[06:25] <derpella> Hello. I am trying to make a new account, and it seems I am stuck on "gathering informations"? The computer is slow, but it takes like 10 minutes... Should I wait, or is it some error?
[06:27] <derpella> I hope it doesn't use Flash?
[08:38] <rye> morning!
[08:46] <jacobrask> Hi! Is this the communication channel for actual ubuntu one development or are most of you guys just hanging out here for support and helping out with the API and such?
[08:59] <mandel> morning all!
[09:00] <mandel> jacobrask, some of use are the actual developers of the clients on the diff platforms
[09:00] <mandel> derpella, no, it does not use flash, everything is python + Qt
[09:02] <mandel> derpella, there are apparently some issues with the ssl certs for when ever you create an account, that might be hitting you... the server guys are trying to get this fix asap
[09:03] <derpella> thanks, I'll try making account on website...
[09:04] <mandel> derpella, please do and let me know if it works, if it doesn't we can take a look at your logs and look closer at what is going on
[09:06] <derpella> :)
[09:07] <jacobrask> mandel: ok, nice
[09:23] <derpella> mandel: sorry for asking, but what is terminal command, so I could have logs?
[09:24] <derpella> because now I am stuck on choosing folders, with this gathering data
[09:28] <derpella> ubuntuone-control-panel-qt  <- well, I guess it works
[09:31] <derpella> now it crashed- IPCError
[09:32] <JamesTait> G'mornin' all! :)
[09:36] <mandel> derpella, so, you can set U1_DEBUG in the terminal before you launch it to get logs in stdout, but there are logs in your XDG_CACHE folder
[09:36] <mandel> derpella, which will have all operations from sso, control panel and the sync daemon used by u1
[09:37] <mandel> derpella, can you paste the ipc error you got in paste.ubuntu.com so I can take a look?
[09:37] <derpella> it'd be
[09:37] <derpella> ubuntuone-control-panel-qt --U1_DEBUG ??
[09:38] <derpella> whhhere is the folder?
[09:38] <derpella> in /var/log?
[09:40] <mandel> derpella, the logs folrder is in ~/.cache/ubuntuone/
[09:42] <mandel> derpella, and setting the debug to stdout is via the env variable U1_DEBUG
[09:42] <mandel> is not an an argument :)
[09:42] <derpella> erm... what do?
[09:43] <derpella> I am not a coder nor a specialist
[09:44] <mandel> derpella, ok, lets look at the logs that are already there, did you find the ~/.cache/ubuntuone/ folder?
[09:44] <derpella> sure...
[09:44] <derpella> well, only interesting things were in the file syncdeamon-exceptions
[09:46] <mandel> derpella, surely you have more info, can you paste the result of ls ls ~/.cache/ubuntuone/log/
[09:46] <mandel> derpella, sorry just one ls :)
[11:09] <gatox> good morning!
[11:10] <mandel> gatox, morning!
[11:10] <gatox> mandel, hi
[11:10] <gatox> mandel, today is your review day? :P (just a small one)
[11:13] <mandel> gatox, sure, shoot!
[11:13] <gatox> mandel, thx https://code.launchpad.net/~diegosarmentero/ubuntu-sso-client/network-page/+merge/105517
[11:15] <mandel> gatox, why self.next() self._next_id = -1 ?
[11:16] <gatox> mandel, we use that everywhere..... we set always in _next_id which the next page should be.... then we told the wizard to move to the next page with next() and we set again to -1 so it's respect the value that the qwizard api return when doesn't know to which page to move
[11:16] <gatox> we don't have a linear flow...... that's why
[11:16] <mandel> ok
[11:24] <mandel> gatox, having to run tests on 3 platforms does not scale..
[11:27] <gatox> mandel, jejeje nop
[11:27] <mandel> gatox, and wait to see my objective-c code..
[11:28] <mandel> gatox, +1
[11:28] <gatox> mandel, great
[11:40] <gatox> mandel, i need to go to the bank and then do an errand..... this may take a while
[11:40] <mandel> gatox, ok, I'll let the other know, no worries
[11:41] <gatox> mandel, thanks
[12:07] <alecu> hello, all!
[12:07] <alecu> ralsina, ping
[12:07] <mandel> alecu, morning!
[12:08] <alecu> hi mandel! I should not be working today; but I forgot to tell ralsina about it, and to put it in canonicaladmin
[12:09] <mandel> alecu, ok, it is because of uds, right?
[12:09] <mandel> alecu, I wanted to show you my cool fsevents server working, but I'll let you rest :)
[12:09] <alecu> mandel: awesome!
[12:09] <alecu> mandel, got any url?
[12:10] <mandel> alecu, I need to push it to lp, and I wont so that you rest :)
[12:10] <alecu> mandel, :-)
[12:11] <mandel> alecu, I need to get the twisted side working to get the events as json, but security bia getpeeredid is in place, events dispatching per user is there and i'm testing it by leaving it running while I compile things etc..
[12:12] <alecu> mandel, how do you handle "dropping" events when the buffers are full? (or how do you plan to do it)
[12:14] <alecu> mandel, the main point of that is that the root daemon should never block on writing to the children SDs
[12:14] <mandel> alecu, we get an special event which is the dropped events one.. I think we should try and get as much info as possible from that and pass it to syndaemon to deal with it, maybe we have to do a local rescan or somehting
[12:15] <alecu> mandel, well, that's not the point that worries me....
