcc: never got that shared folder
[16:47] <mmcc> mandel: so it was never in the wrong place?
[16:48] <mmcc> chaselivingston: you should see it now. web UI hung for some reason
[16:49] <mandel> mmcc, in my case it was always under ~/Library/Application Support/ubuntuone
[16:49] <mandel> mmcc, it was a communication error
[16:49] <mandel> mmcc, the important thing is that we know we have metadata issues
[16:49] <mmcc> ok good, I'm glad I don't have to debug a crazy config path bug
[16:50] <chaselivingston> mmcc: still a no go on that folder, shouldn't it send me an email?
[16:51] <mmcc> chaselivingston: yes it should
[16:51] <mmcc> check spam?
[16:52] <chaselivingston> mmcc: nothing there...
[16:53] <mandel> chaselivingston, one last question, do you have a path with % in it?
[16:54] <mandel> chaselivingston, even in a share
[16:55] <chaselivingston> mmcc: haven't seen one, let me look
[16:55] <chaselivingston> mmcc: nope, doesn't appear that way
[16:57] <chaselivingston> my core 2 duo is struggling w/ this sd running lol
[17:01] <briancurtin> anyone have 1' for a two line MP? https://code.launchpad.net/~brian.curtin/ubuntuone-dev-tools/py3-map-list/+merge/121478
[17:01] <mmcc> chaselivingston: high cpu usage or high memory usage? can you paste a screenshot of activity monitor showing it?
[17:01] <mmcc> briancurtin: I do
[17:02] <chaselivingston> mmcc: it fluctuates a lot, was trying to catch it at 59.6 when I took this shot: http://cl.ly/J4Rc
[17:03] <chaselivingston> mmcc: overall my MBP is running pretty slow right now though
[17:04] <mmcc> chaselivingston: ok, thanks… how much RAM do you have, and are you running low on free mem? activity monitor should be helpful here too
[17:05] <mmcc> whoa, I'm running low on free ram - had no idea thanks to swap on flash, I guess…
[17:05] <chaselivingston> mmcc: I have 4GB, looks like only about 90MB free according to activity monitor
[17:05] <vila_> mmcc: ping, first install ever from u1-120828, copied .app under /Applications, had to reboot as 'sign with existing account' raised an error (DeadReferenceError 'Calling Stale Broker')
[17:05] <mandel> vila_, known issue, close everything I shall have a fix for tom about that
[17:06] <mmcc> vila_: yes, that's a known bug. My apologies that we don't have a README for the builds yet
[17:06] <vila_> mmcc: after reboot, launching the app stays in 'getting information, please wait...' for 11 minutes now
[17:06] <mandel> mmcc, once I have the new metadata everything works like a charm \o/
[17:06] <mandel> vila_, which tab?
[17:07] <mandel> vila, also open activity monitor and confirm that there is a process called ubuntuone-syncdaemon
[17:07] <vila_> mandel: no tab, I just launched the app I haven't sign in yet
[17:07] <vila_> yes
[17:07] <mandel> vila_, kill again the app
[17:07] <mmcc> vila_, I've seen that before too, and now that you mention it, my test app is now hanging there.
[17:07] <mmcc> I think I can try to debug from here
[17:08] <mandel> mmcc, you deal with it, I'm EOD will look at the metadata tom
[17:08] <mmcc> mandel: the folders tab is shown filled in in the background in my case, but the username account info is not.
[17:08] <mmcc> mandel sure
[17:08] <vila_> ok, mmcc , restart app ?
[17:09] <mmcc> vila_: yes, try quitting it. this may also hang, in which case you will have to force quit it using Activity monitor :\
[17:09] <vila_> mmcc: by the way, I have a bunch of tests that want to use u1sdtool, can I haz ?
[17:09] <vila_> mmcc: nope, cmd-Q worked fine
[17:09] <mmcc> vila_: u1sdtool is not currently packaged with the .app, I'm not sure if we have a bug for that or not
[17:10] <vila_> mmcc: no idea, can you check or do you prefer that I file one (and you mark it as dupe if needed ?)
[17:10] <mmcc> vila_: I'm adding one now, no need for you to file. thanks!
[17:11] <vila_> mmcc: great, thanks
[17:12] <vila_> mmcc: seems like I'm blocked again at that 'getting info, please wait...' stage
[17:12] <vila_> anything I can do to help you diagnose ?
[17:14] <mmcc> vila_: log files might be useful. can you zip up the contents of ~/Library/Caches/ubuntuone/log/ ? I'll create a bug and you can attach them there
[17:16] <vila_> Can I haz a cmdline tool to do that ? ;)
[17:16] <vila_> kidding
[17:16] <vila_> but...
[17:17] <mmcc> vila_: the "please wait" thing is continuously animating for you, right? or is it completely frozen
[17:18] <chaselivingston> mmcc: FWIW, mine is continually animating for ~5 mins now
[17:19] <mmcc> chaselivingston: vila_ the bug is https://bugs.launchpad.net/ubuntuone-control-panel/+bug/1042860
[17:20] <chaselivingston> mmcc: was just able to get past that, should I attach logs to that bug?
[17:20] <mmcc> chaselivingston: yes please. what did you do to get past it?
[17:20] <chaselivingston> mmcc: cmd+q and restart app
[17:20] <chaselivingston> mmcc: doesn't always work though
[17:20] <mmcc> yep, that's what I've been seeing too
[17:20] <vila_> mmcc: on first run, I noticed that com.ubuntu.one.fsevents (as root) consumed cpu for quite some time (sry, I didn't note how long, but long enough for me to catch it (minutes))
[17:21] <mmcc> vila_: how much cpu? it shouldn't be doing much at all, but if you have a lot of filesystem activity, it might be logging too much
[17:22] <mmcc> vila_: you can kill that using 'sudo launchctl com.ubuntu.one.fsevents' without affecting the rest of the app, and if you cancel the 'install helper' dialog on startup it won't re-install it
[17:22] <vila_> I didn't have (AFAICS) but I'm not sure there has ever been a fs daemon on this mac (or at least a client consuming from that)
[17:22] <mmcc> vila_: it should be new from starting the build from today
[17:22]  * vila_ remembers his osx basics by the minute ;)
[17:22] <chaselivingston> mmcc: logs attached, thanks
[17:23] <mmcc> chaselivingston: thank you!
