[00:21] hello! [00:21] is there not a way to stay signed into U1 on android? [00:22] I keep noticing when I take pictures, it doesn't auto upload them, only to open the app and it requires me to log in... again... [01:31] roasted: hi [01:31] roasted: I'd be glad to help. We've had a few folks report this problem, and I have no understanding whatsoever where it comes from :< [01:32] roasted: Could you tell me what OS are you running? stock/mod and version [01:33] roasted: The problem usually ocurred when the app was not installed on primary storage (long story short), so we disabled app2sd, as it was AccountManager limitation. [01:34] roasted: It can also be related to time settings. Please make sure your time zone and date and time settings are set correctly (usually automatic settings are good choice.) [01:34] roasted: If none of this helps, please poke me whenever you can (Whenever I'm online), I would love to fix this problem, thanks :) [01:44] karni, I'm running CM7 2.3.5 on an LG Optimus T. U1 is updated (recently installed, in fact). Date/time are looking good. [01:44] * karni scratches on the head [01:50] roasted: k, I'll include a fix, that is time related (OAuth services are time sensitive), I hope it'll help. I'll push the update in ~30min. If 1.2.1 doesn't help, please come by and complain to me, okey :)? [01:51] oh wow, so I should see a U1 update shortly? [01:51] roasted: yes [01:51] karni, oh dang, hang on a second. [01:51] I only now noticed you mentioned apps2sd. [01:51] roasted: yes? [01:51] I'm running U1 with link2sd on an SD card [01:52] symlinked or whatever [01:52] HA! [01:52] roasted: I LOVE you dude xD [01:52] It's the only way I could install more than 5 apps on this poor excuse for an Android phone [01:52] roasted: God damn this is the problem. [01:52] I see. [01:52] roasted: So, I can tell you right away.. the fix won't work :/ ehhh [01:52] roasted: Let me explain. I've got bad news and good news. [01:52] Perhaps I'll unlink U1 and see what happens. [01:52] I mean everything else is linked... [01:53] roasted: Bad news - each time you unmount that storage where U1 is, the account is removed from AccountManager. [01:53] I typically keep only specific apps unlinked, such as keyboards, ADW launcher, etc. Things like that make more sense to keep on the onboard storage. [01:53] roasted: next time - what you could do, is if it tells you please log in - [01:53] well, I never unmount it... [01:53] roasted: could you go to AccountManager and check if the account is still there? [01:53] roasted: This would *really* help. I think the Ubuntu One account will be gone. [01:54] roasted: The problem with app2sd was that each time you unmounted storage / rebooted phone, U1 account is deleted (by AccountManager) [01:54] this sucks, because you have to log in again. [01:54] roasted: have you rebooted your phone lately? [01:55] Well, my phone is bipolar, and sometimes spontaneously shuts off. [01:55] So, if by reboot you mean, it was off and I turned it on, yes, yesterday. [01:55] roasted: aha, that's it :< [01:55] yes [01:55] Would it be "fixed" if I unlinked U1? [01:55] and put it on internal storage? [01:56] roasted: This account removal is "AccountManager by design" [01:56] roasted: yes, that would fix it [01:56] roasted: (nothing personal, just asking) If you left a comment on Google Play (we've had a few pepz with the same problem), consider updating it. It's not something we can fix on our end :( [01:56] I guess I was a lucky sucker who bought a phone with a ridiculous 200MB or whatever of app storage. [01:57] roasted: Oh man :( [01:57] Nah I hadn't left a comment. [01:57] I just installed U1 on it recently. [01:57] Maybe I could add a workaround switch, that could hold the token in app preferences. [01:58] Nah, if that's all it is, then it's a done deal. [01:58] roasted: If it's not much work, I might look at that ↑ in the following weeks. Can't give a date, though. [01:58] Great [01:58] After all the need for link2sd type of setups is becoming less apparent as newer phones have more storage. [01:58] roasted: If it still fails, make sure to come and complain :) [01:58] That is correct. [01:58] I was just cheap and wanted a low end android and realized, oh hi, 5 apps and your phone is continually being flagged as low on storage. [01:58] Ridiculous. [01:59] link2sd saved this phone's life. I would have dropped it asap. :P [01:59] But I have nearly every app linked... so I'm sitting with ~70MB free, whcih is above average for me. (sad, isn't it?) [01:59] I see. Well, not fun certainly :( [01:59] Maybe if the Ubuntu/Android thing kicks in I'll upgrade. :P [02:00] be nice to have a setup like that :P [02:00] :) [02:01] hm [02:02] would this issue be dalvik cache related? [02:02] link2sd says that's the only thing U1 related that's linked [02:02] unless its somehow due to the apps2sd thing under the settings menu [02:03] roasted: it is AccountManager. it's