| xant | hello everyone | 00:08 |
|---|---|---|
| xant | Is ubuntu one back up? | 00:08 |
| pfibiger | xant: it is. | 00:10 |
| xant | My Ubuntu one fails to start | 00:19 |
| pfibiger | xant: you should have a log file, in ~/.cache/ubuntuone/ called syncdaemon-exceptions.log | 00:22 |
| xant | pfibiger, I do | 00:22 |
| xant | failure Connection to the other side was lost in a non-clean fashion. | 00:23 |
| pfibiger | ok, what happens if you quit the ubuntuone client and restart it? | 00:23 |
| xant | pfibiger, it crashes | 00:24 |
| xant | BadTransition: State READY_WITH_NETWORK_WITH_BOTHQ can't handle the SYS_CONNECTION_MADE event | 00:26 |
| xant | 2009-12-22 00:19:18,015 - pyinotify - ERROR - The path /home/xant/Ubuntu One/private of this watch <Watch wd=4 mask=3064 auto_add=False proc_fun=None path=/home/xant/Ubuntu One/private dir=True > must not be trusted anymore | 00:26 |
| xant | from syncdaemon-exceptions.log | 00:26 |
| xant | https://bugs.launchpad.net/ubuntu/+source/ubuntuone-client/+bug/499290 | 00:26 |
| ubottu | Ubuntu bug 499290 in ubuntuone-client "Ubuntu one fails to start" [Undecided,New] | 00:26 |
| xant | ubottu, that one | 00:27 |
| xant | I'm reinstalling ubuntu one | 00:29 |
| pfibiger | xant: ok. i believe that's a known error that's fixed for upload into karmic. i've added a few tags to your bug and assigned it. | 00:30 |
| xant | pfibiger, thanks | 00:31 |
| xant | followed the steps found in https://answers.edge.launchpad.net/ubuntuone-client/+faq/778 | 00:40 |
| xant | Ubuntu one still crashes | 00:40 |
| xant | syncdaemon-exceptions.log doesn't get created this time | 00:40 |
| xant | oath-login.log contains "Starting Ubuntu One client version 1.0.2" | 00:40 |
| pfibiger | anything in syncdaemon.log? | 00:41 |
| xant | It doesn't get created | 00:42 |
| pfibiger | xant: that's the only line in the oauth-login.log? | 00:47 |
| xant | pfibiger, yes | 00:47 |
| xant | I'm tempted to get spideroak :-) | 00:51 |
| aquarius | xant, can you kill ubuntu one, then restart it from a terminal to see if it displays an error? | 00:52 |
| aquarius | xant, in a terminal window, do: killall ubuntuone-syncdaemon; killall ubuntuone-client-applet; ubuntuone-client-applet | 00:53 |
| xant | aquarius, it works, thanks | 00:54 |
| aquarius | xant, and it hasn't crashed? | 00:54 |
| aquarius | xant, I suspect, then, that something was wrong with the syncdaemon before, and restarting it has fixed the problem | 00:54 |
| xant | Sorry everyone, I forgot to kill ubuntuone-syncdaemon | 00:55 |
| aquarius | xant, no problem! Sorry it took us a while to get to the source of the problem | 00:56 |
| xant | Hm | 01:03 |
| xant | I've got encfs set up in my ubuntu one folder | 01:03 |
| xant | it started synchronising after I unmounted my folder | 01:04 |
| xant | I wonder why | 01:04 |
| xant | Ubuntu ones seems to be stuck while updating the 11th file | 01:09 |
| xant | I'd like to suggest a feature | 01:14 |
| xant | When left-clicking the applet, it can show the file path it's currently updating | 01:14 |
| aquarius | the applet is going away in Lucid | 01:24 |
| aquarius | so there'll be more room in the Ubuntu One control panel to show that sort of thing | 01:25 |
| aquarius | (if you want to find out what it's doing right now, you can do "u1sdtool --current-transfers" in a terminal) | 01:25 |
| aquarius | and it would be possible to write a "monitoring app" which talks to syncdaemon via D-Bus and shows in minute detail what it's up to, if someone wanted to :) | 01:26 |
| xant | pfibiger, aquarius, thanks for your help and good night! | 01:50 |
| aquarius | xant, no problem | 01:50 |
| [dutchie] | haven't been able to sync tomboy notes for a while. I clicke "synchronise notes", then I get a dialog box (screenshot in a moment) | 13:16 |
| === [dutchie] is now known as dutchie | ||
| dutchie | http://imagebin.org/76479 | 13:18 |
| aquarius | dutchie, I'm having the same problem, and I think I have the fix; give me one minute | 13:29 |
| dutchie | also, honk :) | 13:30 |
| aquarius | dutchie, OK, you need to re-authorise Tomboy, after some changes we made on the server | 13:30 |
| dutchie | right-o | 13:31 |
| aquarius | so, right-click the Tomboy icon, Preferences, Synchronisation, Clear (and say yes), then Connect to Server and reauthorise. | 13:31 |
| aquarius | Sorry about that. | 13:31 |
| aquarius | mattgriffin, we should put out the word about that, I think | 13:31 |
| dutchie | worked | 13:31 |
| dutchie | it has to be said, this is a very efficient IRC channel | 13:32 |
| aquarius | word put out on twitter and identica | 13:39 |
| Chipaca | aquarius: I did that, and got a bunch of conflicts on the tomboy "default" notes, and a "failed to synchronize" | 13:48 |
| aquarius | chipaca, after clearing your tomboy sync and re-doing it? | 13:48 |
| Chipaca | aquarius: yes | 13:49 |
| aquarius | chipaca, OK, that's worrying | 13:49 |
| Chipaca | aquarius: and now did it again, and got a bunch of conflicts again, but it worked | 13:49 |
| Chipaca | aquarius: and now again, and failed to sync again | 13:52 |
| * aquarius headbutts the desk | 13:52 | |
| Chipaca | aquarius: synchronizing is hard! I'm going shopping | 13:52 |
