[12:16] aquarius, did you see my message yesterday about d-c pairing and couch? [12:16] aquarius, you used "put", not "put_record". [12:17] I haven't run the code, but I think that's not right. [12:34] CardinalFang, I agree, it's not right; that code fails before it gets that far because of the 401 problem [12:35] CardinalFang, how did you guys get on after I left? [12:35] aquarius, well, I'm up early to get code out the door, if that tells you anything. [12:36] aquarius, nothing new to report. [12:36] aquarius, except statik merged some branches. [12:36] kk [12:37] want me to work on some of the stuff with you? [12:37] is there anything you're doing that I can pick up? [12:46] CardinalFang, or have you basically done it all? :) [12:48] aquarius, hmm. I've done nothing with actually getting oauth tokens, so "pairing storing oauth tokens". Also, I plan to move some that code of yours yesterday into the desktopcouch.pair.couchdb_pairing.couchdb_io so it can be shared with the local stuff where possible. [12:49] CardinalFang, hang on, I didn't quite follow that [12:50] aquarius, ignore the second sentence for now. [12:51] so...the things that you were doing, as I understand it, are: have the pairing app exchange oauth tokens as part of the on-the-wire protocol, and the daemon which starts and maintains continuous replication. If I read you right, you're working on the second (the daemon) but not the first (the pairing app)? [12:53] aquarius, we need to get/generate oauth tokens for every pairing, as I understand it (which is really not at all!). If you can make the function that gets the token from U1 (?), I can make the pairing tool exchange them. Yes, I'm doing 2. If you will do 1a, I'll also do 1b. [12:55] ok. I've already done the code which initially generates the oauth token and stores it in the ini file; I'll write a function which can extract that token from the ini file so you have access to it [13:01] Grr. 0800. Must take the kid to the babysitter. Back in a while. [13:01] morning [13:02] aquarius, a nice fat comment saying what the pieces mean would be gratifying too. :( I really must read up on oauth one day. [13:02] kk (which pieces specifically?) [13:03] Er, if I knew that, I might not need any comment. There are three, or two opaque numbers, yes? [13:03] right, so you want to know what the consumer key and token and token secret and etc mean? gotcha [13:04] Yes. Morning, all! [13:05] best dog toy ever: ice cubes [13:06] hey thisfred [13:07] any word from jchris? [13:07] aquarius: none: he wasn't online as far as I could tell yesterday, and on the mailing list nothing specific to our cause yet [13:07] still going through [13:18] kk [14:24] aquarius: so, no mail from chris, and he's not on the channel now. jan____ is away, but we need to verify that this issue is treated with some sense of urgency, since you're blocked by it (correct?) [14:25] correct [14:27] ok, I don't know who on the #couchdb channel we can speak to about this, jason also isn't there. SteveA should we mail/call couch.io? [14:27] rehi CardinalFang [14:28] re CardinalFang. function to get oauth tokens now in a branch for approval [14:29] Rgr. [14:30] * thisfred reviews as well [14:43] aquarius, "oauth_token_secretsx" ? [14:44] pants [14:45] aquarius: your oldest branch, I'd like to review it, or is it obsolete by now? [14:45] CardinalFang, fixed [14:45] thisfred, which one? [14:45] aquarius: if not, could you fix the conflict? [14:45] aquarius: lp:~sil/desktopcouch/create-oauth-tokens-startup [14:46] kk, willfix the conflict [14:46] awesome [14:51] aquarius: also a conflict on your other branch(es) probably the same [14:51] yeah, same one, I suspect [14:51] I hate resolving conflicts [14:52] the quickest way to resolve conflicts is not to generate them [14:52] well, er, yeah :) [14:52] * thisfred runs away before aquarius rips off his arm and beats him over the head with it [14:52] but that ship has sailed, I fear :) [14:53] * CardinalFang high-fives thisfred. [14:53] aha, you two are in cahoots against me, are you? I see how it is [14:54] aquarius: you're gonna hate me even more: the changes that are conflicting I have mostly undone on my latest branch... [14:54] perhaps we