[09:38] <JamesTait> Good morning, everyone! :)
[11:07] <gatox> morning everyone
[11:07] <mandel> morning!
[11:08] <mandel> gatox, we have a small issue with the xdg in trunk and rye
[11:08] <gatox> mandel, fighting with P?
[11:08] <gatox> :P
[11:08] <mandel> can I haz help
[11:08] <mandel> gatox, a little, not too much :)
[11:08] <gatox> mandel, shoot
[11:09] <mandel> gatox, is better to ask rye
[11:12] <mandel> gatox, how is it with the windows api?
[11:12] <gatox> rye, hi
[11:12] <gatox> mandel, a bloody mess
[11:12] <gatox> jejejeje
[11:12] <gatox> and pywin32 is incomplete
[11:12] <mandel> gatox, I know that already :)
[11:12] <gatox> :P
[11:13] <mandel> gatox, go go ctypes!
[11:13] <mandel> gatox, the problem with ctypes is when you use them in an x64  and you used py2exe.. things break
[11:13] <mandel> :(
[11:13] <gatox> crap
[11:15] <mandel> gatox, yeah... I'm getting lots of knowledge out of this project :)
[11:17] <gatox> mandel, i know..... if i search for windows and  python on google..... you are the guy!! jejejeje
[11:18] <mandel> gatox, shhh don't tell anyone :P
[11:19] <mandel> gatox, so what was the issue you had?
[11:21] <gatox> mandel, na..... i'm fighting trying to make a link to a folder...... the way that it's now, don't allow unicode targets, and the one that i'm trying to use that it's kind of more low level don't have the Folder constants, and hardcoding the constants is not working... soo... now i'm going to take a look to pywin if that could be implemented there.... or go with ctypes (that it's kind of more difficult and annoying)
[11:22] <mandel> gatox, which are the constants?
[11:22] <mandel> gatox, and can  see the funtion you are using, for curiosity mainly :)
[11:22] <gatox> yep..... let me give you a paste
[11:25] <gatox> mandel, http://paste.ubuntu.com/769935/
[11:26] <gatox> mandel, according the api, changing "shell.CLSID_ShellLink"  for "shell.CLSID_FolderShortcut" should work... but, shell.CLSID_FolderShortcut is not defined..... and if i put the constant value, directly there, doesn't work
[11:27] <mandel> gatox, what error do you get?
[11:27] <gatox> mandel, the constant value i take it from here: http://www.capo-nord.org/soci/xmav/articolo.php?id=20070614
[11:27] <mandel> gatox, may I see the code qhen you add the constant?
[11:30] <gatox> mandel, i have tested this, and its creating the link, but it's a file link, not a foldeer link: http://paste.ubuntu.com/769940/
[11:30] <gatox> mandel, i'm reading this doc: http://msdn.microsoft.com/en-us/library/aa969393.aspx Using Shell Links section
[11:30] <gatox> To create a folder shortcut using the CreateLink sample function, call CoCreateInstance using CLSID_FolderShortcut, instead of CLSID_ShellLink (CLSID_FolderShortcut supports IShellLink)
[11:31] <gatox> mandel, the other thing is to use IShellLinkW.... but again, it's not implemented
[11:32] <mandel> gatox, why do you say it is creating a file link?
[11:34] <gatox> mandel, if a try to open that link is not resolved and try to find a file
[11:34] <mandel> gatox, hm.. let me 5 mins to play a little with it :)
[11:34] <mandel> I'm waiting for natty to install de build-deps of control panel :)
[11:35] <gatox> mandel, ok..... i'll be looking to pywin code..... if there is any hope to avoid ctypes :P
[11:37] <gatox> mandel, crap...... the create_shortcut method in pywin is doing exactly the same
[11:37] <gatox> EXACTLY!
[11:38] <gatox> why i didn't start looking at that!!??  jeje
[11:38] <gatox> sooo..... pywin is having the same problem as me
[11:40] <mandel> gatox, an the issue is unicode related
[11:41] <gatox> mandel, of course!!! jejejee you are talking to me
[11:41] <gatox> jejje
[11:41] <mandel> gatox, let me take a small look
[11:43] <gatox> AHHHHHHH WAITTTTTTTTTT
[11:43] <gatox> ¬¬ it's not having problems to create folder shortcuts just the way it is...... it's having problem to find the setPath argument
[11:44] <gatox> that happens to test everything in a really weird unicode environment
[11:44] <mandel> gatox, tell me you do have the folder..
[11:45] <gatox> yes yes...... but if i try with another folder with a path that doesn't contain unicode chars, it works fine...... but if the argument in setPath contains unicode chars, there is the problem
[11:48] <mandel> gatox, I've got it working
[11:49] <mandel> gatox, let me paste you what i did :P
[11:49] <gatox> mandel, with unicode?
[11:49] <gatox> mandel, don't play with my feelings
[11:49] <gatox> jejejee
[11:49] <mandel> gatox, paste.ubuntu.com/769957
[11:50] <mandel> gatox, you can see that there I'm using unicode in the program location, right?
[11:50] <mandel> (I needed to use that name because I have a 'decent' username)
[11:51] <mandel> gatox, the problem you have is that the shell in windows does not use unicode but mbcs, so probably when reading the target path it was going, wtf!? I don't understand utf8!!!
[11:51] <mandel> gatox, while python is smart and does the right way when doing an open(unicode) the pywin32 does not and you have to be explicit
[11:52] <gatox> mandel, not working for me
[11:52] <gatox> i'm the corner case of the programming world!
[11:52] <gatox> jjje
[11:52] <mandel> gatox, ok, can you change the script adding the username in the path to let me check in mine?
[11:53] <gatox> mandel, test with this: 雄鳥お人好し ñandú
[11:53] <mandel> gatox, also, open the shortcut, and look at the target property, where does it point?
[11:53] <ralsina> good morning!
[11:53] <gatox> ralsina, morning
[11:54] <mandel> ralsina, buenos dias!
[11:54] <ralsina> mandel, gatox: do either of you have a clean windows 7 vm?
