[08:01] greetings [08:16] hi MacSlow [08:16] hey kvalo [08:16] kamstrup: any tips for a simple (for now) logging in a python daemon [08:22] kvalo: use the 'logging' module :-) [08:22] it's highly runtime-configurable [08:22] we use it in zeitgeist and I think it's ok [08:22] kamstrup: cool, I'll use it [08:22] kamstrup: thanks [08:23] kvalo: there are a few different ways to use it trhough [08:23] kvalo: many use the static API, eg. logging.debug("foo") [08:23] +1 on the logging module :) [08:23] kvalo: but I prefer to use a logger instance, since that allows to have distinct by-component logging [08:23] kamstrup: btw, did you already used one that more logger, TBH? [08:24] kvalo: so something like import logging; log = logging.getLogger("my.service") [08:25] log.debug("W00t!") [08:25] than* [08:25] didrocks: yes, in zg we do [08:25] kamstrup: I never felt the need, but maybe for big project… [08:26] kamstrup: this is for connman-mock and I foresee that I don't need more than one with this daemon [08:26] kvalo: I'm sorry, I didn't get time yet for the review yesterday. It's ok now, can you pastebin the link please? [08:27] kamstrup: what about g_debug()&co in pygtk apps? [08:27] kvalo: nah, don't go there [08:27] kvalo: also you can't easily configure the g logging framework [08:27] didrocks: no worries. give me a minute, I'll find a link [08:28] kamstrup: but what if a library uses g_debug but the app itself something else? [08:28] kamstrup: an academic question, but anyway :) [08:29] didrocks: https://code.launchpad.net/~kvalo/indicator-network/settings-merge/+merge/43183 [08:29] kvalo: that's not an issue [08:29] kvalo: thanks, looking at it now :) [08:29] didrocks: it's already merged, but I would appreciate input from you anyway. thanks a lot for looking at it [08:30] kamstrup: yeah, I just need to configure the dump to the same place, eg. syslog [08:31] kvalo: of course :) [08:31] kvalo: no worry! my pleasure [08:49] kvalo: are you sure of your link, wasn't it for the packaging? [08:50] didrocks: sorry, I meant autofoo packaging. not debian packaging [08:50] oh ok :) [08:52] kvalo: just a comment on ins_PYTHON = bluetooth.py __init__.py mobile.py wired.py wireless.py [08:52] kvalo: -> will better be each module on its own line [08:52] less noise when you add/remove a file, and such [08:53] apart from that, all look good :) [09:10] didrocks: sorry, I had a meeting. hmm, I was supposed to fix that all in one line issue. I'll check again [09:10] didrocks: thanks for the review! [09:10] kvalo: maybe the merge proposal wasn't updated :) [09:10] kvalo: you're welcome === smorar_ is now known as smorar [10:02] hey all [10:04] coz_: hi [10:04] kvalo, hey guy [10:36] is anyone here interleaving g_bus_own_name() with g_bus_unown _name() successfully? I get a weird error where my user data is not passed correctly on the second invocation of g_bus_own_name() [10:37] (there is no idle in between, and that's not gonna happen) === smorar is now known as smorar_ === MacSlow is now known as MacSlow|lunch [11:44] ah oh unity is slightly broken? [11:53] coz_: why do you say so? ;) [11:56] nnnaji, with the current update in ubuntu... including unity updates + compiz updates... first thing I noticed is that desktop cube cannot be enabled with unity... only desktop wall... second I noticed that there is no longer gnome-wm --replace to fix the nux issue with white upper panel on dual monitors with nvidia [11:59] current screenshot http://dl.dropbox.com/u/132551/screenshots/currenunity.png [12:01] also autohide launcher is not working properly either [12:02] eeww I really dont like desktop wall :) === ssj6akshat is now known as akshatj [12:06] coz_: yeah i have some more screenshots of how unity never wakes up properly from Suspend [12:07] i'm collecting information and if i have enough i might report a bug [12:07] nnnaji, oh ok [12:07] ok [12:07] same thin [12:07] g [12:07] white top panel, white dock [12:07] nnnaji, yes that is irritating enough ...but desktop wall is painful [12:07] :) [12:12] if at least each workspace could have its own wallpaper, that would make them identifiable.. [12:12] natty will resolve the cube issue, i'm