[00:05] later all [00:09] nessita, you asked me to fill up a bug. what was it about? [00:09] oh, I remember [00:09] alecu: control panel being redish on quota exceeded [00:10] right. thanks! [00:10] thank you! [00:10] * nessita -> gone! [00:17] bye all [00:45] python looks so clean.. [00:48] karni: compared to dalvik? :-) [00:49] ralsina_babysitt: just compared to java ;> === ralsina_babysitt is now known as ralsina [00:50] karni: well, it sure looks like *less* usually [00:50] ralsina: it looks very good. concise, clean. [00:50] I'm reading action_queue.py from the syncdaemon right now [00:51] I don't know python, but majority is clear to me [00:56] that''s the fun of python. you can pretty much read it [00:56] Of course you can make it awful as you want, but by default it kinda looks nice [00:57] ralsina: hehe :) [01:00] And just in case, this is python: http://lateral.netmanagers.com.ar/static/damned.py [01:00] So don't expect all python to be pretty ;-) [01:01] Much better version: http://lateral.netmanagers.com.ar/static/seven_seg.py [01:01] ralsina: haha. this looks like brainf*ck with few more characters [01:05] It's actually a cool function :-) [01:06] try copying it in an interactive interpreter and then seven_seg("1234") [01:06] ok [01:08] ralsina: now I feel dumb for not understanding the output ' _ \n| ||_ |_ _|\n||| ||_ \n' :< [01:09] ok try print seven_seg("1234") [01:09] sorry I have not used this thing in years ;-) [01:09] it's now a treeliner. [01:09] should I try other values? [01:09] It should be 1234 in large letters ascii art :-) [01:10] * karni opens another terminal [01:10] It should work for any digits [01:11] hahahaha for seven_seg("444") it prints 222, for 555 it prints 333 [01:11] ralsina: anyhow, it's fun ;D cool for a two-liner code in python [01:11] oh, well :-) [01:12] I swear it used to work correctly :-) [01:12] ^ ^ [01:12] python do have a tendency to make backward incompatible changes ;) [01:12] So I've heard. [01:12] still I'll choose it over almost any other language for things it's good at [01:13] for other I'll go with C ;) [01:13] Ok, I'll go have dinner. [01:13] Have nice evening :-) [01:13] ralsina: Thank you (it's 2:13AM ;) ), bon apetit [01:13] evening? it's a bloody middle of the night here ;) [01:13] ralsina: Take care :) [01:13] heheh [01:17] kklimonda: coding? [01:17] karni: I just finished something - now I'm updating 10.10 in vm so I can test one sru [01:18] kklimonda: I see. [01:19] I'm digging into the protocol to understand it deeper and implement the content/meta queue in Java (yeah! ;P) [01:19] Unless I come up with something better. [01:21] * kklimonda is waiting for Qt for android ;) [01:21] either that or I should check Mono for Android [01:22] maybe it's a bit nicer than the Java+XML nightmare Google is calling an SDK ;) [01:22] but that's probably a pythonista in me speaking [01:23] are you serious.. Mono for Android? btw check out the date in the link lol http://tirania.org/blog/archive/2011/Jan-04.html [01:23] future-archive post! [01:24] karni: well, it's already in semi-open beta tests [01:24] kklimonda: I think Java and XML were quite smart choice. [01:24] uhm [01:24] karni: I just find Java, with its checked exceptions, with the accent on XML for doing stuff distasteful. [01:25] karni: I've actually tried to use Android SDK and it was no fun to me [01:25] no fun at all [01:25] kklimonda: it's getting better every release [01:25] I know it works, and that you can do nifty stuff in it [01:25] but for me it completely lacks the "fun" factor [01:25] :D [01:26] yeah, Android should go for Shakespear [01:26] not to mention that the documentation is lacking [01:26] that's true. here and there, yeah. [01:26] karni: it's not that funny - if I have to read and write code for 8+ hours a day it has to fun [01:26] tru. [01:26] karni: I can think in Python [01:26] I actually read python code for fun ;) [01:27] I've started reading CPython interpreter, just to see how it works.. and it's C but still fun (I like C ;)) [01:27] by the way.. I think I'll have to talk to verterok tomorrow. I'm wondering if I should implement the queues more low level then my applicaiton (RequestHandler) looks like a place to start. [01:28] I know that may distaste you, but I've coded GTK with Java, and it was much more pleasant than GTK in C ;) [01:28] karni: btw, are you going to make some sort of authentication provider for other applications to use U1? [01:28] but maybe it was just my incompetence back then. [01:29] karni: Gtk+ in C is.. in C, everything is what it seems - I find it refreshing [01:29] kklimonda: yes. we should have an SSO available to all apps via