=== elehack is now known as elehack|afk === vds_ is now known as vds [12:04] morning all [12:05] duanedesign__, morning! [12:08] gday roman [12:09] gotta spend another day on my Windows machine, yuk [12:10] need to get my ssh key on this machine so i can log into my server and use my irc client === JanC_ is now known as JanC [13:06] What should I do? I bought a song on ubuntuone music store but the file placed in my account is only the sample since its only 38seconds long, the song is really 4mins long. I have tried to contact 7digital with no luck. [13:57] aquarius, vds, beuno: "playlist" records in desktopcouch use ids specific to UbuntuOne. I call this a bug. It breaks the d-c notion of being independent of service domain and probably requires I be online to make new playlists. The ID should be something independent of Ubuntu One, I say. [13:59] CardinalFang, what worries me about that is that there is no canonical unchanging ID that we can rely on. [13:59] CardinalFang, if you want to create a playlist which works in Ubuntu One music streaming, you'll need to read the nodekey *anyway* [14:00] aquarius, that's what application_annotations should be for. [14:00] and the case of people who don't use U1 at all -- imagine importing the playlist into a new music player. How will it identify the songs that it knows about that are in that playlist? Requiring a "support DC playlists" extension to SHA224 every file in the music library just to read playlists is a bit intense :( [14:02] obviously the long-term solution here is musicbrainz ids :( [14:02] O! if we only had those already. [14:02] 'zactly. [14:02] . [14:03] ? [14:03] so [14:03] this sounds like a hard problem [14:03] ralsina: ping [14:03] It is a Hard problem, yes. [14:03] mandel: pong [14:03] ralsina: so we have a standup 2day? [14:04] mandel: yes, on 2 hours and 12 minutes from now [14:04] mandel: so everyone is awake at the same time [14:04] ralsina: oh, it has been chaged, will this be for the rest of the week? [14:04] mandel: yes, I will update the calendar [14:04] mandel: it doesn't make much sense to have standups without 4 of us [14:05] ralsina: but we are the 4 important ones, the rest are just noise ;) [14:05] ralsina: by the way you own me an approval on canonicaladmin :P [14:05] mandel: but we are polite and don't say it out loud :-) [14:05] mandel: canonicaladmin owes me a login [14:05] mandel: and it seems now noone else can approve you, not even chipaca [14:05] ralsina: hahaha ok, no worries I was just wondering if you forgot [14:06] I got word from HR yesterday that they were looking into it [14:06] ralsina: those lotus notes apps are shit, we should write something in django [14:06] shit == polite way of manuel saying something is.. shit :) [14:07] mandel: I have heard that a million times. And then someone tries and it ends being worse shit. Lotus notes is the best in the business, imagine what the rest is like :-) [14:08] CardinalFang, given that we don't have musicbrainz IDs, and given that music players need to be able to read the playlist, I can only think of three types of ID: 1. something the music player already knows without having to calculate it (e.g., path), 2. something the music player can calculate (e.g., SHA of contents), 3. something the music player can ask another service for (e.g., u1 nodekey) [14:09] CardinalFang, I can't think of a type 1 ID that's actually usable; I don't think that type 2 IDs are doable because you have to calculate them for your entire library. [14:10] In an ideal world, a musicbrainz ID would be a type 1 ID (i.e., music players would calculate them already for all your files), but they're actually a type 3 ID (you have to ask an external service for them), and that external service is on the internet (so fetching it for the whole library is prohibitively expensive) [14:10] ugh, rhythmbox/plugins/umusicstore/MusicStoreWidget.py gives a bazillion pylint warnings on Lucid, how did that even get comitted? [14:11] including the import error that makes it break completely [14:14] mandel, vds, CardinalFang: I can't edit the standup appointment, so I'll tell you here: for this week, we do the standup 2:15 hours later so the dallas people can attend [14:14] ralsina: because pylint was not being used 'til a few months ago [14:14] ralsina: ack [14:15] mandel: well,that would explain it. I think we should check most of our codebases then :-( [14:15] ralsina: the evil part of the new u1trial etc is that you will have to fix them, and little by little the code will be nice :) [14:15] ralsina: we had to go through that with desktopcouch [14:15] mandel: that's not evil, that's having a job :-) [14:16] Of course having a job is pretty evil, but still ;-) [14:16] lol [14:16] kenvandine_, I proposed desktopcouch and python-couchdb updates to natty yesterday. This is a noncompatible update for python-couchdb (at least at the level desktopcouch deals with it). If one updates python-couchdb but not desktopcouch, the old desktopcouch breaks. Should the new python-couchdb include a "Conflicts" line going backward? [14:18] goodday everyone [14:31] CardinalFang, aquarius, so, what do we do about this? rename the db? leave as-is? [14:33] might be worth renaming song_ids to ubuntu_one_song_ids for now (and having an explicit record_type_version, which we should have anyway) so that if MB IDs become available later, we can go to a new version of the format with minimal upgrade issues? [14:33] beuno, I don't know. it may be impossible to do better than arbitrary. [14:33] Right, I think that's pretty good, aquarius. [14:33] aquarius, I like renaming [14:33] CardinalFang: using Breaks field may be more appropriate [14:34] Hrm. [14:36] CardinalFang: conflicts is more appropriate for cases when two packages provide same file (or rather a file in the same location, with the same name) [14:37] kklimonda, that's a good point. [14:37] kklimonda, "Breaks" is still new to me. Only used it once. [14:38] CardinalFang: there is a great bullet point list with cases for Breaks and Conflicts in the debian policy, give me a sec and I'll find it for you [14:39] CardinalFang: http://www.debian.org/doc/debian-policy/ch-relationships.html#s-breaks [14:40] and below, in the section about Conflicts there are few pointers when to use them [14:40] also, those two paragraphs are worth reading as breaks and conflicts are similar [14:42] kklimonda, right, thanks. [14:43] CardinalFang: also, some basic testing is a good idea - once, I've made a mess because I added Breaks on a wrong version ;) [14:53] beuno, I was looking to see how to update the server. In tests, playlist = Playlist(name='p2', song_id_list='') [14:54] beuno, that string doesn't look right. [14:56] beuno, I want to make sure there's not something clever going on there. [15:02] beuno, once that change is made, the playlist schema on freedesktop.org needs updating, of course [15:07] aquarius, add another version? [15:08] CardinalFang, well, v1 is the one with ubuntu_one_song_ids; a potential v2 will have some better ID, but we don't know what yet so we can't document what v2 will contain [15:10] CardinalFang, it doesn't, yes [15:40] good morning, #ubuntuone [15:40] meh [15:41] hi alecu, dobey [15:41] dobey: a quick question === kenvandine_ is now known as kenvandine [15:42] dobey: the music store is broken for lucid in nightlies. This is known and we don't care much, right? [15:42] i didn't know, no [15:42] but i'll look [15:43] ok, it's an import that's broken (inline_callbacks) but I don't know if it's just that or if it's a deeper change on APIs that' s hurting it [15:45] dobey: in fact pylint goes ccompletely berserk on the musicstore rhythmbox plugin on Lucid :-) [15:45] ok [15:47] i wonder why apt-get build-dep doesn't seem to be installing what i expect it to [15:51] kenvandine, I pushed an update to keep newer python-couchdb from breaking older desktopcouch. I'm happy with those packages now. Please take a look when you can. [15:51] https://code.edge.launchpad.net/~cmiller/ubuntu/natty/python-couchdb/201101_0.8-0u1 [15:51] https://code.edge.launchpad.net/~cmiller/ubuntu/natty/desktopcouch/1.0.5-0u1 [15:52] CardinalFang, could we please avoid making tarball releases and proposing uploads before the new version bump is committed to trunk? skipping version numbers arbitrarily is weird [15:52] CardinalFang, mind proposing merges for those so i don't forget? [15:53] kenvandine, done. [15:53] thx [15:54] CardinalFang, sprinting