[12:15] <alecu> mandel, that's a slighly different case
[12:15] <mandel> alecu, I hope it does not, I'm using grand central dispatch for the dispatching of the events to the users, that is, to perform the IO to the socket
[12:15] <alecu> mandel, ok, I need to read more about GCD
[12:16] <mandel> alecu, I need to do some proper testing to see when this events are dropped and how to deal with that
[12:16] <alecu> mandel, right.
[12:16] <alecu> mandel, but we should keep in mind that we have two cases of events being dropped.
[12:17] <alecu> mandel, first one is the one you just mentioned: the kernel dropped some events because some of the root daemons reading from /dev/fsevents was slow.
[12:17] <mandel> and we are slow.. true
[12:17] <alecu> mandel, second one is "our" root daemon being the slow one because it blocked on sending stuff to the child Syncdaemons.
[12:18] <alecu> mandel, so, I think we should not block on sending stuff to our SDs.
[12:18] <mandel> alecu, I need to do some experiments, if the GCD adding operations to the q is fast and does not block us we should be dealing with events fast enough, but I need to do some stress testing on this
[12:19] <alecu> mandel, adding "fast" is not enough, because the queue might be full.
[12:19] <alecu> mandel, we should use some syscall to check if the queue is full before adding. And drop some events to just one SD in that case.
[12:19] <mandel> alecu, I'm doing something simlar to the example here: http://en.wikipedia.org/wiki/Grand_Central_Dispatch
[12:19] <alecu> mandel, awesome.
[12:20] <mandel> alecu, with a diff api by creating a q
[12:20] <mandel> alecu, I use a q because I do not want to have events sent in parallel and then loose the correct order which would be very bad..
[12:21] <ralsina> just a quick note on ssl: IS says the ssl may have ordering issues. i could not reproduce via vpn and a user says it only happens with turkish locale
[12:21] <mandel> ralsina, alecu is looking for you
[12:21] <alecu> mandel, using a queue is fine. But we should make sure that the queue has some fixed limit, and that we don't block if we exceed it.
[12:22] <alecu> hi ralsina!
[12:22] <alecu> ralsina, those are some awesome news on ssl
[12:22] <ralsina> i am from Android so i type Slow. alecu, phone?
[12:23] <alecu> ralsina, sure.
[12:23] <alecu> let me find it.
[12:23] <ralsina> alecu it seems gnutls says cert is invalid, too
[12:25] <mandel> ralsina, also, gatox went to the bank to deal with some errands that will take time
[12:29] <mandel> ok, lunch time for me
[12:29]  * mandel lunch
[12:39] <alecu> mandel, btw: what are you using to code the root daemon? python too? objective-c?
[12:41] <alecu> oh, mandel is lunching.
[12:42] <alecu> btw, all: I'm taking the day off, but I'll be around a little bit today at 16z for tech leads call.
[13:24] <ralsina> Hmmm twisted's default reactor for next version is epoll. Would that break anything for us?
[13:40] <mandel> alecu, objective-c
[13:40] <mandel> ralsina, the only place in which the default reactor is used in in syncdaemon and in theory should be ok
[13:44] <gord> hey all, ubuntu one decided to delete the contents of ~/Ubuntu One/ when i setup a new home directory to sync with it. which propagated to my other machines that sync. any way to get this stuff back? or has u1 deleted them for good
[13:45] <mandel> gord, there is indeed a way to get them back, ralsina is there paperwork to be done for this? ^^
[13:56] <ralsina> gord: ask duanedesign or joshuahoover
[13:56] <ralsina> gord: they can get it back for you, don't worry. Also, they are not deleted, they should be in your recycle bin or however it's called
[13:58] <duanedesign> gord: i will PM you for more info
[14:12] <mmcc> good morning folks
[14:26] <ralsina> good morning mmcc!
[14:34] <mmcc> question - I wrote a script to run the network detection watcher and just print whenever it sees a real change - useful for testing IRL, but not a unit test. I think it should be in the SSO client tree somewhere, but where?
[14:36] <ralsina> mmcc: we have one script of similar scope in sso, let me dig
[14:37] <ralsina> mmcc: well, can't find it
[14:38] <ralsina> mmcc: remember that script we had that connected to ussoc and showed the UI? Where was it?
[14:38] <mmcc> hmm
[14:38] <ralsina> oops, mandel ^
[14:38] <ralsina> mmcc: miscmpletion :-)
[14:38] <mmcc> ralsina: no prob. I think I've actually used that script
[14:38] <mmcc> no wait, no I haven't
[14:40] <ralsina> mmcc: we hav one like it at ./ubuntu_sso/qt/tests/login_u_p.py
[14:40] <ralsina> mmcc: so maybe there
[14:40] <ralsina> and the one I remembered was ./ubuntu_sso/qt/tests/show_gui.py
[14:41] <ralsina> mmcc: so, there but without test in the name :-)
[14:41] <mmcc> oh, ok - so maybe mine fits better in ubuntu_sso/networkstate/tests/ , just without test in the name
[14:42] <mmcc> because it doesn't use qt
[14:42] <mandel> ralsina, let me check, I think it was in sso under tests..
[14:42] <ralsina> mandel: found it already :-)
[14:43] <mmcc> now ralsina has anti-lag
[14:43] <ralsina> right, I know what you will type now.