should land that first [14:54] land that first, I suggest [14:55] since I am now confused by which bits need to happen where [14:55] the conflicts will still exist but they'll be smaller [14:56] I did a lot of hoop jumping to make the tests work, and then when the port was readable from the log, I had to unjump some [14:56] or was able to, I should say [14:59] me [15:00] aquarius, outside that type, that branch looks good to me. [15:02] MEETING BEGINS [15:02] If you're here for the stand-up meeting, say "me". Then "DONE", "TODO", and "BLOCKED". [15:02] me [15:03] me [15:03] me [15:03] me [15:03] me [15:05] urbanape, rodrigo_? [15:05] DONE: Figured out that erlang packages can't have tests enabled because tests aren't included. Some reviews/merges. [15:05] TODO: desktopcouch 0.3 release for feature freeze today. Nag urbanape and rodrigo about compat testing with couchdb-0.10 [15:05] BLCK: nope [15:05] dobey, show us the way! [15:05] me [15:05] ☭ DONE: Fixed overactive Fatal Error, Started prefs dialog [15:05] ☭ TODO: Finish prefs dialog, OAuth [15:05] ☭ BLCK: None. [15:05] teknico: bongiorno! [15:05] DONE: completed and landed a branch that improves the handling of login failures from our Funambol Server API code to the Funambol DS server, reviewed one more markgsaye's branch, fought with rabbitmq [15:05] TODO: working on the contacts CRUD web ui [15:05] BLOCKED: rabbitmq-server not starting anymore, but being solved [15:05] next: aquarius [15:05] ⚀ DONE: various smallish desktopcouch branches [15:05] ⚁ TODO: piston oauth in snowy; get DC0.3 done [15:05] ⚂ BLOCKED: couchdb patch which lets OAuth read users from the ini file doesn't seem to work [15:05] In all the world there is only one rodrigo_ [15:05] • DONE: Submitted tomboy package with U1 as default sync server. Submitted branch to fix notes UI in production, and kept looking at it, because it's not clear it fixed it [15:05] • TODO: Add more tests in couchdb-glib test suite. More openSUSE packaging. Add social services accounts config to about-me. Talk to Ara about writing mago tests for evo-couchdb. Propose couchdb-glib/evo-couchdb for GNOME 2.29. Store UUIDs for postal addresses. Conflict resolver tool in pair tool. oAuth authentication and signing of all couchdb-glib requests. Finish adding URLs to contact records in evo-couchdb. Add changes notificatio [15:05] • BLOCKED: none [15:05] DONE: Reviews yesterday. (But I'm on-call-reviewer *today*. Dang.) Replicator nearly finished. Needs testing. [15:05] TODO: Finish replicator. Make pairing-tool exchange OAuth noise. Also, grudgingly review today's branches and mark them all "Disapprove". [15:05] BLOCKED: None [15:05] nobody next, right? [15:06] ah, missed CardinalFang :) [15:06] teknico, How can we help? [15:06] he missed himself i think [15:06] CardinalFang, statik is valiantly taking care of the problem, thanks :-) [15:07] Good man, statik. [15:07] Okay, MEETING ENDS [15:07] dobey, e bonanotte ai sonatori ;-) [15:08] * CardinalFang is glad no one read his TODO. [15:08] bonanotte? little early for that, eh? :P [15:08] you sleep with the fishes [15:09] let's take it to the mattresses [15:14] dangit [15:14] DONE: Proposed an apparently broken branch for multi-downloads. [15:14] TODO: Test bindwood against new CouchDB package. [15:14] BLOCK: None [15:14] urbanape, we sang arias of sadness at your absence from our meeting. [15:16] aquarius, leave the oauth, take the canonical. So, oauth info is client-relative, not server-relative? I get one set of credentials and use that single group of four values to talk to all paired hosts? [15:16] no [15:16] it's server-relative [15:17] each server has its own set of keys [15:17] think of the keys as being a multi-item password, basically [15:18] so when you pair machine A with machine B, machine A needs to get its own keys (by calling the function), and then pass those keys to B (down the connection established by pairing). [15:18] B then stores those keys in a paired-server document on B, with server: A [15:19] aquarius, Right. So get_oauth_tokens() is server end, for transmission at pair-time? I store those somewhere and then retrieve them for a particular host at replication time. [15:20] yes. Machine A calls