[17:23] <chaselivingston> mmcc: np, anything to make the app better :)
[17:24] <mandel> ok, I'm off for real
[17:27] <chaselivingston> mmcc: also, my storage usage stats haven't changed since installing this new build, though something should be uploading, not sure what it's actually doing in the background...
[17:27] <vila_> mmcc: I'm off too but let me know when you have a new bundle to test
[17:27] <mmcc> ok vila_ thanks for your help!
[17:27] <vila_> mmcc: you're very welcome
[17:27] <mmcc> chaselivingston: is syncdaemon.conf showing any activity?
[17:28] <vila_> mmcc: I have a bunch of osx users waiting for u1 ;)
[17:28] <chaselivingston> mmcc: where would that be located/found?
[17:28] <mmcc> you can keep tabs on it in the terminal with 'tail -n100 -f ~/Library/Caches/ubuntuone/log/syncdaemon.conf'
[17:28] <vila_> or should I say: *I*'m waiting for the osx client for a long time so I can "sell" it to a bunch of friends
[17:28] <vila_> now they wait for me ;)
[17:29] <mmcc> vila_: glad to hear it - can't wait to give you something polished to sell :)
[17:29] <chaselivingston> mmcc: tailing it, nothing so far
[17:29] <dobey> mmcc or thisfred_: https://code.launchpad.net/~dobey/ubuntu-sso-client/update-4-0/+merge/121669 please
[17:30] <thisfred_> on it
[17:30] <vila_> mmcc: and a packaged u1sdtools will *immensely* help me for tests
[17:31] <thisfred_> +1
[17:32] <dobey> thanks
[17:34] <mmcc> vila_: ack. it'll help us for debugging too.
[17:34] <chaselivingston> mmcc: getting a no such file or directory for tail -n100 -f ~/Library/Caches/ubuntuone/log/syncdaemon.conf
[17:34] <ralsina> hello again! What did I miss?
[17:34] <dobey> a liger mauling
[17:34] <mmcc> chaselivingston: arg, typo sorry
[17:35] <mmcc> should be syncdaemon.log
[17:35] <chaselivingston> mmcc: ah yes, much activity
[17:35] <mmcc> ralsina: lots of good bug reports, a potential metadata bug
[17:35] <vila> ralsina: steve's ghost bullying the control panel
[17:36] <thisfred_> ralsina: margarita tuesday
[17:36] <ralsina> hehe
[17:38] <mmcc> hmm, share links tab is just frozen on loading overlay for me, even when other tabs work
[17:39] <mmcc> and I get no debug log messages from it
[17:40] <mmcc> margarita*s* tuesday
[17:40] <dobey> so
[17:41] <dobey> what branches are coming for ubuntuone-client and ubuntuone-control-panel that i need to wait on?
[17:44] <ralsina> dobey: I think nonw
[17:44] <ralsina> dobey: I am working on fixing some stuff in the shares tab but nothing crucial
[17:46] <mmcc> ralsina: while you're in there, can you add a bunch of logging? I'm assuming there is no logging, because I'm seeing nothing.
[17:49] <ralsina> mmcc: I could, I guess :-)
[17:49] <dobey> hmm
[17:49] <dobey> i am going to make a branch to skip this test that keeps failing intermittently during nightlies builsd
[17:49] <dobey> builds
[17:51] <ralsina> dobey: ok
[17:53] <vila> mmcc: do you have the bug # for the missing u1sdtools ?
[17:53] <mmcc> vila: https://bugs.launchpad.net/ubuntuone-windows-installer/+bug/1042855
[17:54] <vila> mmcc: you rock, thanks
[17:54] <vila> err, waitaminute u1-windows-installer ?
[17:54] <mmcc> be calm :)
[17:54] <dobey> heh
[17:54] <dobey> don't let the name fool you
[17:54] <mmcc> yes, it's just the project that builds the installers/ packaging. it's the right project
[17:55] <mmcc> just needs a new name
[17:55] <vila> mmcc: np, just want to make sure I file bugs in the right place later ;)
[17:55] <mmcc> vila: if you're ever curious about where to file, ping away
[17:56] <vila> will do
[17:56]  * briancurtin lunch
[17:56] <chaselivingston> mmcc: should my storage space used match what Finder reports as the size of my u1 folder?
[17:57] <mmcc> chaselivingston: it should be close, but it won't necessarily be exact
[17:57] <chaselivingston> mmcc: ok, right now Finder says my u1 folder is ~7.5GB, but only seeing about 5.3GB usage in u1
[17:58] <mmcc> chaselivingston: that's significant. a difference of a few kB could be due to files we don't sync but that's something else
[17:58] <mmcc> maybe just not synced fully yet (i hope?)
[17:58] <chaselivingston> mmcc: hmmm, hopefully that's the case
[17:58] <mmcc> I'm working right now on a build with u1sdtool inside
[17:58] <chaselivingston> mmcc: ok cool
[17:58] <ralsina> chaselivingston: if it's not fully synced you should see it still "sync in prgress"
[17:59] <chaselivingston> mmcc: I do, and now space is decreasing, down to 5.1GB now
[17:59] <mmcc> well, that's unexpected.
[18:00] <mmcc> chaselivingston: so the u1 web site says there's less data now? or it's the display inside control panel?
[18:00] <chaselivingston> mmcc: u1 website, trying to load up the cp now
[18:01] <mmcc> chaselivingston: those two should match. the CP display gets the usage and quota info from the web
[18:01] <chaselivingston> mmcc: right, which I think they do
[18:02] <chaselivingston> mmcc: it does still say in progress, so hopefully that's a good sign
[18:03] <chaselivingston> mmcc: I'm wondering if Finder could be reporting an incorrect size, is there a way to make it refresh its calculation?
[18:04] <mmcc> Finder should be updating live. you could always do a "get info" on the folder
[18:04] <chaselivingston> yeah, that's what I did, it now says it's 6.7GB, which is indeed smaller, but still doesn't match what u1 shows. also still waiting on the builds in your shared folder to download
[18:05] <mmcc> chaselivingston: can you remind me which direction things are going in your case? you've copied a bunch of stuff into your ~/Ubuntu One/ , but you also had some things already in U1, how much was in your account before you started testing the mac client?
[18:05] <chaselivingston> mmcc: there was only a few KB in there before testing, now a lot of it has uploaded, but it doesn't seem to be all
[18:06] <mmcc> chaselivingston: I asked because I don't understand why the size of your local folder is decreasing
[18:07] <chaselivingston> mmcc: well I did remove some things from my local folder, but I'm not sure it was 1GB worth...