[11:54] <gatox> ralsina, nop.... i'm going to install one.... it seems that it's needed :P
[11:54] <ralsina> because the broken release works just fine on XP, seems to me :-(
[11:54] <ralsina> gatox: please
[11:54] <gatox> mandel, nowhere
[11:54] <gatox> just blank
[11:55] <mandel> gatox, so it getting confused by the encoding and ignoring it, lovely windows :)
[11:55] <mandel> ralsina, I do have a clean windows for testing :)
[11:55] <mandel> ralsina, let me finish helping gatox and I'll move to you
[11:55] <mandel> :)
[11:55] <gatox> mandel, ñ and ú.... not much problem....... but: 雄鳥お人好し   frak windows
[11:55] <ralsina> mandel: ack
[11:56] <gatox> mandel, i think that if we could use IShellLinkW maybe would be easier
[11:56] <mandel> gatox, I don't think that method exists..
[11:57] <gatox> not method, interface
[11:57] <gatox> http://msdn.microsoft.com/en-us/library/bb774950(VS.85).aspx
[11:57] <gatox> mandel, and i have the constant value: "{000214F9-0000-0000-C000-000000000046}"
[11:58] <gatox> but it saays that there is not an interface registered that supports that id
[11:58] <gatox> iid
[11:59] <gatox> mmmmmm but it's accepting the id for IShellLinksA
[11:59] <gatox> "{000214EE-0000-0000-C000-000000000046}"
[12:00] <nessita> good morning!
[12:01] <gatox> ahhhh this constants are for windowsXP, they probably change... and that's way the old ones are still valid maybe
[12:01] <mandel> nessita, morning!
[12:01] <gatox> nessita, morning
[12:01] <gatox> mandel, i'll try to obtain the new IID
[12:01] <mandel> gatox, the problem is that you have to be super careful with that, the reason to prefer pywin32 is exactly that..
[12:01] <gatox> mandel, yes, i know......
[12:01] <gatox> but....... it's not working :S
[12:02] <gatox> mandel, i just want to know if it works, and then i'll look for a nice way to obtain that value
[12:04] <mandel> gatox, it is certainly possible to use it, especially because windows explorer can :)
[12:04] <gatox> yep
[12:07] <ralsina> morning nessita
[12:08] <gatox> nessita, hi
[12:08] <ralsina> gatox: +1 on what mandel said, those constants are evil
[12:08] <gatox> ralsina, yep.... i know that part
[12:08] <nessita> gatox: two hellos? wow, I feel special :-)
[12:08] <gatox> but...... pywin is not working just the way it is.... and using ctypes will be the same....... and with some more confusing code
[12:08] <ralsina> gatox: you can instantiate IShellLink using pythoncom
[12:08] <ralsina> gatox: http://timgolden.me.uk/python/win32_how_do_i/create-a-shortcut.html
[12:09] <nessita> hola mandel, ralsina
[12:09] <gatox> nessita, jejeje in this moment i just don't know what i say :P
[12:09] <mandel> nessita, buenos dias :)
[12:10] <mandel> gatox, http://www.codeproject.com/KB/shell/create_shortcut.aspx
[12:10] <gatox> ralsina, yes, that what i'm doing..... but i need IID_IShellLinkW which is not there
[12:10]  * gatox looking.....
[12:11] <mandel> gatox, this is the trick: MultiByteToWideChar
[12:12] <ralsina> gatox: it's there in windows 7
[12:12] <gatox> mandel, wowwwwww :O
[12:12] <gatox> ralsina, not following....
[12:12] <ralsina> gatox: well, if I print shell.IID_IShellLink it's there :-)
[12:12] <gatox> ralsina, yes it is..... but not in pywin i mean
[12:13] <gatox> ralsina, yesss...... that one is there, i mean IShellLinkW
[12:13] <ralsina> gatox: unless you mean you can't instantiate it (sorry I missed the first part of the cnversation)
[12:13] <ralsina> gatox: it's there too
[12:13] <ralsina> gatox: and it's the same value
[12:13] <ralsina> gatox: Wait, not the same value
[12:14] <gatox> yep..... but it's not recognized as an interface with that iid
[12:14] <ralsina> IID('{00214F9-0000-C000-000000000046}')
[12:14] <ralsina> ACK
[12:14] <gatox> i'll try mandel suggestion about MultiByteToWideChar
[12:15] <ralsina> ok
[12:15] <mandel> gatox, try that code using ctypes, if it work, find out how to that with unicode on python and see if you have luck with the pywin32 calls
[12:16] <mandel> if, not, do use ctypes and let me know so that we test the code in an x64 windows machine
[12:28]  * mandel would like guest additions for kvm virtual machines...
[12:30] <mandel> gatox, to run the test for control panel with qt is ./run-tests -qt right?
[12:31] <mandel> nessita, did you have a ppa with the qtreactor that I can use for N?
[12:31] <gatox> mandel, yep
[12:31] <nessita> mandel: yes, is there in the email I sent to u1discuss. The .deb is also there (subject qt4reactor).
[12:32]  * mandel looks
[12:40] <mandel> nessita, got it!
[12:41] <nessita> mandel: yey!
[12:42] <mandel> nessita, I'm suurprised I did a build-dep ubuntuone-control-panel and it did no complain about qt4reactor missing, is that a known issue?
[12:43] <nessita> mandel: we're not packaging anything related qt yet
[12:43] <mandel> nessita, ok :)
[12:43] <nessita> so, strictily speaking is not an issue :-P
[12:43] <mandel> nessita, and.. one more thing, do you have bugs for me, otherwise I'll work on proxy :)
[12:43] <nessita> the issue would be: "no qt packages"
[12:43] <mandel> a more general one :)
[12:44] <nessita> mandel: work on proxy please! I will re-review the system folder branch soon
[12:44] <mandel> nessita, sweet, I'll move to that then, the other branch has also been updated :)
[13:10] <nessita> mandel: when you have a moment, you could also fix this proposal: https://code.launchpad.net/~mandel/ubuntuone-dev-tools/proxy-testcase/+merge/85019
[13:10] <nessita> mandel: so we can have rodney re-reviewing today
[13:10] <mandel> nessita, that is the proxy work I'm doing :)
[13:11] <nessita> mandel: great!!!