a system wide component that apps use to store credentials/authentication keys [02:04] we keep our OAuth token there, it's about sharing and safety [02:06] I see. Good deal. [02:06] Well it's all unlinked now, from link2sd and apps2sd [02:06] we'll see how it goes :P [02:11] roasted: thanks :) [02:12] karni, appreciate the insight! [02:13] roasted: I'm happy I know what link2sd is. This is why I have stock Android on my phone. If something breaks, I *may* have a suspicion it's the ROM. Otherwise I'd also be running CM, although with 16GB on GN no need for link2sd :) [02:14] karni, I hear ya. I'm trying hard not to upgrade... [02:18] Mostly because I have to fix up a bathroom here at the house before I even consider getting a new tech toy I don't truthfully *need* [02:20] roasted: Heheh [02:20] Pushed out 1.2.1. Maintanance update. [02:23] * karni Night all, take care roasted o/ [02:32] take it easy! [08:29] morning all! [08:34] ok, time for a reboot due to updates [09:35] Morning all! :) === rye_ is now known as rye [09:56] erm guys, we have python-django-openid-auth (our package) and python-django-auth-openid which conflict with each other [10:07] rye: yes [10:07] rye: they're both the same thing, pretty much [10:10] Chipaca: and they both install the same files and one of these (our) is at 0.2 version and another is 0.3-2 [10:57] good morniing.... mandel !! jeje you are the only one around :P [10:57] gatox, morning! [10:58] mandel, how are you? [10:58] gatox, fine, I fixed a bug in the nightlies we added when doing the refactoring :( [10:58] mandel, oops.... [10:58] mandel, i saw you take a lot of food pictures during the weekend :P [10:58] gatox, I ow you a number of reviews, can you fix the first mac-fsevents branch followign alecus comment, then I'll move from there :) [10:59] gatox, a couple, some of them are even nice :) [10:59] mandel, yes! on it! [11:21] mandel: is com.ubuntuone.Authentication gone? [11:22] rye, no idea, may have been deprecated, does it start? [11:22] rye, as it, show in d-feet [11:24] ERROR:dbus.proxies:Introspect error on com.ubuntuone.Authentication:/: dbus.exceptions.DBusException: org.freedesktop.DBus.Error.ServiceUnknown: The name com.ubuntuone.Authentication was not provided by any .service files [11:26] rye, hmm I wonder, lets grep sso to see if it is there.. [11:26] rye, I mean u1-client [11:27] rye, seems not to be there but I don't know when it was removed.. com.ubuntuone.CredentialsManagement is there [11:29] mandel, do you know if we are using skipIfNotOS anywhere? i can't find it [11:30] gatox, the decorator? you cannot find it? [11:30] mandel, yes..... i can't find it where is being used [11:30] gatox, it is declared here: ubuntuone/devtools/testcases/__init__.py [11:31] gatox, it could be that we don't use it in u1-client, why? [11:31] mandel, yes, that i found it [11:31] not in u1-client, or u1-cp or u1-sso [11:32] mandel, just wanted to see if i can say SkipItNotOS("linux")...... but it seems i need to say linux2 [11:32] gatox, yes, you need to say linux2.. [11:33] gatox, do this better skipIfOS(('win32', 'darwin'), 'I just ahte anything but linux') [11:34] gatox, but spell hate correctly ;) [11:34] mandel, yes.... i can do that too.... and i prefer it.... thx [11:35] gatox, np [11:39] Am I correct with understanding that there's no document viewer as part of U1? [11:39] I'm noticing I have to download the picture in order to see what it is... for example... [11:43] gatox, may I have a review for https://code.launchpad.net/~mandel/ubuntuone-fsevents-daemon/add-fsevent-tests [11:44] roasted, is that for the web of the desktop client? [11:44] mandel, yep.... in a couple of mins [11:44] gatox, thx! [11:51] mandel, ok..... i have to be sure that the change of the decorator doesn't break anything..... so i'm going to run the tests for that branch in win/linux/darwin....... is going to take a while..... in the meantime i'm going to review your branch [12:00] mandel: either... [12:03] roasted, well, the client is not meant to be that smart, it syncs the files for you to use.. and if you think about it, if you need a viewer to see the doc, you already downloaded it so there is no gain, right? [12:04] roasted, on the other hand, in the web.. yes you should not need to download an img to view it, seems stupid to me.. I wonder who might not about this, rye? JamesTait ? [12:04] mandel: I hate to compare, but given the conversation, I will... I use U1 to back up the pictures on my phone so I can easily view them. I ran into an issue where an image failed to upload, so I wanted to see which one I missed. [12:04] mandel: I ended up downloading a series of images before I gave up and installed Dropbox on my phone. [12:05] I know you say that viewing it is essentially the same thing as downloading it, but viewing it requires 1 click where it appears on the screen. [12:05] roasted, there is not problem (at least for