[18:07] <mmcc> ok
[18:07] <chaselivingston> mmcc: though its size is remaining steady right now at 6.7GB
[18:08] <mmcc> weird. so it was 7.5, then 5.1, now 6.7?
[18:09] <chaselivingston> mmcc: no sorry, local folder was 7.5 then 6.7, perhaps I did remove that much. u1 website showed 5.8GB earlier and now shows 5GB
[18:11] <mmcc> ok, so maybe you removed some files that had already been synced, explaining why the site showed a decrease, while there were still other files that need to be synced, explaining why they're still different
[18:11] <mmcc> ok, we might be safe
[18:11] <chaselivingston> mmcc: true, maybe so
[18:11] <mmcc> do you still have those files you removed, just to sanity check the size?
[18:12] <chaselivingston> mmcc: I don't… though I guess those total sizes would match up between the website and local folder
[18:12] <mmcc> they should eventually, once everything is synced and we reach nirvana
[18:15] <dobey> mmcc, thisfred_: https://code.launchpad.net/~dobey/ubuntuone-client/skip-being-hashed/+merge/121674 please
[18:15] <dobey> is trivial
[18:16] <thisfred_> dobey, +1
[18:16] <mmcc> OK, so the fastest way to package a u1sdtool is to give it a .app wrapper like the others, even though it might not need it (won't waste space due to symlink stuff) Does anyone care what I call that app wrapper? I'm thinking "U1SDTool.app" -- avoiding spaces is nice because this will be pasted at people who need it for debugging from the terminal
[18:16] <mmcc> dobey, approved
[18:17] <chaselivingston> mmcc: name sounds good to me
[18:17] <dobey> thanks
[18:21] <ralsina> mmcc, anyone: semi-trivial branch https://code.launchpad.net/~ralsina/ubuntuone-control-panel/space-the-final-frontier/+merge/121675
[18:23] <ralsina> actually, hold on, I have a much simpler test I can do
[18:23] <chaselivingston> mmcc: so no sd in activity monitor anymore, even w/ cp running...
[18:26] <ralsina> there, reviews please when convenient :-)
[18:26] <chaselivingston> mmcc: …. and of course after I tell you that, it comes back
[18:32] <mmcc> ralsina: ok, looking at that now in my precise VM. actually, looking after it downloads *so many updates*
[18:33] <mmcc> ralsina, while I'm waiting, grammar nitpick in the test docstring: "action are"
[18:34] <ralsina> mmcc: it's copypasted, so I'll fix both ;-)
[18:36] <ralsina> alecu: quick question. In my notebook, it takes *minutes* to fill the list of files for completion. While it does that, you can type in the search box and it shows an empty completion popup, which doesn't seem like a good idea. Agree?
[18:38] <ralsina> mmcc: pushed 6 or 7 fixes of the same typo
[18:39] <ralsina> alecu: also, it seems that if you have a folder with the same name as a UDF but you are not syncing it, it indexes its contents anyway
[18:43] <mmcc> may I request a review of the following two setup-mac merges: https://code.launchpad.net/~mikemc/ubuntuone-windows-installer/build-sign-daemon/+merge/121668 and https://code.launchpad.net/~mikemc/ubuntuone-windows-installer/gimme-u1sdtool/+merge/121679
[18:43] <ralsina> Also, it seems I am syncing over 250K files
[18:43] <ralsina> mmcc: sure!
[18:45] <mmcc> the second one builds a working sdtool, so I'm going to build a new app bundle for anyone interested in giving it a whirl.
[18:45] <ralsina> mmcc: the two branches seem to share changes
[18:45] <ralsina> mmcc: so, any particular order?
[18:46] <mmcc> ralsina: yeah, sorry. build-sign-daemon before gimme.
[18:46] <mmcc> can I retroactively add a prereq?
[18:46] <ralsina> mmcc: no idea, and probably no need :-)
[18:47] <mmcc> ok. I think I did it before by resubmitting the proposal, let me go ahead and do that
[18:47] <ralsina> mmcc: "Joe Developer"?
[18:48] <mmcc> here's the fixed merge proposal for the second branch: https://code.launchpad.net/~mikemc/ubuntuone-windows-installer/gimme-u1sdtool/+merge/121680
[18:48] <mmcc> Joe Developer is the test cert ID from Apple's sample app
[18:49] <mmcc> I figured it's as good as anything until I know what cert ID to actually use
[18:49] <ralsina> mmcc: sounds good to me
[18:50] <ralsina> mmcc: +1 on the branch although I have never tried to build that myself
[18:50] <ralsina> mmcc: so I am trusting you with that
[18:50] <mmcc> ralsina: ok.
[18:50] <mmcc> ralsina: I think there's a lot of that going on in setup-mac :)
[18:51] <mmcc> looking forward to automated builds
[18:52] <ralsina> mmcc: I know the feeling, I was the owner of the only magical windows-build box for 5 months :-)
[18:53] <mmcc> ralsina: yikes :) at least I've successfully moved it from one machine to another already. makes me more comfortable with the magic level
[18:56] <mmcc> speaking of space-the-final-frontier, I just finished an audiobook of Mary Roach's "Packing For Mars: The Curious Science of Life in the Void", and it was hilarious
[18:57] <mmcc> ralsina: what exactly should be happening when I type in the search field? what files am I searching through?
[18:58] <mmcc> I'm seeing nothing happen, basically
[18:58] <mmcc> and no logs
[18:59] <mmcc> (this is in a freshly updated precise VM)
[19:00] <ralsina> mmcc: it should search through all your files you have in u1
[19:00] <ralsina> mmcc: it's doing nothing for me, but that's because I have a metric ton of files
[19:00]  * briancurtin dropping off some letters down the street, be back shortly
[19:00] <ralsina> mmcc: and it doesn't work until it finishes scanning.
[19:01] <ralsina> mmcc: approved the second branch because what could possibly go wrong
[19:01] <ralsina> mmcc: in fact, go ahead and merge
[19:02] <mmcc> ok, thanks for the approves
[19:02] <mmcc> as far as the share links, I have very few files, and I can't tell if it's doing anything
[19:03] <mmcc> this tab really needs some kind of progress indicator to show that it's scanning
[19:03] <mmcc> I also have nothing showing in the table of shared files, which is probably not right. I've definitely shared some files
[19:04] <mmcc> yeah, u1sdtool --list-shared comes back with what I expect
[19:05] <ralsina> mmcc: I will do a branch that does something to show it's still scanning
[19:05] <ralsina> mmcc: and will add some logging
[19:08] <mmcc> ralsina: ok. well, tests pass lint free. should I just approve this and wait for a separate branch then?
