[10:51] The state of the Couch. My keynote presentation from the Erlang Workshop: http://short.ie/628z7e [10:52] oo, cool [10:52] * aquarius reads :) [10:54] hey, cool, we got some ink ;) [11:02] good talk, jan____ [11:05] thanks :) [11:27] hi all [11:27] ping losa? I dunno if these 'out of space' mails are already under control, or if maybe you're not getting them? [11:37] um [11:37] wrong channel, I suspect :) [11:38] * Chipaca needs coffee [11:39] ~ [11:39] c|_| [11:40] here ye go [11:50] jan____: thanks! although it sounds like the coffee machine is ready :) [11:50] :) [11:51] ok, bbl [13:11] Buenos días a todos! [14:47] CardinalFang, ping? [15:12] DESKTOP+ STANDUP MEETING [15:12] IF YOU'RE HERE, SAY "me" [15:12] me [15:13] me [15:13] statik, dobey, rodrigo_, teknico, jblount, cardinalfang? [15:13] me [15:14] aquarius, "MEETING BEGINS", please :-) [15:14] teknico, oh yeah, I knew there was some magic incantation I was supposed to type :) [15:15] aquarius, wow, *all* the people you called are away, or not here (except me) [15:16] cor, it's just us, then, I think :) [15:16] vds? [15:16] should we start? [15:16] ok, go for it vds, everyone else can chip in later if they need to [15:16] DONE: vacations, started a branch to improve funambol exchange logs, testing phone sync [15:16] TODO: finish the branch and propose it, some administration stuff [15:16] BLOCKED: no [15:16] aquarius: please... [15:16] ⚀ DONE: much discussion! [15:16] ⚁ TODO: allow unpairing of servers (bug #419975) -- still working [15:16] ⚁ BLOCKED: nothing [15:16] ⚂ BUG COUNT:https://bugs.edge.launchpad.net/~sil/+assignedbugs?field.tag=ubuntuone-karmic - 4 [15:16] teknico, it's all you [15:16] DONE: landed the branch with the details view for the new contacts web ui, and the one with the improvements to testing contacts creation, fixed bug #424434 [15:16] Launchpad bug 419975 in desktopcouch "for pairing, do not display hosts or servers that are already paired" [Medium,In progress] https://launchpad.net/bugs/419975 [15:17] TODO: land the branch for bug #424434, fix bug #423369, talk with jblount about the new contacts web ui, implement adding and editing contacts [15:17] BLOCK: none [15:17] Bug 424434 on http://launchpad.net/bugs/424434 is private [15:17] next: noone [15:17] Bug 424434 on http://launchpad.net/bugs/424434 is private [15:17] Bug 423369 on http://launchpad.net/bugs/423369 is private [15:17] oh, I forgot that I also tried to sync the phone, in my DONE [15:17] heh, private bugs [15:17] I need to try installing the android client [15:18] MEETING ENDS [15:49] ugh, missed the meeting, sorry [15:50] dobey: do you have some time for some oauth fun? :) [15:57] rodrigo_: today is a us holiday :) [15:59] ah, no oauth fun then for you :) [15:59] I'll poke you tomorrow then [16:00] are you wanting to fix tomboy? [16:00] or something else? [16:00] no, liboauth [16:00] ah, what problem? [16:00] it's generating different signatures than what python-oauth does, so just wanted you to look at the code [16:00] so yeah, forget it, will ask you tomorrow [16:01] the HMAC? [16:01] yeah [16:03] are you just comparing the base64 bits? [16:06] yes, using the same keys/secrets in evo-couchdb and a test python-oauth scripts generates different signatures [16:06] and timestamp and nonce [16:06] CardinalFang, are you on holiday today? [16:07] aquarius, yes. [16:07] timestamp/nonce shouldn't matter [16:07] CardinalFang, ok, no worries then :) [16:08] dobey: well, I mean I use the same data in both programs, and get different signatures [16:09] so liboauth seems to be doing something different, but from what I've seen, it is the same (hmac.sha1 and then base64) [16:09] rodrigo_: i don't know much about the hmac libraries, but i guess you need to compare the code in both libraries, to see how they are calling the HMAC libs, and what the results are [16:09] dobey: they both use openssl, it seems [16:10] rodrigo_: probably, but what API they call, and how they use the results is what matters [16:10] rodrigo_: though i suspect python-oauth is the broken one [16:10] dobey: that's what I wanted you to look at :-) [16:11] dobey: ah, really? [16:11] dobey: but tomorrow :-) [16:11] aquarius, You saw I'm screwing around with the paired records and where we store service-related information, yeah? [16:11] CardinalFang, I in fact did see that :) [16:12] aquarius, blame dobey. He convinced me