this week and DX is making my life hell :) [15:54] but i'll find time to sponsor those later today [15:54] kenvandine, yeah, sorry for my timing. [15:54] no worries, work must go on :) [15:59] dbus 2011-01-11 10:57:56,487 - Failed to load plug-in: desktopcouch.application.plugins.ubuntuone_pairing [16:05] ralsina, ping [16:05] ralsina, when is the standup? [16:05] alecu: in 10 minutes [16:05] cool, thanks. [16:06] 2011-01-16 [16:06] let's make that official ;-) [16:06] aleco dobey mandel thisfred CardinalFang nessita vds standup in 9 minutes [16:06] ok [16:07] i hope my machine is usable again by then [16:11] who is aleco? [16:11] I have a doppelganger! [16:11] alecu: the male you (?) [16:12] strawberry alecu [16:14] me [16:14] nessita, too early! [16:14] CardinalFang, kenvandine: ^^ did you see the desktopcouch error i pasted above? [16:15] me [16:15] me [16:15] me [16:15] you [16:15] so happy together [16:15] me [16:16] me [16:16] me [16:16] * ralsina has fat fingers and is using the small notebook [16:17] thus aleco [16:17] nessita, start [16:17] DONE: on going platform rally. Had meeting with DX team to talk about U1 unity integration. Had improvised meeting with the U1 team about U1 event aggregator. Got u1cp 0.5.1 released. [16:17] TODO: stop having long term running processes such as ussoc and u1cp-backend. Try to meet with Ivanka. [16:17] BLOCKED: recurrent kernel panics when booting. Latest response from kernel team was 'live with the crashes' :-/ [16:17] NEXT: mandel [16:17] DONE: Worked on bug 700951 bug 700981 701049 and proposes a branch to fix those. Finished in another branch the PortAdvertiser for desktopcouch on windows. We need to talk about couchdb on windows, the batch has to be improved as well as the installer they use, it is missing info like installation path, version etc.. which makes it a pain to work with. [16:17] TODO: Propose fixed to PortAdvertiser on windows. Run tests of desktopcouch yet all the instance will point to the default port (reason is the batch of couchdb being crap.) [16:17] Launchpad bug 700951 in desktopcouch "Couchdb process is not started on windows (affects: 1) (heat: 6)" [Medium,In progress] https://launchpad.net/bugs/700951 [16:17] Launchpad bug 700981 in desktopcouch "The function 'run_couchdb' should be moved to the platform module (affects: 1) (heat: 6)" [Medium,In progress] https://launchpad.net/bugs/700981 [16:18] BLOCKED: no, but will need some help to understand how to pass parameters to couchdb [16:18] * mandel looks at the real aleco, alecu is his evil twin, and I know because he has a goatee http://en.wikipedia.org/wiki/Evil_twin#The_goatee [16:18] DONE: meeting with njpatel to talk about unity integration. Discussed with dobey, nessita and thisfred *were* to put the event aggregation code itself. Team dinner with the griff. [16:18] TODO: write a plan for all this with thisfred, run it thru ubunet-discuss; flesh out a list of bugs for it with thisfred. Find out if thisfred is perhaps that aleco guy [16:18] BLOCKED: no [16:18] now it's thisfred turn to mirror my standup notes: [16:18] DONE: same as alecu TODO: same as alecu BLOCKED: same as alecu (basically I follow alecu around all day and say: "me too!") NEXT: the dobester [16:18] DONE: meet with njpatel, bugged pitti, convinced alecu/thisfred/nessita not to do unity integration the complete wrong way, landed branches TODO: land more branches, fix backport tests, more rally BLCK: rally [16:18] next ralsina [16:19] DONE: searching for windows contractor, finally got someone on HR (no results), team leads call, chased people around as usual. [16:19] TODO: reviews, management stuff, have HR fix things for me, someday actual coding ;-) [16:19] BLOCKED: no [16:19] and now the cardinal [16:19] DONE: some work on playlist API, mostly complaining about it and looking at similar code in server. desktopcouch release and package. [16:19] TODO: more playlist public API. [16:19] BLOCKED: None [16:20] CardinalFang: is that release the natty one? [16:20] vds, et tu? [16:20] DONE: continuing on developers RESP APIs, discussed with teknico and aquarius, filed bugs #701029 and #701031 [16:20] reviews: [16:20] lp:~mandel/desktopcouch/add_windows_keyring_import [16:20] lp:~beuno/ubuntuone-servers/parse-m4a [16:20] TODO: continue with the APIs [16:20] vds: Bug 701029 on http://launchpad.net/bugs/701029 is private [16:20] BLOCKED: nope [16:20] nessita, yes, Narwhal. [16:20] nessy narwhal [16:20] it's broken though :) [16:20] CardinalFang: did you see dobey's bug report for the dc release? [16:21] Not yet. I saw the error message pasted above, plugin not loaded? [16:21] * ralsina is starting to really really hate that release [16:21] right sry, that;s what I meant [16:21] i haven't filed a bug yet [16:21] running from trunk the plug-in fails to load [16:22] dobey, what did you do to run it? [16:22] CardinalFang, PYTHONPATH=. bin/desktopcouch-service [16:24] "INFO:root:Loaded Ubuntu One extension for desktopcouch." :( [16:25] Maybe it's an installed-versus-development-tree problem. [16:25] huh [16:26] nessita: stuart asked me to tell you to tell ivanka that he wants to talk to her ;-) [16:26] nessita, also, can you tell Ivanka I will be sending her an email in about 2 hours, to please ack it? :) [16:26] On other news, standup will be at this hour for the rest of the week [16:26] ralsina, re: rbox store on lucid, where are you seeing the failure? the nightlies look like they built fine [16:27] dobey [16:27] dobey: they build fine, but the plugin doesn't load [16:27] ralsina: I tell you to tell aquarius to take a number and do the line :-D [16:27] oh [16:27] * ralsina is just a messenger ;-) [16:27] ralsina, is there a bug filed? [16:27] dobey: line 47 fails. vds was going to file a bug [16:27] beuno: I'll try, though I will try to caught her attention to myself first [16:27] :-D [16:28] * ralsina suggests boleadoras [16:28] she seems just landed (I saw her during breakfast) and she didn't have a many friends face. But maybe is just me. [16:28] dobey ralsina not filed yet, but soon [16:29] Yet other news, I will be starting late and leaving late this week so I can have some time with the Dallas people too [16:29] yey! [16:29] nessita, we'll all add her on facebook [16:29] facebook is for sissies (tm) [16:30] if it's not filed, it's not a bug [16:30] nessita, strawberry girls? [16:30] :-) === beuno is now known as beuno-lunch [16:33] ok, any more comments, or eom? [16:34] eom here [16:35] eom! [16:35] Ohh, forgot! [16:35] shoot! [16:35] Any news on contacting jamesh? [16:35] about bindwood, that is. [16:35] ralsina: who you're asking to? [16:36] ralsina, no news [16:36] alecu, thisfred [16:36] ralsina: I sent another mail yesterday, no reply still [16:36] ralsina: Maybe I'm ending up in his spam, or maybe he's on holiday? [16:36] you are in the worst possible timezone to meet him :-( [16:36] ralsina: or maybe he *really* doesn't want to do bindwood :) [16:36] he's not supposed to be on holiday. [16:37] thisfred: he [16:37] Any of you dares just call him? ;-) [16:38] His phone is available, just try to be awake when he is and bug him. Of course I assume calling .au is expensive [16:40] forget about it, you mailed him, wait until tomorrow morning for a response, or I'll try myself [16:41] ok, thanks, this week is going to be a little hectic, I can call him on monday no problem [16:41] thisfred: don't worry about it [16:41] Perhaps he's just too busy fixing bindwood [16:41] thisfred: yeah, right. I assume he's uploading right now ;-) [16:42] thisfred: in any case, if he doesn't reply before tomorrow's standup, I will handle it. [16:42] somehow [16:42] speaking of bindwood [16:42] i wish the people who have proposed branches would deal with them in some manner (fix or set to rejected or whatever) [16:43] dobey: let me rephrase that for you ;-) [16:44] dobey: guy#1 could you please decide what to do eith branch whatever? Maybe you should set it as rejected or something. Thanks in advance! ;-) === rodrigo__ is now known as rodrigo_ [16:47] ralsina: dobey, excellent point, will reject my 2 branches [16:48] dobey: do you have a way to make tarmac work again remotely or we need to wait until you get back home? [16:48] ralsina, i am running it on my laptop [16:48] dobey: ok [16:48] dobey: is it very CPU or memory intensive? If it's not, maybe I can loan a hosted server to run it [16:49] when it actually has stuff to try and land it is, yes [16:49] dobey: then