| aquarius | I don't know what's wrong with it :( | 13:52 |
| Chipaca | aquarius: urbanape: dobey: CardinalFang: you do know what the platform sprint is about, right? | 13:53 |
| aquarius | Broadly, although getting your take on why we're going would be useful. | 13:53 |
| === CardinalFang changed the topic of #ubuntuone to: Talk to CardinalFang if you have problems | https://one.ubuntu.com | https://launchpad.net/ubuntuone | Please honk if you want a music store | ||
| aquarius | I know why I think I'm going :) | 13:53 |
| aquarius | dutchie, oh, and consider your honk received; I'm working on it :) | 13:54 |
| Chipaca | aquarius: do I have to honk? I'm currently giving my money to 'legalsounds.com', which I'm sure is a 100% legit business | 13:55 |
| CardinalFang | Chipaca, I see the agenda. I don't know if that's what it's *about*. | 13:55 |
| aquarius | chipaca, well, if we don't convert you I shall be somewhat disappointed :) | 13:55 |
| Chipaca | aquarius: :) | 13:56 |
| Chipaca | CardinalFang: you know the 6-month release cycle? it's all a big lie. All the real coding happens in the platform sprint. | 13:56 |
| CardinalFang | Ah, good. | 13:56 |
| CardinalFang | Five days. No drinking, I hope. | 13:57 |
| CardinalFang | aquarius, I saw a python-keyring module this morning. I think I may be able to make desktopcouch support KDE, OSX, and step closer to supporting Windows. | 13:59 |
| aquarius | CardinalFang, aha, there is news on that front | 13:59 |
| aquarius | CardinalFang, ralsina has ported desktopcouch to Archlinux, and in a private copy has moved to python-keyring | 13:59 |
| CardinalFang | Sweet. | 13:59 |
| aquarius | and homeasvs_ was also looking at python-keyring for the n900 | 13:59 |
| CardinalFang | Is that the pypi/keyring/ ? | 14:00 |
| aquarius | http://lateral.netmanagers.com.ar/ | 14:00 |
| aquarius | specifically http://lateral.netmanagers.com.ar/weblog/posts/BB855.html#disqus_thread | 14:01 |
| Chipaca | also, http://lateral.netmanagers.com.ar/weblog/posts/BB858.html | 14:02 |
| aquarius | ya | 14:02 |
| aquarius | and stick http://twitter.com/ralsina/status/6890319668 and http://twitter.com/ralsina/status/6899741456 into http://tweetconvo.com | 14:03 |
| aquarius | CardinalFang, sorry I hadn't updated you on that yet, yesterday was a bit frantic | 14:03 |
| === facundobatista is now known as facu | ||
| === facu is now known as facundobatista | ||
| === Chipaca is now known as Chipakeitor | ||
| === facundobatista is now known as facu | ||
| === Chipakeitor is now known as __chip__ | ||
| === __chip__ is now known as Chipaca | ||
| jblount | Random thought: Has anyone else here looked or played with Juno? http://github.com/breily/juno | 14:26 |
| dobey | i take it that's not a free dial-up e-mail service | 14:27 |
| CardinalFang | statik, https://code.edge.launchpad.net/~cmiller/ubuntuone-servers/monitor | 14:31 |
| * Chipaca looks | 14:32 | |
| CardinalFang | Very old. Not tested in any way. ^ | 14:33 |
| pfibiger | jblount: i have looked at it. yes. | 14:35 |
| statik | hey Chipaca, joshuahoover1: have you guys already done a main inclusion report for python-keyring? If not, could you? It's been synced to universe now, but I don't want to surprise the platform team by trying to pull it into main for ubuntuone-client at the last minute. | 14:43 |
| Chipaca | I haven't. If joshuahoover1 hasn't either, I'd like to learn how :) | 14:44 |
| joshuahoover1 | statik: i have not, i can put together the mir and i'll email you and chipaca once it's there later today so you two can review | 14:44 |
| statik | thanks | 14:44 |
| joshuahoover1 | Chipaca: it's really filling out a wiki template :) | 14:44 |
| statik | and doing a code audit | 14:45 |
| Chipaca | joshuahoover1: if that's all it is, go ahead. If you want help with the code audit, holler :) | 14:46 |
| joshuahoover1 | Chipaca: sounds good, thanks! | 14:46 |
| dobey | brb, washer/dryer delivery... | 14:49 |
| bl8 | Does anyone know why I could get this error : ubuntuone.SyncDaemon.ActionQueue - ERROR - Capabilities query failed: The server doesn't have the requested capabilities | 14:52 |
| bl8 | I'm trying to run ubuntuone-client-1.0.2 on my Gentoo desktop | 14:52 |
| jblount | bl8: Hi! | 14:54 |
| jblount | bl8: I don't think anyone here has any experience with the Gentoo stuff, but that sounds like a error that I've seen before. | 14:55 |
| bl8 | jblount: Hi. Weird thing is that it was working a few days ago, then I re-installed all the ubuntu one related stuff, hoping to write proper installations instructions, now it fails :( | 14:57 |
| jblount | bl8: Yeah, I was just talking to pfibiger and that error is supposed to protect differences between the client and server software. | 14:57 |
| jblount | It's strange that you're seeing it now, after having it working before though. | 14:59 |
| jblount | rmcbride: Any thoughts about problems with ubuntuone-client-1.0.2 ^^ ? | 14:59 |
| jblount | bl8: dobey may also be helpful to figure out what's going on. | 15:00 |
| CardinalFang | It would be nice to know what "capability" it refers to. | 15:05 |
| jblount | CardinalFang: Good point. | 15:06 |