get_oauth_tokens when another machine wants to pair with it, so it knows its own tokens and can give them to the other server [15:20] crap, i think i may have just lost DNS [15:20] sigh [15:20] my network is being really annoying [15:20] CardinalFang, your TODO is why I offered to take reviews for you ;) [15:20] the other server stores those tokens in a paired serve record, and then when the other server starts up replication, it reads the paired server record for A to know which tokens it needs to pass to A as part of the replication process [15:21] statik, :) [15:25] aquarius, thisfred: any news from couch.io on the oauth patch? If I merge lp:~sil/desktopcouch/not-so-random into desktopcouch and include it in a release today, will that break evolution-couchdb, bindwood, and quickly? [15:26] statik: I see no one online that I know is couch.io to ask. Re: the breaking, I defer to aquarius' wisdom [15:27] erm, let me check [15:27] yes. It will break everything. [15:28] Because it specifies that a valid user must be used for all access, and it is currently not possible to access couch as a valid oauth user if that user is specified in the ini file :( [15:30] I think we need a flag-day when we change both at once. We may need code that shuts down running couchdb servers in the preinst. [15:33] aquarius, so, just to make sure, do we need python-couch package to take OAuth info as a key for instantiating client.Server objects, or does couchdb daemon read remote oauth stuff on-demand from its INI file? [15:34] ^key^parameter [15:34] CardinalFang, desktopcouch.records understands how to look up oauth tokens from the keyring to access the local desktopcouch, so you don't need to do anything [15:34] if you use desktopcouch.records, that is [15:35] Good, good. [15:35] if you don't use dc.records, then you get to sign your requests yourself [15:35] the stuff to make dc.records understand oauth and get the keys is lp:~sil/desktopcouch/dc-records-oauth [15:47] aquarius, i wonder how much work it will be to make bindwood work with this oauth stuff [15:55] brb dist-upgrade wants a reboot [16:00] statik, there are already oauth libraries for JS [16:00] bindwood will have to get the keys, of course, and do the signing itself (this is what happens when you don't use desktopcouch.records), but that's relatively easy. [16:01] might need another shell-out-to-python to get the keys, thougth [16:01] makes sense [16:03] aquarius, should we ship a slightly modified version of your branch that doesn't turn oauth on, but makes it very easy to uncomment a line or something in order to enable oauth for testing? then we could do the changes needed in bindwood, evolution-couchdb, etc. in an orderly way, and finally flip the switch to turn on oauth [16:03] i'm kinda worried about breaking the world with the next desktopcouch release, and how to coordinate the introduction of oauth [16:07] We can do that, and I've been debating it in my head [16:07] Question: is shipping DC0.3 without auth and then turning auth on (by uncommenting the line) after freeze a break of the freeze? [16:10] statik, introducing oauth is essentially going to break the world for people who are not using dc.records. I can't think of a way around that. [16:11] because they'll have to start authenticating when they were not before. [16:12] Should we get 0.3 client out the door with trying auth, and then make 0.4 require auth? [16:13] thats what i'm thinking [16:13] 0.3 should let people integrate with auth [16:13] and 0.4 should require it [16:13] or, break things as soon as possible, since we may not be able to avoid breakage for people subverting us and doing it their onw way. [16:14] since we are waiting on the revised COUCHDB-478 patch to even allow desktopcouch to work correctly with the OAuth, I don't see how we can require it just yet [16:15] i'm all for getting the oauth code in so we can all be integrating with it [16:20] the problem is..."integrating with auth" is hard if it's diabled [16:20] disabled [16:20] so I think that putting it out without auth being required is the same thing as there not being any auth at all [16:21] hrm. I don't know what happens if you oauth-sign a request incorrectly when you