maybe we should eventually ask for canonical infrastructure for that [16:50] dobey: you know, avoiding bus_factor=1 and all that [16:50] yes, we need to set up our own private uec space for ubuntuone [16:51] but that doesn't really help with the bus factor issue [16:51] dobey: ok, next week we'll see who we should talk to [16:51] dobey: you can always train me to do it [16:52] * ralsina can do simple tricks and fetch sticks on demand, too. [16:52] well, there isn't really anything that should require human interaction, except for adding handling of new branches/projects [16:53] dobey: cool, so with a bit of docs and access credentials to infrastructure it should be doable. We'll talk about it ... next uesday?t [16:53] next tuesday that is [16:54] dobey: is that the same laptop you take to uds? [16:54] mandel, no, using my older laptop because my newer one has broken video drivers still [16:54] sure [16:54] monday is a us holiday [16:54] cool. Now I really have nothing else [16:55] dobey: ok, so is not the tiny little thing, right? 'cause I was expecting you to say yes :) [16:55] mandel, no, it's the big 10 inch [16:55] haha [16:56] * ralsina is now using a 12" and it feels like IMAX [16:57] MLK day [16:57] dobey: cool, scheduled it [16:58] i wish i could find a nice 9" LCD [16:59] or AM-OLED even [16:59] * ralsina loves his new sphygmomanometer [17:00] or a heliodisplay [17:00] or neural interface [17:00] fonts are so passe [17:02] dobey: noone is building 9" notebooks anymore, right? [17:02] pretty much yeah [17:02] all the 7" are 800x480 [17:02] or 8" [17:03] and 10" are all 1024x600 [17:03] I always wanted one. But when I ditched the 7" I wanted better specs, that's how I ended with a 12" [17:03] sure [17:03] that's how i ended with a 5.6" [17:03] 260dpi ftw. [17:03] it's still 1.6 pounds so at least it's light ;-) [17:03] dobey: what is it? [17:04] i don't think mine weighs that much [17:04] fujitsu u820 [17:04] 1280x800 @ 5.6" [17:04] fujitsu always had cute tiny ones [17:04] And it costs more than mine, too ;-) [17:05] tablet touchscreen, gps, 3g, wifi, 260 dpi, and about 1lb :) [17:05] well it was about $1000 [17:05] but worth it [17:05] You can get it for 888 or so now [17:05] until intel screwed the pooch anyway [17:05] yeah, because they don't make them any more [17:06] so the refurbs are cheaper, sure [17:06] no, that's new. Someone has some stock left, it seems [17:06] dynamism? [17:06] Newworldit [17:07] well i got some extras for free too, so eh [17:07] if i could find a nice 9" display though, i would just make my own laptop [17:07] but getting a 1080p display at 9" isn't easy [17:07] well, you can always buy a mini9 and gut it [17:07] oh, 1080p [17:08] eh, i would make a custom case out of aluminum [17:09] dobey: you could start with this: http://www.hometheaterforum.com/products/panasonic-dmp-bd15-9-inch-portable-1080p-blu-ray-player/reviews [17:16] ralsina, that's not actually 1080p (blu ray movies are though) [17:16] looks like that display is 1024x576 [17:17] dobey: too bad [17:18] yep [17:18] even more reason not to pay $800 for one though [17:20] and there goes the internets [17:21] alecu: got your doc, but google is giving me internal error so I have not looked at it yet [17:22] ralsina, don't worry: we are still working on it. [17:22] ralsina, we'll run it thru the list when it's done. [17:22] alecu: ok [17:22] alecu: even better [17:33] hmm === ralsina is now known as ralsina_lunch === beuno-lunch is now known as beuno === ralsina_lunch is now known as ralsina [18:11] * mandel -> rugby [18:19] thisfred, alecu: https://wiki.ubuntu.com/NotificationDesignGuidelines [18:20] nessita, nice, thanks. [18:25] sigh [18:33] * ralsina wonders if he should spend his sunday evening making aranduka work with couchdb [18:34] And then start talking about a mysterious ubuntu book store or something [18:38] ralsina: I like it! [18:39] ralsina: if you have any (desktop)couch questions, let me know [18:39] thisfred: it has over 3 million books, too ;-) [18:41] I might buy an ereader after all [19:51] hi, i'm having problems getting ubuntu one to actually sync any files. it just created the directories, but won't populate them with the files. it's been at it for about 