| CardinalFang | "The server doesn't have the requested capabilities" -> "The server doesn't isn't capable of ____________________". Silly programmers. | 15:06 |
| CardinalFang | Er, one verb. | 15:07 |
| Chipaca | CardinalFang: file a bug! :) | 15:08 |
| Chipaca | CardinalFang: although I'm not sure the server tells us what it is it doesn't do | 15:08 |
| urbanape | "Does anybody really know what time it is? Does anybody really care?" | 15:09 |
| urbanape | Might have to dig up some Chicago this morning. | 15:10 |
| jblount | heh | 15:10 |
| jblount | urbanape: You pinged yesterday? Are you still in need with my particular version of crazy? | 15:10 |
| CardinalFang | rodrigo_, while I'm thinking about usability, the evolution d-c plugin should set the application name so that the keyring dialog will be nice. | 15:10 |
| urbanape | jblount: yeah, was gonna ping you about bug # 498284 | 15:11 |
| rmcbride | jblount: sorry, sound is off, missed that. I don't know anything about how the client would be installed on Gentoo I'm afraid. | 15:11 |
| rodrigo_ | CardinalFang: hmm, evolution already does it, afaik, or doesn't it? | 15:12 |
| jblount | rmcbride: I'm more interested in if ubuntuone-client-1.0.2 should work or if it should have the capability mismatch errors | 15:12 |
| rodrigo_ | no standup today? :) | 15:13 |
| rmcbride | the current client shouldn't get a capabilities mismatch, but Gentoo isn't a platform I've used for many years, so I don't know what might happen | 15:13 |
| CardinalFang | rodrigo_, I'm judging by the error log in bug#475998. """WARNING **: g_set_application_name not set in server_log_handler""" | 15:13 |
| verterok | bl8: please pastebin the contents of path/to/dist-packages/ubuntuone/syncdaemon/__init__.py | 15:13 |
| bl8 | jblount: rmcbride: I'm installing from the source tarballs | 15:13 |
| jblount | urbanape: # 498284 is in my queue for today | 15:13 |
| urbanape | you rock | 15:14 |
| jblount | rodrigo_: I don't know, I guess we should still have it. How about now? | 15:14 |
| jblount | Desktop+ MEETING BEGINS | 15:14 |
| rodrigo_ | CardinalFang: hmm, might be, yeah | 15:14 |
| jblount | You are the crew, you know what to do: | 15:14 |
| rodrigo_ | me | 15:14 |
| jblount | me | 15:14 |
| urbanape | me | 15:14 |
| CardinalFang | me | 15:14 |
| CardinalFang | Zing. We're quick. | 15:14 |
| * jblount assumes that anyone not "me" ing is on holiday or will catch up | 15:15 | |
| jblount | rodrigo_: Go for it! | 15:15 |
| rodrigo_ | • DONE: Python bindings for libubuntuone. Tested fake music store from aquarius | 15:15 |
| 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? Look at Canola. Add envvar to override music store default location | 15:15 |
| rodrigo_ | • BLOCKED: no | 15:15 |
| rodrigo_ | jblount: go go! | 15:15 |
| jblount | DONE: Got changes for /account-assistance sorted out, got sick, got better | 15:16 |
| jblount | TODO: Finish up some js for /account-assistance/, make a overlay thingie for the forums, land an approved branch post pqm weirdness | 15:16 |
| jblount | BLOCKED: Nope | 15:16 |
| jblount | urbanape: tag | 15:16 |
| urbanape | DONE: Some more ahas about Bindwood. _id handling improved. | 15:16 |
| urbanape | TODO: Brute force the manifest checking on pull to establish new order/location of bookmarks. On-call reviewer today. | 15:16 |
| urbanape | BLOCK: None | 15:16 |
| urbanape | CardinalFang: the comfy CHAIR! | 15:16 |
| aquarius | me | 15:16 |
| * rtgz is late | 15:17 | |
| CardinalFang | DONE: Got d-c stable in my PPA and proposed for Karmic. Started looking up env-agnostic keyring. | 15:17 |
| CardinalFang | TODO: Face duty. Help pfibiger with yesterday's auth madness in ubunet. Assist with small bugs. | 15:17 |
| CardinalFang | BLOCKED: None | 15:17 |
| CardinalFang | aquarius: No one expects it! | 15:17 |
| aquarius | ⚀ DONE: built a fake music store; built a python RB plugin; synced up with rodrigo; investigated why tomboy sync doesn't work; cried | 15:17 |
| aquarius | TODO: more music store; work out why my notes have been eaten (irony!) | 15:18 |
| aquarius | blocked: none | 15:18 |
| aquarius | rtgz, go for it | 15:18 |
| rtgz | DONE: (null) | 15:18 |
| rtgz | TODO: Update ubuntuone-client-diagnose with the latest bug info. Find more bugs. Test UI on Android and Maemo platforms. Create more feature request tickets. Hack more for "Shared with Me" folder emblem support. | 15:18 |
| rtgz | BLOCK: New Year shopping | 15:18 |
| rodrigo_ | aquarius: oh, tomboy sync doesn't work? | 15:19 |
| * rodrigo_ tries | 15:19 | |
| aquarius | rodrigo_, ya. Will fill you in in a moment | 15:19 |
| rtgz | no idea who is next, since ubuntulog will update the log in an hour | 15:19 |
| aquarius | do not try! | 15:19 |
| rodrigo_ | too late, I already tried, and it failed, yeah | 15:19 |
| aquarius | good, if it failed you're OK | 15:19 |
| rodrigo_ | aquarius: so, what's happening? | 15:19 |
| * rtgz is trying to set up lucid lynx in UML | 15:20 | |
| aquarius | rodrigo_, right. Yesterday we did a rollout. After that, our snowy API kept getting 409 Conflict when storing things in Couch, and then dying with a 500 server error, which is what causes tomboy to not sync | 15:20 |
| aquarius | rodrigo_, re-authorising tomboy stops you getting the error | 15:20 |
