[00:23] <urbanape> statik: damnation. Looks that way. Don't know what happened. Must have gotten overlooked with the holiday and bindwood push.
[00:38] <urbanape> hmm, wonder what this means: bzr: ERROR: Public branch "bzr+ssh://bazaar.ubunet/~urbanape/ubuntuone-servers/new-folder-inline-edit" lacks revision "zachery.bir@canonical.com-20100111164746-gw9ghf16yzh3u4f9".
[00:39] <urbanape> ah, nm. Needed to push.
[00:43] <urbanape> statik: submitted
[02:20] <hamlinux> is there a tutorial available for sharing files.  Most notably version control when corroborating work
[05:28] <rico49> hi. i'm running UbuntuOne on Ubuntu 9.10 fully updated. I can see my directory structure under the ~/UbuntuOne "root" folder. None of the file sync. Any help on this would be welcome.
[05:29] <rico49> For example, are any particular file permissions required?
[05:29] <rico49> And does my UbuntuOne name have to match my login name at the Linux level?
[05:29] <rico49> Thanks!!!
[05:29] <mattgriffin> rico49: hi
[05:29] <rico49> mattgriffin: hi
[05:29] <mattgriffin> login names don't need to match
[05:29] <rico49> thanks
[05:30] <mattgriffin> rico49: have you completed the Add a Computer process for the system you are using?
[05:30] <rico49> As far as I know. My computer shows up in the list -- which has only this computer in it.
[05:31] <rico49> I may try deleting it and adding it again. I'm wondering if I reinstalled Ubuntu since my original signup and thus destroyed the certificate
[05:32] <mattgriffin> rico49: if so, you should have an Ubuntu One token under Applications >> Accessories >> Passwords and Encryption Keys
[05:32] <mattgriffin> rico49: are you using the regular Ubuntu One package or have you setup the beta PPA in your Software Sources?
[05:33] <rico49> I have not added beta PPA to my software sources.
[05:33] <rico49> I do see the token
[05:34] <mattgriffin> you can go to one.ubuntu.com/account and remove all of the authenticated machines... this should force your computer to re-authenticate after restarting the ubuntu one sync client. it will also create a new token.
[05:34] <rico49> should i add the PPA to my software sources and upgrade?
[05:35] <mattgriffin> rico49: you shouldn't have to if you are running a pretty standard Ubuntu installation
[05:35] <mattgriffin> are you behind a firewall?
[05:35] <mattgriffin> rico49: sorry... i meant proxy server
[05:35] <rico49> well, i'm running on my home wireless system. i didn't setup a firewall but I didn't go look to see if it was filtering anything due to not running into issues before
[05:36] <rico49> i could check that too.
[05:36] <rico49> do you happen to know the ports and protocol that UbuntuOne uses?
[05:36] <rico49> yup... proxy....
[05:36] <rico49> no proxy server active
[05:36] <rico49> beyond the wireless router
[05:37] <mattgriffin> rico49: https://wiki.ubuntu.com/UbuntuOne/#u1storage
[05:37] <mattgriffin> rico49: ports
[05:37] <rico49> Thanks!!!
[05:38] <rico49> this page is very helpful....
[05:39] <mattgriffin> rico49: np. i'm not really a technical member of the team. i recommend reviewing that wiki page as well as these (https://wiki.ubuntu.com/UbuntuOne/Tutorials/ and https://wiki.ubuntu.com/UbuntuOne/Bugs) if you want to work on fixing it tonight. members of the development team will join the channel tomorrow morning so they can help with deeper diagnostics if needed.
[05:40] <rico49> mattgriffin: I appreciate your help and will follow the links you provide.  The information looks like its just what I need.
[05:49] <rico49> mattgriffin: After I deleted the machine from the list of active machines and then rebooted everything appears to be working.  Thank you very much!!!
[05:50] <mattgriffin> rico49: awesome! no problem. have fun!
[05:50] <rico49> :-)
[11:41] <jussi01> honk!
[11:41] <jussi01> :D
[11:42] <drbobb> hello, my "Ubuntu One" folder is a mess, with *.u1conflict files all over the place
[11:42] <drbobb> including a subfolder that i can't fix in any way
[11:43] <drbobb> this service is simply not working correctly for me
[11:43] <jussi01> drbobb: perhaps its a good idea to start from the beginning, what you did, how you installed, your version of ubuntu etc
[11:44] <drbobb> ubuntu karmic
[11:44] <drbobb> I have only 2 pc's hooked up to my ubuntuone account
[11:44] <drbobb> my laptop and my office desktop
[11:45] <aquarius> jussi01, heh. I'm working on it :)
[11:45]  * jussi01 hi5's aquarius
[11:45] <drbobb> since I never use both at once, there is no reason for anu *.u1conflict files to appear in the first place
[11:45] <jussi01> drbobb: and how did you install it?
[11:46] <drbobb> jussi01: I don't recall whether I used apt-get, aptitude or synaptic, does it make much of a difference?
