=== zyga is now known as zyga-afk [07:46] morning all! [07:46] * mandel is surprise how good 3g works in mallorca... [07:48] morning [07:53] morning [07:59] mandel, 3g... mmmm.... do want... [08:00] rye: I'm surprise it works this well…. I'll upload a pict of my new office :) [08:00] rye: working looking at the mediterranean is great :P [08:01] mandel, I'll send you a postcard full of envy [08:02] hahaha [08:02] Ireland is sunny too [08:02] :) [08:02] rye: at least you are not working on windows and dealing with its stupid encoding errors… god I hate the windows file system... [08:03] mandel: do you not develop on mac and pull into the vm? [08:04] mandel, well, you know... there were 3 standard encoding for cyrillic, koi8-r, windows-1251 and, finally, utf-8. I had all of the filenames encoding on my machines resulting in so many issues... ext fs do not specify the encoding of the filenames at all [08:04] mandel, but not working on windows is good enough [08:05] fagan: I do, but I'm dealing with a bug I can only reproduce irl… I need to understand how windows deals with this kind of paths. Funny thing, I just discovered it because I have some nordic music and the bastards use some funny char in the songs name :P [08:05] rye: yeah.. at least in spanish we just have accents and ñ :) [08:06] mandel: ahhh ok now I get it [08:06] mandel, here, have a cyrillic letter Zh, Ж :) [08:07] rye: that is fucked up hahaha [08:07] wow its raining now but its still really sunny [08:19] * mandel wainsting time while windows updates... [08:29] morning all [08:33] morning duanedesign [08:33] * fagan needs to head to the shop real quick [08:53] I see there regularly are people asking about couchdb sync but never get an answer. Is there any good place to ask about it? [08:54] rye: what is the latest news about the couch sync? [08:55] duanedesign, new couchdb is being tested on staging. The beautiberry shard (user ids 0 - 475000) has replication disabled, all other shards are working [08:57] wimpunk: ^ [08:57] duanedesign: thx! [08:57] your very welcome [08:58] .5 [08:59] rye: are we allowed to answer with such techy things to users... [08:59] * mandel is going to love to answer windows bugs.. [09:00] mandel, well, the user id is public info, the shard names are not secret (and they are not visible from outside) and it is better to give a definite answer for technical question (and replication is somehow a technical term) [09:00] duanedesign, let me update ubuntuone-account-info script so that it works in natty, it will print the account id and other info out === zyga-afk is now known as zyga [09:01] rye: :P [09:01] mandel, speaking of windows bugs... [09:02] * mandel hides [09:04] rye: tell me :) [09:05] mandel, I don't have a hands on experience with windaemon yet and i need some assistance for... [09:06] rye: btw, tnx for those beautiful scripts. Pretty useful for people who wants to know what's going on. [09:07] wimpunk, thanks for the kind words. I just hate when i don't know what's happening [09:07] rye: +1 :-) [09:23] * rye feels writing scripts for windows is quite a bit harder :( [09:38] * fagan tea break [09:42] Hi, no news about that ? "2011-06-16 Syncing of contacts in Evolution is not working for some users due to a server issue. We are looking into this and trying to resolve it ASAP. " My couchDB addressbook crash makes the personal one unusable also after a while. [09:44] lalejand, I'm no expert, but I don't see how the couchdb server outage could cause your addressbook to crash. [09:45] jo-erlend : maybe it's not related then :) [09:45] I will see that when the couchDB one will be fixed. [09:47] reading the comments in launchpad about a related bug, I understood that the problem was fixed for users one by one. I just wanted to say "not yet fixed for me" :) [09:51] lalejand, perhaps it would be wise to say which bug you're referring to? [09:52] on https://wiki.ubuntu.com/UbuntuOne/Status#Contacts it says "'An issue preventing Evolution contacts and Firefox bookmarks from syncing has been discovered and we're working on fixing it. Bug 788206" [09:52] Launchpad bug 788206 in ubuntuone-servers "Oauth authentication against couchdb.one.ubuntu.com fails (affects: 16) (heat: 86)" [High,Confirmed] https://launchpad.net/bugs/788206 [09:55] lalejand, that may not be the current issue. If you look in ~/.cache/desktop-couch/log/desktop-couch-replication.log, you can see if you get errors about a server returning an error 503. I don't think that's caused by OAuth issues. [09:56] jo-erlend : this file ends with "HTTPError: HTTP Error 400: BAD REQUEST 2011-07-05 11:50:16,039 DEBUG finished replicating" [09:57] but file sync looks okay [09:57] ah. Ok. [09:57] ho, but ine U1 window it says that my internet connection is bad :/ [09:57] lalejand, have you tried to delete the key from your gnome keyring and log back on? [09:58] "L'information n'a pas pu être récupérée. Vérifiez votre connexion à Internet." [09:58] jo-erlend : nope, I try right now [09:59] jo-erlend : is it that : "UbuntuOne token for https://ubuntuone.com" I just delete it ? [09:59] yes. [10:00] then, when you open the Ubuntu One client, it should ask you to log on. [10:02] jo-erlend : it says "an exception representing a server-side error" [10:02] lalejand, what says that and when? [10:04] jo-erlend : I deleted the key, disconnected from U1 and reconnected in the U1 param window, it asked me to create an account or click on some button if I have already one, I clicked the button, I entered my email and password (login/pwd that I use to connect to the web interface) and I have a red message in the window saying that [10:05] what does your logfile say? [10:05] jo-erlend, Ho I just tryed again, it worked now [10:05] let me check if things are ok now [10:05] because that is correct. It should now say something about an error 503. That means the OAuth issue was corrected and that you now only have the same problem everyone else has. :) [10:06] no, I don't think your couchdb will sync. That's what the server-side error is about. [10:06] jo-erlend : in the main tab of U1 window it says something like "information couldn't be retrieved. Please check your internet connexion" [10:07] it is much more helpful if you paste the past entries from your logfile. [10:07] jo-erlend : ok so the server-side bug isn't fixed since one month now ? [10:07] ok [10:07] I paste [10:08] I don't think it's been down so long. I encountered it before this weekend. [10:08] jo-erlend : http://pastebin.com/qwLsHWpR [10:10] oh, I haven't seen that before. I don't understand what it means either. [10:10] what version of Ubuntu are you using, btw? [10:11] argh, if I go to the "connected devices" tab, I see 3 times my laptop :( [10:11] how often does tarmac run? mandel, do you know? [10:12] jo-erlend : and all 3 added today pfff [10:12] lalejand, what operating system do you use? [10:12] that is weird, a 400 error… I would have expected something completely diff [10:12] natty [10:13] in the web-interface I see also 3 times my laptop [10:13] mandel, yes, me too. But he said something about an addressbook crash. Perhaps that's caused some issues? [10:13] lalejand, I've seen that too, and I don't know why it happens, but I don't think that's caused any problems for me. [10:14] jo-erlend : ok, but it's weird :) [10:15] what if I delete all the associated devices, and connect again ? It will replicate what is on the server to my laptop and then I will loose last version on my files here ? [10:16] no, I don't think so. What you can do, is to create a new user, log in as that user and open Ubuntu One there. [10:17] but as I said before; couchdb sync doesn't work for anyone at the moment, so your databases won't work. But the error 400 might go away. [10:17] jo-erlend : in order to verify what ? [10:17] the connected devices shouldn't cause that problem. [10:17] jo-erlend : ok [10:18] but if you want to see if it's a problem on your local data, you can create a new user and log onto ubuntu one with that. Since that'll be a blank account, there isn't anything on it to cause any problems so you should be able to log in. [10:18] I mean a new account on your PC, not Ubuntu One. [10:19] in https://wiki.ubuntu.com/UbuntuOne/Status#Contacts they say the bug is present since the 16th of june ! [10:19] jo-erlend : ok [10:19] lalejand, "the bug"? [10:19] there might be several different ones. [10:19] jo-erlend : "2011-06-16 Syncing of contacts in Evolution is not working for some users due to a server issue. We are looking into this and trying to resolve it ASAP." [10:20] Can someone translate this for me? Google Translate can't handle it well: "Вовремя появилась. Аналогичные удалил." [10:20] right. But now, no couchdb sync works at all, and I think that's a new issue. But that will also affect Evolution. So I think they're two different bugs. [10:20] two different issues, I should say. [10:21] jo-erlend : ok [10:21] lalejand, can you try what I suggested, just to confirm? [10:21] jo-erlend : ok [10:23] jo-erlend : should I close my session ? [10:23] lalejand, no, that's not necessary. [10:23] ok [10:23] you can just switch to that other user and log onto Ubuntu One. [10:24] I would be more than a little surprised if that didn't just work. [10:27] jo-erlend : ok I have the same error message on the main tab in U1 window ("information could not be retrieved, ..."). U1 is downloading stuff to my HD, looks like things that are in my U1 folder. [10:28] lalejand, please paste the logfile