/srv/irclogs.ubuntu.com/2010/03/24/#ubuntu-desktop.txt

leftyfbSo has it been decided that the window control buttons are going to remain on the wrong side?02:51
bratscheCheck the bug report.03:08
bratscheOr the flamefest on the mailing list.03:08
leftyfbbratsche: I don't know the location of either of these03:09
bratscheGoogle does.03:09
leftyfbOh, I'm sorry. I was told this would be a helpful place to get these sort of answers.03:10
bratscheSorry, I guess I should have just not answered.  I don't know anything newer than what's in those two places, and I'm not interested enough to actually follow them.03:11
bratscheI guess I don't really understand why it's such a big deal, so I haven't been paying attention.03:12
lifelessbratsche: :P03:27
lifelessbratsche: its a big deal because its shallow and easy to think about03:27
lifelesscomplex things noone has an opinion on03:27
bratscheThis is why I'm not on jono's team.  I don't know how to deal with people. :)03:27
lifelessbratsche: can be taught03:28
Nafaibratsche: :)03:28
lifelessbratsche: I don't think you're bad at it :)03:28
bratscheI don't want to learn.  I hate people. :)03:28
* Nafai looks forward to meeting bratsche in person03:29
Nafai:)03:29
bratschehaha03:29
lifelessbratsche: haha LIES!03:39
pittiGood morning06:04
igoryonyahello, is there a way to do in gnome something similar to windows explorer extensions and where do I find that information?. I need to add some programs to context menu that will appear, when you right-click on the certain file or folder or always.06:36
Nafaiigoryonya: #ubuntu-app-devel might be a better channel for this question, but you'll want to look at nautilus extensions06:38
igoryonyaNafai: thanx06:39
robert_ancellpitti, can you have a quick look at a UI change in simple-scan (bug #543384) and see if that would be a problem06:44
ubottuLaunchpad bug 543384 in simple-scan "Toolbar confusing for new users" [Medium,Triaged] https://launchpad.net/bugs/54338406:44
robert_ancellpitti, also there's a number of grammar fixes for gbrainy that should be applied.  Will these be a big issue for a freeze break?06:45
pittirobert_ancell: hey, how are you?06:48
robert_ancellpitti, hey, good :)06:49
pittirobert_ancell: gbrainy> is it possible to apply the grammar fixes in all the .po files as well, to avoid breaking translations?06:49
robert_ancellpitti, as in take the .po changes from upstream?06:49
pittirobert_ancell: ah, did they fix it there?06:50
pitti(or "you"?)06:50
robert_ancellpitti, yeah, the fixes are coming because gbrainy is getting a lot of testing from being in the betas.  I wanted to review all the puzzles for Lucid as the author is not a native english speaker but that got skipped when I went to OEM.  So he's basically been fixing like crazy based on feedback from Ubuntu users06:51
pittisweet :)06:51
pittirobert_ancell: so, grammar fixes are alright, it'd just be nice to not just do them in the code, but also in the .po files06:51
pitti(upstream, preferably; we don't want to break it for everyone else)06:52
robert_ancellpitti, np, I'll just make a big patch with meld and open a bug report for the freeze request06:52
pittirobert_ancell: simple-scan> replied to the bug (easier to keep discussion there)06:54
didrocksgood morning07:06
pittibonjour didrocks07:08
didrocksGuten Morgen pitti07:09
didrockspitti: Albin pinged me about bug #538680, can you have a look at it when you have some time? (he is the DD responsible of that package and a MOTU too)07:12
ubottuLaunchpad bug 538680 in e17 "Please remove e17 from the archive" [Undecided,Confirmed] https://launchpad.net/bugs/53868007:12
pittididrocks: sounds like ubuntu-archive should be subscribed to that one?07:13
pittididrocks: ah, it is already; seems sensible07:13
didrockspitti: right, he's just pinging me to ensure it's done by beta 2, I guess :)07:14
didrocksthanks pitti :)07:38
pittide rien07:38
seb128hello!08:46
pittibonjour seb12808:46
pittiseb128: after fixing control-center, I don't want to hear about keyboard layouts again. EVAR! :-)08:49
seb128lol08:50
* seb128 hugs pitti for winning the great keyboard battle08:50
pittididrocks: new p-distutils-e uploaded, with the buffering fix, FYI08:52
seb128pitti, I did subscribe you to some extra gvfs automount bugs btw, hope it's ok08:53
didrockspitti: oh sweet!08:53
seb128lut didrocks08:53
didrockshey seb128 ;)08:53
pittiafter those two, that should get me on #2 again, one more than kirkland :-P08:53
seb128pitti, I understand the stack now but I'm still not sure what to look for in the logs08:53
pittiseb128: absolutely; I saw them, will look at them soon08:53
pittiseb128: also, automounting is my pet project, I'm happy to do those bugs08:53
seb128ie I don't see what is wrong08:53
seb128didrocks, sorry about gtk08:54
seb128didrocks, I did the update, I rebased bzr though08:54
seb128didrocks, but you might have to refresh some patches now08:54
didrocksseb128: yeah, I saw, no pb, will do08:54
seb128thanks08:54
didrocksseb128: can you NEW atk and pango when you have a sec, please?08:54
seb128yes, I was just thinking about it while writting there ;-)08:55
seb128I forgot yesterday08:55
didrocksno pb ;)08:55
seb128didrocks, had a good night otherwise?08:55
didrocksI'm finishing/testing the gdm mess and then, jump into that08:55
seb128cool08:55
didrocksseb128: oh yes, a needed one :)08:55
didrocksseb128: and you?08:55
seb128lut davidbarth08:55
seb128didrocks, quite good thanks08:55
seb128didrocks, I managed to get to bed a bit late but not very late ;-)08:55
pittiseb128, kenvandine: I'm sorry if that sounds dumb, but my messaging indicator now has "Broadcast"; that seems to open gwibber; I've heard "twittering" and "microblogging", which seem to be the usual terms for it; but "broadcast"??08:56
didrocksseb128: I had a look at precisely that this morning. But as I told "I'm off now" at 8PM and finally left at 10PM, I can't say anything on that :p08:56
didrockspitti: in French, it sounds better (it was an unexisting word before "microbloggage"). Sounds ugly :)08:57
seb128pitti, we discussed that with mvo and ted yesterday too08:59
seb128pitti, "twitter" is copyrighted08:59
seb128we can't use it08:59
pittiand microblogging?08:59
mvoI got feedback like this too from upgrade tests08:59
seb128that's what we used before08:59
mvobroadcast is really confusing08:59
seb128not sure why design decided on broadcast rather than microblogging08:59
seb128I would prefer microblogging too for what is worth09:00
seb128it's probably feedback to bring to sabdfl or mpt though09:00
* mvo is currently in the world-of-pain of bzr format incompatbilities and in a bad mood because of that09:01
pittimvo: bzr upgrade?09:02
mvopitti: webui upgrade does not work, bzr+ssh upgrade does not work, I'm trying sftp upgrae now, for each upgrade it downloads 60mb of data before it fails09:03
mvobzr+ssh fails with "backup.bzr" exists *even* if I deleted it before via nautilus (that takes 5min as well)09:03
pittiyes, the webui upgrade doesn't work for me either09:04
pittibut rm -r backup.bzr in lftp (or nautilus) and bzr upgrade bzr+ssh:// usually does09:04
pittihmm09:04
pittiso I bought an Album from the music store, and "My downloads" says "showing 1 to 0 of 0 downloads"09:05
mvonot for me, I asked in #bzr already and no luck. oh well, I will keep trying. bzr and I used to be best buddies, but this is a real dent in our relationship09:05
* pitti hugs mvo09:07
vishmvo: hi.. could you have a look at the patch in Bug #393358 ... [, it's a small/simple patch ]09:07
ubottuLaunchpad bug 393358 in update-manager "Synaptic/Update manager show weird download dialogue towards end of download" [Undecided,Confirmed] https://launchpad.net/bugs/39335809:07
didrocksok, just need to connect gdmsetup to the dbus signal and I don't want to hear this damned sound ever :)09:08
seb128didrocks, :-)09:08
mvovish: sure, will do that now09:10
vishthanks.09:10
mvovish: hm, its a string change, we need a UI freeze exception for it09:11
vishhmm.. :s   , i should probably have pinged you earlier :(09:11
mvovish: otherwise its trivial, I can import it into the ubuntu-ui-changes branch that I created to accumulate fixes like this (so that we only need one UI freeze exception)09:11
mvovish: sorry from me too, I'm not great at triage09:12
vishmvo: that would be awesome  :)   do all the accumulated fixes get UIffe at the same time or each bug needs to be dealt separate?09:13
mvovish: I think one "big" one, its not that much that I have currently, so I hope it will be ok09:15
vishneat..09:16
pittiseb128: did you already dist-upgrade and restart session today?09:32
seb128pitti, the mini yes, not my laptop09:34
seb128pitti, why?09:34
pittiseb128: if you have a second, mind to connect your ipod and see what happens?09:35
pittipreviously you got two dialogs (gphoto and afc), right?09:35
seb128pitti, oh yes, wanted to check that09:35
seb128brb door bell09:36
seb128re09:40
seb128easier than that09:40
seb128let's connect the ipod to the mini now09:40
seb128pitti, right, I only get 2 instead of 3 now09:41
seb128and the rhythmbox change to not consider afc devices as mtp devices reading the new udev properties works too09:41
seb128sweet09:41
pittiseb128: one for camera, one for music?09:42
pittiok, that's back to the old problem then09:42
seb128yes09:42
seb128what old problem?09:42
seb128I'm not even sure it can be considered buggy09:42
pittinot really buggy, just a little confusing09:43
seb128it's slightly, but technically the ipod touch has the camera mtp device09:43
pittiI don't have a better idea either09:43
pittisame for my G1; after all, it is a camera _and_ a music player09:43
seb128it just lacks the optic device09:43
seb128well you do have a camero on the g1 though?09:43
seb128like the iphone09:43
pittiit'd be slightly nicer to merge those two dialogs into one09:44
seb128the touch doesn't have the electronic part for it09:44
pittiseb128: yes, I do have a camera there09:44
seb128but it has the same software stack09:44
seb128the photo import dialog works too09:44
seb128you can only import screenshots though09:44
glatzormvo, hello09:58
glatzormvo, I just pushed debconf unittests to aptdaemon's main branch09:58
glatzormvo, my latest patch seems to catch the race conditions09:59
mvoglatzor: woah, you *rock*09:59
* mvo hugs glatzor very hard09:59
mvoglatzor: I will update the branch after lunch (need to finish some other stuff first)10:00
didrocksseb128: is there a specific bug for the sound thing in gdm?10:03
seb128didrocks, yes10:03
seb128it's on the gdmsetup spec wait10:04
didrocksI don't find it on https://blueprints.edge.launchpad.net/ubuntu/+spec/desktop-lucid-gdmsetup10:04
seb128didrocks, look on the right side10:05
seb128"ug #437429: No GUI to configure/disable login sound"10:05
didrocksseb128: oh right, never noticed you can link blueprint and bugs10:05
seb128"Bug #437429: No GUI to configure/disable login sound"10:05
didrocksseb128: sweet, thanks10:05
ubottuLaunchpad bug 437429 in ubuntu-sounds "No GUI to configure/disable login sound" [Undecided,Invalid] https://launchpad.net/bugs/43742910:05
seb128didrocks, np10:05
Nghrm, after this morning's updates rhythmbox has broken for me, and starting it (from docky in this case) has left me with a weird situation where it keeps starting and crashing and somehow starting again10:18
NgI've killed docky and it's still starting and crashing every 2s10:21
Ngwhat on earth is restarting it?!10:21
Ngoh no, this is going to be dbus, isn't it10:21
