=== pochu_ is now known as pochu [01:24] can anyone help I want to get a win game to work in wine can anyone help? === rickspencer3 is now known as rickspencer3-afk [02:15] mpt - tut tut, spaces in folder names ;) [02:19] chrisccoulson: Totally agree. [02:19] I think spaces should be invalid characters in file names. [02:20] heh. i was referring to the current discussion on UF. have you seen that? [02:33] chrisccoulson: I'd comment on this discussion but it's dangerously close to feeding trolls ;) [02:34] chrisccoulson: no haven't seen it. [02:34] TheMuso - http://ubuntuforums.org/showthread.php?t=1239982 [02:35] meh forum? I don't care. [02:35] There is only a couple of forums I frequent, and they are not Linux related. Most Linux forums are too bloated, with too much crap to wade through. [02:35] basically, the short story is "developers are all stupid and ubuntu one is useless because someone put a space in the folder name" [02:36] The two forums I frequent are very slim and allow you to find what you want fast. [02:36] oh [02:36] i'm not sure if the discussion is serious or not [02:36] yeah, i'm not a big forum fan either [02:41] using spaces in dir names should not be done by the OS, but thats just my opinion (and apparently the opinion of ~ 62% of UF) [02:41] using them even in directories users create isn't a particularly good idea [02:41] why? [02:41] IFS is a pita to use [02:41] ccheney: but most users shouldn't really care about terminal and problems connected to it [02:41] if you don't know what IFS is look it up in man sh [02:42] most users don't use linux [02:42] only something like 4% of users do [02:43] Half the problems in shell scripts related to quoting arrise from characters that shouldn't be allowed in file names IMO. [02:43] tab complete works fine for just simple command line stuff, but using scripts would be a pita [02:44] ccheney: fwiw I think that people who are going to use terminal to mess with files in their ~/ are most likely smart enough to figure out how to work around space issues [02:44] TheMuso: yep, its a pita to deal with invalid characters in names [02:44] kklimonda: even using IFS correctly (which is what you have to do) can be problematic at times [02:45] unless you end up wanting to write your scripts in perl or python i suppose, but yea spaces really cause pain for shell scripts [02:45] I even remember seeing bugs filed against programs in C that behaved badly with spaces in file names. [02:45] ccheney: I haven't had any problems with spaces sine I've stopped using ls ;) [02:45] but then I use python for most scripts more advanced than iterating over all files in folder [02:46] ccheney: folders with spaces looks nicer, more natural and that's it imo. [02:47] ccheney: should we also ban parenthesis and exclamation mark? [02:48] Yes. [02:48] square bracket should also be gone as it can cause problems for some users.. [02:48] IMO [02:49] I think the only characters that should be allowed are upper/lowercase leters, numbers, dots, dashers, and underscores. :) [02:49] I kinda thought the same few years ago [02:49] I had folders named like Machinae.Supremacy.-.Arcade_2008 [02:49] but now I think it looks ugly ;) [02:50] And looks are all that is important of course. [02:52] TheMuso: I'm certainly not going to name my files and folders to easy developers' job. [02:54] Each to their own. [03:11] well i know for certain you can't iterate over a directory with a space in the name without munging with IFS at minimum [03:11] kklimonda: users can name their dirs whatever they want, apps shouldn't use spaces in their own stuff though, which was what my argument was, if the user wants to inflict pain on themselves so be it ;-) [03:12] ccheney: for i in *; do something; done works just fine [03:12] doesn't if there is a space in the dir name [03:12] or at least not with normal shells [03:12] oh maybe it might work for the special case of * [03:13] ccheney: you hav eto quote it later obviously but it works [03:13] but something like for i in `generated list of dirs` won't [03:13] ccheney: the only problem is when people do something like for i in `ls .