me) to compare :) [12:05] * JamesTait reads scrollback [12:05] downloading it has to close down U1, go int he app drawer, open gallery, and find the downloaded picture [12:05] aka, not a solution [12:05] roasted, so, the exat device in which you wanted a preview is on your phone, is that right? [12:06] roasted, sorry to ask, but we know support android, ios, windows, linux and mac in the near future so I want to get the details right :) [12:06] mandel: I'm referencing my android phone, yes. [12:07] I know downloading vs "caching" an image to view is essentially the same thing, but at the same token, it's not. [12:07] roasted, ok, perfect, the person that can do something about android is karni [12:07] ah yes, he helped me yesterday with a separate issue [12:07] :P [12:07] roasted, I could do something if ti was related to the 'desktop' versions of the software :) [12:07] roasted: So for the Android use case, karni is your man. [12:07] mandel: well, isn't U1 the same on the web? [12:08] hang on let me try it [12:08] I thought I remember it doing the same thing [12:08] roasted, that is out of my scope, as in, my head is too small to know of the bits :) [12:09] roasted: For web, I think there might even be a bug already open about this. [12:09] yep, same deal [12:09] I click a picture, it downloads [12:09] sigh... [12:09] roasted, yeah, that on the web is ugly.. [12:09] JamesTait, any idea of the bug number? [12:09] mandel: Just checking. [12:10] I quite like the interface, to be honest. [12:10] I don't think it was specific to images though. [12:10] But downloading a file to simply view it is so ass backwards. [12:10] JamesTait, I think at some point we should have a wiki like.. If you can deal with code monkeys poke this people depending on the client: and then a list with code monkeys :) [12:10] JamesTait, plus the, always ask rye first hehehe [12:10] roasted, yes yes, I think we all agree on that one [12:16] Hm, can't seem to find it now, but I'm sure this has come up before. [12:29] ok, I have to feed the brain aka I'm off to have lunch :) [12:29] * mandel lunch [12:30] Firefox, for me, attempts to open it in a new tab (JPG and MP3), then closes the tab and prompts for an app. [12:31] roasted: Are you comfortable filing a bug for this? It's certainly something that could be improved. [12:32] If it turns out there already is a bug open, we can always mark it as a duplicate. [12:32] mandel: you guys.... http://paste.ubuntu.com/1047158/ [12:32] good morning! [12:33] 4 generations of authentication methods [12:33] rye, wow! that is terrible! [12:33] rye, did we write that? [12:33] rye, as in, is in the code? [12:33] mandel: no, that's what I am using to get lucid...precise/nightlies working with my support scripts [12:33] mandel: lucid..precise [12:33] rye, ok, lets move dbus_service to linux [12:34] rye, add that as a bug I'll make sure it is get fixed by today [12:34] rye, is horrible, idiotic and we broke an API which should not happen [12:34] roasted: well have this as soon as server side is ready. we are looking into it. [12:34] mandel: well, what was the reason behind moving it under separate dbus_service ? [12:34] For example, clicking an MP3 file in the web UI could open the web music streaming functionality: http://voices.canonical.com/ubuntuone/2011/12/15/happy-holidays-celebrate-with-web-music-streaming/ [12:35] rye, was though to be a better name in a diff branch [12:35] rye, having said that.. we should be able to allow you to import without stating that is linux [12:35] mandel: the API - CredentialsManagement of SSO has not changed - i am using a low-level api [12:36] rye, and if you import it in windows (that is the dbus paths) we should raise an import error syaing, nah on windows/darwin [12:36] rye, everything that can be imported is an API ;) [12:36] * JamesTait -> Lunch [12:37] rye, so, yes, file a bug I'll take care of it today [12:37] CredentialsManagement moved? [12:37] karni: good to know! I want to stick with U1 because I'm a big fan of it but the lack of viewing on the fly is a bit of a hindrance. [12:37] wait, i don't understand all the implications so far [12:37] roasted: stay tuned [12:38] rye, no credentialsmanagement is not moved, actually yu should be able to import it without stating the platform [12:38] rye, let me go for lunch look at how you use the code and decide, ok? [12:39] rye, I don't think well with a noise stomach.. (if we where in an office I'd drag you to the canteen/kitchen :D ) [12:46] mandel, +1 to your branch [13:00] hmm [13:02] karni: is this going to be those "hang on for 10 minutes while I whip up a magical solution" type of scenarios? Or shall I "stay tuned" in the following weeks? :P [13:03] roasted: The latter. Sadly, it's more dependant on the server. Although I wish I could whip it up :) [13:03] karni: haha, all good. At least when my phone reboots I'm still signed in to U1. :P [13:05] roasted: ha! :) [13:10] mandel, this is ready: https://code.launchpad.net/~diegosarmentero/ubuntuone-client/darwin-fsevents-1/+merge/110382 [13:12] mandel: can you fix your fix-tcp-activation branch please? looks like there were lots of lint errors [13:14] roasted: I'm glad we talked this last night (at least night at my place), knowing about link2sd is crucial on support side. [13:15] Yeah, I hear ya. I'm sure the link2sd conversations will begin to diminish even more and more as time passes and people upgrade to phones that don't have piss poor excuses for storage space. [13:18] Hello, all [13:21] hola alecu [13:24] karni: just to make sure, U1 on my laptop is more of a 1 way sync, right? I can't just put a file in the U1 folder in Nautilus and expect it to sync upwards -no? [13:24] roasted: If you put a file in U1 folder on your laptop, it will upload to all _PC's/machines_ connected to U1 [13:25] roasted: mobile clients (Android/iOS) are sort of "on demand" [13:25] roasted: U1 syncs files between your computers, so that when you create/change/delete a file on one computer, it will sync that state to other computers connected to U1. [13:26] Oh, oh wow. [13:26] I wasn't aware it was that... nice. [13:26] alecu, hi [13:26] alecu, i updated my branch [13:26] I thought it was more of a 1 way street/manual push type of thing [13:27] gatox, alecu, mandel, dobey: if anyone needs a review, I have half an hour to kill [13:28] i hope you're not using an extremely inefficient instrument for that [13:28] gatox: great, I'll take a look [13:29] alecu, thx [13:30] ralsina, all: I'm feeling all righty, but I've got an old dog's cough that's been lasting over a week. So I'll be going to check with some doctor in an hour or so. [13:37] mandel: thanks for the review. Regarding the "author and email" comment line, since these are branches for the stable trees, I think those comments should still be there. === zyga is now known as zyga-food [13:41] alecu: take care [13:41] dobey: very inefficient, it kills half an hour in half an hour. [13:42] dobey: in average, it takes 78 years or so to kill a person. [13:43] life is a very inefficient means of killing [13:46] dobey, really philosophical of you :P [13:47] oh wow. xubuntu's new theme is like a direct copy of OSX [13:48] dobey, jaaaaaaaa you think? [13:48] i guess OSX is the new CDE [13:49] just needs a giant printer sitting on the dock [13:51] * mandel back [13:51] alecu, sure, then approve them :) [13:52] alecu, uh, just read you are broken, I'll approve if you didn't [13:52] ralsina, if you are free, 1-1? [13:52] mandel: in 3' [13:52] ralsina, great, I'll me in mumble waiting [14:00] ralsina, right now: http://paste.ubuntu.com/1047158/ [14:07] mandel or ralsina: trivial branch to review: https://code.launchpad.net/~dobey/u1sync/fix-api-break/+merge/110827 [14:07] dobey, on it! [14:07] dobey: sure [14:07] u1sync? [14:07] are we still shipping that? [14:08] no [14:08] it's only in bzr [14:09] but the security fix last week broke api [14:09] ok [14:09] dobey: you have +2 [14:18] also, tarmac is not finding u1trial, which is a new one [14:18] which tarmac? [14:19] dobey: u1sync, it seems [14:20] weird, i didn't see the jenkins irc notice [14:21] dobey: I just got the email [14:23] ah, i guess u1sync probably shouldn't be landing on lucid [14:23] or it shouldn't use u1trial [14:25] * dobey makes it use plain trial/pyflakes [14:27] it has horrible coverage anyway [14:35] hi folks, catchin' up... [14:35] mmcc, hi === zyga-food is now known as zyga [14:47] hi team, i'm here. sorry for the late start. [14:48] wtf gtk [14:54] hello briancurtin [14:55] hi ralsina [14:55] briancurtin: I saw at what time you uploaded the binaries, so don't worry about the late start :- [14:56] hopefully it's a quicker process this time to get them back and then get the installer signed :) === yofel_ is now known as yofel [14:59] team, standup! [14:59] in .2 minutes! [15:00] me [15:00] me [15:00] me [15:00] meh [15:00] me [15:01] me [15:01] alecu is out, right? [15:01] mandel: probably still at the doctor's [15:01] gatox: standup! [15:01] me [15:01] ok, mandel go [15:02] DONE: bug 110792 and bug 1014094. Reviews, reviews. Got sd to start on mac os x and get events from fs I'm blocked looking at an sso error. [15:02] TODO: find sso bug and fix it. Reviews, take a look at fsevents daemon branches. [15:02] BLOCKED: no [15:02] briancurtin, please [15:02] Launchpad bug 110792 in Ubuntu "[needs-packaging] Beyond The Red Line" [Wishlist,Invalid] https://launchpad.net/bugs/110792 [15:02] Launchpad bug 1014094 in Ubuntu One Client "ubuntuone-login crashed with ImportError in __main__: No module named linux" [High,Fix committed] https://launchpad.net/bugs/1014094 [15:02] DONE: in my constant battle with dependencies/buildout, i figured out the py2exe problem and got binaries sent off to be signed [15:02] TODO: follow up whenever signed binaries are available and create the