from that account. [10:28] replicating log ? [10:29] yes. [10:34] jo-erlend : http://pastebin.com/gSQ2d8kD [10:36] no errors there. That's the complete file? [10:36] yes [10:37] well, as I said, I'm no expert. But at least now you have something to show the experts. :) [10:37] ups, now I got errors [10:38] http://pastebin.com/Cta6SJ2Z [10:38] this is the complete file from 1 min ago [10:39] yes, that looks more like the ones we're all getting... Except, I get 503. [10:40] arf [10:41] but that probably means there are some issues on your clientside. But I have no idea how to troubleshoot them, so that's for someone else. [10:41] jo-erlend : on my alternative session I have error 500, and on mine I have error 400 :/ [10:44] yes, but 500 is not a problem. That's on the serverside, so that'll be fixed. :) [10:44] I have to go. Good luck :) [10:44] @ping [10:44] pong [10:46] jo-erlend : thx for yout help [10:46] "your" [10:49] where are all the argentinians? [11:06] is there a specific help/hotline for paying U1 users ? [11:08] lalejand, /me is reading the scrollback [11:09] hmmm [11:11] ok, ATM the sso/u1 login is working, reading further [11:13] good morning! [11:13] karni, Can someone translate this for me? Google Translate can't handle it well: "Вовремя появилась. Аналогичные удалил." - "Appeared in time. Removed the similar ones" but i need more context [11:13] karni, for cyrillic you can poke me directly [11:15] ralsina: morning! [11:15] I was wondering where everyone was... [11:15] lalejand, so, regarding 400 BAD REQUEST (which should be 401 UNAUTHORIZED soon, actually) - this means that the account tokens that are stored locally are not usable, i.e. they have been removed from the server etc. Since it appears like you have reauthorized your machine the desktopcouch might not have picked up the new credentials (which looks like a bug) [11:16] rye : just filed a bug report : https://bugs.launchpad.net/evolution-couchdb/+bug/805897 [11:16] Launchpad bug 805897 in evolution-couchdb "couchDB adressbook doesn't work (affects: 1) (heat: 6)" [Undecided,New] [11:16] lalejand, you can restart desktopcouch - /usr/lib/desktopcouch/desktopcouch-stop; killall desktopcouch-service; and then run /usr/lib/desktopcouch/desktopcouch-get-port - that will restart replication with new token for sure [11:16] lalejand, what I am very interested is the internal server error message [11:16] rye : I try that on my alternative session ? or mine ? [11:17] lalejand, let me check your account [11:19] rye : one precision : I have nightly PPA packages for U1 [11:20] lalejand, that should not matter for server-side errors, so that's ok [11:23] mandel: sorry, I started working at 8AM today! Lazy me! ;-) [11:24] ralsina: you lazy!!! [11:24] hehe [11:24] ralsina: you can always say there are lazier and get up at 7AM [11:25] ralsina: I fixed correctly the encoding issue in that last branch and I'm waiting for the bzr push to finish… I hope it does not brake due to my connection :P [11:26] * fagan needs to do a big find and replace on this rst thingy [11:27] muffinresearch: little question I have the names of all of the images but I dont know what to put in before it so I can have the right link to it. What will the url be for the images? [11:28] I can just do a s/::/::url and it would be good [11:28] (at least for all the images) [11:29] ralsina: can you please take a look at https://code.launchpad.net/~mandel/ubuntuone-client/fix-805372/+merge/66887 [11:29] rye: sweet, thanks. it means the person installed Ubuntu One Files, and removed similar software ;) \o/ [11:29] ralsina: and remind me to ping alecu or nessita about it :P [11:29] * karni goes back to sleep [11:30] night karni [11:30] karni, night :) [11:30] \o =) [11:30] lalejand, ok, it is a server-side bug, currently filed as critical and i am starting to poke developers [11:31] rye : ok I will have news via my bug report ? [11:32] mandel: looking [11:32] rye : is there something I can do ? [11:32] cheers [11:34] mandel: much better [11:35] lalejand, could you please confirm that it was working before thursday/friday last week? [11:35] ralsina: yes, that does the trick in a nicer way, and will only put funny chars in the logs since __str__ is just used for logging… [11:36] lalejand, OTOH the couchdb/evolution nightlies may be an issue that prevents it from working in natty. rodrigo_, ping - do you have any way to test latest ubuntuone nightlies with evolution couchdb? [11:36] rodrigo_, on natty [11:37] rye, the nightlies don't build afaik, but once they do, they should be available for natty, shouldn't they? [11:37] rye : nope, it was not working before thursday/friday las week [11:38] mandel: +1 but wait for someone that knows more syncdaemon's approval [11:39] rye : mi couchDB addressbook is broken from a longer time (maybe related to https://wiki.ubuntu.com/UbuntuOne/Status#Contacts "2011-06-16 Syncing of contacts in Evolution is not working for some users due to a server issue. We are looking into this and trying to resolve it ASAP") [11:40] rodrigo_, hm, i mean that latest nightlies for natty do not work with evolution at all, leading to weird timeout problem dialogs [11:40] * fagan break [11:40] lalejand, let me finally update the script to find out the account info... which will not work due to the bug [11:40] grrr [11:40] rye, ah, haven't tested in natty for long, so will have a look [11:41] rye : should I leave PPA packages and use regular ones ? [11:41] lalejand, to get ubuntuone addressbook locally working, i am afraid, yes [11:42] rye : ok, How can I downgrade ? [11:42] lalejand, there is a ppa-purge application in ubuntu repositories [11:43] ralsina: ok, although is more of a 'someone that know more sd on windows...' [11:43] and we are not that many ;) [11:43] mandel: exclusive club that one :-) [11:43] mandel: oi i hope you have a lot of bugs there today mister!! cheeky [11:43] precisely that one :P [11:44] czajkowski: hahaa it was an easy comment to make ;) [11:44] * mandel thinks opera turbo is a placebo... [11:44] mandel: may the karma gods be evil to you today and fill your inbox full of bugs! :p [11:44] czajkowski: or a very slow connection hehehe [11:45] hello everyone! [11:47] hola nessita! [11:49] @ping [11:49] pong [11:49] rye : I just noticed that I have also rye/ubuntunone-extras :) [11:50] rye : this is for the icon in the tray, right ? I don't remember [11:50] lalejand, that will work... most likely [11:50] ry : I can leave it then [11:50] rye : I purge the ppa:ubuntuone/nightly, should I do something special now ? [11:50] yeah, /api/quota is not affected, accessing /api/account crashes [11:51] lalejand, u1sdtool --quit; and stop desktopcouch - /usr/lib/desktopcouch/desktopcouch-stop; killall desktopcouch-service [11:51] rye : ok [11:51] rye : stop: Unknown job: desktopcouch [11:52] ah sorry [11:52] lalejand, um... === teknico is now known as teknico_away [11:52] ok [11:52] what do I have to write , [11:52] stop desktopcouch - /usr/lib/desktopcouch/desktopcouch-stop ? [11:52] or /usr/lib/desktopcouch/desktopcouch-stop ? [11:53] * ralsina <=======[coffee] [11:53] lalejand, just /usr/lib/desktopcouch/desktopcouch-stop [11:53] ralsina << coffee; [11:53] * rye hates bitwise overloading of C++ for streams [11:53] rye: I am C++ incompatible ;-) [11:53] rye : ok everything done [11:54] I start it again ? [11:55] lalejand, start ubuntuone-control-panel, it will start syncdaemon as well, but will show errors regarding account information, desktopcouch replication may be working and evolution.. you need to restart evolution too -- evolution --force-shutdown [11:55] and start evolution again [11:56] mandel: hi there! Could you please send your report from yesterday? [11:56] rye : my laptop appears 4 times now in "connected devices" [11:57] rye : \o/ couchDB addressbook looks ok [11:57] lalejand, that's ok, you can delete other laptop entries from control panel/devices [11:57] rye : ok [11:58] nessita: sure, sorry I left early and forgot [11:58] rye : about mobile contact sync --> so we should not use it ? [12:00] "on https://wiki.ubuntu.com/UbuntuOne/Status#MobileContacts it says that Mobile contacts are inactive. On http://voices.canonical.com/ubuntuone/?p=943 it repeats that Mobile Contact is OFF. But on https://one.ubuntu.com/mobile/ Contact sync is still present. So it is working or not ? We should use it or not ?" [12:03] facundobatista: hola! and ping [12:03] lalejand, hm, i am editing the wiki now and the service has been shut down, the /mobile/ page will need a copy update which I am sure is being done at the moment [12:04] rye : so no mobile contact sync for now. Do we have an idea of when the new service will be available ? [12:04] facundobatista: regarding your review in https://code.launchpad.net/~mandel/ubuntuone-client/fix-803984/+merge/66758, the solution that mandel proposed is what we discussed in a call with Chipaca and ralsina. We decided we will force the 'b' everywhere, except when the caller explicatively states that he does not want the binary mode on. [12:05] lalejand, i don't have a timeframe at the moment [12:05] facundobatista: the sentence "those who know how to use open" has proven to be False, since the syncdaemon code did not have the open calls with the proper mode [12:05] rye : ok [12:07] mandel: you reverted the changes in the branch? [12:09] nessita: yes I did [12:09] mandel: why? [12:10] nessita: I mentioned yesterday, the changes are not big and in python 3 you will have to pass the b on