[14:43] <mandel> ralsina, cool :)
[14:44] <mandel> mmcc, ralsina, so I have the fsevents server cheking for the ownership of a path by the connected client.. atm I'm not checking at the group ownership, is that ok for the time being?
[14:45] <ralsina> mandel: sure
[14:45] <mandel> ralsina, sweet! I'm nearly done, I have to write the example for the python code.. oh and I'm not using launchd to create the handle for the domain socket... I'm to lazy to write that atm
[14:46] <ralsina> mandel: baby steps
[14:46] <mmcc> mandel: sounds good. that launchd checkin stuff is boilerplate that might be best to add later.
[14:47] <mmcc> in my experience, testing during development with launchd is a pain. adds a bunch of steps between compile and run
[14:47] <ralsina> mmcc, mandel: we need to start planning the packaging, now that we have a binary that does something ;-)
[14:47] <mmcc> ralsina: I have some thoughts. maybe I should write them up in an email?
[14:47] <mandel> ralsina, mmcc, let me push this to a branch and we can take a look if you have the time
[14:48] <mmcc> mandel, got nothing but time for you man
[14:48] <mandel> mmcc, specially with you, you might cry at my code (I have not written objective-c in ages...)
[14:48] <ralsina> mmcc: please
[14:48]  * ralsina likes long mails explaining stuff
[14:49] <ralsina> in the right dose, of course
[14:49] <mmcc> ralsina: ok, cool
[14:51]  * mmcc adds draft packaging plan to TODO list
[14:51] <mandel> mmcc, I'm going to push the code with the xcode userdata.. I have no clue which files to ignore, sorry
[14:52] <mmcc> mandel: it's a huge pain, don't worry. For the actual project tree I want to suggest that we don't use XCode...
[14:54] <mmcc> you can always use it yourself and just not check in your .xcodeproj bundle, but I want to propose using a makefile - XCode's files are annoying to read & diff
[14:54] <ralsina> mandel, mmcc, dobey, gatox, thisfred: standup in 6'
[14:54] <ralsina> briancurtin: too of course
[14:54] <briancurtin> i'm already typing, not falling behind this time
[14:54] <ralsina> thisfred: you are on sprint this week or next? // need to take memory pills...
[14:55] <mmcc> and for a project this size and with no UI, I don't think it adds much to use XCode for building
[14:55] <thisfred> ralsina: DONE: traveled to London TODO: sprint ;)
[14:55] <ralsina> thisfred: ok, excused ;-)
[14:55] <mmcc> me
[14:55] <ralsina> mmcc: early! :-D
[14:55] <mandel> mmcc, probably not, it just makes it easier to build I hate make files :P
[14:56] <mmcc> mandel, I'm with you on makefiles, but at least they're readable. XML gives me hives
[14:56] <mmcc> mandel especially *undocumented* XML build files!
[14:56] <mandel> mmcc, yes, those are even worse..
[14:57] <mandel> mmcc, you can find some initial code in lp:~mandel/+junk/fsevents /cc ralsina
[14:57] <mmcc> mandel, and per-user preference files that XCode saves in the .xcodeproj that cause spurious diffs... ugh
[14:57] <mmcc> thanks mandel, looking in a sec
[14:57] <mandel> mmcc, yeah, those are the ones I wanted to add to the bzr ignore..
[14:58] <mandel> mmcc, we can go over the code after the standup in mumble, there are some thinks I want to discuss as well as the fact that right now the config iss hardcoded with my paths...
[14:58] <mmcc> mandel: sonds good
[14:58] <mandel> which ideally is sent my the clients when connected.. but that is work to be done in the python side..
[14:58] <mmcc> sonds very good indeed
[14:59] <mmcc> mandel, ack
[15:01] <dobey> ralsina: lies
[15:01] <ralsina> me
[15:01] <briancurtin> me
[15:02] <mandel> me
[15:03] <ralsina> mmcc, mandel: use cmake
[15:03] <mmcc> I think gatox is out?
[15:03] <mandel> ralsina, you kde people..
[15:03] <ralsina> mandel: he
[15:03] <ralsina> mandel: ok, use xmkmf
[15:03] <mandel> yeah, gatox when to the bank.. so he is probably in a q
[15:03] <ralsina> ok, then mmcc go
[15:03] <mandel> ralsina, I'll write my own version of make in ruby!
[15:03] <mmcc> DONE: cleaning up test suite for darwin network detection, refactored
[15:03] <mmcc> a bit for testability. Saw a really big ostrich at the San Antonio Zoo.
[15:03] <mmcc> TODO: finish tests for network detection, land that branch, write up
[15:03] <mmcc> draft OSX packaging plan
[15:03] <mmcc> BLCK: no
[15:03] <ralsina> mandel: you mean rake?
[15:04] <mmcc> next, uh? ralsina
[15:04] <ralsina> DONE: weekend, reviews, SSL debugging, 1-1s TODO: fix something, reviews, help around, tech leads call BLOCKED: no NEXT briancurtin
[15:04] <briancurtin> DONE: propose installer branch, screwed up jenkins branches but just got them in this morning
[15:04] <briancurtin> TODO: look into the get_rootdir exception people have been reporting
[15:04] <briancurtin> BLOCKED: None
[15:04] <briancurtin> NOTE: i have half days on tuesday, thursday, and friday
[15:04] <briancurtin> NEXT: mandel
[15:04] <mandel> DONE: Goot root process to filter peruser getting the user uid when connected. Now the service send a json representation of the event over the domain sockets (need a good separator for each message, I think \n is good since there are no \n in any of the strings). Added security to ensure that a client does not request events from paths that he does no own.