signed installer and upgrade the rest of the files and links [15:02] BLOCKED: no [15:02] NOTE: i may or may not have a PSF board meeting/call this morning, but it looks like a few people are trying to move it [15:02] NEXT: thisfred [15:02] DONE: Bug #1006882 Bug #1006889 TODO: Bug #1006872 BLOCKED: no NEXT: dobeyh [15:02] Launchpad bug 1006882 in U1DB "sync_exchange should include txid on each doc" [High,In progress] https://launchpad.net/bugs/1006882 [15:02] Launchpad bug 1006889 in U1DB "api for validating transaction_id of target_replica" [High,In progress] https://launchpad.net/bugs/1006889 [15:02] λ DONE: reviews, more 3.99.0 release work [15:02] λ TODO: finish releases/uploads, triage, tarmac tweakery [15:02] λ BLCK: New gtk+ in quantal broke some sso tests. [15:02] Launchpad bug 1006872 in U1DB "sync_exchange should transmit the txid it thinks the target was at" [High,In progress] https://launchpad.net/bugs/1006872 [15:02] mmcc [15:02] DONE: controlpanel bugs [15:02] TODO: awaken controlpanel [15:02] BLOCK: none [15:02] NEXT: ralsina [15:02] DONE: calls, reviews, askubuntu, canonicaladmin, talked with people, etc. TODO: tech leads call, reviews, find a bug to have fun with. BLOCKED: no, NEXT gatox [15:02] DONE: [15:02] Reviews, propose the 4 branches (split the big one), start working on fsevents remaining features. [15:02] TODO: [15:02] Finish with fsevents remaining features. [15:02] BLOCKED: [15:02] No [15:03] Comments anyone? [15:04] noop [15:06] EOM then [15:07] gatox, ralsina, can you refresh my memory? do you remember what should happen in u1-client when sso emits a CredentialsNotFound signal? [15:08] mandel: it should switch to the wizard's "signin/signup" page [15:08] i just ♥ API stability guarantees, that aren't actually guaranteed [15:08] ralsina, well, that is the bit that is not working atm, everything else works (which is just getting the events in sd :( ) [15:09] mandel: not working on mac you mean? [15:09] mandel: wait, in u1-client? [15:10] mandel: I don't think it needs to do anything besides disconnecting fromt the server and switching to no user status [15:10] mandel: unless I am misunderstanding you [15:10] ralsina, yes, that is what it is doings, jus runs with no user [15:10] ralsina, I though u1-client would request the creds, doesn't it? [15:10] mandel: nope [15:11] ralsina, no? oh well [15:11] mandel: but if you could start u1cp, it would :-) [15:11] ralsina, ok, I'll set my creds and see what is going on :) [15:11] mandel, remember that my branches are ready for review....... [15:12] mandel, today is your review day.... so i have the right to annoying you [15:12] :P [15:13] gatox, yes you do, but did you update the branch following alecus comment and told me? ;) [15:13] mandel, yes.... while you were having lunch [15:13] mandel, this is ready: https://code.launchpad.net/~diegosarmentero/ubuntuone-client/darwin-fsevents-1/+merge/110382 [15:13] mandel, and i gave the +1 to your branch [15:13] gatox, I missed it :) [15:13] mandel, forgot to set it globally approve [15:13] gatox, looking in a sec [15:14] done [15:14] gatox, mmcc, in sso I'm getting a 'There was a problem trying to setup the page' yet this used to work, right? [15:15] mandel, yes, it used to work [15:15] mandel, let me try here [15:15] hrm. mandel, what error are you seeing ? [15:16] mmcc, backend is None [15:16] that is not familiar. you're doing '--login_only'? [15:17] mmcc, gatox, found the reason I was trying to use the sso lib in the site-pacakages dir and not in the sso path, means that the find bin is not working, not too hard to fix :) [15:17] :P [15:18] mandel, yes.... is working here from sources [15:18] mandel: can you explain a bit more? which site packages dir? [15:19] mandel: from the .app package? or did you have it installed somewhere? [15:19] mmcc, did a python setup.py install on the sso project which installed the lib and therefore it cannot find the bin [15:19] mmcc, my fault [15:19] ah, ok. [15:19] if only these stupid computers would just know what we *wanted* [15:22] mmcc: they know. They are reusing. [15:22] refusing* [15:23] ralsina: you're thinking of cats [15:23] ok..... lunch for me..... brb === gatox is now known as gatox_lunch [15:23] mmcc: there is a reason why OSX versions are named after cats. It's all genetics. [15:23] * ralsina looks forward to OS X 10.8 "boots" and 10.9 "fluffy" [15:24] ralsina: hah. 10.10 "reboots" [15:32] ralsina: because Steve Jobs is the old guy living in the overgrown wastes of Washington D.C. in Logan's Run? [15:33] dobey: OMFG, I have forgotten Logan's Run [15:34] dobey: or you mean the series? [15:34] the movie [15:35] in that case, looks like I have forgotten about that bit [15:35] http://www.youtube.com/watch?v=5QlJSqpBjiE [15:35] Logan's Run came out when I was 5 after all [15:35] http://www.youtube.com/watch?v=sax6J8n1AiE is much better even [15:36] dobey: what's that captured with? A barbiecam filming