[11:46] <jussi01> (or is it installed by default on ubuntu?)
[11:46] <jussi01> drbobb: ok, just checking it was from the repos, not from some other download
[11:47] <drbobb> in any case it's official deb packages
[11:47] <jussi01> drbobb: so, when did thes files appear?
[11:47] <jussi01> drbobb: yes, but sometimes official packages can be old, beta, etc.
[11:48] <drbobb> jussi01: the ones I installed were from karmic/main
[11:48] <jussi01> right. so when di the files appear?
[11:48] <drbobb> uh let me see
[11:49] <drbobb> I'm getting them all the time, since late november or so
[11:49] <jussi01> ok, and could you tell your use case? ie. how you use the service?
[11:49] <drbobb> in one case I have a subfolder called L1, that gets renamed to L1.u1conflict whenever I sync
[11:50] <drbobb> and doesn't appear in the web interface no matter what
[11:50] <drbobb> although it is present on both my client machines
[11:50] <jussi01> argh, I need to disappear for a time, back in a little while. meantime, I hope someone here can help
[11:50] <jussi01> (work... just lovely)
[11:51] <drbobb> generally I create or edit a file on either of the machines, and sync ie. connect when I'm done
[11:52] <drbobb> then after a while I'm at the other machine, and I see that file is there except it has .u1conflict appended to its name
[11:54] <jussi01> drbobb: oh, and Ive still like 1 sec, but take a look here: https://answers.edge.launchpad.net/ubuntuone-client/+faq/677
[11:55] <drbobb> jussi01: I read that, but it doesn't seem to work for me
[11:55] <drbobb> for instance, i rename the folder L1.u1conflict back to L1 manually
[11:56] <drbobb> as soon as I connect it gets renamed back
[11:57] <drbobb> now I have L1.u1conflict on both client machines, even though their contents are the same
[11:57] <drbobb> and no L1 on either
[11:58] <rtgz> drbobb, hi. In case you put new file to the directory, does it get synced or it is immediately renamed to .u1conflict?
[11:59] <rtgz> drbobb, I mean if you are connected to the server
[11:59] <drbobb> rtgz: I just did that yesterday, put a new file in on one client
[11:59] <drbobb> now on the other client I see only a .u1conflict file
[12:00] <drbobb> and none with the original's name
[12:00] <rtgz> drbobb, ok, say, you put the file to computerA. Does it get synced?
[12:00] <drbobb> supposedly yes, I get a notification that says 1 file updated
[12:01] <rtgz> drbobb, okay, then computerB tries to sync, downloads something that is immediately renamed to conflict file, right?
[12:01] <drbobb> so it seems
[12:01] <rtgz> drbobb, okay, do you have the access to both machines at this time?
[12:01] <drbobb> right now only one of them
[12:01] <drbobb> the other is at home and switched off
[12:02] <rtgz> drbobb, ok. Does computerA synces u1conflict files if you put them to computerB?
[12:02] <drbobb> oh and one more thing
[12:02] <drbobb> I tried to fix the L1 folder by moving it out of Ubuntu One on both machines
[12:03] <drbobb> and copying it to the server via the web interface
[12:03] <drbobb> but that doesn't work
[12:03] <drbobb> I click on create new folder, provide a name and confirm
[12:03] <drbobb> but nothing happens - no folder is created
[12:04] <rtgz> drbobb, hm... Is there anything special with the file name (aka, could you provide it here for me to test the web ui and client) ?
[12:05] <rtgz> drbobb, I mean the path name, etc.
[12:05] <drbobb> the (sub)folder name is L1
[12:06] <rtgz> drbobb, anyway, here's how that can be diagnosed - first create a new folder, say 'testing' on ComputerA and create the directory structure mimicking the one that you already have. This will allow us to get info faster when browsing for relevant log entries.
[12:07] <drbobb> ok, I guess I should come back when I'm able to contract both computers at once
[12:07] <rtgz> drbobb, then it would be good to start ubuntuone-syncdaemon with debug enabled on the computerB ( https://wiki.ubuntu.com/UbuntuOne/Bugs#Temporary%20debug%20mode ).
[12:07] <drbobb> now I gotta get working, c.u. again
[12:08] <drbobb> ok, thx for that suggestion
[12:08] <rtgz> drbobb, then try connecting via applet and check whether files in testing dir get u1conflict
[12:09] <drbobb> rtgz: some files do get u1conflict
[12:09] <drbobb> but in the case at hand it's the folder itself
[12:09] <rtgz> afterwards post your computerB's ~/.cache/ubuntuone/log/syncdaemon.log to the ticket (or paste.ubuntu.com if you get to IRC from computerB as well)
[12:09] <drbobb> ok will do
[12:09] <rtgz> drbobb, thanks
[12:10] <rtgz> grrr, we need better conflict notification, i.e. "This file is considered to be conflicting because there appears to be newer file in the web interface and local copy was changed"
[12:12] <rtgz> and the ability to chose whether to resolve conflicts manually (there appears some file conflicts in your U1 directory. How do you want to handle them? [Rename my to u1conflict], [Examine all files]... etc.