[15:04] <mandel> TODO: talk with mmcc about the code, find room for improvement. Ask ralsina about some copyright in a lib used to the example (we could use many others)
[15:04] <mandel> BLOCKED: no
[15:04] <mandel> COMMENTS: jenkins is not yet there due to some lint issues in the skipJenkins branches :*(
[15:04] <mandel> I think I'm last, right, dobey is not working.. is he?
[15:05]  * ralsina would have to check and canonicaladmin is too stupid
[15:05] <ralsina> doing it anyway
[15:05] <ralsina> That smart site shows the leave requests in alphabetical order since I joined the company. And it only works or page 1.
[15:06] <ralsina> So to know who's on leave I have to ask for the monthly company-wide request board
[15:06] <ralsina> So yes, rodney is on swap today ;-)
[15:06] <mandel> ralsina, alphabetical order is what hr uses, is soooo much better!
[15:07] <mmcc> er, what does "on swap" mean?
[15:07] <ralsina> mandel: alphabetical order of the first name, so I nly see Alejandro's requests for 2011
[15:08] <ralsina> mmcc: when we make you travel on a weekend, you get a day off
[15:08] <ralsina> mmcc: so we don't feel like slave drivers, and all that
[15:08] <mandel> mmcc, you moved a national holiday from its day to an other, for example, tom is a holiday in madrid I swaped it for july + what ralsina said
[15:08] <briancurtin> is there any process for updating dev-tools on tarmac? a fix of mine was merged into dev-tools, but a u1client branch which uses it claims to fail because it can't import a name (skipIfJenkins)
[15:08] <mmcc> ah, ok.
[15:09] <mandel> briancurtin, you dev-tool branch did not land due to a lint error
[15:09] <briancurtin> mandel: i just fixed that one
[15:09] <ralsina> briancurtin: may need to be done manually, since dobey is on swap, ping sidnei
[15:09] <mandel> briancurtin, oh.. then we have to wait for dobeys machine to pick up nightlies..
[15:09] <mandel> ralsina, is tarmac not longer in _dobeys servers?
[15:10] <briancurtin> if it'll eventually just work, that's probably fine
[15:10] <ralsina> mandel: since like, months ago!
[15:10] <ralsina> mandel: it's in canonistack
[15:10] <dobey> tarmac is cloudy
[15:10] <mandel> ralsina, he, good news!
[15:10] <mandel> dobey, cylon, go back to the swap day!
[15:11] <mandel> mmcc, give me 10 mins, mumble and we take a deep look at the code for the fsevents, is that ok?
[15:11] <mmcc> mandel: sure that works
[15:14] <beuno> ralsina, any ideas why I wouldn't be getting notifications for uploads/downloads to U1 in ubuntu?
[15:16] <ralsina> beuno: nope
[15:16] <gatox> FINALLY BACK!!!!
[15:17] <salgado> beuno, "show_all_notifications = True" on syncdaemon.conf?  (just a wild guess, really)
[15:17] <ralsina> beuno: am I filling you with confidence or not? ;-)
[15:17] <gatox> i hate errands!!!!!!!!
[15:17] <beuno> salgado, ralsina, it's set to true
[15:18] <ralsina> gatox: err-ands are better than err-ors. Basic boolean algebra.
[15:18] <gatox> ralsina, jejeje
[15:18] <ralsina> beuno: then there may be something on syncdaemon-exceptions.log
[15:18] <ralsina> beuno: we have had a few similar things in the past
[15:19]  * gatox goes back to work.... is going to be a long day :P
[15:19] <beuno> ralsina, nada
[15:19] <beuno> empty
[15:19] <ralsina> beuno: weirder then
[15:21] <ralsina> beuno: we are a bit shorthanded today because of UDS swaps and stuff, can you remind me tomorrow?
[15:22] <beuno> ralsina, sure
[15:22] <ralsina> beuno: thanks
[15:25] <mandel> mmcc, I'm back, mumble?
[15:25] <mmcc> mandel, just a sec, testing
[15:25] <mandel> sure
[15:44] <ralsina> gatox: can you assign bug #986328 to rtgrant for feedback on texts for the buttons (or if we should just keep them as they are?)
[15:44] <gatox> ralsina, yes
[15:44] <ralsina> gatox: thanks
[15:44] <ralsina> gatox: put a comment explaining exactly what you want feedback on, then assign to him
[15:53] <gatox> and now i have internet issues...... connecting with 3g..... not a very good day!
[15:54] <ralsina> gatox: take it easy
[15:54] <ralsina> gatox: go to a bar
[15:54] <ralsina> gatox: have some tea or something
[15:55] <gatox> ralsina, yeah...... trying to relax..... i'm with 3g now (in the closest bars internet is even worse than my 3g connection).... i'll debug some u1-client stuff that doesn't require internet in the mac..... and wait for fibertel to fix the problems
[15:55] <ralsina> Grmbl, so *units* should be transleatable. Makes sense.
[15:56] <ralsina> gatox: cool
[16:03] <mmcc> so mandel, you're using OS X 10.7 I see... your code doesn't build for me on 10.6, I'll take a look at the docs and see if it's something simple I can change in the build or if you've used any 10.7-only api...
[16:03] <mandel> mmcc, ouch!
