=== hallyn is now known as hallyn_afk [03:59] hi [04:04] why would xchat close itself? [07:49] mornings [08:24] morning all! [08:39] morning mandel [08:39] ralsina: hello hello! [08:40] ralsina: guess what, yesterday walking the dog the owner of the chiguagua came to talk with me, the little fellow is dead and started crying on my shoulder [08:40] I'm considering to write a short movie about this [08:40] mandel: :-( [08:40] dark comedy seems to match the full week perfectly [08:41] well, at least she didn't stab you or anything === rye is now known as rtg [08:44] ralsina: true, did I tell you I believe she is a 'mujer de dudosa forma de vivir' ? [08:44] yes you did [08:44] thus my concern about chihuahua-avenging-stabbiness [08:45] ralsina: she was there crying I was wondering, should I hold her hand or something? Will she charge me for that? Do the people around know that? [08:45] groan [08:45] regardless, dead puppy, crying woman.... hand holding is appropiate, I assume [08:46] I suppose [08:47] I don't knoqw, Ill get back to code, at least windows is behaving now :) [08:48] cool, me too [09:17] morning im still doing twisted stuff today ping me mandel if you have a merge [09:43] ralsina: one question about the design of the windows port, we got a IOCPReactor that can use namedpipe, unfortunatly there is not integration between the IOCP twisted loop and Qt main loop. [09:44] ralsina: with a SelectThreadedReactor that can be easily done, but makes the UI and the twisted main lopp be the same, which migh make the UI less responsive [09:45] how about using two diff processes, communicate via some easy IPC supported in pythons level between the UI process and the sd and sso twisted main loops? [09:45] mandel: scary? [09:46] how unresponsive are we talking about here? [09:46] ralsina: sso wont be a problem, sd might be an issue because the reactor is taking care os file uploads, downloads etc… + the UI [09:46] where the UI of sd is the control panel [09:47] but the reactor is not really doing uploads, it's just telling sd to do them? Or am I confused [09:47] ralsina: the reactor is taking care of dispatching the async ops [09:48] so if we have lots of them, it will slow the UI [09:48] mandel: yes, but dispatching should take milliseconds [09:48] I had the same conversation at #twisted, the consense there was, if you do lots of operation you will have issues with the UI [09:49] ;-( [09:49] ok then, if we must we must [09:50] are you thinking a simple queue or something like real RPC [09:50] because if it's a queue we can just use multiprocessing [09:52] Hello people! :D [09:55] ralsina: I was considering a q or something of the like [09:55] ralsina: would simplify the solution, adding extra complexity in the reactor sounds like a dangerous thing to do [09:55] agreed [09:55] let's keep it as simple as possible [09:55] anyway, I think we can do it after we have it working [09:59] ralsina: well, sd is working atm, I can change the reactor to the IOCPReactor and use named pipes for the 'dbus' like calls, lqter we make the control panel two processes, one uses the reactor for the calls, and they communicate via a q to a second process that has a very easy UI [09:59] but then we are back to the question of, is this overeng? [10:01] I'm split with the multiprocess idea or attempting to make the IOCPReactor smarter… [10:02] I suspect the multiprocess is "simpler" n the sense that there is nothing hard about it, but it does make the whole thing more complicated. [10:03] OtOH, the IOCPReactor changes would be local, but are probably two orders of magnitude harder to get right [10:03] ralsina: exactly, I'm going to talk with our windows expert to see which is the best option [10:03] cool [10:03] he might say something like, 'thats piss easy' [10:03] maybe he has a great idea on how to do it [10:03] ralsina: other option is to move from a IOCPReactor to a SelectThreaded one [10:04] mandel: that's beyond my depth in twisted [10:04] and do there the named pipe operations, if that is the case I can add the Qt inegration in like 20 min [10:04] ralsina: I'm just typing random words with the word reactor attach to them to sound smart ;) [10:05] mandel: if it's 20 minutes, see you in half an hour! ;-) [10:06] ralsina: I already have the integration with the threaded reactor done, I need the named pipes :D [10:12] hehehe, cool :-) === rtg is now known as rye [11:03] * fagan break (going to buy la noire) [11:37] morning all [11:43] Damn la noire is out friday in Ireland, american people on youtube showing off that you got it first arent cool [11:46] fagan: is that a playstation exclusive or are they also releasing it for xbox? [11:46] duanedesign: rockstar do everything on both [11:46] they cant afford not to [11:48] * duanedesign nods [11:48] duanedesign: I have an xbox so I hate that I missed out on heavy rain [11:48] got to play a little bit of it and loved it [11:49] duanedesign: I just bought brink too [11:49] its very cool === teknico is now known as teknico_away [11:58] i have not gotten to play heavy rain yet [11:59] duanedesign: its really interesting you can lose missions and that doesnt make you go back and have to do it again you just continue and it changes how the game turns out [12:00] ralsina: ping [12:38] fagan: pong [12:39] ralsina: just wondering am I going to the sprint too [12:39] fagan: good question. I didn't do the list, Chipaca did [12:39] I got the email about me not having my travel details and it reminded me too [12:39] and he's away today [12:39] ralsina: ah its cool we can sort it when he is around tomorrow [12:40] I suppose that if you are in marianna's list you are attending, but yes, check with chipaca tomorrow early [12:40] sure thats cool [12:40] ok, back to the salt mines. [12:55] idea for Ubuntu One: Add a shell account! [13:01] oops, lost link, did anyone get my post? [13:03] hello === teknico_away is now known as teknico [13:18] hey guys, just wondering - are there any special requirements ubuntu one needs of a filesystem? basically i want to sync directly over to my samba share with a symlink and samba can't do much special filesystem stuff - all my music is too much for this small ssd drive [13:51] gord: basically, ubuntu one doesn't follow symlinks [13:52] ralsina, hrm.. so there is no solution to the problem that arises when you have a large amount of data to sync, but a small home directory? [13:52] you could sync on the samba server [13:53] explanation: create a UDF (ie: a folder you sync that's outside ~/Ubuntu One) [13:53] gord: you can set the samba share/music as a syncable directory [13:53] sync that folder to the samba server and not to the SSD [13:53] I am not 100% sure udfs can be defined on mountpoints [13:54] whenever i have tried to do that in the past (and currently happens when i try now) all the ubuntu one options are greyed out in nautilus [13:55] gartral, UDFs cannot be defined on mountpoints, outside of $HOME and attempting to do that with bind mounts or mounting the folder within home can lead to sudden data removal from the server if syncdaemon notices the files are not there, i.e. folder is not yet mounted [13:55] hm [13:56] gord, so no udfs on mountpoints, I'd suggest syncing a UDF on the samba server instead then. [13:56] i definitely need to come up with a simple one-file u1 uploader/downloader that can be used on the server [13:56] of course if it's a headless server doing that may be tricky [13:56] it is :) [13:57] gord: someday you will be able to just use streaming to the desktop ;-) [13:57] gord, u1 now has a REST HTTP file api that can make it extremely easy to do, actually [13:57] ralsina, like with Chipaca's u1-music? [13:57] rye: well, for downloads, but not for syncing [13:57] ralsina, aaaaand uploads! :) [13:57] rye: right ;-) [13:58] gord: bad news are, that costs you a few bucks a month [13:58] canonical employee ;) [13:58] hello [13:58] ralsina, gord works on the unity team [13:58] gord: good news are, it's free for you then ;-) === m_conley_away is now known as m_conley [13:59] * ralsina will never remember to check directory first [13:59] gord, you may try u1-music to see how it can work, but we need to poke Chipaca so that he fixes it to enable it working alongside of ubuntuone installation [13:59] ralsina, we need either to poke Chipaca or fix the package :) [14:00] poking