2-3 hours now, so i think it's safe to say it's just not slowness in the system [19:52] anyone have any clues? [19:52] * beuno nudges rye [19:53] rather fresh install of maverick btw [19:53] cryogenfx, if you have put a fairly large amount of files then that's expected (and this is being fixed, actually, so this is not a permanent "feature" of Ubuntu One). What does running u1sdtool --waiting-metadata do ? [19:53] cryogenfx, btw, https://wiki.ubuntu.com/UbuntuOne/FAQ/WhyAreOnlyFoldersUploadedButNotFiles [19:57] the commans spews out alot of lines, not sure what they are, but i recognize alot of filenames in there. but 2-3 hours, and not a single little txt file is transfered? [19:57] how long does it usually take then? [19:57] sorta going on a business trip tomorrow, and am wondering whether the files will be synced before i go or not [19:58] 2719 lines of text actually [19:58] cryogenfx, if you're on Maverick, you can install "magicicada" [19:58] probably one for each file then [19:59] which will give you more visibility into what's happenning [19:59] oH [20:03] aha, the linecount is decreasing as time goes on, now it's down to 2609 [20:03] so, the metadata creation is what's taking so long [20:03] wow [20:03] right, that's exactly what will be fixed in Natty :) [20:03] some server improvements will apply for all clients [20:03] but Natty will get the bulk of the improvement [20:03] ok, cool. thanks, at least now i know that it's doing something :) [20:04] looking forward to natty then [20:05] but it only needs to create this metadata once per file, right? so next time, will be quicker with new files on top of these? [20:05] yeap [20:05] phew :P [20:07] ok, i salute you, thanks alot. now to go to #ubuntu and find out why i must enter my keyring password 3-5 times on login :) [20:11] cryogenfx, i suppose you will need to check whether you are using gdm and that its keyring password is the same as your login one [20:38] beuno: __lucio__: is U1 slow ATM or is it just me? [20:43] beuno: U1 is definitely slow. if you want to track that down, it's the perfect moment. [20:44] beuno: I lost some time on debugging transactions and it turns out it was just terribly slow server :/ [20:45] <__lucio__> karni, let me check [20:46] __lucio__: thank you [20:47] dobey, https://wiki.ubuntu.com/NotificationDesignGuidelines [20:48] <__lucio__> karni, we have a couple of slow slaves. new connections will suffer until it comes down. old connections are ok. [20:49] __lucio__: That's really bad for me :< You can estimate how long will this take? [20:55] so, in an effort to speed up the process, so i can actually have some uploaded files before i leave tomorrow, i moved away one of the folders, that contains thousands of files, and deleted the folder from the webinterface, but still it grinds through the metadata of the thousands of files. is there any way to "flush" them, or refresh the process or sometimes? [20:59] <__lucio__> cryogenfx, the way id do it is to stop syncdaemon, install nightlies so you have the speed updates on metadata, and restart syncdaemon so it just gets the diff from local rescan. but i dont do support, and maybe there is a reason for that :) [20:59] beuno: __lucio__ found me a little less loaded server [21:00] ok, i'll try that, thanks [21:09] ok, looks like I have to babysit for a while. Will be here intermittently for a while. === ralsina is now known as ralsina_babysitt [22:14] nessita, thisfred, dobey: I'm going over the list of events, and I think we pretty much every other event we should display could be shown as notifications. [22:15] the only type I think we should discuss is synchronization conflicts [22:15] yes [22:15] I believe those should not be shown just as notifications. Perhaps in the messaging menu, or something. [22:15] conflicts? [22:15] probably more urgent than that, and not really a message [22:16] maybe the urgent hint to bounce the icon is enough [22:16] dobey, right. [22:16] alecu: since we're not offering an action to take over the files, not sure if we should provide anything else other than a nofitication... [22:16] alecu: so, you would bounce the icon... and then? [22:16] nessita, but notifications are transient, and this should not be