=== dendrobates is now known as dendro-afk === dendro-afk is now known as dendrobates [00:32] hmm http://aeric.poon.my/?p=496 might mean something to someone. === dendrobates is now known as dendro-afk [01:14] ajmitch, I apologize for the delay https://bugzilla.gnome.org/show_bug.cgi?id=626962 [01:14] Gnome bug 626962 in Community Extensions "Ubuntu One Music Store extension fails to build" [Normal,Resolved: notgnome] === dendro-afk is now known as dendrobates === JUMPA is now known as jumpa === jumpa is now known as JUMPA === teknico is now known as teknico_away === dendrobates is now known as dendro-afk [14:20] apachelogger: nessita: rodrigo_: ping [14:21] Chipaca: ahoy [14:22] Chipaca: pong [14:23] nessita: rodrigo_: [14:23] nessita, verterok: any objections against uploading my gsoc KDE stuff to universe for maverick? [14:23] (that is if it gets a FFe) [14:23] I wanted to check apachelogger was working with all the latest bugfixes [14:23] Chipaca: +1 [14:24] Chipaca: I'm not aware of what apachelogger has been working on [14:24] apachelogger, what does that include? [14:24] besides "KDE" :-) [14:24] nessita: quite. So, talk with him :) [14:24] Chipaca: we need to be sure about the sso dbus signals :) [14:24] http://apachelog.wordpress.com/2010/08/21/ubuntu-one-technical-aspects/ [14:24] apachelogger, remember the kwallet patch you submitted is going to be removed in the next upload (today), so you might want to make sure you add that to ubuntu-sso-client [14:25] do I have something to patch against? [14:25] seeing as my patch was rejected due to sso being in a constant state of flux [14:25] apachelogger: also note we can't merge your patch upstream until you sign the canonical contributor agreement (if you want to do that) [14:25] apachelogger, ubuntu-sso-client, should be stable now for you to patch it [14:27] apachelogger, about the UI, we even added DBus interfaces to u-sso-client to allow other UIs to be implemented [14:28] Chipaca: last I saw the agreement I got advised not to sign it... so... [14:28] apachelogger: by whom? [14:28] a lawyer friend of mine [14:29] apachelogger: okie then :) [14:30] Chipaca: I'm not sure what you want me to look :-) [14:30] Chipaca: do we have a merge proposal to review? [14:30] nessita, http://apachelog.wordpress.com/2010/08/21/ubuntu-one-technical-aspects/ [14:31] supposedly we could just use ubuntu-sso-client as is in KDE too? [14:32] i.e. only deploy the UI components (based on syncdaemon's dbus interface) in maverick [14:32] apachelogger, there is a low-level DBus interface in u-sso-client that you could use to provide a QT ui [14:33] apachelogger: we provide a DBus interface and you could build a KDE GUI on top of that [14:33] apachelogger, we added that in fact based on your needs to have a QT UI [14:33] nessita: so... I'm not sure what apachelogger's current work entails, but I thought the easiest way to get him an answer as to whether it should/could go in universe. [14:33] apachelogger: let me give you a pointer [14:33] nessita: or rather, I'm pretty sure it shouldn't, for the reasons I think you're explaining (that I already explained), but you could get into more detail than i can [14:33] alecu: do you have the wiki pointer for ussoc? [14:34] nessita: do we need to create a Qt UI on top of it? [14:35] nessita, yes... just a sec. [14:35] * apachelogger finds it unlikely to roll out a somewhat stable UI in time to accumulate testing for maverick [14:35] apachelogger: what we could do is split ubuntu-sso-client into ubuntu-sso-client and ubuntu-sso-client-gtk, and you patch the former and replace the latter with -qt [14:36] Chipaca: the latter you meant? [14:36] nessita: patch the former to also use kwallet [14:36] ah [14:36] :-) [14:36] https://wiki.ubuntu.com/SingleSignOn/UbuntuSsoClient [14:37] nessita, but that's the doc for programs that need credentials [14:37] apachelogger: ok, the link that alecu pasted shows the high level DBUS interface. I forgot we haven't documented yet the "low-level" interface that the GUI's use [14:38] nessita, yup, we still need to write the docs for someone wanting to login/register [14:38] * to