`; do; done [03:13] i've had the problem not with ls . but in other generated lists of dirs before [03:14] since the space is the considered the separator by default, which is why i said maybe * does something special i didn't know about [03:15] * ccheney is getting hungry, off to find food === Amaranth_ is now known as Amaranth === tkamppeter_ is now known as tkamppeter [07:32] Hi im having a problem with the volume, I have ubuntu 9.04 jaunty and realtek hd integrated audio [07:32] the volume is EXTREMELY low [07:32] I have tried alsamixer and of course the gnome volume control applet neither of which work [07:33] Alexia: #ubuntu for help [07:34] ok [08:04] hey glatzor :) [08:05] morning mvo! [08:20] mvo, In the latest PackageKit apt backend I implemented a safe upgrade feature, which allows to install also upgrades which would require the installation of further packages [08:20] glatzor: nice [08:22] glatzor: btw, is there a way to attach additional meta-data to a aptdaemon transaction? something like the gboject.set_data() (or will that work)? so that I can create the transaction in one place, add meta-data (like a app-icon and the name (that may be != pkgname)) and get that from the GetTransactions command later? [08:23] (I could of course use a tid: metadata mapping, but it would be cool to have it directly in the transaction) [08:24] The AptTransaction is a gobject.GOBject [08:29] glatzor: what I mean is that I create a transaction with e.g. "commit_package()", attach data and later (in a independant part of the app or even in a seperate one) use ActiveTransactionsChanged to get the data again. I will have two different objects, so there needs to be some dbus magic I guess to make that work ? [08:30] we could also use a singelton transaction cache in the client module [08:31] mvo, the client.get_transaction(tid) could look up the the transaction in a cache [08:31] glatzor: yeah, I played with that and added a very simple cache [08:32] glatzor: I just need to figure out when to destroy them again - do you want me to commit my approach or do you have something (almost) ready already :) ? [08:32] glatzor: that will also solve (partly) the problem that the bus limits for the matchers is pretty low [08:34] I would like to kill the transaction on the server side. but I haven't looked at removing them manually from the bus. [09:37] morninge everyone [10:12] kenvandine: are you susbcribed to xsplash bug tracker? [10:12] mvo: welcome back man"!! [10:12] hey asac! [10:12] apt-get dist-upgrade is crashing for me ;) [10:12] hehe [10:12] *wehh* [10:12] asac: segfault? [10:12] it started on saturday ;) [10:12] yes [10:12] let me check if its still the case [10:12] i definitly have a .crash file still [10:12] asac: does it make a difference if you move away /var/cache/apt/*.bin ? [10:13] asac: please file a bug with it [10:13] mvo: i cannot file a bug with it. [10:13] apport refused to do that [10:13] we need a --force option there :/ [10:13] yep [10:13] I had the same problem in the past [10:14] too bad it doesnt crash anymore. not sure what changed. it crashed all the time the other day (even after reboot) [10:14] let me copy the .crash file somwehere [10:14] mvo: how does it feel to be back to work? [10:15] asac: I wouldn't have minded to stay on leave for another two weeks ;) - but its also nice to write code again :P [10:16] hehe [10:16] i can guess that [10:16] mvo: https://chinstrap.canonical.com/~asac/_usr_bin_apt-get.0.crash [10:16] guess you need to apport-unpack and then gdb against it [10:16] i think it was some kind of database state [10:17] what a shame that it doesnt work anymore [10:17] i guess it was update-manager in the background updating the apt databases [10:17] You don't have permission to access /~asac/_usr_bin_apt-get.0.crash on this server. [10:17] * asac wants to uninstall it ;) [10:17] :( [10:17] mvo: can you see my home? [10:17] mvo: you should be able to log in through launchpad et al [10:17] yes [10:17] that works [10:18] let me see if I can login [10:18] mvo: try again [10:18] mvo: it rememberd the minimal permissions of the local crash file [10:18] have it now [10:18] permissions I guess [10:19] * asac hopes that something gets out of it ;) [10:20] * asac cannot killall apt-get [10:20] cannot ctrl-C it either :) [10:20] its a remote ssh thing though [10:29] asac: ahoy, so kubuntu is probably going to ship a simple installer appliation for firefox, so it becomes easier to obtain firefox, while we will stick with konqueror by default (at least for now that is) ... but what is needed to use the official firefox logo? get permission from trademarks@mozilla.com? also, since it is not any kind of free, wouldn't it need to go to restricted anyway? [10:32] .oO ( simpler than 'apt-get install firefox' !? ) [10:44] asac: hrm, hrm, bt is full of ?? [10:48] mvo: about the gnome-codec-install bugs, they are just bugs right ,as in not working properly rather than ,"that was how they have been designed" ? [10:48] mac_v: I did not had a chance to test it yet, but yes, it sounds like bugs [10:49] mvo: hmm... ok... sure no probs, i just wanted to invalidate the papercuts tasks associated :) [10:54] mvo: too bad [10:55] apachelogger: a simple installer application? [10:55] apachelogger: what does that mean? [10:55] why dont you make that install the packages? [10:56] just use the packages from the archive and you should be fine wrt to trademarks (assuming we will get approval for our slight adaptions this time again ;)) [10:57] asac: http://aplg.kollide.net/screencasts/kubuntu-firefox-installer2.ogv [10:58] apachelogger: so you install the package? [10:58] aye [10:58] firefox [10:58] apachelogger: should be ok. one question we should figure though is if -gnome-support package would work in theory on kubuntu [10:58] otherwise firefox has no xdg mime database integration [10:59] which is a bit of a suboptimal user experience [10:59] asac: IIRC it does [10:59] apachelogger: then be sure to install -gnome-support package too [10:59] or i can make that a recommends again [10:59] (if thats not the case) [10:59] currently it installs firefox with all recommends, so that gets in [11:00] apachelogger: i will check with my trademark contact if he has anything to say about the use of the firefox logo in the installer ... but i wouldnt think its a problem [11:00] apachelogger: i think i moved it to suggests after getting pressure from the kubuntu folks [11:01] hm [11:01] if the integration finally works [11:01] e.g. kde admin tools modify xdg database in such a way that gnome-support works [11:01] we can move it to recommends [11:01] again [11:02] need to take a look at that [11:02] also you dont install ubufox ... which is a suboptimal expereince on ubuntu ;) [11:02] mvo: ... qt frontend for apturl :) [11:02] bug bug [11:02] well, I am quite sure it installs ubufox at this point [11:02] oh [11:02] right i think i didnt demote that yet ;) [11:02] though as it looks we will get started on a kubuntu firefox [11:03] yep 3.5 has recommends [11:03] apachelogger: someone needs to write a qt frontend for this apturl [11:03] mvo said its fairly trivial [11:03] with that all that synaptic stuff wont get pulled in again [11:05] yeah, I was looking at it some time ago, seemed easy enough to do, but i was lacking motivation ... but since I am even considering firefox to become default for the kubuntu netbook edition it seems like a good idea ;-) [11:06] i think its the right decision to make firefox easier accessible for kubuntu users [11:06] apachelogger: this installer entry in the menu ... will it go away when firefox is instlaled? [11:06] asac: that is the idea ... though I need a sensible approach to it [11:07] yeah [11:07] the installer package can't simly conflict firefox, because that would most likely remove firefox in favor of the installer upon upgrades to 9.10 [11:07] though it works just fine in a non-upgrade scenario :) [11:08] maybe putting a Replaces: ...-installer in the firefox package would help [11:08] and using the same .desktop file name [11:08] not so sure how that behaves or if there are bugs in dpkg that would cause issues on upgrades of -installer pcakage later on [11:09] well, I think conflicts+replaces in firefox would overrule the conflicts in the installer [11:09] so as I see it once firefox is installed the installer would not get in at any point, since both declare conflicts but firefox actually replaces the installer [11:11] apachelogger: what i meant was a Replaces without a Conflicts ;) [11:11] and installer/firefox sharing same .desktop file name [11:12] asac: well, they do not need to share the same desktop file to have a replaces ;-) [11:15] If one package is to be installed, the other must be removed first - if the package being installed is marked as replacing (see Overwriting files and replacing packages - Replaces, Section 7.6) the one on the system, or the one on the system is marked as deselected, or both packages are marked Essential, then dpkg will automatically remove the package which is causing the conflict, otherwise it will halt the installation of [11:15] the new package with an error. [11:15] asac: I suppose a replaces in firefox would work :) [11:19] apachelogger: what i meant was a Replaces: without a Conflicts [11:20] so the -installer package stays on system, just the .desktop file is dominated by the firefox package one if instlaled [11:20] oh [11:21] asac: though, that would not have much benefit over having the package removed, would it? [11:22] apachelogger: if users remove firefox, they get the installer back ;) [11:22] but no. there is not much of a benefit [11:22] except that you dont need to worry about apt suddenly thinking the -installer package might be more important [11:24] k [11:27] asac: please poke me once you hear back from the trademarks contact, Riddell tells me I need a package dispensation from TB to get the package with firefox logo into main, since the logo is non-free [11:28] I'll start the main inclusion process with the globe icon from mozilla vcs for the time being [11:39] bratsche: do you know about xsplash? [11:50] apachelogger: yes. do that [11:50] i wouldnt expect the icon to be a problem [11:50] apachelogger: actually the icon is already shipped inthe app-instlal-data package afaik [11:50] just refer to that one [11:50] if the size is ok [11:59] Riddell: I