chipaca will probably not work for a few days [14:00] or at least not achieve much, he's a busy bee :-) [14:00] me [14:01] or we don't have a stand up [14:01] bug #776386 - ping joshuahoover - do you recall where you have posted the instructions? [14:01] Launchpad bug 776386 in ubuntuone-client "File Sync error (IOError with tritcask metadata) (affects: 5) (heat: 24)" [High,Triaged] https://launchpad.net/bugs/776386 [14:01] me [14:01] alecu? [14:01] me [14:01] (writting notes) [14:01] fagan, standup if you are around [14:02] oh didnt think we had one [14:02] me [14:02] hrm, interesting - i'll have a poke later and see if i can come up with a solution that doesn't require re-uploading everything - streaming sounds perfect though :) [14:02] * fagan writing notes [14:02] mandel, go ahead [14:03] gord, we have magic uploads, so if you've uploaded files before, it won't upload them again [14:03] DONE: Work on the idea of the IOCPReactor. Looked at the integration with Qt [14:04] TODO: keep looking [14:04] BLOCKED: no [14:04] DONE: mgmt call, hacked in designer a lot (installer wireframes), poked everyone in the right order (hopefully) to get everything signed off, misc stuff. Probably achieved my first food poisoning in this trip. [14:04] TODO: finish the damn wireframes, receive the final signoffs on dashboard & installer, start passing things around, see doctor [14:05] BLOCKED: see food poisoning above. [14:05] alecu? [14:05] DONE: started a control panel branch to use qt + twisted [14:05] TODO: work on that. Mumble with mandel. [14:05] BLOCKED: no [14:06] FORGOT: have two candidates for the "windows guy" position already [14:06] cool! [14:06] alecu: sadly it's not going to happen today, as we would really need ;-) [14:06] λ DONE: nightlies fixes, UDS [14:06] λ TODO: 11.10 nightlies [14:06] λ BLCK: Still recovering from cold. [14:06] now me then [14:07] DONE [14:07] * got the basics of twisted figured out [14:07] TODO [14:07] * do some coding to learn it proper [14:07] Blocked [14:07] * nope [14:07] fagan, what are you reading for twisted? [14:07] alecu: just learning the basics of how it works and how u1 is using it [14:08] alecu: I didnt know much about it before other than it does networking mumbojumbo [14:08] fagan, right. Do you plan to use any book or tutorial? [14:08] ok, alecu, mandel, nessita will be around later today [14:08] please ask her for any assistance you may need and she may be able to provide you [14:08] alecu: I am using the tutorial but ordering the book later today [14:09] I will stop to go see the doctor, will try to be back ASAP [14:09] ralsina, ack [14:09] ralsina: double ack [14:09] Now, if you will excuse me, I have to look up the turkish conjugation of "to vomit" [14:11] sigh, rain. [14:12] Hello U1 team....I'm having a bit of an issue with my file sync...anyone available to take a look? [14:12] dobey: I love the rain :) [14:12] cgregan: im sure if you said your issue someone will take a look when they have the chance [14:12] so work away :) [14:14] So I got my hardware refresh on monday and signed into my account...I have my Documents folder synced and now it appears that although U1 is saying it only has 2 files left to sync (after 3 days BTW) half of my files are still missing. [14:15] and the web interface crashes when I attempt to use it to see what is up there [14:15] cgregan: wow sounds bad [14:16] yeah fagan.....luckily I have my old drive so I can manually pull the files from it, but not quite the point [14:16] hoping there is merely an issue with something on the server side [14:17] cgregan: could you pastebin the syncdaemon log from .cache/ubuntuone/log/syncdaemon.log [14:17] fagan: sure [14:18] that should give more info on whats going on with the sync from the computer [14:18] the web interface should be cought by the oops thingy we have built in [14:18] fagan: it was [14:19] fagan: https://pastebin.canonical.com/47574/ [14:21] cgregan: well im not the greatest with the syncdaemon logs so ill let someone else answer it but there is a gateway timeout down at the bottom that may be it [14:22] fagan: I saw that...seemed bad [14:22] cgregan, re: web interface crash, could you please