[02:56] I think I like coding during night because nobody's disturbing me. That's much more comfortable. [02:56] No phone calls, no gtalk chats, no nothing. [03:04] karni: ping [03:04] >:) [03:04] duanedesign: hey! [03:04] heheheh [03:04] it's ok ;) [03:04] :D [03:04] it's just music, code, and.. not so evil duanedesign [03:04] karni: i was aobut to catch a few zzzz [03:04] :) [03:05] karni: yeah i am having trouble getting used to ircclouc [03:05] irccloud* [03:05] karni: i have used irssi for so long [03:05] what do you mean :)? oh. it's another irc client? [03:05] oh yeah [03:05] oh, yeah. That is the evilduanedesign [03:06] I see heheh :) [03:06] this is me on my irssi/screen setup on my VPS. evilduanedesign is me when I am using irccloud.com, or my phone [03:07] it has a neat interface and if i did not have a VPS or server I would really consider using it all the time [03:07] oh, it looks kinda nice [03:08] it has the advantage of being always on [03:08] so you get the scrollback like you do with ruuning irssi in GNU Screen [03:08] oh this is cool, like screen! [03:08] right [03:08] * duanedesign nods [03:08] does it integrate with desktop notifications ;D ? [03:09] ahhhh [03:09] I might give it a try one day [03:09] I kinda like the raw edge and terminal-like funcitonality of screen+irssi, but this is definitely worth checking out [03:10] yeah [03:10] and irssi is very fast [03:11] as fast as you can type, indeed :) [03:11] GUI tend to, in my experience, lag a second or so [03:12] yup [03:13] gotta get back to eclipse, I'm getting more red underlines than I should [03:13] ;) [03:13] have some good sleep duanedesign! [03:13] thanks happy coding karni [03:13] thanks duanedesign ! [03:13] o. [03:14] (-.-)~zzZZ [03:26] oh god [03:26] it's after 4 am [03:26] I've just spent over 4 hours writing code in Qt.. for fun.. [03:34] kklimonda: :) [03:35] kklimonda: any eye-candy results? [03:38] I guess you just fell asleep ;] [03:38] karni: not really, it's a simple media player [03:38] to watch movies from gomtv [03:38] :) [03:39] but I have to say I've never had so much fun with C++ in my life. [03:39] It's like I've discovered it for the first time. [03:41] and I managed to break the notification, ops. [03:42] :D [03:47] triaged. [03:51] wow, I just ran out of ram.. [03:52] ok, I still have like 600M left.. but still.. [04:22] okey, time to get some sleep. night guys! [10:24] Ah, good morning, all! :D [10:34] hello all [10:41] 09:40:35 < jesterina> what's the new dos2unix called again? [10:41] 09:40:56 < idimmu> isnt it just called [10:41] 09:40:58 < idimmu> dos2unix ? [10:41] Yes :-/ [10:42] Unless you want to re-implement it in sed or something :-s [10:45] Er, wrong channel/o\ [11:32] aquarius: good morning [11:37] hi duanedesign [11:40] aquarius: I am looking for some advice on an application I am making. I realize you are awfully busy but on the off chance that you had time in the near future I thoght I would ask. [11:45] aquarius: so if you find yourself with the rare opporunity of a little free time maybe take a peak https://launchpad.net/stipple [11:47] duanedesign, I discovered that at the weekend; haven't had a chance to try it yet! [11:47] aquarius: [11:47] oops [11:50] duanedesign, I like the idea of it :) [11:51] aquarius: cool. Just kinda stuck thinking "This could be doing its job in a better fashion". Just not sure what that better way is. I thought this is where I could benefit from some advice from someone who has been down the road before [11:51] duanedesign, talk to me about the issues you're thinking about [11:52] (I don't want to say: this is how I would design such a thing, because it's not my project. But if you say, I am struggling with this problem, I am more than happy to give you my thoughts) [11:52] aquarius: my biggest concern is you currently launch the app. Click 'Sync to DB' That puts the .config files for the selected apps in Couch [11:53] aquarius: then you launch the app on Computer B. Open the app and click "Sync from DB' [11:54] ok [11:55] aquarius: I think it should probablly be a daemon? [11:56] I do not think so. [11:56] If you write a daemon which (a) runs all the time (b) requires Ubuntu One (c) monitors config files for changes (d) syncs them between machines... [11:56] then what you've done is reinvented the syncdaemon. :-) [11:56] storing the occasional attachment in desktopcouch is OK, but it's really not for storing many many files (that's what file sync is for). [11:57] * duanedesign nods [12:01] aquarius: that makes sense. might be a better idea to switch to file sync. Because some of the plugins are syncing directories. like the emacs plug in syncs your .emcs file and emacs.d directory [12:01] indeed. [12:02] if you're syncing files, using filesync is the