think bratsche does, didn't he wrote most of it? [11:59] asac: apturl UI> no time yet :( [12:01] apachelogger: ^^ ;) [12:01] mvo: didnt you drop all your specs and have plenty of freetime ;)? [12:05] asac: eh, sort of ;) [12:05] asac: I got some more actually (the crashdump one suddenly appeared) [12:13] anyone here with hardy instlal? === al-maisan_ is now known as al-maisan [12:25] mvo: that's why I'm asking :) [12:30] ArneGoetje: i wanted to enable devmode today [12:30] ArneGoetje: are you there? [12:31] Riddell: :) [12:36] ArneGoetje: ok enabled devmode. so pass "devmode" as forth argument for karmic until after beta. and tell me when the first packages are in the ppa so i can verify a few [12:57] hi [13:18] asac, i am === MacSlow is now known as MacSlow|lunch [13:27] hm, what was the magic again to make Gdk-CRITICALS a breakpoint in gdb? [13:28] mvo: G_DEBUG=fatal_criticals [13:29] pochu: rock! thanks a lot [13:29] that will cause a SIGABRT iirc [13:29] mvo: np :) [13:29] pochu: yep, exactly what I need [13:29] :) [13:33] kenvandine: what was the question i asked :-P? [13:34] hehe [13:34] if i was subscribed to xsplash bugmail [13:34] the whole dx team is [13:34] which includes me [13:37] ah ;) [13:49] guys , Bug #387796 ,has been wrongly marked as fix released , since the icons wont be shown[default is to not show the icons] unless the latest patch is applied [13:49] Launchpad bug 387796 in nautilus "Display application icons for "Open With" entries in the context-menu" [Wishlist,Fix released] https://launchpad.net/bugs/387796 [13:50] mpt: ^ shall i change the status? [13:56] mac_v, yes please [13:56] ok. [13:56] chrisccoulson1, I'm surprised that kind of "omg spaces" discussion is still going on after the 1980s [13:57] mpt - yes, me too. i'm mostly indifferent on the matter really, but it seems a lot of people don't like spaces [13:58] i think the thread starter just wanted an excuse to rant though [13:58] yeah [13:58] where "a lot" < 0.1% of the user base [13:58] yeah, i agree [13:59] he did make me laugh when he said that ensuring filenames don't have spaces in is a good developers number one priority [14:01] chrisccoulson1 mpt: bug # pls :) [14:01] mac_v - we're referring to a topic on UF discussing ubuntuone, not a bug ;) [14:02] chrisccoulson1: oh... nevermind then :) [14:04] chrisccoulson1, that's common Bugzilla behavior :-) [14:13] asac: karmic langpacks go directly into the archive. I belive with devmode enabled, we will get a whole bunch of new languages for mozilla trans;ations. Do we need them to be in the -base langpacks first? Then I would need to generate new -base langpacks manually. === rickspencer3-afk is now known as rickspencer3 [14:26] Riddell: Hi. [14:26] yo bratsche [14:26] ArneGoetje: i would reall yprefer to do a ppa round ... otherwise we might end up with all firefox broken for all languages but en-US [14:26] What's up? [14:26] ArneGoetje: i just want to see how bad it is ;) [14:26] bratsche: xsplash, someone (presumably me) needs to do a KDE equivalent [14:26] asac: heh [14:26] ArneGoetje: you think you can do a manual run ? [14:27] bratsche: so I need to know what starts and what stops xsplash and any other useful information [14:27] Riddell: Bummer. [14:27] asac: I can [14:27] so yeah. we need a -base update for sure. but lets do a manual run first, so we can see how bad it is [14:27] asac: lemme see the schedule... [14:28] Riddell: It's started by gdm scripts.. if you see /etc/gdm/Init/Default and /etc/gdm/PreSession/Default those are what launch it. It listens to dbus signals from the desktop to know when to exit, and it runs on a timer for like 10 or 15 secs in the event that something goes wrong and it doesn't receive the signals. [14:28] asac: oh... there is a Launchpad export due today... I can request it to be a full export and then to a ppa run first. [14:29] ArneGoetje: that would be great. at best keep the bits in case we need to rerun (e.g. because of bugs in the script or something) [14:29] asac: of course [14:30] ArneGoetje: also please double check that you dont pass devmode to anything != karmic ;) ... but i guess thats clear [14:30] but since its going to ppa we would probably notice ;) [14:30] asac: how about the ff3.5 templates? [14:30] ArneGoetje: http://bazaar.launchpad.net/~mozillateam/rosetta/po2xpi/revision/48 [14:30] thats the commit i did [14:31] ArneGoetje: i think you have to manually upload the template once.... seems launchpad didnt eat them [14:31] i remember something vaguely of doing the first .xpi upload manually or something. maybe check with jtv [14:31] https://translations.edge.launchpad.net/ubuntu/karmic/+source/xulrunner-1.9.1 -> not existing [14:32] asac: what to do with the existing firefox template? [14:32] ArneGoetje: as we discussed. just keep it. the new one will be with full version [14:32] asac: [14:32] asac: ok [14:32] or doesnt that match what you remember? [14:33] ArneGoetje: i think we also wnated to copy those over that have no upstream .xpi [14:33] asac: it does. just want to make sure. :) [14:33] sure thats what we want [14:33] ArneGoetje: i think malay and the ones that are whitelisted should be copied over at least [14:33] thats ast [14:33] fi [14:33] and oc [14:33] afaik [14:33] PLUS malay [14:33] the whitelist is where? [14:34] in the data/.../whitelist.txt file [14:34] Hi, can anyone tell me if Karmic will have telepathy-idle (IRC) in main? [14:34] ArneGoetje: but first check what we need to do to get the template uploaded [14:34] Riddell: Doesn't KDE already have some kind of splash screen? Someone told me it does. [14:34] ArneGoetje: i can produce them manually for you if you need to upload manually [14:34] ArneGoetje: then we can copy stuff over before we do the first upstream import i guess [14:35] asac: did the ff3.5 and xul1.9.1 source packages contain the en-US.xpi? and did Rosetta eat that one? === MacSlow|lunch is now known as MacSlow [14:36] bratsche: yes, ksplashx (which doesn't have any dependencies outside of xlib) [14:36] nocturn: Seems seb128 is not here atm [14:36] ArneGoetje: i think it ate them yes. [14:36] nocturn: probably you'll have to ask when he is back [14:36] bratsche: if xsplash is started by gdm, surely there's not much time for it to be shown before gdm is up? [14:37] asac: ok. then I can manually copy over the ff3.0 translations and after that import new upstream xpis. [14:37] ArneGoetje: i just remember that jtv said something about doing the first manually [14:37] Riddell: gah.. I wish I knew about that before, I would never have wasted my time. :) [14:37] Riddell: Okay so here's the whole picture.. [14:37] ArneGoetje: right. if the template is set i think thats the right approach. but at best only copy those over that have no upstream xpi [14:38] i am not sure how launchpad behaves if you copy them over first ... maybe the upstream upload will never override anything [14:38] ;) [14:38] asac: ok, I upload the upstream ones first and then only copy over the rest [14:38] yeah sounds reasonable [14:39] ArneGoetje: do you know if the copy will drop strings that have no matching id in the tempalate (thats what we want i guess) [14:39] Riddell: We'll be changing the startup sequence in the near future so that in most cases you don't see usplash, and we get into X early in the bootup sequence. That's when xsplash gets started by /etc/gdm/Init/Default (I think, unless someone tells me otherwise).. and system startup is happening. When that's done and the gdm greeter is ready to be shown then xsplash gets a signal from gdm greeter and hides itself. [14:39] asac: should I copy over all languages w/o xpis? or only the ones you mentioned? [14:39] Thanks Zdra, will do that\ [14:39] asac: yes, not matching strings will be ignored [14:40] ArneGoetje: i think all [14:40] the ones i mentioned have high prio though [14:40] asac: ok, will do that now. [14:40] because those are whitelisted or known to have translations completed [14:40] ArneGoetje: do you have the template? [14:40] asac: which template? [14:40] ArneGoetje: its xulrunner-1.9.1 and firefox-3.5 in case you can find them in the upload queue [14:40] bratsche: why is it any slower to loda gdm greeter than it is to load xsplash? [14:40] ArneGoetje: the templates that got eaten [14:41] asac: lemme see [14:41] ArneGoetje: if not - and if we need to do the first template upload manually - we need to produce them manually [14:41] just let me know. but maybe they are just dangling in some moderation queue [14:42] tedg: did you the new libindicate version that agateau needed? [14:43] Riddell: The idea is that since we're starting X earlier, some of the system startup that was happening under usplash is now happening under X. So we start gdm greeter once all that is ready, I believe. [14:43] Riddell: But I'm not familiar with the startup sequence stuff. Keybuk, or anyone from Foundations team would know the answers to this. [14:44] Keybuk: so if xsplash is started by gdm, why not just start gdm greeter? [14:45] asac: template is in the queue, need to manually approve it [14:46] nice [14:47] asac: template name and translation domain should be "firefox-3.5" and "xulrunner-1.9.1" respectively? [14:47] yep [14:47] asac: ok... approving [14:47] * asac hits reload [14:47] https://translations.edge.launchpad.net/ubuntu/karmic/+source/xulrunner-1.9.1 [14:47] guess it takes a bit? [14:47] https://translations.edge.launchpad.net/ubuntu/karmic/+source/firefox-3.5 yay ;) [14:48] hmm 3.5 is there ... (not imported) .. 