open the source of the page and look for "OOPS" word and copy the whole identifier in the comment and paste here? [14:22] fagan: it is currently trying to sync a rather large file, and maybe that it causing some trouble [14:22] rye: sure [14:22] gw timeout? [14:23] rye: yeah [14:23] file size should have no bearing on the web ui "crashing"; number of files might though [14:24] dobey, unless the shard the account on is under a heavy load and it crashes no matter what [14:25] rye: hmm...so just was able to get into the web UI [14:25] and it actually looks like the problem was when the files synced from my old machine [14:25] some directories have no content [14:25] so of course then no content is on my new machine [14:26] cgregan, some directories on the web ui don't have content, do you have the access to the old machine? [14:26] So although it has taken 3 days for the files to sync down (with 2 remaining) it is doing what it thinks is right [14:26] rye: I have access to the drive from the old machine [14:26] not the whole machine [14:27] rye: but that has nothing to do with file size :) [14:27] I will double check the original content and get back to you with more details [14:27] rye: that's just "always breaks, for everything" [14:27] dobey, true [14:27] cgregan, could you please also grab the ~/.cache/ubuntuone/log directory from the old machine drive so that we can see whether it believed it has synced everything [14:28] rye: certainly [14:28] rye: hehe thats what I thought too but I only asked for the syncdaemon log :/ [14:29] I didnt believe that the notifications were telling the truth [14:38] rye: i think he's referring to removing ubuntuone which we have in this faq: https://wiki.ubuntu.com/UbuntuOne/FAQ/HowDoICompletelyRemoveAndReinstallUbuntuOne [14:41] mandel: ping [14:41] Chipaca: on the phone one sec [14:42] oh Chipaca you are around 1 quick question am I on the list for the team sprint ralsina didnt know [14:43] fagan: there is no list. When there is, I'll let you know either way. [14:44] Chipaca: cool [14:44] rye: so confirmed...a ton of docs not synced from the old Documents folder. I have not copied them directly and OSDNotify tells me that over 4000 files are now being uploaded to U1 [14:44] I also got the log dir [14:44] s/not/now [14:44] cgregan, could you please archive the log dir and send it to ubuntuone-support@canonical.com ? [14:45] rye: sure [14:46] joshuahoover, btw, i talked to launchpad team and flacoste said they have notified IS about the spam [14:46] rye: good :) [14:47] Chipaca: I'm back sorry [14:47] Chipaca: personla stuff with the ex-suegra etc... [14:47] mandel: in Ubuntu you can't sync folders outside of your home. What's the restriction on Windows? [14:47] mandel: uncomfortable essential conversations, i guess [14:47] Chipaca: same thing, we did a map of the xdg_home dir to C:\Users\username [14:49] is there a way to refer to "any folder under taht" that is comprehensible to windows users? [14:49] Chipaca: let me find a msdn doc that uses the term [14:50] 'cause I call it home folder [14:53] i think they call it user profile folder or something like that [14:54] Chipaca: dobey is right, the call it profile folder [14:54] <- smrt [14:55] brb [14:57] ok [14:57] shen me niao! [14:57] where did my Alt+F2 go?! [14:58] anyone else using classic on up-to-date 11.04? [14:59] yep im on classic cos i HATE unity ... [14:59] dobey: using unity in oneiric ... and settings daemon seems to be busted [14:59] grr [14:59] so gtk2 is all "YAY ITS 1992" [14:59] stupid keyboard [14:59] dobey: I love unity [14:59] :) [14:59] fagan: good for you [15:00] that's quite an accomplishment for an irishman [15:00] * Chipaca runs [15:00] lol [15:00] dobey: try the same shortcut as used in unity, maybe the got that screwed up [15:00] don't make me laugh, my chest hurts :( [15:00] dobey: ow, sorry :-/ [15:00] mandel: no no. my stupid keyboard [15:00] Chipaca: lol we love unity in ireland too north and south :) [15:00] fagan: i hate unity .... why why why did they mess up a perfect desktop with unity ... ARGH!!!!! i used unity for a couple of days and got pissed off with it... switched back to classic [15:00] my