=== dendro-afk is now known as dendrobates | ||
=== dendrobates is now known as dendro-afk | ||
=== dendro-afk is now known as dendrobates | ||
=== dendrobates is now known as dendro-afk | ||
=== dendro-afk is now known as dendrobates | ||
=== dendrobates is now known as dendro-afk | ||
gekken11 | Does anyone know if there is a workaround/fix for no "add this computer" issue in UbuntuOne? | 03:17 |
---|---|---|
gekken11 | honk | 03:19 |
duanedesign | gekken11: hello | 03:21 |
gekken11 | ya e | 03:22 |
duanedesign | gekken11: yes there is | 03:22 |
gekken11 | heya duane | 03:22 |
gekken11 | if it is u1sdtool -q; killall ubuntuone-login; u1sdtool -c | 03:22 |
gekken11 | that doesn't work | 03:22 |
duanedesign | can you open applications > accessories > passwords and encryption | 03:23 |
duanedesign | click the arow to expand the folder | 03:23 |
duanedesign | and look for an Ubuntu One key | 03:23 |
gekken11 | not seeing that as a menu choice | 03:24 |
duanedesign | Passwords and Encryption Keys? | 03:24 |
gekken11 | found it. not in accessories | 03:25 |
gekken11 | ok, delete the desktop couch keys? otherwise, nothing | 03:27 |
duanedesign | no | 03:27 |
duanedesign | ok | 03:27 |
duanedesign | if there is no key open http://one.ubuntu.com/account/machines | 03:28 |
gekken11 | ok | 03:29 |
duanedesign | and see if your machine is listed | 03:29 |
gekken11 | nope | 03:29 |
duanedesign | ok | 03:29 |
gekken11 | lol, that's why I'm here! hehe | 03:29 |
duanedesign | are you using firefox as your default browser? | 03:30 |
duanedesign | System > Preferences > Preffered Applications Web Browser | 03:31 |
duanedesign | make sure Firefox is selected | 03:31 |
gekken11 | yes | 03:31 |
duanedesign | ok | 03:32 |
duanedesign | can you close Ubuntu One preferences. Close Firefox. | 03:32 |
gekken11 | yeah, did all that | 03:33 |
gekken11 | and a reboot | 03:33 |
duanedesign | open a terminal and just run: u1sdtool -q | 03:33 |
gekken11 | saemon stopped | 03:33 |
duanedesign | once that goes back to prompt try: killall ubuntuone-login | 03:33 |
duanedesign | then open Ubuntu One from the Me Menu | 03:34 |
gekken11 | hmm, no process found | 03:34 |
duanedesign | you can run | 03:34 |
gekken11 | let me see if there is a PID | 03:34 |
duanedesign | ps aux | grep ubu | 03:34 |
duanedesign | that will return any running Ubuntu One processes | 03:35 |
duanedesign | you should get just : grep ubu | 03:35 |
gekken11 | yeah, there are several others | 03:35 |
gekken11 | sso-login | 03:35 |
gekken11 | some python process | 03:36 |
duanedesign | well lets quit those | 03:36 |
gekken11 | dead | 03:39 |
duanedesign | once they are quit you can try opening Ubutn One Preferences | 03:40 |
duanedesign | from the 'Me Menu' | 03:40 |
gekken11 | yeah, no love | 03:41 |
duanedesign | gekken11: what does it do? | 03:41 |
duanedesign | does it open the browser at all | 03:41 |
gekken11 | nope | 03:41 |
duanedesign | are you behind a proxy, or use a script blocking application like no script | 03:42 |
gekken11 | nope | 03:42 |
gekken11 | fresh install | 03:42 |
duanedesign | ok can lets try | 03:43 |
duanedesign | can you run: sudo apt-get install ubuntuone-client-tools | 03:44 |
ajmitch | 0 | 03:44 |
duanedesign | after that installs | 03:44 |
gekken11 | installed | 03:45 |
duanedesign | try: u1sync --authorize | 03:45 |
gekken11 | yatzee | 03:45 |
gekken11 | or however that is spelled | 03:46 |
duanedesign | just for future references the u1sync tool is used for testing and not recommended for daily use | 03:46 |
duanedesign | it has proven useful in instances like this however | 03:46 |
gekken11 | right on, man thanks! | 03:46 |
duanedesign | great | 03:47 |
gekken11 | now I feel I must help some other n00b | 03:48 |
gekken11 | N00BS come to me. I will provide mediocre advice | 03:48 |
gekken11 | and if it fails, I will send you to Duanedesign | 03:48 |
duanedesign | :D | 03:48 |
gekken11 | sorry, no good deed goes unpunished | 03:49 |
=== dendro-afk is now known as dendrobates | ||