transient. [22:17] nessita, perhaps show a list of events that failed. [22:17] isn't cp supposed to provide some way to deal with conflicts? [22:17] alecu: where? [22:17] I'm not sure if bouncing the icon is the right way to go. [22:17] dobey: not this cycle [22:17] a list of *files and folders* that failed. [22:17] alecu: right, but where would you show that? [22:17] cp [22:18] the launcher icon can not show a list of messages [22:18] yes it can [22:18] it can [22:18] but you don't want that [22:18] although requiring right-click on the icon is wrong [22:18] nessita, I believe the right way would be opening gnome-activity-journal, filtered for u1 conflict events (if installed) [22:18] neil said it doesn't support that... [22:18] alecu: that makes sense [22:18] alecu: excellent solution [22:18] i don't think opening activity journal makes sense [22:19] you can't do anything about it there [22:19] I don't know if that's at all possible, though :-) [22:19] we should find kamstrup in the rally, from the zeitgeist team, and have a little talk with him... [22:28] CardinalFang: What information could I usefully provide you to debug why my android phone can't log in to Ubuntu One (either contacts or music)? When I launch contacts it asks me to login, fires up the browser which does the OpenID stuff, then takes me to the “You've got an android phone” page and there's no obvious way to get anywhere from there. [22:47] alecu, thisfred: can you please review https://code.launchpad.net/~nataliabidart/ubuntu-sso-client/shutdown/+merge/45926 and https://code.launchpad.net/~nataliabidart/ubuntuone-client/credentials-shutdown/+merge/45927 ? [23:01] sure thang [23:03] thanks! [23:24] verterok: I know it's pretty late probably. You around? [23:26] __lucio__: you know your way around the protocol right? [23:26] __lucio__: ok, will catch you next time :) [23:29] <__lucio__> a bit :) [23:30] <__lucio__> karni, do you have any questions? [23:31] __lucio__: oh, cool. yes, I'm wondering how is it solved in ubuntu one. if a file changes, and the software uses delete-and-write (not 'modify'), you get few generation bumps [23:31] __lucio__: how is that handled by the syncdaemon? oh.. it's nothing in common with the protocol really. [23:31] you know syncdaemon internals maybe? I've got the sources, but I'm just wondering [23:31] if you could pull the answer out of the hat :) [23:31] <__lucio__> karni, what do you mean delete-and-write? [23:32] if you rename a folder, other clients will receive one notification [23:32] <__lucio__> i do know how that stuff works, but i need to understand the question first :) [23:32] if you use gedit to edit a text file, you'll [23:32] <__lucio__> yeah, one notification saying: there is new stuff for you [23:32] create a tmp file, origninal will be removed, and the temp renamed to the original's file name [23:33] so other clients will receive "foo isLive=false" and next "foo isLive="true" [23:33] so basically, 2-4 notifications instead of one. interestingly, one of them can say that there's a new content [23:33] <__lucio__> karni, are you using generations? [23:33] i think that's handled by ignoring the tmp files [23:33] while I got a protocol error that "node has no content" [23:34] __lucio__: yes, I'm using generations [23:34] dobey: beuno said U1 syncs temporary (i.e. hidden) files too [23:34] <__lucio__> first, a node is created, it is created with content=None (makefile), then content is set (putcontent) [23:35] aha. so this is probably 2 bumps of generation [23:35] is it? [23:35] <__lucio__> we sync all files we dont ignore, we have a pattern that we use to define what to sync [23:35] oh [23:35] now that's something useful [23:35] <__lucio__> well, there is also move and maybe delete, depends on the case of course [23:35] __lucio__: right [23:35] <__lucio__> but you should not worry about that really [23:35] karni, syncdaemon explicitly ignores some [23:36] <__lucio__> you get N notifications or one, you only care about interpreting the delta [23:36] the question is.. if I'm getting generation callback such as.. say, 5, 6, 7, 8 -- I'm only interested in the last one [23:36] <__lucio__> yes [23:36] should I use queuing with a timeout or there's a smarter solution [23:36]