=== lucio_ is now known as __lucio__ [04:35] Whats the difference from Ubuntu One and DropBox? [04:37] besides shipping default with ubuntu [04:40] FFForever: Ubuntu One, is a lot more than just file sync :) [04:40] like? [04:41] FFForever: contacts (evolution), notes (tomboy), and bookmarks (bindwood) sync [04:41] FFForever: and more stuff is comming down the pipe [04:42] I see, sadly none of it i use [04:42] FFForever: also desktopcouch, which is the backend of contacts, notes and bookmarks [04:43] cool [11:47] rodrigo_: hi. Is automatical desktopcouch and oaut handling implemented in couchdb-glib unstable http://www.freedesktop.org/wiki/Specifications/desktopcouch/Documentation/SimpleGuide#head-a3d1fb0683499a0ba3ec111fb8115b7296ba54fc ? [11:51] adiroiban, I think rodrigo's having some connection problems [11:51] yeah, seems so === rodrigo is now known as rodrigo_ [12:03] rodrigo_: what is the scope of couchdb-glib? do we realy need to support couchdb_document_contact_* ? In the future we are going to see similar function for each document type? [12:04] adiroiban: yes, for supported document types [12:05] adiroiban: think that the field names need to be consistent, so I don't want people to try to guess [12:05] also, we might change the field names, or their position in the record [12:06] „supported document types” are the one listed in Formats http://www.freedesktop.org/wiki/Specifications/desktopcouch/ ? [12:06] so far, yes [12:07] can I query a view using couchdb-glib? [12:08] not yet, I didn't add it because I really didn't need, but will add it soon [12:08] do you need it? [12:08] yes [12:08] ok, file a bug to couchdb-glib and I'll add it [12:08] I'm trying to migrate my toy app from sqlite to desktop couchdb [12:09] cool [12:10] brb [12:10] OAuth is automaticaly handled ? or do I need to use dbus and gnomekeyring like in evolution? [12:10] you need to do it like in evolution, but I am about to add a DesktopCouch object that handles it [12:10] so if you're going to write that code, please write the DesktopCouch object [12:11] i filled a bug [12:11] ok [12:11] and I was waiting to be accepted [12:11] and then try to fix it [12:11] I was not sure if it was valid or not [12:11] I just need a bit of time to work on several things in couchdb-glib, so any help would be appreciated [12:11] adiroiban, assign it to me please [12:11] ok, now really brb, need to buy some food :-) [12:24] is is feasible to create design document by storing them on the filesystem ? http://www.freedesktop.org/wiki/Specifications/desktopcouch/Documentation/DesignDocsFilesystem [12:24] the requirement of restrating couchdb looks like a showstopper [12:26] do you know of an application using this feature? I would like to know how it handles desktopcouchd restart and how it is packaged [12:28] adiroiban, it is feasible, but that code is quite new and not being used much yet. jml was working with that; he may know more. jml, ping [12:29] hello? [12:29] jml, you were using the design-docs-on-the-filesystem stuff, yes? [12:30] aquarius, adiroiban, yes. I've got a directory in my bzr branch that I've been manually symlinking into my desktop couch dir for local testing [12:30] aquarius, adiroiban, for unit testing, I've got a thing that sets it up automagically. I guess that could be used to do a deployment thing [12:30] I haven't thought at all about packaging yet, still early days [12:31] (you'd be surprised how little time I get to muck around with code) [12:31] jml: thanks. I will not use that feature for now [12:31] adiroiban, lp:~jml/+junk/lister has the code [12:32] adiroiban, I did it because it was easier than 'if db.hasView("foo")' (or however it's spelt) [12:33] jml: for my point of view, rather than storing the design document in XDG dirs, for a real app it would make sense to have them stored in a single place, somewhere in usr/share [12:34] jml: don't worry. thanks for the link. I will try to look at the code and see if I can use it [12:35] adiroiban, ok. [12:35] good luck. let me know if it helps, or you can see ways to improve it. [12:35] adiroiban, that's why it uses xdg dirs. There is a system-level xdg folder for apps too -- /etc/xdg/appname on Ubuntu, for example [12:35] so your package installs the design docs to /etc/xdg/appname/blahblahblah [12:36] aquarius: neat. I was not aware of that part :) [12:36] aquarius, upgrades are a million times more interesting than installation [12:36] adiroiban, the idea is that a package, when installed, will upgrade its design docs (in /etc/xdg), and then send a signal to restart all running desktopcouches, so they re-read the design docs [12:37] the second part of that is...not brilliantly implemented at the moment, but we're eager to get it working well now that people are using the filesystem design docs :) [12:37] part of the reason that it's not brilliant right now is that I wasn't sure