[16:04] <mandel> mmcc, yes, I'm on Lion... I wonder what changed
[16:04] <mmcc> mandel, if you want the complete list: https://developer.apple.com/library/mac/#releasenotes/General/MacOSXLionAPIDiffs/index.html#//apple_ref/doc/uid/TP40010630
[16:05] <mmcc> mandel: but that's just api, I'm getting complaints about missing objc properties too :\
[16:05] <mandel> mmcc, I tried to compile it with 10.6 support and threw a bunch of errors regarding the autorelease pool..
[16:06] <mmcc> mandel, I think maybe the @autoreleasepool{ ...} block thing is new in 10.7's ObjC
[16:06] <mandel> mmcc, could perfectly be, then I'll have to change all the memory management to the old style, not hard but boring
[16:18] <mmcc> mandel, I have an old XCode - @autoreleasepool was added for ARC (Automatic Reference Counting), which is supported in 10.6 and 10.7, but you need XCode 4.2 or later (and the corresponding new compiler) to compile code that uses ARC.
[16:19] <mandel> mmcc, does that mean that you just need to update xcode?
[16:19] <mmcc> mandel So I'll upgrade, but I also want to think a bit about whether or not we should use ARC. I've heard some complaints about it, but maybe it's not so bad in a relatively simple daemon
[16:19] <mmcc> mandel, yes
[16:19] <mandel> mmcc, I think removal is not a thing that should scares us much
[16:20] <ralsina> mmcc: I got a third hand rumour from UDS that running things as root is going to be harder/different in the next OSX, you heard anything?
[16:22] <mandel> ralsina, that thing is, them having iCloud might mean us being screwed up with rules like that..
[16:22] <ralsina> mandel, gatox: can I get a trivial review here https://code.launchpad.net/~ralsina/ubuntuone-control-panel/octetes/+merge/105684
[16:22] <gatox> ralsina, on it
[16:22] <ralsina> gatox: thanks
[16:22] <mandel> ralsina, sure, review day!
[16:23] <mandel> ralsina, people translate bytes?
[16:23] <ralsina> mandel: bytes -> octetes in french, for example
[16:23] <ralsina> le logiciel c'est 900000000 octetes
[16:23] <mandel> ralsina, ok, one single thing to that: putos gabachos!
[16:23] <gatox> everyday you learn something new :P
[16:24] <ralsina> sorry, octets
[16:24] <mandel> ralsina, but does not surprise me from a contry that telecharges things from the internet
[16:24] <ralsina> so they are boy bytes, not girl bytes
[16:24] <ralsina> mandel: you spaniards translate "mouse" and "cd"
[16:24]  * alecu loves "cederrón"
[16:24] <ralsina> mandel: and cd, you officially translate to cederrón which is unforgivable
[16:25]  * ralsina still doesn't understand why that ends in N
[16:25] <gatox> jeje
[16:25] <gatox> ralsina, +1
[16:25] <mandel> ralsina, gatox, because they tranalated cd-rom and not cd, there must be a cederran somewhere.. :P
[16:25] <ralsina> gatox: thanks!
[16:26] <ralsina> mandel: devederrón coming soon. Y blurayo
[16:26] <alecu> cedeerreuvedoble
[16:26] <gatox> jejejejejejeje
[16:26] <ralsina> pásame el reproductor de rayoazul!
[16:26] <mandel> ralsina, alecu, gatox, I prefer: uh!! se ve!
[16:26] <alecu> pues cógelo tu mismo.
[16:27] <mandel> lol
[16:27] <ralsina> ah, no hace falta, lo apareo mediante dienteazul
[16:27]  * ralsina will stop now
[16:27] <gatox> LOOOL
[16:28] <mandel> ralsina, I'll run the tests on windows and +1
[16:28] <ralsina> mandel: awesome. I am not adding tests for this because... I have no idea how
[16:28] <mandel> ralsina, I'm not requesting tests for this because I'm not after 80% :)
[16:29] <ralsina> hehe
[16:30]  * mandel lost a contact lens as soon as he switch to the windows machine...
[16:30] <mandel> my boy is trying to say something..
[16:31]  * mmcc went away for a sec and thought I came back to the wrong channel
[16:31] <mmcc> ralsina, I know they're doing new things with an application sandbox, and developer certificates. I need to read up more about how that'll affect us.
[16:32] <mmcc> ralsina: I think it will still be possible for some time to do what we want, but the user experience might be a little less smooth - ie, we may not be able to be in the Mac App Store, because we probably can't run in the app sandbox
[16:33] <mandel> ralsina, approved
[16:34] <mmcc> ralsina: But I think running as root should be OK as long as we use the new launchd API for installing our root daemon
[16:34] <ralsina> mmcc: right
[16:34] <ralsina> mmcc: I suspect we may have to do a sandboxed version that is less useful, for the store
[16:35] <mmcc> ralsina, I'll make a note to see how much less useful it'd have to be. At one point I thought you had to pre-register (at app submission time) a separate security capability for each file you want to access...
[16:36] <ralsina> mmcc: ok, that would be massively less useful
[16:36] <mmcc> ralsina: it would be all the way less useful
[16:36] <ralsina> mmcc: unless we could register a folder
[16:36] <mmcc> ralsina: and then not access its files
[16:36] <ralsina> hehe
[16:36] <mmcc> ralsina: but before we go too far here, they have changed things a bit and I need to catch up
[16:37] <ralsina> mmcc: great, thanks
[16:37] <mmcc> ralsina: I'll make a list of things to get urbanape to find out about at WWDC...