=== dendrobates is now known as dendro-afk | ||
=== dendro-afk is now known as dendrobates | ||
=== chrowe_ is now known as chrowe | ||
=== JUMPA is now known as jumpa | ||
=== yofel_ is now known as yofel | ||
=== lifeless_ is now known as lifeless | ||
=== didrocks1 is now known as didrocks | ||
mkarnicki | afternoon all | 12:24 |
rye | mkarnicki, hello! | 12:25 |
mkarnicki | hi rye =) | 12:26 |
duanedesign | hello mkarnicki rye | 12:46 |
mkarnicki | hi duanedesign | 12:46 |
=== teknico is now known as teknico_away | ||
duanedesign | rye: 'pyinotify - ERROR - add_watch: cannot watch' error in exception.log, is that enough information to determine what is going wrong or is additional information needed? | 12:52 |
duanedesign | bug 614161 | 12:53 |
ubot4 | Launchpad bug 614161 in ubuntuone-client (Ubuntu) "ubuntu one applet crash on login (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/614161 | 12:53 |
rye | duanedesign, that usually means that max_user_watches is set to a low value and it most probably will be unless trackerd is installed which increases this value | 12:53 |
rye | duanedesign, sysctl fs.inotify.max_user_watches | 12:53 |
duanedesign | rye: ahh, thats right. thank you! | 13:00 |
facundobatista | ok | 13:04 |
facundobatista | confirmado: me tengo que comprar una UPS | 13:04 |
nessita | facundobatista: why? | 13:05 |
nessita | (and hi) | 13:05 |
facundobatista | nessita, had an energy micro-cut, and my PC went down | 13:05 |
=== teknico_away is now known as teknico | ||
Chipaca | apachelogger: ping | 14:44 |
apachelogger | Chipaca: pong | 14:44 |
Chipaca | apachelogger: re: your two branches for kwallet | 14:45 |
Chipaca | apachelogger: the work is still in flux, but we're getting rid of the gnomekeyring dep | 14:45 |
Chipaca | apachelogger: so it kinda obsoletes your work :-/ | 14:45 |
apachelogger | Chipaca: what does the new solution look like? | 14:46 |
Chipaca | apachelogger: we talk to ubuntu-sso-client over dbus | 14:46 |
Chipaca | now, the current ussoc *does* dep on gnomekeyring | 14:47 |
* apachelogger reimplemented ussoc in Qt and that beast does not depend on gnomekeyring nor kwallet and essentially can implement encrypted token query in like 50 sloc :P | 14:48 | |
Chipaca | apachelogger: you did? did you let us know? :) | 14:48 |
apachelogger | I am pretty sure I a mentioned it a couple of times ;) | 14:49 |
Chipaca | nessita: ^ | 14:49 |
apachelogger | btw, I think a mailing list would be a good idea to setup | 14:49 |
apachelogger | https://code.edge.launchpad.net/~apachelogger/ubuntu-sso-client/gsoc | 14:49 |
nessita | Chipaca: looking, had no idea | 14:50 |
nessita | apachelogger, Chipaca: is C code | 14:51 |
nessita | is a completely different project, with the same name, right? | 14:52 |
Chipaca | wut? | 14:52 |
* Chipaca looks | 14:52 | |
nessita | bzr: ERROR: Branches have no common ancestor, and no merge base revision was specified. | 14:54 |
apachelogger | it is a completely own implementation to prevent RAM from getting eaten but has the same dbus interface | 14:55 |
apachelogger | also it is C++ with QtCore (i.e. non-gui dependant which also works in favor of light weight) | 14:57 |
apachelogger | in particular it currently uses 700 KiB non-shared memory, of which around 176 KiB are actual heap data | 14:58 |
apachelogger | in a gnome environment it would come to about 1.5 MiB I would assume | 14:59 |
apachelogger | (that is if no other Qt apps are running) | 14:59 |
helo | is it possible to buy music from ubuntuone using only a web interface? | 15:26 |
mkarnicki | apachelogger: what does gsoc stand for in /ubuntu-sso-client/gsoc? is it GSoC realted :D? | 15:40 |
apachelogger | mkarnicki: side product | 15:42 |
mkarnicki | apachelogger: sorry, didn't get that.. is it an ancronym? | 15:43 |
mkarnicki | apachelogger: btw I'll be writing the same thing for android really soon, and I feel like this moment is perfect to ask that question - should I name it ubuntu-sso-android-client or android-ubuntu-sso ? ideas :)? | 15:44 |