[13:11] <mandel> nessita, getting twisted to use the uath was giving me some headaches :(
[13:13] <nessita> mandel: any reason to do this:
[13:13] <nessita>         expected_result = []
[13:13] <nessita>         expected_result.extend(self.dirs)
[13:13] <nessita>         expected_result.extend(self.files)
[13:13] <nessita> instead of:
[13:13] <nessita>         expected_result = self.dirs + self.files
[13:13] <nessita> ?
[13:14] <mandel> nessita, no real reason, bad code from my side
[13:15] <nessita> mandel: another question, any reason to prepend "native" to this method? native_is_system_path
[13:15] <mandel> nessita, so that we know that the are no decorators in it
[13:15] <mandel> I followed the naming convention we used in the other methods with no decorators
[13:15] <nessita> ah, I see
[13:18] <mandel> nessita, pushing a new revno with that stupid list operation fixed
[13:18] <nessita> thanks
[13:19] <mandel> revno 1169
[13:27] <nessita> re branching
[13:34] <mandel> alecu, is there a way to know why I'm getting a 400 error when using an HTTPClientFactory?
[13:35] <alecu> mandel, surely there is
[13:35] <mandel> alecu, can you point me to it?
[13:35] <mandel> alecu, docs should be enough :)
[13:35] <alecu> mandel, is the server a test twisted.web.server running in the same process?
[13:36] <mandel> alecu, no.. is bloody squid the one that is giving me that :(
[13:40] <mandel> alecu, te molesto post lunch
[13:40]  * mandel lunch
[13:42] <Yanch0>  guys anyone managing to fix the 'Calling Stale Broker' problem on a windows 7 client when trying to synch to ubuntu one ?
[13:43] <mandel> nessita, ^
[13:44] <gatox> brb!
[13:44] <nessita> Yanch0: hi there, not ATM, but as far as I know it means syncdaemon has stopped working
[13:44] <nessita> Yanch0: is this happening all the time?
[13:45] <Yanch0> yes nessita rarely works ifne
[13:45] <nessita> Yanch0: what's you windows locale and version?
[13:46] <Yanch0> ver 6.1 build : 7601 - windows 7 x64 professional - en locale
[13:51] <alecu> mandel, I found out how to do that, by digging into the twisted sourcecode. Here's how: http://pastebin.ubuntu.com/770054/
[13:59] <nessita> alecu: any idea how to debug further the problem that Yanch0 is having?
[14:00] <nessita> Yanch0: the "Calling Stale Broker" appears on the controlpanel?
[14:00] <nessita> I mean, in the graphical UI
[14:00] <Yanch0> yes nessita
[14:00] <nessita> Yanch0: can you tell if that happens at a certain point, of after doing something specific? can you reproduce?
[14:01] <alecu> looking
[14:02] <Yanch0> File Sync starting ... then after minutes of nothing happening that error
[14:03] <alecu> Yanch0, we should take a look at your logs, to see what's causing syncdaemon to die
[14:04] <nessita> Yanch0: can you please monitor the memory consumption?
[14:04] <Yanch0> cool alecu - how can i get u those?
[14:04] <nessita> alecu: perhaps killed by the OS by mem usage? :-/
[14:04] <alecu> Yanch0, let me find the right folder...
[14:04] <nessita> alecu, Yanch0: C:\Users\<username>\AppData\Local\xdg\cache\
[14:05] <alecu> nessita, sounds reasonable, and that might be it if there's no other error
[14:05] <nessita> alecu: right
[14:05] <Yanch0> ubuntuone-syncdaemon.exe *32 is currently @ 26 cpu and 400 mb
[14:05] <nessita> Yanch0: can you monitor if the mem usage grows, and see the value when the process is killed/terminated?
[14:06] <Yanch0> ok nessita will keep an eye on that .. in the mean time which logs u need?
[14:06] <Yanch0> going to 530mb
[14:06] <alecu> Yanch0, and whenever you get that message it means that the control panel is not connecting to syncdaemon anymore. You should try killing the control panel process and running it again.
[14:07] <Yanch0> alecu, thats what im doing now after i asked the qs again - to try to reproduce
[14:07] <alecu> Yanch0, useful logs would be the syncdaemon logs just before the problem started (you can check that by looking at the date&time inside the files)
[14:07] <nessita> Yanch0: everything under C:\Users\<username>\AppData\Local\xdg\cache\
[14:07] <nessita> (I need to run a quick errand, brb in 15 minutes)
[14:07] <alecu> or just everything, right :-)
[14:08] <Yanch0> we're at 1gb memory
[14:08] <Yanch0> (documents folder is 34mb all in all)
[14:11] <gatox> back
[14:12] <Yanch0> 2gb memory - out of 8 total
[14:13] <Yanch0> stuck at 1949620K memory for a bit .. cpu sticking to 25 /26
[14:24]  * nessita is back
[14:25] <nessita> Yanch0: let us know if the syncdaemon process dissapears (gets killed)
[14:26] <Yanch0> nessita, still alive, but stuck to 1960276
[14:26] <nessita> mandel: branch approved, IRL tested the listdir function and works as expected
[14:26] <nessita> Yanch0: is controlpanel failing, if you open it?
[14:27] <Yanch0> no nessita .. top right still says starting
[14:27] <Yanch0> i now just clicked on settings .. and pam
[14:27] <nessita> Yanch0: pam == stale broker error?
[14:27] <Yanch0> eror appeared  - ram 1960 364
[14:27] <Yanch0> yep
[14:27] <Yanch0> ---------------------------
[14:27] <Yanch0> Ubuntu One experienced an error
[14:27] <Yanch0> ---------------------------
[14:27] <Yanch0> Sorry, an error has occurred and Ubuntu One needs to close.
[14:27] <Yanch0> ---------------------------
[14:27] <Yanch0> Close   Hide Details...
[14:27] <Yanch0> ---------------------------
[14:27] <nessita> Yanch0: is the syncdaemon process still running?
[14:28] <Yanch0> sorry for multi line paste!!
[14:28] <Yanch0> yes still running
[14:28] <nessita> but seems is not responding...
[14:30] <nessita> alecu: any ideas?