it's better. [16:13] It is better [16:13] those americans are crazy, working on a holiday :D [16:13] but I've made some pretty sweeping changes in my branch [16:13] so merging yours in basically makes the whole world conflict. [16:13] annoyed with myself [16:13] and I've lost track of how it all works. [16:13] I suck [16:13] :( [16:14] Alright, lunch is nearly finshed cooking. Laters. I'm leaving this client running, but not here. [16:15] I am seriously thinking about going with CardinalFang's branch and doing repairing a different way. [16:25] rodrigo_: can you verify that the PLAINTEXT signature is the same in both? [16:25] dobey: ok [16:29] dobey: hmm, almost identical, except for a % escaped char, it seems [16:30] rodrigo_: oh? which one has the extra escaped char? [16:31] hFbTtoEagx%2526uyMzGrgtLT -> liboauth [16:31] hFbTtoEagx%26uyMzGrgtLT -> python-oauth [16:31] I'm not encoding/decoding anything, so that's what liboauth returns [16:31] liboauth is broken [16:32] I guess in the encoding code, right? [16:32] it's encoding & as %26 and then-reencoding that as %2526 [16:33] escaping the % [16:33] but where, i don't know [16:34] with hmac though: [16:34] bv6w2jDu0hbPWwVkQbUQpAcRsOc%3D -> liboauth [16:34] fO7lX5eRxVZVirX4rViaSxQjOlE%3D -> python-oauth [16:35] * rodrigo_ looks at liboauth [16:37] it doesn't encode the signature, AFAICS [16:37] hrmm? [16:38] oh, yes, it does [16:39] it looks to me like either the liboauth base64 implementation is wrong, or HMAC is returning a different dijgest in liboauth for some reason [16:54] rodrigo_: make a simple test that just does the HMAC() call in C with hardcoded values, and another simple test script in Python that uses the same hardcoded values, and compare the result string in C with the result.digest() from python [16:56] rodrigo_: if those are equal, i suspect the problem lies in the base 64, or the values being passed to the HMAC call [16:56] ok [16:56] rodrigo_: should be easy enough to figure out :) [16:56] * dobey goes back to his holiday now :) [18:21] hi [18:21] Hi ppq [18:23] i copied a folder (about 5000 small files, 450mb) into my ~/Ubuntu One/. it started to synchronize, then the applet disappeared. ubuntuone.com shows 14,6mb of 2gb as used space, this number doesnt increase. how do i resume the upload? starting the ubuntuone client only says its already running... [18:27] ppq, the applet disappears because it stopped an active task (it should be all uploaded now) [18:31] facundobatista, it doesn't continue. it showed 14,6mb an hour or so ago and until now, nothing has changed :( the folder has been copied completely to ~/Ubuntu One/. i am the owner, it's drwxrwxrwx. [18:32] ppq, but it's ok that it doesn't continue if it finished [18:32] why do you say it didn't finish? [18:33] because i can access only 14,6mb of those 450mb via ubuntuone.com [18:33] there are many files missing [18:34] ppq, oh, I got it... [18:34] :) [18:35] ppq, two things can happen: a) it didn't upload everything b) it uploaded everything and is not shown in the web [18:35] ppq, we can know that by looking into the logs... do you want to give them a look? [18:37] i can't find the logfile, where is it located? [18:40] ppq, ~/.cache/ubuntuone/syncdaemon.log [18:42] this file doesn't exist. but theres a folder ~/.cache/ubuntuone/syncdaemon/ with one file metadata_version and two sub-folders: fsm/ and vm/ [18:43] ppq, bu, sorry, bad copy and paste [18:44] ~/.cache/ubuntuone/log/syncdaemon.log [18:44] ppq, ^ that's it [18:46] wow, 33mb :) [18:46] i think i should zip it [18:48] the last lines: 2009-09-07 19:46:54,564 - ubuntuone.SyncDaemon.Main - NOTE - ---- MARK (state: EXCESSIVE_TIMEOUTS; queues: metadata: 4643; content: 4718; hash: 0, fsm-cache: hit=414351 miss=5579) ---- [18:49] i think thats not healthy.. 33mb logs after first use [18:50] facundobatista, do you want to see the log? http://ppq.ath.cx/syncdaemon.log.tar.bz2 [18:53] wait, thats to slow, i will upload it to a webserver [18:57] ppq, we're having some strange issue with the server now [18:57] ppq, it should exit from that EXCESSIVE_TIMEOUTS alone, and continue working when the issue is fixed [18:58] ppq, I'll update you the status when know better [18:58] muchas gracias por ayudarme :) [18:58] ppq, :) [21:27] facundobatista, ahora está sincronizando :) [22:15] I seem to be having trouble with my files being synced, but not the data in them