mkarnicki | apachelogger: oh shoot, I get it. sorry! | 15:44 |
mkarnicki | apachelogger: it's like branching the project, and pushing ~username/project/feature right? | 15:45 |
* mkarnicki feels silly | 15:45 | |
apachelogger | aye | 15:47 |
apachelogger | just that mine is not really a branch ;) | 15:47 |
apachelogger | mkarnicki: btw, I would go for ubuntu-sso-client-android | 15:48 |
mkarnicki | apachelogger: thanks :) | 15:48 |
=== dendrobates is now known as dendro-afk | ||
=== dendro-afk is now known as dendrobates | ||
=== jumpa is now known as JUMPA | ||
=== JUMPA is now known as jumpa | ||
rodrigo_ | apachelogger, ping | 16:05 |
verterok | apachelogger: hi | 16:05 |
verterok | apachelogger: just commented on the kwallet support branch for syncdaemon | 16:06 |
verterok | apachelogger: we think it's better to add support in ubuntu-sso for kde than in syncdaemon as we are about to drop the gnomekeyring dependency really-soon-now | 16:06 |
mkarnicki | CardinalFang: easy question this time :D - should I use Log.isLoggable(TAG, LEVEL) or a static boolean LOCAL_LOG flag to turn on/off logging? (I read that string building in log messages is expensive) | 16:13 |
CardinalFang | mkarnicki, Yes, it can be too expensive IF you're logging in a tight loop and constructing String objects. I don't turn off logging usually. Log constants or not in a loop, and I think you're okay. | 16:29 |
CardinalFang | mkarnicki, if you're concerned about a loop, a bool check is very cheap, so I'd do that. | 16:30 |
CardinalFang | ...or just comment out that code. | 16:30 |
CardinalFang | ...or make two functions of code (with and without logging), and programatically switch them earlier based on whether you want logs or not | 16:32 |
CardinalFang | a preprocessor would help with that last dumb idea. | 16:32 |
mkarnicki | CardinalFang: very cool, thanks for all hints :) | 16:36 |
mkarnicki | CardinalFang: is there anything like a preprocessor in Java ;d ? I like #ifdef's in C/C++ ;) | 16:36 |
mkarnicki | CardinalFang: nah.. bool checks should do fine. thanks! | 16:37 |
CardinalFang | mkarnicki, Nope, not AFAIK. Best not to ask me, or I'll have you using m4. | 16:37 |
mkarnicki | CardinalFang: M4? hahah thanks :) I have enough info now, both from you and #android-dev | 16:38 |
mkarnicki | cool. | 16:38 |
helo | are there plans to allow purchasing music through ubuntuone using a web browser? | 17:06 |
CardinalFang | helo, good question. It's technically possible. Do you object to using it embedded in a music player in principle, or do you dislike Rhythmbox? | 17:10 |
=== beuno is now known as beuno-lucnh | ||
helo | CardinalFang: i'd like to be able to browse/purchase new music while not using ubuntuone-client and associated dependencies (linux, synched home pc, etc) | 17:19 |
CardinalFang | helo, ah. It's not really hard, but not near the top of our to-do list. 10.10 beta is really close, and we have too much to do for the next few weeks to even think about it. | 17:26 |
CardinalFang | helo, the pieces are all there. Web UI to purchase, Web UI to download from Ubuntu One storage. | 17:28 |
Chipaca | verterok: rodrigo_: I suspect kenvandine pinged too many people about those branches, and we're all getting back to apachelogger about them :) | 17:29 |
Chipaca | verterok: rodrigo_: at least we're all saying the same thing (phew) | 17:29 |
rodrigo_ | :) | 17:29 |
rodrigo_ | sorry apachelogger for the delay, anyway | 17:29 |
kenvandine | hehe | 17:30 |
kenvandine | i just pinged joshuahoover :) | 17:30 |
=== dendrobates is now known as dendro-afk | ||
=== dendro-afk is now known as dendrobates | ||
=== dendrobates is now known as dendro-afk | ||
=== dendro-afk is now known as dendrobates | ||
Chipaca | apachelogger: so, about this qt ussoc, is it going into the maverick? | 18:25 |
=== beuno-lucnh is now known as beuno | ||
=== dendrobates is now known as dendro-afk | ||
=== dendro-afk is now known as dendrobates | ||
mkarnicki | beuno: you around? | 18:56 |