[13:18] <rtgz> tomboy note web ui greates uuids w/o dashes. Is it ok?
[13:23] <verterok> dobey, Chipaca: ping
[13:24] <verterok> dobey, Chipaca: I know we already defined the UDFs DBus API, but was wondering if you guys needs something specific for the Shares/d bits
[13:35] <rtgz> verterok, is support to obtain info on directory planned? I.e whether all files / folder under it are synced or not. Maybe it is better described as qecursive query for a path info.
[13:36] <rtgz> i missed several days here being sick, breaking/fixing router and now I am back to ask weird questions and provide weird solutions to nonexistent issues :)
[13:36] <verterok> rtgz: not for the time frame of the UDFs feature, but it should be simple to add such method
[13:37] <verterok> rtgz: :)
[13:39] <rtgz> verterok, since nautilus cannot be told to gather dbus requests this may mean a lot of metadata querying. ATM the only thing I know about metadata is something somewhere is slow...
[13:39] <rtgz> verterok, I mean these requests will need to be sent for every folder and file. every folder request should trigger recursive internal checks.
[13:40] <rtgz> Btw, applet is going away for Lucid, what's going to replace it?
[13:40] <verterok> rtgz: I think the problem with doing a lot of dbus requests to get the metadtata of each file/dir is the cost of the dbus call itself
[13:41] <verterok> rtgz: querying the metadata should be fast in the order of ~100 objects (files/dirs)
[13:41] <verterok> s/100/1000/
[13:41] <rtgz> verterok, this is done in the nautilus extension at the moment, so we cannot speed up this part.
[13:41] <verterok> but don't have any benchmark :)
[13:41] <rtgz> verterok, really need to see how that magical metadata looks like.
[13:42] <rtgz> Btw, one.ubuntu.com is not yet IPv6-ready :-P
[13:42] <verterok> rtgz: this is a very simple way to get all the metadtata of a tree: http://pastebin.ubuntu.com/355495/
[13:43] <verterok> rtgz: the metadata it's just a bunch of pickled python dicts
[13:43] <verterok> rtgz: ls -R ~/.local/share/ubuntuone/syncdaemon/fsm
[13:44] <rtgz> verterok, every file has its own corresponding medatada file?
[13:44] <verterok> pfibiger: do you know if there are any plans to make one.ubuntu.com IPV6-friendly? ^
[13:44] <verterok> rtgz: yes
[13:45] <rtgz> verterok, I am already prepared to test IPv6 connections :), that's why I am poking everyone everywhere and it seems that not much people out there are excited as much as I am with this :)
[13:50] <pfibiger> verterok: i wouldn't say there are specific plans, but we ought to. file a bug? :)
[13:50] <verterok> pfibiger: ok, thanks! :)
[13:51] <verterok> rtgz: ^
[13:51] <rtgz> verterok, he he... bug bug bug
[13:58] <urbanape> jamesh: ping
[13:58] <urbanape> or jblount: ping
[14:00] <urbanape> jamesh, jblount: nm, found it
[14:17] <joshuahoover1> dobey: ping
[14:25] <dobey> hi joshuahoover1
[14:26] <dobey> verterok: i don't know that it needs any changes? and chipaca is on holiday this week
[14:26] <joshuahoover1> dobey: hi! do have any ideas on bug 395483?
[14:26] <verterok> dobey: oh, ok.
[14:27] <verterok> dobey: just wondering if there are any plans to handle shares in the UI and you guys might need something extra from the API
[14:27] <verterok> dobey: we can't take a look to it next week
[14:27] <verterok> s/can't/can/ :p
[14:27] <dobey> ok
[14:28] <adiroiban> rodrigo__: hi, do we still need to provide the unnamed main couchdb and desktopcouch classes / headers?
[14:32] <dobey> joshuahoover1: commented + Incomplete :)
[14:34] <joshuahoover1> dobey: thanks!
[14:35] <dobey> i guess i do have to do bug 'day' today, since i'm not in a meeting room somewhere :)
[14:48] <rodrigo__> adiroiban, no, we are breaking the API like crazies on the unstable branch (CouchDB->Couchdb), so no need to, I think
[14:48] <rodrigo__> adiroiban, well, maybe we could add a couchdb-compat.h file that #define's all the new stuff to match the old one
[14:48] <rodrigo__> adiroiban, but since only evolution-couchdb uses it that I know, I don't think we need it
[14:49] <adiroiban> rodrigo__: yes. same opinion here
[14:49] <rodrigo__> adiroiban, but if you want, write that file, if evolution-couchdb from gnome-2-28 branch in git compiles with it, then it's working
[14:49] <adiroiban> I think it should be better to fix evolution-couchdb
[14:49] <rodrigo__> adiroiban, unstable branch of evo-couchdb now compiles fine with unstable coucjdb-glib
[14:50] <adiroiban> job done :)
[14:50] <rodrigo__> adiroiban, it fails to run on my system because I've by mistake installed e-d-s 2.30, which has some corba->dbus changes
[14:50] <rodrigo__> but I should be fixing it soon to work with both 2.28 and 2.30
[14:50] <adiroiban> ok
[14:50] <diverse_izzue> hey all! is there (other than hanging out on irc) a way of knowing what's going on in the ubuntuone-world, such as a blog?