didn't need to sign it at all [16:22] and...your couch is visible on the network, and unauthenticated. [16:22] which is Ungood [16:22] aquarius, it's the same as there not being any auth at all for a normal user, but for a developer integrating they could tweak the local.ini to enable auth, right? [16:23] they could, as long as that developer is prepared to break every app that talks to DC that they're not working on, like, say, bindwood [16:24] statik, but I am beginning to think that it's a good idea to do what you're suggesting [16:24] I did not think that having working oauth would take this long :( [16:25] still none of the usual suspects online on #couchdb. I think jan____ is en route to Dublin [16:26] aquarius, cool! i always wonder whether my ideas are crazy. when you start to agree with me, then I *know* they are. so, how should we do this? should I merge all your pending branches into desktopcouch, then you do one more to comment out the auth, then I release 0.3? [16:27] CardinalFang, do you have an in-progress branch that needs to go into desktopcouch 0.3? [16:27] disabling oauth means two things: comment out one line in the ini file, and then tell dc.records to not oauth-sign requests. I'm not even sure if we need to do the second [16:27] statik, Yes! [16:28] I have to go pick up my daughter and then have dinner with her and so on, so I'll be gone for about four hours, but then I'll be back. [16:28] aquarius, Can we make records try without oauth on failure? [16:28] CardinalFang, could do, but I'd be inclined to not do that [16:28] since that's extra code we write now and then remove in 0.4 [16:30] aquarius, i hope we don't need to do the second [16:31] if i send an oauth-signed request to something that doesn't require oauth, i would hope it just ignores the extra header [16:31] I think that's what happens. I think. [16:31] jblount, urbanape: either of you ever see an error in the web ui that says something along these lines: "this file is already being created by somebody else" [16:31] cool, we have a plan. could thisfred do the branch to comment out the ini file so you don't have to work so late tonight aquarius? [16:33] aquarius: could start on that right now, so I can ask you if I run into unforeseens... [16:33] when you get back [16:34] yes [16:34] ok, starting on that [16:34] I think that all you should need to do is remove the [couch_httpd_auth] require_valid_user = true part from create_ini_file [16:34] but I have not tested this. [16:35] that way there will still be an admin account and so on [16:35] this will, I should note, not affect people who already *have* an ini file with that in [16:35] but I think it's reasonable to get those people to delete the ini file [16:35] excellent, I' try, test and deliver [16:37] joshuahoover, you can ask gafton about that, i bet its from a stale upload reservation in the updown server [16:38] right, off to get Niamh from second day of dance school :) [16:38] bbl [16:38] statik: thanks...just found a previous bug with the same issue [16:52] * statik runs tarmac-lander on desktopcouch [16:54] * dobey runs to get lunch [16:58] aquarius: (running commentary, ignore until you're back) [admins] section *does* need to go in addition to require_valid_user, or we get the 401s. with those two sections commented out, the tests all run. Now merging all open branches into this to see if it doesn't break anything else. [17:04] done [17:08] prop0wzed [18:20] pfibiger, https://code.launchpad.net/~urbanape/ubunet/tooltips-for-long-filenames/+merge/10674 [18:22] hi thisfred, i've got a conflict when trying to merge lp:~sil/desktopcouch/not-so-random [18:23] thisfred, should I just merge your noauth branch instead, and that will pull in the others? [18:24] statik, I updated my system (including CouchDB) and Bindwood still seems to run okay. I get no errors messages in my JS console (well, not about connecting to Couch, anyway) [18:25] urbanape, fantastic! desktopcouch 0.3 will have optional oauth, and then later on we'll make it mandatory. that should give you a window to fix bindwood to support oauth [18:28] statik: exactly [18:28] I saw that exchange earlier [18:29] statik: I merged them all into mine and resolved