Authenticator framework, and possibly expose the service itself [01:29] great [01:30] I've taken a look how to write it [01:30] and then decided to make soap bubbles - much more entertaining ;) [01:30] yup, it's nothing complex. hahahah [01:31] examples were using inner classes.. [01:31] really, I have nothing against them - but there was no obvious reason to do that [01:32] and no documentation :( [01:32] ok, my Java rant is over [01:32] kklimonda: Romain Guy says static inner classes are smaller and better on Android. [01:32] kklimonda: I know it sometimes means you've got ten classes in one file, which sounds terrible. [01:33] kklimonda: But you gotta take into consideration it's a mobile environment. You're even expected to [01:33] minimize the number of created objects. [01:33] karni: that's why I don't understand why did they go with Java [01:33] And use primitives as much as possbile instead. [01:34] kklimonda: Maybe because of sandboxing and dalvik? [01:34] karni: choosing Java, and then saying that you have to do all this weird stuff, because normal Java is too bloated seems like a backward thinking. [01:34] karni: iOS have sandboxing [01:34] You're running multiple, separate, sandboxed dalvik virtual machines on your phone. [01:34] karni: and dalvik is just an implementation detail [01:35] kklimonda: it's not that normal Java is too bloated. They have stripped it themselves anyway. It's not bloated anymore as available to Android. [01:36] kklimonda: it's just language specifics - such as those inner classes. [01:36] but seriously.. it's advised to use two primitive ints instead of a Pair object of some sort. because it's an object, and GC has to kick in whenever neccessary. [01:37] Ok.. it's getting late. And my eyes already hurt ehh. [01:37] kklimonda: It's been good talking to you. Take care mate [01:37] karni: have a good night [01:38] kklimonda: tnx, you too. [02:17] just can't stop reading protocol stuff ;d [02:17] the protocol should be documented, seriously :< [02:20] ;D [02:20] karni: haven't you heard? source code is all the documentation you need [02:23] kklimonda: come on.. I even recently read somewhere that source code as documentation is not the way to go. [02:23] karni: meh, that's evil people talking ;) [02:23] kklimonda: plus, it's acceptable if you know the language of the source well. [02:23] kklimonda: xD [02:25] karni: but the ugly truth is that you should get used to it [02:25] karni: there aren't many project - open source or commercial ones - that have a decent documentation. [02:25] kklimonda: I know.. [02:26] projects* [02:27] kklimonda: heading to bed this time for real. take care! [02:27] karni: cya mate [03:07] sigh [03:07] hello latency and fail [11:47] 'lo all [12:04] duanedesign: morning! === teknico is now known as teknico_away [13:34] good morning people! === teknico_away is now known as teknico [14:25] ¡holá! === teknico is now known as teknico_away [14:35] hola thisfred [14:35] hola jefa [14:41] jua [14:41] thisfred: since we're it, could you please start working on splitting the spec you and alecu did yesterday into bug items to attach to the blueprint? [14:41] si! [14:42] thisfred: there are 2 blueprints, one for showing status of shares and udfs, and the other to integrate with unity. I think some work items will go into the first one, and some others into the second one [14:42] https://blueprints.edge.launchpad.net/ubuntu/+spec/other-ubuntuone-n-shares-status-udfs [14:42] sure [14:44] hum, our internet is a bit bursty today. thisfred, I can reach the second blueprint myself... [14:44] do you have the url? [14:45] I'll find it, or we can look it up after the plenary [14:45] I'm translating into a list of work items first, I'll create bugs after [14:46] perfect === dobey_ is now known as dobey [15:57] * nessita -> meeting with Ivanka, cross fingers! [15:58] heh [16:04] Hey guys, I don't suppose someone can figure out why my U1 client won't connect. Running ubuntuone-syndaemon --debug works for a while, then I get "process next in queue (len 0)" and it kinda sits there for ages [16:07] sigh, why is it that U1 waits for me to ask a question, then decides to work? :) [16:07] it's working now === teknico_away is now known as teknico [16:09] kazade, that's a heisenbug [16:09] xyz, yeah, this is the second time I've asked a question in here and then the client has decided to start doing stuff! [16:11] spooky === ralsina is now known as ralsina_lunch === teknico is now known as teknico_away [16:36] kazade, i suppose that's because the servers got restarted recently after the upgrade, so it is not that spooky _now_. hm === beuno is now known as beuno-lunch === ralsina_lunch is now known as ralsina [17:11] hi everyone [17:12] Hey karni. :) [17:12] beuno-lunch: Having read some syncdaemon sources yesterday, it's not feasible (at the moment) to implement all that stuff. Therefore, plan for now: handle gracefully multiple generation bumps with timeouts. [17:12] JamesTait: hello James! [17:13] beuno-lunch: not feasible* in context of our ultimately conservative schedule. we can always get back to it. === teknico_away is now known as teknico [17:36] ralsina: ping === beuno-lunch is now known as beuno [17:57] karni, ack [17:57] beuno: I'm working on it ATM. [17:58] cool [17:58] beuno: excuse my previous .apk teases. I really want to give it to you. [18:01] karni, no worries [18:52] oops, I missed mandel [18:56] beuno: I've also hidden UDFs, SHARES, dirs and files starting with a dot. [18:57] karni, nice. Easily unhidden? [18:58] beuno: yes. will add as an app setting right now. [19:00] beuno: I implemented 10 second timeout before invoking sync on a volume, who's generation is being bumped. (we can try smaller timeout, but it can cause more syncs than neccessary) [19:00] karni, sounds good [19:00] The (resonably) more, the better. We really don't care about those intermediate generation broadcasts. [19:02] beuno: currently, i've taken off the 'sync' option from folders. will get back to that. would be best to start playing with it once we've got some scheduling mechanisms, but I'll consult with verterok first. [19:02] karni, +1 [19:02] beuno: all in all, it's nothing complex. but I care about handling all that stuff in a proper manner (instead of fire-and-forget) [19:36] beuno: unhiding done. [19:37] * beuno w00ts [19:44] beuno: JamesTait: Today I read on twitter that Ubuntu SSO does not work on Firefox with NoScript on. Is that true/you aware of that? [19:44] karni, could be, at least for creating accounts [19:44] uhm [19:45] * karni starts to think that U1 server load increases significantly in the evenings [19:46] karni, yeah, this is something we should nudge __lucio__ about more [19:47] __lucio__, graphs seem to show that around this time, every day, things get pretty slow [19:49] <__lucio__> beuno, we just had a rollout. and graphs dont show what you say. [19:49] __lucio__: that is, you've been updating server software, correct? [19:49] <__lucio__> karni, yup [19:49] __lucio__, well, I see a drop on graphs every single day [19:50] __lucio__, I also see karni complain around the same time every day [19:50] __lucio__: autehtication and client setup takes around a minute or more at the very moment [19:50] <__lucio__> beuno, what graphs are you looking at? [19:50] __lucio__, daily storage graphs [19:51] <__lucio__> beuno, in particular? [19:51] I know I could use a mock client/server, but testing in live environment brings that second touch to results. [19:51] <__lucio__> karni, you should use staging [19:52] __lucio__: what if we deploy the app and people don't use staging :< ? [19:52] anyhow, __lucio__ , I've got a question [19:53] __lucio__: should I modify any part of authentication, you mentioned I shoud use staging.one.ubuntu.com instead of one.ubuntu.com [19:53] *that's a question [19:53] __lucio__: I know I should connect to the IP you gave me. Just asking if anything concerning authentication changes too. [19:55] In the "Desktop Couch Scratch" folder, do I have to manually add the bookmarks I want saved, or does the system do it automatically? [19:56] nirazio: the system does it automatically; adding things to that folder will accomplish nothing [19:57] Chipaca: Thankyou :) [19:58] nirazio: you're welcome [19:58] nirazio: is it working for you? [19:58] it == bindwood [19:58] Yeah it's working [19:58] <__lucio__> karni, same stuff should work as far as i know [19:59] nirazio: sweet === AlanChicken is now known as AlanBell [20:40] beuno: What should I do with downloaded files (not favourite), which have been changed on the server? AU1 removed them and it looked as if it was never downloaded, thus you always accessed the newest version. But I'm not sure if it's the best solution. CardinalFang, you can also speak your mind on that :) [20:40] What would you expect to happen. [20:40] If the file has been changed, even if it's not your favourite, it should not be shown to the user anymore. Since it's not up to date, right? [20:40] Hmm. [20:41] Prompt on attempt to open.. ? [20:42] "A new version of the file is available." | [20:42] What's the prompt? Open local old version, or download newer server copy first?" [20:42] Right. [20:42] AlertBox, right. [20:42] *Dialog [20:42] karni, sounds good to me. [20:43] CardinalFang: sounds sensible. thanks Chad. [21:02] I need