[14:51] <rodrigo__> diverse_izzue, there is a twitter and a blog yeah
[14:51]  * rodrigo__ looks for the blog
[14:51] <dobey> there is a blog, and twitter/identi.ca accounts, and a facebook page
[14:51] <rodrigo__> http://voices.canonical.com/ubuntuone/?feed=rss2
[14:51] <adiroiban> I build the source package for the latest couchd-db. works fine for me https://edge.launchpad.net/~adiroiban/+archive/ppa/+packages
[14:51] <rodrigo__> ah, and facebook, yeah
[14:52] <rodrigo__> adiroiban, cool!
[14:52] <rodrigo__> adiroiban, works fine, as it installs fine, or the API really works fine?
[14:52] <adiroiban> both
[14:52] <rodrigo__> woohoo!
[14:53] <rodrigo__> ok, so as soon as evo-couchdb is fixed for the 2.28/2.30 issue, I'll submit it for lucid
[14:53] <rodrigo__> adiroiban, well, or you can submit it yourself, if you want
[14:53] <adiroiban> yep. my packaging changes are here https://code.edge.launchpad.net/~adiroiban/couchdb-glib/ubuntu-0-5-99
[14:53] <rodrigo__> adiroiban, but I'll do a 0.6.0 release before that
[14:54] <rodrigo__> adiroiban, you haven't pushed the branch
[14:56] <adiroiban> rodrigo__: true, it looks like there were some error durring the push
[14:57] <adiroiban> solved
[14:58] <adiroiban> rodrigo__: what about the disable-debug merge proposal?
[14:58] <adiroiban> is it ok?
[14:58] <adiroiban> or you would like some other changes
[14:58] <rodrigo__> adiroiban, yeah, sorry, it's in my queue, just busy with some other stuff
[14:58] <adiroiban> ah. np
[14:59] <rodrigo__> haven't looked at your last change, so don't worry for now, I'll add a comment as soon as I review it
[14:59] <diverse_izzue> rodrigo__ & co: thanks. what are you planning for lucid?
[15:00] <rodrigo__> adiroiban, the package branch looks great, so keep it around, and we'll submit that for lucid once we release 0.6.0 and evo-couchdb is fixed
[15:01] <rodrigo__> adiroiban, oh, working on vala bindings?? -> https://code.edge.launchpad.net/~adiroiban/couchdb-glib/vala-bindings
[15:01] <rodrigo__> adiroiban rocks :D
[15:01] <adiroiban> well, I need to do that work for building my vala app
[15:02] <adiroiban> and I will plan to maintain them
[15:02] <adiroiban> so I pushed that branch in case someone else find them useful
[15:02] <adiroiban> if I have time, maybe I will write some vala tests for the vapi files
[15:03] <rodrigo__> adiroiban, a friend of mine started writing vala bindings a long time ago, maybe you can get in touch with him, not for his code, which is out of date now, but for his plans, which I think were to be part of the vala bindings project/package
[15:03]  * rodrigo__ looks for his email
[15:05] <jblount> Desktop+ MEETING BEGINS
[15:05] <jblount> Say "me", document status, etc
[15:05] <jblount> me
[15:06] <teknico> me
[15:06] <rodrigo__> me, document status, etc :)
[15:06] <CardinalFang> me
[15:06] <teknico> oh right, sorry: "document status, etc"
[15:07] <urbanape> me
[15:08] <jblount> DONE: Worked on layout issues tagged gd in launchpad
[15:08] <jblount> TODO: FACE
[15:08] <jblount> BLOCKED: Nope
[15:08] <jblount> teknico: tag!
[15:09] <teknico> DONE: finished fighting for net access, situation stable now; more work on two branches for configuring funambol for sending sms messages (#418048); dealt with a dormant branch, waiting for desktopcouch's new version
[15:09] <teknico> TODO: propose and land two branches for configuring funambol for sending sms messages (#418048); finish setting up a development environment in a virtual machine; start implementing the mobile sync REST API for client app (#504689)
[15:09] <teknico> BLOCK: none
[15:09] <teknico> next: CardinalFang
[15:09] <teknico> uff, gnome slows me down ;-P
[15:09] <rodrigo__> eh, I'm next!