| rodrigo_ | hmm | 15:20 |
| rodrigo_ | aquarius: ok | 15:20 |
| aquarius | but I did that, and then told it to overwrite my local notes with ones from the server, because I'm an idiot | 15:21 |
| pfibiger | rodrigo_: chad is looking at the revisions to see if he can figure out what caused that behavior. it was some change in ubuntuone-server trunk between revisions 2064 and 2098 | 15:21 |
| rodrigo_ | aquarius: when you re-authorize, best thing is to remove all notes from the server | 15:21 |
| aquarius | rodrigo_, I know. Now. :( | 15:21 |
| rodrigo_ | pfibiger: ah, ok | 15:21 |
| mandel | aquarius: ping | 15:23 |
| CardinalFang | aquarius, beware the fake music store doesn't become The Real music store. We may have you wire up the back end to your GrooveShark account. | 15:23 |
| aquarius | pong | 15:23 |
| aquarius | CardinalFang, nya haa! | 15:23 |
| aquarius | CardinalFang, part of the master plan :) | 15:23 |
| mandel | aquarius: mind if I give it a go to port desktopcouch to windows? | 15:23 |
| CardinalFang | Rock! | 15:23 |
| aquarius | mandel, yeah, that's a rubbish idea, don't do that, we hate Windows and everything it stands for, and if you do it I'll shout at you | 15:24 |
| aquarius | :-) | 15:24 |
| aquarius | That'd be fantastic, mate. Anything at all I can do to help, let me know :) | 15:24 |
| CardinalFang | mandel, we were talking earlier about an fait accompli effort to replace gnomekeyring. | 15:24 |
| aquarius | I've done *some* Windows hacking, in the dim and distant past | 15:24 |
| aquarius | the installer will be the hard thing, I think | 15:24 |
| mandel | aquarius: he, I have to do windows hacking everyday, so no problem | 15:25 |
| CardinalFang | Port dpkg first. | 15:25 |
| aquarius | "run it in wubi" is not a port :) | 15:25 |
| mandel | hehe | 15:25 |
| CardinalFang | mandel, I owe you a review of contacts-wrapper. | 15:25 |
| mandel | no I already have a number of ideas to allow the interprocess communication with the process, kind DBUS, the issue is the keyring | 15:26 |
| mandel | CardinalFang, no problem, dont worry to much about the review, I know what it takes you so long, you... you hate mey ODM (Object document mapper) hehe | 15:26 |
| aquarius | mandel, python-keyring is the answer for the keyring support | 15:27 |
| rodrigo_ | brb | 15:27 |
| aquarius | mandel, ralsina (on twitter) has a private codebase that uses it, and homeasvs_ has been looking at it too | 15:27 |
| aquarius | mandel, if you get together with them they may have alreayd done th ework, and landing a branch that does that first will get pretty quick approval from us :) | 15:27 |
| rtgz | Question - why does couchdb contacts do not match with vCard/xCard format? The guys were talking about vCards in U1 on the famous podcast two weeks ago and it looks like dc format is somehow evolution-oriented. | 15:28 |
| mandel | aquarius, sweet, so there is a python-keyring one working, 'cause that is just one of the tqo problems | 15:28 |
| aquarius | I believe so, yes | 15:28 |
| aquarius | the d-bus magic will also have to be changed to a Windows-native way of doing it | 15:29 |
| aquarius | Xalior, ping? | 15:29 |
| aquarius | Xalior, you're thinking about a desktopcouch port to OS X? | 15:29 |
| mandel | yes, maybe using DDE on windows | 15:29 |
| aquarius | dde? party like it's 1999 :) | 15:31 |
| aquarius | isn't DDE deprecated these days? Raymond Chen is always a bit mean about it | 15:31 |
| mandel | DDE is supported out of the box in the python version for windows, it kinda old but could do the trick, the other is usedin RPC which is complient with OSF DCE which means it " remotly might" be multiplatform | 15:31 |
| aquarius | doesn't have to be multiplatform, it just has to be windows :0 | 15:31 |
| rodrigo_ | DDE still exists? | 15:32 |
| aquarius | unless "multiplatform" means "windows vista *and* XP" | 15:32 |
| aquarius | rodrigo_, hell yes. (a) MS never break backwards compatibility, and (b) Word still uses it, for a start :) | 15:32 |
| mandel | aquarius, hehe yes DDE is old, but it means no dependencies to other modules for python on windows, and installing python mocules on a win machine is not that easy | 15:32 |
| aquarius | mandel, agreed | 15:32 |
| aquarius | mandel, but...I don't know if you can call a function over DDE and get an answer back, or whether you can only use DDE for starting an app? | 15:33 |
| aquarius | mandel, totally agree about not installing other python modules | 15:33 |
| rodrigo_ | it's for RPC iirc, so yeah, you can call a method and get an answer | 15:33 |
| mandel | CardinalFang, getting back to the wrapper, reject it, I think it is too close to OO for a core lib like desktopcouch, maybe it has place in an other project | 15:33 |
| mandel | old school DDE allows to exchange data, and RPC allows to make calls to functions | 15:35 |
| aquarius | then score. | 15:35 |
| aquarius | DDE sounds like the way; it's certainly the (well, a) native way to do it | 15:35 |
| aquarius | and I think it's really important that desktopcouch works in the native way on each platform, and doesn't try and build a little mini Linux environment on other platforms | 15:35 |