a super8 show filmed off a TV? [15:36] stupid people recording videos of televisions with a camera [15:36] apparently [15:38] briancurtin: when you have a sec, could you reevaluate your 'needs fixing' on https://code.launchpad.net/~mikemc/ubuntuone-windows-installer/setup-mac/+merge/110155 -- I fixed the style issues... [15:38] mmcc: will do [15:39] thanks [15:54] ralsina, what needs to be setup to see all the logs from u1-client in the stdout? [15:55] U1_DEBUG=True [15:55] mandel: and passing --debug to syncdaemon [15:55] ralsina, ok, thx! [16:12] pfibiger: heya! [16:14] pfibiger: so, yeah, I don't have high hopes, but I think this is a semi-accurate description of what happened: http://paste.mitechie.com/show/699/ === salgado is now known as salgado-lunch [16:16] greg-g: so it sounds like it wasn't u1 doing anything wrong, it was a dying ssd having files disappearing, and then u1 saying "hey, deleted files. let's propagate that delete." is that fair? [16:17] rye: can you give greg-g a hand? he's lost some critical files in u1, and i think we might be able to restore the previous revs for him [16:17] pfibiger: yeah, fair. [16:18] pfibiger: totally a crappy situation that isn't anyone's fault except mine for not having a real offline backup :/ [16:18] greg-g: it just helps us figure where we need to start looking for the files :) ...which email address is hooked to the account? [16:18] greg@grossmeier.net [16:18] greg-g: ok, let me have a look [16:18] rye: thanks man [16:19] (I assume, you may in fact be a woman, who am I to know? ;) ) [16:19] mandel: actually you'll have to hold off on your fix-tcp-activation branch landing, as a gtk+ update broke sso tests on quantal. [16:20] greg-g: I'm Roman Yepishev, [rye], so, what's the location of the files, ~/Ubuntu One/ ? [16:20] dobey, ok, let me know when is ok to merge [16:21] rye: ~Ubuntu One/Grad_School/* [16:21] ralsina, gatox_lunch, mmcc, I've reach the point in which syncdaemon does work but local rescan is never executed.. which means nothing is really done, I'll look closer at the issue [16:21] mandel: i will. you still have to fix all the lint errors though :) [16:21] ok [16:21] rye: the folders are all there, just not the files, which is why I am blaming the dead SSD [16:21] dobey, oh, I forgot about those.. [16:21] mandel: whoohoo [16:21] mandel: breaking further in the road is better, so you get the whoohoo anyway [16:22] nice mandel! [16:22] greg-g: i am running a recovery task. This is suspicious though [16:22] gah, u1sync [16:22] :( suspicious doesn't sound good [16:22] i'll deal with you later [16:24] ralsina, mmcc, gatox_lunch, in case you care: http://paste.ubuntu.com/1047567/ [16:24] ok, i need to get lunch. bbiab [16:24] mandel: that looks strngely familiar [16:24] mandel: also, it looks like you arepretty close to having it running, if it got that far [16:26] ralsina, yes it does, it got me a little uneasy because it turns out that TextEdit defaults file format is a dir, weird yet the events are right [16:27] mandel, an .rtfd? so the events should show a dir, right? [16:28] mmcc, yes they do, I had no idea it was a dir and was really worried I screwed things up! [16:28] hmm, weird. unless you pasted an img, it should be just an .rtf, one file. maybe it's different on 10.7 [16:28] mmcc, is a dir with a TEXT file in, well the good news is that it works [16:30] mandel: is it a .rtf that's a dir? or an .rtfd? if .rtf is showing up as a dir, that's strange to me [16:31] mmcc, .rtfd [16:31] ok cool [16:50] hey, I'm a little stuck here in a soup of twisted/pyqt/qt - been looking at function traces of controlpanel, but nothing's jumping out... starting up, it gets past talking to SSO client to check for credentials, and we call on_credentials_not_found in controlpanel/gui/qt/controlpanel.py, which should set the current Widget to the wizard, and turn off the overlay, showing the wizard. those functions all get called but have no effect - [16:50] the UI hangs on one frame of the overlay, and we just get a trace of QTimerEvents forever after, with no effect. [16:51] I'm feeling my lack of twisted & QT knowledge here :\ [16:54] mmcc: the QTimerEvents may be because of the overlay animation [16:54] mmcc: is it animated? === salgado-lunch is now known as salgado [16:54] mmcc: also, pastebin of backtraces much appreciated :-) [16:54] ralsina: yes the timer events are for the overlay. no, it's not animating. [16:55] ralsina: no backtraces, but I'll paste my function trace. using sys.settrace, so the format might not be familiar [16:55] mmcc: it's ok, just want something to look at :-) === gatox_lunch is now known as gatox [16:55] https://pastebin.canonical.com/68371/ [16:56] briancurtin: we have signatures! :-) [16:56] YES [17:00] dobey, I'm getting the following paste.ubuntu.com/1047624 any idea? [17:02] back, had to step away for a sec [17:03] on that trace, I filter out calls to/from the logging