way forward; that's what it's for :) [12:02] maybe creating a custom UDF and still having manual "back up my config"/"restore my config" (which copies them in and out of the UDF)? [12:04] aquarius: yeah, that makes sense. Thinking through i think that solves a lot of my iinitial concerns [12:05] aquarius: thank you that was very helpful. [12:10] good morning everyone [12:12] o/ [12:19] \o [14:02] me [14:02] dobey: mandel: stand up in the place where you are! [14:02] me [14:03] thisfred: it's the end of the world as we know it. [14:03] * DONE bug #723227 https://code.launchpad.net/~thisfred/ubuntuone-client/use-single-notification/+merge/52147 [14:03] * DONE bug #702183 https://code.launchpad.net/~thisfred/ubuntuone-control-panel/attention-seeking/+merge/52131 [14:03] * NEEDSREVIEW bug #729055 https://code.launchpad.net/~thisfred/ubuntuone-client/filenames-in-notifications/+merge/52483 [14:03] * INPROGRESS bug #702172 https://code.launchpad.net/~thisfred/ubuntuone-client/quota-notifications [14:03] * INPROGRESS bug #728722 https://code.launchpad.net/~thisfred/ubuntuone-control-panel/dbusify [14:03] Launchpad bug 723227 in ubuntuone-client (Ubuntu Natty) (and 2 other projects) "SD creates new notifications instead of updating one (affects: 1) (heat: 283)" [Medium,In progress] https://launchpad.net/bugs/723227 [14:03] * TODO bug #702007 [14:03] * TODO bug #730661 [14:03] NEXT: mandel [14:03] Launchpad bug 702183 in ubuntuone-client (Ubuntu Natty) (and 2 other projects) "Syncdaemon needs to open the control panel in the background and change the launcher icon to urgent when the user exceeds their quota (affects: 1) (heat: 6)" [Medium,In progress] https://launchpad.net/bugs/702183 [14:03] dobey: I feel fine [14:03] Launchpad bug 729055 in ubuntuone-client (Ubuntu Natty) (and 2 other projects) "File names should be shown on notifications (affects: 1) (heat: 6)" [Medium,In progress] https://launchpad.net/bugs/729055 [14:03] Launchpad bug 702172 in ubuntuone-client (Ubuntu Natty) (and 2 other projects) "Syncdaemon needs to send a notification when a folder shared to the user exceeds the owning user's quota (affects: 1) (heat: 6)" [Medium,In progress] https://launchpad.net/bugs/702172 [14:03] Launchpad bug 728722 in ubuntuone-control-panel (Ubuntu Natty) (and 5 other projects) "control panel should have a .service file so it can be opened through dbus (affects: 1) (heat: 6)" [Medium,Triaged] https://launchpad.net/bugs/728722 [14:03] Launchpad bug 702007 in desktopcouch (Ubuntu Natty) (and 2 other projects) "get_all_records does not return records with their attachments (affects: 1) (heat: 6)" [Medium,Confirmed] https://launchpad.net/bugs/702007 [14:03] Launchpad bug 730661 in ubuntuone-client "progress bar does not show up in Unity or something (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/730661 [14:03] DONE: bindwood reviews. ubutu_sso teting. [14:03] TODO: ubuntu_sso UI nad installer UI [14:03] BLOCKED: no, confused by twisted trial, yes [14:03] dobey: please [14:04] λ DONE: started mp3 install checking [14:04] λ TODO: bug #727558, mp3 install in banshee [14:04] λ BLCK: Need webdev help. [14:04] Launchpad bug 727558 in libubuntuone (Ubuntu Natty) (and 3 other projects) "Need to notify user when Purchased Music folder is not subscribed (affects: 2) (dups: 1) (heat: 14)" [Medium,Confirmed] https://launchpad.net/bugs/727558 [14:11] dobey, thisfred: can I have a review for https://code.launchpad.net/~mandel/ubuntu-sso-client/run_tests_windows/+merge/52081  [14:11] mandel: sure, can I have one (or two) for https://code.launchpad.net/~thisfred/ubuntuone-client/filenames-in-notifications/+merge/52483 ? [14:11] thisfred: on it [14:27] wtf keeps causing ubuntuone-syncdaemon to keep respawning when it crashes? [14:39] dobey, looks like gir from libdee to me === Chex_ is now known as Chex [14:49] gord: it's because gir1.2-unity-3.0 isn't depending on the dee gir package, from what i see, and seems to be correct from what kenvandine said [14:52] there is a warning in the libunity build about the dee namespace [14:52] dee-1.0.vapi:4.1-4.13: warning: Namespace Dee does not have a GIR namespace and version annotation [14:52] namespace Dee { [14:53] Ugh. natty-update reboots aren't making things much better lately. [14:53] which seems to be why the include isn't make it into the gir [14:53] the question i have is if installing that gir is enough [14:53] gord: or were you suggesting it is what is causing the crashing? [14:53] dobey, yup, but it seems that the gir wasn't installed here as you suggested, installing it stops the crashing [14:53] gord: it might cause the crash, but it shouldn't cause a respawn. it should just crash and burn [14:54] gord, great.. so until we can figure out why is keeping the include