how people would use it! [12:37] aquarius: well. I can see 2 use cases. deployment and development [12:38] rodrigo_, are you avoiding canonical irc? :) [12:38] adiroiban, yep -- design docs being in xdg folders means that dev is relatively easy because you store them in your local one for testing (or for overriding the system ones) [12:39] for deployment storing them in /etc/xdg and killall -HUP desktopcouch should do the job [12:39] aquarius, no, can't connect [12:39] rodrigo_, weird [12:39] rodrigo_, so, then, I'll ask here instead; how do I build the widgets? :) [12:39] for developement you can invoke a manual restart [12:40] aquarius: ./autogen.sh && make [12:40] adiroiban, well, desktopcouch doesn't currently re-read on a HUP signal; that's part of the way I planned to do it, sure, but I'm open to suggestions on this stuff [12:40] rodrigo_, I don't have to specify --prefix? [12:40] brb, hopefully with the connection problems solved [12:40] rodrigo_, I don't have to specify --prefix? [12:40] ah ) [12:40] :) [12:45] g'night folks. [12:47] night jml [12:48] aquarius: you can specify a --prefix if you want, but you can run the test programs without installing [12:49] ah, cool; I want to test importing it [12:49] aquarius, if you want to install, use the --prefix that you want [12:49] aquarius: ah, then you need to install yes [12:49] aquarius: just use a --prefix where you have a python installed [12:50] but I guess you have it in /usr only right? [12:50] oh, well, just install to /tmp and set PYTHONPATH [12:50] that's what I'm doing :) [12:50] what's the music store widget called? [12:50] ah, MusicStore [12:50] Nice [12:51] erm [12:51] PYTHONPATH=build/lib/python2.5/site-packages/ python -c "import ubuntuone.gtkwidgets, gtk; w=gtk.Window(); w.add(ubuntuone.gtkwidgets.MusicStore); w.show_all(); gtk.main()" [12:51] TypeError: GtkContainer.add() argument 1 must be gtk.Widget, not GObjectMeta [12:51] am I using it wrong? [12:52] it should *be* a widget, shouldn't it? [12:53] yes [12:54] ah, you're missing the () after MusicStore [12:54] ha! [12:54] I am very stupid :) [12:55] cheers. works now. nice! [12:55] approved :) [12:55] python -c "import ubuntuone.gtkwidgets, gtk; gtk.gdk.threads_init(); w=gtk.Window(); w.add(ubuntuone.gtkwidgets.MusicStore()); w.show_all(); gtk.main()" [12:56] it also complains about threads not being initialized [13:24] rodrigo_: can I use couchdb-glib with HTTP Basic Auth ? [13:25] adiroiban: it doesn't have code to do that, but it should be easy to, just need to add a HTTP header to the SoupMessage [13:26] couchdb-bugs are handled in Launchpad ? I could not find the gnome bugzilla module [13:26] yes, in LP [13:26] what app are you porting to it, btw? [13:27] https://launchpad.net/lucruri [13:27] trying to build couchdb-glib on Karmic i get this error http://paste.ubuntu.com/352912/ [13:30] adiroiban, can you paste line 127 of dbwatch.c? [13:35] adiroiban, you shouldn't basic auth to desktopcouch, though [13:35] aquarius: true. but since couchdb-glib is a generic lib. I think it should do basic auth [13:36] adiroiban, true enough, I suppose. I'd like to see desktopcouch-glib, then, as a separate library :) [13:36] adiroiban: try again with trunk, fixed it [13:38] rodrigo_: thanks. I'm using the bzr import. let's see if I can work around and get the latest git [13:38] adiroiban: I've got a patch to rename most of the objects/functions to couch instead of couchdb, for introspection to work [13:39] adiroiban: so if working with trunk, use git please [13:40] ok. np [13:43] adiroiban: ok, committing the patch now, so just rename CouchDB to Couchdb [13:48] hmm, it makes sense what aquarius says about desktopcouch-glib, we could have a separate library and have there all the dc code, like oauth/keyring and contacts/etc record types [13:57] rodrigo_: I agree, or have a ./configure flag to build desktopcouch support in couchdb-glib [13:58] well, I think we want to build it always, if the dependencies are available [13:58] (keyring and dbus) [14:00] * rodrigo_ -> lunch [14:00] bbl [14:20] rodrigo_: it looks like in couchdb-glib, make install does not copy all the required headers [14:21] I only got couchdb-glib.h couchdb-document-contact.h and couchdb-types.h [14:31] rodrigo_: also couchdb-glib-1.0.pc is missing Requires: json-glib-1.0 [14:31] should I fill a bug for those problems? [15:02] adiroiban, yeah, fixing it now, trunk has been a bit out of maintainership [15:03] rodrigo_: np. I'm working on HTTP Basic Auth on master [15:03] Desktop+ DESKTOP MEETING [15:03] Welcome team! We're doing a standup meeting, if you'd like to state your status please say "me" in order to grab a slice of time. [15:03] just as an exercise to get used with the code [15:03] me [15:04] adiroiban, cool [15:04] me [15:04] me [15:04] me [15:04] adiroiban, I'd suggest you add a _enable_http_auth function, as there is the _enable_oauth [15:04] adiroiban, or a set_authentication_whatever [15:05] adiroiban, but I'll let you have a look and suggest what you think is best [15:05] rodrigo_: ok. I will follow oauth names [15:05] me [15:05] DONE: _changes polling stalled a bit, trying to get some help from mozilla places devs [15:05] TODO: Finish up _changes polling, and dispatching behavior for various types (bookmarks, folders, separators, livemarks containers, dynamic containers) and event handlers. [15:05] BLOCK: None [15:05] rodrigo_: go go go [15:06] • DONE: Packaging of libubuntuone/RB plugin. Added selection API to contacts picker. Match aquarius music store prototype in C widget. Added libu1 dependencies to ubunet-developer-dependencies package. On-call review. Applied introspection patch to couchdb-glib master [15:06] • 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. Add libubuntuone dependencies to dev deps. Separate dc-specific code in [15:06] couchdb-glib [15:06] • BLOCKED: no [15:06] next jblount [15:06] me [15:07] DONE: Got ui bits done for public files [15:07] TODO: Make javascript API for rest API for public files, attach web ui, put up for review and do a dance [15:07] BLOCKED: Nope [15:07] aquarius: YOU! [15:07] ⚀ DONE: music store review; talk to pfibiger about daemon deployment; talk to rodrigo about libubuntuone packaging; work out chart API; talk to mentalguy about amqp [15:07] ⚁ TODO: spec file delivery downloader daemon; have music library page send message to downloader daemon; make workitems of outstanding todo items; make tomboy first-sync experience nicer [15:07] ⚂ BLOCKED: [15:07] CardinalFang, hit it [15:07] no MEETING BEGINS? [15:07] DONE: still boggling over Bug#499595. [15:07] TODO: more on that. [15:07] BLOCKED: None, except snowy is kind of silly. [15:08] teknico, MEETING BEGINS, then. Go! [15:08] DONE: more bug triaging, more configuring funambol for sending sms messages with vds (#418048), started setting up a development environment in a virtual machine [15:08] TODO: finish configuring funambol for sending sms messages (#418048), finish setting up a development environment in a virtual machine [15:08] BLOCK: none [15:08] next: ? [15:08] teknico: Sorry, I forgot :) [15:11] EOM (I think) [15:18] adiroiban: git pull, should be fixed now [15:22] hello. for some reason when i drop files into my Ubuntu One directory on my desktop, it says it is updating but the files do not show up when i go to the one.ubuntu.com site [15:27] rodrigo_: back to main chat. thanks for the update. [15:27] adiroiban, np, I should have done it some time ago [15:30] anyone have a suggestion of what I can look for as the problem? [15:33] * popey wonders if someone could look at bug 500975 and see if it's specific to ubuntuone or a tomboy generic sync issue [15:33] Launchpad bug 500975 in tomboy "Synchronization failed with the following exception: Unexpected character '<' at [1:1]" [Undecided,Confirmed] https://launchpad.net/bugs/500975 [15:34] i am getting the failure on both ubuntu and windows syncing to u1 [15:51] popey, can you sniff the network traffic to see what document that is? My guess is that it's an HTTP 4xx or 5xx, and an acompanying body that looks HTML-y. [15:51] what tool would you propose i sniff the traffic with? [15:51] If you're on Linux, wireshark. [15:51] or tcpdump [15:52] what am I looking for? [15:53] Well, that's a HTTP request, so filter out all but HTTP. Look for traffic to one.ubuntu.com . [15:53] Connect, ask for https://one.ubuntu.com/notes/api/1.0/op/ . Get back some response. The response is the interesting bit. [15:54] I'd need to figure out the tcpdump fu to do that [15:54] * popey googles [15:54] wireshark is much easier. [15:54] i am on 3g, installing wireshark will take $time [15:54] Ah. [15:57] popey, man pages would use no $bandwidth. :) [15:57] popey: $ sudo tcpdump -v -A -l host one.ubuntu.com |less [15:57] i have two machines, one running windows has network, machine running ubuntu has 3g.. well gprs :S [15:57] thanks! [15:59] popey, If your active network interface is one of many, you may have to specity which to use with "-i gsm0" or something. [15:59] yeah, using -i usb0 [15:59] getting syntax erros with -l though.. having a play [16:00] yay [16:01] the -A gives me some 'output' but it's not human readable [16:01] hrm. [16:02] Okay, Let's get more of the packet. [16:02] popey, $ sudo tcpdump -v -s 1000 -A -l host google.com |less [16:03] google.com you say? [16:03] :) [16:03] Er, no. o.u.c Sorry. [16:03] :) [16:05] i see no URLS [16:05] will play more [16:06] this is https so surely tcpdump isnt' going to see any data [16:07] maybe hostnames, certificate names and so on [16:08] Aw, I missed that. You're right. [16:08] Dang. I want to know what Tomboy is getting. [16:08] Set up a MITM proxy [16:08] :) Easier to get source and print it. [16:09] You might be surprised... [16:10] I am adept at "apt-get soruce tomboy; apt-get builddeps tomboy; $hack; dpkg-buildpackage -rfakeroot" [16:11] tomboy --debug may show me the note that is barfing [16:11] it shows the urls :) [16:12] oauth_nonce always makes me smile.. I'm such a child [16:12] It may be a bug that I'm working on. [16:13] I am getting 500's back from u1 [16:13] in the debug, so I'm not surprised that tomboy is barfing [16:13] Ah! Good. [16:13] Well, sort of good. You know. [16:13] ya [16:14] Okay, I presume it is my bug. (Not one I caused. One that I'm fixing.) [16:14] you put google analytics js in your 500 messages [16:14] thats not really necessary now is it? :) [16:14] We like statistics. Ha ha. [16:14] clearly! [16:15] what would be more useful is perhaps a tracking ID that I could give you and you could then find the error on your end.. like an OOps code [16:16] anything I can do? info you need? [16:16] The problem I'm working on is that Tomboy/Snowy storage/rtansactions doesn't map onto couchdb very well, and there's a hole that we fall into. [16:17] It won't be today that it's fixed for you. Perhaps tomorrow, if I'm lucky. [16:18] you're convinced that the bug I am hitting is the one you're working on? [16:18] There's secrets on those servers, so mundane mortals like developers can't go poke and prod. [16:18] I'm pretty sure. [16:18] do you have a bug number I can follow? [16:20] Hrm. Yeah. It's private for the next 10 seconds... https://bugs.edge.launchpad.net/ubuntuone-servers/+bug/499595 [16:20] Error: This bug is private [16:20] rodrigo_: ping [16:23] popey, I don't think I have the power to make this bug public, even though there's almost nothing secret in it. [16:23] can you subscribe me? [16:24] popey is my nick on lp [16:26] popey, in the mean time, you should file a bug on tomboy so that it demands an HTTP 200 before parsing anything, and for HTTP 4xx and 5xx, it displays the text. [16:41] hi Chipaca [16:41] man my inet connection today sucks [16:42] rodrigo_: can you update your blueprint? [16:45] the contacts picker one? [16:46] updated [16:56] is there a way to authenticate from the first time from command line? [16:57] X is required currently [16:58] dobey: ok, só how can i check why the applet never connects? [16:59] dobey: can you take a look? http://paste.ubuntu.com/353004/ [16:59] or someone else [17:00] do you have an ~/.config/ubuntuone/syncdaemon.conf file? [17:01] Hello! I am trying to use ubuntu one in a live cd, but I can't. I select "connect" in the applet (or in nautilus) but it does not even ask my password. I have never used ubuntu one before. my syncdaemon-exceptions.log: http://paste.ubuntu.com/353004/ [17:01] (Pretto is actually trying to help me in #ubuntu-br) [17:01] dark: i told them about this [17:02] ah ok:) [17:02] dark: check if you have ~/.config/ubuntuone/syncdaemon.conf file [17:02] cat: /home/ubuntu/.config/ubuntuone/syncdaemon.conf: No such file or directory [17:07] dobey: so, no file, we are trying to create it by hand [17:07] does oauth-login.log show anything? [17:08] Pretto: creating it won't help :) [17:08] but there was an issue with it causing network i/o to be blocked, which we've since fixed [17:08] dobey: nothing but Starting Ubuntu One client version 1.0.2 [17:09] ok [17:10] can you enable karmic-proposed in System->Settings->Software Sources on the Updates tab, reload the list, and install the new ubuntuone-client-gnome, then try again please? [17:10] yes i can [17:10] dobey: ok, hold on [17:10] dark: do that :p [17:11] dobey, i am on a live cd - is this a known issue? [17:13] dark: there are known issues with fixes waiting for review, in karmic-proposed, yes [17:14] dark: i'm not sure exactly which issue you're having, though [17:15] oh, great, i am also having problems with i/o (unrelated to ubuntuone) [17:15] [ 5167.189293] SQUASHFS error: squashfs_read_data failed to read block 0x4cf0f9c9 [17:15] i think i should just give up for now [17:15] hrmm [17:16] that could be an issue with the burn, or with your cdrom drive, perhaps [17:16] hm maybe it's ok now... (it's probably the motherboard, it's also failing randomly to detect sata) [17:16] or your mobo :) [17:17] i was also with a pendrive.. that died right now (in a power failure.. i don't quite understand) [17:19] it happens [17:19] i made a backup few hours before [17:20] is there a way to tell Ubuntu One to push changed files to the server side only periodically? [17:21] I tend to save a lot while I work on a document, and every time I save, UbuntuOne pushes to the server [17:21] disconnect, and only connect when you want to push the changes [17:21] heh, yeah I guess that would do it, wouldn't it ;) [17:22] The real problem is that, half the time, u1 tries to update a file when I open it in vim. [17:22] dobey, tell me: why not fuse? [17:22] Actually, not half the time. All the time. === nettrot_ is now known as nettrot [17:22] dark: it