confident [12:12] nnnaji, well when someone packages the plugins extra that shouldnt be an issue [12:12] then you would have wallpaper plugin [12:13] coz_: ok, thx ;) [12:13] http://i53.tinypic.com/nqy4bt.png [12:13] that's how i get stuck after suspend [12:14] whoa [12:14] it still works, but every click takes about half a minute to create a response [12:14] nnnaji, see if you can run gnome-wm --replace & disown in terminal [12:14] ok, that's what i'll do next time [12:15] coz_: so far, my approach was to killall gnome-session :P [12:16] nnnaji, mm I know the gnome-wm --replace worked yesterday for remedying the white panels [12:16] coz_, now i'd like to have a way of saving the state of my session, a "freeze" method or something.. [12:17] then it wouldn't be so painful to test the "gnome-wm --replace & disown" [12:17] ah ok [12:30] ronoc: hi. again a small review: https://code.launchpad.net/~kvalo/indicator-network/settings-backend-begin/+merge/43635 [12:30] sure [12:37] kvalo, done [12:42] ronoc: thanks again! === MacSlow|lunch is now known as MacSlow [12:48] njpatel: parteeey! [12:49] njpatel: finally all dee unit tests pass again [12:49] njpatel: had to fight some gdbus quirks+bugs to get there [12:49] but I'm there :-O) [13:05] kamstrup, WOOHOO! [13:11] njpatel: the wicked thing is that dee now support complex column types [13:11] njpatel: eg. dee_model_set_schema (model, "s", "u", "as", "(asaasyb)", "b", NULL); === bregma__ is now known as bregma [14:09] kamstrup, no way... [14:09] kamstrup, dude that's just sexy [14:14] klattimer: ping? [14:15] dbarth: hey [14:16] klattimer: hi karl, feeling better? [14:16] not really [14:16] still feverish :/ [14:36] i would like to help with unity [14:37] is there something that i could do here ? Am interested in the bug fixing aspect, i have not done packaging as such [14:42] G__81, hi, that's awesome, we've got some docs on contributing bug fixes here http://unity.ubuntu.com/getinvolved/ [14:44] thanks njpatel, nice meeting you. You work for Canonical right ? i have read some of your blog posts as such. Great job [14:45] G__81, yeah, thanks :) [14:45] morning [14:45] njpatel, i ll read through the docs and get back in few minutes [14:45] lamalex, g'morning [14:46] sounds good [14:47] njpatel: here's something you'll find "sexy"... https://code.launchpad.net/~unity-team/dee/gdbus/+merge/43648 [14:47] njpatel: 5000 lines of sexy! [14:48] kamstrup, hah, you win. I'm now sure what you win, but you win it. [14:48] njpatel: and note ninja-leetness 5000 lines exactly! [14:48] kamstrup, 10857 lines (+4010/-3787) 42 files modified (has conflicts) [14:48] njpatel: so I guess a cursory review of the API is what's needed really [14:48] and then if there are obvious bloopers [14:49] njpatel: you get 10857 lines?! I have 5k here [14:49] kamstrup, that's lines changed, ~4 new ~4 deleted [14:49] er, +1 on both [14:50] anyway, taking a look [14:55] kamstrup, if you don't mind, I'll review this beast in the evening, over a cup of tea [15:10] are these the bugs in unity https://bugs.launchpad.net/unity [15:10] i dont know Vala. I know C and i know Python to an extent [15:10] G__81, those and also https://bugs.launchpad.net/ubuntu/+source/unity [15:11] lamalex, so does it involve knowledge of Vala ? [15:11] G__81, not really [15:12] Places can be written in Vala [15:12] but Unity itself is C++ [15:16] G__81: welcome! Check the two links at the end of the topic to learn how to get started [15:16] those are smaller bugs for new contributors [15:19] njpatel: sure no worries - it's you and kenvandine that are in need of it - so unless kenvandine is panicking there is no rush :-) [15:21] ha [15:21] 5000 lines exactly [15:21] :-D [15:23] kamstrup, ninja-leetness -1 since it doesn't build on natty :-P [15:25] njpatel: kenvandine reports that the build is broken on natty [15:25] dee-model.c: In function ‘dee_model_build_row_valist’: [15:25] dee-model.c:401:9: error: passing argument 3 of ‘g_variant_new_va’ from incompatible pointer type [15:25] /usr/include/glib-2.0/glib/gvariant.h:232:33: note: expected ‘struct __va_list_tag (*)[1]’ but argument is of type ‘struct __va_list_tag **’ [15:26] njpatel: ^^ [15:26] kamstrup, nice [15:27] just needs a cast [15:27] njpatel: this is about as far from "nice" as we can get [15:27] so I have a va_list args; [15:27] then pass it to a function expecting a va_list* [15:27] so I do fct(&args); [15:28] kaboom! ^^ above error [15:28] (on natty) [15:28] that's the problem in a nutshell [15:28] I'm amazed that glib compiles at all [15:29] there may be some nasty casting magic required to get a pointer to a va_list, because strictly speaking that's not possible afaik [15:29] (because a va_list can be a compiler-dependent macro) [15:29] yep [15:29] * njpatel looks closer [15:31] kamstrup, http://groups.google.com/group/comp.lang.c/msg/8657dc15d7cb0989 ? [15:33] njpatel, i just warned kamstrup of the fun to expect from GI stuff after we get dee to build on natty [15:34] Heh, I'm sure kamstrup will love that [15:34] he completely adores javascript [15:39] njpatel: so you're telling me the va_list impl in gcc changed entirely in natty? [15:39] don't make me cry [15:39] kamstrup, I'm still looking and actually don't know. It's a bit scary [15:40] njpatel: i wonder what desrt has to say [15:43] so the question is can g_variant_new_va() work at all, since it takes a pointer to a va_list? [15:46] kamstrup, the copy method should work, right? === ubot5` is now known as ubot5 === Pendulum_ is now known as Pendulum [16:44] G__81: were you able to find the information you needed? [16:47] i am looking into the page, trying to find whether i could add some value to the project [16:48] :) [16:48] by fixing bugs of course :) [16:56] G__81, welcome! :-) [16:57] thanks jono :) [17:00] but just wanted to know should i need a VM to test all these bugs after fixing ?? [17:01] G__81, VMs often don't work as well as you need 3D support for Unity [17:01] you may want to consider getting Natty set up on a USB stick or different machine [17:02] i dont have an other machine unfortunately so the USB stuff would that work ? [17:05] njpatel: eeek va_copy is C99 === dbarth_ is now known as dbarth-afk [17:53] how to restart bamf ? [17:53] (I forget it) [17:54] njpatel, ? [17:59] for interested, libbamf can't get the .desktop file of emacs [17:59] I will report that bug [18:24] hi guys! is there anybody willing to review my branch for bug #683466?? [18:24] Launchpad bug 683466 in unity (Ubuntu) "There are 2 Quit menu options in Quicklist" [Low,Triaged] https://launchpad.net/bugs/683466 [18:24] https://code.launchpad.net/~cando/unity/fix683466/+merge/43372 [18:29] cando, I am sure someone can help === kklimonda1 is now known as kklimonda [18:29] DBO? [18:29] ^ [18:29] jono? [18:29] yeah sure [18:29] read up [18:29] I got it [18:29] :-) [18:30] cando, what is your contrib agreement status? === kklimonda is now known as Guest60816 === Guest60816 is now known as kklimonda [18:31] cando, so this will work for english [18:31] awesome... [18:31] cando, so I am looking at your branch [18:31] your approach is a good start [18:32] but it only works for english [18:32] you need to compare the actual label on both menu items [18:32] not just where I indexed it [18:32] also please prefer the non-standard Quit over our jiggered in quit [18:33] its possible the application will have implemented a smart quit routine and not made it work right if you send it a close event [18:33] DBO, ok..thanks!i'm going to fix these thing...:) [18:33] awesome [18:33] cando, have you sent in a contributors agreement? [18:34] DBO, nope..^^ [18:34] cando, okay you will need to do so before I merge [18:35] DBO, ok...what/where should i agree ? [18:35] DBO: how does the patch look like? [18:35] sounds like apps should be fixed as well [18:35] dbarth_, minor issues but hes almost there [18:35] DBO, http://www.canonical.com/contributors ? [18:35] dbarth_, we cant force apps to fix, so this patch at least prevents identical strings [18:35] but i guess it's reasonable to catch the obvious duplicates in the launcher code as well [18:35] cando, yeah thats what I was looking for [18:36] DBO, ^^ good [18:36] cando: please send the agreement to the official canonical address, for administrative reasons [18:36] cando, please cc me (jason.smith@canonical.com) and