[19:09] <mmcc> also, whoops - I accidentally snuck the plist change to include the icon into that last setup-mac merge… harmless but kind of sloppy, my apologies
[19:12] <ralsina> I smell a UI Freeze exception request coming for the share tab
[19:17] <ralsina> mmcc: if you get an empty list on the shares tab, try this: 1) start syncdaemon manually. 2) wait a little 3) start u1cp
[19:17] <ralsina> mmcc: it seems if u1cp asks quickly while sd is processing the command pool, it doesn't get the info
[19:18] <mmcc> ralsina: hm. well, I have had sd on for a whle, and restarted cp a few times, same issue
[19:19] <ralsina> mmcc: ok, then it's something else
[19:19] <ralsina> mmcc: can you run u1cp with U1_DEBUG=True and send me the output?
[19:19] <mmcc> ralsina: coming right up
[19:21] <mmcc> ugh, got the "No JSON object could be decoded" error while waiting for the account panel
[19:22]  * briancurtin back
[19:23] <dobey> mmcc, thisfred_: https://code.launchpad.net/~dobey/ubuntuone-client/update-4-0/+merge/121684 please
[19:23] <mmcc> ralsina, http://paste.ubuntu.com/1172399 has that error, hold on and I'll paste another run where maybe it won't show up
[19:24] <thisfred_> dobey, stop doing all this work. Tuesdays are now your official day off.
[19:24] <ralsina> dobey: I got it
[19:24] <thisfred_> either that, or I'm switching my review day to february the 29th
[19:25] <ralsina> thisfred_: yes, tuesdays are a bad day for review day :-)
[19:25] <thisfred_> I'm just kidding
[19:25] <dobey> all my tuesday reviews are usually easy, since they're just "merge from trunk to stable" branches
[19:25] <ralsina> mmcc: oh, you have some cute exceptions there
[19:25] <thisfred_> exactly
[19:26] <mmcc> ralsina: http://paste.ubuntu.com/1172404 - this one avoided the JSON exception and showed the folders panel correctly, but nothing in the share links panel
[19:26] <mmcc> ralsina: yeah. on the run w/those exceptions, I got no username/quota info and then that crash when trying to load the accounts tab
[19:26] <ralsina> mmcc: ok, I'll have to add some logging there
[19:26] <thisfred_> dobey, +1
[19:29] <mmcc> ralsina: just saw your bug update, fwiw, I also see no empty completion popup
[19:33] <dobey> thanks
[19:33] <dobey> brb
[19:33] <ralsina> mmcc: ok, then it's something else in your case
[19:33] <ralsina> mmcc: is everyone on mac getting the same thing?
[19:34] <mmcc> ralsina: the same thing?
[19:34] <ralsina> mmcc: nothing on that tab
[19:34] <mmcc> oh, I don't know. I've been testing it on precise just now…
[19:35] <ralsina> mmcc: oh on precise?
[19:35] <ralsina> mmcc: ok then!
[19:35] <ralsina> mmcc: just give me 10' and we'll look at that :-)
[19:36] <ralsina> dobey: if I were to propose a branch that adds a string and a button to u1cp tomorrow early, would that make you hate me a lot?
[19:37] <mmcc> ralsina: sounds good. since you asked, on the mac I get the endless loading overlay when I switch to the links panel. also nothing in the log though
[19:37] <mmcc> going to be afk for a few minutes, need to switch locations
[19:38] <ralsina> mmcc: the overlay is just waiting for one IPC call to get the shared files list from sd, so theremust be some IPC problem there
[19:48] <ralsina> mmcc: https://code.launchpad.net/~ralsina/ubuntuone-control-panel/disable-enable/+merge/121691 should give you slightly better logging while you review it
[19:49] <ralsina> mmcc: it will log when it starts the thread to collect the file list, and when it's finished, and how many files it found
[19:55] <ralsina> vila: do you have a report of your testing?
[19:55] <ralsina> vila: or should I just read the backlog very carefully? ;-)
[19:59]  * alecu is back, after lunch, errands, and a broken internet connection.
[19:59] <alecu> ralsina: ping
[20:00] <alecu> ralsina: I agree that it's not a good idea for the search box to have so much delay while it fills up
[20:00] <alecu> ralsina: but I don't think searching a partial list is any better
[20:01] <mmcc> alecu, ralsina how is it building that list? can we save a precomputed list to a cache file somewhere or something?
[20:01] <ralsina> alecu: I have a branch to disable it while it's useless
[20:01] <alecu> ralsina: my guess is that we should not be scanning the disk again, but using the metadata instead
[20:01] <ralsina> alecu: makes tons of sense
[20:01] <mmcc> oh you're way ahead of me, I'll go back to testing :)
[20:01] <ralsina> alecu: I am sending a mail about things I see in that page we should fix immediately
[20:01] <alecu> ralsina: that would still be slowish
[20:02] <alecu> ralsina: but will solve the udf issue
[20:02] <ralsina> alecu: we can still fix that by being slightly smarter
[20:02] <alecu> mmcc: yes, the metadata in this case is exactly that: a precomputed list of files.
[20:02] <ralsina> alecu: but that's a bug and doesn't have a deadline of today/tomorrow ;-)
[20:03] <alecu> one thing: I'm not sure if the metadata has a reasonable way to do full text search of filenames...
[20:04] <ralsina> alecu: we can just use the metadata because it's faster than reading a bazillion folders
[20:04] <ralsina> alecu: even if it can't do the searching
[20:04] <alecu> ralsina: that's for sure, though :-)
[20:04] <alecu> ralsina: how do you plan to disable this? the whole tab?
[20:04] <ralsina> alecu: no, just the search field
[20:04] <ralsina> alecu: since you can do other things in the tab
[20:05] <ralsina> alecu: https://code.launchpad.net/~ralsina/ubuntuone-control-panel/disable-enable/+merge/121691
[20:05] <alecu> ack
[20:05] <mmcc> ralsina: with your disable-enable branch, I still get no log messages. same exact behavior. Am I doing something wrong with deps or something?