[15:10] <rodrigo__> • DONE: Submitted latest stable couchdb-glib/evo-couchdb for Lucid. Changed DBus's create_share to accept more than one user/email. Looked at tomboy syncing problems. Some renaming of classes in couchdb-glib unstable. Almost got new XML<->HTML note converter
[15:10] <rodrigo__> • TODO: Conflict resolver tool in pair tool. Look at becoming a MOTU (https://wiki.ubuntu.com/UbuntuDevelopers). Make sandy's snowy test suite work with our server (http://git.gnome.org/cgit/snowy/tree/api/tests.py). Discuss with jdo and aquarius about oauth token per app, not per machine? Send Otto (otto.greenslade@canonical.com) a screencast of contacts picker. Retrieve OAuth tokens from keyring in music store widget. Ch
[15:10] <rodrigo__> eck with Ken status of music store packages
[15:10] <rodrigo__> next: CardinalFang
[15:11] <teknico> rodrigo__, oops, sorry
[15:11] <teknico> I blame gnome once again ;-)
[15:12] <rodrigo__> teknico, yeah, sure, gnome's fault :D
[15:12] <teknico> :-D
[15:13] <urbanape> CardinalFang?
[15:13] <CardinalFang> sorry -- got distracted.  urbanape, you first.
[15:13] <urbanape> DONE: Bindwood's default operation is now complete. Local client understands and performs changes in location - moves happen! Submitted and landed new-folder-inline-edit branch.
[15:13] <urbanape> Started helping jblount with the public URLs branch for web UI.
[15:14] <urbanape> TODO: On-call review today, finish up public URLs branch, get migration and subsequent client code paths done for Bindwood.
[15:14] <urbanape> BLOCK: None.
[15:14] <urbanape> eom?
[15:14] <urbanape> oops, had that last one on the clipboard.
[15:14] <urbanape> CardinalFang is next
[15:15] <teknico> glipper, klipper, it's all a dripper
[15:17] <CardinalFang> DONE: slept off some of cold.  got sprint travel booked.  tomboy/snowy bug defense.
[15:17] <CardinalFang> TODO: reviews. triage. coffee. desktopcouch update into karmic.
[15:17] <CardinalFang> BLOCKED: None.
[15:17] <CardinalFang> Sorry for delay.
[15:17] <CardinalFang> EOM.
[15:17] <urbanape> jblount: should we maybe put the "make public" stuff on the info overlay?
[15:21] <jblount> urbanape: Yep. It should already be there in that branch I sent you.
[15:27] <dobey> oh i thought we didn't do the standup on tuesdays
[15:27] <dobey> because of the bug day thing
[15:28] <urbanape> oh, hey look, it is.
[15:29] <jblount> urbanape: :)
[15:29] <voytech> Hi people, what do You think about synchronizing akonadi in DesktopCouch/ UbuntuOne ? I'm trying to find a topic for my master thesis and I've found out that Aconadi is server for storing PIM data. In my opinion it could be nicely integrated with ubuntuOne. But as for now I don't have a knowledge how it works exactly. So I'm asking You If my idea id good
[15:30] <dobey> voytech: someone was working on an akonadi backend to store contacts in desktopcouch, but i don't know if that's gone anywhere or not yet
[15:30] <dobey> rodrigo__: do you know anything about that?
[15:30] <rodrigo__> I know the kdeab people were working on it, not sure about the status though
[15:31] <rodrigo__> statik should know, I think
[16:00] <voytech> statik: Do You know something about synchronizing akonadi in ubuntuone/desktopcouch ? If there is someone working on it ?
[16:02] <jblount> voytech: statik 's laptop is not working atm, so he might take a long time getting back to you
[16:06] <aquarius> voytech, there's already been some work done by Till Adam and the KDAB team to provide a CouchDB back end to Akonadi
[16:06] <aquarius> voytech, that code's in the upstream akonadi codebase, as I understand it
[16:06] <aquarius> I don't know if anyone's still working on it and keeping it up to date with current desktopcouch, though
[16:07] <aquarius> voytech, if you wanted to pick that code up and bring it up to date and provide really good Akonadi integration, that'd be great! We'd be happy to give advice about how to talk to desktopcouch
[16:08] <till> I've not kept up with desktop couch, lateluy, no.
[16:08] <till> The resource still worked, though, last time I tried.
[16:08] <aquarius> till, that's what I thought, on both those points :)
[16:08] <till> voytech: we (KDAB and the Akonadi team) will be happy to help
[16:08] <aquarius> excellent!
[16:08] <till> voytech: we just don't have the bandwidth to push it, atm
[16:08] <till> (and desktopcouch was a bit of amoving target)
[16:09] <aquarius> it's moving more slowly now than it was now that it's a stable API :)
[16:09] <till> voytech: please to meet you, btw, I'm Till, I coordinate the KDE and Maemo work at KDAB, and I wrote the akonadi/couch resource :).