1.9.1 still missing [14:48] now there too [14:48] great. [14:49] ArneGoetje: http://releases.mozilla.org/pub/mozilla.org/firefox/releases/3.5.2/linux-i686/xpi/ ... there are the upstream .xpis [14:50] * asac checks what the po2xpi thing actually does when multiple versions now show up [14:51] should work: [14:51] translation_domain_dirs=`ls -d $SOURCE_DIR/xpi/*` [14:51] nice [14:51] ArneGoetje: ^^ ... so it should work oob ;) [14:51] if they get exported as SOURCE_DIR/xpi/firefox-3.5 etc. [14:52] Riddell: Yes, but it also involved splitting things apart and renaming things. So kenvandine was going to look at it, and then ping you. [14:52] tedg, that is next on my plate this morning [14:52] Riddell: It's done from my perspective. But it turns out the distro team has higher standards than I do :) [14:53] groovy [14:53] kenvandine: what's the crack? [14:53] kenvandine: Cool, ping me if you notice anything fishy. [14:53] asac: looking at the upstream xpis... I see there is a ta-LK.xpi ... we don't have that code in Rosetta AFAIK [14:54] asac: nor do we have a locale for that one [14:54] Riddell, anything tedg does is like crack :) [14:54] asac: (that's Tamil in Sri-Lanka, btw.) [14:55] ArneGoetje: yeah. i remember that tamil folks were happy to have a translations. [14:55] ArneGoetje: how is the process of starting a new langauge? does that require some tamil guy driving it? [14:55] asac: well, we can put it into the tamil language-pack... but users won't be able to use it without a proper locale... [14:55] otherwise i think we can skip that translations until someone from sri-lanka starts a full ubuntu translation [14:56] ArneGoetje: yeah. they can set LANG=ta-LK though [14:56] ArneGoetje: let me know how we can do /not do this [14:56] i would love to have that in ... and maybe we can enable the locale and make it fallback for other apps to normal "tamil"? [14:57] asac: hum... ok, I think that's not our problem... so I just try to import it into Rosetta first [14:57] yeah [14:57] do that for now [14:57] everything else comes later ;) [14:57] asac: ok... downloading... [15:05] tracker inclusion thread is getting too long :D [15:05] on d-d-l [15:07] kenvandine: is there a wiki or a documentation about xsplash? i cant find any info anywhere [15:08] rodrigo_ - yeah, i've seen that [15:08] what are your opinions on tracker? [15:09] mac_v, https://wiki.ubuntu.com/Artwork/Incoming/Karmic/Boot/Demo [15:09] mac_v, xsplash is just part of the boot experience [15:10] kenvandine: i'm aware of the boot/demo wiki , but no sure how to customize it or change the settings [15:11] and https://wiki.ubuntu.com/DesktopExperienceTeam/BootExperienceProject [15:11] there isn't a way to customize it yet [15:11] not sure how customizable it will be [15:12] kenvandine: well i wanted to change settings in my system , nothing major , thanx for the link :) [15:12] :) [15:12] think of it like usplash, which isn't really customizable [15:12] xsplash will be easier to change though... [15:13] kenvandine: any update on the testsuite from erlang MIR? [15:13] it just isn't defined yet, as to how [15:13] kenvandine: usplash *is* customizable , [15:13] i think we are adding a command arg to set the background image [15:13] mac_v, not easily [15:13] asac, no... i will check on that [15:13] hehe , ;) [15:14] kenvandine: no problem. just wanted to remind you that the MIR is stuck on that part ;) [15:15] kenvandine: BTW, is xsplash an Ubuntu project? [15:16] mac_v, yes, part of ayatanna [15:16] ayatana [15:17] chrisccoulson1: well, for only indexing, it seems a bit useless for me, since I never search for files the way it allows you to do it [15:17] hmm... nice :) [15:17] chrisccoulson1: but I think as a central storage for desktop apps, it makes sense [15:17] chrisccoulson1: specially if we add couchdb support to it, so that all data gets replicated :) [15:19] yeah, that sounds good. i would probably use it for indexing and searching as well though. 0.6.95 in ubuntu currently sucks a lot though, so i don't use it now, although i would if it worked well [15:20] i did start maintaing some packages of the 0.7 branch, but they've stagnated for a little while now [15:20] awe: good morning ;) [15:21] asac: good afternoon! [15:21] thx [15:21] in the middle of a mtg, will ping you when i'm done... [15:21] yep [15:25] chrisccoulson1: 0.7 is a WIP, and changing a lot, so better to wait, I think [15:30] rodrigo_ - yeah, i've been following the development of it quite closely. i started maintaining some packages to have some idea of how the packaging would evolve (seeing as it's likely to be me who packages it ;) ), and to test it and fix bugs too, but i just ran out of time in the end. [15:32] rodrigo_: Adding couchdb support to tracker would also be cool as other things (U1) could add index data there even if the files weren't brought down locally (small disk scenario) [15:32] tedg: yeah, it's ion my TODO list :) [15:32] rodrigo_: Cool :) [15:33] just need to make the TODO list shorter to