from getting added, i can just add a depends [14:54] kenvandine: well you had another crash, so it's probably not enough, but not sure what is :( [14:54] that seems different [14:54] maybe i should change the recommends to a conflicts [14:54] :( [14:55] kenvandine: the include needs to be specifide in the makefile probably [14:55] it is something that changed yesterday [14:55] dobey, no, that isn't the problem [14:55] dee-1.0.vapi:4.1-4.13: warning: Namespace Dee does not have a GIR namespace and version annotation [14:55] that is why it isn't getting added to the gir [14:55] kenvandine: i was looking into fixing the webkit gir similarly, but all the ones it depends on are also broken [14:55] dobey, well syncdaemon-exceptions definitely had an exception on looking up the gir for dee and it was definitely restarting constantly - thats all i know ;) [14:55] kenvandine: the vapi has nothing to do with the gir [14:55] yes it does [14:56] libunity is written in vala [14:56] it generated the gir from the vapi [14:56] the gir is generated from the vapi? [14:56] ugh [14:56] yes [14:57] so the dee gir is broken also [14:57] is libdee in C or Vala? [14:57] C [14:58] where is libdee hosted? [14:58] lp [14:58] https://code.launchpad.net/libdee [14:58] no it isn't :) [14:58] https://code.launchpad.net/dee [14:59] sigh [14:59] yes it is [14:59] :) [15:00] grmbl [15:00] dobey, it looks like the gir generation stuff can be simplified quite a bit [15:00] this was all done a while back with lots of work arounds [15:00] oh yeah [15:00] the dee gir build is nasty [15:01] yeah [15:02] this was done back when gir generation was undergoing lots of change [15:03] i'll fix that up in a few [15:03] in a meeting now [15:35] guys. you've just blown my mind. I know I've been in here and said this before, but you have so much NON. Stuff I can't even find on blogs. [15:35] I am amazed :D screw the haters, u1 rules [15:36] heh [15:41] and $2 cheaper than amazon [15:41] ive never bought digital music before, this .. doesn't feel right [15:41] where is the packaging.. where is the disc? [15:41] * lamalex shudders [15:43] hmmm but it has me logged in with my gmail account and not my u1 account which is on my canonical.com address [15:43] heh [15:43] dobey, how do I re-sign in? [15:44] I've been toying with starting a company that sells slips for digital downloads, but in LP packaging. So the millenials can have their digital music but still have some of the good parts of buying records [15:44] uhm. do you have 2 accounts? [15:44] or in 7" packaging more likely [15:44] dobey, i dont think so [15:45] actually I guess I maybe could [15:45] the canonical.com one is the one signed in according to the control panel [15:45] heh, that's why Apple did that thing with the box set junk, where you get a bunch of artwork and junk with the mp3s [15:46] ugh freaking apple. i swear i thought of it first :P [15:46] amazon also has that for some albums now too [15:46] cd packaging though, right [15:46] I'm talking LPs. Full 12" record sleeves [15:46] it's all digital, no physical anything [15:46] OH [15:46] it's like a magical tarball with a bunch of useless crap in it :) [15:47] yeah I'm talking about the physical packaging [15:47] you know, it's like open source, but you pay for it :P [15:47] anyway [15:47] a 12" record sleeve with a slip of paper inside the size of a record with yor download code on it [15:47] is your gmail address your default contact on lp? [15:47] nope [15:47] canonical [15:47] oh weird [15:48] so where are you seeing gmail being used? [15:48] in the checkout [15:49] interesting [16:01] soo if I buy this will it be associated with my account? [16:01] i dont want to pay $8 and then lose it, you know? [16:02] i am not sure [16:02] joshuahoover: ^^ do you know anything about that? [16:03] lamalex: did your account used to have the gmail address as the default at one time? [16:06] lamalex: did you see my last question? did your account used to have the gmail address as the default at one time? [16:07] joshuahoover, no I hard locked :( [16:07] my lp account had gmail as default months ago [16:08] lamalex: probably when you first signed up for a u1 account it did [16:08] so that's what got pulled into the u1 accounts db [16:08] ok [16:08] so it should be the same account [16:08] u1 shows canonical though [16:09] lamalex: what does it show on https://edge.one.ubuntu.com/account/ [16:09] on the web and in the gtk client shows canonical [16:09] which is why i'm unsure [16:10] lamalex: let me look at your account [16:26] lunch time, bbiab [16:26] Chipaca, dobey: i seem to recall a bug about this issue lamalex is experiencing but i can't find it :( [16:26] joshuahoover: 7d are keeping a copy of the account details [16:27] joshuahoover: not sure