make a login/register ui [14:39] apachelogger: so, basically, SSO provides a DBUS service with 2 main interfaces: the Credentials which is what we call "high level" (apps like syncdaemon use them) [14:39] apachelogger: and the UserManagement, which is a lower level interface, that is used by the GTK GUI [14:40] apachelogger: and that any other interface (let's say, KDE), can use to provide the same functionality [14:57] nessita, hi, i wanted to ask what is the correct way for up-to-date maverick system to authorize the machine. In my case a browser window is opened when I run ubuntuone-preferences, should it be this way or I should get a GUI for that? [14:58] nessita: sso-client package misses recommends on gnome-keyring [14:58] rye: we're landing the latests versions for everything, do you have trunk for all? [14:58] apachelogger: why should it recommend it? [14:59] nessita: because you only support gnome-keyring === zyga is now known as zyga-us-visa-stu [15:00] apachelogger: we have the requirement of python-keyring, which requires gnome-keyring [15:00] apachelogger: so I don't understand why we need to "recommed" gnome-keyring being that we need it to work (at least the current version) [15:00] me@osiris:~$ apt-cache show python-gnomekeyring|grep gnome-keyring [15:00] Depends: libc6 (>= 2.4), libglib2.0-0 (>= 2.12.0), libgnome-keyring0 (>= 2.20.3), python (<< 2.7), python (>= 2.6), python-support (>= 0.90.0), python-gtk2 [15:00] Suggests: gnome-keyring [15:00] I beg to differ [15:01] Chipaca: what was the command you taught me to see package dependencies? [15:01] nessita: apt-get rdepends? [15:01] let's see [15:02] nessita: or apt-cache putty :) [15:02] not putty, dotty [15:02] off-by-180 [15:02] rdepends has a misleading name, it lists reverse relations rather than depends in particular BTW [15:03] Chipaca: E: Invalid operation (for rdepends) :-) [15:03] nessita: apt-cache, not get :P [15:03] apachelogger: "libgnome-keyring0" -> "gnome-keyring"; [15:03] apachelogger: so, "python-gnomekeyring" -> "libgnome-keyring0"; and "libgnome-keyring0" -> "gnome-keyring"; [15:04] apachelogger: we do depend on gnome-keyring :-) [15:04] me@osiris:~$ apt-cache show libgnome-keyring0 | grep gnome-keyring | grep -v libgnome [15:04] Suggests: gnome-keyring [15:04] Breaks: gnome-keyring (<< 2.30) [15:04] I again beg to differ [15:05] apachelogger: what I showed you is the output of apt-cache dotty python-gnomekeyring [15:05] dotty also tracks relations and not dependencies in particular [15:06] apachelogger: I'm new at packaging, so I probably made a mistake (or ten), but honestly I'm not convinced of this [15:06] apachelogger: how come python-gnomekeyring does not depend of gnome-keyring? that sounds weird [15:07] because it only depends on the library and the library only suggest gnome-keyring [15:07] I am sure there is a reason for that [15:07] but why exactly you will have to ask someone who has to do with those packages [15:07] not exactly my area of work either :P [15:09] heh [15:26] rye: hey, you never answered [15:28] nessita, i am running bleading-edge maverick , not the version from trunk === JanC_ is now known as JanC [15:52] anyone knows about glibc brokenness on latest maverick packages? I seem to be having weird memory corruption bugs, and after 1 hour looking at my code, it doesn't seem to be my fault [16:06] rodrigo_, are you talking about nautilus plugin? [16:07] rye, I am talking about a nautilus plugin branch I'm working on, yes [16:08] it's segfaulting on a g_free call [16:11] rodrigo_, related to _CBData ? [16:12] rye, no [16:12] rye, it's crashing here: [16:12] if (finfo->priv->path != NULL) [16:12] g_free (finfo->priv->path); [16:13] rodrigo_, is finfo->priv != NULL ? [16:13] rye, it should, yes [16:14] it's crashing on the g_free === zyga-us-visa-stu is now known as zyga [16:23] aha, I found it!! [16:24] stupid copy&paste === dendro-afk is now known as dendrobates [16:52] alecu, nessita, Chipaca: https://code.edge.launchpad.net/~rodrigo-moya/ubuntuone-client/check-published-files/+merge/33660 === rye changed the topic of #ubuntuone to: Ubuntu One services will