| mandel | ok, then I speak with ralsina and homeasvs_ about keyring, and then move to get the IPC working | 15:37 |
| mandel | well, I'll start from tom (holidays whowhoww) | 15:37 |
| CardinalFang | mandel, I was thinking of a contrib directory or a new project python-dekstopcouch-orm{,-contacts} . | 15:37 |
| mandel | I was thinking to about it python-dekstopcouch-orm sounds great, 'cause I've already made changes to make it less focus on contacts | 15:38 |
| aquarius | mandel, the reason I also pointed at Xalior is that he's thinking about a desktopcouch port to OS X, and it would be good (if he has time to work on it) if the two of you co-ordinated your efforts to factor out platform-specific stuff int he same way | 15:38 |
| mandel | CardinalFang, the mapping is looking like FluentNHibernate for mapping config | 15:39 |
| aquarius | mandel, so you and he don't factor out platform-specific stuff ina different way :) | 15:39 |
| mandel | aquarius, sure, I understand, we should do that | 15:39 |
| mandel | I'll get in touch with them | 15:39 |
| dobey | hmm | 15:49 |
| CardinalFang | mandel, I was thinking about some metaclass hackery that would read some schema description. Maybe this is an ORM version-2 idea, though. | 15:50 |
| mandel | CardinalFang, probably... I'll finish with what I have and will ask you then, I'm sure you can teach me something new about python :D | 15:51 |
| mandel | nevertheless, reject the merge it fill better in its own project :D | 15:52 |
| CardinalFang | Rgr. | 15:52 |
| CardinalFang | mandel, FWIW, I liked everything except the "import *" and "__all__ = ( big long tuple )" . | 15:54 |
| rodrigo_ | statik: any reason to have libubuntuone project closed? mandel can't access it | 15:54 |
| statik | rodrigo_, no it should not be closed. i think it was an incorrect default thats old. I'll get it changed | 15:56 |
| mandel | CardinalFang, is there a nicer way to do it? | 15:57 |
| mandel | statik, very appreciated | 15:57 |
| statik | mandel: no problem, publishing the code is like the very least we can do. thanks for all your work on this project! | 16:00 |
| CardinalFang | mandel, Er, maybe. I'm not sure. You have a module/namespace that has all that info. I would have to test, but *perhaps* "import foo" "__all__ = foo". This is more if-I-had-designed-Python than I-have-done-this-before advice. | 16:00 |
| mandel | statik, thanks | 16:01 |
| mandel | CardinalFang, I know.. is not nice, but after doing the more general orm it would be nicer, I hope | 16:01 |
| statik | mandel, all the branches should be public now | 16:01 |
| mandel | :D | 16:02 |
| bl8 | jblount and others: Wiped all Ubuntu One stuff, installed 1.1.0, and now everything seems to work quite fine. Woohoo ! | 16:06 |
| jblount | bl8: Great! | 16:11 |
| sandy|lurk | rodrigo_: any news on the new converter, and switching U1 to use the XML from the REST API as the canonical format? | 16:13 |
| statik | we've got to use canonical formats | 16:16 |
| statik | anything else would be outrageous | 16:16 |
| sandy|lurk | :-P | 16:16 |
| rtgz | dobey, I guess simple ~/.local/ubuntuone/shares <-> ~/Ubuntu One/Shared With Me/ will do for now, do you have any objections if I start such branch? | 16:18 |
| dobey | rtgz: you mean to make emblems work correctly for shared with me folders? | 16:20 |
| rtgz | dobey, yup | 16:20 |
| dobey | rtgz: not really sure at the moment. it might be better to fix nautilus to do the right thing with symlinks | 16:21 |
| rtgz | dobey, we can get the real path for the file name passed to us. | 16:22 |
| dobey | rtgz: if we maintain a tree with references to NautilusFileInfos, right? | 16:23 |
| rtgz | rtgz, not exactly via nautilus routines, but via canonicalize_file_name call. | 16:23 |
| dobey | which is? libc? | 16:24 |
| rtgz | dobey, however, this will be only valid for Shared With Me folder case, not for arbitrary symlink | 16:24 |
| rtgz | dobey, yep, gnu libc | 16:24 |
| dobey | rtgz: i'd rather not use that method then. if it were some API in glib maybe, as it should handle win32 and such correctly. | 16:26 |
| Xalior | aquarius / mandel: Yeah, I'm gonna look at DesktopCouch over the holidays. There's a few bits that need tweakery. Like MacOS really doesn't have any "on demand service loading", etc. | 16:50 |
| Xalior | (start at login, start at startup, start by hand, don't start) | 16:50 |
| aquarius | I thought launchd did that | 16:51 |
| aquarius | ? | 16:51 |
| aquarius | obviously not :) | 16:52 |
| mandel | Xalior, I need to look at that in windows too, let me know which generic changes you need for that... we might as well create a branch for us (I think I can find a mac os somewhere ofr testing) | 16:52 |
| Xalior | mandel: minimum of 2G ram, Intel CPU. Since I'll be targeting snowleopard, and that's minimum specs :) | 16:53 |
| mandel | Xalior, dammed, you mac people. hehe, I'll do my best, so, should we create a branch? I already got a look at this: http://pastebin.ca/1723816 from ralsina is a good start | 16:54 |
| dobey | i think dbus is already working on osx | 16:55 |
| Xalior | A STUPID dependancy. And will be removed. | 16:55 |
| Xalior | dbus is (a) not part of standard macos; (b) overkill for a single application to use and (c) a right royal pain in the arse to setup. | 16:56 |