the conflicts [18:29] thisfred, that is the best news i have gotten all day! [18:29] the conflicts are on all 3 of aquarius' branches [18:29] urbanape, thanks for testing bindwood and for the tooltips! [19:06] thisfred, i'm pushing a 1-line branch now that fixes a problem with writing the bookmark file [19:07] lp:~statik/desktopcouch/fix-bookmark [19:07] statik: will review and then I have to go pick up my stuffs [19:07] thisfred, awesome. hope all your stuff arrived safely! [19:08] there's nothing fragile or very expensive in there. If the CDs are ok, I'm ok :) [19:10] statik the tests pass but there's no proposal to +1 or is there? [19:10] thisfred, i'm just proposing it now, one sec i'll paste the link [19:13] https://code.edge.launchpad.net/~statik/desktopcouch/fix-bookmark/+merge/10677 [19:16] hello how to ad files to the ubuntuone folder [19:18] statik: I approved and approved [19:19] ta [19:21] know one knows how to ad folders to ubuntuone? [19:21] Mandrew: on the desktop? [19:21] on the computer [19:22] right, but with the local desktop client, or via the web browser? [19:22] i didnt find any awnsers in the faq [19:22] on the desktop [19:22] locally you just put new folders in ~/Ubuntu One/My Files (though soon that will just be ~/Ubuntu One) [19:23] i tried to drag n drop into the ubuntuone folder it dint like that :) [19:24] you need to put it in "~/Ubuntu One/My Files" currently [19:24] gotta go, be back in a couple hours [19:24] thanks m8 you solved my probz [19:25] Dobey yr d man [19:40] apropos my somewhat stalled multi-download branch: do we preserve permissions on files in U1? [19:44] is there anyone that knows if there is possible to get a email adress that look like this mandrew@ubuntu.com or some like it? [19:44] urbanape, i don't think we touch permissions [19:45] Mandrew, email addresses like that are available for Ubuntu members or developers [19:46] can i become a member or how does that work? [19:48] Mandrew, absolutely you can. membership is available to anyone who makes a significant and sustained contribution to the ubuntu community. You can contribute in any number of ways. There is a lot more information available here: https://wiki.ubuntu.com/Membership [19:48] one good way to become a member is by working with the nearest LOCO team [19:48] nice work ill check it out [19:49] loco team? [19:49] it's the name for local community teams of ubuntu advocates [19:49] they do all kinds of different work, from training to marketing to programming to answering questions in the forums [19:49] is there one here in sweden? [19:51] yes. https://wiki.ubuntu.com/LoCoTeamList [19:52] man your the greatest thanks alot m8 [19:52] happy to help, good luck with ubuntu [19:53] statik, how do we not handle permissions? What happens on a second machine or on a shared machine? Do we stamp files with some default permission? [19:53] im new to ubuntu and linux but i just love it the concept with it [19:54] urbanape, maybe i gave a wrong answer...chipaca or verterok or mentalguy would know for sure [19:55] urbanape: we don't handle permissions :) [19:58] urbanape, dunno if you saw my comments in the other channel about zipstream - i can help get that set up as a 3rd party sourcedep if you want [20:01] statik, we're not using the bare ZipStream. [20:02] I had to munge it a lot [20:02] there's probably room for a generalized solution that we can contribute back. I'd be happy to work on that. [20:02] ah, even more fun. so we still need to maintain it as a separate lib, and send those patches back to spideroak [20:02] yes [20:03] okay, that's no problem. [20:03] What I'll do then is to extract the bits that deal with the file selection (via the checkbox) and propose that separately. [20:03] I think it'd be nice to get the zipstream stuff generalized and contributed back. [20:03] and let someone else maintain it. [20:04] urbanape: In case I was clear when I was looking at it, I pretty much want to make babies with that branch. It makes me _very_ happy. [20:04] yeah, I was silly not to have actually tried to open the resulting files. [20:04] I just saw that the proper files were in the zip archive and were non-zero length. [20:04] urbanape: :), that's what