Ngis there any way for me to inspect what's trying to talk to rhythmbox's dbus service?10:23
Ngeither that thing doggedly keeps trying, or dbus isn't sensible to back off services that don't start correctly10:23
Ngit's not the indicator apple, or u1-syncdaemon10:28
Nghuh, unplugging my iphone let it start up correctly10:29
seb128Ng, run in a command line and see if it crashes?10:30
pittiseb128: btw, are you aware of a bug about the unmount/eject/save removal menu cleanup? if not, I'll create one10:32
seb128pitti, there is one about having 3 options being confused10:32
seb128with a zillion comment10:32
pittiah, that sounds appropriate10:32
seb128and a hundredpapercut task10:32
seb128let me get you the number10:33
Ngseb128: annoyingly it started just fine with -d both with the phone already plugged in, or with plugging it in after rb started, but I have a crashdump to feed to LP10:33
pittiseb128: bug 453072 ?10:33
ubottuLaunchpad bug 453072 in gvfs "Context menu for an USB pendrive shows "Unmount", "Eject" and "Safely Remove Drive"" [Low,Triaged] https://launchpad.net/bugs/45307210:33
seb128pitti, yes10:33
pittiseb128: heh, incidentally that was the first hit when I wanted to file a new bug; LP search is pretty good these days10:33
seb128hehe10:34
* pitti targets to lucid and assigns10:34
Ngseb128: fwiw launchpad suggested a bunch of dupes, 5 of which were opened in the last 14 hours and they all seem to relate to iphones or new ipod touches, so I'll dupe&confirm :)10:35
seb128Ng, ok thanks10:35
seb128Ng, can you give the bug number?10:36
Ngseb128: the lowest number is bug #545077. they're mostly all still awaiting retrace, so still private10:36
ubottuError: Could not parse data returned by Launchpad: list index out of range (https://launchpad.net/bugs/545077)10:36
* Ng grumbles at his LP duck10:36
seb128I bet they crashed again10:37
* seb128 looks10:37
seb128indeed10:37
* seb128 removes lock10:37
seb128lazr.restfulclient.errors.HTTPError: HTTP Error 500: Internal Server Error10:37
seb128they keep crashing on that10:37
seb128wououh, didrocks rocks10:38
seb128didrocks, congrats on getting the gdm change in!10:38
didrocksseb128: heh thanks, *finally* :)10:38
seb128didrocks, can you email the translators about the string added too? or talk to dpm about emailing those10:39
didrocksseb128: sure, was on my todo, just finish to launch a gtk build in the meantime (better to do other things when gtk is building) :)10:39
seb128;-)10:39
* seb128 grabs coffee when gtk is building usually10:40
seb128speaking of which10:40
seb128brb grabbing some coffee now10:40
didrocksseb128: a long coffee, so :)10:40
seb128lol10:40
didrocksheh10:40
seb128didrocks, pitti: do you use gthumb usually? do you consider 2.11 as unstable?10:48
pittiseb128: I have used it once only since we upgraded 2.11 (I don't do a lot of photos these days), and it was pretty buggy10:49
pittiQ-FUNK pointed out some good points yesterday10:49
seb128ok, I sort of not trust him though so I would other opinion before downgrading version10:49
pittiseb128: unless you object, I'd like to do the downgrade and instead fix the libgphoto unmounting10:49
pitti(if it was broken)10:49
seb128especially we should sync on debian first, it seems it would fix stability issues10:49
seb128well I've no strong opinion either way, just that I don't trust him on such things10:50
pittiseb128: we do have the latest Debian version already10:50
seb128he often argue thinking his workflow is the only one which counts10:50
seb128ok10:50
seb128that was supposed to fix some of the crashers10:50
seb128I don't know enough about the changes to judge otherwise10:51
didrocksseb128: pitti: if we downgrade, we should still fix the gvfs patch for gthumb so that we can import photo10:51
pittididrocks: yes, what I said above10:51
pitti2.10 had been rock solid10:51
baptistemmgood morning10:51
seb1282.11 is not?10:52
seb128it uses recent technologies and is maintained10:52
seb128but we can probably delay getting it to next cycle10:52
pittireportedly it dropped a lot of features and has a half-written gio port10:52
pittiseb128: I can play around with it a little more later today10:52
seb128it's in universe10:52
dpmseb128, ok, I'll take care of that (gdmsetup string), didrocks just talked to me10:53
seb128I think we should let motu deal with it10:53
seb128dpm, hey10:53
seb128dpm, thanks!10:53
pittiuniverse> yes, but it's sooo much better than f-spot :-P10:53
* pitti ducks10:53
seb128lol10:53
seb128you try shotwell10:53
dpmno worries, thanks for the heads up guys10:53
seb128that's the futur apparently ;-)10:53
pittiseb128: I didn't, no; you?10:53
seb128I did10:53
seb128it's quite nice10:53
seb128it's written in vala too10:54
* pitti installs10:54
pittiok, so (1) fix gvfs, (2) get lunch, (3) try shotwell, (4) check gthumb10:54
seb128;-)10:55
seb128gvfs is broken?10:55
pittiargh, it has again this "import photos" thing?10:55
pittiwhy can't I just point programs to "this is my photo dir"10:56
pittiah, it just thumbnails10:56
pittiseb128: so, if it would use ~/.cache instead of ~/.shotwell, the import seems quite okay :)10:56
pittiseb128: hm, it seems to lump them together in one big list?10:57
pittiok, seems this still isn't for me then -- I organize by directory :(10:58
pittiseb128: thanks for pointing out10:58
pittiseb128: gvfs> bug 48152810:58
ubottuLaunchpad bug 481528 in gvfs "Volumes with photo MIME type do not have a camera icon" [Low,In progress] https://launchpad.net/bugs/48152810:58
seb128Ng, the crash should be fixed with today updates10:59
pittiseb128: trying to clear my +assignedbugs list, and this should be easy to fix10:59
seb128Ng, it was an issue with old usmuxd10:59
seb128the new one set the property required10:59
seb128pitti, ah nice10:59
pittiseb128: ugh, gvfs crashes with old usbmuxd?11:00
seb128pitti, one list + by date using tags11:00
pittithat sounds like a bug11:00
seb128pitti, rhythmbox does11:00
seb128pitti, it's a bug, teuf fixed it in git11:00
pittiah, good11:00
seb128pitti, I was wondering why it doesn't happen there11:00
cassidyseb128, you probably one tp-sofiasip 0.6.2 in Lucid; it workarounds an Empathy freeze : https://bugzilla.gnome.org/show_bug.cgi?id=61246011:00
ubottuGnome bug 612460 in VoIP "Empathy freezes on attempting to place a SIP call" [Critical,Resolved: notgnome]11:00
cassidys/one/want11:00
seb128pitti, that's because it was happening only with the udev rules you added yesterday11:00
pittiseb128: we shouldn't count on devices having udev properties; but the check looked quite robust (in gvfs)11:00
seb128cassidy, thanks11:00
cassidyseb128, I think it's packaged in Debian11:01
seb128pitti, right, http://git.gnome.org/browse/rhythmbox/commit/?id=5a4c30a2b1c939b48eb1a8f86ebe95b052d5ad0311:01
pittiseb128: ah, I see; so it crashed _because_  it was now using the correct backend? :-)11:01
pittisweet11:01
seb128cassidy, ok good11:02
pittididrocks: *hug* for gdmsetup; great job!11:06
* pitti hugs seb128 as well11:06
* didrocks hugs pitti11:07
* seb128 hugs didrocks pitti too11:07
* didrocks hugs seb128 too ;)11:08
Ngseb128: hrm, maybe usbmuxd was already running because my phone was plugged in all morning, but I had already updated11:08
ThraulI am trying to remaster a ubuntu desktop cd for personal use. I can make small changes such as desktop background, theme etc on the boot (livecd), now how do I edit the menus and add desktop icons to the live cd?11:08
Ngseb128: thanks :)11:08
seb128Ng, np11:10
seb128Thraul, hi, not really a question for this channel, try #ubuntu11:10
Thraulok thanks11:11
aquariuspitti, ping11:11
seb128hey aquarius11:12
pittiaquarius: please don't ping, just ask the question11:12
pittiaquarius: hello, how are you?11:12
aquariuspitti, heya, wasn't sure if you were around :) You deliberately bought a WMA album despite the warning?11:13
pittiaquarius: well, it said "we do not recommend", not "it's broken"; it sounded like you'd rather want to avoid trouble with codecs?11:13
aquariusYou have uncovered a bug by doing this, which is that the back end doesn't show you WMA files because it assumes no-one will buy them :P11:13
pittiaquarius: that album was only available as WMA..11:13
seb128james_w, there?11:13
pittiaquarius: ah, heh :)11:14
seb128james_w, I need your help on a bzr issue if you are around and have a minute to help me on this one11:14
aquariusgood catch :)11:14
pittiaquarius: I was going to download it and convert to ogg11:14
pittiI don't hate mp3 enough to go through the hassle/quality loss of reconverting them, but I convert everything else11:14
=== MacSlow is now known as MacSlow|lunch
aquariuspitti, unfortunately that album will remain in limbo and undownloadable until that bug is fixed :(11:18
aquariushey seb12811:18
pittiaquarius: not a big deal; it's not _that_ urgent :)11:19
aquariusthank you for being so considerate. Most people are not :)11:19
pittiaquarius: my primary goal was to test the music store, and Katie Melua was one of the first artists I thought of (my Pink Floyd collection is already complete..)11:19
pittiaquarius: happy to be testing :) the store looks great really, and nice11:20
pitti... and large choice11:20
pittiaquarius: they even have "Wise Guys", a relatively small German A-Capella band11:20
seb128aquarius, how is feedback looking so far for the store?11:20
aquariuspitti, cool! The German store is one of the three complete stores :)11:21
pittiaquarius: while I'm at whining; do you guys plan to add a progress dialog of some sort? when I click on "U1 music store", nothing happens for about 20 or 30 secs, then I finally see it11:21
aquariusseb128, so far...pretty good, actually. The bugs that have come in are either largely cosmetic or they're around the inegration with logging in11:21
seb128cool11:21
aquariuspitti, yep. Already done in dev. When a package upload request with a UI freeze break comes through from kenvandine, you'll know why ;)11:22
pittisweet11:22
pittiaquarius: jamendo etc. have a "loading song list..." or so in the standard progress bar in RB, AFAIR11:22
seb128"    - Show a temporary view when loading the initial store page (LP: #530247)"11:22
seb128was uploaded yesterday11:22
seb128in libubuntuone11:22
pittihm, I have that11:23
pittioh, it's just a small text, I might have missed it11:23
aquariusah, no, there's two aspects to the loading stuff. one is the temporary view, which is now in; the other is a proper loading indicator, which we have mockups for but no code yet11:26
didrocksseb128: thanks for the NEWING btw. New gtk uploaded in ubuntu-desktop team ppa11:47
seb128didrocks, np; thanks, any reason to not upload to lucid now?11:47
seb128didrocks, you want to to review the change before?11:47
seb128hum11:48
didrocksseb128: not sure if enabling introspection can break something, but at least having a test on clutter and clutter-gtk will be good11:48
seb128how do people download tarballs with merge-upstream?11:48
seb128in the old workflow I used to bzr merge lp:11:48
seb128dch -11:48
seb128bzr-buildpackage11:48
seb128didrocks, ok11:49
didrocksseb128: bzr bd -e or just running uscan11:49
didrockswould be good if merge-upstream can do that for us so that we just have to point to upstream bzr branch, nothing else :)11:50
seb128right11:50
seb128didrocks, well bzr bd -e relies on the changelog no?11:52
seb128but if you update the changelog before merging it doesn't work11:52
seb128it doesn't merge if there are uncommited changes11:52
seb128or you need to update changelog, commit and merge11:53
didrocksseb128: oh right, uscan so11:58
seb128righto11:58
seb128time for lunch11:58