[14:30] <Yanch0> in taskmgr still running ( 1960 500) ram
[14:30] <Yanch0> and top right still says it's starting
[14:31] <nessita> Yanch0: if you close and re-open the UI, same error?
[14:31] <alecu> nessita, no idea either :-(
[14:31] <Yanch0> ok i closed the ui .. syncdaemon still alive
[14:32] <Yanch0> different error now
[14:32] <Yanch0> attribute error , "nonetype object has no attributed 'get_root_dir'
[14:32] <nessita> Yanch0: and is the syncdaemon process still "alive"?
[14:32] <Yanch0> same if i click Settings
[14:33] <Yanch0> yes still cpu 26 and same ram as before 32mb more
[14:35] <nessita> Yanch0: let's stry this. Please kill syncdaemon, and close everything related to U1
[14:36] <nessita> Yanch0: then, move the current logs to another folder, so they don't get rotated. Once that's done please re-open the controlpanel and in the Folders tab (the first one shown), unclick the Documents folder
[14:36] <nessita> Yanch0: I'm guessing you're suffering from https://bugs.launchpad.net/ubuntuone-client/+bug/883926
[14:36] <ubot4> Launchpad bug 883926 in ubuntuone-client "[windows] Breaks on system directories My Music, My Videos dirs in ~\Documents folder on Win7 (affects: 6) (dups: 3) (heat: 55)" [Critical,In progress]
[14:37] <nessita> alecu: in that bug, syncdaemon never finished Local Rescan, it keeps cycling and cycling over the system folder, so that's why Yanch0 may be seeing "File sync starting" all the time
[14:38] <Yanch0> unchecked documents
[14:38] <nessita> Yanch0: ok, let me know how that goes, if file sync progress, then we have more info :-)
[14:38] <Yanch0> click settings get that error again
[14:39] <nessita> Yanch0: can you kill everything and re-open? and confirm the Documents folder is not selected to be synch
[14:39] <Yanch0> ok will do
[14:39] <nessita> Yanch0: thanks for helping debug!
[14:40] <alecu> nessita, oh, I see. I remember the discussion regarding this bug at the sprint
[14:40] <Yanch0> file sync in progress
[14:40] <Yanch0> and syncing ubuntu one (empty folder)
[14:40] <nessita> alecu: we're working on 2 fixes: on windows, ignore system folders, on linux, avoid LR cycling for ever (have a limit)
[14:41] <nessita> Yanch0: then, despite this not being good news for you, you're suffering from the bug linked above. We're working on fixing this, as we speak
[14:41] <Yanch0> great :)
[14:41] <nessita> Yanch0: wanna click on "affects me too", in the bug? (at the top)
[14:42] <Yanch0> funny thing is that i cant sync other folders except the documents
[14:43] <nessita> Yanch0: yes, that's because the DOcuments folder have system folders in it
[14:43] <nessita> Yanch0: which are making a given stage of syncdaemon startup fail, and retry indefinitely
[14:43] <nessita> Yanch0: that's why it never progresses after the "File sync starting" stage
[14:43] <Yanch0> gotcha .. same issue as this guy with cpu usage
[14:44] <Yanch0> where is the affects me ?
[14:44] <nessita> Yanch0: in the bug report, at the top left corner
[14:44] <nessita> This bug affects 6 persons. Does this bug affect you?
[14:45] <nessita> Yanch0: so, a immediate workaround would be to move the things you want to be synched to another folder and choose that to be synched
[14:45] <Yanch0> i wanted also to sycn d:\projects
[14:46] <nessita> Yanch0: and that worked?
[14:46] <Yanch0> but i can only choose a folder inside c:\users\matthew pulis directory and not overlapping with existing folder
[14:46] <nessita> Yanch0: ah, yes, sorry, I misread the path you pasted before
[14:46] <nessita> Yanch0: only things in inside your profile path
[14:47] <Yanch0> pity
[14:47] <nessita> Yanch0: master bug for that last thing is https://bugs.launchpad.net/ubuntuone-client/+bug/854143
[14:47] <ubot4> Launchpad bug 854143 in ubuntuone-client "Can't sync symbolic links, or folders outside my profile (home folder). (affects: 6) (dups: 4) (heat: 38)" [Wishlist,Confirmed]
[14:47] <nessita> Yanch0: you can also click affects me to to that bug :-/
[14:47] <Yanch0> waiting the confirmation code
[14:47] <Yanch0> yes i tried to act funy and create a symbolic link to d:\projects
[14:47] <Yanch0> but received complaints too
[14:49] <Yanch0> anything u want me to submit ?
[14:50] <Yanch0> as in logs
[14:50] <nessita> Yanch0: no need, thanks. The first issue is already diagnosed, and the second one involves a redesign of current syncdaemon and sadly it will not happen before the next ubuntu release
[14:51] <nessita> Yanch0: but thanks, a lot, for helping debugging this
[14:52] <Yanch0> wish u luck on debugging this .. seems a pain
[14:53]  * mandel back
[14:53] <mandel> nessita, super, thx alot!
[14:55] <Yanch0> nessita, im gonna idle here - feel free to ping if u want more debugfs
[14:55] <nessita> Yanch0: great, thanks :-) let me know if you run into another issue
[14:55] <nessita> alecu: I'm loving the webclient sso branch :-)
[14:56] <mandel> gatox, alecu, care to take a look at: https://code.launchpad.net/~mandel/ubuntuone-client/ignore-system-folders/+merge/84848
[14:56] <nessita> alecu: looks very clean and understandable, in particular the tests
[14:56] <gatox> mandel, no......
[14:56] <gatox> :P
[14:56] <gatox> jeje
[14:56] <mandel> gatox, te odio!
[14:56] <gatox> jeje
[14:56] <alecu> nessita, thanks for the compliment :-)
[14:56]  * gatox looking....
[15:00] <gatox> me
[15:00] <nessita> me
[15:00] <nessita> che, y ralsina?
[15:00] <gatox> niiiiii idea
[15:01] <mandel> nessita, he was here a while back
[15:01] <mandel> me
[15:01] <mandel> maybe a network issues?
[15:02] <nessita> alecu: stdp?
[15:02] <alecu> me
[15:02] <nessita> dobey: welcome back! standup?