[16:38] <ralsina> mmcc: even better :-)
[16:47] <ralsina> mandel, gatox: how would you, who have dirtier minds, create a file that fails as in bug #959447?
[16:47]  * gatox looking....
[16:48] <gatox> ralsina, something like this? my_私のファイル.txt
[16:48] <gatox> ralsina, i don't know if your irc client support unicode chars
[16:48] <ralsina> gatox: yes, it does
[16:48]  * ralsina tries
[16:49] <ralsina> ohhhhh got it
[16:49] <ralsina> if I do walk('.') it works, but walk(u'.') fails :-)
[16:50] <gatox> so.....with u is trying to convert automatically or something instead doing a decode
[16:51] <ralsina> gatox: yes, something like that
[16:51] <ralsina> Since in this case we don't care about the names, I could just use ascii paths
[16:51] <mandel> gatox, you are faster..
[16:51] <mandel> ok, EOD for me, also I need to get my glasses..
[16:51] <ralsina> mandel: all that ninja training
[16:51] <mandel> lol
[16:51] <ralsina> bye mandel! Good luck!
[16:51] <gatox> jejeej
[16:51] <gatox> mandel, bye!
[16:52] <mmcc> bye mandel
[17:03] <urbanape> mmcc: did you see Daniel Jalkut's recent blog post on the catch-all entitlement?
[17:04] <urbanape> http://www.red-sweater.com/blog/2438/the-sandboxs-big-red-button
[17:04] <mmcc> urbanape: I'm reading his sandbox posts right now...
[17:04] <urbanape> it says temporary, but who knows
[17:05] <mmcc> urbanape: I'd skimmed it a while ago, and that's what I had in the back of my head when I said it'd changed recently...
[17:05] <urbanape> right
[17:06] <mmcc> so urbanape - the apple dev account / membership stuff has changed since I last paid for it - is there a separate mac vs. ios dev account? What membership do you have?
[17:07] <urbanape> there are separate.
[17:07] <urbanape> We've got a team membership for both.
[17:07] <urbanape> I'll invite you into the team for both, if you like.
[17:07] <mmcc> urbanape: yes, that'd be good, thanks!
[17:08] <mmcc> if there are only a few slots, you can leave me out of the ios membership I guess
[17:08] <urbanape> I don't know if you can have separate agents for mac and iOS, but if so, it would probably make sense to make you the agent on the mac side.
[17:08] <urbanape> nah, I don't think we're close to any limit.
[17:08] <mmcc> ok, good. What's the agent's responsibility?
[17:08] <mmcc> (and who's the ios agent now? you?)
[17:08] <urbanape> the agent is the only person who can actually build releases for upload to the app store
[17:08] <urbanape> i'm the iOS agent
[17:09] <urbanape> I mean, you can share the credentials and stuff.
[17:09] <mmcc> ah, ok - so it's the account they use for the code signing then?
[17:09] <urbanape> but the agent creates the initial certs
[17:09] <urbanape> right
[17:09] <mmcc> got it
[17:09] <urbanape> ralsina: lemme know who else should be on the mac developer team
[17:10]  * ralsina reads backlog
[17:11] <ralsina> urbanape: doing develpment work on mac (even if ntot full time) gatox, me, alecu, mmcc, mandel
[17:11] <urbanape> k
[17:19]  * mmcc leaves for lunch
[17:38] <karni> Hiya all o/ Any Samsung S II users out there :)?
[17:39] <karni> If yes - check out the links at the bottom of https://bugs.launchpad.net/ubuntuone-android-files/+bug/993683
[17:39] <karni> oh, cool bot :)
[18:45] <ralsina> gatox: I am about to propose a unicode branch, you owe me one ;-)
[18:45] <gatox> ralsina, jeje ok
[18:58] <ralsina> gatox, briancurtin: reviews please? https://code.launchpad.net/~ralsina/ubuntuone-control-panel/fix-959447/+merge/105704
[18:58] <gatox> ralsina, on it
[18:58] <briancurtin> ralsina: looking
[18:59] <ralsina> thanks!
[19:07] <gatox> ralsina, mmcc can i have a review for tihs please? https://code.launchpad.net/~diegosarmentero/ubuntuone-client/refactor-remove-windows-linux/+merge/105705
[19:07] <ralsina> gatox: sure
[19:07] <mmcc> gatox, ok
[19:07] <gatox> thx
[19:11] <gatox> ralsina, i have this error running the tests of your branch on windows: v
[19:11] <gatox> http://paste.ubuntu.com/987703/
[19:11] <ralsina> gatox: that is nowhere near my code :-)
[19:12] <gatox> crap....... so we have something broken there
[19:12] <ralsina> gatox: well, it's the same file, so mybe it is my fault
[19:12] <ralsina> gatox: let me check
[19:12] <gatox> ralsina, let me test in trunk
[19:13] <ralsina> gatox: thanks
[19:14] <gatox> ralsina, it's in trunk too....... so approve
[19:15] <ralsina> gatox: and we need a second opinion on the trunk failure on windows to open a bug for it. Because that test is not even close to passing :-)
[19:17] <mmcc> gatox, in your branch, you have ubuntuone/syncdaemon/volume_manager.py line 1629 importing 'allow_writes' from linux -- I haven't tried but I bet this won't work on windows.  windows has its own os_helper.allow_writes()...