seb128bbl11:58
didrockssame here, bll :)11:58
didrocksbbl*11:58
asac_bryceh: http://pastebin.com/FzYXbnkW ... should i commit that anywhere before uploading?12:07
tjaaltonasac_: there's 1.7.6-1u1 in the works (git.d.o)12:24
asac_tjaalton: can you try to apply the patch there?12:26
asac_and see if it still applies cleanly?12:26
* asac_ would hope so12:26
tjaaltonasac_: if you are in a hurry you can upload it as-is, and we'll make sure it applies on top of the new version (no reason why it wouldn't). rename the patch though, "111" is available :)12:33
tjaaltonasac_: also, send it upstream http://wiki.x.org/wiki/Development/Documentation/SubmittingPatches12:34
kenvandinegood morning folks12:34
seb128hey kenvandine12:34
asac_tjaalton: its not ready for upstream (see changelog)12:34
seb128kenvandine, I'm updating indicator-me just for information12:35
kenvandineseb128, thanks for the sponsoring12:35
seb128kenvandine, np12:35
kenvandinecool, thx12:35
asac_tjaalton: whats the problem with 250 ;)12:35
seb128kenvandine, I'm switching to the ubuntu namespace too12:35
tjaaltonasac_: bloat! :)12:35
seb128kenvandine, so make sure you down pull from the ubuntu-desktop one next time12:35
kenvandineseb128, will do12:35
kenvandinethx12:35
seb128np12:35
asac_tjaalton: i usually add stuff on top12:35
asac_when using 111 it means i inject it in the middle ;)12:35
tjaaltonasac_: so 20112:35
seb128I will update the changelog with a warning too in the old location12:35
seb128to avoid errors12:36
asac_tjaalton: ok. that makes more send ;)12:36
asac_sense12:36
tjaaltonprobably should just drop the numbers12:36
tjaaltonthey are meaningless anyway12:36
tjaaltonbut after lucid12:37
asac_yeah12:37
asac_ok renamed ... also using - for the word splits and added to series which i previously forgot12:37
asac_will push12:37
tjaaltonok12:37
=== MacSlow|lunch is now known as MacSlow
LaneyDo you guys think it's a bug that http_proxy isn't set in the graphical session? ie when launching apps from the gnome menu or gnome-do13:11
pittiLaney: it sounds like a bug; but last time I tried this, it worked for me13:12
pitti(when configuring the proxy with the capplet)13:12
Laneyyes, I used that13:12
seb128launchpad has zillion of proxy bugs listed13:13
LaneyI tried to look on bugzilla13:13
seb128look to gnome-control-center bugs13:13
seb128not sure how much our code is different from upstream there13:13
seb128there is quite some changes from mvo to set system values13:14
LaneyI guess gnome-terminal reads from gconf to export it there, right?13:14
mvoLaney: yep13:15
mvoseb128: I did not touch the general mechanism, just the "write stuff out to system", so there should be no change in behavior13:15
Laneymvo: Is there a bug here then?13:16
mvoLaney: my understanding is that the gui apps are responsible to get the proxy via gconf (unless libproxy is used)13:17
mvoLaney: stuff in the terminal is lucky because gnome-terminal sets http_proxy based on gconf13:17
Laneyoh, that's annoying behaviour then13:17
Laneyhttp_proxy would be a good interface13:17
Laneyalthough I guess there are corner cases with no_proxy13:18
seb128mvo, oh, good to know13:19
mvowell…13:19
mvoI guess it could be exported session wide as http_proxy13:20
mvobut that is not how its done currently13:20
LaneySo the reason I'm looking at this is because I noticed that docky doesn't work behind a proxy13:20
Laneyand that's because mono's DefaultWebProxy implementation relies on the env vars13:20
mvoheh :)13:24
mvolibproxy seems to be what is meant to solve the issue13:24
Laneyindeed13:26
Laneybut it's not part of gnome, and consequently isn't used very much13:26
Laneyis exporting the env vars to the session something that can be done in lucid?13:27
Laneyor does it require more discussion?13:27
mvothe disadvantage is that its much less dynamic, the env would be set at session startup and set in stone13:30
mvogconf is much better in that regard13:30
mvoso it would have to be "a) look at gconf and use that b) look at http_proxy env and use that"13:30
Laneycould it not be updated when the user edits it?13:30
mvothe advantage of course is that it would make all apps suddenly work, even if they do not look at gconf13:31
mclasenLaney: libsoup uses libproxy13:31
mvoit would only work for newly started apps if anything, not for already running ones. so I guess its not that useful13:32
LaneyI would argue that working somewhat is better than working never13:32
Laneyat least until libproxy is widely used13:33
Laney...needs the mono bindings enabled too :)13:33
Laneythis non-dynamic behaviour is what broken apps experience when launched from the terminal anyway13:34
mvotrue, there are corner cases however. its not uncommon that http_proxy takes precedence over gconf in a app (so that user can run "http_proxy=some-proxy" gedit). those would break badly13:35
mvowell, not badly, but break when dyanmically switching the proxy settings13:35
mvoso my take is that its not appropriate (at this point) for lucid and needs discussion13:36
Laneyalright13:36
mvosorry13:36
Laneythere is some code in f-spot to read from gconf anyway that I can copy to docky13:36
mvocool13:37
mvoif that could travel into mono itself, that would rock13:37
mvo(not sure if mono can dynamically import gconf easily and deal with the case when its not there, but I guess that is possible)13:37
Laneyyeah I don't know how that would work13:38
Laneyit would need to be generalised too to have proper events13:38
seb128pedro_, hey13:43
pedro_hey hey seb12813:43
seb128pedro_, how are you?13:43
pedro_seb128, things are going good, thanks. what about you?13:44
mvohey pedro_13:44
pedro_hola mvo!13:44
seb128pedro_, good, thank you13:44
seb128pedro_, thanks for picking those lucid bugs you assign to our team btw13:45
pedro_seb128, you're welcome, i've been reviewing some on the telepathy stack, we were missing some that were filed at the pymsn product13:46
pedro_i've reassigned those to papyon now13:46
seb128pedro_, speaking of those, the telepathy-butterfly ones, could you check if those are crashes or just apport noise?13:46
seb128we probably not need to put high for those which are just noise13:46
pedro_seb128, by apport noise you mean the ones that generate a stacktrace but doesn't crash the application (empathy) ?13:47
pedro_because most of them are like that, they don't crash empathy but they do generate a stacktrace13:48
seb128pedro_, yes, I guess you reassign and set high due to duplicates?13:49
seb128ie the number of duplicates those get13:49
pedro_seb128, correct, based on number of dups and users affected13:49
seb128but if those don't have visible effect that's not high since apport doesn't run on stable13:49
seb128you see what I mean?13:49
pedro_seb128, yes, got it13:50
seb128pedro_, thanks13:52
seb128pedro_, oh, other thing13:52
pedro_you're welcome13:52
seb128pedro_, I got the impression by reading bug emails than gedit is crashing in lucid, did you notice that too?13:52
seb128would be good to upstream those of those13:53
pedro_seb128, are you reading my computer screen?13:53
seb128lol13:53
pedro_seb128, i'm in fact looking at some gedit crashes now hahah13:53
* pedro_ looks around13:53
* seb128 hugs pedro_13:53
* pedro_ hugs seb128 back13:53
pedro_don't spy me! :-P13:53
seb128you got me, I'm spying on you13:53
seb128lol13:54
pittihey pedro_, how are you?14:00
pedro_pitti, hello! I'm good, thanks. how's everything going for you?14:01
pedro_morning rickspencer314:01
rickspencer3good morning pedro_14:01
seb128hey rickspencer314:01
pedro_is raof looking at f-spot bugs?14:01
=== robbiew_ is now known as robbiew
pokyieHi All14:02
seb128pedro_, not sure but he got the milestoned one assigned now, why?14:03
pokyieDoes anyone got this files in their machine14:03
rickspencer3hi seb12814:03
pokyie1. /lib/firmware/iwlwifi-1000-1.ucode14:03
seb128pedro_, he knows the f-spot code a lib14:03
pokyie2. /lib/firmware/iwlwifi-1000-2.ucode14:03
pedro_seb128, ok good, just to confirm that someone is actually looking at the issue ;-)14:03
pedro_thanks seb12814:03
nigelbpitti: thanks (g-p-m string bug)14:12
baptistemmHi thre14:14
baptistemmI would be interested to have bug 545551 fixed for lucid to be able to have some hints to triage bluez and bluetooth related bugs14:15
ubottuLaunchpad bug 545551 in bluez "apport hook for bluez" [Medium,New] https://launchpad.net/bugs/54555114:15
nigelbbaptistemm: I can do it :)14:18
nigelbwhat information do you want the hook to collect?14:19
baptistemmI think the hook is quite complete now14:19
nigelbah :)14:20
nigelbbaptistemm: the hook is supposed to be in your package I think14:20
nigelbi.e. the package which its supposed to report bugs on14:20
pittibaptistemm: hm, why is it an apport branch and not a bluez one?14:20
arandIs there a specific change in indicator-applet/application that got rid of the tooltips? I've been looking at changelogs with little luck...14:20
Laneydid it ever have them?14:21
Laneyi thought that was a design feature14:21
pittibaptistemm:14:21
pitti1914:21
pitti    if command_available('rfkill'):14:22
pitti2014:22
pitticommand_available('rfkill list')14:22
pittibaptistemm: I think you mean command_output() here :) (same for getfacl)14:22
baptistemmdoh14:22
nigelbpitti: what it into debdiff?14:23
nigelbs/what/want14:23
pittibaptistemm: and you don't need to call attach_conffiles for a package in its own package hook, that's already done automatically (in general-hooks/ubuntu.py)14:23
arandLaney: hmm, looking at it in karmic... you're right, maybe not.14:24
pittinigelb: stuffing it into the bluez package is trivial; I mean, if you want to (with nice changelog and dh_install and all that), as a packaging exercise, then please go ahead of course :)14:24
nigelbpitti: packaging exercise was the intention :) will do :)14:24
pittinigelb: ok, then I'll review it with that in mind (i. e. point out errors and let you fix them instead of doing that myself)14:25
pittinigelb: thanks14:25
pittinigelb: (but please coordinate with baptistemm to fix above two issues first)14:25
nigelbthanks for the chance ;)14:25
nigelbyes14:25
baptistemmpitti, fixed in my branch, thanks for the review14:25
nigelbbaptistemm: fixed? shall I package?14:26
pittibaptistemm: otherwise this looks fine14:26
pittibaptistemm: for bonus, you could do the last ui.information a bit differently14:26
baptistemmnigelb, if you're interested there is bluez 4.62 waiting sponsoring as well :)14:26
nigelbbaptistemm: I'm not a sponsor, yet14:26
pittinigelb: ok, I subscribed to the bug now; so just attach it there, and I'll followup14:28
nigelbpitti: okay :)14:28
nigelbpitti: unrelated, but does apport automatically compress the logs?14:38
baptistemmnigelb, If you need help for packaging (even thought I'm quite a newbie) don't hesitate to ping me14:40
nigelbbaptistemm: sure :)14:40
baptistemmpitti, you meant I should change the wording, or something else?14:41
pittinigelb: binary ones yes, text ones not automatically14:41
nigelbpitti: aha, anyway to disable it?14:42
pittibaptistemm: oh, sorry; I meant to write a followup, and hten realized that it doesn't work yet14:43
pittibaptistemm: there's root_command_output(), but that doesn't work as a backgroud "monitor" thing, just synchronously14:43
pittibaptistemm: I was going to suggest starting the log, then asking the user to reproduce and click OK, and stoppin git14:43
pittinigelb: disable compression?14:43
nigelbpitti: is it better left in? (debug log for cheese) if not, how to disable?14:44
pittinigelb: text attachments aren't compressed14:44