linux, so why not doing it? [12:11] mandel: not following. What we will do with all the new bugs because we forgot a 'b' until we migrate to python 3? :-) [12:11] nessita: fix them? [12:12] mandel: we can *avoid* them, remember, "it just works". If we forget even one 'b', we will have thousand of users that will not be able to use the service [12:12] nessita: I mean, I dont know, I had a needs fixing with a very decent reasoning behind it from a devel from chicharra… [12:13] mandel: but in the meeting we agreed something else, and we even discussed what facundobatista mentioned [12:13] nessita: and then I mentioned it on the meeting on monday with ralsina and Chipaca, and we dicussed the comment from facu [12:14] mandel: and what did you agree on? [12:14] nessita: to follow facus comment... [12:14] nessita: reverting, as suggested by one of the chicharra developers. [12:16] nessita: can you take a look at the vm plz [12:16] lalejand, the server release today will fix the account issue [12:16] Chipaca: which one? :1 or :3? [12:16] nessita: :1 [12:16] nessita: :3 is gone [12:18] Chipaca: that looks like the contents of :3 [12:19] nope, :3 was win7_base, this is win7_001 [12:20] Chipaca: ah, true, my confussion [12:21] Chipaca: well, that looks really good :-) [12:21] nessita: still a lot of issues with it, but yes [12:22] Chipaca: you started only the control panel there? [12:22] nessita: yes [12:22] ok, I'll try with the other services [12:22] nessita: syncdaemon gives a bunch of key errors [12:22] Chipaca: do we have that in trunk? [12:22] (the styling) [12:22] nessita: not yet, pep8 issues [12:22] speaking of the devil [12:22] * Chipaca grins at DiegoSarmentero === zyga is now known as zyga-food [12:26] Can I get a review for https://code.launchpad.net/~ralsina/ubuntuone-windows-installer/fix_804600/+merge/66731 ? It has screenshots! [12:32] alecu: hi there! could you please send the report from yesterday? I'm hungry to know what happened with the u1cp :-) [12:32] ralsina: sure [12:33] nessita: thx! [12:34] nessita, hi there. My branch to make signals removeable needs one more review: https://code.launchpad.net/~alecu/ubuntuone-client/removable-signals/+merge/66807 [12:34] nessita, it's for the error where deferreds where being called twice. [12:35] alecu: I'll review after ralsina's [12:35] DiegoSarmentero: ping [12:35] s/where/were/ [12:36] nessita, pong [12:36] DiegoSarmentero: hi there! how is it going? [12:37] nessita, fine, i've to polish some ui, but it is on the path :P [12:37] ralsina: these two should be swapped, right? [12:37] 181+ congratulations_ui, [12:37] 182+ choose_sign_in_ui, [12:37] nessita: alphabetical! Forgot :-( [12:37] they are so close ;-) [12:37] nessita: pushing in 1' [12:40] nessita: pushed [12:40] ralsina: ack [12:41] ralsina: you could say 'ch' comes after 'c' [12:41] ralsina: and then we'd all laugh at how *old* you are [12:43] Hmmmm.... ch came after c until 15 years ago. I amover 15, indeed :-) [12:51] mandel alecu fagan nessita thisfred dobey DiegoSarmentero standup in 9' [12:51] ack [12:51] ack [12:52] I wonder why my nick doesnt highlight comments if my name isnt the first line [12:52] weird [12:52] meh === teknico_away is now known as teknico [12:52] because your irc client is lame [12:52] fagan: i had that issue too...trying to remember how i fixed it [12:52] dobey: its irssi [12:53] hi dobey! question: was this fixed by your branch in ussoc? bug #805244 (report says ussoc 1.3.1 was used, so I guess no?) [12:53] Launchpad bug 805244 in ubuntuone-control-panel (Ubuntu) "Method "Create Collection" with signature "a{sv}" on interface "org.freedesktop.Secret.Service" doesn't exist (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/805244 [12:54] ralsina: btw, your branch yesterday with the voting issue; it was because you requested two reviews, not because two are required. [12:54] dobey: I know [12:54] nessita: that is a different issue than the one i fixed, so maybe an update in gnome-keyring broke some other API :( [12:54] dobey: I couldn't find how to remove one after I asked, though, so I voted it myself :-) [12:54] nessita, ping [12:54] dobey: can I assign that bug to you? [12:54] fagan: crappy irc client? [12:54] nessita: sure [12:54] facundobatista: pong (about to have standup) [12:54] dobey: thanks! [12:54] nessita, hola! [12:55] ralsina: is irssi crappy? [12:55] * fagan always thought it was nice [12:55] nessita, che, should I tell visual c++ to install mssql server? [12:55] fagan: no idea, never used it. But if it doesn't highlight your name, by definition, it is [12:55] facundobatista: hola! and nopes [12:55] nessita, or it is not needed? [12:55] nessita, thanks! [12:55] facundobatista: anytime [12:56] ralsina: I think it might have something to do with my config [12:56] ralsina: why do you always choose an odd number to let us know about the stand up? [12:57] fagan: irc clients are like cars. stock is lame. [12:57] mandel: the alarm starts at 10, then I notice, the I finish something, then I ping. Usually takes 1-3 minutes :-) [12:57] dobey: yeah irssi would be pretty lame if you didnt config it at all [12:57] :D [12:57] ralsina: that is what she said :P [12:57] lol [12:57] dobey: I was looking into plugins but it seemed like a lot of effort [12:58] hehe [12:58] fagan: you don't need an extension for that [12:58] fagan: /help hilight [12:58] ahhh ok then [12:58] thanks dobey [12:59] fagan: this is what I had in my ~/.irssi/config. Looks like I had to specify my nick. Hope that helps http://pastebin.com/WgVwGvkq === m_conley_away is now known as m_conley [12:59] mandel: aft5er an issue that DiegoSarmentero had during the weekend with our xdg, I reported: #805927 [12:59] bug #805927 [13:00] Launchpad bug 805927 in ubuntuone-client "UnicodeEncodeError when importing windows' version of xdg (affects: 1) (heat: 6)" [Medium,Triaged] https://launchpad.net/bugs/805927 [13:00] me [13:00] me [13:00] me [13:00] me [13:00] duanedesign: nice thanks [13:01] me [13:01] me [13:01] me [13:01] go ralsina [13:01] dobey: me? [13:01] DONE: windows call, closed #805461 #804562 #804569 #804271, proposed branches for #805405 #805290 #803929 [13:01] TODO: close the ones I have proposed, #804610 #800376, calls [13:01] YOUR TODOS: objectives [13:01] BLOCKED: no [13:01] NEXT: nessita [13:01] fagan: bit premature there [13:01] ralsina: ah sorry forgot about dobey [13:01] me [13:02] DONE: day off. The day before that: support to DiegoSarmentero so he had his setup ready to go for development, tested and fixed instructions at https://wiki.ubuntu.com/UbuntuOne/Contribute/WindowsTesting (they are certified now!) [13:02] TODO: make control panel tests pass, assitye the rest, have 'account api' call with web team? [13:02] BLOCKED: nopes [13:02] NEXT: DiegoSarmentero [13:02] DONE [13:02] ralsina: that's what she said ;) [13:02] Some minor improves in the UI, ProgressBar in the right upper corner replaced for an overlay animation while Ubuntu One is connecting to the server. [13:02] TODO [13:02] Working on Table style. Polish ConnectingWidget and add Ubuntu Font. [13:02] BLOCKED [13:02] no [13:02] DONE [13:02] * Fixed some of the photo bits in the rst [13:02] TODO [13:02] * Finish off the branch [13:02] Blocked [13:02] * nope [13:02] thisfred: go [13:02] DONE: Celebrated the birth of grilling hamburgers or something like that TODO: Check mail | Bug #762722 BLOCKED: no NEXT: mandel [13:02] Launchpad bug 762722 in ubuntuone-control-panel (and 1 other project) "Opening the Control Panel during sync removes progress indicator (affects: 1) (heat: 9)" [High,In progress] https://launchpad.net/bugs/762722 [13:03] DONE: finished bug 803984 and bug 805372, looked at the other encoding bus, we should be suing str for paths, pass the to sd, and encode them to mbcs later so ensure things do not brake [13:03] TODO: Ensure filesystem notifications do not use utf8 [13:03] BLOCKED: no [13:03] alecu, go [13:03] Launchpad bug 803984 in ubuntuone-client "Windows: Inifite loop between upload started -> upload error (TRY_AGAIN) (affects: 1) (heat: 6)" [High,In progress] https://launchpad.net/bugs/803984 [13:03] Launchpad bug 805372 in ubuntuone-client "Windows pyinotify implementation cannot deal with utf8 (affects: 1) (heat: 360)" [High,In progress] https://launchpad.net/bugs/805372 [13:03] DONE: a branch to fix signals not being removed (bug #805471); a few reviews for ralsina; more debugging on u1cp->sso [13:03] TODO: more debugging. expenses! objectives? [13:03] BLOCKED: no [13:03] Launchpad bug 805471 in ubuntuone-client "Signals not removed means deferreds called twice (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/805471 [13:03] NEXT: dobey [13:03] λ DONE: hack day, us holiday [13:03] λ TODO: bug #805244, start the shiv^Hm?, more fixes/triage/etc [13:03] λ BLCK: None. [13:03] Launchpad bug 805244 in ubuntuone-control-panel (Ubuntu) "Method "Create Collection" with signature "a{sv}" on interface "org.freedesktop.Secret.Service" doesn't exist (affects: 1) (heat: 6)" [Medium,Triaged] https://launchpad.net/bugs/805244 [13:04] ok, comments! [13:04] Please everyone, objectives? [13:04] thisfred and dobey, got yours, need to approve them [13:04] * fagan has a standing objective to be awesome if that counts [13:05] ambitious! [13:05] ralsina: yes, objectives. Can we mumble about them after our daily mumble? [13:05] nessita: of course [13:05] ralsina, nessita, alecu: when do we have mumble? [13:05] fagan: you are excepted from the objectives TODO [13:05] I need to have