if there is an api to update we're not using, or if we need to ask them to stop caching [16:27] Chipaca: ah, ok...which is (obviously) confusing...i'll make a note to ask 7d on our call with them this week [16:29] joshuahoover: ty [16:30] lamalex: so, this is something that needs fixed (either by us or 7digital), but your purchases will be on your ubuntu one account under your canonical email address === zyga is now known as zyga-afk [16:31] joshuahoover: ah, sorry, should've said: what i just said was after i went over it all with lamalex in private :) [16:31] (just to make sure this was the case) [16:31] Chipaca: ah, ok :) [16:34] thanks guys :) [17:20] CardinalFang: ping [17:39] joshuahoover, hi [17:49] CardinalFang: hi! anyway we can get your branch approved since you incorporated the feedback from the first review? https://code.launchpad.net/~cmiller/ubuntu/lucid/desktopcouch/lp726597/+merge/52249 [17:50] joshuahoover, I'll ping the previous reviewer. I thought he'd come back by now. [17:50] CardinalFang: cool, thanks! :) [17:54] kenvandine, who should I ping for SRU reviews? [17:55] CardinalFang: you should subscribe ubuntu-sru to the bug iirc [17:56] https://wiki.ubuntu.com/StableReleaseUpdates#Procedure [17:57] CardinalFang: hrmm, the merge is wrong though. it needs to get uploaded to lucid-proposed (which i think anyone can do), an archive admin approve it there, bugs verified, then it will get released to -updates if everything passes [17:58] CardinalFang: also, there's already a 0ubuntu3.1 in -updates i believe [17:58] Since yesterday's natty updates, ubuntuone-syncdaemon uses all my CPU and crashes and restarts constantly. [17:58] nhaines: apt-get remove gir1.2-unity-3.0 for now [18:00] dobey: that fixed the problem immediately. :) [18:00] CardinalFang: https://launchpad.net/ubuntu/lucid/+source/desktopcouch/0.6.4-0ubuntu3.1 [18:00] It doubtless introduced other problems, but since this is a laptop I'll deal with that for a bit. :) [18:01] nhaines: i don't think anything else is using unity via gir right now, so it shouldn't have caused any more issues by removing it [18:02] dobey, hi, I am not really here but what's the bug number for current unity-introduced syncdaemon crashing? [18:02] i don't remember [18:03] looks like bug #731023 [18:03] Launchpad bug 731023 in libunity (Ubuntu) "ubuntuone-syncdaemon crashed with ImportError in get_interfaces_for_object(): No module named Dee (affects: 104) (dups: 42) (heat: 450)" [Critical,Confirmed] https://launchpad.net/bugs/731023 [18:05] dobey, thanks, will do a bugpattern now [18:08] great [18:08] vim is broken [18:09] i've been saying that for years! [18:09] [18:10] dobey, no, well, today it is broken due to package dependency and this is the first time my lovely vim is handled by command-not-found [18:13] 10:08 < rye> great [18:13] 10:08 < rye> vim is broken [18:13] 10:09 < dobey> i've been saying that for years! [18:13] 10:09 < dobey> [18:14] Yay, I hate using PuTTY and Windows [18:14] ok, bugpattern is set [18:19] http://ubuntuone.com/p/gjv/ [18:19] dobey, will add your workaround to the bug report [18:20] i did [18:20] what i really want is for it to stop respawning [18:22] dobey, hm, yes, argh, did not update the page before posting :( [18:22] are we aware that U1 is DOSing our users on login? [18:22] apw, yes, bug #731023 [18:22] Launchpad bug 731023 in libunity (Ubuntu) "ubuntuone-syncdaemon crashed with ImportError in get_interfaces_for_object(): No module named Dee (affects: 105) (dups: 43) (heat: 458)" [Critical,Confirmed] https://launchpad.net/bugs/731023 [18:22] apw, unity is to blame [18:24] it is restarting the daemon on crash every 30s? [18:24] every 30 seconds? [18:24] that seems like a long time [18:24] and not in line with what everyone else is seeing [18:25] dobey: so i tested bug #661292 with the latest ubuntuone-client package installed 1.4.6-0ubuntu2 and it still doesn't appear to work without killing nautilus [18:25] Launchpad bug 661292 in ubuntuone-client (Ubuntu Maverick) (and 4 other projects) "Nautilus is not aware of published files (affects: 7) (dups: 4) (heat: 40)" [Medium,Fix committed] https://launchpad.net/bugs/661292 [18:25] dobey, thanks for the tip about the version. It appears there's a lp:ubuntu/lucid-updates/.... source-package tree. [18:25] well it may be more often, but whoever is respawning it is fail regardless of why its restarting [18:25] joshuahoover: did you wait? [18:26] apw, sudo apt-get remove gir1.2-unity-3.0 [18:26] CardinalFang: yeah there is, also one for lucid-proposed [18:26] dobey: yeah...i waited...tried it a few times just to be sure [18:26] huh [18:26] le sigh [18:29] joshuahoover: are you using i386 or amd64 in the vm? [18:29] dobey: i386 [18:30] ok, bugpattern is already in action [18:31] joshuahoover: hrmm, something is not right then, because my patch is being applied in the build log, so it's there [18:31] dobey: is there something i can check? [18:32] dobey: i know i have the latest proposed package installed according to apt-cache policy ubuntuone-client but is there something else i should check? [18:33] joshuahoover: and it's 0ubuntu2? [18:33] dobey: yep [18:34] i don't know [18:34] well [18:34] we won't receive any more bug reports for this issue [18:34] i think something is maybe not right in your vm perhaps [18:35] because apport breaks on that [18:35] so sweet [18:35] i wonder if rick can verify it [18:52] * dobey uploads a quick change to u1client to make the unity caused crashiness stop [19:28] dobey: yay [19:30] help ubuntuone-syncd is taking 100% CPU. it is respawning every 1.5 seconds. I cannot work. What can I do? [19:30] I could uninstall all ubuntuone, but I just want to stop it today [19:31] sinzui: uninstall gir1.2-unity-3.0 [19:31] okay [19:32] * dobey wonders if the u1 update is published yet [19:32] looks like it is for i386/amd64 at least [19:33] hooray me for making that upload [19:34] \o/ thanks dobey [19:35] :) [19:35] Chipaca, it's been 3 hours and I still don't have my album :\ [19:35] it made a music folder [19:35] but that's it === facundobatista is now known as rule [19:42] lamalex: in ubuntuone-control-panel-gtk in the folders tab, is the music folder subscribed? [19:42] i suspect it isn't :) [19:44] that appears to be the problem [19:44] known bug? [19:46] yeah [19:46] well "bug" [19:47] but yeah, i know of it :) [19:53] beuno: hi there! I started late today and would like to finally move the sync to action queue, which is hell important - how about I provide the apk then? I'll be working late today, so perhaps I'll manage before tomorrow evening. [19:53] karni, hey, I've been off yesterday and today [19:53] sure, send away :) [19:53] rule: available for one question :) ? [19:55] beuno: ok, I'll continue to work on that and sync should be much more robust. I'll get back to you once that's ready. [19:55] __lucio__: perhaps you're available for one question? === ajmitch_ is now known as ajmitch [19:55] karni, I'll be back later, when my girlfriend isn't looking [19:56] otherwise, tomorrow [19:56] beuno: haha sure! I don't think I'll finish in 4-5 hours at least, so is safe to assume I'll ping you by tomorrow [19:57] karni, tell me === rule is now known as facundobatista [19:57] facundobatista: thanks. Is there anything holding me from thinging about sync as ActionCommandQueue commands? Let me shortly justify. [19:58] thinging? [19:58] facundobatista: u1 desktop client does some work in the background, and then uploads/downloads some data, etc. so it's not very imporant what's it's detailed state. [19:58] thinking, sorry ;) [19:59] facundobatista: what I want to do is implement such commands: (apart from ListVolumes and ListShares), SyncVolumeFromScrach, SyncVolume [19:59] facundobatista: that way, I can put neccessary Sync* commands on the queue, and each will further add any up/downloads that are neccessary to sync [19:59] karni, for those there is the GetDeltas [19:59] *are [19:59] right, yes. that's not what I mean. [19:59] ok, I'll try to be even more precise :) [20:00] what the SyncVolume will do? [20:00] all the sync logic for that volume, and add any uploads/downloads neccessary to the queue [20:01] what this gives me is much easier handling of the service that is running to do all that work - when the queues are busy [20:01] I easily update the UI, add notifications, etc [20:01] karni, well, we explicitly put the sync logic in Sync, not in the queue [20:01] karni, it was our design decission, you can mix that.... but I don't know if it's a good idea (maybe it is!) [20:01] if I do the sync logic off the queues (i.e. exactly, in Sync, not queue) then my queue is idle [20:02] facundobatista: I see. Right. Too bad I don't know as much as you guys! ;) [20:02] facundobatista, karni, I'm about to need to create a volume, if one does not exist. The path part worries me, in that I don't know if I need the full path as u1sdclient shows it to me. On a phone, i don't know what comes after /home/- and before -/Music (or whatever). I guess I need to read up how this works. [20:02] facundobatista: I'm quite sure that handling sync logic that way might streamline how I manage all actions (commands) [20:03] karni, I wouldn't take the decission of where to put the Sync logic based in the notification of when the queue is idle or not... the former is way more complicated and the decission should be based on more heavy stuff [20:03] CardinalFang: I get it, I'll get back to you shortly once I finish with facundobatista ! ok? [20:04] facundobatista: That