nigelboh, so make it .txt file, ah14:45
pittinigelb: hang on, I think we are talking past each other14:45
pittiit doesn't care at all about the file name, it cares about contents14:45
pittinigelb: what does currently happen, and what do you want to happen?14:45
nigelbin bug 545079, the debug log gets attaached as "HardwareInformation.log.gz" (naming mistake) is it because I gave the file name .log that it ended up as .gz?14:46
ubottuLaunchpad bug 545079 in libgphoto2 "Cheese stopped working" [Undecided,New] https://launchpad.net/bugs/54507914:47
pittinigelb: ah, because it has binary stuff in it14:48
nigelbpitti: ah, so leave it alone?14:48
pittinigelb: which hook does that come from?14:49
* pitti doesn't see/have a cheese hook14:49
nigelbpitti: I'm writing it14:49
pittioh, I don't have cheese installed14:49
pittiseems we don't do that by default on desktop14:49
baptistemmpitti, is ther a way to call an apport hook from another package, for instance gnome-bluetooth triggering bluez apport hook14:49
pittinigelb: so you should be able to force it, hang on14:49
didrockspitti: no, it's only on the netbook nowdays14:49
pittinigelb: what do you call right now for this particular field?14:49
nigelbCheeseDebug14:49
pittinigelb: I mean, you are using r['HardwareInformation.log'] = command_output() or so?14:50
nigelbpitti: http://pastebin.com/AyWxUhgE14:50
nigelbpitti: report["CheeseDebug"] = apport.hookutils.command_output(['env', 'GST_DEBUG=*cheese*:3', 'cheese', '-v'])14:50
pittibaptistemm: not easily, no; sorry14:51
pittinigelb: ah, hmm14:51
pittinigelb: you can force the non-compression of files, but not of arbitrary string values14:51
pittinigelb: so you could theoretically output it to a file and then do14:51
baptistemmpitti, can a user trigger the hook by itself to sent the information to a bug number ?14:52
pittireport['CheeseDebug'] = ('/tmp/mylog', False)14:52
pittinigelb: the False forces it to not get encoded/compressed14:52
pittibaptistemm: yes, apport-collect -p bluez 1234514:52
pittibaptistemm: (or you reassign the bug to bluez, then apport-collect 12345 will do)14:52
nigelbpitti: so add a line below the previous line?14:52
baptistemmapport-collect -p bluez 12345 will do the trick14:53
baptistemmthanks14:53
pittinigelb: well, you have to change command_output, too; you want to use subprocess.Popen to have it output to a temporary file14:53
nigelbpitti: ah, too much complication.  I'll leave it as gz ;)14:53
pittinigelb: :)14:53
* vish pokes nigelb to get it right ;p14:54
nigelbvish: its nonessential (my belief)14:54
vishnigelb: well , it makes life easier than having to download the log.. and you aernt going to be redoing this again ;)14:55
nigelbvish: okay okay14:55
* nigelb adds to long list of things to do14:55
* kenvandine is really amazed how fast lucid boots in kvm14:55
* baptistemm is really amazed how fast lucid boots on ssd :)14:56
nigelblucid doesn't boot in vbox for me at all (with testdrive)14:56
* baptistemm is really amazed how fast lucid boots on kvm on ssd :)14:56
kenvandinehehe14:57
vishpitti: btw , Bug #542091 , for the cheese apport hook [in case you want to keep track]14:58
ubottuLaunchpad bug 542091 in cheese "Add apport hook for cheese" [Wishlist,In progress] https://launchpad.net/bugs/54209114:58
vishamong the long list of other things ;p14:58
nigelbpitti: bluez is done! built successfully, attached debdiff15:06
=== MacSlow is now known as MacSlow|alf
=== MacSlow|alf is now known as MacSlow|afk
=== MacSlow|afk is now known as MacSlow
vishseb128: hi , did you already pull from the packagers branch for the humanity update?16:05
* vish just added a bug# to the changelog a few mins ago16:05
seb128vish, I'm uploading right now, I can ctrl-C if required16:05
seb128I've r4616:06
seb128which one did you add?16:06
vishseb128: right thats the one , no problems :)16:06
seb128ok good16:06
seb128uploaded16:07
vishseb128: thanks. :)16:07
seb128np16:07
seb128mvo, where can the s-c categories names be translated?16:20
mvoseb128: they should be in the s-c pot16:23
mvoseb128: unless they come from a ".directory" file in which case that is used16:23
seb128mvo, the .directories being used are the gnome-menus ones? if for accessories or graphics?16:24
mvoyes16:25
seb128mvo, seems buggy16:26
seb128mvo, those are not showed translated there16:27
mvobut they are translated?16:27
seb128the gnome-menus are correctly translated though16:27
dpmseb128, mvo, we've got bug 545102 about those, the categories' translations don't seem to be loaded from data/software-center.menu16:27
ubottuLaunchpad bug 545102 in software-center "Department names shows untranslated, albeit they're translated" [Undecided,New] https://launchpad.net/bugs/54510216:27
seb128well I guess so, those strings are the gnome-panel menus categories16:27
seb128dpm, thanks16:28
seb128mvo, does it work for you?16:28
seb128mvo, I get the same issue in german16:29
chrisccoulsonpitti - do you think that bug 544139 should be added to the list of RC bugs?16:47
ubottuLaunchpad bug 544139 in consolekit "Active VT tracking can fail at startup" [High,Triaged] https://launchpad.net/bugs/54413916:47
chrisccoulson(we should certainly try and fix it before lucid is released anyway)16:47
pittichrisccoulson: yes, I think so16:48
chrisccoulsonthanks, would you mind doing that for me? :-)16:48
nigelbpitti: will you be sponsoring bluez hook? or want me to subscribe sponsors?16:48
pittinigelb: please do subscribe sponsors16:48
pittiI'll probably get around to it soon, but no need to block on me16:48
nigelbok, thanks :)16:49
mvohm, rhythmbox having VIRT 1,5Gb is not good, is it?16:51
seb128mvo, I will not reply to you as long you ignore me and my comment :p16:56
seb128mvo, but no, doesn't seem correct16:57
mvoseb128: sorry, I'm in a meeting16:57
seb128mvo, I was just joking ;-)16:57
* seb128 hugs mvo16:57
mvoseb128: I can check buggy translations16:57
seb128mvo, did you use the music store?16:57
seb128mvo, you should check buggy code, read the bug dpm indicated16:57
seb128mvo, the translations are there16:58
mvoI think its because of the store16:58
mvothat is very impressive btw16:58
seb128;-)16:59
mvoseb128, dpm: i18n issue should be fixed, thanks17:01
mvodpm: please hit^Wping me if you notice similar issues, I'm not on top of my bug lists17:02
seb128mvo, you rock!17:02
seb128mvo, thanks17:02
mvoI wish17:02
* mvo hugs seb12817:02
seb128pedro_, those gnome-appaerance-capplet crashes in pango are not pango bugs they are bug #27491517:16
ubottuLaunchpad bug 274915 in gnome-control-center "gnome-appearance-properties crashed with SIGSEGV in pango_shape()" [Medium,Triaged] https://launchpad.net/bugs/27491517:16
dpmmvo, thanks!17:18
pedro_seb128, will review them back again, it's a bit tricky sometimes to determine where is actually crashing with those pango lines, but will have a look again and mark those as dup of that one17:18
pedro_seb128, thanks for ping me ;-)17:18
seb128pedro_, I'm cleaning those now17:18
dpmmvo, no need to hit you, I know everyone is busy, I prefer hugging17:18
seb128pedro_, just letting you know for next ones17:18
* dpm hugs mvo17:18
seb128pedro_, there is quite some of those weird bugs on other softwares too though17:23
seb128pedro_, but most have been triaged or forwarded already by now17:24
dpmmvo, while you are at it, you might want to have a look at bug 545095 as well, it seems that the "Featured Applications" string is not marked for translation in s-c17:25
ubottuLaunchpad bug 545095 in software-center ""Featured Applications" untranslated and untranslatable" [Undecided,New] https://launchpad.net/bugs/54509517:25
ccheneyi think i have a real fix for the annoying OOo won't upgrade due to running OOo problem17:26
pedro_seb128, true, i've seen a few pango crashes for a couple of cycles already but those are already submitted upstream (most of them)17:26
pedro_IIRC behdad fixed a bunch during the first weeks on this cycle though17:26
brycehasac, yes see https://wiki.ubuntu.com/X/GitUsage17:26
seb128pedro_, ok17:27
vishmvo: hi , reminding the list of SC icons > http://paste.ubuntu.com/400652/  would be great if you can update the .menu file :)17:29
kenvandinedpm, think there might be any more translation related requests for gwibber?17:35
dpmkenvandine, let me have a look at the existing i18n bugs, I think they had all committed fixes, IIRC...17:38
kenvandinedpm, thx, i am getting close to releasing with a bunch of bug fixes17:40
kenvandinewould like to fix up any known i18n issues too17:40
seb128time for sport and dinner, bbl17:43
dpmkenvandine, there is only bug 538981 as pending IIRC. All the others I knew of were Fix Committed. I tagged them all with gwibber-i18n -> https://bugs.edge.launchpad.net/gwibber/+bugs?field.tag=gwibber-i18n17:45
ubottuLaunchpad bug 538981 in gwibber "Strings in the side panel are not translatable" [Medium,Confirmed] https://launchpad.net/bugs/53898117:45
dpmI haven't looked at that bug yet, I'm not sure how difficult is to make those strings translatable there17:45
kenvandinedpm, humm, yeah ideally we want to keep that list dynamic but that means it won't appear in gwibber.pot17:48
kenvandinedpm, have you seen any other application do that?17:48
kenvandinedpm, we could just wrap "name": aname.capitalize() with _()17:48
kenvandinebut then it still won't make it into the template17:49
kenvandinedpm, can you manually add strings that need translating to the template?17:49
dpmkenvandine, you could use this perhaps -> http://docs.python.org/library/gettext.html#deferred-translations17:50
dpmthe N_() trick17:51
dpmbecause otherwise adding strings to the template manually will be a pain for you, I think17:52
kenvandinedpm, it would be cool if we could add strings to the template in a way that they don't get over-written when you rebuild the template17:54
kenvandineor i guess just a list of possible values stuck in the code somewhere :)17:54
dpmkenvandine, would N_() not work for you in the same way?17:56
kenvandinedpm, either way... it would be a pain to maintain, if we ever change the default values, we would need to remember to update that list17:56
kenvandinedpm, yeah it would17:56
kenvandinejust extra code, would think there would be a way to create a static list of possible strings to translate somehow17:57
* kenvandine isn't an i18n guru17:57
* dpm just fakes it17:57
didrocksthis time enjoying an early evening :) see you tomorrow everyone!17:59
kenvandineenjoy didrocks!17:59
kenvandinedpm, if i added something like this to gwibber.pot17:59
didrocksthanks kenvandine, have a nice afternoon :)17:59
dpmyep, have a nice evening didrocks!17:59
kenvandinemsgid "Messages"18:00
didrocksdpm: you too ;)18:00
kenvandinemsgstr "18:00
kenvandinethat would get removed next time we re-generate the template right?18:00
dpmyeah, it would get removed18:00
kenvandineok18:00
kenvandineso in code is the lesser of the evils :)18:00
* kenvandine takes a stab at it18:01
dpmyeah, otherwise you'd have to have a patch to modify the template at build time, I guess18:01
kenvandinei really want to make sure gwibber is localized :)18:01
* kenvandine hugs the awesome translators18:01
* dpm hugs kenvandine for the rocking work in getting i18n fixes done or merged!18:02
dpmkenvandine, the fixes you already did will get gwibber in top shape i18n-wise, but there might be more small issues coming up as translators keep testing the app. This is usually how things are detected: translators finish the translation, the language pack is released, and then translators detect untranslated or untranslatable strings18:05