david (david.barth@canonical.com) on the agreement [18:36] cando: and cc Jason on the submission [18:36] right [18:36] so that we can start reviewing the patch and all [18:37] DBO, dbarth_ Consider it done! [18:37] it just helps us get the process running in parallel; tomorrow the form should have been processed anyway [18:37] dbarth_, that canonical contrib page could use updating [18:37] bamf and nux should probably be added [18:37] cando: thanks for the contribution anyway [18:37] lamalex: bamf, not nux, nux is (c) inalogic [18:37] ah ok [18:37] dbarth_, :) [18:41] hello [18:41] are there any plans to include a perl API for the Application Indicator? [18:44] evaluate, no [18:44] evaluate, not that they are not welcome, but they serve no benefit to us right now [18:45] indicators speak dbus though [18:45] you dont use the dbus api directly [18:45] there are C libs to use [18:45] right, but you COULD use dbus right? [18:46] not really [18:46] just like you *could* use the BAMF dbus api [18:46] its not stable [18:46] and we dont support it [18:46] but you COULD [18:46] sure [18:46] ok [18:46] apologize [18:46] never [18:46] you're ugly [18:46] now you're asking for it [18:47] what are you guys to do? eat some spinach to make your bench grow? [18:47] hmm. are there really that many more python programs then perl ones? [18:47] that people use [18:48] python is a drop in replacement for perl in much the same way that Optimus Prime is a drop in replacement for a truck [18:48] lol [18:48] Perl is used heavily in scientific applications however [18:48] god I wrote so much of it back in the day... [18:48] but the trend there was to see perl replaced with python, that was 3 years ago [18:49] and I was too young to legally drink... [18:49] ah the good ol days [18:49] python is ugly [18:49] not as ugly as DBO [18:49] evaluate, yeah but you can tell if its been run through an MD5 algo or not [18:49] anyway, is there any documentation about the messaging menu (or the appindicator menus in general)? [18:50] evaluate, for the C API? [18:51] DBO, ohh, I'm sorry, I don't know _anything_ about python, I have just looked through some code (mainly to adapt some of the appindicator applets to my needs) and it's just ugly compared to anything else I've seen (C, C++, Perl, PHP, whatever) [18:51] it just doesn't seem to follow any logic [18:51] DBO, yeah, preferably [18:51] oh I am just having fun dude [18:51] trust me this is a rare case of me not ripping python a new one for not being able to run a ****** thread properly [18:52] evaluate, hold on trying to find the documentation, ted, any help? [18:52] ted is on holidays [18:52] https://wiki.ubuntu.com/DesktopExperienceTeam/ApplicationIndicators [18:52] ^ [18:52] evaluate, http://people.canonical.com/~ted/libappindicator/current/AppIndicator.html this too [18:53] it seems google can read it, maybe you can too [18:53] https://wiki.ubuntu.com/MessagingMenu [18:53] seb128, DBO, I have seen both of those, I just couldn't find any documentation about how to make a new entry in the messaging menu for example [18:54] those two only show how to make a completely new icon as far as I have seen [18:54] oh [18:54] messaging menu... [18:54] uhm [18:54] where is ted... [18:54] yeah where is ted [18:56] see the url I copied just before [18:56] it's about the messaging menu [18:56] seb128, I'm going through it right now [18:56] ted is on holidays as said 10 lines before [18:56] until eoy [18:56] seb128, for some reason I keep failing to read what you say today [18:56] ;-) [18:56] brb [19:00] ok, that last link seems to have some useful information. One last question, did you have any reports about the fact that using custom icon paths in C would yield blurry icons? (not sure if this is related to awn, or to the Application Indicator in general) [19:03] If I set a stock icon (like 'indicator-messages' for example) it looks fine, but if I set a custom path, like '/path/to/icon.png', it looks really blurry, like if it would be resized to 16x16 or something and then back up (I use a 32 px large dock). Like I said, I didn't test this in the gnome panel yet, just in awn [19:03] evaluate, not that I know about but