presented several problems which we couldn't solve or workaround, or we would have used it [17:22] I would like to see a periodic update implemented as a feature...I imagine I wouldn't always remember to reconnect the client when I want to push changes [17:22] dobey, there is a rationale somewhere? [17:23] it would be nice to have this problems actually fixed [17:23] but for now, connecting only when I want to push will suffice. thanks dobey [17:23] nettrot: what's that again? [17:23] even if the main implementors weren't willing to move to it again [17:23] dark: i don't know. [17:24] now i am curious :( [17:25] Chipaca: I haven't gotten around to filing this, but Ubuntu One tries to sync 'changed' files when I open the files in vim. Except that it's probably syncing the swap file.... [17:25] So really I'm probably just not thinking this through [17:26] nettrot: sync as in upload, right? [17:26] correc [17:26] dark: most of the problem was python-fuse, iirc [17:26] yes, vim does a weird three-file dance which perplexed us, but the behavior is now correct (sure, not really needed, but correct) [17:27] nettrot: at some point we might add in smarts so it realizes it's "just" vim doing that and ignore those files, but we have no app-specific rules yet [17:28] Which is fair. I wonder if syncing dotfiles should be optional though. [17:28] dobey, and you weren't willing to use other language? [17:29] nettrot: that's a good idea, I think. Could you create a wishlist bug for that so we don't forget it? [17:29] dark: *i* wasn't writing that part [17:29] Chipaca: Yeah, I'll go file that now. [17:29] nettrot: thanks! [17:29] dark: or i would have done it in C anyway :) [17:29] the command line interface seems to be simple enough to implement in C, and the few ui could just use it too [17:30] am [17:30] dobey, you are a canonical employee? [17:30] Chipaca: Which project should I file against? [17:30] yes [17:31] nettrot: ubuntuone-client [17:31] the ui and command line interfaces aren't the syncdaemon [17:31] the syncdaemon is where all the file sync logic is [17:31] dark: the syncdaemon would not have been an easy beast in C, at all [17:31] the nautilus extension is in C [17:31] the syncdaemon is written in python? [17:31] hmmm [17:31] everything but the nautilus extension is in python [17:31] Chipaca, it uses some kind of standard protocol? === jdaynes is now known as feutete [17:32] it uses ubuntuone-storage-protocol [17:32] which is built with protocol buffers [17:32] dobey, I would be happy with an optional nfs interface, really. Just like many people was happy when google gave them pop access [17:33] i don't even require cryptography! :p [17:33] i know its flaws, but it would be just... nice.. i hope it's not difficult to use both nfs and the ubuntuone-storage-protocol on the same data [17:34] .-. [17:34] i think rewriting in C using fuse would be easier than making an nfs implementation work :) [17:35] i guess you could nfs mount the Ubuntu One directory if you wantecd to [17:35] not sure if there are issues with it or not though [17:35] but can I do it remotely? from the ubuntuone servers? [17:36] my local Ubuntu One directory isn't very useful right now [17:36] do what remotely? mount it as ssuch on another machine? no [17:36] not through u1 anyway [17:37] oauth over nfs might be an issue [17:37] :) [17:38] ssh could be a good place to begin [17:39] i think if we were going to go that route, we'd just do webdav [17:39] which we have talked about possibly doing, so windows/mac users can mount their share [17:40] ideally it should be possible to access ubuntu one in a standard unix installation [17:42] dark: there's a bug to move us to gio from inotify, so it should run on solaris :) [17:42] no, i meant, without the ubuntu client [17:43] using already established unix tools [17:43] dark: I was pulling your leg. There are plans for a webdav client, which might take you closer. I don't see us writing an NFS server, but other people could. [17:44] i don't think we'll be doing nfs or sftp support [17:44] if you don't, how could other do? [17:44] or uucp [17:44] dark: the protocol is free and open [17:45] Chipaca, but your server isn't [17:45] if you want nfs support to avoid using our server, you're not going to run our server anyway [17:45] so i don't see how that argument fits? [17:46] no, i want nfs support *in your physical server* (i want the 2gbs) [17:46] dark: I mean, using the protocol you could write an nfs server that was also a client of our server, proxying [17:46] hmm yes, but this doesn't address my 'ideal' [17:46] but ok ^^ [17:56] dark: I'm not sure how this differs from your ideal, could you expand? [17:58] it would be nice that in any reasonable unix platform i would get a way to access ubuntu one without further installation; for the same reason it is nice to gmail to also have pop access (so that in any