[15:02] <dobey> no :)
[15:02] <gatox> cylon
[15:02] <dobey> lol
[15:02] <alecu> hehehehe
[15:03] <nessita> dobey: let's play a game, where you can't say no :-D
[15:03] <nessita> gatox: go
[15:03] <gatox> me go?
[15:03] <gatox> DONE:
[15:03] <gatox> Worked on make link.
[15:03] <gatox> TODO:
[15:03] <gatox> Keep working on make link and hope to finish it
[15:03] <gatox> BLOCKED:
[15:03] <gatox> No
[15:03] <gatox> nessita, go
[15:03] <nessita> DONE: reviews, started with bug #834730 (is getting interesting)
[15:03] <nessita> TODO: more reviews, more on bug #834730, 1-1 with ralsina when he comes back
[15:03] <nessita> BLOCKED: nopes
[15:03] <nessita> NEXT: mandel
[15:03] <ubot4> Launchpad bug 834730 in ubuntuone-client (and 1 other project) "Use SignalBroadcaster and RemoteMeta from ubuntu sso client (affects: 1) (heat: 6)" [High,In progress] https://launchpad.net/bugs/834730
[15:03] <mandel> DONE: Fixed MPS for ignored-system-paths and autoupdate-looping-call. Looked at my ProxyWebClient code to firxt the proxy testcase branch to use twisted webclient and not urrlib2
[15:03] <mandel> TODO: more on the ProxyWebClient work (nearly got it).
[15:03] <mandel> BLOCKED: no but got to ask some stupid questions to alecu about twisted :)
[15:03] <mandel> alecu, please!
[15:03] <alecu> DONE: many fixes for merge proposals, more webclient for proxy
[15:03] <alecu> TODO: finish some internal branches so they can be reviewed tomorrow
[15:03] <alecu> BLOCKED: no
[15:03] <alecu> NOTE: today we are celebrating amelia's birthday with her friends at the pelotero, so I'll be out a bit earlier than usual
[15:03] <alecu> NEXT: dobey
[15:03] <dobey> λ DONE: sprint, swap days
[15:03] <dobey> λ TODO: get twisted patch upstream, help nessita with qtreactor packages
[15:03] <dobey> λ BLCK: No.
[15:03] <gatox> mandel, +1
[15:03] <gatox> branch
[15:03] <mandel> gatox, thx!
[15:04] <alecu> mandel, "<alecu> mandel, I found out how to do that, by digging into the twisted sourcecode. Here's how: http://pastebin.ubuntu.com/770054/"
[15:04] <nessita> any comments anyone?
[15:05] <nessita> dobey: we have a problematic bug on precise, regarding a glib flag
[15:05] <nessita> dobey: I've assigned the bug to you, I would guess is not difficult to solve
[15:05] <alecu> guys, tomorrow's my EOY, so make sure you point me at all reviews I should be doing
[15:05] <dobey> flag? no. typo
[15:05] <alecu> mandel, gatox ^
[15:05] <gatox> alecu, ack
[15:05] <dobey> it's because precise has the new glib which just aborts
[15:05] <nessita> mandel: ^ make sure alecu re-reviews the autoupdate branch
[15:05] <nessita> dobey: ah, ok. Can you fix? :-)
[15:05] <mandel> alecu, so, I have another stupid question, when I do deferred.errback(error.Error) it get converted in a Failure with an error.error in it, is that the correct way to do it, or is there a way to re throw the error.Error
[15:06] <mandel> nessita, ack
[15:06] <nessita> dobey: when you have a moment, please
[15:06] <gatox> alecu, nothing from me
[15:06] <mandel> alecu, ack :)
[15:06] <dobey> yes
[15:06] <alecu> mandel, can you point me at a bit of code?
[15:06] <nessita> dobey: thanks!
[15:06] <mandel> alecu, sure let me push it so that we are talking about the same :)
[15:06] <alecu> mandel, yes, it's the right thing for exceptions to be wrapped in a Failure when they are sent thru errbacks
[15:07] <alecu> mandel, but I'm not sure what you mean by "re-throw"
[15:07] <mandel> alecu, let me push it :)
[15:07] <nessita> ok, eom!
[15:07] <alecu> ralsina, go!
[15:08] <ralsina> DONE: mgmt call, canonicaladmin, worked on debugging windows brokenness (rye found it!)
[15:08] <ralsina> TODO: build a working windows release
[15:08] <ralsina> BLOCKED: no
[15:08] <gatox> ok, lunch for me!! go back in a while...
[15:08] <nessita> ralsina: todo bien? we had the 1-1 at 11:30
[15:08] <ralsina> sorry about not being here, but windows rebooted :-)
[15:08] <dobey> MS has had the same TODO for 20 years ;)
[15:08] <ralsina> dobey: ha!
[15:08] <ralsina> nessita: the reboot was not exactly smooth ;-)
[15:09] <mandel> alecu, can you pull from lp:~mandel/ubuntuone-dev-tools/proxy-testcase and look at line 92 of ubuntuone/devtools/testcases/tests/test_squid.py
[15:09] <nessita> ralsina: the reboot took 35 minutes? :-/
[15:09] <nessita> ralsina: what happened?
[15:09] <mandel> I guess that probably updates..
[15:09] <ralsina> nessita: yes. Windows needed to reconfigure stuf because of an update. Took 38 minutes
[15:09] <nessita> ralsina: oh my god...
[15:09] <ralsina> nessita: indeed
[15:10] <mandel> nessita, yeah, I usually avoid updated til EOD, otherwise you waist soooooo much time :(
[15:10] <nessita> right
[15:11] <alecu> mandel, you probably mean this file, right? ubuntuone/devtools/testcases/tests/test_squid_testcase.py
[15:11] <mandel> alecu, sorry, yes..
[15:11] <ralsina> mandel: me too, but ubuntuone's uninstaller asked me to reboot and I clicked without thinking :-(
[15:11] <alecu> mandel, ok, looking at line 92... you are trying to re-throw the failure, right?
[15:11] <mandel> alecu, so there i get an error.Error then I check if it is 407, if not I want to pass it to the deferred_result
[15:12] <mandel> alecu, exactly, is that the correct way to do it?