Chipaca | apachelogger: the api will probably change or grow another leg over the next day or so, and the freeze is upon us, so *please* talk with us :) | 19:02 |
beuno | mkarnicki, hi | 19:05 |
mkarnicki | hi beuno, one question if I may | 19:05 |
mkarnicki | aquarius is not back and I have that interesting question | 19:05 |
mkarnicki | it's about UDFs and where to download them | 19:05 |
beuno | good question | 19:05 |
mkarnicki | normally, files land under '/sdcard/Ubuntu One' but it's not the proper place for UDFs to land | 19:06 |
mkarnicki | since they're not *in* that folder | 19:06 |
beuno | mkarnicki, so, in the webui, we show them in the root folder | 19:06 |
mkarnicki | and I'm more and more thinking about... yes yes, but the question is | 19:06 |
mkarnicki | where to save them. | 19:06 |
beuno | so | 19:06 |
mkarnicki | and my solution, or suggestion | 19:06 |
mkarnicki | would be to use android function getApplicationFilesDir() or something like that | 19:07 |
beuno | you could save them in /sdcard/Ubuntu One/full/path/of/the/udf | 19:07 |
mkarnicki | which will return something ala-program-files on windows, which is /sdcard/Android/data/com.ubuntuone.androidu1 | 19:07 |
mkarnicki | beuno: if we introduce sync to AndroidU1 | 19:07 |
mkarnicki | beuno: it would be a mess to distinguish whish files under '/sdcard/Ubuntu One' are '~/Ubuntu One' files and which are UDF files | 19:08 |
beuno | yeah, I think I lack a bit of knowledge on how android works here | 19:09 |
mkarnicki | beuno: naturally, only starred items will be synced, but I think it needs consideration where should we keep that UDF content | 19:09 |
mkarnicki | CardinalFang: ^ what do you think? | 19:09 |
mkarnicki | beuno: normally, ~/Ubuntu One is ~/Ubuntu One and UDFs are somewhere under ~/ -- but we don't want to make a mess on the sdcard with UDFs | 19:09 |
mkarnicki | beuno: therefore I'm considering moving *all* files (both ~/Ubuntu One and UDFs) under /sdcard/Android/data/com.ubuntuone.androidu1/_here_ | 19:10 |
beuno | mkarnicki, I would make /sdcard/Ubuntu One the root folder, aka, ~ | 19:10 |
mkarnicki | it's little bit cryptic, but so is C:/Program Fiels/Canonical/stuff (I'm making this up) | 19:10 |
beuno | sure, that could work as well, but would be harder for people to find the files on the sdcard | 19:10 |
mkarnicki | true. that's the problem :< | 19:11 |
mkarnicki | I can think of holding everything under /sdcard/Ubuntu One , it seems resonable plus easily accessible | 19:11 |
mkarnicki | beuno: you know what. when I talk much, I clear my mind. thanks :) it just needs to be a lil' backed-up by the database, and that should do. | 19:13 |
beuno | mkarnicki, having to communicate ideas helps shed light on to them | 19:13 |
mkarnicki | beuno: :) | 19:13 |
beuno | I can't tell you how many times we start questions and end up answering them ourselves | 19:13 |
mkarnicki | =D | 19:14 |
mkarnicki | I'm glad I'm 'not alone' ;) | 19:14 |
mkarnicki | or, should I say, it's not just me hehe | 19:14 |
mkarnicki | I'll rethink how to handle UDFs - aquarius really wants them ;) | 19:15 |
mkarnicki | I'm currently cleaning up code and working on i18n | 19:15 |
beuno | that sounds great, you're an unstoppable machine of awesome | 19:16 |
beuno | I've had a few Canonical co-workers ask for for this app, and they said it was awesome | 19:16 |
mkarnicki | you are too kind, thank you :) | 19:16 |
mkarnicki | beuno: you mean, they have tested it \o/ ? | 19:16 |
mikeconcepts | wondering if the free level of ubuntu one permits sync of firefox bookmarks once the associated plugin is installed | 19:17 |
beuno | mkarnicki, yes they have | 19:17 |
beuno | mikeconcepts, it does | 19:17 |
beuno | but we have couchdb replication disabled at the moment | 19:17 |
mkarnicki | beuno: that's great ^ ^ I'm glad I can bring a little awesomeness to Ubuntu! | 19:17 |
beuno | as we finish upgrading it to a more scalable version | 19:18 |
mikeconcepts | also wondering if oneconf once installed will show inventory of connected computers | 19:18 |