dpmanyway, need to take a break, bbl18:05
kenvandinedpm, thx!18:06
mvovish: thanks, I thought that mpt was working on that18:18
mvovish: or am I misrembering18:18
mptmvo, hi, good to see you online again18:18
mptmvo, working on what?18:18
mvompt: the new icon integration with software-center.memu. I thought you mentioned that you want to ensure it will also be in the directory files (or rather that it gets names that match the directory files)18:19
mptmvo, I don't know where to put the icons to achieve that myself. I thought kwwii was doing that18:20
mvompt: ok, I will talk to him18:21
mptvish might be able to do the corresponding software-center.menu changes to look in the right places18:21
vishmvo: mpt: we considered all options and finally named them so http://paste.ubuntu.com/400652/18:21
mptUnfortunately I'm not being given any time to do that18:21
mvompt: ok18:22
vishmpt: they just landed in Lucid and should be available18:22
pittigood night everyone!18:23
vishmvo: i went through the .menu file and right now all those categories use the "applications-other" icon , you can just re-name them and they should be good to go18:25
cjohnstonkwwii: ping19:07
seiflotfyhi rickspencer3 :)20:12
rickspencer3seiflotfy, 45 minutes, rigt?20:13
rickspencer3I'm on a call now20:13
seiflotfyyes sir20:13
kwwiicjohnston: pong20:33
cjohnstonhey kwwii, were you sent into about a visual problem for the scrollbars that I am having?20:35
seiflotfyhey guys21:01
robtaylorseiflotfy: rickspencer3 : so, shall we start?21:01
rickspencer3yuppers21:01
robtaylorseiflotfy: oh, hi :)21:01
seiflotfyhi robtaylor :)21:01
rickspencer3hi robtaylor and seiflotfy21:01
robtaylorseiflotfy: so, this is your meeting, what would you like to cover?21:01
rickspencer3all - seiflotfy, robtaylor and I are going to spam the channel for a bit21:02
rickspencer3sorry21:02
rickspencer3go ahead seiflotfy21:02
robtaylorrickspencer3: :)21:02
seiflotfyrickspencer3, robtaylor i would like to cover the topic of semantic technologieso n the desktop21:02
seiflotfyespecially for lucid +121:02
seiflotfycrap my keyboard is not responsive21:02
seiflotfyrobtaylor is the CEO of codethnk one of the main companies behind tracker21:03
seiflotfyme and kamstrup are here representing Zeitgeist21:03
rickspencer3"semantic technologies"21:03
robtaylorseiflotfy: ah, thanks for the intro.21:03
rickspencer3this sounds rather, mmmm,21:03
rickspencer3how should I say, "grand"?21:03
seiflotfywell its a step lucid +1 will have to introduce21:04
seiflotfyubuntu is going a nice path by redesinging stuff to be more user friendly and enrich the user experience21:04
robtaylorrickspencer3: so, I guess i should give you s shre overview of where I'm coming fron. COdethink's been doing a lot of teh core work for tracker, which is athe core data storage compoennt for Maemo6 and Meego21:04
seiflotfyyet those are going to be limited by the current state of techznologies21:04
robtayloruh, sorry, high latency here21:05
robtaylorrickspencer3: and of corse, as part of this work, we're very interested in increasing the adoption of the technology21:06
thorikehey sei21:06
thorikehey seiflotfy , robtaylor , rickspencer321:06
rickspencer3how does zeitgeist and tracker interact?21:06
rickspencer3hi thorike21:06
rickspencer3are they related at all?21:06
RainCT_netbookHi all21:07
seiflotfyrickspencer3, yes and no21:07
robtaylorrickspencer3: at the moment, there's some relation at the UI level, GAJ combines data from tracker and zeitgeist to present the timeline view21:07
* robtaylor has been pushing seiflotfy and kamstrup on this for a while ;)21:07
seiflotfyrickspencer3, while tracker knows about ur data zeitgeist only covers how ur data is used21:07
seiflotfyboth can co-exist nicely21:08
rickspencer3ok21:08
seiflotfyas demoed in GAJ21:08
rickspencer3so what specifically do envision for Ubuntu 10.10?21:08
robtaylorseiflotfy: you want to go on that?21:08
seiflotfyrobtaylor, sure thing21:08
thorikehi RainCT_netbook  :-)21:09
* mclasen hopes that trackers stops ruining system responsiveness by then...21:09
seiflotfyrickspencer3, lets start with documents21:09
seiflotfyrickspencer3, currently there is no such thing as document organizer21:10
seiflotfyagain we r fiddling through the hierichal filesystem21:10
robtaylormclasen: have you installed a recent 0.7 tracker? Its crawing is good enough for mobile devices ;)21:10
seiflotfytracker would allow us to browse our data using tags, authors, annotations etc21:11
rickspencer3robtaylor, to be honest, I have never used a tracker like feature that wasn't either:21:11
rickspencer31. weirdly consuming resources at inconvenient times21:11
rickspencer32. run manually, therefore alway too out of date to be useful21:11
rickspencer3has something changed there?21:11
robtaylorrickspencer3: absolutely21:11
seiflotfyyes alot21:11
seiflotfytracker being considered the datastore for meego and maemo says it all i guess21:12
seb128robtaylor, the issue is mainly inotify no?21:12
rickspencer3so what has changed?21:12
robtaylorseb128: that's the main remaining issue.21:12
seb128robtaylor, or did you stop index the disk and rely on softwares to push infos?21:12
seb128indexing21:13
robtaylorrickspencer3: so, several things here21:13
seb128robtaylor, the io load handling under linux just sucks21:13
robtaylorrickspencer3: the archoitecture was redesigned so there is a cetral 'semantic'; store, then miners for various data sources21:13
seb128indexing will slow down your system a lot, I don't see how tracker can solve that with current linux and inotify limitations21:14
robtaylorrickspencer3: of which the filesystem is one21:14
robtaylorrickspencer3: we did a lot of fine grained interaction tweaking to keep the system resonsive21:14
robtaylorrickspencer3: you can test it out, the tracker-team ppa has the latest version of the 0,7 system in21:15
rickspencer3so that issue #1 is gone in your mind?21:15
seb128robtaylor, did you stop running the indexer?21:15
robtaylorrickspencer3: absolutely.21:15
rickspencer3ok21:15
robtaylorseb128: no, indexing is still needed to get a lot of information in, but it's secondary to the aim of tracker21:15
RainCT_netbookseb128: Last time I tried it on my netbook it indexed everything I have there in a couple hours and I didn't notice it at all21:15
RainCT_netbookit's definitelly better than 0.621:16
robtaylorseb128: the ideal long-term plan would be that indexing isn't needed and all software uses it as the data oraganisation systenm. but that's a loong way off. if ever..21:16
rickspencer3robtaylor, so what value does it bring now?21:17
rickspencer3why would users want this?21:17
seb128well I fail to see how you solve the issue from tracker since the issue is mainly inotify and linux io load handing21:17
robtaylorseb128: as you pointed out, the main remaining issue is kernel-level. inotify sucks in some interesting ways21:17
seb128but I'm happy to be proved wrong21:17
seb128I will check on 0.721:17
robtaylorseb128: basically, its all done by making sure the systemhas time to run and we keep low priority. could be made better with more kernel suport, but atm that's outside our development remot (watch this space thoiugh)21:18
seiflotfyrickspencer3, let alone ubuntu is missing a proper search tool21:18
robtaylorseiflotfy: one minute please21:18
rickspencer3robtaylor, go ahead21:19
seb128robtaylor, ok, please continue we can discuss that later21:19
seb128ie technical details are interesting but not important now ;-)21:19
robtaylorrickspencer3: so, what tracker brings at the architectureal level is linked data21:19
robtaylorrickspencer3: the model is bsaed on  relationships between items of data21:20
robtaylorrickspencer3: in practice, the means we can support such things as 'show me all emails related to this document'21:20
* rickspencer3 nods21:20
robtaylorrickspencer3: or, combined with zeitgeist, ' show me the work i did last week with the ubuntuone team'21:21
rickspencer3ok21:21
rickspencer3I would like to discuss each a bit seperately first, if that is ok21:21
rickspencer3so I can understand a bit where you are coming from21:21
thorikerobtaylor, what kind of relationships do you mean?21:21
robtaylorrickspencer3: absolutely21:21
thorikeok I pull back21:22
rickspencer3robtaylor, how does the UI reveal this functionality in your mind?21:22
robtaylorthorike: that's a big question ;)21:22
thorikerobtaylor, just examples for now21:22
thorikeI know it's a big question :-)21:22
rickspencer3robtaylor, are envisioning something like places -> Search for Files21:23
robtaylorrickspencer3: so, in Ui designs that are being actioned here, we have a core concept of people. so for example, you can tag face in photos in either facebook or on a device.21:23
rickspencer3but it works better because the daa is indexed21:24
rickspencer3?21:24
robtaylorrickspencer3: and from your contacts vuiew, go to 'pictures of this person'21:24
rickspencer3hmmm21:24
rickspencer3that sounds rather esoteric21:24
rickspencer3does it help for when you are trying to get work done, or is the people concept only an example?21:24
robtaylorrickspencer3: i'm afraid ihave to be a bit inventive here, lots I can't directly talk about ;)21:25
thorikerickspencer3, well re zeitgeist and the people thing21:25
thorikewhat we do is log the interaction with the people21:25
thorikeand then can tell the user21:25
thorikewhat did you do last week with dave21:26
robtaylorrickspencer3: so, in terms of getting work done, lets say you're working on a document or project. You now need to check up on some details you know you discuessed, but you can't remember when21:26
rickspencer3ok21:26
thorikeemails, chat, synchronous tools21:27
rickspencer3robtaylor, this happens to me all the time21:27
rickspencer3like someone pasted me a link in IRC for a wiki page, and I can't find it21:27
rickspencer3would tracker help me with that in some wya?21:27
seb128:-)21:27
robtaylorrickspencer3: right, so we're on this border between searching and data foraging21:28
rickspencer3robtaylor, I'm just trying to understand your vision here, I'm not questioning the value of tracker21:28
thorikerobtaylor, as far as I know tracker doesn't really support the "when ting"? am I wrong?21:29
robtaylorrickspencer3: your brain is saying, say, i know i got that link whenwe were talking about 'x'21:29
* rickspencer3 nod21:29
robtaylorso you'd use teh global search for x, then from that search, get all the messages (email/im/irc) around whn 'x' was mentioned21:29
rickspencer3ok21:30
rickspencer3and then presumedly I'd see the link21:30
robtaylorthinking on my feet for the ui, you could have actions from the initial results (e.g. arrowing off from the result -> messages -> get a time based view)21:30
rickspencer3and not look like a total ass, as usual21:30
robtaylorthat's the idea21:30
kenvandineseb128, can you sponsor libubuntuone again?21:31
rickspencer3robtaylor, so I take it there is no UI for desktops yet then?21:31
kenvandineseb128, fixed the missing dep Riddell reported21:31
robtaylorby the system knowing how items are realted, it helps model how a user thinks21:31
robtaylorrickspencer3: exactly21:31
rickspencer3ok21:31
robtaylorrickspencer3: there's the search-bar in tracker, and there's GAJ21:31
seb128kenvandine, ok21:31
rickspencer3ok21:31
rickspencer3so a bit of UI21:31
robtaylorrickspencer3: but nothing that has focused ui designs and tight-teams21:31
rickspencer3ok21:32
rickspencer3let us put UI aside then21:32