sounded a little childish of me. What I meant is that, syncdaemon on Android is very different form the one on desktop. [20:04] facundobatista: It is quite often shutdown/killed, mostly when the app is idle and not in the foreground [20:05] karni, so? [20:05] facundobatista: I was thinking about it for quite some time. And I thought that if I can put 'Sync' command on the queue, and it'll do the rest for me [20:05] karni, ok [20:05] then all sync-logic is not done on the background, and kind of meaningless for the rest of the app [20:06] if I ran for few hours, then I wouldn't mind having the sync logic totally disconnected from all the rest, and just dispatching events [20:06] but in that case, I feel it gives the sync-logic a somewhat strong presence and importance to the service [20:07] purely based on how it's implemented (as ActionQueueCommands) [20:07] facundobatista: What's your opinion? [20:07] karni, I'm not following how where the logic is put is affected by the application being often shutdown/killed, or running idle [20:08] Can I tell u1 not to send notifications? it sends so freaking many [20:08] facundobatista: The service has a 90-second idle timeout (*if* it is not in the foreground) [20:08] facundobatista: The sync (currently) is run on a separate thread. [20:08] lamalex: install the nightlies [20:09] dobey, thanks [20:09] oh [20:09] lamalex: i think it has an option, and it's also a lot better about notifying [20:09] facundobatista: If the user leaves the app, and the sync becomes slow (which has been the case for some time), then the service itself shutdown [20:09] although the sync was running. You could think 'it's a bug' - certainly [20:09] facundobatista: It's easy for me to check a boolean flag isSyncRunning or something like that [20:10] facundobatista: but it's much easier to put something on the queue, which, on it's own and automatically not only keeps the service running, but also [20:10] provides feedback to the UI (*if* it's present) [20:10] ok [20:10] facundobatista: I could decouple those things. I certainly want to improve my sync logic at present. [20:11] karni, ok, but you're taking distance of the desktop's client architecture [20:11] dobey, our ubuntuone-client-crashdb.conf crashes apport now [20:11] dobey, preparing branch now [20:11] uhm [20:11] rly? [20:11] why? [20:11] also [20:11] shen me niao [20:11] shen me niao: command not found [20:12] it's "wtf" in mandarin [20:12] facundobatista: I'm aware of that. That's why I'm asking of this rather big steop is sensible. Still, having in mind, that Android is indeed _very_ far from the desktop impl. [20:12] dobey, https://bugs.launchpad.net/ubuntu/+source/apport/+bug/731526 [20:12] rye: Error: Bug #731526 is private. [20:12] huh? [20:13] bug #731526 ? [20:13] Launchpad bug 731526 in apport (Ubuntu) "apport-gtk crashed with KeyError in get_crashdb(): 'bug_pattern_url' for ubuntuone bugpattern (affects: 1) (heat: 10)" [Undecided,New] https://launchpad.net/bugs/731526 [20:13] CardinalFang: I'm getting around to your question. I'll check if makeVolume is implemented in u1-java-sp [20:13] karni, ok [20:13] kuhm [20:13] rye: that seems like a regression in apport [20:14] CardinalFang: ok, CreateUDF, there it is. Indeed it takes the path as argument. [20:14] dobey, no, they have switched to bug_pattern_url from bug_pattern_base since now they have single bugpatterns.xml file. [20:14] rye: yes [20:14] rye: and if old files cause a crash, that is a regression [20:14] karni, a relative path, maybe. [20:14] because it didn't crash before [20:15] if it's not using it, it should ignore it, not crash. [20:15] CardinalFang: so your question is, where to put this on the phone and what path to provide to CreateUDF, is that right? [20:15] karni, yes. [20:15] CardinalFang: ok, gimme 1" [20:15] dobey, it also affects indicator applet and indicator-appmenu [20:15] dobey, well, i think we need to update our thing and let apport guys fix the default values for bugpatterns [20:15] yes, apport needs fixed [20:16] the code is just broken [20:16] it should try: except KeyError: there [20:16] CardinalFang: com.ubuntuone.android.files.syncdaemon.SyncDaemon.init() -- on the phone, you can use sharesDir (just a name. same as UDFs dir) [20:17] CardinalFang: this, in fact, is simply /mnt/sdcard/u1/ [20:18] now how about the path provided to the server.. [20:20] CardinalFang: com.ubuntuone.android.files.syncdaemon.SyncManager line 502 - when _receiving_, I trim the "~/" from path. would it be safe just to append that to a folder name and submit as the path of your UDF ? [20:23] facundobatista: how about I spend few hours on that and see how well this serves the Android app? I can always improve this later. [20:23] karni, ok [20:24] CardinalFang: please