reasonable platform one could read email) [17:59] but i was thinking. megaupload/etc is full of porn.. even with those annoying captchas (and delays, and etc). i wonder what will happen with ubuntu one when people just notice it [17:59] and no, not having ftp access will not stop anyone from abusing the servers :) [18:03] well i hope in the karmic+1 ubuntu one work flawlessy in livecd [18:04] i think it's very compelling to have my data everywhere but also protect it from eavesdropping [18:59] hi all , i'm seeing a "Something has gone wrong (500) Server Error" when trying to view my contacts from the webUI. is this a known bug? [19:13] hi---why does the folders in the ubuntu one folder backuped only by their names and not content???/ [19:14] Say that another way. [19:14] hamzaatova2: what do you mean? (we don't) [19:16] Chipaca, the content is not backuped just the folders names [19:16] hamzaatova2: do you mean that you're not seeing files on the web ui, only folders? [19:17] Chipaca, yeah [19:17] hamzaatova2: try this: u1sdtool --waiting-content [19:18] Chipaca, what is u1sdtool? [19:18] (That should realy be an option in the applet or whatever.) [19:18] hamzaatova2: u1sdtool is a little tool that talks to the ubuntuone-syncdaemon over dbus [19:18] (Early adopters might understand shell commands. After that, all bets are off.) [19:18] hamzaatova2: the ubuntuone-syncdaemon is the process that does all the file synchronization [19:19] CardinalFang: yes. I mentioned this to John Lea, and I think his mind exploded, or something :) [19:19] Chipaca, i dont understand but how to do it??' [19:19] CardinalFang: I've got to get back to him [19:20] hamzaatova2: Put hte command Chipaca suggested into a terminal window. [19:20] hamzaatova2: ah! ok. Applications -> Accessories -> Terminal [19:20] hamzaatova2: and copy what I said to try into that window, and press enter [19:21] Chipaca, nettrot ---u1sdtool: error: no such option: --waiting-content [19:21] hamzaatova2: ah, ok [19:21] verterok: when did --waiting-content get into u1sdtool? [19:21] Chipaca: :) [19:21] Chipaca: I think a few weeks ago [19:22] Chipaca: before the xmas holidays [19:22] verterok: was it released? [19:22] verterok: i.e. into ubuntu? [19:22] Chipaca: probably in the nightly/beta ppa [19:22] ah [19:22] It's not in my version, and I'm running the latest ubuntu alpha here. [19:22] verterok: do you remember the dbus-send ? [19:22] Chipaca: yes, I have it my ZimWiki :) [19:22] gimme 1' [19:22] nettrot: strange, I'm running ... what is it I'm running [19:23] 1.1.0+r300-0ubuntu1~ppa1~karmic [19:23] I've never installed u1 out of the ppa. [19:23] ah, ok [19:23] nettrot: never mind then :) [19:23] hamzaatova2: give me a second, verterok will haev another command for you in a minute [19:23] Chipaca: dbus-send --session --print-reply --dest=com.ubuntuone.SyncDaemon --type=method_call /status com.ubuntuone.SyncDaemon.Status.waiting_content [19:24] hamzaatova2: that ^ (from dbus-send to the end) [19:24] in the terminal Chipaca ? [19:24] hamzaatova2: yes please [19:25] Chipaca, thats it??? [19:25] hamzaatova2: yes, it's long [19:25] ok---will it start to backup now? [19:26] hamzaatova2: did it output anything? [19:26] hamzaatova2: it's never stopped backing up, if it's working correctly [19:27] yes--- there is very long output [19:27] hamzaatova2: ok, all those files are waiting to be uploaded [19:27] ok---lets see if it will back it up [19:27] americans always say please---thank you and such [19:28] dbus-send --session --print-reply --dest=com.ubuntuone.SyncDaemon --type=method_call /status com.ubuntuone.SyncDaemon.Status.current_status [19:28] what does that ^ print ? [19:28] hamzaatova2: especially in the "name" entry [19:30] "READY_WITH_NETWORK_WITH_BOTHQ" [19:30] hamzaatova2: ok, so it's not connected [19:30] hamzaatova2: click on the icon, and click "connect" please :) [19:31] if you don't have an icon, you can do [19:31] dbus-send --session --dest=com.ubuntuone.SyncDaemon --type=method_call / com.ubuntuone.SyncDaemon.SyncDaemon.connect [19:31] there is no connect please :) [19:31] (the dbus-send way is more of a "connect dammit!") [19:32] the cloud is colorful after the former command [19:32] but there is no connect [19:33] the cloud is not there anymore [19:33] hamzaatova2: if you did the .connect, current_status should have changed [19:33] hamzaatova2: you can check that. Also, current_downloads should start showing movement :) [19:34] current_...something [19:34] 1 sec [19:34] hamzaatova2: u1sdtool --current-transfers [19:34] i didnt did connect and the cloud is not launching anymore [19:34] hamzaatova2: that should start showing progress [19:35] ok--i see a file in upload process [19:35] hamzaatova2: ok, so *do* do the .connect, and do then do the «u1sdtool --current-transfers» thing also [19:35] ok. Now, if you have a lot (10k) of small files, it will be slow. Other