lunch... [13:05] Chipaca, ralsina, alecu, mandel: mumble at 10:30 ART? [13:05] nessita: ok [13:06] ralsina: yeah I know :) [13:06] agg objectives.. [13:06] mandel: you have 25 minutes, if that's enough [13:07] nessita: would not mind 10/15 more, just because I'm at my parents place and they are old people... [13:07] they could choke ;) [13:08] mandel: so, not sure what you prefer then [13:08] nessita: at 45 is ok for you? [13:08] nessita, did you submit how to fix the bug reported about xdg? [13:08] it is to me, let me ask Chipaca [13:09] Chipaca: can you make it to the mumble in 33 minutes? [13:09] DiegoSarmentero: yes, is bug #805927, you can subscribe :-) [13:09] Launchpad bug 805927 in ubuntuone-client "UnicodeEncodeError when importing windows' version of xdg (affects: 1) (heat: 6)" [Medium,Triaged] https://launchpad.net/bugs/805927 [13:09] DiegoSarmentero: I fixed it, you have to pull from lp:~mandel/+junk/xdg [13:10] mandel, ok! [13:10] DiegoSarmentero: wait, I'm talking about a diff bug, is there a unicode error in xdg? [13:10] mierda.. [13:10] puto fs de windows y la madre que lo pario [13:10] DiegoSarmentero: I'll take a look at it asap [13:10] mandel, no problem, i fix it on my machine... so it's working here [13:11] mandel: go have lunch, we will wait for you [13:11] DiegoSarmentero: can you put that changes in the bug, that way I dont have to think much ;) [13:12] ralsina: ping [13:12] nessita: pong [13:12] ralsina: run-tests will not finish in my machine [13:12] nessita: weird [13:12] * ralsina checks [13:12] mandel, nessita already put the changes... it was just converting the string to unicode: https://bugs.launchpad.net/ubuntuone-client/+bug/805927 [13:12] Launchpad bug 805927 in ubuntuone-client "UnicodeEncodeError when importing windows' version of xdg (affects: 1) (heat: 6)" [Medium,Triaged] [13:12] ralsina: any clues? [13:13] nessita: what's the last thing it printed? [13:13] mandel: so, how long you estimate for us to have the meeting? [13:13] ralsina: all test run OK, then,. nothing else was printed, and the command never returned [13:14] nessita: if you ctrl+c/ [13:14] I meant what happens if you ctrl+c? [13:14] ralsina: "Terminate batch job? Y/N" [13:14] nessita: it happened to me once, I re-ran and it finished [13:15] ralsina: so, the script is running [13:15] let me check [13:15] alecu: about to approve your branch [13:15] and now it stuck again. Seems to happen "sometimes", although pretty often [13:16] nessita: for some reason u1trial never ends [13:17] ralsina: is it u1trial or the style chek? [13:17] ralsina: is never ending here [13:17] nessita: u1trial [13:18] nessita: run "C:\Python27\\python.exe" "C:\Python27\\Scripts\u1trial" ubuntuone_installer --gui --reactor=qt4 [13:18] yeah, doing that right now [13:19] ralsina: yes, same behvior here [13:19] Since all the tests end, I have no idea why u1trial does that [13:19] ralsina: did you ask dobey? [13:19] nessita: no, but let's :-) [13:19] dobey, I hav a problem! ^ [13:20] hrmm? [13:20] nessita: never noticed because I always ctrl+c the run-tests to avoid the style checks when testing :-( [13:21] dobey: for some reason u1trial is not stopping after all the tests are finished [13:21] ralsina: you should not avoid style check while developing!!!!!! [13:21] ralsina: on windows, or on linux? [13:21] nessita: style checks on windows show hundreds of useless lines (about Ui files) and push the tests out of the buffer [13:21] dobey: windows [13:22] weird [13:24] nessita: I forgot chipaca and I have the mgmt call in 6 minutes. [13:26] dobey, ralsina: I get the same when running "u1trial --reactor=qt" in windows. === zyga-food is now known as zyga [13:26] yay qt :( [13:27] ralsina: how long is that meeting? can you please talk about plans and account API? [13:32] ralsina: ping? ^ [13:33] alecu: your branch approved [13:33] ralsina: how can I IRL test yours? [13:34] ralsina: also, I don't think we should offer "Back" in that screen [13:34] nessita, thanks. I saw that a few minutes ago :-) [13:34] nessita: it's a one-line change nt to offer it, but why? [13:35] alecu: ah! re: the unicode stuff, we should talk about that during the mumble call, which will take place soon in the future (we need to align mandel's and bosses' agendas) [13:35] ralsina: because at that point, I think, it makes no sense? [13:35] re looking the screenshot [13:35] nessita: it just sends you back to the license page. Unless the user did something that can't be undone, there's not much reason not to offer the expected "Back" [13:36] ralsina: the screen says "UBUntu one installed!", what can teh user do if he goes back? [13:36] nessita: read the license again :-) [13:37] ralsina: but the message say "ubuntu one has installed" <- is done and can't be undone by going back [13:38] nessita: ok, convinced. Will push without a back button in 1' [13:38] ralsina: can you please add IRL testing isntructions? I will run it here [13:38] nessita: sure [13:39] ralsina: also, any answer to: "how long is that meeting? can you please talk about plans and account API?" (I want to make some mate if I have time) [13:39] nessita: 1 hour usually. I will ask about accounts [13:39] thanks, please let us know when you guys are free [13:40] hey hey all, its now got to the point where i have so much stuff in u1 i can't sync it all on to my /home partition for lack of space. any plans of following symlinks or something this cycle so i can use my slow HDD's to sync to rather than the fast SSD i use on my home partition? [13:41] nessita: pushed without "Back" and added IRL instructions [13:42] volunteers for bug #805899 ? [13:42] Launchpad bug 805899 in ubuntuone-client (Ubuntu Oneiric) (and 1 other project) "Can't sign in to U1 from Live session (affects: 1) (heat: 8)" [High,Confirmed] https://launchpad.net/bugs/805899 [13:42] thisfred, dobey? ^ [13:43] gord, I think there's no plan for symlinks during this cycle. [13:43] * thisfred looks [13:44] ralsina: dup [13:44] ralsina: hmm, looks like we're missing dependencies on the CD [13:44] nessita: I'm bacl [13:44] back [13:44] thisfred: eh? [13:44] dobey: can you mark it dup? [13:45] ralsina: already done [13:45] dobey: cool [13:45] dobey: thanks! [13:45] alecu, oh thats a shame, i guess there is no solution to my problem other than stop syncing things or get a larger SSD right? [13:45] nessita: ups, the irc client let me down... [13:45] dobey: well if the dbus service we need isn't running, I'm guessing something on the live CD is missing that we need (and that may not be explicitly listed in our dependencies) [13:46] gord: is it all in ~/Ubuntu One, or in other folders that you've synchronized as well? [13:46] dobey: oh, already known, nm [13:46] gord: because you can choose to not sync some of those other folders to that disk [13:46] dobey: I think gord wants to sync more, not less, [13:46] gord, perhaps mounting your hdd as /home (or /home/gord) and having symlinks to the sdd for the folders you want to be fast. But it sounds like a lot of trouble. [13:47] mandel: you had lunch already? [13:47] alecu: so, regarding the encoding, the problem is that pyinotify gets the paths as a sequence of bytes (str in my email, ups!) and sd has the logic to deal with them and use utf-8 against the server, on the windows side we get the paths as a encoded string in mbcs, so we have to make the pyinotify implementation to use bytes so that we need to change nothing on the other side [13:47] nessita: yes I have [13:47] mandel: ok. Bosses remembered they have the weekly management call, so we're waiting for them [13:47] nessita, do you know if we can have udfs for folders symlinked in /home from another filesystem? [13:48] alecu: you can't [13:48] alecu: symlinks get nothing [13:48] yeah i need to sync more but not on my /home partition, what are the requirements u1 needs to allow syncing? does it have to be in /home? certain filesystems? [13:48] alecu: afaik no links, sym or hard will ever be synced [13:48] alecu: nopes, and the reason is we don't allow UDFs to be symlinks [13:48] gord: has to be under /home [13:48] alecu: part of UDF validation is checking that the path is not a link [13:49] which I still disagree with, but there you go. I think dropbox shows it can be done in a non-confusing way [13:49] nessita, so there's no solution to gord's problem above? having a small fast ssd as /home, and no way to sync a bigger hdd? [13:49] gord: because there is no way for us to guarantee we can sync to the same place outside of /home on other systems. [13:49] precognition is not a feature of u1 :) [13:49] dobey: which is why we should sync them as real files [13:50] alecu: not at the moment, unless mounting the FS via some fancy mount method [13:50] thisfred: sync what? [13:50] thisfred, gord, I agree that we should sync symlinks too. [13:50] nessita, who may know what that fancy method is? [13:50] syncing symlinks is useless [13:51] dobey: sync symlinks as actual files they represent, just like dropbox does (I think). Unless they exist as symlinks on the target system [13:51] dobey, s/syncing symlinks/syncing stuff from other filesystems/ [13:51] thisfred: i see no actual value in it [13:51] dobey: I've used it and I do [13:52] yeah i don't actually want to sync symlinks, i just want to "mv ~/Music/ /storage/Music; ln -s /storage/Music ." and have u1 play nice with it [13:52] right [13:52] which is basically impossible to do [13:52] so u1 should follow symlinks [13:52] why? [13:52] well if the disk is not mounted you want