[19:19] <gatox> mmcc, the tests works on windows....... if you take a look, you will see that the line removed was: from ubuntuone.platform.linux import allow_writes
[19:20] <mmcc> gatox huh. you're right, I misspoke - but the new line says from 'ubuntuone.platform.os_helper.linux import allow_writes'
[19:20] <ralsina> gatox: look at line 511 of the diff
[19:20] <mmcc> gatox which was the problem I meant to point at
[19:20] <ralsina> gatox: you are importing a linux-specific thing in volume_manager.py
[19:21] <mmcc> gatox, should work if you just leave off ".linux", right?
[19:21] <gatox> ralsina, mmcc yes, i did it that way, to respect what was already there....... that was: from ubuntuone.platform.linux import allow_writes
[19:21] <ralsina> gatox: yeah, but it looks wrong :-)
[19:21] <mmcc> seems like it was a bug
[19:21] <mmcc> was it being tested?
[19:21] <gatox> mmcc, ralsina ok, fixing that
[19:22] <ralsina> other than that it looks good to me
[19:23] <gatox> ralsina, mmcc i'm not sure about this..... i mean..... who could know why this is being that this way??...... because the module is already importing allow_writes....... but for some reason, inside this function is being imported again in a really specific way
[19:23] <ralsina> gatox: bzr blame knows
[19:24]  * mmcc to the bzr gannotatemobile
[19:24] <ralsina> gatox: 99% of the time, the obviously broken code is broken obviously
[19:24] <ralsina> gatox: but yes, let's run a check on it with the guilty dev and not merge before getting info
[19:26] <gatox> ralsina, if i understand bzr blame correctly..... i need to talk with mandel :P jeje
[19:26] <ralsina> gatox: SURPRISE! ;-)
[19:27] <ralsina> gatox: mail him so you have an answer when you start tomorrow
[19:27] <gatox> ralsina, mmcc so..... i'll leave this branch on hold, until tomorrow, so i can ask mandel if that is there for any particular reason
[19:27] <mmcc> gatox: ralsina yep I see mandel's name there too. ok
[19:27] <gatox> ack
[19:32] <mmcc> ralsina, gatox I had another question about that branch, so I was going to add a comment - noticed it's still approved by ralsina, should I unapprove or just comment
[19:32] <ralsina> mmcc: you can't unapprove me anyway ;-)
[19:32] <gatox> mmcc, or you can wait
[19:32] <ralsina> mmcc: just comment, or put a needsinfo
[19:32] <gatox> mmcc, and i'll ask for a review to mandel tomorrow
[19:32] <mmcc> ok. sounds good
[19:33] <gatox> mmcc, i already sent him an e-mail asking about the import tihng
[19:33] <gatox> thing
[19:33] <mmcc> ralsina, I can *dis*approve you
[19:33] <mmcc> er, of you. that joke needs work
[19:33] <ralsina> mmcc: hehe
[19:33] <ralsina> disapproving of me is indeed almost required.
[19:34] <ralsina> but then again, you have a performance review in ... 11 months! ;-)
[19:34] <ralsina> tick tock
[19:35] <mmcc> ralsina: I have a big red clock counting down
[19:35] <mmcc> So my question about the branch: Why add a separate filesystem_notifier module for platform agnostic things that are related to filesystem notification? That seems like it should go in filesystem_notifications, either in __init__ or in a separate file that __init__ includes...
[19:36] <mmcc> It might seem picky, but I was a little confused by the naming. I saw changes to "notifier" and wondered where the files in "notifications" had gone
[19:39] <gatox> mmcc, yes, i added first inside filesystem_notificatoins, but
[19:40] <gatox> i wanted something that you can import always in the same way, and in that case that would be: from ubuntuone.platform.filesystem_notification import pyinotify....... but if you add the file inside filesystem_notification....... and linux.py and windows.py wants to import pyinotify..... you have a circular import and doesn't work
[19:41] <urbanape> dobey: have you tried gmap-pedometer? Last I recall, it provided elevation, and I think it can do a KML
[19:41] <urbanape> brb
[19:45] <mmcc> gatox, sorry, I don't see how it would cause a circular import.
[19:45] <mmcc> gatox so, "filesystem_notifier" has code borrowed from pyinotify that we use on all platforms, right?
[19:47] <gatox> mmcc, no..... mmm i'm looking at the circular import thing..... and it doesn't seem that is the problem..... but that happen when i was trying to run the tests..... maybe was a different approach..... i'll try to add it again inside filesystem_notifications.... i also prefer that place
[19:48] <mmcc> yeah it's a little nicer all in one place I think. As I look at it more, the module you mande 'filesystem_notifier' is pyinotify on linux, but on windows, it's just the bits of pyinotify we need to wrap the events coming from the code in the filesystem_notifications module...
[19:48] <mmcc> seems like it fits better together.
[19:49] <dobey> urbanape: wow, that site is awful :)
[19:49] <dobey> urbanape: and the gpx i got from making a route doesn't seem to have elevation there either
[19:50] <urbanape> did you turn elevation on?
[19:50] <dobey> yes
[19:50] <dobey> to "large" whatever the heck that means
[19:50] <urbanape> hm. sorry.