ask as much as you need, as in 3 days I will be probably unavailable for 9 days [I'll _try_ to chech my mail though] [20:24] karni, Okay. Thanks. I think I'm mostly good. [20:25] CardinalFang: I'll add the CreateUDF action command queue -- you'll definitely need that! won't you :) [20:25] dobey, hm, it looks like it is really broken in apport, the config does not get merged, the internal one gets simply overwritten, yes, apport is broken [20:25] karni, yeah, I will. [20:25] CardinalFang: I'm on it [20:33] IMPORTANT: This changed the format of crashdb.conf: bug_pattern_base is now obsolete, and the new attribute bug_pattern_url now points to the full URL/path of the patterns file. T [20:40] CardinalFang: I understand I can assume that preferred path is unique, since we can't create two UDF's with the same path. is that correct? [20:40] CardinalFang: So I'll also check if the Volumes table already contains such UDF and, in case it does, not queue the command. [20:52] CardinalFang: Please note the db structure await's cleanup. Most imporant table is the Files table, which contains the UDF preferred path in the 'file_name' column, and 'UDF' string as file_parent -- this is due to the fact, that [20:53] CardinalFang: both list of regular volume contents _and_ list of volumes is presented using same CustomListActivity activity. [20:54] CardinalFang: this gave much streamlining to fetch and populate data, but I'm considering making UDF and Shares list a separate activity, and they may use the Volumes table at that time. So Files table is the one you want to be looking at, if you need. [20:55] CardinalFang: in case you wonder why things are that and not the other way around sync and meta entries in the database - it's probably because I was strongly optimizing, and brought down the sync time ~7x IIRC [20:57] verterok: what is the 'name' argument in public CreateUDF(IRequestHandler handler, String path, String name) [20:57] verterok: I assume the UDF name from the last segment of the preferredPath (path, in this case) [20:58] karni, ah. Wow, 7x. [20:58] CardinalFang: I used transactioning -- and I have separate methods for initial sync, and regular sync (initial sync writes the meta to tables without checking for their presence) [20:59] I was quite happy back then, I must admit :) [21:06] karni, just to talk out loud: The remaining things I worry about: 1) Downloaded photos being caught and sent back to the server. 2) Path of photos. Discovered locally can be anywhere. 2a) Name collisions if put in one dir on server. 2b) Eventually we might want automatic sync upload of other UDFs. 3) ... something else I can't remember now. [21:07] karni, so for #1, I need to watch what's requested somehow and ignore it when the media watcher wakes up to mention it. [21:07] CardinalFang: thank you, I really appreciate your feedback [21:08] * karni processes [21:08] karni, for #2, I think I need to hash or include the path as a prefix to the file stored in the UDF. [21:09] Er, 2a. [21:09] 2b, probably can't do anything until we even think about it. Ignore for now. [21:09] CardinalFang: #1 -- you're right. That's your solution, and I'd love to see the sources. You might want to have a look on an alternative .util.MediaUtilities.java [21:10] 2b) is -- if I understand correctly -- local rescan of /sdcard/u1 for custom folders, treated as UDFs and auto synced-up ? [21:11] or *to* other UDFs, and give user the choice. that's probably what you meant, right? [21:11] karni, well, the photos will be saved by some camera app, for instance. The will be in /sdcard/Photos or something completely different. [21:11] * karni nods [21:11] So, local path can't be a factor in what I upload as. [21:12] right. and if you upload all stuff to 1 folder, we can have filename conflicts [21:12] dobey: ping [21:12] All photos generated should be candidates, regardless of local path. I don't really want to make a /sdcard/u1 . [21:13] CardinalFang: oh right. that's what is started in .util.MediaUtilities as a draft - it counds _all_ media taken since timestamp X [21:13] *counts [21:13] karni, right. And that's what I made earlier this year too. Your code is cleaner. :) [21:14] oh.. now that's a compliment ^ ^ thank you. [21:14] Mine is tightly bound to the media Observer. Needed refactoring. [21:14] CardinalFang: I'm also wondering if we should consider all media. icons and stuff populated but 3rd party apps - we obviously don't want to sync-up that. [21:14] CardinalFang: I see [21:15] ralsina: yo [21:15] hi, dobey, do you have a bug# for the "banshee doesn't tell you the music store udf is not being synced"? [21:16] CardinalFang: from Uri perspective, 1) is indeed non-trivial, unless we decide on a fingerprint that could be build from the Files meta data