[15:12] <dobey> whoot. new RAM is "out for delivery"
[15:13] <dobey> i wonder how fast pbuilder would be, if i ran it all in ram
[15:13] <mandel> dobey, bohh I though you bought a new pet :(
[15:13]  * mandel imagines the cylon riding a ram
[15:14] <mandel> dobey, this one to be precise: http://the-tech-guy.net/wp-content/uploads/2011/11/ram-merino.jpg
[15:15] <alecu> mandel, well, _process_auth_error is already an errback handler (from factory.deferred in _get_auth_page), right?
[15:15] <nessita> alecu: tiny needs fixing on https://code.launchpad.net/~alecu/ubuntu-sso-client/proxy-web-client/+merge/85135, with that I'll approve
[15:15] <mandel> alecu, yes it is
[15:15] <ralsina> mandel, nessita: bug #904300
[15:15] <ubot4> Launchpad bug 904300 in ubuntuone-client "Don't start syncdaemon on windows vista or later if running with escalated privileges (affects: 1) (heat: 6)" [High,Triaged] https://launchpad.net/bugs/904300
[15:15] <alecu> mandel, so, what happens when factory.deferred throws an error that is not a error.Error?
[15:15] <alecu> mandel, say, a connection error
[15:16] <mandel> alecu, exactly, I want to be able to pass it to the deferred_result, and there is where my test fails
[15:16] <alecu> mandel, I'm not sure I like passing the deferred_result between _get_auth_page and _process_auth_error
[15:16] <mandel> ralsina, nessita that looks like a feature that we are missing on windows yet have on linux, do not run as root (unless you really really want to)
[15:17] <ralsina> mandel: yes, the check should be made crossplatform
[15:17] <mandel> alecu, the issue there is that I want to be able to do the retry so that in the second round I send the base auth
[15:17] <mandel> alecu, so, returning the original deferred does no good, right?
[15:18] <nessita> ralsina: but how can we make that check? there are some "regular" users that are also admin
[15:18] <nessita> ralsina: and we want to let those users running syncdaemon, no?
[15:18] <ralsina> nessita: you have to check for privilege elevation
[15:18] <ralsina> nessita: admin users don't start processes as admin, normally
[15:19] <nessita> ah, I see
[15:19] <ralsina> being admin on windows is more like being in sudoers than being root
[15:19] <nessita> ralsina: well, FYI, in trunk/ubuntuone/platform/windows/os_helper.py, there is:
[15:19] <nessita> def is_root():
[15:19] <nessita>     """Return if the user is running as root."""
[15:19] <nessita>     # TODO: Do check if we are running as admin
[15:19] <nessita>     return False
[15:19] <ralsina> well, that's the mising bit :-)
[15:19] <nessita> ralsina: so if we just implement that, the rest is solved
[15:20] <ralsina> exxxxxactly
[15:20] <nessita> ralsina: I'll add that to the report
[15:20] <ralsina> cool, thx
[15:20] <mandel> herb, at least we did forsee the issue :)
[15:20] <ralsina> Or we can switch that False to True and problem solved forever ;-)
[15:21] <mandel> ralsina, no no no, that would not work, it wont let any user run sd
[15:21] <nessita> ralsina: another FYI, the "official" windows tag is "u1-windows"
[15:21] <ralsina> nessita: I know, sorry, my mistake
[15:21] <ralsina> mandel: I know. Solves *all* windows bugs ;-)
[15:21] <mandel> ralsina, well we get back to the first bug, it does not work on windows :P
[15:22] <ralsina> mandel: so, only one bug? progress!
[15:22] <mandel> managers...
[15:22] <ralsina> metrics!
[15:23] <mandel> alecu, so, is there a common pattern to follow in this cases?
[15:24] <dobey> bah, trying to find a picture of something in the original battlestar series, is really hard to do now on google
[15:24] <dobey> "oh hi, here are more pictures of tricia helfler"
[15:24] <alecu> where!
[15:26] <alecu> mandel, I'm trying to figure out the pattern for this
[15:26] <alecu> mandel, anyway, I don't like the deferred_result being passed back and forth, and I don't like the .called being checked each time.
[15:26] <mandel> alecu, sweet thx! I've been thinking about it and I always get a fail in the assert because I get a failure and not an error
[15:27] <mandel> alecu, true, I don't think is needed to be checked, is just for safety
[15:27] <alecu> mandel, I think that this could be much simpler if we used the deferred chaining in a different way.
[15:28] <mandel> alecu, and return the original deferred? the problem is the second request by the second HTTPClientFactory
[15:29] <alecu> nessita, just pushed the fix to https://code.launchpad.net/~alecu/ubuntu-sso-client/proxy-web-client/+merge/85135
[15:29] <nessita> alecu: thanks!
[15:30] <alecu> mandel, yup, I think returning the original deferred would be the right thing, and chaining the callbacks so the result is passed through
[15:33] <mandel> alecu, I'll also be looking down that path..
[15:35] <alecu> mandel, I think 1) _get_auth_page should be returning the factory.deferred, 2) _process_success is not needed
[15:35] <alecu> 3) the retrying should happen in an errback on that deferred
[15:35] <mandel> alecu, that was just used to return the other deferred, if you remove it, it goes awaya
[15:37] <alecu> mandel, right. And the main problem with the current code is that if there's an error other than error.Error, then deferred_result.errback never gets called.
[15:37] <mandel> exactly
[15:37] <alecu> mandel, and it's very verbose too :-)
[15:37] <mandel> well,m I dont mind it being verbose, I'm more worried about the error :)
[15:46] <ralsina> IN YOUR FACE, UAC!
[15:46] <ralsina> mandel, rye, elopio: I think I solved everything :-)
[15:46] <mandel> ralsina, no, I'm still single and not rich :(
[15:46] <nessita> alecu: bug A for you on webclient branch!
[15:46] <ralsina> mandel: file a bug!