seiflotfyhi kamstrup21:32
kenvandineseb128, thx21:32
rickspencer3I think we can discuss the user value without designing specific UI21:32
robtaylorrickspencer3: some of Seth's ideas at the UI hackfest, and the nautilus discussions started going in this direction though21:32
seb128kenvandine, yw!21:32
rickspencer3so, in terms of Zeitgeist, how does that relate?21:32
* rickspencer3 acks the nautilus designs from hack fest21:32
seiflotfyrickspencer3, not the same21:33
seiflotfywhile tracker runs of statis metadata21:33
seiflotfyzeitgeist provides personal experience data21:33
seiflotfyas in "what do i do when reading x" based on a log21:33
rickspencer3so the proposal is to have desktops running two trackers all the time?21:33
seiflotfyuhm its a daemon21:34
seiflotfyand they are both leightweight21:34
seiflotfyi have them both here21:34
RAOFGood morning, ipv6 TheMuso :)21:34
TheMusoRAOF: heh good morning.21:35
robtaylorfrom my point of view, time-series data is very important for tracker. ITs the only way we'll be able to offer a lot of the functionailty we want21:35
robtaylorand its forndamentel for racking21:35
TheMusoRAOF: Totally unintentional on my part.21:35
robtaylor*ranking21:35
rickspencer3seiflotfy, ok, that's a "yes"21:35
rickspencer3seiflotfy, can you tell me why a user would want zeitgeist if they are already running tracker, or visa versa?21:36
seiflotfykamstrup, wanna take over or should i?21:36
robtaylorrickspencer3: so, zg is a component that deals with store what actions the user perfomed, or had happen to them21:37
robtaylorrickspencer3: it can then analyse relationships between data by the time-locality of actions21:37
robtaylorrickspencer3: so, stricly speaking, zg isn't anindexer in any real sense21:38
rickspencer3robtaylor, that sounds exactly like Zeitgeist to me :/21:38
robtaylorrickspencer3: zg = zeitgiest21:38
robtaylor:)21:38
rickspencer3ok21:38
rickspencer3I can do that21:38
rickspencer3sorry to be dense, but I'm not seeing what zg brings extra, I know I am missing something21:39
* robtaylor gets lazy ;)21:39
seiflotfyzeitgeist logs activity21:39
RainCT_netbookYeah, Zeitgeist isn't an indexer, it's just a daemon providing logging functionallity over D-Bus (insert, get, find, relationships)21:39
rickspencer3ok, so it logs actions, rather than indexes data?21:40
RAOFBut isn't tracker *also* meant to be (among other things) a store for such things?21:40
RainCT_netbookAnd the information gets inserted by applications themselves, plugins or the zeitgeist-datahub daemon (which currently just copies over stuff from GtkRecentManager)21:40
seiflotfyrickspencer3, zeitgeist provides contextual relationships while tracker does contentual :)21:40
* didrocks would like zeitgeist and tracker being officially blessed by GNOME release managers first (as the discussion for tracker as a module in GNOME is currently stuck)21:40
seiflotfydidrocks, i am proposing this week21:40
rickspencer3ok, so I think I understand a bit better21:40
didrocksrickspencer3: don't blame yourself, this was a high point of discussion for a lot of people in GNOME ML "differences between tracker and zg is not obvious" :)21:40
rickspencer3so, robtaylor if I could do 1 think to help you, what would that be?21:41
seiflotfytracker would be optimally the main storage for tags/annotations etc on the desktop21:42
seiflotfyallwoing cross desktop applications to re-use these tags21:42
seiflotfyzeitgeist on the other hand allows cross application exchange of activities21:42
rickspencer3seiflotfy, same question for you, what would you ask me to do?21:43
seiflotfylucid + 121:43
seiflotfysome tracks21:43
seiflotfythat would push the integration of both in 10.1021:43
seiflotfydo u want specific use cases ?21:43
rickspencer3seiflotfy, you want to ship zg and tracker by default in 10.10?21:43
seiflotfythat woudl be optimal21:44
seiflotfyits the only way we can catch up a bit with KDE21:44
rickspencer3ok21:44
rickspencer3so21:44
rickspencer3I get this question all the time21:44
TheMusoHow would average users benefit? Sorry coming in late to this convo.21:44
rickspencer3"please help my project by delivering it in Lucid"21:44
rickspencer3TheMuso, hmm, covered that a bit, maybe scrollback?21:44
TheMusorickspencer3: oh ok21:44
seiflotfyrickspencer3, ok21:45
rickspencer3seiflotfy, so this is not something that any one person decides21:45
seiflotfyi know21:45
seiflotfybut we can start with natuilus21:45
rickspencer3what you need to do is to demonstrate some awesome user value AND21:45
* TheMuso gets the benefit now21:45
rickspencer3a track record of execution21:45
seiflotfywhy not provide "most used documents per folder"21:45
seiflotfyzeitgeist covers this easily21:46
rickspencer3then you could potentially influence folks21:46
seiflotfyas seen in docky21:46
rickspencer3I would suggest targeting a zg-based feature in universe for 10.10 that you can get some excitement about21:46
rickspencer3think of the path that Gwibber took, for example21:46
seiflotfyyeah21:46
rickspencer3tracker seems different to me21:46
rickspencer3seems to be shipping on multiple platforms, and in some ways is tried and true code21:47
rickspencer3robtaylor, seiflotfy does that seem about right to both of you?21:47
robtaylorrickspencer3: well, the core of tracker is tried and true. There's a lot that still needs to be done around desktop UI designs, programmer-freindly api's that sort of thing21:48
rickspencer3mmmm21:48
robtaylorrickspencer3: and a big question we have at the moment is how todo really great ranking for global search21:49
thorikewhat' global search?21:49
robtaylorthorike: seen an android phone?21:49
seiflotfyrickspencer3, u can always get back at kamstrup for zeitgeist questions21:49
thorikeyep21:49
rickspencer3hmmm21:49
seiflotfyhe is a maintainer and lead architect21:49
rickspencer3so sounds like tracker features might need a bit of back time21:49
rickspencer3bake time, even21:49
seiflotfyrickspencer3, kamstrup is employed by canonical so u have a direct connection to him21:50
* kamstrup solutes rickspencer321:50
rickspencer3seiflotfy, robtaylor I think we need some UI concepts that some developers are committed to deliver ...21:51
rickspencer3and bring this conversation to UDS21:51
robtaylorrickspencer3: exactly21:51
seiflotfyrickspencer3, we have GAJ21:51
seiflotfy:)21:51
seiflotfyit uses both21:51
seiflotfy:)21:51
rickspencer3seiflotfy, is it in universe?21:51
seiflotfylucid?21:51
thorikeand with the next release we'll add "related items"21:51
robtaylorrickspencer3: do you know if the Canonical UX team be there?21:51
rickspencer3robtaylor, yes, they will be21:51
seiflotfyrickspencer3, i think it is21:51
rickspencer3seiflotfy, what is the package name?21:52
robtaylorrickspencer3: i think the main thing for me is to make sure 10.10 doesn't ship with the old tracker in main again ;)21:52
chrisccoulsonrobtaylor - it won't21:52
rickspencer3robtaylor, that can for sure be arranged easily21:52
chrisccoulsoni'm already on the case with that ;)21:52
robtaylorchrisccoulson: awesome :D21:52
chrisccoulsonthe 0.7 packaging is in good shape already21:52
chrisccoulsoni will upload it as soon as lucid + 1 opens21:52
seiflotfygnome-activity-journal21:52
seiflotfyrickspencer3, ^21:52
seiflotfywe are relasing again in a week or 221:53
robtaylorrickspencer3: and some simple low-hanging fruid i'd liek to see is getting some tracker-based vfolders in nautilus by default. e.g. 'All Videos' 'All Music'21:53
rickspencer3robtaylor, interesting21:53
chrisccoulsonthat would be awesome21:53
seb128hum21:53
rickspencer3this seems something that can be installed as a seperate package21:53
seb128I'm wondering if we can demote tracker in lucid21:54
rickspencer3from universe21:54
rickspencer3and bake21:54
seb128since nautilus is using it at runtime now21:54
robtaylorrickspencer3: there's a little that needs to be done in nautilus, to support direct tracker querying in it's vfolder query language21:54
seb128no offense to the tracker team but 0.6 is outdated anyway21:54
seb128chrisccoulson, do you know?21:54
robtaylori've chatted over thatwith alexl, and he's up for is, as the tracker backing is pretty much the only supported vfolder backend for the next nautilus release21:54
robtaylorseb128: yeah, i really want to see 0.6 dissapear for good21:55
chrisccoulsonseb128 - i don't mind. i was going to upload the latest 0.6 point release at soem point, but i never got round to it21:55
chrisccoulsoni haven't used 0.6 for ages ;)21:55
seb128chrisccoulson, I don't think we should was efforts on this version21:55
seb128let's do better in lucid+121:55
chrisccoulsonyeah, lucid+1 will get 0.7/0.821:55
seb128tracker 0.6 in lucid is of little anyway for users21:55
rickspencer3seb128, are you suggesting demote tracker to universe for lucid, and then do tracker .7 in main when 10.10 opens?21:56
robtaylorseb128: its pretty damn aweful. I'm still pushing for the 0.8 stable series to get a Tracker *2* name. We need some signalling that is really is a completely different beast with different aims21:56
seb128rickspencer3, yes21:56
seb128rickspencer3, well not especially in main if nothing use it21:56
seb128rickspencer3, having it in universe means it's easier for other contributors to work on it too21:56
rickspencer3if we demote it to universe, could we not replace it with .7 in lucid?21:56
rickspencer3or is that too aggressive?21:56
seb128chrisccoulson has been looking at it21:57
seb128but 0.7 had stopper issues21:57
chrisccoulsonrickspencer3, AFAIK, the current 0.7 branch does not handle ontology changes well, which would mean users having to manually remove their index after upgrading21:57
seb128I'm not sure we should spend efforts on that now21:57
seb128nothing in going to use it in lucid anyway21:57
chrisccoulsonwhich was the blocker issue for not uploading 0.721:57
chrisccoulsonrobtaylor can probably explain better though ;)21:57
rickspencer3ok21:58
rickspencer3never mind21:58
rickspencer3it was just a suggestion, it's not my call anyway21:58
robtaylorbtw, if anyone needs a bit of an overview of the linked data structure we use at the ce core of tracker, I have a nice slide cast on my blog (http://blog.floopily.org/2010/03/01/rdf-beginners-guide-and-competition/)21:58
rickspencer3robtaylor, I'll try to find the link later21:58
rickspencer3j/k21:58
chrisccoulsonrickspencer3, i sort-of maintain 0.7 in a PPA though, with mbiebl (the debian maintainer)21:58
rickspencer3ok21:58
chrisccoulsonrickspencer3, https://edge.launchpad.net/~tracker-team/+archive/tracker-unstable has 0.7 packages21:59
rickspencer3ok21:59
rickspencer3thanks21:59
chrisccoulson(which i need to update with last weeks release at some point)21:59
robtaylorchrisccoulson: *nod*. I'm not sure of the status on that either. It's at least scheduled to be fixed by the 0.8 series21:59
seiflotfyrickspencer3, i would like to add http://seilo.geekyogre.com/uploads/2010/03/image7.png21:59
rickspencer3robtaylor, seiflotfy do we need to decide anything now?21:59
* didrocks adds the link to "should be read when slackering tomorrow"21:59
seiflotfycoudl we make this a track for lucid21:59
robtaylorrickspencer3: no, not at all :)21:59
seiflotfy?21:59
didrocksoupss, my boss is there :)21:59
rickspencer3hehe21:59
robtaylorrickspencer3: i just want to get some shared bain cells going on the possibilities ;)22:00
seiflotfyi mean lucid + 122:00
robtaylor*brain22:00
robtaylorrickspencer3: and make sure we have the right basis in place going forward22:00
robtaylorsomeone might need to explain to me what a 'track' is ;)22:01