keyboard apparently ended up having the Fn lock turned off [15:01] * dobey glares in the general direction of Redmond [15:01] Neoti: I love it mainly because it suits how I like to work [15:01] dobey: stop looking at me! [15:01] Neoti: I use a lot of hotkeys in games and stuff so now I can apply that to my entire desktop [15:02] uhm, ok, yall can stop arguing about unity and totally misusing the term "desktop" now [15:02] although I really hate some stuff I love the multitasking and how good the search is [15:02] lol games what games ... ubuntu and games lol ... im only joking .... but they need to get Magic the gathering working on ubuntu .... then my move will be complete [15:02] Neoti: heroes of newerth [15:02] Chipaca: yeah, stupid cold :( [15:03] dobey: does it also hurt if i TICKLE YOU!!! [15:03] Neoti: if you can get a higher action per minute game other than starcraft ill give you a picture of a green dog [15:03] Chipaca: physically, and emotionally, yes [15:04] Starcraft is cool... so im not going to even bother trying ....:) [15:04] Neoti: well heroes of newerth and savage 2 are my fav games on linux [15:04] savage 2 is free too so its even better :D [15:05] yall are weird [15:06] dobey: if you havent met me yet then you are only getting the half of my weirdness [15:06] i met you, yes [15:06] thankfully only briefly :) [15:07] dobey: haha [15:07] * fagan doesnt remember a few of the names of the people he met at lucid UDS [15:08] I did meet as many people as I could so that was fun [15:09] fagan: i use to like playing recoil... but can not get wine to work properly with it... [15:09] Neoti: I never use wine for games [15:09] fagen: what do you use then ? [15:10] fagan: what do you use then ? [15:11] Neoti: native games [15:11] like what other games do you need than world of goo and heroes of newerth they both cover all of te major bases [15:12] Magic the gathering, Starcraft, Recoil [15:12] command and conquer [15:12] House of the dead III [15:12] alecu: ping [15:12] Starcraft I can agree with [15:13] but the others aren't my thing [15:17] there are native engines for C&C and Starcraft, but I don't know how well they work any more, or how complete they are [15:22] there is an interesting one in lua called spring [15:22] but all the games I saw that use it look really ugly [15:23] maybe thats me being shallow about graphics though [15:24] lua is a cool language [15:24] not really [15:26] dobey: well no language that I have seen looks more like my psudocode than ada :) [15:26] I did a little bit of lua and liked it but the bindings arent there and adobe is the only one "maybe" using it [15:27] uh [15:27] lots of people are using lua [15:27] most notably, blizzard. [15:27] dobey: I havent seen much of it around [15:27] and most decent tams that develop games [15:27] great scripting lang [15:27] I didnt know blizzard use it [15:27] well, proprietary code isn't for you to see [15:28] true [15:30] Chipaca: more nightiles in O coming [15:31] dobey: O my [15:31] cgregan, and now I know why weren't they syncing. [15:32] cgregan, bug #683116 ... grr I really need to come up with something here [15:32] Launchpad bug 683116 in ubuntuone-client (Ubuntu) (and 1 other project) "Syncdaemon silently crashes at startup (affects: 9) (dups: 2) (heat: 146)" [High,Triaged] https://launchpad.net/bugs/683116 [15:33] rye: thanks for the info [15:33] rye: keep me in the loop [15:33] I will subscribe [15:35] man, budapest killed me [15:36] dobey: ubuflu [15:37] all the ubuntu developers carry their local flus with them [15:37] it is not ubuflu [15:37] got a problem, syncing worked fine for quite some time, but now it wont, i dont know why :S [15:37] uploading over the webinterface works, though. [15:41] strange crap... [15:41] working now again :S [16:02] yay, time to get some lunch [16:02] bbiab [16:10] alecu: alecu_: ping [16:11] mandel, pong [16:11] mumble? === hallyn_afk is now known as hallyn [16:55] EOD === beuno is now known as beuno-lunch === teknico is now known as teknico_away === m_conley is now known as m_conley-lunch [17:05] mandel: ping [17:21] ah bother [17:33] hello! [17:34] hola [17:35] fagan: early? [17:50] ralsina: hey there, how is it going? [18:02] alecu: ping === beuno-lunch is now known as beuno === hallyn is now known as hallyn_afk [18:29] nessita: can i get a review on https://code.launchpad.net/~dobey/ubuntuone-client/allow-gtk3-compile/+merge/61448 ? :) [18:29] dobey: sure1 [18:29] ! [18:33] karni, CardinalFang, DanRabbit here can help us with the artwork for the Android Files App [18:33] could you let him know what we need? [18:33] whoa nice [18:33] I know there's the startup image, the icon's orange is different to music [18:33] what else? [18:33] DanRabbit, do you have an android device? === m_conley-lunch is now known as m_conley [18:33] beuno: I do not sorry [18:34] it's ok [18:34] beuno: DanRabbit: I think we could use a properly sized icons (resized look really bad), they're too small in the UI [18:34] okay, we're talking icons for Android right? [18:34] for 2.3 inclusive, yes [18:35] and maybe a monochrome version as well? [18:35] beuno: is the logo on the action bar fine? (I took it somewhere from design resources) [18:35] okay, I had supplied several sizes previously do you know which ones are missing? [18:35] slow down guys, slow down [18:35] I meant the *file* icons in the UI [18:35] file, folder, published file, etc [18:35] unless we can find those already on the net with proper license with CardinalFang [18:36] ah [18:36] we were missing the icon (we have it), missing splash screen (landscape and portait), better _file_ icons on the list [18:36] karni: if I could get a list of icons needed, the sizes you need them at, and bonus points for screenshots of the current ones, I can do that :) [18:36] DanRabbit: sure thing! [18:37] DanRabbit: I'll give you the sizes of splash screens as well [18:37] DanRabbit: I'll get back to you shortly, is that ok? [18:38] karni: that would be great :) [18:38] DanRabbit: thanks. want to quickly triage one problem. [18:38] karni: you could also email me: daniel.p.fore@gmail.com if you'd like [18:38] DanRabbit: perfect, saving your mail [18:38] thanks [18:39] dobey: approved [18:39] nessita: thanks [18:39] anyone else want to give a quick second review? [18:39] https://code.launchpad.net/~dobey/ubuntuone-client/allow-gtk3-compile/+merge/61448 [18:55] nessita: oh, btw; your cp branch supposedly fixes a bug, but it wasn't committed with a --fixes= ? [18:55] dobey: the orange branch? let me confirm [18:56] yes [18:56] nessita: the comment you removed has a bug # in it [18:56] dobey: true, sorry for forgetting that :-( [18:56] nessita: doh, of course, it just merged a few minutes ago :) [18:56] * nessita updates the bug status [18:58] DanRabbit: under little fire here, sorry [18:59] karni: no pressure :) === DanRabbit_ is now known as DanRabbit [19:29] nessita, beuno: ended up needing larger changes, so i resubmitted; could you please re-review? https://code.launchpad.net/~dobey/ubuntuone-client/allow-gtk3-compile/+merge/61455 [19:30] dobey: ack [19:32] dobey, still +1 [19:43] nessita: did you mean to poke here about that? :) [19:44] dobey: yes, sorry! too many opened channel and the same prefix :-D [19:46] nessita: anyway, the test-gsd-ubuntuone.c didn't actually do any testing itself. it was to use for testing the display of the quota dialog, which isn't there any more. it was used by running it, and then running the accompanying python script to send the signal, while it was running [19:46] nessita, pong [19:46] dobey: right, I now remember [19:46] alecu: hi there! how is it going? [19:47] dobey: approving [19:47] nessita, fine fine [19:47] alecu: I'm one test distance of proposing a branch to have multiplaftorm-ready code in u1cp [19:47] alecu: what are you working on? [19:48] nessita, using qt+twisted in the control panel [19:49] nessita, mandel told me that no designs have been approved yet for the control panel ui. [19:49] alecu: what does that mean? (other than being scary :-P) [19:49] what?!?!?! [19:49] * nessita passes out [19:49] nessita, there have been a few designs approved for the "installer" wizard [19:49] nessita, but no cp yet. [19:50] alecu: is mandel working on that? (installer wizard) [19:50] nessita, afaik ralsina is working on that [19:50] ah [19:50] ok, I