[20:06] <mmcc> (this is in precise)
[20:06] <ralsina> mmcc: no idea
[20:06] <ralsina> mmcc: how old is your syncdaemon process/
[20:06] <mmcc> ralsina: no older than an hour ago when I updated the world and had to reboot
[20:06] <ralsina> mmcc: ok
[20:10] <mmcc> ralsina: here's more info - on precise, u1sdtool -s tells me the queues are idle, and --waiting prints nothing. But on darwin, it is stuck on "doing local rescan", queues: WORKING, and --waiting says this:   "GetPublicFiles(running=False, _iri='https://one.ubuntu.com/files/api/public_files')"
[20:11] <ralsina> mmcc: could you tail -f the log? Maybe it just is doing that local rescan
[20:11] <ralsina> mmcc: or it exploded doing it
[20:12] <ralsina> mmcc: seems to me that your sd hanged and u1cp is not getting back the info it requested (on darwin)
[20:12] <ralsina> mmcc: on precise, I am stumped
[20:12] <dobey> ralsina: i can do the u1cp release tomorrow; but let's make sure it's in trunk before noon.
[20:13] <mmcc> ralsina: yeah, ignore darwin, there's other stuff going on there
[20:14] <ralsina> mmcc, alecu: I checked vila's logs
[20:15] <ralsina> he's failing to get the credential from wherever it is stored in darwin
[20:16] <mmcc> ralsina: hmm. is that the bug with SSO interaction on creating / logging in the first time?
[20:16] <mmcc> he did mention that in chat
[20:16] <ralsina> mmcc: looks like it
[20:16]  * ralsina goes to vnc to logout of u1
[20:17] <ralsina> dobey: before noon, sure
[20:18] <dobey> ralsina: were you reviewing my update-4-0 client branch btw?
[20:18] <ralsina> dobey: I saw eric reviewed it
[20:18] <dobey> ralsina: ah ok. you said "I got it" so I was waiting for your review as well :)
[20:18] <ralsina> dobey: sorry :-)
[20:18] <dobey> but i'll mark it approved with just eric
[20:23] <alecu> ralsina: +1 to enable-disable. I tested it irl also on P.
[20:23] <ralsina> alecu: awesome
[20:23] <ralsina> grmbl, I can't vnc into the mac anymore!
[20:24] <mmcc> alecu: so it's something weird on my end that I can't make that branch work on P?
[20:24] <ralsina> rebooting the freedom-hating piece of aluminum
[20:24] <ralsina> mmcc: yes
[20:24] <alecu> mmcc: does trunk work on your P dev env?
[20:24] <ralsina> mmcc: it's not us, it's you ;-)
[20:25] <mmcc> trying trunk now
[20:25] <chaselivingston> mmcc: would this indicate problems with syncing? 2012-08-28 16:15:51,222 - ubuntuone.SyncDaemon.StorageClient - DEBUG - Ping! rtt: 1.599 segs
[20:25] <mmcc> those changes don't depend on trunk from any other project do they?
[20:26] <mmcc> chaselivingston: no, that's just a periodic heartbeat message
[20:26]  * mmcc wonders why does it say '1.599 segs' instead of 'secs'? 
[20:26] <alecu> mmcc: not the small changes in that branch
[20:26] <chaselivingston> mmcc: ah gotcha
[20:26] <mmcc> chaselivingston: I have a build for you with u1sdtool, just a sec
[20:26] <chaselivingston> mmcc: awesome, thanks
[20:27] <ralsina> mmcc: you are surrounded by argentinians (and a spaniard?)
[20:27] <alecu> mmcc: probably some spanish speaker misstyped "segundos" instead of "seconds"
[20:27] <mmcc> chaselivingston where by 'sec' I mean 'a couple of minutes'
[20:28] <chaselivingston> mmcc: haha, I knew what you meant ;)
[20:28] <mmcc> segundos, claro que si
[20:34] <dobey> thisfred_, mmcc: https://code.launchpad.net/~dobey/ubuntuone-dev-tools/update-4-0/+merge/121704 please
[20:35] <thisfred_> on it
[20:37] <thisfred_> dobey, looks good to me
[20:37] <dobey> gracias
[20:39] <dobey> doh, and i'll have to merge the client branch manually, again
[20:42] <mmcc> chaselivingston: I didn't forget. just killing some logs that probably cause the fsevents daemon to slow down on your poor old mac
[20:42] <chaselivingston> mmcc: no worries, thanks :)
[20:42] <chaselivingston> mmcc: got a new quad core i7 coming thursday :)
[20:43] <mmcc> chaselivingston: also, if you're not testing the daemon (and you shouldn't), just cancel the dialog that pops up asking about a helper tool. that'll avoid installing it in the first place
[20:43] <chaselivingston> gotcha, will do
[20:45] <dobey> chaselivingston: 3770?
[20:46] <chaselivingston> dobey: ? not sure what you mean
[20:46] <dobey> chaselivingston: "new quad core i7"
[20:46] <chaselivingston> dobey: yes… what does the 3770 mean? lol
[20:47] <dobey> chaselivingston: it's one of the i7 model numbers
[20:47] <chaselivingston> dobey: oh sorry, wasn't aware. not sure which one it is, let me see if I can find out
[20:47] <dobey> chaselivingston: or you bought a machine that has an i7 and you have no idea which one?
[20:48] <chaselivingston> dobey: that's more the case, not sure which i7 the new 15" pros are shipping with
[20:48] <mmcc> chaselivingston: retina display or not?
[20:48] <dobey> oh, macbook
[20:48] <chaselivingston> mmcc: no retina
[20:48] <chaselivingston> mmcc: just went w/ the high-res
[20:49] <dobey> "high res" it's like what, 120 dpi?
[20:49] <mmcc> yeah, they don't even say the processor model number for the laptops
[20:49] <chaselivingston> dobey: not sure, it's like 1680x1200 or something I believe
[20:50] <chaselivingston> dobey: vs. 1440x900 regular res
[20:51] <ralsina> grmbl my mac dev env seems to have spontaneously combusted
[20:51] <ralsina> mmcc: how's the bundle coming?
[20:51] <dobey> 1680x1050
[20:52] <chaselivingston> dobey: ah 1050, that's right
[20:52] <dobey> so yeah, 128 dpi
[20:52] <dobey> low res
[20:52] <chaselivingston> dobey: lol, but higher than the "low res" 1440x900
[20:52] <ralsina> dobey: at 2 feet from your face, that's the same as the latest iphone's resolution ;-)
[20:53] <mmcc> ralsina: chrome says it uploaded, but the webui is stuck on 'loading' overlay
[20:53] <mmcc> story of my life
[20:53] <ralsina> mmcc: open the page n another tab?