| Xalior | If you put a dbus dependancy on macos you've removed the ability for most MacOS users to install it, straight out the box. | 16:56 |
| Xalior | ie: last time I looked it required darwinports, which is a bunch of patchfiles and requires you to have xcode/gcc installed :) | 16:57 |
| aquarius | nonono, no d-bus on OS X. | 16:58 |
| Xalior | :-D | 16:58 |
| aquarius | that'd be like requiring someone to install launchd on Ubuntu; it's possible, but stupid. | 16:58 |
| aquarius | use the native platform mechanism. | 16:58 |
| Xalior | my point exactly | 16:59 |
| aquarius | mandel, I should note that ralsina's 3-line port to python-keyring has minor fail, in that he hasn't patched desktopcouch.records :) | 16:59 |
| mandel | aquarius, I know.. but we have a starting point | 17:00 |
| aquarius | *nod* yes indeed | 17:00 |
| mandel | about dbus, same on win platforms wethat is why aquarius mentioned that we should work together, try and make changes as generic as possible | 17:01 |
| mandel | Xalior, how can you do IPC in Mac OS X? | 17:02 |
| aquarius | Xalior, yeah; since there are platform-specific bits ("how do I get started if I'm not running", "where do I store the keys", "where do I store my files"), I'd like to see those factored out in some kind of sane way; i.e., no "if sys.platform=='mac'" all over the place :) | 17:02 |
| aquarius | and factoring those bits out should be cool if you two work together; that way, we've got the "platform-specific back end" stuff all nicely in separate modules, and someone who wants to port desktopcouch to another platform just implements that backend | 17:03 |
| Xalior | mandel: start here: http://www.puredarwin.org/curious/ipc#TOC-Forms-of-IPC :) | 17:03 |
| Xalior | There's some similar bits to dbus in Core-Foundation, used by notifyd. | 17:04 |
| Xalior | but that's just a wrapper around Mach-IPC | 17:04 |
| Xalior | http://www.chromium.org/developers/design-documents/os-x-interprocess-communication makes for some very interesting reading, actually. They're WAAAAAY more up to date than my last IPCpokery, 18mo ago, when I was hacking stuff the museum touchscreens :) | 17:07 |
| mandel | Xalior, can u use RPC? since DBUS is not use too much we might both be able to use it | 17:09 |
| Xalior | mandel: apple provide an rpc.statd and an rpc.lockd | 17:11 |
| mandel | Xalior, nah I dont think we can use that at all | 17:12 |
| mandel | I'll look at what can be done on the windows size and will let you know :D | 17:13 |
| Xalior | My interest in windows is truely less than none. :) | 17:13 |
| Xalior | And having to contort to windows methods will not so much hinder my port, but stop it dead in its tracks :) | 17:13 |
| Xalior | I'm happy to work with someone else who is working on one :) | 17:14 |
| Xalior | But if I can do it in macos in 2 lines of code, and it takes windows 22 lines, I'll be writing two and leaving the other twenty to someone else :) | 17:14 |
| Xalior | and, with that in mind, I completely and openly state that I'm doing a port for "my own needs", and if no-one else wants it, I'll cry, but I'll live :D | 17:14 |
| Xalior | mandel: eg: My first port of call would be http://en.wikipedia.org/wiki/Apple_events | 17:18 |
| mandel | Xalior, for what I can understand you can an even manager, we might be able to do something with that, use the idea to wrap how the communication is done and use a common interface | 17:21 |
| Xalior | yep. abstract it to a send message/get message call, that is implimented for each platform. | 17:21 |
| mandel | Xalior, I can accommodate my code to the APi you use. | 17:22 |
| aquarius | I wouldn't abstract down that far if I were you | 17:22 |
| aquarius | you don't need a generic sendmessage | 17:22 |
| mandel | aquarius, indeed, I more precise api would be better | 17:23 |
| mandel | but we can abstract that much and then let the platform code work, should be a good starting point | 17:23 |
| Xalior | aquarius: how much I actually would abstract to a platform generic API would, quite literally depend how much something is used, eg: one call can be if/else/etc - if it's something that's hammered a lot, then wrap it. | 17:23 |
| aquarius | desktopcouch only sends/responds to a few messages: "start running if you are not already running", "what is your port number", "are you still running", "what are the oauth keys". Implement those as functions, and then back-end them however's most appropriate for the platform. I advise against trying to build a cross-platform messaging abtraction :) | 17:23 |
| * Xalior orders a taxi to take him home. Taxi, in snow. Let the angry Glasgwegian swearing from someone who thinks that he gets to complain AND get paid at the same time begin. | 17:28 | |
| sandy|lurk | rodrigo_: this might also be of interest to you: https://bugzilla.gnome.org/show_bug.cgi?id=604953 | 17:30 |
| ubottu | Gnome bug 604953 in General "Tomboy XML is whitespace dependant" [Normal,Unconfirmed] | 17:30 |
| rodrigo_ | sandy|lurk: oh yeah, related to some of the bugs we have, so adding me to the CC | 17:31 |
| sandy|lurk | a user doing some 2-way wiki conversion noticed it the other day | 17:31 |