cases work ok. [19:35] i dont understand----and there is no cloud---why? [19:36] i dont have many files [19:36] hamzaatova2: the applet disappears when there is nothing interesting to say, unless you configure it differently [19:37] hamzaatova2: try: ubuntuone-client-preferences [19:37] what for??? [19:37] verterok: has the bandwidth throttling preferences been fixed? [19:38] Chipaca: don't know what's the status of the preferences GUI [19:38] Chipaca: it's fixed in the syncdaemon DBus API [19:38] hamzaatova2: It will let you set an option to always show the cloud. [19:38] hamzaatova2: just in case, don't touch the throttling options (there was a bug, and it's rather insidious to clean up after if you touch it) [19:38] ok--it is ok without it [19:39] ok---byyyyyyyyyyyyyyyyyyyyyyyyyyyye [19:40] any idea why libsoup and epiphany can not access couchdb ? [19:41] maybe they've got muddy feet [19:41] my mom never let us on the couch with muddy feet [19:41] CardinalFang: question for you ^ [19:42] adiroiban, I need more information. What happens? [19:43] soup_session_send_message return SOUP_STATUS_CANT_CONNECT [19:44] trying the link in firefox or chromium, everhing is fine [19:44] trying the link in Epiphany (webkit) i get libsoup error [19:44] I was wondering if I am the only one with this problem [19:44] the problem seems to be from libsoup [19:45] adiroiban, it looks just like a HTTP server. Can't connect seems pretty low-level. I say strace it to see what the socket is doing. [19:46] it's probably because there's an ipv6 address set up for the loopback interface [19:46] in /etc/hosts [19:47] I presume he's not connecting to ip6-localhost . [19:47] adiroiban, what is your URL you're using? [19:47] well, libsoup is trying to [19:48] Feel free to clobber the password, a. [19:48] ::1 localhost ip6-localhost ip6-loopback [19:48] http://localhost:60748/_utilshttp://localhost:60748/_utilshttp://localhost:60748/_utils [19:48] http://localhost:60748/_utils [19:48] sdfasdf [19:48] is probably in /etc [19:48] err [19:48] sorry [19:48] /etc/hosts [19:48] yep. I will look for ipv6 [19:49] in epiphany I'm using the bookmark file [19:49] it's the ipv6 thing [19:49] dobey, ?! Where did you get your hosts file? I have two machines and netighet look like that. [19:49] "neigher" [19:49] CardinalFang: it's the default in ubuntu afaik [19:50] Damn my fingers are cold and inflexible. [19:50] i've never changed it [19:50] adiroiban, Easy to find. $ grep " localhost" /etc/hosts [19:50] that did it [19:50] verterok: ping [19:50] dobey: thanks! [19:50] sure [19:50] dobey: pong [19:50] ugly stuff [19:51] verterok: hey, i'm trying to get some tests written which use DBusTwistedTestCase, and running into some problems, wondering if you could help [19:51] Weird. I'm on a breezy->lucid and a jaunty->lucid. [19:51] dobey: sure, shoot [19:51] i'm on karmic, upgraded from jaunty [19:51] i have a clean karmic install [19:52] verterok: i set up a FakeLogin(), and am trying to call 'login' on the dbus iface, but i keep getting the dbus error saying no such method [19:52] * CardinalFang boggles. I have the IPv6 logic in desktopcouch already, but for only one address. Maybe I should change it. [19:53] dobey: there is a test that uses a fake oauthdesktop... [19:53] * verterok looks [19:53] yeah [19:53] dobey: TestDBusOAuth [19:54] dobey: as there can be only one object exposed for each path, we need to chaneg the oauthdesktop path :( [19:54] *change [19:54] dobey: take a look to TestDBusOAuth.setUp [19:55] ok [19:55] dobey: probably your test isn;t finding the method because you already change the path? [19:55] s/;/'/ [19:58] CardinalFang: so you don't have an IPV6 localhost in your /etc/hosts? [19:58] "::1 ip6-localhost ip6-loopback" [19:58] "127.0.0.1 localhost" [20:04] verterok: hrmm, so now i'm just getting twisted timeout errors :/ [20:05] dobey: should be something in the logs about what's causing the timeout [20:11] CardinalFang: can you try to add localhost for ipv6 and see if you can use libsoup [20:18] rodrigo_: the http auth code is working, should I send it to you for review as a MP ? [20:40] verterok: nope, might just be due to something in the script i'm trying to test [20:40] dobey: could you pastebin the code? [20:43] not right now. gotta go to the airport [20:43] but we can discuss more tomorrow i guess :) [20:43] thanks [20:46] dobey: sure [21:11] adiroiban, send it as you want, as a patch, as a merge proposal in LP, etc [21:12] rodrigo_: ok. I have created a bug in LP, and attached a branch [21:15] adiroiban, send it as you want, as a patch, as a merge proposal in LP, etc [21:15] grr, inet connection sucks [21:15] rodrigo_: ok. I have created a bug in LP, and attached a branch [21:16] assign the bug to me please [21:17] done: bug 504452 [21:17] Launchpad bug 504452 in couchdb-glib "Implement Basic HTTP Auth" [Undecided,In