[17:13] <voytech> I was disconnected for some time
[17:14] <aquarius> voytech, http://irclogs.ubuntu.com/2010/01/12/%23ubuntuone.html#t16:06 might be useful then :)
[17:14] <voytech> But I see my idea isn't bad :)
[17:18] <aquarius> jolly good idea, is what it is.
[17:37] <voytech> till: So you wrote the resource.. Where can I get some information about it ? I have to understand many thing fisrtly, becouse as I said before I'm new to ubuntuone/desktopcouch
[17:51] <voytech> till: And I would like to know how your resource exactly works ? Maybe I could download it to become familiar with it ?
[19:10] <gourgi> gourgi: hi jblount, hi all, for days now i'm seeing a "Something has gone wrong (500) Server Error" when trying to view my contacts from the webUI.  http://img709.imageshack.us/img709/4572/screenshotwv.jpg  . what i have to do now ?
[19:13] <jblount> gourgi: Hi! I think something is temporarily wrong with the contacts web ui. Let me ask around and see what the progress is on it.
[19:14] <gourgi> thanks jblount
[19:18] <jblount> gourgi: So that error page should have had a number on it that would help us track down the problem. I'm going to get this fixed, but I don't have any good way to track down your error.
[19:18] <jblount> gourgi: How long have you been seeing this?
[19:18] <gourgi> about 6-8 days now
[19:19] <jblount> gourgi: Wow, that's a long time. Thanks for letting us know.
[19:20] <gourgi> it is the 3rd day i'm coming here reporting this
[19:20] <gourgi> :)
[19:21] <rtgz> jblount, looks fine from here :-/
[19:21] <jblount> gourgi: Do you see the same problem here? https://edge.one.ubuntu.com/contacts/
[19:21] <jblount> rtgz: Yeah, it works for me, but I've only got a few contacts in.
[19:21] <gourgi> jblount yes
[19:24] <gourgi> same problem on the edge too.. jblount do you need to see my ~/.cache logs ?
[19:25] <rtgz> shot in the dark: gourgi, do you have some non-latin1 characters in the first/last name of at least one of your contacts?
[19:26] <jblount> gourgi: While I can't fix this for you now, I can make that number I mentioned show up on edge in a few hours so we can track down the problem. Have you submitted a bug about this yet?
[19:27] <gourgi> rtgz i have almost all my contacts in my native language (greek). jblount i haven't yet, but i will if needed.
[19:27] <rtgz> gourgi, good. will try with greek symbols
[19:27] <jblount> gourgi: It would be very helpful if you could submit a bug, that way we could track the progress of the fix.
[19:28] <gourgi> ok , i 'm filling it now
[19:28] <jblount> gourgi: Thanks!
[19:30]  * rtgz inserted an entry, waiting for replication...
[19:40] <rtgz> hm.. how does desktopcouch use Avahi?
[19:41] <aquarius> rtgz, for LAN sharing
[19:41] <rtgz> aquarius, and if there is no Avahi running?
[19:41] <rtgz> aquarius, 2010-01-12 21:41:16,446 - We seem to be running already, or can't publish our zeroconf advert.  org.freedesktop.DBus.Error.ServiceUnknown: The name org.freedesktop.Avahi was not provided by any .service files
[19:42] <aquarius> then it'll throw an error saying that we can't publish our zeroconf advert
[19:42] <aquarius> as you have discovered :)
[19:42] <rtgz> aquarius, will this error prevent, say, replication ?
[19:42] <gourgi> rtgz jblount thanks for looking into this!  bug filed  https://bugs.launchpad.net/ubuntuone-servers/+bug/506584
[19:42] <aquarius> we only publish a zeroconf advert if LAN sharing has been set up, so it doesn't affect most people
[19:42] <rtgz> aquarius, hm.. I have LAN sharing set up? O_O
[19:42] <aquarius> erm. I'm not sure whether replication as a whole is killed by that happening. CardinalFang would know better.
[19:43] <aquarius> rtgz, there is a non-zero chance that I am lying about that. If I am lying, though, and we publish zeroconf regardless, then it's a bug. :)
[19:43] <rtgz> aquarius, my ISP started to use .local domain (yep, will fight with them... SOON). My Avahi does not like this and dies.
[19:43] <aquarius> what?
[19:44] <aquarius> that's flat out wrong
[19:44] <aquarius> kill your ISP :)
[19:44] <rtgz> aquarius, cannot kill ISP. onIspKill="disconnect()"
[19:44] <aquarius> but if you're not doing LAN sharing then DC *shouldn't* publish zeroconf. But...it might do.
[19:44] <aquarius> wish cardinalfang was around, he knows more about that sharing bit than I do. I'd have to grovel through the code and work it out
[19:46] <dobey> so
[19:46] <dobey> publish != search
[19:46] <rtgz> aquarius, desktopcouch-service does not seem to return me the terminal prompt...