feels free to open one with a test case if you have one [19:03] I thought that maybe the indicator does a resize of the icon automatically if it comes from a custom path... [19:03] evaluate, https://bugs.launchpad.net/ubuntu/+source/indicator-application/+bug/533439 [19:03] Ubuntu bug 533439 in indicator-application (Ubuntu) "Application Indicators are using fixed icon size" [Low,Triaged] [19:04] seb128, ok, I'll investigate this in gnome-panel too and see if it applies there and open a bug [19:04] not sure if that's the exact same issue though [19:04] seb128, looking at it right now [19:05] going back on the perl binding question without trolling [19:05] the current indicator team doesn't plan to work on it since that's not something requested a lot [19:05] but if someone want to works on those there is no reason they could not be accepted [19:06] out of curisity...no more gnome-wm --replace available in current updates [19:06] seb128, regarding that bug report, it seems to be the same issue. AFAIK awn resizes the icons automatically to fit the dock size, so if the indicator uses 22x22 icons, that would explain the problem. I wonder why it's low priority though... [19:07] cozziemoto, did that work before? [19:07] seb128, yep it worked fine and I had to use it because of dual monitors [19:07] seb128, when enabling twinview the upper panel turned white and gnome-wm --replace started nux again ...I believe and fixed the issue [19:07] evaluate, likely because people set the priority considering that the main user of indicators is the gnome-panel and unity-panel [19:07] which don't have that issue [19:08] cozziemoto, unity --replace? [19:08] mmm let me try on the unity machine hold on' [19:08] seb128, I have to restart x on that machine...apparenlty nvidia-settings is causing issues [19:09] seb128, well, they do have the problem, but it's not the same. As in, they don't see the resized icon, but the small one. I would find that equally annoying :p [19:09] I'm not the right person to argue over bug settings ;-) [19:09] I also noticed on current updates the launcher no longer auto hides unless a maximized window is opend however the maximized window now only opens "under" the launcher panel [19:09] talk to ted when he's back from holidays ;-) [19:09] hey evaluate, out of curiosity what app are you looking to port to appindicators thats in perl? [19:10] cozziemoto, you can turn off the autohiding in ccsm if you want [19:10] seb128, no i want auto hide but it no longer just hides it only hides when i window touches it but now all windows open beneath the launcher [19:11] not sure what is your usecase [19:11] it's being smart now [19:11] if nothing needs the space it's displayed [19:11] if you claim the screen space it goes hiding [19:12] in which case that doesn't work for you? [19:12] "smart" and "intelli" are now functions of DBO. Think about that for a minute. [19:12] seb128, previous to yesterday.... when autohide was ticked it was just always hidden until you mouse over upper left corner [19:12] right [19:12] it's smarter now [19:12] if just hides if you need the space [19:12] jcastro, I'm not looking to port anything particular, my problem is that I'm looking forward to adopt the application indicator, but most of the stuff doesn't really fit my needs (for example, before I had an add-on in thunderbird that showed me the number of unread mails in the icon in the taskbar) so I would like to write some small apps to get some of that functionality back. It is very hard for me to do all of that in C (for example to get un [19:12] read emails from IMAP through C), that's why I would welcome a perl API [19:12] why hiding if you don't put anything near it [19:13] it's not claiming space you need [19:13] seb128, well it is in fact opening on top of everything [19:13] and everything is opening under the launcher [19:13] well it was as well before [19:13] no it was noty [19:13] all windows opened above the launcher when maximized [19:13] it was pushing softwares placed on the left of screen? [19:14] jcastro, ohh, and like I said, I don't know any python. =) [19:14] evaluate: thunderbird is getting real message indicator support