guess today is too late to bitch about not having approved designs :-) [19:50] nessita, mandel is working on making the iocp reactor for windows play nicely with qt [19:50] tomorrow it is [19:51] sigh [19:51] alecu: ok, I'll be finishing this (separating linux-specific into linux modules), and then I'll be moving the little logic we have in the gui module to the backend [19:51] nessita, we've been discussing with mandel if it would be fine to keep the separation between the u1cp gui and backend, and we came to the conclusion that keeping that separation is not useful for the windows port. [19:51] alecu: I wanted to be sure we're not overlapping in our work [19:52] alecu: wait [19:52] alecu: we talked about this in Budapest. We said we will not have 2 processes, not that we'll mix logic and view [19:52] nessita, the separation you are doing is ok, though. [19:52] nessita, right [19:52] alecu: is that what you mean? [19:52] nessita, yes. [19:52] ah, great. I almost pass out twice :-P [19:53] nessita, but we (mandel and I) discussed having two processes again, to work around the iocp+qt issues [19:53] nessita, but we would need iocp in the gui layer as well [19:53] nessita, so one process is better. [19:53] nessita, anyway: please let me know when you need reviews for that. [19:54] i am scared [19:54] alecu: yessir [19:54] dobey: I'm scared of asking why [20:02] wonder if i should upgrade yet [20:43] hi, My u1 folder appears to be in sync with the cloud, but the application reports 159,7MB of usage, whereas du -h reports 180MB. How is this possible? [20:43] alecu, someone else: https://code.launchpad.net/~nataliabidart/ubuntuone-control-panel/multiplatform/+merge/61465 [20:43] sorry, du -h reports 206MB [20:45] nessita, will do. [20:45] thanks! [20:45] philsf, du reports on block usage, not exact file usage. So if a file is 1byte, but the filesystem block size is 4kb, then du will report 4kb. [20:46] philsf, add that error for all the files in the filesystem.... [20:46] philsf: what alecu says :-) [20:46] brb [20:49] wtf [20:49] beuno: uhm, so, i am *pretty* sure that i do not have "578.8 files" in u1 [20:50] dobey, I don't think we tell you the number of files anywhere [20:50] beuno: the dashboard "Files" rectangle just says "578.8" and doesn't classify it as MB or anything [20:51] beuno: which is confusing, given contacts/notes are counts of [20:51] dobey, sounds like a bug. FF4? Have you fudged with the font size? (can you control+0?) [20:51] it is FF4, and doing Ctrl+0 doesn't make MB appear [20:52] alecu, nessita: but both nautilus and "disk usage analyzer" (baoba?) also report similar sizes to du. Why is u1 the only one reporting different total size? [20:52] and yes my default font size is not "humongous" like the actual default in firefox is [20:53] philsf: are you *certain* all your files are on u1? [20:53] dobey, strange, I can't reproduce with the same values [20:53] hmm [20:54] dobey, no. I'm sure they're on one computer (confirmed by rsync), not all of them but the total size reported by u1 does not match anything else I use to measure it [20:54] dobey, http://ubuntuone.com/p/u8q/ [20:55] beuno: it also seems weird to show the quota usage in that spot anyway, to me [20:55] dobey, it is [20:55] philsf, try "du -bh" [20:55] beuno: since i don't think we have any means of distinction for "quota used by files" vs. "quota used by everything else" [20:56] alecu, 164MB, u1 reports 159,7MB. [20:56] dobey, we _only_ have quota used for files [20:56] shhhhhh [20:57] oh well [20:58] philsf, well, we are getting closer, right? Probably u1 is summing the size of directories differently to what du does. [20:59] philsf, du -hb still seems to count full blocks for directory sizes. === hallyn_afk is now known as hallyn [21:15] nessita, what happens here? test_add_oauth_headers.skip = "Why is the patch being ignored?" [21:19] alecu: ah, I gave up :-( [21:19] bah! [21:19] alecu: patch will not patch the function [21:20] alecu: I tried debugging for a while, but I failed miserably [21:20] ack. will take a look there. [21:20] stupid storm [21:20] alecu: thanks :-) [21:35] nessita, I don't like the "MockWebClient.failure = 404" and