be unavailable for 30 min starting at 17:00 UTC | Need assistance? Review the Status and the FAQ first: https://wiki.ubuntu.com/UbuntuOne/Status and https://wiki.ubuntu.com/UbuntuOne/FAQ === dendrobates is now known as dendro-afk [17:13] rodrigo_, reviewing your branch === beuno is now known as beuno-lunch === nessita is now known as nessitalunch [17:26] nessitalunch: I am getting a segfault with ubuntu-sso-client [17:32] hey, is there any way to see which file ubuntuone is syncing at the moment? === dendro-afk is now known as dendrobates === beuno-lunch is now known as beuno === dendrobates is now known as dendro-afk === nessitalunch is now known as nessita [18:39] apachelogger: can you please be more specific? :-) [18:45] nessita: trying to login -> backend goes down with segfault [18:45] apachelogger: how are you trying to login? [18:45] last log entry: 2010-08-25 18:26:15,602:602.580070496 ubuntu_sso.main keyring_store_credentials: app_name dbus.String(u'Ubuntu'). [18:45] nessita: using the GTK UI [18:45] apachelogger: how are you running the GTK UI? [18:46] from terminal without args [18:46] apachelogger: you're having a mismatch between the version of the UI and the version of the daemon [18:47] be sure you're running both from the same branch [18:47] I am running from the packages [18:47] apachelogger: you shouldn't be running the GUI other than from a dbus call [18:47] apachelogger: can you paste how are you running the GUI? [18:47] the exact command [18:48] me@osiris:~$ ubuntu-sso-login-gui [18:48] and in another tab [18:48] me@osiris:~$ /usr/lib/ubuntu-sso-client/ubuntu-sso-login [18:49] apachelogger: where are you getting the ubuntu-sso-login-gui from? is not shipped in the package [18:49] apachelogger: that script is not supported (just FYI) [18:49] apachelogger: we will be removing it soon [18:50] me@osiris:~$ dpkg -L ubuntu-sso-client | grep login-gui [18:50] /usr/bin/ubuntu-sso-login-gui [18:50] it is shipped in the package [18:51] apachelogger: it shouldn't be, I'll file a bug. Thanks for pointing this out! [18:51] nessita: how should one be starting the gui? [18:51] apachelogger: anyways, you should raise the GUI using Dbus [18:51] apachelogger: use d-feet or something similar to call the dbus method [18:52] apachelogger: the methods you want are in the Credentials interface. Either "login_or_register_to_get_credentials" or "login_to..." [18:52] k [19:00] nessita: http://paste.ubuntu.com/483561/ [19:00] apachelogger: you don't have the gnome keyring available [19:01] apachelogger: I'm a bit lost what your question is [19:01] what part is supposed to create it though? [19:01] apachelogger: the keyring is a dbus service [19:01] * apachelogger has a default keyring, it just seems that it falls over not having a login keyring [19:01] alecu: look ^ [19:01] me@osiris:~$ qdbus org.gnome.keyring [19:02] / [19:02] gnome keyring is there [19:02] apachelogger: ah yes, sorry [19:02] also in seahorse I can access it properly === teknico is now known as teknico_away [19:02] apachelogger: I'm just passing this to al [19:02] alecu* [19:02] aye [19:02] apachelogger: we have a bug reported for that [19:02] I'm about to start working on this issue, yes [19:03] simply creating a login keyring should suffice to work around it? [19:03] it's #624033 [19:03] * apachelogger finds "login" a bit of a generic name TBH [19:03] apachelogger: not sure, let the master answer [19:04] nessita: segfault again [19:04] last line of log: 2010-08-25 20:04:16,309:309.218883514 ubuntu_sso.main keyring_store_credentials: app_name dbus.String(u'ubuntuone-syncdaemon'). [19:05] apachelogger: may be related to not having the keyring available? [19:05] I created one :P [19:05] hum [19:05] that is after I obtained a token [19:05] apachelogger: just FYI, app_name should be "Ubuntu One", which is a constant defined in ubuntuone.clientdefs [19:05] apachelogger: so the token actually works with syncdaemon [19:06] k [19:06] well, since it crashes anyway... ;) [19:07] apachelogger, "login" is the keyring that's automatically unlocked by some pam module using your login in password when you have just logged