package to make it shorter [17:03] mandel: you should preceed your links with protocol and :// :) [17:04] mandel: and you have an old storage-protocol [17:04] dobey, and how will I trick you to go to lemonparty.org? :) [17:04] dobey, superb, thx! [17:04] dobey, oh, DO NOT VISIT THAT LINK [17:04] mandel: i'm lazy, not illiterate [17:05] ralsina, mmcc, gatox, the reason why I was not moving in sd is because set_application_name was raising and exception, I've set it up to do nothing so we can move foward :) [17:05] dobey, lol [17:05] and the lines saying 'event' are from loadingoverlay.py's timerEvent(). the settrace function tracing only starts at the beginning of on_credentials_not_found, so that's why the trace isn't as verbose at the beginning [17:06] mandel, that is the correct implementation. we're not setting the app name programmatically on osx anyway [17:06] mmcc: what is in your ubuntuone-control-panel/ubuntuone/controlpanel/gui/qt/controlpanel.py:135 ??? [17:06] mmcc, the exception was crashing the daemon but the reactor kept running :( [17:06] mmcc: because for me that's a harmless call to a function that parses storage sizes :-) [17:07] ralsina: the line numbers are off because I have a bunch of debug crap (the settrace function) at the beginning :( [17:07] ohhhh [17:07] 135 is self.ui.wizard.restart() inside on_credentials_not_found for me [17:07] mmcc: ok, then I assume the problem is caused by the threads that calculate folder usage for the wizard [17:07] oRLY [17:08] mmcc: which you should be able to test by... runnig the tests for them, which are there somewhere [17:08] and now really, I need to have lunch [17:08] hrm, but all the tests pass :< [17:08] mmcc: could be a race condition, that code is fragile [17:08] ok I will let you eat, I have my clue [17:09] mmcc, are you having problems trying to execute u1-cp? [17:09] rye, which is the correct info to pass to sd from the keyring as the oauth? [17:09] rye, comsumer_key:consumer_secret? [17:09] gatox: yeah, see above. it doesn't manage to show the setup wizard... [17:10] mmcc: hope it's a good one, but it's the only thing involving threads there [17:10] mandel: a:b:c:d or c:d where a- consumer key, b-consumer secret, c-token, d- token secret [17:10] rye, thx [17:10] mmcc, if i try to execute it here i'm going to see the same, or do i need something? [17:11] gatox, I think trunk should get this far, checking [17:14] gatox: yes, my local changes are only debug, so if you start controlpanel with U1_DEBUG you should see it getting the credentials, then just hanging... and the UI should be frozen on the first dot of the overlay [17:15] mmcc, can you show me how are you running it..... i'm having some problems with the pythonpath [17:15] gatox, yeah - U1_DEBUG=1 PYTHONPATH=.:../ubuntuone-client:../ubuntu-sso-client python ./bin/ubuntuone-control-panel-qt [17:16] from ubuntuone-control-panel/ [17:18] mmcc, do you add something from PySide?? [17:18] did you.... [17:18] ralsina: installer uploaded for signing [17:19] mmcc, i'm seeing an error that says that can not import PySide.QtCore [17:19] gatox: no I haven't done anything with PySide. [17:21] mmcc, it was a problem with the path.... already fix it..... weird [17:21] mmcc, so...... i see the loading overlay getting stuck.... but nothing else.... should be any particular traceback? [17:22] gatox: no, no traceback. if there was a traceback, I'd be happier :) [17:23] * briancurtin lunch [17:24] mmcc, ok, i can debug this for a while if you want, i've fought with something similar symptons in the past [17:26] oh wow. this case is awesome [17:31] all, I'm off to walk the dog, gatox I'll get all your reviews ready during my dinner :) [17:31] mandel, do it tomorrow! [17:31] but early tomorrow [17:31] jeje [17:38] * mmcc was offline for 10 minutes and didn't notice... [17:39] mmcc, oops.... i ask: i can debug this for a while if you want, i've fought with something similar symptons in the past [17:40] gatox: I replied to that, did it get lost? [17:40] mmcc, never receive the reply [17:40] gatox ah, ok - I said "sure, I wouldn't mind the help. I'll keep looking too, so let me know when you learn any clues" [17:40] mmcc, great [17:40] gatox: and "I don't think it's the folder size calculation threads as ralsina suggested, they don't seem to get launched until the folders tab is shown." [17:41] mmcc, i'll let you know as soon as i find anything :D [17:44] mmcc: approved setup-mac [17:45] briancurtin: great, thanks. btw, I wonder if it's worth thinking about factoring out common things between the two platforms' setups, like copying in the certificates? It'd be nice to not have to worry about that kind of thing getting out of sync. === zyga is now known as zyga-afk [17:55] mmcc: sounds good to me, similar to how we already break up platform imports elsewhere. setup.py is the main one, then setup-mac and setup-win do special stuff [17:55] or something like