progress] https://launchpad.net/bugs/504452 [21:17] cool! [21:18] adiroiban, can you propose the branch for merging into couchdb-glib trunk please? [21:18] it makes it easier to review [21:19] adiroiban, for next time, it's easier if you bzr branch couchdb-glib/trunk, and then: [21:19] bzr commit --fixes lp:xxxx -m "comment" [21:19] and then propose the branch for merging [21:20] that would link the branch to the bug when you bzr push [21:20] true. I was playing with bzr-git [21:20] ah, ok [21:20] That's good to know... [21:21] but it looks like it is not working as expected [21:21] as I can not push the branch to couchdb-glib [21:22] adiroiban, no it will push to your user [21:22] ah, right, you need some magic in ~/.bazaar/locations.conf [21:22] [/opt/extra/src/canonical/couchdb-glib] [21:22] push_location = lp:~rodrigo-moya/couchdb-glib [21:22] push_location:policy = appendpath [21:22] public_branch = lp:~rodrigo-moya/couchdb-glib [21:22] public_branch:policy = appendpath [21:22] replace /opt/extra... with the path where you have the couchdb-glib branches dirs [21:23] and my username with yours [21:23] then bzr push will just push to lp:~adiroiban/couchdb-glib/branch [21:27] how can I store using desktopcouch an empty list? My records have a list of "tags" that I use to organize, but those tags are optional and when there are none it raises a ValueError because the list is empty [21:30] well guess I can just store None instead by manually checking with an if, but that doesn't really look pretty [21:32] rodrigo_: still not able to combine git with bzr http://paste.ubuntu.com/353129/ [21:33] adiroiban, yay, don't worry about git, I'll do the merge manually myself [21:33] so keep your patch, and bzr branch... etc from what I said above [21:33] well. in the future I will create a plain bzr branch [21:34] and request the MP using that branch [21:34] the strange thing is that the branch was created (even if it was complaining) [21:35] so you can see the MP https://code.edge.launchpad.net/~adiroiban/couchdb-glib/basic-auth/+merge/16990 [21:35] cool [21:36] rodrigo_: there was a strange thing in couchdb-glib/couchdb.h . Even thou the header specifies 8 tab [21:36] there were some space indentations [21:36] ah [21:36] and I was not sure about the coding convention [21:36] feel free to complain about that and I will fix it [21:37] adiroiban, I'm thinking that the API might look better if, instead of _enable_oauth, _enable_basicauth [21:37] if we did something like what libsoup does with its SoupAuth object [21:37] that is, have a couchdb_set_authentication (couchdb, CouchdbAuthentication *auth) [21:38] and have an object derived from CouchdbAuthentication that does oauth and another that does basic auth [21:38] yes. we can created some wrapping objects for CouchdbAuthentication *auth [21:39] but if we are not going to support other authentication methods [21:39] if you want, I can create the CouchdbAuthentication interface, and then you implement the basic http on top of it [21:40] well, we might [21:40] not sure right now, but having this API would make it easier to extend [21:40] rodrigo_: sure. It should be easy to implement the basic http auth using the new interface [21:41] ok then, tomorrow you'll have the interface [21:41] no hurry [21:41] or maybe now, it indeed just needs to have a _sign_request or similar method [21:42] well, basic http auth just authenticates once, right? [21:42] ah, it's a signal on SoupSession [21:42] well... the header of each request will have the basic auth info [21:43] and the username and password are „cached” [21:44] in libsoup it is implemented as I signal to make it easy to implement a password prompt [21:45] it's cached in libsoup or in your code? [21:45] in my code [21:45] so, the authenticate signal on SoupSession is called for every SoupMessage that is sent? [21:45] and that requires authentication. yes [21:46] but I think the Auth object is also cached by libsoup [21:47] hmm, ok, sounds like we can have an interface with a 'authenticate' or 'sign_message' virtual method [21:47] yep. the Auth object is also cached [21:48] by libsoup [21:48] that should cover at least the needs for oauth and basic http [21:49] libsoup auth tests are here http://git.gnome.org/browse/libsoup/tree/tests/auth-test.c [21:49] yeah, looking at them now [21:49] they should cover various other authentication methods [22:03] not sure though how the SoupAuth stuff fits oauth, I guess we could add a SoupOAuthAuth object [22:04] or easiest thing would be to not use SoupAuth and have couchdb-glib call the authenticate virtual method for each msg [22:04] that should work for both basic and oauth, right? [22:07] I'll think a bit about it, so please do the same and let's discuss tomorrow, ok? [22:11] * rodrigo_ goes to think and relax :-D [22:26] :) [22:27] or we can just have OAuth [22:27] and the rest of Auth supported by libsoup