[19:51] <dobey> and i made a route in Harrisonburg, where there are obviously lots of hills :)
[19:52] <dobey> eh, oh well. i don't think google maps exposes elevation either, though google earth might
[19:53] <dobey> i doubt OSM has elevation for all points, either
[19:54] <briancurtin> ralsina: is there a way to cancel a holiday in canonicaladmin? the half-day i put in for yesterday is cancelled, i'll be here the full day
[19:54] <briancurtin> er, tomorrow...no idea how i thought "yesterday"
[19:54] <ralsina> briancurtin: if I have not approved it, yes, if I have, we have two paths
[19:55] <briancurtin> its already approved
[19:55] <ralsina> then we can 1) go to HR and ask them to change it
[19:55] <ralsina> or 2) you tell me what day you will really be gone, and I pretend we did it
[19:55] <ralsina> correctly
[19:56] <ralsina> in this case, you have a holiday on a sunday, so HR is probably a better idea
[19:56] <briancurtin> i'll email them to change it
[19:56] <ralsina> briancurtin: thanks
[19:59] <gatox> si?
[19:59] <ralsina> gatox: tal vez!
[19:59] <gatox> jeje wrong channel
[20:10] <gatox> mmcc, if i try to import pyinotify inside filesystem_notifications __init__ there is actually a problem..... i need to create a different module inside filesystem_notifications to do this..... i wasn't so crazy :P jeje
[20:11] <mmcc> gatox, what problem does it cause?
[20:13] <gatox> mmcc, we are importing linux/windows file from the __init__ and those files are trying to obtain pyinotify from the same init... is just necessary to do the import somewhere else...... the REAL problem, is how to do the import and avoid the lint issues (which shouldn't be there anyway..... but..... that's how the lint checker works)
[20:15] <mmcc> gatox, only windows uses the pyinotify_agnostic stuff, right?
[20:15] <gatox> and mac
[20:15] <mmcc> right
[20:15] <gatox> i think i figure it out
[20:16] <gatox> the real problem here, is that i need to do something kind of unnatural...... because the lint checker is going to complain... when it's actually something valid
[20:16] <gatox> but it seems to be working now
[20:18] <mmcc> I'm curious what unnatural things you had to do - not familiar yet with the corners of lint.
[20:18] <mmcc> you can just push it and tell me to look if you want
[20:19] <gatox> mmcc, i mean when it fails for a double import..... when you actually are importing something in the if part or the else part..... same problem here.....
[20:19] <gatox> mmcc, no problem...... it's not going to win! :P
[20:19] <ralsina> gatox, briancurtin: another easy review please! https://code.launchpad.net/~ralsina/ubuntuone-control-panel/fix-865688/+merge/105713
[20:19] <gatox> ralsina, ack
[20:19]  * ralsina is picking low-hanging fruit today
[20:19] <briancurtin> ralsina: looking
[20:20] <mmcc> gatox oh yeah, I ran into that, dumb lint.
[20:20]  * mmcc wants flow sensitive lint!
[20:35] <gatox> ralsina, +1
[20:42] <dobey> bah, i don't think i have time to do any real yard work before the storm gets here :(
[20:47] <ralsina> gatox: thanks!
[20:59] <toabctl> there a study about cloud services from faunhofer instutite. just for the record: http://www.sit.fraunhofer.de/de/cloudstudy.html . the study is in english
[21:17] <mmcc> hey, does twisted's UnitTest failUnlessRaises() not do what I think its docs say it does?
[21:18] <mmcc> I am testing a regular function that doesn't return a deferred and I want to test that it raises a NetworkFailException
[21:18] <dobey> mmcc: you should use self.assertRaises()
[21:19] <mmcc> dobey, ok. I saw that somewhere but it isn't in twisted's docs ... is it our extension? I'll go look
[21:20] <dobey> mmcc: it's in one of the TestCase implementations we inherit from
[21:20] <dobey> mmcc: probably either twisted, or unittest
[21:20] <dobey> mmcc: the failFoo APIs are for implementing assertFoo APIs, generally
[21:28] <mmcc>  dobey, thanks. it comes from unittest... still not clear why failUnless doesn't seem to match its docs tho
[21:29]  * dobey doesn't have good experiences with python docs
[21:31] <mmcc> OH. PEBKAC, of course you pass it the callable, not the *result* of calling the callable. yargh, docs are fine
[21:32] <dobey> and now, having issue with python testing libraries that have circular dependencies on each other
[21:33] <dobey> wish python had a LogicError for such a case
[21:36] <mmcc> dobey, aren't you off today?
[21:36] <dobey> yes
[21:36] <mmcc> oh, ok then
[21:39] <dobey> was looking at getting a few simple non-u1 packages (but which u1 stuff sort of depends on) packaged for python3 in ubuntu
[22:00] <mmcc> leaving now for a few hours, will be back this evening to wrap up my network state branch
[22:03] <gatox> ok...... i'm off for today...... brain is blocked!!! see you tomorrow!!
[22:11] <duanedesign> can someone help me with somewindows logs
[22:11] <briancurtin> duanedesign: i'm about to head out of here shortly, but i can take a look
[22:12] <duanedesign> \o/
[22:14] <duanedesign> briancurtin: https://pastebin.canonical.com/65993/
[22:14] <duanedesign> got some more i might email
[22:16] <briancurtin> duanedesign: that doesn't ring any bells, never seen anything like that...
[22:16] <briancurtin> duanedesign: i can take a deeper look tomorrow
[22:17] <duanedesign> briancurtin: yeah i emailed you
[22:17] <duanedesign> have a good night
[22:17] <briancurtin> you too