dumb reviewers (like me) are for! [20:04] but yeah, gibberish. [20:07] * jblount goes back to long meetings with limited breaks === CardinalXiminez_ is now known as CardinalFang [20:12] * CardinalFang yells at twisted/dbus and its main-loop madness. [20:30] * aquarius returns [20:32] ...to discover that all his branches have been rejected. Nobody loves me. [20:32] :) [20:33] statik, so...what remains to be done for DC0.3 now? [20:33] aquarius, rejected and merged via thisfreds branch where he fixed up the conflicts [20:33] dvcs ftw [20:33] aquarius, i think we need a branch from CardinalFang, then we do the release [20:33] yep. he's a lovely Dutch hero, yes he is [20:33] omg it's done? [20:34] * aquarius breathes a mighty sigh of relief [20:34] aquarius, don't take my word for it [20:34] thisfred did the branch we talked about before you left [20:34] * aquarius grins [20:34] aquarius: you decided on "Gangsta Rap Made Me Do It"? [20:34] and CardinalFang says hes working on something mysterious, it's connecting to my machine [20:34] CardinalFang, how are you getting on with stuff? Need any help? [20:34] dobey, Cop Killer. :) [20:35] aquarius: do you have to sing also? [20:35] and the syncdaemon team is making AMAZING progress, with bandwidth throttling and resumable uploads and downloads [20:35] dobey, I do not. I wouldn't want to embarrass all those professional singers at stage school by being way better than them [20:36] and dobey is making a ui that everyone will love [20:36] blimey, we're gonna rock harder than Ayers Rock for karmic. Coolness. [20:36] aquarius: oh. you totally should have picked "Still Alive" then [20:36] aquarius: the credits song for Portal [20:37] one of the reasons that I'm looking forward to being out of crunch for the freeze is that I might get a chance to play Portal ;) [20:38] heh [20:38] although I don't thin kthere's a wii port :( [20:39] doesn't look like it [20:39] but you can buy it on steam and play under wine [20:40] plays under wine? nice. [20:41] i believe so [20:41] i have it on 360, so i can't confirm yet [20:42] however, i have been tempted to buy the awesome deal that valve had for all their games on steam [20:42] which is to say "every valve game, ever made, all for $100" [20:44] i think it was $100 [20:44] whatever it was, it was a great deal [20:44] that's pretty cheap [20:44] yeah [20:44] considering L4D is still like $50 in the box [20:44] although when I (infrequently) play games it's on the wii; that's why I bought a wii :) [20:45] i bought a wii for the cool toys [20:45] (wii fit) [20:47] aquarius, It's going. Struggling with DBus weirdness. [20:48] I'm happy to chip in if you'd like me to, elsewise I shall log off for the night [20:49] aquarius, I think just trying will take less time than synching with you would. Thanks though. [20:50] aquarius, G'night. [20:50] no worries :) [20:50] I cherish my status as an impediment to progress ;) [20:50] if you don't manage to finish (thisfred, statik, this applies to you too) drop me a mail with anything you'd like me to pick up on in the morning [20:52] Rgr. [21:05] grr, gtk+ [21:07] joshuahoover: btw, the applet no longer has the --signup/-s option. passing it is harmless, but it's not necessary :) [21:08] dobey: ah, thanks :) [21:11] dobey: is there an issue with the client crashing on startup? [21:11] dobey: i'm seeing quite a few of those being filed within the past 5 days now [21:12] yes and no [21:13] joshuahoover: some of them are fixed by the branch i landed this morning [21:13] dobey: very good...is there something i should look for in the logs that will tell me/clue me in that it's the same issue you fixed? [21:13] joshuahoover: others have python traces from syncdaemon, which seem to be the cause of failure [21:14] joshuahoover: if it doesn't have the syncdaemon-exceptions.log, it's probably just the fact that syncdaemon is being slow at start-up, and would be "fixed" by my branch from this morning [21:14] but that isn't in the nightlies/beta yet [21:15] and that bug is #414635 [21:15] dobey: good, i'll use that as the one to link others to [21:31] dobey: i'm seeing quite a few that have this as the last line in their syncdaemon-exceptions.log: State