soon, but yeah, I get your point [19:14] cozziemoto, can you describe some steps for your issue? [19:14] seb128, right now...for example...firefox...at maximized opens underneath the launcher panel so that much of firefox is hidden [19:14] cando, you still working? [19:14] seb128, simply open firefox maximized [19:15] the launcher autohide [19:15] it's just displayed when nothing is on the left of the screen [19:15] you might hit a bug... [19:15] jcastro, well, thunderbird actually has some kind of indicator support, the gets it into the messaging menu, but I found that icon that showed the number of unread emails really useful, I don't really like the messaging menu... [19:16] seb128, I understand that but these are 2 issues... one...the launcher when put into autohide would always be hidden...now it is not...before firefox opened on top of the launcher panel now it opens underneath it [19:16] yeah we talked about this a bit with the tbird guys, like if you could have a number of new mails etc. on the launcer [19:16] er, launcher [19:16] cozziemoto, well, when firefox open the launcher should hide since firefox claim the space... [19:16] that's what it does here [19:16] seb128, it does noty [19:16] not [19:16] DBO, ^ [19:16] seems a bug [19:17] yes smspillaz said to but DBO :) [19:17] bug [19:17] can you describe the bug in clearer terms [19:17] I dont understand what you are trying to describe [19:17] Hi, would i just set up a Unity development environment for the first time. will i still need to get and build a customized version of Compiz if im running on a ubuntu nbr? [19:17] it's time like this when I wish we had screen recording built into unity [19:18] jcastro, right now for example I use a "program" that consists of a bash script (that executes all of the other stuff), a perl snippet (which gets the number of unread mails from my IMAP accounts) and a C program (that generates and updates the indicator icon), and I find this really, REALLY ugly... [19:18] cozziemoto, ok in fact I can confirm [19:18] DBO, the launcher doesn't go into autohiding when firefox open [19:18] okay [19:18] DBO, well ...previously...2 days ago... the panel set to autohide did just that at all times until moused over...now it does not....the second issue is that ,,,for example...firefox opend at maximized opens beneath the launcher and the launcher does not autohide unless i take firefox out of maximized and then re maximize it [19:18] so intellihide is failing the intelli-test === akshatj is now known as akshatj|bed [19:19] right [19:19] it also open beneath the launcher initially [19:19] like it doesn't test on application start [19:19] seb128, okay will fix [19:19] cozziemoto, can you open a bug about it? [19:19] seb128, sure [19:20] seb128, if this were closed source, I would just add a 1 second timer to double check [19:20] it will be opend under coz- [19:20] cheating FTW? [19:20] lol [19:20] is it laucnhpad.net/natty? [19:20] I'm sure you would wake up a night feeling bad about it ;-) [19:20] seb128, not as bad as I would about shipping it broken [19:20] cozziemoto, https://launchpad.net/unity/+filebug [19:20] ah yes thanks [19:22] cozziemoto, in fact it's already open [19:22] bug #686419 [19:22] Launchpad bug 686419 in Unity "Intellihide: Fails to hide when restoring an overlapping window" [Medium,Confirmed] https://launchpad.net/bugs/686419 [19:22] cozziemoto, so no need to open a new one [19:22] ah ok cool [19:22] hum [19:22] DBO, ^ can you check the fix there will fix the start issue as well? [19:23] now if sam would enable desktop cube again in unity i would be a happy man :) [19:23] I had the cube on for a while with unity [19:23] it's still my preferred way to roll [19:24] yeah but sam said it was intentional to use wall,,,and unity is disabled with cube [19:24] cozziemoto: +++..infinity ;) [19:24] :) [19:25] seb128, what? [19:25] seb128, yeah I will handle all the intellihide bugs tonight [19:26] DBO, bug #686419 [19:26] Launchpad bug 686419 in Unity "Intellihide: Fails to hide when restoring an overlapping window" [Medium,Confirmed] https://launchpad.net/bugs/686419 [19:26] DBO, there is a merge request for sam on it [19:27]