entry and the file itself [21:16] maybe [21:17] ralsina: yes, bug #727558 [21:17] Launchpad bug 727558 in libubuntuone (Ubuntu Natty) (and 3 other projects) "Need to notify user when Purchased Music folder is not subscribed (affects: 2) (dups: 1) (heat: 14)" [Medium,Confirmed] https://launchpad.net/bugs/727558 [21:17] thx dobey [21:18] Right. There's a good scheme for this. Blacklist some locations. Watch for .nomedia, assuming the system doesn't already do this for the Observer implementation. [21:19] ...a good scheme for "all media". [21:19] CardinalFang: I'm not sure how heavy is file computation, but duruing each upload we first compute the hash and provide that along other meta when uploading a file. We could in theory compute a hash of the file that is candidate for upload and check if we already have it stored on the server (using Files table) [21:19] ralsina: what do you think about branching libu1, and fixing trunk to use the new webkit dom/js API, and stopping the build of nightlies of it on lucid/maverick? [21:19] * karni nods [21:19] karni, Mmm, maybe. [21:19] dobey: and avoid all that weird stuff about the title, for example? [21:20] CardinalFang: I would assume the hash-computing load could be visible when a user started to download/sync whole folder with 200 pictures [21:20] CardinalFang: but I can't really say for sure. We've been uploading single files till now, and that's when I compute the hash. [21:21] ralsina: yeah, do it The Proper Way (TM) [21:21] dobey: it's ok by me. lucid/maverick would be ok because they still use rhythmbox, if I follow your reasoning correctly? [21:22] karni, I suspect IO uploading takes far longer than computing the hash. We can chunk the work to be nice to the CPU and still be ahead of the radio. [21:22] oh sure, that's definitely true [21:22] ralsina: no, but lucid/maverick have older webkit which doesn't have this API exposed (hence why the nasty hacks using the status text are in there) [21:23] CardinalFang: (u1-java-sp) com.ubuntuone.storageprotocol.HashUtils.getHashInfo is the one [21:23] dobey: well, ok, I guess. [21:24] karni, I'm going to branch that and pore over the source. It's too much of a mystery to me to think about this clearly. [21:24] CardinalFang: then I guess we can compute the hash for all sync-up candidates before we even try to upload them. [21:25] CardinalFang: whatever you need. Please let me know if I can be of any help, I'll be happy to help. [21:25] karni, thanks. You're already a huge help. [21:25] CardinalFang: my pleasure [21:26] ralsina: not building new nightlies won't break them, but i don't want to waste time backporting webkit either [21:27] and then having webkit break everything else [21:27] dobey: yeah, that would be kinda useless. [21:27] Lucid is lucid, you get what you get. [21:27] * karni switches to windows for a short while [21:30] ok, that should make life somewhat better [21:30] dobey: got any thoughts on bug #711317 ? [21:30] Launchpad bug 711317 in rhythmbox-ubuntuone-music-store "Banshee - Link to purchased music folder sometimes duplicates songs to ~/Music (affects: 1) (heat: 8)" [Undecided,Invalid] https://launchpad.net/bugs/711317 [21:30] Basically it bothers me that if you sync the music store UDF, you don't get the songs on banshee's library [21:33] ralsina: it's a known bug. well, you do get them if you buy them in banshee, but yes, it is an annoying bug :) [21:34] can we tweak banshee's config to add the udf to their library? [21:34] So, if you buy one song in banshee the whole lot of songs appear? [21:34] no [21:34] banshee doesn't work that way [21:35] it's a known issue, and it is on my radar [21:35] ok then [21:35] but you know, i have like 544353454 things to do :( [21:35] is it in your radar as "ohmygodthatplaneiscomingrightthisway"? ;-) [21:35] Yes, I know [21:37] mp3 codec install is a bigger change, so i'm trying to do it first [21:37] but the 3 main things left for banshee u1ms are mp3 codec install, udf subscription, and the non-appearing songs issue [22:07] * karni is back [22:22] alright, i'm off. later all [22:23] bye dobey [22:56] CardinalFang: perhaps you know, I don't think verterok is around. createUdf in u1-java-sp takes 2 strins, name and path. I provided {name: 'custom-udf', path: 'custom-udf-path'} and it created such UDF on my PC: ~/custom-udf/custom-udf-path . We could use this to create ~/UbuntuOneFiles/udfs_here but if you want to crate ~/a_custom_udf with the same path, I can't tell how you can do that with verterok's implementation at the ... [22:56] ... moment. [23:18] CardinalFang: oops. at the moment UDFs (i.e. synced folders) will land under /sdcard/u1/ anyway. We don't support custom UDF paths, but that's certaly place for future improvements. One thing at a time :)