beuno | good question, I don't know a lot about oneconf | 19:18 |
mkarnicki | mikeconcepts: btw, you can check status of services here https://wiki.ubuntu.com/UbuntuOne/Status | 19:18 |
mikeconcepts | currently it isn't showing inventories but folder sync is working just fine | 19:19 |
mikeconcepts | mkarnicki: thank you | 19:19 |
mkarnicki | mikeconcepts: you're welcome | 19:19 |
=== mkarnicki is now known as mkarnicki-afk | ||
=== jumpa is now known as JUMPA | ||
=== teknico is now known as teknico_away | ||
=== dendrobates is now known as dendro-afk | ||
=== dendro-afk is now known as dendrobates | ||
apachelogger | Chipaca: it is very likely that the Qt version will be in maverick | 20:20 |
apachelogger | api enhancements are really a no brainer | 20:20 |
Chipaca | apachelogger: ok. Should we ping you if/when the dbus api changes? | 20:20 |
Chipaca | apachelogger: well.... ... | 20:20 |
Chipaca | apachelogger: we might be using a plug+socket | 20:20 |
Chipaca | apachelogger: I know qt and gtk are supposed to interoperate well using a plug+socket, but have never seen it in practice afaik | 20:21 |
Chipaca | apachelogger: fingers crossed! | 20:21 |
apachelogger | Chipaca: well, IMHO the fdo secrets solution is very sane and portable ... using dbus and encryption ... from a portability POV I would also opt for dbus really | 20:23 |
Chipaca | apachelogger: not tracking you, sorry. What? | 20:23 |
apachelogger | instead of using a socket I would suggest enhancing the dbus api to become a bit like http://www.freedesktop.org/wiki/Specifications/secret-storage-spec | 20:24 |
apachelogger | which essentially would mean introducing sessions and handshaking an encryption type (or maybe just use a fixed one?) | 20:25 |
Chipaca | I don't think we have time right now, but it's an excellent point. nessita ^ | 20:25 |
Chipaca | apachelogger: but socket/plug is orthogonal to that | 20:25 |
Chipaca | apachelogger: it's to embed the ussoc in the client, for example | 20:26 |
apachelogger | well, I never ported a socket across operating systems, but I imagine it a lot less platform agnostic than dbus | 20:27 |
apachelogger | why would you embed ubuntu-sso in the client? | 20:27 |
nessita | Chipaca: reading | 20:27 |
nessita | Chipaca, I'm not following, could you please provide some more context? | 20:30 |
Chipaca | apachelogger: to be prettier? behave more like the user expects? | 20:31 |
dobey | i don't get it | 20:32 |
Chipaca | nessita: apachelogger says we should make our api follow a freedesktop spec more. I'm not sure if we're talking about the same thing, but wondered | 20:32 |
Chipaca | I haven't read the spec :) | 20:32 |
Chipaca | dobey: which it? | 20:32 |
dobey | well, the emedding bit, to start with | 20:32 |
Chipaca | dobey: well... design wants the login dialog to be *inside* the control panel or whichever app it comes up in first | 20:33 |
apachelogger | Chipaca: well it must not be the exact spec (since that will become a backend to ubuntu-sso at some point anyway) but I think the approach very sane and protable (not only across desktops but also operating systems) | 20:34 |
apachelogger | Chipaca: why do you need to embed sso though? or what particular do you mean by embedding? | 20:34 |
dobey | yeah i don't get that | 20:34 |
apachelogger | from my POV ubuntu-sso-client does the non-gui stuff and you can stack whatever gui you want ontop of it | 20:35 |
apachelogger | well, not really ontop of it, but attached to it ;) | 20:35 |
dobey | how does embedding it vs. just popping up the dialog correctly, make sense? | 20:35 |
Chipaca | apachelogger: ussoc is both things | 20:35 |
dobey | apachelogger: no, it does gui stuff. the whole point is that it's the gui for doing the log-in | 20:36 |
apachelogger | what gui do you have? | 20:36 |
dobey | if everyone has to create their own log-i ui, it's broken. the log-in ui should be the same for everything using it | 20:36 |