in with gdm. [19:07] alecu: I c [19:07] that is bound to fail on KDE systems [19:07] or you are missing a dep for that [19:08] apachelogger, which dep do you think might solve this? [19:08] dunno [19:08] I never heared of such a pam module before [19:09] apachelogger, also it's failing on some gnome systems that have a keyring named "default" but no "login" keyring. [19:09] apachelogger, it's lib-pam-gnomekeyring [19:10] installed [19:10] http://live.gnome.org/GnomeKeyring/Pam [19:10] login gets created if not present [19:10] but it seems KDM will not include the PAM module [19:11] (which IIRC was also the problem that early on lucid was not able to start gnome-keyring in KDE - in GNOME it did get autostarted frome the module but in KDE not) [19:11] apachelogger, I think that's your pam setup, nothing to do with kdm [19:11] but yeah, you'd need gnome-keyring autostarted === dendro-afk is now known as dendrobates [19:16] me@osiris:/etc/pam.d$ grep -ri gnome_key . [19:16] ./common-password:password optional pam_gnome_keyring.so [19:16] no autostart -> no login keyring [19:17] alecu: I would check for existance of the keyring and if not manually create it === rye changed the topic of #ubuntuone to: Need assistance? Review the Status and the FAQ first: https://wiki.ubuntu.com/UbuntuOne/Status and https://wiki.ubuntu.com/UbuntuOne/FAQ | http://bit.ly/caHbOf for help on adding your computer | Then say "honk" for more help. Please be patient :) | https://one.ubuntu.com/ === yofel_ is now known as yofel [19:46] rye: any chance I'll get access to my contacts before I get old and wrinkled? ;) [20:09] rodrigo_: so you've managed to track down how to fix up the mono bindings? [20:10] * ajmitch was looking for a fix yesterday for it [20:13] nessita: want me to get a backtrace? [20:14] ajmitch, not yet, busy with other stuff, but will look at it soon [20:19] apachelogger: for the segfaulting? yes please [20:20] apachelogger: also, you're running kde, right? (just to know) [20:21] rodrigo_: ok, you'll probably end up cursing gapi2-codegen's lack of useful output as I did :) [20:23] nessita: yes, I am running KDE [20:24] ajmitch, yeah :( [20:25] ajmitch, only output I get is something about not knowing what GtkVBoxClass is [20:25] ajmitch, do you get the same? [20:25] yep [20:25] adding in the syncdaemon header header into ubuntuone.source generated C# files for at least that part, but still no useful u1ms code [20:25] & it of course failed to compile, not unsurprising [20:27] apparantly the gapi stuff hasn't seen a lot of love recently [20:31] ajmitch, it used to work ok, and there have been no changes, so not sure what's up [20:32] the only code changes to the u1ms widget recently have been adding in the syncdaemon support, but I'm not sure how that should affect it [20:32] nessita: http://paste.ubuntu.com/483608/ [20:33] diving into the gapi code to get some useful answers takes a little while [20:33] ajmitch, in fact, it started happening when I removed the previous files that were talking to syncdaemon, to replace them with libsyncdaemon [20:33] ajmitch, so not sure what doesn't make gapi happy [20:33] the previous code didn't link to an external library (libsyncdaemon)? [20:34] it would seem rather strange if you did need to have any information about libsyncdaemon in the code for gapi to look at [20:35] hmm [20:35] but it's just internal usage, not in the headers it is supposed to be parsing [20:35] or does gapi look at the .c files? [20:36] the headers afaik [20:36] yeah, that's what I thought [20:37] without useful output from the tool this is just speculation really [20:38] & frustrating at that :) [20:38] Field ParentClass has unknown Type GtkVBoxClass in Struct UbuntuOne.U1MusicStoreClass [20:38] yeah, that's the only thing I get [20:39] that's boilerplate widget code you've got there too [20:39] but I guess he should know UbuntuOne.U1MusicStoreClass is a GTK class struct [20:39] apachelogger: I have no idea what's going on there, seems like a dbus issue [20:39] apachelogger: maybe you have a too old/ too new dbus? [20:39] apachelogger: maybe gnomekeyring depends on a different