[20:53] <dobey> heck, the 'retina' is low res compared to my laptop :)
[20:53] <ralsina> mmcc: and see if it's there
[20:53] <ralsina> mmcc: bad overlay day huh
[20:54] <mmcc> for REAL
[20:54] <mmcc> anyway, http://ubuntuone.com/1CENTFFDneAt3ESZdZmcgd
[20:54] <alecu> dobey: what's the ppi on your laptop?
[20:54] <dobey> alecu: 260
[20:54] <alecu> dobey: that's on the fujitsu? awesome!
[20:55] <dobey> yep
[20:55] <mmcc> that is now a zipped folder with a README
[20:55] <chaselivingston> mmcc: cool, downloading now
[20:55] <alecu> mmcc: is that the mac bundle?
[20:56] <mmcc> dobey: is that the lifebook u820?
[20:56] <mmcc> alecu yes
[20:56] <dobey> mmcc: yep
[20:56] <mmcc> alecu a new build with u1sdtool
[20:58] <dobey> i want a couple of the new panasonic ips panels for my workstation though
[20:59] <mmcc> dobey: is that really a 5" screen? yow
[20:59] <dobey> mmcc: the u820? yep. 5.6"
[20:59] <mmcc> btw, this is why the web was invented: http://en.wikipedia.org/wiki/List_of_displays_by_pixel_density
[21:00] <mmcc> yeah, really high DPI is spoiling. my phone and ipad make the laptop look crappy, and make my big LCD look really crappy
[21:02] <chaselivingston> mmcc: here's my u1sdtool output. that last bit doesn't seem too good
[21:02] <chaselivingston> mmcc: https://pastebin.canonical.com/73206/
[21:03] <vila> ralsina: sry EOD was fafk (far away ;) should I read the log here ? In a nutshell: first install, couldn't sign in. Will try the new bundle first thing tomorrow
[21:03] <ralsina> mmcc: I was getting eternal overlay on mac: was a certificate error on sd
[21:03] <ralsina> vila: awesome, thanks
[21:03] <mmcc> chaselivingston: yeah, the warning isn't an issue but the 'fatal error' is…
[21:03] <ralsina> mmcc: fixed that now I get only eternal overlay on the share file tab
[21:03]  * vila can't find what sd means... single daemon... simple density (oops, showing my age ;)... super dynamic...
[21:04] <mmcc> ralsina: please, to explain. how can I achieve this fixing as well
[21:04] <vila> simple diagnostic ?
[21:04] <mmcc> vila: syncdaemon :)
[21:04]  * vila facepalms, time to sleep ;)
[21:05] <chaselivingston> mmcc: does that error mean I'm not actually syncing?
[21:05] <mmcc> chaselivingston: can you tell me what u1sdtool --waiting says ?
[21:05] <chaselivingston> sure, one sec
[21:05] <mmcc> chaselivingston: we'll see…
[21:05] <mmcc> ping alecu
[21:05] <chaselivingston> mmcc: https://pastebin.canonical.com/73207/
[21:06] <ralsina> vila: sorry, syncdaemon
[21:06] <mmcc> chaselivingston: and u1sdtool --current-transfers should also be enlightening
[21:07] <ralsina> mmcc: check syncdaemon logs, if you see any mentions to .pem files
[21:07] <ralsina> mmcc: if you have none, it's not the same problem
[21:07] <mmcc> btw, that ascii codec error in : https://pastebin.canonical.com/73207/ looks like a good bug
[21:07] <chaselivingston> mmcc: current uploads and downloads: 0
[21:07] <chaselivingston> mmcc: sounds like there's a problem
[21:08] <mmcc> chaselivingston: yes it does…
[21:08] <chaselivingston> mmcc: one sec, let me relaunch the cp and then run that just for fun
[21:08] <mmcc> ok
[21:08] <mmcc> brb myself,need to escape the rain
[21:09] <chaselivingston> mmcc: now I get: Current uploads: 0
[21:09] <chaselivingston> Oops, an error ocurred:
[21:09] <chaselivingston> 'ascii' codec can't decode byte 0xe2 in position 50: ordinal not in range(128)
[21:10] <ralsina> mmcc: when running from sources, at least, ubuntu-sso-login is not started
[21:11] <dobey> oh man, unicode vs. ascii
[21:11] <dobey> it's like a celebrity deathmatch
[21:11] <chaselivingston> mmcc: now I got something useful: https://pastebin.canonical.com/73209/
[21:12] <mmcc> chaselivingston: what changed, did you just re-run u1sdtool again ?
[21:13] <chaselivingston> mmcc: yeah, but now it's back to 0 on uploads and downloads, no change. cp is still running, but it doesn't look like sd is
[21:13] <vila> mmcc: going to sleep now, ping me with the url for the last bundle you produce before leaving and I'll test it (ping will be easier than me parsing the log before coffee ;)
[21:13] <mmcc> vila: you bet. I could also just email?
[21:14] <vila> mmcc: even better
[21:15] <chaselivingston> mmcc: hm, and now I'm back to useful output. any idea why sd quits and restarts seemingly randomly?
[21:17] <mmcc> chaselivingston: I thin u1sdtool will start it if it's not running, but the quitting is troublesome. do you see any exceptions or errors in syncdaemon.log?
[21:17] <alecu> mmcc: pong
[21:18] <mmcc> alecu, mandel noticed metadata corruption issues on the mac builds - sd quitting frequently, and it was fixed by killing the metadata and restarting…
[21:18] <alecu> mmcc: was that the cause of the disconnections?
[21:18] <mmcc> yes, he thinks so. in another channel, the possibility of thread issues (tritcask being thread unsafe) was brought up
[21:19] <chaselivingston> mmcc: not seeing any
[21:19] <alecu> mmcc: are we using threads in SD?
[21:19] <ralsina> mmcc: do you have the bundle? I just can't login anymore on mac using sources :-(
[21:19] <alecu> mmcc: well, I know we are using them for the macfsevents
[21:19] <mmcc> so, macfsevents uses threads and I wanted to see if we were messing that up somehow
[21:19] <mmcc> right
[21:19] <mmcc> chaselivingston: ok, thanks
[21:19] <dobey> sigh; this rain sucks
[21:19] <alecu> mmcc: but I checked that they were being used in a thread safe way
[21:20] <alecu> mmcc: I mean, when the first macfsevents branches landed.