apachelogger | it really just shows the auth dialog? | 20:36 |
Chipaca | dobey: well, even simple things like popping the dialog up on top of the app that requests it (so the user sees it, instead of having to notice it blinking in the taskbar) | 20:36 |
dobey | Chipaca: i did say *correctly* for a reason :) | 20:37 |
apachelogger | Chipaca: for that one supposedly could send the xwinid via dbus? | 20:37 |
dobey | i don't think doing it wrong is a good excuse for doing it more wrong | 20:37 |
Chipaca | apachelogger: you can't create a gtk window with just a xwinid | 20:37 |
dobey | and plug/socket is generally discouraged in gnome | 20:38 |
Chipaca | see, you should've started there :) | 20:38 |
Chipaca | dobey: you're saying that disabling the control panel and waiting for the user to realize it's because they have the sso dialog waiting for their attention is better? | 20:38 |
dobey | well, a widget is a widget. if you don't make it plug/socket, you'll just make it a GtkWidget you embed instead :) | 20:38 |
dobey | Chipaca: no | 20:39 |
Chipaca | see, now *I'm* not following | 20:39 |
dobey | Chipaca: i'm saying if the user is waiting to notice that there is another dialog, you're doing it wrong :) | 20:39 |
nessita | apachelogger: a GTK one | 20:39 |
dobey | Chipaca: why isn't the window id of the parent being passed in over the API to whatever pops up the dialog? | 20:39 |
Chipaca | dobey: and what do we then do with the window id? | 20:40 |
Chipaca | set_transient_for doesn't want a window id | 20:40 |
Chipaca | it wants a gtkwindow | 20:40 |
Chipaca | and you can't build a gtkwindow with just a xwinid | 20:40 |
dobey | Chipaca: stop looking at GtkWindow :) | 20:40 |
dobey | Chipaca: you can set a parent with xwinid, if you use the lower level api to do it | 20:40 |
* Chipaca weeps | 20:40 | |
dobey | ie, gdk | 20:40 |
Chipaca | ah! | 20:40 |
Chipaca | whee | 20:41 |
dobey | so you build the dialog with NULL parent | 20:41 |
Chipaca | dobey: today, I hate that you give me the good news after the bad | 20:41 |
Chipaca | None, but I think it boils down to the same :) | 20:41 |
Chipaca | I sure hope this bit of the api isn't cuddled away cutely by pygtk | 20:41 |
dobey | set the transient parent via the gdk api, and then realize/show the window | 20:41 |
Chipaca | excellent | 20:42 |
Chipaca | dobey: and could this same level of trickery be used to actually add the dialog to a widget in the parent, if we're asked to do that next cycle? | 20:42 |
dobey | hmm, so in pygtk it might be harder to do :-/ | 20:44 |
dobey | gdk_window_set_transient_for() wants a GdkWindow as the parent | 20:44 |
dobey | ah, ok | 20:45 |
dobey | but you can do gtk.gdk.window_foreign_new(xwinid) | 20:45 |
Chipaca | no problem because | 20:45 |
Chipaca | yeah, exactly | 20:46 |
dobey | and that should give you a gtk.gdk.Window, to use for transient_for | 20:46 |
dobey | right | 20:46 |
Chipaca | nessita: alecu: are you guys following this, or do you want an executive/panicked hacker summary? | 20:46 |
dobey | Chipaca: re: embedding it as a widget in the parent is the same trickery, because that is how plug/socket works, really | 20:46 |
nessita | Chipaca: second option, starting on Monday? | 20:46 |
nessita | :-) | 20:46 |
Chipaca | it's the level of change I can do if you're too busy making things work, but I'd rather spend my time fixing emblems | 20:47 |
Chipaca | apachelogger: what you have in qt then is just the backend, not the actual dialog? | 20:47 |
alecu | Chipaca, not following | 20:47 |
dobey | it should never be embedded though. if it is, it's being done wrong. there is no reason for it to happen in that way for the user workflow | 20:48 |
Chipaca | dobey: I can show you designs, but I know it would kill you | 20:48 |
dobey | i've seen the designs | 20:48 |
Chipaca | alecu: we'll change the api to take a x window id, and if/when we show the dialog, we set_transient_for that window id using the low level api | 20:49 |
Chipaca | alecu: no socket/plug needed (phew) | 20:49 |