similars [21:35] nessita, they used to be self.be.wc.failure = 401 [21:36] alecu: right, but since failure is a class variable, is "more correct" if we use the class to access the variable [21:36] alecu: and also, pylint will not stop complaining about that (even if the disable are in place) [21:37] nessita, I find that touching the class variable is touching global state and problems problems etc. [21:37] alecu: I added a cleanup line on setUp... [21:37] nessita, a class variable belongs to a class till it's changed, and then it belongs to the instance. [21:37] alecu: if you find how to make pylint happy, I'm happy to change it back [21:38] nessita, right, I saw the cleanup on setUp. But we'll need to remember to do that in every other TestCase where that mock is used... [21:39] nessita, anyway, I see that this style is in all this file (not only in the changed parts), so let's skip this issue. [21:40] alecu: which, from my POV, makes some sense. But again, I'm happy to change it back, but the pylint warning will not go away [21:40] ok [21:42] what pylint warning? [21:42] nessita, from my POV instances of that mock default to not erroring, and only specific instances are set up to error. So there should be no need to reset the class behaviour on every TestCase setUp. [21:42] alecu: right [21:43] nessita, if pylint complaints, it's because pylint prefers instance variables to be created in the __init__ [21:43] nessita, which I find verbose and ugly when I want a default that I can set on the class. [21:43] alecu: yeah, but the problem is that even if I add the disable clause, pylint will not shut up [21:48] maybe we should write our own python checker thing [21:50] dobey: you think? I would like to be able to fix pylint... [21:52] i think "fix pylint" means "completely rewrite pylint" [21:52] pylint is way too complex of a system [21:52] dobey: I haven't looked at it, to be honest [21:53] nessita: well it's not just pylint; it's astng, logilab-common, pylint, and who knows what else [21:56] dobey: right. Would you be able to re-write it? :-) [21:57] i already have enough to rewrite right now i think === m_conley is now known as m_conley_away [22:09] nessita, I think the patch being ignored has to do with the import code that makes stuff from linux.py show up in a level above. I'm working on fixing it now. [22:09] alecu: fixing the import on __init__? [22:10] nessita, no, the import should be fine. I'll work on the tests. [22:12] nessita, in "ubuntuone/controlpanel/integrationtests/test_web_client.py", try "from ubuntuone.controlpanel.web_client import linux as web_client" (instead of "from ubuntuone.controlpanel import web_client") [22:14] alright, well i think it's time for me to bounce [22:14] have a good evening all! [22:14] nessita, the explanation is: the test code is patching the reference to build_oauth_headers in __init__.py, but not the real definition in linux.py, which is what add_oauth_headers uses. [22:15] bye dobey! [22:15] seems to fix it. [22:19] alecu: but... [22:19] alecu: I was trying to have the webclient tests multiplatform [22:20] alecu: since they don't depend on any platform specific (but the implementation does) [22:20] aha, let's see. [22:20] alecu: what about moving the oauth helper methods to the __init__.py module? that is multiplatform [22:20] and leaving the Soup-specific under linux.py [22:24] alecu: makes sense? [22:24] nessita, makes sense. I'm trying to find out now what it is that OAuthRequest.to_header did. [22:25] ack [22:27] well, it looks like it returns a dict, so this should apply to the twisted web client or the qt web client as well, so I'm +1 with that change. [22:28] alecu: ok, adding that to my branch [22:33] alecu: any other change you would like me to add? [22:34] nessita, I can't think of any other thing. It looks nice!!! [22:35] :-) [22:37] alecu: fixed and pushed to revno 148 [22:49] alecu: I need to run some errands in a few minutes (I need food!). But I don't want to block the review, would you need anything else from me? [22:50] nessita, not at all. go erranding [22:51] great, thanks! [22:53] nessita, approved. [22:53] :-) [23:03] ok I'm gone [23:04] see ya crowd!