that [17:58] yeah, that'd work. I was thinking more like having a 'common.py' or something that both import, and common has things like copy_datafiles(dest) ... but I don't have a good reason. seemed like the minimum change? [18:13] mandel: what? Is syncdaemon on mac working (just read your tweet) [18:43] gatox: any clues yet on this controlpanel UI problem? I'm learning plenty, but nothing immediately useful yet... [18:43] mmcc, not yet [18:43] ralsina: fyi, gatox offered to poke around on this - and I don't think it's the threads in the folder size calc code, we're dying before they'd get launched [18:44] mmcc: ack [18:44] mmcc: just guessed that because of posixpath and threads being involved [18:44] guys & gals: I'm taking the rest of the day off... But first, if anybody needs me to finish any review, it's time to let me know. [18:44] I know gatox did. [18:45] alecu, i have a couple :P [18:45] alecu, are you ok? [18:45] ralsina: ah, ok - my bad. I can explain that. that's called from the logging code, which I've filtered out very crudely by saying "if "/logging/" in func_name".... so once it gets deeper into posixpath stuff, the filter doesn't work [18:46] ohhhhhhh ok [18:47] sorry about that [18:49] gatox: send them my way! [18:49] gatox: I'm feeling fine, just with an ugly cough since last week. [18:50] i'm pretty tired too...... but...... i had a busy weekend :P [18:50] alecu, https://code.launchpad.net/~diegosarmentero/ubuntuone-client/darwin-fsevents-1 - https://code.launchpad.net/~diegosarmentero/ubuntuone-client/darwin-fsevents-2 - https://code.launchpad.net/~diegosarmentero/ubuntuone-client/darwin-fsevents-3 - https://code.launchpad.net/~diegosarmentero/ubuntuone-client/darwin-fsevents-4 [18:50] torture was the name of my weekend jeje [18:50] * gatox likes to complain [18:53] alecu: in case you missed it, mandel claims to have syncdaemon running on mac after he fixes one bug [18:53] pic or it didn't happen! [18:53] alecu: I am not believing yet ;-) [18:54] but apparently bug #1014752 is the last one before it "works" [18:54] Launchpad bug 1014752 in Ubuntu One FsEvents daemon "An error occur when a path is added twice" [Medium,Triaged] https://launchpad.net/bugs/1014752 [19:07] gatox: ping [19:07] gatox: I'm reviewing the -1 branch [19:07] gatox: and I'm trying to understand why your branch skips some of the tests [19:08] gatox, like for instance: test_commit_middle_events [19:08] * gatox checking [19:09] alecu, grrrr.... it shouldn't, that was a consequence of changing skip_if_win to skip_it_not_linux [19:09] alecu, i think that we need to have both to avoid this [19:10] gatox: right. Furthermore, I think we should have two different decorators for darwin: one for your implementation, and one for mandel's [19:10] * mmcc has a late lunch [19:11] alecu, so...... i need to revert the last changes to have skip_if_not_linux.... and leave it as before? [19:11] school run, brb [19:12] gatox: probably. But that's as deep as my knowledge of the matter goes. [19:12] alecu, yes...... i'm pretty sure that we need to differenciate that [19:12] so....... changing it back [19:14] alecu, i'm going to revert the last revision and push it [19:27] alecu, darwin1 reverted [19:53] alecu, so, i already revert all the branches..... and saw your comment on branch-2...... sounds logic...... [19:58] because the darwin part is based on windows..... with some differences based on the lib we are using....... and some others parts were darwin behaves like linux [20:04] well, that was neat: http://pastebin.ubuntu.com/1047991/ [20:07] alecu, i'm at eod now..... and reallyyyyyyy tired :S, but please send me an email if you considere with should unify windows and darwin implementations so i can start working tomorrow morning on that...... there are part where they are really similar and others where they are not.... as in paths handling, ancestors, and stuff [20:09] ok.... i'm leaving..... alecu please send me that email if you can, i need to know that before going deeper in the other implementations..... bye all! [20:28] gah, --help-reactors isn't being helpfully crashy :( [21:01] hrm, still banging my head against this GUI freeze - if I comment out the code that creates the backend proxies in backend.py (and hardcode a 'no' answer to getting credentials), I get it to show the signin page, but it's not responsive. the UI just spins, but there *are* qt events being sent (they're timer events being filtered out by the (now disabled) loadingoverlay... [21:02] * mmcc hopes that typing that out will jog something [21:04] mmcc: no, doesn't ring any bells :-( [21:04] hrm. [21:04] (thanks) [21:15] EOD for me, bye ppl! [21:37] anyone familiar with getting nonsensical stack backtraces from pyqt ? [21:39] I was using inspect.stack() to dump a trace, but this trace: http://paste.ubuntu.com/1048159/ doesn't make sense between line 11 and line 10... [22:03] alright all, have a good evening!