[21:20]  * alecu checks again
[21:21] <mmcc> ralsina: the last built bundle is here:  http://ubuntuone.com/1CENTFFDneAt3ESZdZmcgd
[21:21] <ralsina> mmcc: thanks
[21:23] <chaselivingston> alright guys, it's EOD for me, I'll continue testing and talk with you all tomorrow
[21:23] <mmcc> thanks chaselivingston
[21:26] <ralsina> AHA!
[21:26] <ralsina> there is something weird going on with process launching on darwin
[21:29] <mmcc> ralsina: do tell…
[21:31] <ralsina> mmcc: the only way I found to login is to start all three processesmanually at the same time
[21:31] <ralsina> mmcc: u1cp can't start ubuntu-sso-login for some reason
[21:31] <mmcc> that is really weird. are you using an old buildout?
[21:31] <ralsina> mmcc: it's suppose to be the latest, I'll update it
[21:32] <mmcc> well, if it's the new one, no need to do that. but I'm curious about your logs
[21:32] <alecu> mmcc: I don't find anything wrong in the way we handle callbacks from the macfsevents thread.
[21:32] <mmcc> btw, I have to go help with/eat lunch, but I'll check back every once in a while
[21:32] <mmcc> alecu, me neither. provided the main event loop thread is also what's supposed to be touching the metadata
[21:33] <alecu> mmcc: yes, that's for sure.
[21:33] <dobey> ok, i need get get out of my office for a while. later all!
[21:33]  * ralsina is starting to worry. A lot.
[21:34] <alecu> mmcc: I don't think this might be related, but we found some issues on the handling of metadata recently. Some metadata having an empty value where a python pickle was expected.
[21:34] <alecu> this was on Precise. Or was it on windows?
[21:34]  * alecu looks.
[21:39] <alecu> it looks like it happened on Ubuntu. We should ask rye, but it's 1am in kiev, so we might ask him tomorrow as well...
[21:58] <ralsina> alecu, mmcc: the share file tab on darwin is broken somewhere inside IPC, the callback triggered by getting the published files list from sd is never triggered. I don't much care, we can just hide that tab for this release
[21:58] <ralsina> or leave it broken
[21:59] <ralsina> I am more concerned about the issues logging in. I managed, but had to fiddle with manually starting daemons, which is not a good idea
[21:59] <alecu> ralsina: how did you fix the ssl issue?
[21:59] <ralsina> alecu: removed storage-protocol from site-packages added its sources to the PYTHONPATH
[21:59] <ralsina> so tomorrow priority #1 is ensuring login / logout works
[21:59] <ralsina> I will send a mail about it to mandel, mmcc
[22:00] <alecu> ralsina: but that was running from trunk, or from mmcc's latest installer?
[22:00] <ralsina> alecu: trunk
[22:00] <alecu> ralsina: ok. I can run u1cp from trunk, but I get the loading screen from death
[22:01] <alecu> sorry
[22:01] <alecu> ralsina: I meant: I can run u1cp from the latest bundle but...
[22:01] <ralsina> alecu: hmmm... do you have credentials?
[22:03] <alecu> ralsina: seems like I don't
[22:04] <alecu> but this seems to be dying when doing the tcpactivation.
[22:04] <ralsina> alecu: start syncdaemon manually
[22:04] <ralsina> alecu: there's some sort of ... thing... going on there
[22:05] <ralsina> and it worked better a few days ago
[22:07] <alecu> ralsina: seems like SD was started OK. It's when starting the SSO that's failing
[22:08] <ralsina> alecu: I had sso shut down, then when I started SD u1cp got the "no creds" signal and unblocked itself
[22:08] <ralsina> alecu: so, maybe start SSO manually, then SD
[22:10] <ralsina> mmcc: sent you alecu and mandel a mail about priorities for the next 24 hours. Comments encouraged!
[22:13] <Costeelation> hi
[22:13] <Costeelation> anyone speak spanish?
[22:14] <Costeelation> someone?
[22:17] <ralsina> Costeelation: sí
[22:17] <ralsina> Costeelation: que necesitas?
[22:17] <Costeelation> a si mira,
[22:19] <alecu> ralsina: "File Sync in progress..."
[22:19] <alecu> ralsina: it's going thru a UDF with photos, it might take some time :-)
[22:19] <ralsina> alecu: good
[22:19] <ralsina> alecu: what did you do?
[22:20] <Costeelation> es q queria saber si las carpetas compartidas, por decir tu compartes una conmigo, q tenga mmm un archivo, y entonces esa carpeta aparezca en mi carpeta local de ubuntu one, ya noté q aparece en la web pero no en la carpeta local si me entendiste? es posible en este momento o esta caracteristica aun  no se ha agregado
[22:27] <Costeelation> ralsina, si entendiste? :S
[22:27] <ralsina> No va a aparecer en tu carpeta Ubuntu One
[22:28] <Costeelation> aaa bno
[22:28] <ralsina> Costeelation: pero si abres el control panel, vas a ver la carpeta que te compartieron, y puedes elegir sincronizarla
[22:28] <ralsina> y de ahí en más la vas a tener
[22:30] <Costeelation> bueno muchas gracias lo intentare haber si se sincroniza :)
[22:31] <ralsina> Costeelation: cualquier cosa pregunta por aquí nomás
[22:34] <ralsina> I'm beat. See you all early tomorrow
[22:35] <alecu> ralsina: mmcc: it seems u1sdtool is starting many copies of SD
[22:36] <alecu> ralsina: mmcc: I wonder if the u1cp does the same.
[22:36] <mmcc> ok, back. ralsina I saw your email and basically agree. I can start looking at the login stuff now and try to give mandel a head start
[22:36] <alecu> ralsina: mmcc: that may explain the metadata corruption we are seeing.
[22:36] <mmcc> catching up
[22:36] <mmcc> I haven't seen multiple SD copies…
[22:37] <mmcc> that'd definitely explain that though. looking…
[22:37] <alecu> mmcc: http://pastebin.ubuntu.com/1172733/
[22:38] <ralsina> mmcc: ok
[22:38] <ralsina> alecu: indeed it would
[22:38] <mmcc> yeah, that should not be happening
[22:38] <ralsina> alecu: if tcp activation is failing to detect there is a copy running, that may explain the multiple copies
[22:38] <mmcc> alecu, any tips on reproducing that?