to not delete it all I guess [13:53] in case /storage is a mountpoint and it gets unmounted, syncdaemon will be happily removing the files from online storage [13:53] been there, done that [13:53] alecu: let me look the bug report with some instructions [13:53] but you could maybe autounsubscribe folders on filesystems that are not mounted [13:53] because we don't necessarily have write access to same place on ohter systems [13:53] and what happens if it's an external disk, and it gets plugged into another system, that's also synced with u1 [13:54] dobey: If we treat it as a folder in the place it is mounted, it will just create a new folder on the other system [13:54] until we are providing a proper back-up service, handling anything outside home is not feasible [13:54] well you don't sync the symlink surely? in my case the music folder on system B would just get created and populated as a directory Music in home [13:54] unless you already have that folder there, and then it will sync into it [13:55] what gord says [13:55] fagan: ping [13:55] the problem is that pretty much anything we decide to do with symlinks/folders outside home, will never actually be what the user wants, as long as we are a sync service, and not a back-up service. [13:55] gord: This discussion has been had 3 years ago, though, so I don't hold out much hope for a policy change [13:56] thisfred, we should push for one, because gord is not the first user to ask for it. [13:56] at least a policy review. [13:56] dobey: it will be what the user wants: if the user knows what symlinks are, they are sophisticated enough to understand the solution we provide, which will have some corner cases, I admit [13:57] I have the exact same issue on my new desktop. [13:57] the only people really asking for it are hackers at canonical, and ubuntu/debian developers. :) [13:57] yes, but if we can help them without breaking it for the other users, why not? [13:57] thisfred: it will never be what the user wants, unless the user already knows what we're going to do and they consciously decide that is what they want, after the fact [13:57] alecu: can't find the bug report, but I think jdo was a commenter there [13:57] dobey: again, like it is with dropbox [13:57] thisfred: but going into it, it's almost never what they want [13:58] dobey: I agree you have to find out how it works, and then adapt to that [13:58] i don't know what dropbox does with symlinks, but i guarantee it's the wrong thing [13:58] dobey: but it makes things possible that are not possible now [13:58] if it doesn't Just Work (TM), then it's broken. [13:58] dobey: no it's not, it works just fune [13:58] fine [13:58] no it doesn't [13:58] fine [13:58] how do you sync a symlink to windows? [13:59] It does what I want it to do [13:59] yes, and you are not anywhere near a typical use case :) [13:59] dobey: As I keep saying, it's only a symlink on the original system [13:59] dobey: which is the problem people keep asking us to solve, nothing more [13:59] that is only the first time you've said that [13:59] no [13:59] i don't actually care about symlinks, its just there doesn't seem to be a solution to having a lot of data in ubuntu one and trying to use ubuntu one with small drives [13:59] yes :) [14:00] and it's not the problem people keep asking us to solve, unfortunately [14:00] but anyway, whatever [14:00] gord, so, as a temp solution, you may "mount /dev/hdd ~/mount"; then make udfs out of ~/mount/Music, etc, then symlinks in ~/Music to ~/mount/Music [14:00] there are much worse problems to deal with [14:00] dobey: read what gord is asking, that is exactly the problem he is having, he wants to sync stuff not in his home. He doesn't care about symlinks [14:01] gord, but be very careful with the thing that rye commented above, about not unmounting that while syncdaemon is running [14:01] i read what gord is asking. [14:01] alecu: still have to be very careful with that: unmount your external HD, and it might delete all your music [14:01] gord, because if it detects that those files were deleted, it will start deleting them on the cloud. [14:01] right [14:01] gord, it's not external, right? [14:02] alecu, huh and that should work? will have to have a play, and yeah its internal [14:03] alecu: I think if SD knew about symlinks, or was just plain able to sync stuff outside home, we could detect unmounted filesystems and unsubscribe them or mask them temporarily pretty easily [14:03] gord: if you're going to do the mount thing, make sure it gets mounted before you log in from gdm/lightdm [14:03] This is what my music player does for disks that aren't mounted [14:03] thisfred, absolutely. Sounds great. [14:04] the problem is that sd uses pyinotify directly [14:04] the problem is that "sync stuff outside of home" is not feasible. [14:05] Well I don't know much about pyinotify, but I very much doubt it sends 'deleted' events for files on disks that are unmounted [14:05] there is "back-up stuff outside home" which is a different thing [14:05] and we are not a back-up service yet [14:05] thisfred: what i mean, is that if it used gio, a lot more useful info would be available [14:05] It's totally feasible, as long as you let go of the idea that all UDFs have to live in the same place on each device [14:05] * nessita -> brbs [14:05] (plus it would be cross-platform already) [14:06] which assumption the mobile clients have already broken [14:06] thisfred: as long as you ignore the fact that we are a sync service, yes [14:06] dobey: well let's switch to that [14:06] the mobile client doesn't sync [14:06] dobey: we're syncing files and folders, not filesystems [14:06] dobey: yes it does, on a file by file basis [14:07] no, it browses and lets you download a file. it's the same as the web client [14:07] anyway, discussion interesting, but not going anywhere :) [14:07] so hell, just use the web client then; it's a solved problem [14:07] well in that case it should be made to [14:07] ∎ :) [14:08] i'm sure making 10000 folders and uploading 5000 files individually would make an interesting 6 years of one's live [14:08] life [14:09] I don't understand what you're saying there [14:09] If I explicitly download a file from U1 to my mobile client, I don't see why it shouldn't be kept in sync [14:09] i'm saying the argument that android breaks the sync assumption in some way is moot, because it's the same as the web client [14:10] well, that's great for a first iteration, but I'm sure we're more ambitious than that in the long run [14:10] no i think it does keep it in sync on android [14:11] but it's nothing at all like solving the problem of syncing external drives on a PC [14:11] they are very different problem spaces [14:11] well it's all sync, it has nothing to do with backup [14:11] it isn't all sync [14:11] anyway, this is going nowhere [14:12] dobey: well then you use confusing language since that's what you just called it [14:12] kiss kiss kiss [14:12] :P [14:12] :) [14:13] pistols at dawn? ;) [14:13] too early [14:13] true [14:13] drinking contest at midnight? [14:14] that works [14:14] hehehe [14:21] mandel: where are logs now, in windows, using latest xdg? [14:21] AppData\Local\xdg\cache\ [14:22] nessita: we are using the xdg_cache path for logs and partial files [14:22] mandel: ack, thanks. I m about to report 2 not-minor bugs in syncdaemon. One makes local rescan crash so syncdaemon does nothing else [14:23] nessita: ok, I'm close to be done with the UploadCMd tests so that facundo can give the green light and will move to those before I deal with the utf-8 issue [14:23] ack [14:24] nessita: that way we can talk about the udf8 issue at the meeting, but it is a no brainer :P [14:24] mandel: no brainer? [14:24] not sure what that means in this context [14:24] nessita: facil [14:25] mandel: hum, I doubt it :-) unicode and file path names are never easy [14:26] nessita: well, I know where the code has to go, and I should be able to understand it… and I have no brain :P [14:27] nessita: have you had any news regading the mumble? [14:28] mandel: as soon as the manager finish their call. ralsina, Chipaca: any ETA for our daily mumble? [14:28] sorry guys, it started a bit late, and we are all a bit late [14:29] mandel: I'm about to report paste.ubuntu.com/638446 [14:29] ralsina: you added IRL testing instructions to the commit message in https://code.launchpad.net/~ralsina/ubuntuone-windows-installer/fix_804600/+merge/66731 :-) [14:29] grr, power blip [14:30] nessita: can you do me a favour, can you get into mumble and let me know if you can hear me? [14:30] nessita: yikes, fixing [14:30] mandel: yes, let me report this bug first [14:30] nessita: sure [14:31] nessita: there [14:31] mandel, back to the unicode issue. Above you said "on the windows side we get the paths as a encoded string in mbcs" [14:31] alecu: yes [14:32] mandel, in python, it means either a secuence of bytes in "UTF-16/UCS-2" encoding, or either a unicode object. [14:32] mandel, what do you mean by mbcs? [14:32] mandel, so, how does pyinotify gets the paths on windows? [14:32] alecu: http://docs.python.org/howto/unicode.html#unicode-filenames [14:32] looking [14:33] ooook. [14:33] alecu: the paths are returned by a call to COM [14:33] alecu: does it make sense? [14:35] mandel, ok. But that COM call returns a type(unicode) object, right? [14:35] alecu: yes [14:36] alecu: the code works as following, we get the events from COM, get the to a Q so we do not loose any events, in a diff thread, read the Q and