seiflotfyrickspencer3, my problme with the current desktop is that it is too static22:02
seiflotfywe should add a track or 2 in lucid+122:02
seiflotfyto discuss enriching new possibilities of experience22:03
seiflotfyadaptive and personal experience22:03
seiflotfyby static i mean everything on the desktop is based on the hierarchal filesystem22:04
rickspencer3seiflotfy, my point to you is that you are empowered to realize your vision for the desktop22:05
rickspencer3but you'll have to prove it with your code and prove it to the users22:06
seiflotfyi get what u mean22:06
chrisccoulsoni think the idea of virtual folders in nautilus is awesome. i have music scattered all over my desktop (some private, and some shared with other users), and it would be cool to have all that in a single virtual folder22:06
TheMusoVirtual folder for me would be pointless, I just organise things into folder proper.22:07
TheMusofolders22:07
thorikechrisccoulson, and with zeitgeist you would then be able to ask: what did I listen to yesterday, while I was writing on this paper22:07
thorikeand in general, zg will relate the music files according to how u use them22:08
chrisccoulsonTheMuso - but sometimes you have to have things in different folders, but don't want to care about that when you're accessing them22:08
thorikeso, our aim is to surprise the user with the "related items" thing, but in a positive way22:08
TheMusochrisccoulson: I can see how this is true for some.22:08
thorikeas in "ah yeah  that's true, these songs I like to hear together"22:08
robtaylorchrisccoulson: how's your c? ;) I could do with a volunteer for that stuff :)22:08
seb128robtaylor, did alex agree on virtual folders in nautilus?22:09
chrisccoulsonrobtaylor, i like C ;)22:09
thorikewe are also thinking of expanding this to include recommender functionality22:09
robtaylorseb128: its functionailty that's already there22:09
seb128he has usually been against virtual locations to display that sort of things22:09
seb128hum k22:09
seb128I though the vision was that nautilus is a file manager22:09
thorikerickspencer3, and in general, as compared to general semantic technologies such as tracker22:10
robtaylorseb128: what he's agains it virtual file systems22:10
seb128that ie music management should be done in ie rhythmbox22:10
robtaylor*against is22:10
seb128not in nautilus22:10
thorikewhich try to express similarity and semantics of contents independent of users22:10
thorikezg personalizes this semantics22:10
robtaylorseb128: absolutely, I agree :)22:10
rickspencer3guys, tbh, when I hear a term like "semantic technologies" I think of solutions looking for problems22:11
seb128I think I miss the nuance there22:11
didrocksso do I :)22:11
thorikein that it will be capable of reflecting personal meaning of computer-represented objects, as emerging and developing from them being involved in a user’s activities and experiences.22:11
rickspencer3I think it might be easier to explain zg if you could sum it up in terms that put user valuke forward22:11
robtaylorseb128: possibly. have you got a unstable nautilus build to hand?22:11
thorikerickspencer3, oki :-)22:11
thorikewhen using zeitgeist you don't need to remember where / in what folder or alike you have put an item22:12
robtaylorseb128: if you install the tracker0.7 package from the tracker-team ppa, you should be able to do a search in nautilus and save it as a vfolder22:12
thorikeit's just there, for example, in gaj22:12
chrisccoulsonrobtaylor - oh, i've not tried that yet22:12
thorikeyou can recognize the name instead of recalling it22:12
thorikewhich is especially good for sporadic users22:12
TheMusoI don't like that, that doesn't promote keeping organised files. :)22:13
didrocksI think real and precise test cases are the best way to show the added value.22:13
seb128robtaylor, not right now but I will try later for sure22:13
thorikeTheMuso, there is aplace for keeping files organized22:13
thorikebut it is known that there are filers and pilers alike22:13
thorikeand none is better or worse22:13
robtaylorchrisccoulson: yeah. unfortunatly the ui doesn't support making an 'all video' search properly, but you can do a search for  'foobar', type Video, and hand edit the .vfolder file22:13
robtaylorseb128: awesome :)22:13
thorikeand in particular for the "short term memory of computer"22:13
thorike1 or two days22:13
thorikemax22:13
chrisccoulsonrobtaylor, i just tried searching in nautilus, and it crashes ;)22:13
chrisccoulsonheh22:14
robtaylorchrisccoulson: doh !22:14
thorikezeitgeist will make sure, you'll get back to your data22:14
thorikewithout overhead of organizing them22:14
thorikerickspencer3, does that make more sense to you like that?22:14
chrisccoulsoni should probably modify the tracker packages actually to build the gdb symbols too22:14
chrisccoulsonor rather package the symbols22:14
robtaylorchrisccoulson: well, so, when i'm back in GMT we should meet up and make a plan for stabilising it and growing teh query format22:15
robtaylorchrisccoulson: is that with nautilus git head?22:15
thorikerickspencer3, so another use case would be empowering the open-file dialogue with zg22:15
chrisccoulsonrobtaylor, this is with 2.29.92.1, so i'm not sure how far that is from git now22:16
thorikewhere you'll have the "most-used with this app" items22:16
thorikeor recently used22:16
thorikewhere zg is different from recently used as it captures, keeps, and exploits the old events too22:16
thorikeso it doesn't loose interaction memory22:16
thorikewhich is one of the design flaws with "recent documents"22:17
robtaylorchrisccoulson: thats pretty much git head, i think22:17
chrisccoulsonyeah, i just looked at the recent commits, and there doesn't seem to be much interesting there22:17
seb128chrisccoulson, the most recent one could be22:18
thorikeopen-file dialogue or the zeitbutton could also make use of related-items fct. of zg22:18
chrisccoulsonseb128 - http://git.gnome.org/browse/nautilus/commit/?id=a92a7bfcb2c7bb9f171724bc97b1c163f52c0fd7 ?22:18
robtaylorthorike: i think we need projects and code for these ideas22:18
seb128chrisccoulson, yes22:18
thorikeso what is most-related by usage (using together with the doc. displayed in the window) with the document22:18
seb128chrisccoulson, not really in fact22:18
thorikerobtaylor, it's coming22:19
robtaylorchrisccoulson: you could be hitting some issue in the magic dynamic SO loading22:19
seb128chrisccoulson, I though they fixed a tracker crash but it was probably before tarball22:19
thorikethe related-items fct. will be out with the next release22:19
thorikeand we are already starting to build on it22:19
robtaylorchrisccoulson: though it just worked for me, git head 1ab3721422:19
thorikewhich will allow to capture something like working sets22:20
RainCTgood night guys22:20
thorikewhich is "documents you used together working on a task"22:20
thorikewhich often users dont keep tidy and clean within folders22:20
chrisccoulsoni should probably update the tracker version in the PPA really22:21
chrisccoulsonbut that's a job for tomorrow now22:21
chrisccoulsoni may as well wait for tomorrows tarball ;)22:21
kklimondasorry to ask but how does it track what task is user currently working on?22:21
thorikeso, we relieve users from the burden to always right away copy stuff together when they finish working on something, for example, because they are interrupted by their boss and have to switch to sth else22:21
seb128chrisccoulson, did you ping asac about nm-applet btw?22:21
seb128asac, asac_: ^22:22
thorikeand afaik, this working-set notion is not supported by anything currently22:22
seb128asac, asac_: can we get those nm-applet changes in lucid now?22:22
thorikebut often users work with several items on a task22:22
thorikeand task switching is known to be a heavy burden mental-load wise22:22
thorikeit's already difficult enough to "think-into" the task again22:23
thorikethe desktop should relieve users from searching the relevant items again, if they have already been used22:23
thorikefailure of re-finding is especially annoying people, because they know it's there22:24
thorikebut they cant get a hold of the stuff again22:24
chrisccoulsonseb128 - we mentioned it in the meeting on tuesday, but i just pinged him again22:25
chrisccoulsonit might have got lost during the team meeting ;)22:25
thorikekklimonda, it does by analysing the usage and re-usage patterns22:26
thorikeand you are right there is not a UI yet22:26
thorikebut that will soon come22:26
seiflotfykklimonda, we implemented A Priori and we are working on also implementing Winepi and Minepi22:26
seiflotfytwo other algorithms22:26
thorikethey are basically associating the events on the documents22:27
thorikeand thereby also relating the documents22:27
kklimondadoes it mean all applications have to be aware of tracker/zg to make use of it?22:27
thorikeand one UI is already in trunc of zg22:27
kklimondawell, to make use of the task tracking22:27
seiflotfywell this part is coverd by zg22:28
seiflotfyand yes they can be aware of it using dbus22:28
kklimondaI'd love to see a rhythmbox and banshee ported to tracker for example - right now I've all my music rated and I can't neither change application or move it on the disk without loosing this data. So I love the idea but without putting a lot of resources into this we won't get a good results.22:28
thorikewell for rhythmbox for example there is or soon will be a zg data provider22:29
robtaylorkklimonda: +1 from me :)22:29
seiflotfythekorn, correction its there :)22:30
seiflotfythorike, kklimonda rhytmbox dataprovider exists now22:30
RAOFThe Banshee dev's concern about that is whether tracker is fast enough.22:30
robtaylorRAOF: *nod*22:30
thorikekklimonda, yep your point is a good example of making use of tracker22:31
thorikezeitgeist, after listening to you for a while and checking which songs you skipt etc22:31
robtaylorRAOF: its an intersting one that. I see quite often people saying 'we put a lot of work into making our database fast, tracker couldn't possibly be fast enough'22:31
RAOFBecause quite a large set of Banshee features involves querying their sqlite database.22:31
thorikewould then also be able to design playlists according to this data and not only according to static song properties22:32
robtaylorRAOF: we've got a speed test suite now, so i think that's a task for me to show them the money ;)22:32
RAOFrobtaylor: I don't know if tracker *is* fast enough; I don't think there's any architectural reason why it couldn't be.22:32
robtaylorRAOF: tbh, I don't really know either. I'm looking forward to running the test suite when i get a chance ;)22:33
robtaylorRAOF: its very much about the use cases22:33
RAOFIt would be quite easy to sync banshee's data with tracker, but it'd be quite a lot of work to actually replace the DB backend with tracker.22:34
thorikeand these playlist recommendations would also change over time22:34
robtaylorRAOF: yep. I was just about to say, the real part of the story that's missing with tracker rigt now, is great developer apis for gnome22:34
thorikewhat you liked a year ago is different from now sometimes22:35
robtaylorthorike: absolutely22:35
thorikebut zg also allows to scope back to "last year" and tell you22:35
thorikesame for documents and mails, websites22:35
thorikeyou were using on a project22:35
thorikethey are meaningful mostly while the project is running22:36
thorikeand then querying what was related to them while the project was running is important22:36