| rodrigo_ | I guess I should trim the strings when converting from HTML to XML | 17:33 |
| rodrigo_ | or fix it in tomboy, how hard would it be? | 17:33 |
| aquarius | ooh cool http://bl-log.blogspot.com/2009/12/ubuntu-one-on-gentoo.html | 17:34 |
| aquarius | that should be linked from somewhere. Do we have a "how to install Ubuntu One on other distros" page? I've got one for desktopcouch | 17:34 |
| sandy|lurk | rodrigo_: another stable Tomboy release isn't due for months, so the answer is probably "fix both" :-) | 17:35 |
| rodrigo_ | sandy|lurk: ok, I'll add the trimming thing then for now | 17:36 |
| rodrigo_ | sandy|lurk: we can remove it later | 17:36 |
| sandy|lurk | I haven't looked into what Tomboy's actually doing here (hence the UNCONFIRMED state), but since we support new lines within list items, it might be a bit tricky | 17:36 |
| rodrigo_ | yeah, shift+enter, I just know about it | 17:38 |
| rodrigo_ | now I don't have to put all in 1 line!! :D | 17:38 |
| rodrigo_ | aquarius: in the tutorials page maybe? | 17:39 |
| sandy|lurk | does the file sharing part of U1 require desktop-couch? | 17:41 |
| sandy|lurk | I haven't tried installing any of this stuff yet | 17:41 |
| sandy|lurk | (openSUSE) | 17:41 |
| metropolis | Is there a "right" way to talk to my desktop couch instance, given that it apparently has a randomly-generated username/password combo? | 17:46 |
| rodrigo_ | sandy|lurk: no, the file sharing doesn't use desktopcouch | 17:48 |
| sandy|lurk | seems like it should be easy to build, then | 17:48 |
| bl8 | aquarius: you're welcome ;) | 17:48 |
| aquarius | bl8, just trying to work out where to link your mastery from :) | 17:48 |
| bl8 | Err, "mastery" might be too strong... | 17:49 |
| sandy|lurk | I should install portage on openSUSE | 17:50 |
| aquarius | bl8, hey, if it works, I'm impressed. I'm nowhere near capable of writing ebuilds :) | 17:51 |
| bl8 | sandy|lurk: Eh eh, just switch to Gentoo ;) | 17:51 |
| sandy|lurk | bl8: this weekend I'll be decommissioning my last gentoo box | 17:51 |
| rodrigo_ | sandy|lurk: homeasvs_ has been doing fedora packages for it, so maybe you can get his packages and try to make them run on suse | 17:52 |
| sandy|lurk | yeah, I think I saw something about that...will google a bit | 17:52 |
| bl8 | aquarius: Bah, it's just shell scripts :) | 17:56 |
| aquarius | bl8, homeasvs_, https://wiki.ubuntu.com/UbuntuOne/Ports | 17:59 |
| mandel | can anyone take a look at this: http://pastebin.ca/1723951 is what I've done so far to allow map a python object to a desktopcouch record in a generic way | 18:02 |
| Chipaca | mandel: instead of "def _get_full_name ... full_name = property(_get_full_name)", you could use @property | 18:16 |
| mandel | Chipaca, I know (ups). I was trying to show an example where you can map a readonly property, I should have use the @ for the example... The idea is that you can map properties or fields to set the fields in the record, the mapper will do that for you so you just have to care of defining how to place the object in a record | 18:18 |
| Chipaca | mandel: where does ClassMapper come from? | 18:19 |
| mandel | I provide that, would be the abstract class that a mapper should extend I added the code here: http://pastebin.ca/1723976 | 18:23 |
| mandel | Chibata, I forgot one line in the copy past, the read record returns the created instance | 18:25 |
| Chipaca | mandel: one thing I'd do differently is change ClassMapper.init to take *args and **kwargs that are passed into the callable | 18:26 |
| Chipaca | mandel: that would make the api imho nicer, because you'd do | 18:26 |
| Chipaca | self.init(Contact, -1, "", "", "") | 18:26 |
| Chipaca | instead of the lambda | 18:26 |
| mandel | Chipaca, brilliant idea! | 18:27 |
| Chipaca | mandel: however, I don't know if I'm answering the question you're asking, at all | 18:27 |
| mandel | Chipaca, you are, I'm just looking to improve the api, that way people can used desktopcouch by just providing the mappings, no need to add extra code | 18:28 |
| sandy|lurk | So I just installed the 1.1.0 tarballs and starting the applet I get this: http://paste2.org/p/574036 | 18:29 |
| sandy|lurk | and nothing else happens. any ideas? | 18:29 |
| sandy|lurk | not sure if I was supposed to install another tarball for the storage daemon? | 18:30 |
| bl8 | sandy|lurk: The sync daemon might be crashing because of missing deps | 18:42 |
| Chipaca | sandy|lurk: try running the syncdaemon manually; what happens? | 18:43 |
| Chipaca | mandel: would all mapper classes be like that? an __init__ with a mapping declaration spread out across a few method calls? | 18:44 |
| sandy|lurk | Chipaca: how do I run the sync daemon manually? | 18:45 |
| sandy|lurk | I don't know the process name | 18:45 |
| sandy|lurk | command name, whatever | 18:45 |
| * sandy|lurk looks for a .service file | 18:45 | |
| Chipaca | sandy|lurk: /usr/lib/ubuntuone-client/ubuntuone-syncdaemon | 18:46 |
| sandy|lurk | ah, libexec | 18:46 |
| Chipaca | sandy|lurk: or that was what it was here :) | 18:46 |
| sandy|lurk | the tarball installs to $prefix/libexec/ubuntuone-syncdaemon | 18:46 |
| Chipaca | oh, ok | 18:46 |