create pyinotify events as the ones you would get on Linux [14:36] mandel, ok, and internally syncdaemon handles everything as type(unicode). The conversion from "linux filesystem encoding" to type(unicode) is done on the linux platform specific stuff. [14:37] mandel, ahhhh ok. [14:37] mandel: bug #805981 [14:37] mandel, then that means that the "synthetic pyinotify events" you are creating should have "bytes" [14:37] Launchpad bug 805981 in ubuntuone-client "Windows: KeyError during local rescan, which makes syncdaemon do nothing else (affects: 1) (heat: 6)" [High,Triaged] https://launchpad.net/bugs/805981 [14:37] alecu: yes, exactly [14:37] mandel, should have "bytes in some encoding" as the filenames [14:38] mandel, that's why you'll convert the type(unicode) filenames as returned from COM into the "bytes" required by the pyinotify events using some encoding. Say "utf-8". [14:39] did I understand correctly? [14:39] Eww, found why Android music app doesn't load some album art. "GraphicsJNI E VM won't let us allocate 5120000 bytes" [14:39] "bitmap size exceeds VM budget" [14:39] * CardinalFang boggles. [14:39] CardinalFang: WTF is that image? A 1024x768 BMP/ :-) [14:39] alecu: yes, that is what we have to do [14:39] CardinalFang, nice! [14:39] CardinalFang, here's ten bucks. Get a gig'o'ram. [14:39] CardinalFang, that's because we don't resize the images on the server [14:40] alecu: in a phone? Good luck with that :-) [14:40] we send whatever is embeded [14:40] alecu, in my phone? Yes please. [14:40] alecu: but ofcourse we also have to make sure that the os_helper layer knows about this too [14:40] CardinalFang, in 3 months or so we should be able to resize images, but, you know, 3 months [14:40] beuno, I can do it at download time. [14:41] ralsina, as long as the image has to be uncompressed before displaying, packing it with jpeg2000 would not be useful either :-) [14:41] CardinalFang, that'd be awesemoner [14:42] mandel, why does the os_helper layer need to knows about this too? [14:43] * alecu does not know about the os_helper layer. [14:43] nessita: now? [14:43] grr, evil broken APIs [14:43] mandel, btw: please let me review those branches. [14:44] alecu: sure, I'll add you as a requested review [14:44] alecu: the os_layer wants to be writing to the fs in the correct encoding, that is mbcs which ever that is in your installation [14:44] not plain utf8 [14:44] I mean, not the utf8 we use in the server side [14:45] mandel, let's not care about server side. [14:45] mandel, server side we use type(unicode), we use utf8 for the network, and we use type(unicode) internally in syncdaemon as well. [14:45] Chipaca: yes! [14:46] alecu: god… [14:46] mandel, but for writing on windows from python we should be using the same type(unicode) objects we use inside syncdaemon. [14:46] sorry, was fetching coffee [14:46] alecu: are you sure about that? [14:46] mandel, we should not be converting that to the encoding specified by mcbs, because python should do that in the best way. [14:47] alecu: ha.. should do that, we need to confirm that assumption :P [14:47] mandel, as a general python knowledge, I am sure. But perhaps syncdaemon does something different I am missing. [14:47] Chipaca: coming to mumble? [14:49] nessita: am there [14:52] ugh [14:52] nessita: why is sso tests printing log messages to console in middle of tests? [14:53] dobey: got a trace? (we changed nothing related AFAIK) [14:53] test_delete ... ERROR:dbus.connection:Unable to set arguments ({'org.freedesktop.Secret.Collection.Label': dbus.String(u'sample_keyring', variant_level=1)}, '') according to signature u'a{sv}': : Fewer items found in D-Bus signature than in Python arguments [14:53] [OK] [14:54] i am getting that in trying to fix bug #805244 [14:54] Launchpad bug 805244 in ubuntu-sso-client (Ubuntu) (and 1 other project) "Method "Create Collection" with signature "a{sv}" on interface "org.freedesktop.Secret.Service" doesn't exist (affects: 2) (dups: 1) (heat: 16)" [Medium,Triaged] https://launchpad.net/bugs/805244 [14:54] and trying to maintain compat with older version [14:54] and i am trapping TypeError and not re-raising it, so not even sure why it would end up in a log message, since i'm not logging it [14:56] fagan: ping [15:21] manuel: lost you on mumble [15:21] dobey: we saw this behavior a lot in the past, when something "breaks" in the background somehow the debug mode gets enabled [15:26] makes no sense :( [15:27] soi'm surprised that people are actually hitting this bug [15:27] it implies there's no default keyring set up yet, which would be very odd [15:28] i guess maybe on a fresh install is possible, but still quite odd [15:28] especially if on wifi [15:32] nessita: https://code.launchpad.net/~dobey/ubuntu-sso-client/fix-805244/+merge/66918 [15:32] nessita: i am pretty sure that solves the bug, but the test thing is bugging me still :-/ [15:36] dobey: well the latest report was from someone running the live CD, there it's less surprising I gyess [15:37] eh, i just wish people would stop breaking API with no sane way to keep compat [15:44] alecu: ping [15:44] mandel, pong [15:45] alecu: got an encoding related branch for review, got time for it? [15:45] this is the one: https://code.launchpad.net/~mandel/ubuntuone-client/fix-805372/+merge/66887 [15:45] Chipaca: the lint issues need to be fixed in the branch from DiegoSarmentero you are trying to land. why is it not proposed/owned by DiegoSarmentero directly? [15:45] mandel, looking [15:46] thx [15:49] dobey: ack, sorry for the delay was on a call [15:49] * mandel walks dog [15:49] dobey: because his 'push' was failing. But he's got it working, and fixed the lint issues, and will be pushing it later today. [15:50] ok [15:56] why does computers show up several times in the u1 client? [15:57] nessita: no worries; i'm going to head off to lunch right now. thanks :) [15:57] bbiab [15:59] jo-erlend: http://askubuntu.com/q/19792/711 [16:00] * ralsina goes eat === ralsina is now known as ralsina_eating [16:00] Chipaca, that seems to indicate that the computer will be listed once per application [16:01] ? That can't be right? [16:01] jo-erlend: not sure what you mean === DiegoSarmentero is now known as DiegoSarmentero_ === DiegoSarmentero_ is now known as DiegoSarmentero [16:01] jo-erlend: what part indicates that? once per what application? what do you mean by 'application'? [16:02] Chipaca, not your answer, but the other one. Your answer doesn't really answer the question, I think. Why is a computer authorized several times? [16:02] jo-erlend: because you deleted the key from your keyring, for example [16:03] jo-erlend: several ways, but that's probably the most common [16:03] Chipaca, «If you have multiple device entries for Ubuntu One you can use the following steps to determine which devices to keep and which to delete. Be aware that it is not uncommon to have two entries. One for Ubuntu One and one for Tomboy.» [16:04] that's what I meant. That didn't sound right. [16:04] that is absolutely correct [16:05] jo-erlend: *in seahorse*, there'll be one key per application [16:06] ah! I misunderstood then. [16:06] jo-erlend: if you log in to software centre (to purchase things, for example), you'll have a key for it also [16:06] the ubuntuone-control-panel (or in older ubuntus, ubuntuone-preferences) will list only those that are for Ubuntu One [16:07] jo-erlend: server-side, on the sso site you'll also have one key per app, whereas on the u1 site you'll only have the u1 keys [16:09] right. Think those keys in gnome-keyring should be better labeled. [16:14] mandel: poing [16:14] mandel: ping [16:30] mandel: ping [16:33] ok, lunchtime [16:45] nessita: pong [16:45] * mandel back [16:46] mandel, ping [16:46] alecu: pong [16:46] mandel, I'm trying to understand the branch that fixes pyinotify.py [16:47] ok [16:47] mandel, it seems that the only place used for that output_format is in the __repr__ of the event, right? [16:47] mandel, that means that the pyinotify in the windows dir is handling file names as type(unicode) [16:49] alecu: yes, that is correct, which was giving an exception when we wanted to write the logs [16:49] mandel, ok. And btw: why do we include pyinotify inside our own code? [16:50] alecu: because that is a subset of pyinotify, is just the events, i has no code regarding how those events are gotten from the OS [16:50] alecu: so instead of implementing something similar, I grabbed the minimun I needed and used it [16:51] the license etc is ok, right? [16:51] mandel, I think not. If the owner signed a contributor agreement... blah blah blah. [16:52] alecu: really? mierda contributor agreement... [16:52] alecu: so wants the point of open source if we cannot use it, and ofcourse I cannot do easy_install pyinotify on windows... [16:52] mandel, also... I understood from what we talked before that we were going to handle all the filenames in the "windows synthetic pyinotify events" as byte. But this clearly shows that we are handling them as unicode, so I'm puzzled. [16:53] mandel, good point on the open source bit. [16:54] alecu: well, right now, the formatter brakes if the data is unicode, which only happens with path, but nevertheless it should ot brake [16:54] s/ot/not [16:55] alecu: so, we fix that in case we have unicode comming from somewhere, next we deal with the file paths so that they use byte and not unicode, which the formatter per encapsulation, should not care about [16:55] * mandel hopes he makes sense [16:55] mandel, oh, ok. that