thorikeand is currently not supported22:36
thorikee.g., you asking a year later22:36
kklimondathorike: how easy would it be to back all this data up? most users don't have a full disk backup running and zg sounds like something that shines only if it has enough data from long period.22:37
thorikeI have this document at hand, which I know I generated during that project22:37
thorikewhat was related to it then22:37
thorikeand not with the new "noise" that has come since the project22:37
thorikekklimonda, it's one sqllite DB22:37
thorikeso it's an easy backup22:38
thorikeor what do you mean?22:38
thoriketo bootstrap also "old data" can be imported / simulated wherever possible22:38
thorikee.g., chat logs are easy22:38
thorikekklimonda, and you're right,22:38
thorikezg shines most, when there is a long history22:38
thorikeand with our design, it's not a problem to have years of data in there22:39
thorikeand query it efficiently22:39
didrocksthe concept is interesting but I still really think that for UDS, you should bring some existing and precise use cases/scenarios working (or at least some proof of concept of them) to make the added value more visual22:39
thorikedidrocks, that's true22:39
thorikewe'll try22:39
kklimondathorike: well, is it possible to for example sync the database using desktopcouch (or another service) and then, after the format, sync it back and get back on the track - how would it handle missing files or files that are missing right now but are going to be restored from backup in a moment.. same for tracker - does it remove "file" from the database completely or can it be restored if the22:40
kklimondafile with the same hash shows up eventually.22:40
didrockskklimonda: I was thinking that at first, but maybe my workflow on my laptop isn't the same than the one on my netbook :)22:40
kklimonda(that's how I clean my desktop for example - I move $HOME to ../BACKUP and then move back only important files like music or photos or some of the configuration I actually care about22:41
thorikekklimonda, well couchdb is rather slow compared to our DB22:41
thorikebut synching between several of your personal DBs on devices is on our list22:42
thorikeit's important considering also mobile phones22:42
thorikere missing files: they are shown and considered22:42
thorikeshown in grey so you realize they are not there22:42
robtaylorok guys.I'm afraid I need to head to sleep now!22:43
rickspencer3bye robtaylor22:43
rickspencer3thanks for your time!22:43
robtaylorrickspencer3: seiflotfy: thanks for this. Its great to get the discussion flowing :)22:43
thorikebye robtaylor22:43
robtaylorrickspencer3: thanks for your time too22:43
robtaylorthorike: thanks, see you on #zeitgiest ;)22:43
thorikeyep :-)22:43
thorikekklimonda, and you have both options22:44
thoriketo delete files completely from database or just not showing them, aka blacklisting22:44
thorikeand yes, they can be restored if the file with the same hash shows up again22:45
thorikezg DB keeps the activity on the items, not the items itself22:45
thorikeand the activity is still meaningful to oneself, even if the items are temporarily not available22:45
kklimondathorike: only name? no hash? for example if I save an attachment from email, then remove it and then get the same attachment from some site under the different name it wont show up?22:45
robtaylorrickspencer3: oh, i forgot to say, take a look at this TED talk for some interesting MS stuff on linked data - http://www.ted.com/talks/gary_flake_is_pivot_a_turning_point_for_web_exploration.html22:45
* robtaylor really goes to sleep now ;)22:45
thorikei.e. for remembering "what have I done two weeks ago"22:46
kklimondaok, that feature sounds like O_PONIES ;)22:46
thorikecurrently people use their calendar, mailtool, filesystem and whatsoever for that (fragmented user experience)22:46
thorikein zg, it's all there22:46
rickspencer3ah gary flake22:46
rickspencer3the one, the only, the fired22:46
thorikekklimonda, same hash, it would show up22:46
kklimondathorike: oh? great22:47
thorikebut you're touching an important issue22:47
thorikewhat is "the same file"22:47
thorikesooner or later, we'll need some layer that will clarify this22:48
thorikebeyond what we have now22:48
thorikedocuments develop and people might want to have this shown22:48
kklimondaright - for example user can send a document to a friend and get it back - the content is different, the name can be different and so hash but the document is probably the same..22:48
kklimondamhm22:48
thorikeas in two weeks ago, a paper I was writing had some content I had later deleted22:49
thorikeyep, we'll tackle that22:49
thorikebut have not a real working solution yet22:49
* didrocks waves goodbye, it's late!22:49
* thorike waving back :-)22:50
kklimondag'night didrocks22:50
didrockskklimonda: thanks, you too ;) bye thorike!22:50
thorikekklimonda, so we need to work on "document identity"22:50
thorikeneeds versioning22:50
thorikemaybe22:51
thorikecan be found out sometimes by similarity22:51
thorikecontent wise22:51
thorikebut sometimes also needs the trace22:51
thorikeas in copied from here to there22:52
thorikethen sent22:52
thorikeand back, again working on it22:52
thorikethe usual stuff you do, when working with someone on a paper or presentation22:52
kklimondaright.. and the remote system may not even know about zg and all this stuff.22:52
thorikesadly yes22:52
thorikein case the remote system is your system you can control22:53
thorikebut in case not it's also by design, because, you shouldn know about others file use22:53
thorike...22:53
thorikebut we have teamgeist for that22:53
thorikewhere the sharing is tracked too22:53
thorikethat will probably a gsoc project22:53
thorikehaving a data provider for telepathy22:54
thorikewhich is the protocol we use for teamgeist22:54
thorike... that's for trusted communities22:54
thorikelike google wave, but simpler :-)22:54
thorikeyou want to share s.th. with someone else, e.g. you are working on a project with him/her22:55
thorikeyou use the file and if you want your collaborator/peer22:55
thorikewill a) be aware immediately that you do22:55
thorikeand b) also have access to the file (exactly your version, which has been sent over a tube)22:55
thorikeit's implemented, but the group stuff is tricky22:56
thorikebecause you don't want to harm privacy of others22:56
thorikeit needs to be easy, predictable, expectancy conform and all that22:56
thorikeand  mutually22:56
kklimondamhm22:56
thorike?22:57
kklimondaright, I see :)22:57
thorikeI think, it's a UI issue22:57
thorikeit's amazingly cool to work with it22:57
thorikebut users might be fearing loosing control22:58
thorikethat has to be handled on ui level22:58
thorikesimple customizing, easy to understand22:58
kklimondathorike: btw, is zg something that both gnome and kde developers are interested in?22:58
thorikegnome: definite yes, kde: interests many developers, but the nepomuk people claim that we should be using their ontology for events too22:59
kklimondathorike: right - the whole concept is great but a bit overwhelming especially when you add the sharing into that :)22:59
thorikewhich doesnt realy work22:59
=== robbiew is now known as robbiew_
kklimondaoh right, they have nepomuk :/23:00
thorikenepomuk is nice for documents, static information23:00
thorikealthough introducing quite some overhead23:00
thorikeand there is not really an event ontology as we designed in the last two months23:00
thorikewhich is simple and thin, and does the job greatly23:01
thorikeas soon as we have the feeling it's stable, we'll talk back to sebastian23:01
thorikethere is ongoing connection and discussion, although not daily :-)23:02
thorikekklimonda, how do you mean overwhelming?23:02
thorikegood or bad? (polarizing ;-) )23:03
thorikekklimonda, and if you want to put the concept idea in one line23:04
thorikeit's "everybody knows, what I do, but I don't" :-)23:04
thorikegoogle, amazon, travel agencies, ...23:04
thorikethey all log your behavior23:04
thorikeand analyse it to their benefits23:05
kklimondathorike: both zg and tracker sound like something great if they get enough tracktion but both present a huge change to the way users use their computers.23:05
thorikethere is no similar means for that on the desktop yet23:05
kklimondathorike: maybe I feel like that because it's all fragmented and when the missing UI bigs emerge it will look clearer23:05
kklimondas/bigs/bits23:06
thorikeyep, well using location information was also pretty much a change and people adapted quickly23:06
thorikewe do the same, but with the time-dimension23:06
kklimondathorike: it sounds much more exciting that gnome-shell in my opinion :)23:06
thorikekklimonda, wow thanks for the flowers23:06
thorike:-)23:06
thorikekklimonda, and I can see the fragmented feeling23:07
thorikedid you use gnome activity journal, yet?23:07
thorikeand the zg docky, which shows most or recently used23:07
thorikethat gives you a preview feeling I think23:07
thorikeespecially with zeitgeist's next release, which will show related items23:08
thorikeneeds integration to gaj (gnome activ. journal) though23:08
thorikewe're working on it23:08
thorikeif we get people surprised like "oh yeah cool, that's true this file which I put in another folder is really related to the doc at hand"23:09
thorikethat will work, if we get this "aha effect" (as we say in german) :-)23:09
thorikelike heureka23:09
kklimondathorike: my usage patterns sucks and I do way too much things using terminal so it doesn't work for me :/23:10
thorikewell, some of us do too23:11
thorikeso eventually we'll need to tackle that23:11
thorikebut that will need kernel support23:11
thorikefirst, we need to proove it useful23:11
thorikethen come the diamonds :-)23:11
thorikekklimonda, you usage pattern sucks? .. you mean you do everything at once? ... then you'd be our hard core test case for segmenting with the association algorithms .. apriori, etc23:12
kklimondathorike: yeah - when I don't do anything work-related I just jump from one topic to another :)23:14
kklimondait's a terrible habit I'm trying to overcome though23:14
thorikewell, why .. you'd be the ideal test case ... and why change if it works for you?23:16
thorike:-)23:16
thorikethe computer should change to support that behavior better23:17
thorikeit's like with filers and pilers23:17
thorikeit's reproduced in studies again and again .. people are different23:17
thorikeand computers should cope with it23:17
thorikeand the full power of zg will even better show on devices more mobile than netbooks23:21
thorike"what did I do three days ago at this bus stop (gps position)?"23:21
thorikethat's easy for zeitgeist23:21
thorikeand is not possible without events23:21
thorikebut even now we have enough usefullness23:22
thorikemy parents use gnome activity journal on their desktops23:22
thorikeas they have trouble with folders and remembering file names23:22
thorikegaj does the job for them23:22
thorikeif they want to re-find what they did23:22
thorike... well it's getting late23:22
kklimondaok, I guess I'm at my limit of processing any informations :)23:23
thorikewas nice talking to you, kklimonda, rickspencer323:23
* thorike waves good by23:23
thorikekklimonda, me too :-)23:23
thorikelets take it up at some other time23:23
thorikemeet you at #zeitgeist or #gnome-zeitgeist23:23
kklimondathorike: g'night :)23:24
thorikekklimonda, you too :-)23:24

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!