get to it, but it will happen sooner or later :D [15:33] asac: upstream xpis uploaded [15:34] ArneGoetje: did the template import work? [15:34] seems so [15:34] nice [15:34] https://translations.edge.launchpad.net/ubuntu/karmic/+source/xulrunner-1.9.1 [15:42] asac: do you want to adjust the priorities for the firefox/xulrunner templates? [15:43] ArneGoetje: we can make xulrunner prio higher. [15:43] but once stuff is enabled they will see that its partly untranslated and will soon understand it i guess [15:44] asac: I mean the 3.5/1.9.1 templates should have a higher priority than the old ones for karmic... [15:44] kenvandine: not sure if you are aware of this , but xsplash flickers once or twice and shows the user gtk_color_scheme's window background color , or should i file a bug about it? [15:44] ArneGoetje: yeah ... thats for sure [15:44] we can also add a note to the old templates that those are not the current default [15:45] mac_v, we know, but a bug report would be nice [15:45] ok [15:45] thx [15:46] kenvandine: just asked since you commented in a bug saying xsplash is still in development :) [15:46] asac: ok, once I'm finished with all the imports, I will do that [15:47] thx [18:58] hey bryce, xtrace is pretty useful (I got it working now) [18:58] i can't trigger the error i'm trying to catch now though :( [18:59] heh [18:59] chrisccoulson, heisenbug eh? [18:59] it seems so! [18:59] chrisccoulson, sometimes that happens when the issue is a race condition [18:59] and i don't have much experience debugging these kinds of issues either ;) [19:00] chrisccoulson, try throwing some sleeps at your original (non-xtrace) code [19:00] yeah, i could do [19:00] thanks:) [19:01] just as i say that, it has crashed now under xtrace! [19:04] bryce - the documentation for xlib lists the errors that XGetWindowProperty can throw, and BadMatch is not one of them - but I'm seeing BadMatch randomly (and the xtrace log is showing this too) [19:04] do you know what conditions would trigger this errror? [19:04] s/errror/error [19:06] the bug i'm looking at btw is bug 321041 [19:06] Launchpad bug 321041 in gnome-settings-daemon "gnome-settings-daemon crashed with signal 5 in xkl_process_error()" [Medium,Incomplete] https://launchpad.net/bugs/321041 [19:07] and a backtrace leading up to that crash: http://launchpadlibrarian.net/30283660/Stacktrace === rickspencer3 is now known as rickspencer3-afk [19:19] heh, bratsche - this g-s-d crasher i was talking about last week is somehow exposed by xsplash [19:19] (i've got no idea why though) [19:20] Any idea what window this is that's crashing? [19:20] it calls XGetWindowProperty on the xsplash window at the start of the session, and that returns an error which is not expected. and it's always the xsplash window that this falls over with :-/ [19:20] wierd [19:21] bratsche - http://pastebin.com/m77704866 [19:21] 0x00400004 is xsplash [19:22] as shown earlier in the log here: http://pastebin.com/m338d3229 [19:22] i'm completely confused now [19:23] Yeah me too. [19:23] this is starting to cause me some sleepless nights, and the number of duplicates is increasing daily [19:23] i think it's time to rebuild Xorg with some debug statements in after dinner ;) [19:26] chrisccoulson: How can I duplicate this? [19:27] bratsche - i'm not sure. for me, it happens repeatably on login if the session fully loads with xsplash still running. that doesn't happen often though because my VM is pretty slow, and xsplash seems to disappear before the panel loads. in this case, i don't see the crash [19:27] chrisccoulson: Stupid question, but what is the klavier stuff? [19:28] it's all to do with keyboard layout profiles, and allows you to set per-window layouts and stuff [19:28] i'm not 100% sure exactly how it works, but all i know is that it is used by g-s-d in the keyboard plugin [19:28] i'm having to learn how it works though [19:29] chrisccoulson, BadMatch is a common function return #define value in xserver [19:29] chrisccoulson, there's also a similar one for xkb [19:29] ./include/xkbfile.h:#define _XkbErrBadMatch 17 [19:29] bryce, thanks. i'm not sure what the call is doing wrong to trigger it though [19:30] it seems undocumented in xlib for XGetWindowProperty [19:31] chrisccoulson, grepping through xserver code, BadMatch shows up all over the place [19:31] bryce - i'll have a look in a minute, and see if i can pinpoint where i think it might be occurring [19:32] grep returning lots of instances in libx11 as well [19:32] aha! [19:33] http://pastebin.ubuntu.com/255871/ [19:34] hmm, XErrorDB... [19:34] XProtoError.8: BadMatch (invalid parameter attributes) [19:35] bryce - am i right in thinking that when i do a GetProperty call from a client, that this is handled by ProcGetProperty() in dix/property.c ? [19:35] chrisccoulson, hmm that's a bit vague, but in