dobey | ux design should only dictate the behavior and experience, not how that experience is achieved | 20:49 |
dobey | :) | 20:49 |
apachelogger | Chipaca: yes | 20:50 |
apachelogger | so I still do not know what dialog you are having in the GTK version | 20:50 |
apachelogger | also | 20:50 |
nessita | apachelogger: wanna try it? | 20:50 |
apachelogger | by passing the xwindow id you can supposedly define a parent for the dialog | 20:50 |
nessita | apachelogger: bzr branch lp:ubuntu-sso-client | 20:50 |
apachelogger | hence let the window manager figure out sorting | 20:50 |
apachelogger | nessita: I am on a netbook right now | 20:51 |
nessita | apachelogger: ah | 20:51 |
Chipaca | apachelogger: how does your ubuntu-sso-client get the user details to talk to sso and get the tokens? | 20:51 |
apachelogger | one that is upgrading right now ^^ | 20:51 |
alecu | Chipaca, afaik sending a id thru dbus is all that's needed for plug/sock as well. | 20:51 |
apachelogger | Chipaca: oauth | 20:51 |
Chipaca | alecu: yes | 20:51 |
alecu | Chipaca, so it's pretty much the same | 20:51 |
=== mkarnicki-afk is now known as mkarnicki | ||
Chipaca | apachelogger: bananas | 20:51 |
apachelogger | client app requests token | 20:52 |
Chipaca | apachelogger: how does it *get* it | 20:52 |
apachelogger | my implementation goes looking for the token in every keyring it can find a plugin for | 20:52 |
Chipaca | apachelogger: and when it isn't there? | 20:52 |
apachelogger | if that fails it gets an oauth request token | 20:52 |
apachelogger | then opens a browser window with the auth | 20:52 |
apachelogger | then receives a callback with token (yes I implemented a http server in Qt :P) | 20:53 |
apachelogger | stores token in whatever keyring it found a plugin for and emits NewCredentials or whatever the dbus signal is | 20:53 |
Chipaca | ahhhh | 20:53 |
Chipaca | apachelogger: so you've rewritten ubuntuone-login, not ubuntu-sso-client | 20:53 |
Chipaca | you called it the latter just to confuse us | 20:54 |
Chipaca | sneaky | 20:54 |
apachelogger | from what I have ssen ubuntu-sso-client is ubuntuone-login with the ubuntuone stripped in all but the keyring reading and saving :P | 20:56 |
dobey | ugh, browser windows | 20:57 |
apachelogger | Chipaca: so what does ubuntu-sso-client do differently? | 20:57 |
dobey | apachelogger: it's migrating to the actual ubuntu sso stuff | 20:57 |
dobey | so it will do registration, etc.. without opening a browser | 20:57 |
dobey | (or may do already, i haven't been paying a lot of attention to it lately) | 20:58 |
apachelogger | is the web ui so crappy? | 20:58 |
dobey | yes | 20:58 |
apachelogger | why not fix that then? | 20:58 |
dobey | because going to the web is the wrong way to do things | 20:58 |
dobey | if the answer is "go to the web" why do we have a local client at all? :) | 20:58 |
mkarnicki | dobey: the API for browserless oauth is ready iirc | 20:59 |
mkarnicki | dobey: it serves a captcha (you probably know more then me anyway :) ) | 20:59 |
Chipaca | apachelogger: our biggest problem with new users is that the browser doesn't appear | 21:00 |
Chipaca | apachelogger: or is not noticed | 21:00 |
dobey | "oh hi chrome, you opened a tab on workspace #32 where i use the web from." | 21:01 |
dobey | "thanks for nothing" | 21:01 |
apachelogger | well, wether essentially duplicating strings from the web ui or not is a good idea is a topic of its own | 21:01 |
apachelogger | but I do not see the problem with defining a parent via xwindow id | 21:01 |
apachelogger | embedding is all sorts of nasty if the thing you are embedding is GUI and backend | 21:02 |
dobey | there is no problem | 21:03 |
dobey | i just explained that :) | 21:03 |
apachelogger | well then I vote for that + dbus only api ;) | 21:04 |
=== dendrobates is now known as dendro-afk | ||
=== dendro-afk is now known as dendrobates | ||
=== teknico_away is now known as teknico | ||
=== sonicbadger is now known as ChrisWoollard |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!