[19:46] <dobey> and searching for other shared things still requires working avahi
[19:46] <dobey> even if you aren't publishing anything over it
[19:47] <rtgz> the problem that it looks like my replication is now broken due to this :-/
[19:47] <aquarius> dobey, yeah, the point is that we only use avahi to detect the current location of already-paired LAN desktopcouches (well, and to actually *do* pairing). We know whether we are paired with LAN desktopcouches or not by looking in the management database, so *if* we are not paired with anything, we shouldn't publish a zeroconf advert (because it'll never be used. also, security.)
[19:48] <aquarius> rtgz, run desktopcouch-pair
[19:48] <aquarius> rtgz, to see if you're already LAN paired.
[19:48] <aquarius> I think that you're not, though, which suggests that we're doing zeroconf advertising when we don't need to be, and if that fails we're crashing. Thus, epic fail
[19:49] <rtgz> aquarius, it used to work, yet the last time i got replication worked was on 2010-01-04. Since then - ERRORs related to avahi and nothing went to U1... Scaaary
[19:49] <aquarius> rtgz, yeah. I think (a) we're publishing zeroconf when we shouldn't, and (b) failing to publish zeroconf stops replication, which is two separate bugs. If you could file those bugs that'd be really nice
[19:50]  * rtgz  - yeah, right - tried to test 1 bug, got 2 separate issues... yeah, right...
[19:52] <dobey> aquarius: if i buy a new laptop with ubuntu on it, and i bring it hope, and start up some ubuntu one thing (which in turn starts up desktopcouch), how does it know whether or not it should be paired with things on the LAN that are already shared?
[19:52] <aquarius> dobey, it is not paired. DC doesn't do LAN pairing unless you specifically ask it to, with the desktopcouch-pair tool
[19:52] <dobey> aquarius: it doesn't check to see if something's already published, and try to connect to it if so?
[19:52] <aquarius> it doesn't magically set up a mesh network
[19:52] <dobey> boo.
[19:52] <dobey> so why are we using zeroconf? :)
[19:52] <dobey> (maybe we should call it oneconf)
[19:53] <rtgz> aquarius, http://paste.ubuntu.com/355670/
[19:53] <aquarius> how else do you propose that paired machines A and B find one another again?
[19:53] <rtgz> dobey, noconf in my case
[19:53] <aquarius> hahaha. So desktopcouch-pair *and* desktopcouch itself both assume that avahi works. Oops.
[19:54] <dobey> aquarius: connect()? :)
[19:55] <aquarius> dobey, connect to...where?
[19:55] <jblount> aquarius: Are you going to push your Boxee ui to lp?
[19:56] <dobey> "Host: [             ]"
[19:56] <dobey> aquarius: my point is more that if we're going to use zeroconf, we might as well use it right :)
[19:56] <aquarius> jblount, it's on my svn server because setting up LP projects is a pain in the arse. I have, now, another task on my list: create a bzr lp-new-project command that creates a project.
[19:56] <dobey> (and make the user experience better)
[19:57] <aquarius> dobey, you don't know the host (DHCP changes IPs, foo.local names don't work on Windows), and you don't know the port (DC randomly generates it every time it starts).
[19:57] <aquarius> we can't ship Ubuntu with a CouchDB exposed to the universe listening by default. pitti would have me boiled alive.
[19:57] <dobey> huh?
[19:58] <dobey> i never said anything about having couchdb listening remotely by default
[19:58] <aquarius> so how do you get your magic mesh network started, then?
[19:58] <aquarius> well, OK, our little daemon listening by default, not couch itself
[19:58] <aquarius> but it's still listening by default, which is a no-no.
[19:59] <rtgz> aquarius, ubuntuone-client or desktopcouch - who is going to be the unfortunate owner of the bug?
[19:59] <dobey> there's policy that apps can't find stuff shared by zeroconf from other machines, by default?
[19:59] <jblount> gourgi: That fix is on it's way to edge now, is it ok if I ping you in a few hours (if you are around) to try to get us a erorr number?
[19:59] <aquarius> rtgz, desktopcouch. assign it to cmiller, if you would, which he will give me a fishy look for but he's not here :)
[19:59] <dobey> aquarius: i was saying we should let the user see "Oh, these are being shared on your network, pick one if it's yours, or ignore"
[20:00] <dobey> and then let the user authenticate and pair if one of them is theirs
[20:00] <aquarius> yeah, but if we don't listen by default then there will never be any listening!
[20:00] <dobey> what are you talking about listening?
[20:00] <gourgi> jblount thanks! yes, ping me please , i'll be around and i hope not in bed :)
[20:00] <dobey> avahi is doing the listening, not us
[20:00] <dobey> it's ALRADY doing it
[20:00] <dobey> we're just not doing anything with the info
[20:00] <aquarius> dobey, yeah, but avahi needs to advertise a port that we can connect to
[20:00] <dobey> aquarius: no it doesn't
[20:00] <gourgi> jblount if i don't reply please add a comment on my bug report
[20:01] <aquarius> avahi has remote service activation?