[15:46] <ralsina> mandel:  ;-)
[15:47] <alecu> \o/
[15:47] <ralsina> mandel: I found a way to start an unelevated process from an elevated process
[15:47] <dobey> don't file a bug; you might assign it to someone
[16:12] <dobey> nessita: https://code.launchpad.net/~dobey/ubuntuone-client-gnome/convert-typo/+merge/85688
[16:16] <ralsina> mandel, gatox, nessita: I have a hack, but the "real" solution for starting unescalated processes is probably something like this: https://pastebin.canonical.com/57119/
[16:16] <ralsina> mandel, gatox, nessita: and that's a real pain
[16:17] <alecu> sounds hurtful
[16:19] <mandel> ralsina, and you plan to do that with pywin32 , ctypes or c?
[16:19] <ralsina> mandel: good question
[16:19] <gatox> or Go?
[16:19] <gatox> :P
[16:19] <alecu> ralsina, mandel: probably ctypes + patience!
[16:19] <ralsina> The current hack I prposed for testing uses an external DLL to do the trick
[16:19] <ralsina> here's the hack: http://www.codeproject.com/KB/vista-security/RunNonElevated.aspx
[16:20] <mandel> ralsina, what about the legal issues?
[16:20] <ralsina> mandel: it's under MIT license
[16:20] <mandel> and, if we are doing that, we might as well compile a module in python rather than adding the dll
[16:20] <mandel> :P
[16:21] <ralsina> mandel: I have to start a non-elevated process from InstallBuilder
[16:21] <ralsina> if we do it in python, it has to be added to -installer so it "de-elevates" itself and respawn
[16:21] <mandel> ouch
[16:23] <ralsina> mandel: thus my "that's a real pain" comment
[16:38] <mandel> ralsina, is there a way to tell squid I'm debugging and that I want to see the auth header he got?
[16:39] <ralsina> mandel: not that I know of
[16:39] <ralsina> mandel: if you are using firefox, you can check using firebug
[16:39] <nessita> dobey: approved
[16:40] <dobey> muchas gracias
[16:41] <mandel> ralsina, I'm using twisted
[16:41] <ralsina> mandel: no idea then
[16:42] <ralsina> mandel: but why are you using twisted as client? I thought the choices were libsoup and qtnetwork?
[16:42] <mandel> ralsina, I was preciously using urllib2 for the tests for the squid testcase but I was required to use twisted instead
[16:43] <ralsina> mandel: ack
[16:56] <mandel> ok, I need to go a say happy birthday to the little brother, bbl
[16:56] <dobey> make sure you pay disney for the rights to sing the song
[17:18] <dobey> lunch time; bbiab
[18:17] <dobey> brb, new RAM :)
[18:27] <dobey> hrmm
[18:27] <dobey> that didn't work out so well
[18:28] <dobey> i just put another 6GB in, and linux thinks i only have 3.2
[18:29] <dobey> :-/
[19:41] <alecu> an early EOD for me today
[19:41] <alecu> bye all!
[19:41] <rye> gatox, btw, the issue with home dir was me, i failed to update the bzr tree
[19:41] <rye> dobey, 32bit system?
[19:41] <gatox> alecu, bye
[19:41] <gatox> rye, fiuuuuuuuuu :P
[19:41] <rye> f
[19:41] <dobey> rye: yes
[19:41] <rye> gatox, i know, i know... but then i found why sd was hanging so it makes me good again.
[19:41] <gatox> dobey, you can update the kernel to recognize the whole memory
[19:42] <gatox> rye, good! :D
[19:42] <rye> gatox, pae?
[19:42] <dobey> gatox: not if it doesn't boot :)
[19:42] <gatox> dobey, ahhhhh that's a problem :P
[19:42]  * rye never used pae kernels, either simple i386 or x86_64
[19:42] <gatox> rye, pae?
[19:42] <gatox> ahhhhh
[19:42] <rye> physical address extension
[19:42] <gatox> rye, once i updated to a server kernel
[19:42] <dobey> and i can't install the amd64 kernel on oneiric
[19:42] <gatox> but i start using 64bits
[19:43] <rye> gatox, once loooong ago i spent a week recompiling the whole slackware for 64 bit cpu. It was fun.
[19:43] <rye> will never do again.
[19:43] <gatox> jejeje
[19:44] <rye> ZOMG battery weights almost as the rest of the netbook
[19:44] <gatox> dobey, cof cof clean install cof cof
[19:44] <dobey> yeah no
[19:44] <gatox> rye, well..... i'm still trying to get my new laptop to work properly
[19:44] <rye> it turns out it is not that much ZOMG
[19:44] <gatox> damn nvidia
[19:44] <dobey> i don't want a 64 bit userspace
[19:44] <rye> gatox, nouveau?
[19:44] <rye> dobey, why? RAM usage?
[19:45] <dobey> ram, proprietary software, no need for it
[19:45] <gatox> rye, i've try nouveau..... but no..... i think that  maybe the intel driver is missing.... this laptop uses both.....
[19:45] <gatox> it's weird
[19:45] <rye> i guess i don't have any proprietary software except for nvidia driver... and some bits for tax office running under wine
[19:46] <gatox> it's working..... but i can not use the vga or hdmi properly..... i'm fighting with xorg.conf
[19:46] <gatox> among other things
[19:46] <dobey> rye: flash
[19:46] <dobey> rye: the plug-ins are the real pain
[19:46] <rye> dobey, works using magic on my machine (nspluginwrapper, as I read your description above)
[19:47] <rye> well, at the moment it does not work using magic because i uninstalled it
[19:47] <dobey> nspluginwrapper isn't magic. it's insanity
[19:47] <dobey> i will not install that thing
[19:48] <rye> thankfully, neither of my banks uses flash for anything but banners. One uses Java for certificate check and signing though
[19:48] <rye> and youtube... if it is not popular and not available in webm, then I am not waching it or use minitube
[19:49] <rye> ralsina, when you tweeted about lighting up the battery, did you meen it looks like it is broken?
[19:49] <ralsina> rye: 6 hours to charge means it's broken
[19:49] <rye> ralsina, it is nearly completely depleted
[19:49] <ralsina> rye: unless it's like 8800maH or something
[19:49] <rye> ralsina, 7000mAh
[19:49] <ralsina> rye: then maybe it's ok :-)
[19:50] <rye> therefore it is called ZOMG-cell battery.