help to debug a serious bug: Ubuntu One is filling my /tmp directory with hundreds of files named like "tmpxDFL3N" until my hard drive fills up. How can I see what process writes these files? What is happening? [21:04] duanedesign: rye: Could you help? ↑ [21:06] otto__, hm, it should do one at a time, could you please check lsof ? [21:10] that doesn't sound like ubuntuone [21:12] rye: reading lsof man pages, restarting u1, I'll be back when I get the result [21:15] u1sdtool -s shows server rescan, no tmpXXXX files made yet.. waiting for upload to start [21:17] CardinalFang: You knew a ListPreference cannot be backed by xml integer-array, and it throws an NPE? Found it on android/issues page. Gotta use Strings, and then valueOf() to get int. [21:20] rye: server scan complete, uploading started and lsof shows line: ubuntuone 3839 otto 23u REG 9,0 197406720 1572787 /tmp/tmpl06QfW [21:21] otto__, is there more than one tmp file? [21:22] rye: not yet, please wait until u1 freaks (may take a few minutes) [21:23] rye: correction: it is now uploading yet, since network stats don't show outgoing traffic - we need to wait a bit more [21:25] rye: we are some file listings I captured earlier today: http://otto.kekalainen.net/u1-bugs/ [21:25] rye: I run ls several times so you can see how the tmpXXXX files multiply [21:28] rye: background info: running lucid with all updates, u1 account with paid 4x20GB, recenty added folder ~/Kuvat with 40 GB of stuff, user otto@sange.fi [21:29] otto__, ok, i think you can shut down u1 then remove these temporary files which should not be multiplying actually. facundobatista can syncdaemon create multiple deflated files in /tmp ? [21:30] rye, it should not, but it could on some strange situation [21:37] hmm.. now it is stuck in State: QUEUE_MANAGER, no upload happens [21:38] otto__, is u1sdtool --waiting-content ? [21:38] karni, yeah, I've run into that before. [21:38] otto__, is u1sdtool --waiting-meta or u1sdtool --waiting-content returning anything ? [21:39] rye: Failure: dbus.exceptions.DBusException: org.freedesktop.DBus.Error.NoReply: Did not receive a reply. [21:39] otto__, ah, State: QUEUE_MANAGER, what are the queues ? [21:40] u1sdtool --waiting-content | wc -l works now and shows 7193 [21:40] otto__, is queues: IDLE ? or WORKING_ON_BOTH ? [21:40] queues: WORKING_ON_CONTENT [21:41] otto__, ok, so it is uploading something [21:42] otto__, are there many tmp files in /tmp ? [21:42] only tmpl06QfW at the moment. It isn't uploading or at least there is no network traffic going out [21:47] rye, nothing happens at the moment, I'll try restarting it. [21:48] otto__, nothing as in no upload at all? [21:48] otto__, u1sdtoo --current-transfers ? [21:48] rye, yes no upload [21:49] rye, I already pressed the button. Now State: LOCAL_RESCAN [21:49] nessita, http://www.pygtk.org/pygtk2tutorial/sec-TreeModelSortAndTreeModelFilter.html#sec-TreeModelFilter [21:50] alecu: thanks! [21:52] rye, I'm getting tired (local time 23:52). I'll just post a bug next time i notice this happening, likely some time tomorrow. Thanks for the lsof tip. With it I can prove the the bug is in ubuntuone-process. [21:53] otto__, ok, thanks! [21:56] beuno: I want to track down one last thing today. Transactioned sync broke (it works without transactioning). So I'm looking at it ATM. [22:08] I poked around the FAQ's, but I can't seem to find a definitive answer for this. My computers don't stay connected to Ubuntu One file syncing service that I'm paying for. Files don't get synced. I have to Restart the syncing service manually if I want to get files to sync. Every time. On every computer I have. It's always been like this. I appreciate that there are issues that are being ironed out, but why can't the client just reconnect on [22:08] its own? Why do I have to go in and kick it every single time I want some files to get synced? (I'm running Ubuntu 10.10 on all my machines). [22:26] postfuturist, is your client not connecting on startup of the machine or it gets disconnected during the operation? [22:29] i wonder whether i can recommend nightlies for startup and metadata storage... [22:29] I don't know, I've never changed any settings. I just notice that at some point files stop syncing--for days at a time. I have to open the Ubuntu One Preferences window, the Connect button is greyed out, so I click Restart and then eventually the Connect button appears and I click on it. Then it connects and files sync. [22:31] Does it not connect on startup? [22:32] Chipaca, can "nightlies" be recommended for startup connections/metadata loading (tritcask aswesomeness) ? [22:46] * karni eod, have a good day/evening everyone