[20:01] <dobey> aquarius: you're conflating SERVING and PAIRING
[20:01] <jblount> gourgi: Will do, it'll take some time, but I'll catch up with you here or there. Thanks again for your persistence :)
[20:01] <dobey> aquarius: i'm only talking about the client side discovery part
[20:01] <aquarius> dobey, so what does avahi advertise, then?
[20:01] <dobey> aquarius: we should obviously only ever publish if the user enables the "[ ] Publish" checkbox
[20:01] <gourgi> jblount no problem , i wish i can help more and solve this
[20:01] <dobey> aquarius: why do you think it's advertising something?
[20:02] <aquarius> dobey, I don't get what shows up in the list of "these machines are being shared on your network"
[20:02] <dobey> aquarius: nothing, if nothing is publhsing stuff on your network
[20:02] <aquarius> right. Then, that's what we currently do.
[20:02] <dobey> *sigh*
[20:03] <dobey> you're conflating SERVING and DISCOVERING, which are separate things :)
[20:03] <dobey> stop it :)
[20:04] <aquarius> I don't get what's supposed to happen once I've DISCOVERed that machine2 is sharing. What do I do then?
[20:04] <dobey> aquarius: well if we discover something, then we let the user pair with it
[20:04] <aquarius> how?
[20:04] <dobey> uhm, how do we pair with it now?
[20:05] <aquarius> by connecting to the advertised port on that machine
[20:05] <aquarius> which means that it's SERVING.
[20:05] <dobey> aquarius: it doesn't mean the lcoal machine is serving
[20:05] <dobey> it means some other machine is serving
[20:06] <dobey> avahi doesn't need to be serving anything to discover other machines which are
[20:06] <aquarius> yeah, I agree. What you're describing is basically what we do, except that we don't give machines the ability to serve unless they need to be, in the same way that you turn off bluetooth discovery after a minute or so
[20:08] <dobey> i was saying we should make it easier to pair with other machines that are already publishing
[20:08] <dobey> you were saying that we should "only ever use avahi if we intend to publish"
[20:08] <dobey> which doesn't make sense to me :)
[20:09] <dobey> aquarius: and https://edge.launchpad.net/groundcontrol might have stuff to create projects... i don't remember if it does or not :)
[20:09] <aquarius> ah. What I meant was: since we don't allow machines to be publishing unless they're already paired, we only need to serve if we already have pairings. If we do not have pairings, we don't need to serve, and thus we don't need avahi.
[20:10] <aquarius> your disagreement is because you want machines to be able to serve even when they're not paired, and I don't.
[20:16] <rtgz> OK, asked ISP to bring the zone down, since this is only used for BitTorrent retracker
[20:17] <dobey> aquarius: and how do you pair two machines, if neither of them are serving?
[20:18] <aquarius> dobey, you run desktopcouch-pair on machine 1, it tells you to run desktopcouch-pair on machine2, they then find one another.
[20:19] <dobey> ugh :)
[20:19] <aquarius> you have to do something on machine2, otherwise I can just pair with your shared DB without your consent and steal all your data.
[20:20] <dobey> i wonder how well couchdb deals with > 2 machines paired together
[20:20] <aquarius> if your answer is: don't turn on publishing if you will ever, ever be on the same LAN as any untrusted computer ever, then, well, fail.
[20:20] <dobey> so if i pair with my machine at home, and then go to a coffee shop, my couchdb can be violated by anyone?
[20:21] <dobey> there's no authentication at all?
[20:21] <aquarius> No, because you don't have the oauth tokens.
[20:21] <aquarius> The pairing process exchanges them.
[20:21] <dobey> but after i pair, and i go to another network, any machine there can run desktopcouch-pair, and it'll see my couch?
[20:22] <dobey> and can try to exchange tokens with it?
[20:27] <rtgz> since there are many ISPs that have bad .local habit I wonder why people did not complain earlier...
[20:27] <dobey> there probably aren't that many?
[20:27] <rtgz> and since there is no info from replication thread that something is not working we give false hope... again
[20:31] <rtgz> Bug #327362
[20:38] <rtgz> Ok, fixed that on the router...
[20:43] <aquarius> dobey, no. we advertise like this: if we are paired with another machine, we advertise couch itself (protected by oauth). If we are currently *in the process of pairing*, we also advertise the pairing daemon.
[20:45] <rtgz> ok, greek letters went through fine
[20:51] <dobey> hmm
[21:11] <dobey> anyone seeing http://pastebin.ubuntu.com/355715/ when running make test in ubuntuone-client trunk?
[21:15] <rtgz> dobey, checking...
[21:28] <rtgz> dobey, you are not alone
[21:34] <dobey> indeed
[21:37] <dobey> hmm
[21:45] <dobey> well at least I already filed the bug for it... in September...
[22:05] <urbanape> jblount: made some good progress today on the branch. I think things will work now, but still want to add some niceties.
[22:13]  * rtgz oO( this is the first time I ran `make test` on ubuntuone-client )