START_CONNECTING can't handle the SYS_SET_CAPABILITIES_ERROR event ...you think this is unrelated to the bug you submitted a fix for? [21:32] it is a separate issue [21:32] pick one, mark the others a duplicate of it, and assign it to chipaca :) [21:34] * Chipaca hugs dobey [21:34] * Chipaca embraces dobey, even [21:34] * Chipaca extends dobey [21:34] et cetera [21:35] heh [21:36] dobey: heh...got ya...thanks! [21:36] https://updown.ubuntuone.com/5d6e8680-5da0-4aff-9330-42caa0399a85 [21:39] dobey, i LOVE it! ship it [21:43] sigh [21:43] well now i have to hook it up [21:49] and we're back, with 24 boxes of crap in the otherwise empty living room! :) [21:54] dobey: ping [21:54] verterok: hi [21:54] dobey: hi [21:55] dobey: I'm finishing the dbus api for bandwidth throttling, and re-realized that dbus don't support dict with multiple types :/ [21:56] dobey: I was trying to return a int as the value, to make your life easier [21:57] well, it supports {sv} no? [21:57] yeah, a{sv} works [21:57] using a variant for the value? [21:57] (although that is annoying [21:57] dobey: yes, but v don't support NoneType :p [21:58] dobey: so, what do you think of not including the key:value when it's disabled (None)? [21:58] verterok: what about a separate set of methods for enabled/disabled? [21:59] verterok: or returning a negative value for disabled? [21:59] dobey: whatever that makes you and the UX team happy :) [22:00] dobey: using perl isn't an option ;) [22:01] i wouldn't say perl [22:02] and we don't have time to rewrite in C [22:02] dobey: so: {download_limit:, upload_limit=}, with int: -1:off, 0-n: on? [22:03] ok [22:04] verterok: hrmm, though separate enable/disable might be better [22:04] dobey: what are we going to do with 0, that would just stop using the network :/ [22:04] verterok: not use the network. it's the correct behavior :) [22:04] dobey: ok, that was the output of get_updown_limit (or something) [22:05] verterok: right, but as preferences, it would be nice if i set it, and those settings were saved, independent of me enable/disable the feature [22:05] verterok: i presume these settings are getting stored in ~/.config/ubuntuone/syncdaemon.conf ? [22:06] dobey: no [22:06] dobey: there is no work on syncdaemon side, I'm just starting [22:06] s/are/will be/ then :) [22:06] dobey: also, syncdaemon don't have persistent mechanism for preferences :O [22:07] verterok: you don't have configglue.write()? [22:07] dobey: we don't have the plumbing to do that [22:07] verterok: so i have to keep track of the settings, and hope syncdaemon doesn't get started without the applet? [22:09] * dobey doesn't like where this is going :( [22:09] and it's already 5 pm :-/ [22:10] dobey: good point [22:11] dobey: I don't think the applet should track syncdaemon settings, as you pointed out, that will be a mess if SD is restarted [22:12] i wasn't even thinking about it restarting out from under the applet, but yeah, that's even worse :) [22:12] "I set it to use 128K/s, but now it's using 2M/s" [22:12] fun! [22:13] dobey: yes, a *lot* of fun [22:16] dobey: btw, configglue don't have a configglie.write method :( [22:17] configglue doesn't have any way to save settings? [22:17] sounds like it needs more glue [22:17] dobey: no, or at least I can't find it [22:19] :-/ [22:20] does it give you a ConfigParser object? [22:20] dobey: no, an OptionParsr one :( [22:20] :/ [22:21] dobey: exactly [22:21] can you create a ConfigParser and just cache the throttling settings? [22:22] dobey: also, we shouln't store all the options configglue load at startup, as some are CLI options [22:22] right [22:22] dobey: ok, let's file a bug about this, looks like it wants it's own branch [22:22] dobey: IMO syncdaemon should support this [22:23] it kind of has to [22:23] so, let's keep with what we'r doing now, it's bug not a feature...right ;) [22:23] otherwise the throttling config is entirely useless :) [22:24] "doesn't save my settings" is a pretty big bug, yeah :) [22:25] dobey: :) [22:26] dobey: I'm filling the bug ATM [22:26] great [22:29] dobey: bug #418882 [22:29] Launchpad bug 418882 in ubuntuone-client "syncdaemon doesn't save my settings" [High,Confirmed] https://launchpad.net/bugs/418882 [22:30]