| sandy|lurk | bl8: you're right, thanks | 18:46 |
| Chipaca | sandy|lurk: start it with --debug, for extra win | 18:46 |
| dobey | hmm | 18:47 |
| mandel | Chipaca, there can be a single map code that takes a dict, that maps everything at once, plus init and id | 18:47 |
| mandel | Chipaca, such as self.map({"field":"attr","field2":prop}) | 18:48 |
| dobey | Xalior, aquarius: is there some API like python-keyring, but for IPC methods instead? or do we have to do all that abstraction ourselves? and how will that play with the various bits we use it for that aren't internal communication (like network status)? | 18:49 |
| sandy|lurk | great, seem to have gotten through deps. now I'm running into https://bugs.launchpad.net/ubuntuone-client/+bug/487333 , glad I'm not the first :-) | 18:54 |
| ubottu | Ubuntu bug 487333 in ubuntuone-client "oauth traceback when token is None" [Undecided,New] | 18:54 |
| dobey | sandy|lurk: what version are you using? | 18:59 |
| sandy|lurk | dobey: 1.1.0 | 18:59 |
| rtgz_ | i think I know the reason why N800 has problems with the web site... ;-) | 18:59 |
| dobey | sandy|lurk: you got that same error dialog? | 18:59 |
| sandy|lurk | dobey: I assume I just need to find the location of that .crt file | 18:59 |
| dobey | sandy|lurk: you're on opensuse, right? | 18:59 |
| sandy|lurk | yes | 18:59 |
| sandy|lurk | the error dialog is useless...the python error is the same | 19:00 |
| sandy|lurk | so I figure it's the same bug | 19:00 |
| dobey | hrmm, i thought it used /etc/ssl/certs | 19:00 |
| sandy|lurk | well, but that same file is not there | 19:00 |
| dobey | sandy|lurk: useless in that it says something about SSL verification failing? | 19:00 |
| sandy|lurk | Authorization Error, and then a bunch of "Errno" stuff | 19:01 |
| sandy|lurk | [Errno socket error][Error 185090050...tines:X509_load_cert_crl_file:system lib | 19:01 |
| dobey | ok | 19:01 |
| sandy|lurk | I shouldn't say "useless" | 19:01 |
| dobey | so yeah, that missing file then | 19:01 |
| sandy|lurk | nod | 19:01 |
| sandy|lurk | /etc/ssl/certs has a bunch of *.pem files | 19:03 |
| sandy|lurk | site-specific certs | 19:03 |
| sandy|lurk | and files with weird names like f3cf1e8e.0 | 19:03 |
| dobey | right | 19:03 |
| dobey | yeah i don't know why that is | 19:04 |
| dobey | sandy|lurk: is there a ca-certificates package? | 19:04 |
| sandy|lurk | good thinking | 19:04 |
| sandy|lurk | openssl-certs | 19:05 |
| sandy|lurk | but it installs all of those *.pem files | 19:05 |
| dobey | and no ca-certificates.crt file? | 19:05 |
| sandy|lurk | nope | 19:05 |
| dobey | sandy|lurk: what version is it? | 19:07 |
| dobey | sandy|lurk: it looks like the ca-certificates.crt file gets generated on ubuntu by an update-ca-certificates command | 19:08 |
| sandy|lurk | 0.9.8h-28.2.1 | 19:08 |
| dobey | hmm | 19:08 |
| sandy|lurk | it's extracted from firefox, apparently | 19:09 |
| dobey | that sounds like it's straight from openssl | 19:09 |
| dobey | since it has that same version number | 19:09 |
| CardinalFang | pfibiger, where are my oopsen? | 19:09 |
| dobey | but maybe that's just suse putting the sources in the openssl source rpm | 19:09 |
| sandy|lurk | maybe | 19:09 |
| sandy|lurk | the pkg description says "This package contains some CA root certificates for OpenSSL extracted from MozillaFirefox" | 19:09 |
| dobey | sure | 19:19 |
| dobey | sandy|lurk: hrmm, i wonder what to do about that for opensuse then | 19:29 |
| sandy|lurk | dobey: I was just looking at the python docs | 19:29 |
| sandy|lurk | I guess a single file is required | 19:29 |
| sandy|lurk | so I should probably file a bug against the openSUSE package | 19:29 |
| rtgz | okay, N800 does not work with Web UI because GoDaddy Class 2 CA certificate is not installed there and it is not that visible to the end-users | 19:30 |
| dobey | sandy|lurk: i wonder if there's some way we can get all the distros to ship the ca certificates in the same manner and locations | 19:30 |
| dobey | rtgz: you mean, the ca-certificates.crt isn't there? | 19:30 |
| sandy|lurk | dobey: that would be nice, I think | 19:31 |
| rtgz | dobey, well... haven't found one yet, but I am still browsing the FS... | 19:31 |
| sandy|lurk | dobey: so, of course some apps bundle their own .crt...I tried pointing at one of those and got this error in the daemon when trying to connect from the applet menu item: http://paste2.org/p/574087 | 19:32 |
| dobey | sandy|lurk: i don't see any valid reason that they're all different. this really should be something in the LSB i think | 19:32 |
| sandy|lurk | I take it that Firefox should be opening, asking me to authorize my computer, but that has not happened | 19:32 |
| dobey | sandy|lurk: hrmm. weird | 19:32 |
| dobey | rtgz: dpkg -l|grep ca :) | 19:33 |
| urbanape | have to run out on a quick errand. bbiab. | 19:37 |
| rtgz | In order to make maemo work with u1 web server, certs for media.one.ubuntu.com, files.one.ubuntu.com and one.ubuntu.com must be added. But I can't find the CA package in the emulated device. Will borrow the real one (hopefully) next week, might have more info then. | 20:08 |
| dobey | later all! | 22:08 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!