[22:38] <ralsina> mmcc: did you switch it to unix domain sockets these days?
[22:39] <ralsina> mmcc: I remember reading about it :-)
[22:39] <alecu> mmcc: when I use "u1sdtool -s" from your bundle, sometimes it prints "Another instance is running"
[22:39] <mmcc> ralsina: it has always been using unix domain sockets…
[22:39] <alecu> mmcc: that means that SD was started again.
[22:39] <mmcc> alecu: yeah I've seen that too
[22:39] <mmcc> alecu aha. ok
[22:40] <alecu> hhmmmm...
[22:40] <alecu> maybe not.
[22:40] <alecu> :-(
[22:40] <mmcc> aside from adding u1sdtool, I'm not sure what else might have changed wrt. login since last week
[22:40] <alecu> I just got that message again, but no new copies.
[22:42] <mmcc> so ralsina, I think I'm going to disable the root daemon installation for release. definitely lower priority than sorting all this out
[22:42] <ralsina> alecu: you can try to just start it manually a few times. It should detect it and not do it
[22:42] <ralsina> mmcc: ok
[22:42] <ralsina> mmcc: we can do another release in a couple of weeks anyway
[22:42] <mmcc> ralsina: yep
[22:43] <briancurtin> i'm out of here. i will be seeing byte strings in my nightmares.
[22:44] <briancurtin> they will ruin my dreams of unicode
[22:44] <ralsina> briancurtin: hehe it's funny 'cause it's true
[22:44] <alecu> ralsina: if I run SD from the cmdline, it detects that it's already running, and quits. Except that it did not do that at least two times, because I have three copies running from before.
[22:45] <ralsina> alecu: so it looks like a race condition :-/
[22:48] <mmcc> looking for exactly how we launch SD again. I wonder if there's something going on wrt the fact that it shows up as 'ubuntuone-syncdaemon' instead of "UbuntuOne Syncdaemon.app" in the process list
[22:49] <mmcc> like, am I not launching it the way I thought I was launching it? are there two ways?
[22:49] <mmcc> let's see…
[22:51] <mmcc> ralsina: when you were having problems running sub-processes when running from source, is it possible that you hadn't sourced the env-mac script and were getting errors related to python not being the buildout-wrapped python?
[22:51] <ralsina> mmcc: I did source it
[22:51] <ralsina> mmcc: without sourcing it, nothing worked
[22:51] <mmcc> ok, that's true. well, check that off
[22:53] <ralsina> mmcc: could be that I needed to export PYTHONPATH?
[22:53] <ralsina> mmcc: I was just setting it in the CLI
[22:54] <ralsina> saly my son is now watching power rangers dino thunder on the mac mini ;-)
[22:54] <mmcc> hrm, so - which one of platform/tools/perspective_broker and platform/ipc/perspective_broker is used to launch SD... either one could be doing it
[22:54] <mmcc> I wonder why that last line of text got cut off there
[22:54] <mmcc> hrm, so - which one of platform/tools/perspective_broker and platform/ipc/perspective_broker is used to launch SD... either one could be doing it
[22:54] <mmcc> hrm, so  which one of platform/tools/perspective_broker and platform/ipc/perspective_broker is used to launch SD... either one could be doing it
[22:54] <mmcc> there we go
[22:55] <mmcc> ralsina: you should not need to export, it'll be inherited.
[23:06] <chaselivingston> ping mmcc: back on for a while if you need help testing anything
[23:06] <mmcc> pong chaselivingston thanks… knee deep in some IPC code right now, will let you know
[23:06] <chaselivingston> mmcc: cool, sounds good
[23:13] <chaselivingston> mmcc: is there a way to see upload transfer speed w/ u1sdtool?
[23:14] <mmcc> chaselivingston: I don't think so. I think you're stuck with --current-transfers in a loop and a spreadsheet (aka pain)
[23:15] <chaselivingston> mmcc: cool, was just wondering, no worries
[23:21] <alecu> wow! SD just died saying:
[23:21] <alecu> Fatal Python error: GC object already tracked
[23:21] <alecu> Abort trap: 6
[23:21] <alecu> ughhh
[23:22] <mmcc> alecu: yeah mandel saw that earlier too
[23:22] <mmcc> this came up wrt that:  http://pyrit.wordpress.com/2010/02/18/385/
[23:22] <mmcc> but it doesn't seem to point at a solution
[23:23] <alecu> mmcc: SD threw a fsevents warning line just before dying
[23:23] <ralsina> well, fsevents is the only pyobjc we are using, right?
[23:24] <mmcc> alecu: yes, chaselivingston saw that too, and that warning is not likely to be related. it probably shouldn't even be a warning
[23:24] <mmcc> more like 'info'
[23:24] <mmcc> ralsina: not pyobjc but C extension, yes
[23:27] <alecu> so, if I run two copies of SD on mac *very* close in time, both will start.
[23:27] <mmcc> pyobject_del not called in macfsevents
[23:27] <mmcc> alecu, that's interesting because the share links tab appears to be repeatedly calling ipc_client_connect
[23:27] <alecu> and they both end up dying with that fatal python error eventually.... within 10 to 20 minutes.
[23:28] <alecu> mmcc: interesting, yes!
[23:29] <mmcc> I put a traceback.print_stack() in platform/ipc/unix: get_u1_domain_socket(), and I get this: http://paste.ubuntu.com/1172800/
[23:30] <mmcc> note that the first traceback is from share_links_search, but the next several are from share_links.py, lines 85, 86, 88 , 89
[23:30] <mmcc> look like each line ends up trying to start syncdaemon…?
[23:35] <mmcc> yeah, they all do a test for UbuntuOneClient is_connected and they must each fail because they continue on to tell it to connect
[23:36] <mmcc> it looks like maybe ipc_client's UbuntuOneClient needs a 'connecting' state as well as a 'connected' state
[23:36] <mmcc> since it does a self.client = yield ipc_client_connect(blah) in connect() and also uses (self.client is None) to represent whether or not it's connected
[23:37] <mmcc> so if that hasn't returned instantly, multiple calls will get in there
[23:37] <ralsina> mmcc: so, the thing is, those requests are done at launch instead of when the tab is opened like most others
[23:37] <ralsina> mmcc: and good catch, BTW :-)
[23:38] <mmcc> ralsina: kudos all around