dbus than the one that kde provides? [20:40] there is only one dbus [20:40] so either what is currently in maverick is broken or it is something else [20:40] apachelogger: what's currently in maverick is working on gnome desktops [20:40] apachelogger: do you have a gnome VM? maybe you can try there and see if it fails for you [20:42] dont have one around [20:42] ajmitch, well, that's the usual boilerplate for GTK widgets, so I'd assume gapi knows about it [20:48] * ajmitch shall have to get back to it later [20:51] ajmitch, in /usr/share/gapi-2.0/gtk-api.xml, there is no Gtk*Class at all, so not sure why it tries to parse the music store one :( [20:52] or more, why did it work up until now/ [20:53] * ajmitch will be back later [20:57] * apachelogger pokes his dbus with a long pointy stick of fluff [21:53] apachelogger, I just removed your kwallet patch from the u1-client package, since it didn't apply anymore [21:54] ajmitch, for when you get back, I've succeeded in not having the GtkVBoxClass warning, but still doesn't create any .cs files [21:54] * ajmitch is around now [21:54] so that warning may have been somewhat spurious? [21:55] fwiw, I can get a U1MusicStoreClass.cs created by commenting out the line that refers to parent_class in the .xml file [21:55] ajmitch, unfortunately, seems so [21:55] I haven't checked to see if the assembly that is created is at all useful though [21:56] ajmitch, hmm, maybe we can mark it to be removed on the ubuntuone-api.metadata file [21:56] let me try... [21:56] that would probably remove it, but I don't know if it should be [21:56] worth a try [21:58] no, no luck :( [21:58] rodrigo_: cruel world [21:59] apachelogger, patch ubuntu-sso-client, as you talked with nessita [22:08] http://simplest-image-hosting.net/jpg-0-plasma-desktoplm3446 [22:09] ajmitch, so, you get something in ubuntuone-api.xml? [22:09] ajmitch, right now, I get it empty [22:10] rodrigo_: empty after changing what? [22:10] at the moment I'm rebuilding the previous libubuntuone in a chroot to have a look at it [22:10] ajmitch, empty when it gets generated [22:10] ajmitch, ok [22:15] assuming that pbuilder will cooperate & let me into the chroot after it built [22:15] but I can at least see that the gapi2-codegen output is different in the build log [22:15] * ajmitch just needs to look at the xml file [22:19] ajmitch, I'm looking at the gtk-sharp package, to see if anything changed there [22:19] nothing has changed there [22:20] ah, you checked already? [22:20] it hasn't changed for quite awhile (last upload 14 weeks ago) [22:21] yeah, I've been complaining about the gapi stuff in the debian mono channel as well :) [22:21] ok [22:25] * ajmitch is wanting to get this as working as possible before uploading to debian experimental [22:40] * ajmitch wonders why the ubuntuone-api.xml file is so very different between the 2 releases [22:40] in 0.3.2, it has object name="U1MusicStore", and in 0.3.3 it only gets the struct name="U1MusicStoreClass" [22:41] nessita: the gui uses gobject signals? [22:41] yet more evidence that gapi is just crap :) [22:45] eh, let's just rewrite it all in vala, and not provide any bindings to anything ;) [22:45] dobey: can you get it done by next week? :) [22:46] apachelogger: yes [22:48] sure, but i don't measure time in earth days, so one week is actually 3 years in earth time, for me ;) [22:56] dobey: we were just discussing how it'd be nice to use gobject introspection for mono bindings, if you have a free evening to work on it :) [22:59] me and mono don't mix well [23:02] it was worth a try [23:03] don't worry, the gapi code that parses C at the moment is written in perl, which is such an improvement [23:07] nessita: why not dbus? [23:07] eh, i like perl [23:08] apachelogger: the GUI part is not a dbus service, is just a plain GTK interface, so it signals thru gobject [23:09] nessita: but the backend is a dbus service so the gui could just call a method of the backend's dbus api [23:13] apachelogger: yeah, propably. We didn't design it that way, we should be able to re-think that for the next cycle === dendrobates is now known as dendro-afk