makes sense. [16:55] :D [16:56] mandel, I'm approving the branch, and we'll ask everybody on the contrib agreement after they return from lunch. Or tomorrow. [16:57] alecu: ok… but this contributor agreement is a PITA... [16:57] mandel, I know. === beuno is now known as beuno-lunch [17:01] mandel: any updates regarding the bug you filed to pywin32 upstream? [17:01] mandel: I'm running a non patched version and I'm having no issues at all [17:02] nessita: I was going to send the version of pywin32 I was using, but I need to use a clean vm.. will do later this evening [17:02] I don't want to loose time on the vm right now [17:03] mandel: ok, for now I removed those steps (the patching of the win32) from the instructions [17:04] dobey: approved [17:04] ok [17:04] ralsina_eating: you need more reviews? [17:04] mandel, alecu: you need reviews? (IRL @ win7 in particular) [17:04] if not, I'll start making the u1cp tests run in win7 [17:05] I'm fine, thanks. [17:05] nessita: so far I don't [17:07] nessita: I'm moving to your rescan issue [17:07] mandel: yey! [17:07] mandel: I had to remove the file in order to have a working syncdaemon [17:07] mandel: so you may need to create your own ugly?.png file to play with :-) [17:08] nessita: got plenty of those :) [17:08] mandel: then how come you did not hit this bug? [17:09] nessita: because I never did I rescan… I always got them and remove the files because I was testing other things... [17:09] nessita: you got further than I did in the IRL tests [17:09] I think I hit something else before that... [17:10] mandel: the rescan is just the LOCAL_RESCAN stage, which takes place before any other action (INIT -> LOCAL_RESCAN -> [LOCAL_RESCAN_DONE] -> everything else [17:10] so I'm puzzled [17:10] puzzled? why> [17:10] ??? [17:11] mandel: why you did not hit this [17:11] I got to the point where the files where written on the fs, then removed everything Ubuntu One/ xdg etc... [17:11] and never got to do a rescan with data in the hd [17:11] mandel: ah, I see4 [17:13] * alecu lunches === alecu is now known as alecu-lunch === alecu-lunch is now known as alecu === ralsina_eating is now known as ralsina === beuno-lunch is now known as beuno [18:15] nessita: I found the issue with the local_rescan, listdir has to be fixed to ensure that the paths are not returned with utf8 chars, got the test done will push as soon as I have fixed the method [18:18] mandel: on mumble, one sec [18:20] mandel, what do you mean "with utf8 chars"? [18:21] facundobatista: fear not ;) I mean that we swap ? for an interrobang so that the user can see his illegal path on windows, the pyinotify implementation we have takes care of that case and swaps it, but listdir was not doing so [18:22] facundobatista: so metadata ignores the fact that we have funny utf8 char, and we do not loose files so that the user can have them synced.. also the chars are very strange ones I'd would be very very strange if the create a file with that path on windows [18:23] mandel, we have files causing oopses having \n in them [18:23] mandel, i mean real newline [18:24] rye: really… people should take a test to be able to use a computer… why would they have a \n? [18:24] that most be an app doing something stupid... [18:30] well, I'm going to call it a day, catch you all tom [18:30] * mandel EOD [18:30] mandel: report pliz? [18:37] I'm back but will have to babysit for a little while. Will make up for it late tonight. Ping me because I am around! [18:38] hrmm [18:38] is there a nifty way to get the byte difference between two strings in python? [18:39] hrmm, or maybe that's not what i want [18:44] byte difference? As in "what bytes are different?" [18:45] yes, or if there are any different bytes [18:45] am having problem with launchpad api :( [18:46] alecu: different bytes? s1 != s2 :-) [18:46] and str() of an entry object in the launchpadlib api, gives the URL for the resource, but to my eyes they look the same [18:46] ralsina, what? [18:46] though python is saying 1 != 2 [18:47] dobey: using filter? [18:47] ralsina: since humanize is defined in ubuntuone/controlpanel/gui/__init__.py, the test should be in ubuntuone/controlpanel/gui/tests/ (and you put it inside qt tree). I'll move it [18:48] ralsina: no. i'm just doing if foo == bar: [18:48] dobey: apply a filter function over zip(s1,s2) and you should get the exact difference [18:48] and it's failing, when it shouldn't [18:49] it's quite weird [18:49] nessita: I put it there because the gtk tree doesn't use this humanize, but that's ok [18:49] because it's not failing for everyone :( [18:49] dobey: but you have a case that fails? [18:50] ralsina: yes, multiple it seems, and multiple that work. is very weird [18:51] dobey: if s1 and s2 are the allegedly alike strings: filter(lambda t: t if t[0] == t[1] else None, zip(s1,s2)) [untested like crazy] [18:51] dobey: sorry I meant you have a specific example copied somewhere that fails :-) [18:52] WTF: https://api.launchpad.net/1.0/~adorilson != https://api.launchpad.net/1.0/~adorilson [18:54] ralsina: did you talk to Claire re: sso spec? [18:55] nessita: pinged her, got no response [18:56] ralsina: could you please write her an email? [18:56] nessita: sure, with you an copy [18:56] thanks! [18:56] dobey, are those two copied and pasted? [18:58] ralsina: that line is a copy and paste of a direct print statement. i added quotes around them after that, and no visible character differences within quotes :( [18:58] dobey: could there have been a trailing space? [18:58] dobey: can you print repr of the strings? [18:59] WTF: "" != "" [18:59] is repr [19:00] dobey: you're using == for comparing objects that do not define __eq__? [19:00] that will compare id's of the objects [19:00] i don't know if they define __eq__ or not [19:00] seems like no? [19:01] then why would it work sometimes? [19:02] dobey print thingie.__eq__ and you will know [19:02] > [19:02] so it is defined [19:02] so wtf :) [19:02] dobey, therefore, it sometimes fails because __eq__ is buggy ;-) [19:18] ugh [19:18] the etag is different for some reason [19:21] dobey: could be different apaches answering? [19:22] i hope not [19:23] ralsina: and i'm not sure it matters as i think this is a value stored in the db [19:23] etag? In the db? weird. [19:25] well not sure [19:25] but the etag seems to have two hashes in it [19:25] and the first one is different, but the second is the same [19:25] it's quite odd [19:26] and this is a very annoying bug for us :( [19:28] etags are for cache invalidation, comparing objects based on their etags is kinky, at least. [19:31] sigh [19:51] and #launchpad-dev is totally dead [19:52] dobey, it's post-thunderdome sprint [19:52] I'd guess people are jetlagged, on holiday and tires [19:59] dobey: did you and ralsina find any solution to the issue where u1trial gets "stuck" when using qt4reactor in windows? [20:00] nessita: I have been looking at it [20:00] nessita: no, i'm not sure what to say about it. and i fixed the sso bug, and ran into this lp bug [20:00] nessita: without much success. It seems there is a thread leftover or something like that, based on a message I get when stopping it [20:01] it would be nice to know what is keeping it running exactly, and if the same thing happens on linux with qt4 [20:01] ralsina: QOject ::killTimers: bla? [20:01] nessita: right [20:01] dobey: it does not happen on linux, confirmed [20:01] ok [20:02] beuno: people should not do sprints in the thunderdome. Not ergnonomic at all. [20:02] nessita: and debugging stuff on windows is... semi-impossible? [20:03] ralsina: well, it makes room for new hires [20:03] ralsina: we can always move the style check to the top, before the tests are run [20:03] dobey: yes, but HHRR complains about the turnover, and maintenance about the blood.... not worth it. [20:03] nessita: agreed [20:03] That way It's "fixed" with a Ctrl+C [20:04] nessita: want me to add it to the branch you review? [20:04] ralsina: does your branch pass stule checks? [20:04] nessita: yes! [20:04] ralsina: then please! [20:05] ralsina: is something in the windows tests creating another qt main loop or thread or something? [20:05] and not properly disposing of it [20:05] dobey: in the control panel case, it's the exact same suite on both ends [20:06] dobey: could be [20:06] dobey: so not sure how that can be happening [20:07] nessita: pushed with the tests at the bottom [20:10] ralsina: that branch was already approved by me, you knew that, didn't you? :-) [20:10] nessita: for some reason some launchpad mails take foreeeeeeever to arrive [20:10] So no [20:11] ralsina: ah :-) [20:11] But hey, that means I can merge it ;-) [20:11] nessita: are you ok with 1-review for the installer? [20:12] ralsina: yes (given the shortage of reviewrs) [20:12] nessita: ok, +1ing myself and will stop asking for 2nd reviews [20:13] ralsina: heh, I mean I'm ok with this branch, but i think in general ew should aks 2 reviews [20:13] nessita: ok ;-) [20:13] let's take it case by case [20:22] nessita, alecu: review please (not urgent): https://code.launchpad.net/~ralsina/ubuntuone-windows-installer/fix_803929/+merge/66827 === Stanley is now known as Guest71626 [20:30] Hello what's exactly the problem "Our Notes, Contacts and Bookmarks database is not responding at the moment." [20:30] nessita, ralsina: I need your opinion on some matter... [20:30] alecu: sire [20:31] alecu: shoot [20:31] nessita, ralsina: in syncdaemon tool, I have a long number of methods. Long. [20:31] nessita, ralsina: I need to apply a new decorator to *most* of them [20:31] but not all? [20:31] *all* of syncdaemon tool public methods, except for __init__ and a few others. [20:31] alecu, nessita: and a tiny tiy branch for SSO ready for review: https://code.launchpad.net/~ralsina/ubuntu-sso-client/fix_805405/+merge/66784 [20:31] alecu: want to do it via instrospection? [20:32] (__init__ and _get_dict and wait_connected are the ones that do not need it) [20:32] ralsina, I was thinking metaclasses. Or something like that. [20:32] alecu: override __get_attribute__? [20:32] perhaps I can even rename wait_connected to some other thing... [20:32] alecu: well, it does bring python black magic into syncdaemon which is scary. I prefer nessita's idea [20:33] ralsina: needs tests, also, question: you know that error['__all__'] will not show all the errors? [20:33] nessita: how unintuitive :-) [20:34] ralsina: heh [20:34] ralsina: __all__ are those errores at form level [20:34] nessita: it's used on a specific case (bad answer to recaptcha) and there, the only field is __all__ [20:34] (just counted, it's 51 methods :P) [20:35] ralsina: not sure what you meant. The __all__ is sent by the SSO server and can have a variety of errors, not only bad captchas [20:35] nessita, so, __get_attribute__ sounds reasonable, but I wonder if it is a bit of black magic also. [20:35] perhaps adding the 51 decorators makes more sense. Or perhaps not, because if we add yet another method we'll have to remember to add the decorator. [20:35] nessita: then what should I display? There is nothing else when the problem is the recaptcha [20:36] alecu: yes to __get_attribute__ being a bit of black magic, but since the def will be in the same module, I prefer this solution to 51 decorators [20:37] I prefer __get_attriute__ to metaclasses [20:37] ralsina: you need to handle all the fields from that error dict. You can have specific errors for specific bits, such as 'email' and 'password' [20:37] ok, you guys have convinced me. Thanks! [20:37] ralsina: when you get a __all__ error, you *also* need to show an error at form level [20:37] guys and gals :-) [20:37] :-) [20:38] ralsina: not sure if what I say is clear... [20:38] nessita: I am getting *only* a __all__ saying "Bad recaptcha solution" [20:38] ralsina: well, that is your particular case. But the spec is different [20:38] Or I may be wrong, will test different errors [20:38] ralsina: the API says: [20:38] "we'll send a dict with errors back to you" [20:39] just like django validation does [20:39] (and, implementarion detail, right now is the same dict, AFAIK) [20:39] nessita: will try it and get back to you on the report then :-) [20:39] ralsina: so, despite your IRL tests can't display any other field, we need to handle the error dict according to the spec [20:40] ralsina: GTK version already does this, we may need to move that to a common location [20:40] nessita: I will look at that code, too [20:41] ralsina: see gtk/gui.py _build_general_error_message and [20:41] on_user_registration_error for example [20:41] nessita: but if it's like a django validation, the recaptcha error shouldnot be form level, so there is some other bug there. [20:42] ralsina: we handle error['email'], error['password'], error['message'] and error['__all__'] [20:42] ralsina: There may be a bug there, yes. In that case we need to handle in client and report to upstream [20:42] nessita: yep, will double-check and report [20:42] thanks [20:51] alright all. have a good evening [21:02] nessita: here's the current behaviour: If the captcha is wrong, only the "wrong captcha" message is given in __all__, even if other fields are wrong too. Then again, I suspect that is intentional. [21:03] ralsina: not sure what your conclusion is then [21:04] (I still think we need to properly handle the error dict, since this can change server side and our client will blow up) [21:04] That I had to reimplement the method, and it will be done in 2' ;-) [21:04] OTOH, the GTK code doesn't seem to handle __all__ in on_user_registration_error and the docstring for that method is just wrong. [21:05] Ok, it does handle __all__ sorry. The docstring is still wrong. [21:17] ralsina: why? [21:17] (may be outdated) [21:17] Because it says """""Captcha image generation failed.""" [21:18] """Concatenate __all__ and message from the errordict.""" <- seems correct... am I too sick? (this running nose is killing me) [21:18] And that's the docstring for another method [21:18] ah, that's another line, let me find it [21:18] what method? [21:18] on_user_registration_error ? [21:18] on_user_registration_error, right [21:19] ralsina: definitely bad copy and paste [21:19] I can add that change to this branch [21:19] please [21:23] nessita: care to take another look? I'm afraid I crossed your "needs tests" threshold on https://code.launchpad.net/~ralsina/ubuntu-sso-client/fix_805405/+merge/66784 [21:24] ralsina: from my POV, that change needs tests [21:24] jajajaja [21:24] * ralsina goes write a test [21:25] I hate writing tests in sso-client with the replay mocker [21:25] ralsina: I hate that too. I hate that record-and-play mocker lib [21:25] I never use it [21:26] Me neither. Last night I wasted 1 hour trying to find the right incantation to make the number of calls match [21:29] can I have a couple of reviews for https://code.launchpad.net/~nataliabidart/ubuntuone-control-panel/run-tests-windows/+merge/66972 ? [21:29] nessita: sure [21:29] and, test for u1cp now run in windows! [21:29] yay! === m_conley is now known as m_conley_away [21:30] nessita: I have no idea how to test that branch. Should I fake critical() and write tests that send different kind of error dicts? [21:30] I meant my branch [21:30] looking [21:31] ralsina: yes, and you should call on_user_registration_error with different error dicts [21:31] nessita, all tests pass on windows???? [21:31] nessita, congrats! [21:31] nessita: ok, will do that [21:31] NOOOOO I HATE YOU BZR ON WINDOWS! [21:31] bzr: ERROR: Failed to rename C:/Users/ROBERTO/canonical/run-tests-windows/.bzr/c [21:31] heckout/limbo/new-7 to C:/Users/ROBERTO/canonical/run-tests-windows/data: [Error [21:31] 5] Access is denied [21:32] That means I have to whack my whole ~/canonical where all my branches are [21:34] dog + walk = later [21:35] alecu: yes! the tweaks needed were very little :-) [21:35] nessita: indeed they pass! cool! [21:36] And changes are minimal, so +1 from me [21:38] thanks [21:38] I'm gonna to bed now, I'm, not feeling particularly well [21:39] (the running nose is morphing into a not minor cold) [21:39] see ya later! [21:59] alecu: still here? [21:59] still [22:00] I don't understand the mocker we use in ubuntu-sso-client :-( [22:00] alecu, I save a file, SD uploads that file, the bubble notifies about a file uploaded three days ago [22:01] facundobatista, juas! Please open a bug, let thisfred know about it. [22:01] alecu: if you do understand that mocker, could you tell me how to write a single test for on_user_registration_error from this branch? https://code.launchpad.net/~ralsina/ubuntu-sso-client/fix_805405/+merge/66784 [22:01] ralsina, my recommendation is "don't use that mocker for new tests" [22:02] alecu: it's an idea. But I don't have patch, and the test case makes me use replay and so on. [22:03] And self.view.message_box is already mocked [22:03] And I want to put that file on fire :-) [22:03] ralsina, then create a different TestCase class, make it inherit from twisted.trial.unittest.TestCase and you'll have patch... [22:03] ok, I will go that way [22:03] ralsina, but if you really want to use mocker.... [22:03] ralsina, let's see. [22:04] Honestly, I don't but crating a whole testcase different from all the others just to avoid the mocker... fun, but it smells [22:05] ralsina, that's the issue with mocker. Every time we have to go and fix a test, or create a new one, it's a long time getting again in the mocker mindset. And the weird error messages and... [22:05] facundobatista: that is very strange, please assign the bug to me and tell me the number [22:06] ralsina, so, let's say you make it with mocker. What do you want to assert in the test? [22:06] ralsina, do you have a specific question or just need a pointer into how to use mocker? [22:07] alecu: a simple test where I call on_user_registration_error with error={} would be a good start [22:07] facundobatista, does the "file uploaded three days ago" shows on the latest logs? [22:08] alecu: I also suspect we are not testing much beyond "things are called in the right order" in lots of tests [22:11] ralsina, you said "self.view.message_box is already mocked". Where is that? [22:11] alecu: SetUpAccountControllerTestCase.setUp === m_conley_away is now known as m_conley [22:15] ralsina, well, that's just an instance of the mocker class, it's not mocked the way we mock it when we are not using mocker. [22:16] alecu: ooooooooooook ;-) [22:16] ralsina, so, that object that's already mocked is no big deal. [22:16] (I know that phrase above sounded very confusing :-) ) [22:16] no kidding ;-) [22:16] ralsina, so my recommendation stands: no more mocker :-) [22:17] ralsina, do you still want me to build an example with it? [22:17] alecu: ok, I will do a non-mocker testcase. Hope nessita doesn't hate it. [22:18] ralsina, nessita hates mocker. I started using it, and she made me see how awful it was. [22:18] alecu: ha [22:18] alecu: ok, should be no problem then [22:18] but now, I EOD [22:19] bye! [22:20] thisfred, alecu, ok!