[19:51] <rye> gatox, yeeeah, i still want to grab my friend's laptop to see how it does this internal/discrete gpu magic
[19:51] <cjohnston> lisettte: ping
[19:53] <gatox> rye, i've to tested...... with a really powerfull game :P
[19:53] <gatox> test it
[19:53] <dobey> heck, solaris got this right 8 years ago; why can't linux? :-/
[19:54] <rye> dobey, define:this
[19:54] <dobey> rye: 64 bit kernel with 32 bit userspace
[19:54] <rye> dobey, well, i had this when i started converting the slackware... it is definitely possible
[19:55] <dobey> yes i know it's possible
[19:56] <dobey> actually, i think we did do it on suse way back when, but all of my own machines were 32 bit then, so didn't really matter
[19:56] <dobey> and firefox didn't constantly eat all my ram, either
[20:03] <gatox> EOD for  me!! see you guys tomorrow..... or later if i can't resist and get back to fight with the f$%#ing make link issue :P
[20:04] <dobey> make link issue?
[20:04] <dobey> lint i presume
[20:04] <gatox> link
[20:05] <gatox> create folder links with really weird unicode paths is not working
[20:05] <dobey> oh
[20:05] <gatox> yep.....
[20:05] <dobey> not make as in /usr/bin/make
[20:05] <gatox> not, make as: mklink :P
[20:16] <JanC> dobey: you can already install many 32-bit packages on a 64-bit Ubuntu system right now
[20:17] <dobey> JanC: yes; but i *only* want the kernel to be 64 bit
[20:18] <dobey> and you can't do that.
[20:18] <dobey> at least, not easily
[20:18] <JanC> dobey: I guess once all packages are modified to support multiarch, that should be doable
[20:19] <JanC> I'm not sure if they are far enough to start a 32-bit desktop system though
[20:19] <JanC> 32-bit userspace desktop
[20:20] <JanC> partly depends on Debian too, I guess
[20:20] <dobey> you don't need anyting special to start the 32 bit userspace
[20:21] <dobey> of course, if the pae kernel worked, i would be fine with that too
[20:21] <dobey> but alas
[20:21] <JanC> dobey: I think that at least libc needs to be different?
[20:22] <JanC> also, some applications can use 64-bit support too
[20:22] <dobey> libc doesn't need to be different afaik; lest something in the kernel requires the 64bit libc to be installed
[20:23] <dobey> i guess i might need it for the nvidia driver to work or something
[20:23] <dobey> but i don't much care about that; as long as it works
[20:23] <dobey> all applications "can use" 64 bit, so long as they are compiled for it, yes
[20:24] <dobey> but for 99.9999999999999999999% of them, it's totally useless
[20:24] <JanC> by "can use" I mean it's useful for them
[20:24] <JanC> things like video/image editing in some cases
[20:25] <dobey> in what cases?
[20:26] <JanC> 64-bit is much faster for many such tasks (actually, amd64 can be much faster than ia32--this is not true for all architectures)
[20:26] <dobey> they would benefit more from developers not being afraid of threads
[20:27] <ralsina> dobey: most developers *should* be afraid of threads.
[20:27] <JanC> I am afraid of developers who aren't afraid of threads ;)
[20:27] <dobey> ralsina: most developers shouldn't be writing code :)
[20:27] <ralsina> dobey: just like they should be afraid of lightning and big rocks on hills ;-)
[20:28] <dobey> i love how CPU manufacturers started adding more cores, and developers stopped writing multithreaded code
[20:28] <dobey> more CPUs you can't use!
[20:28] <JanC> dobey: that's because suddenly their multithreaded code broke down with more than one core...  ;)
[20:29] <dobey> anyway, when i get 10 million dollars to make a movie, i'll buy a million dollar SGI to edit it on
[20:29]  * JanC thinks about the infamous ralink drivers for example...
[20:31] <ralsina> dobey: just in case question: if I delete this branch, that pisses you off, right? https://code.launchpad.net/~ralsina/ubuntuone-control-panel/unique_snowflake
[20:32] <dobey> deleting branches is annoying, yes
[20:32] <dobey> but whatever
[20:32] <ralsina> with many cores comes great deadlocks
[20:32] <ralsina> dobey: so, want me to delete the merge proposal first instead? Just want to do it right
[20:32] <dobey> if your threaded code breaks on 2 cores, it probably breaks on 1
[20:32] <ralsina> dobey: well... it breaks faster!
[20:33] <dobey> ralsina: if you're going to delete it, then just delete it; deleting the proposal first makes more work
[20:33] <ralsina> dobey: if there is something better instead of deleting it, since it's useless now, I am all ears
[20:33] <JanC> dobey: 1 core can never run 2 threads at the same time, so there certainly is code that only breaks on multi-core systems...
[20:35] <dobey> ralsina: not really
[20:35] <ralsina> dobey: ok, then sorry dude!
[20:36] <dobey> JanC: breaks and breaks in an obvious way aren't the same thing; it was still broken on the single core, even if it wasn't obvious
[20:52] <ralsina> can I get reviews forthis one of these days? https://code.launchpad.net/~ralsina/ubuntuone-windows-installer/quote_autostart
[20:54] <dobey> one of these days maybe
[20:55] <ralsina> dobey: ok, no rush ;-)
[20:57] <dobey> looks ok to me. i won't even try to pretend that windows would be sane about that
[21:01] <ralsina> dobey: yeah, so quoting doesn t hurt. I expect.
[21:03] <dobey> well, now that lp let me, +1 from m
[21:03] <dobey> me
[21:05] <ralsina> thanks dobey!
[21:09] <dobey> ugh contacts picker
[21:30]  * nessita -> eod
[21:30] <nessita> bye all!
[21:31] <ralsina> bye nessita!
[21:31] <nessita> ralsina: have a nice holiday!
[21:31]  * nessita -> run to pilates
[21:31] <ralsina> nessita: thanks!
[21:49] <dobey> why did gtktable have to get deprecated
[21:49] <dobey> grubmle
[22:00] <dobey> i think i'll just add a -Wno-error= to ubuntuone-client-gnome for now at least
[22:06] <mfisch> Is there a way to set U1 to sync files who's name that start with a # sign?  like my bip logs?
[22:08] <dobey> later all