poking around under the hood it looks like it might return that if for instance you passed a float instead of an int, or so on [19:36] chrisccoulson, not sure but sounds like a feasible assumption [19:36] you could stick some print statements in there to check [19:36] i can see a code path in ProcGetProperty() that could return this error, but i can't see how it should happen [19:37] yeah, i think my next step is to rebuild with some print statements [19:39] chrisccoulson, have you looked at man XGetWindowProperty? [19:39] bryce - yeah, i looked at that [19:39] BadMatch is not documented as an error that this call can return [19:40] huh, yeah you're right [19:41] looking at the code path, dixLookupDrawable is called from dixLookupWindow with type=M_WINDOW. BadMatch is returned if this type does not match the type of the returned drawable [19:41] it's mentioned for some of the other calls but not that one [19:41] but i'm speculating here now, i need to rebuild it with some print statements [19:41] yeah [19:41] ok well solid hypothesis so far though, that's good [19:42] once this is sorted, bug me and I'll make sure we update the man page and notify upstream [19:42] thanks:) [19:43] bryce: could you first ask upstream about bug 401055 ... [19:43] Launchpad bug 401055 in firefox "[MASTER] Gdk-ERROR **: The program 'firefox-3.5' received an X Window System error (BadWindow)" [Unknown,Confirmed] https://launchpad.net/bugs/401055 [19:43] its the XError I showed you in berlin if you remember ;) [19:44] i have a band aid, but would like to ask upstream why XQueryTree now gives BadWindow [19:44] heh, we're flooding bryce with X errors now ;) [19:44] if they have an idea that would be great. could be that its some window not properly cleaned and query iterating over it [19:45] mine is old ;) [19:45] asac, ok [20:04] asac, weird. Well the libx11 code is like a maze of twisty passages, I don't see where XQueryTree can generate a BadWindow. I'll just forward the issue upstream. [20:05] bryce: thanks. please add the right x package to the bug too [20:06] bryce: would like to wait with bandaid in case they want more info [20:06] asac, if you can produce a smallish test case that upstream could use to reproduce the issue, it would likely help [20:07] no way [20:07] well, otherwise, I've a sense my upstreamed bug is going to just end up lost in the pile of xserver bugs [20:07] you can ask folks directly ;) [20:07] its firefox. i dont think they want firefox to be broken like that ;) [20:08] hrhm [20:08] bryce or chrisccoulson: Do either of you know much about what's going on in compiz when it starts? xsplash starts before compiz and once compiz starts xsplash disappears and then reappears. Do either of you have some clues whether it's possible to solve this, and where I might look? [20:08] at least get a first input. if they have no clue we apply bandaid and i will let mozilla upstream deal with it [20:08] bryce: ^^ [20:09] bratsche - i'm not sure. i'm not running compiz here [20:09] but yes, you're right - xsplash is running before the window manager [20:09] which is something that the phases of gnome-session try to avoid [20:09] bratsche, are you sure xsplash is running, or that it's just whatever was left on the framebuffer when xsplash exited? [20:10] right, i've got to disappear for some dinner [20:11] bratsche: I guess it is compiz mapping the cow [20:12] bryce: I'm sure it's running.. because once the rest of the session has loaded the xsplash changes its window opacity. [20:13] mclasen: Sorry, what is the cow? [20:13] composite override window [20:13] or overlay, maybe [20:13] bratsche, then could it be xsplash and compiz both trying to update the screen and stepping on each other's toes? [20:14] sorry, I'm just throwing out random ideas at this point, I don't have any useful advice [20:14] I don't really know how window managers work yet, tbh, so I'm not sure how to fix this yet. [20:26] mclasen: Is the COW created with XCompositeGetOverlayWindow? The man page says if it hasn't yet been mapped then this will map it. [20:26] Not sure if this is the same thing. [20:27] yeah, I guess that is it, but my memory of the issues with that is vague [20:27] Cool, I'll at least look in this direction and see what happens. Thanks! [20:29] asac, done. [20:34] asac, firefox no longer is showing the busy cursor when I click form submit buttons, while it reloads the page [21:01] bratsche - have you triggered the crash yet? [21:02] Nope. [21:02] hmmm [21:33] chrisccoulson, asac, please review https://wiki.ubuntu.com/X/Troubleshooting/XError which I've just drafted to help document how to debug X client errors. [21:34] chrisccoulson, asac, since you two have gone through the process many times before, I hope you would each add some nuggets of wisdom to it. (seb128, maybe you too) [21:35] bryce - i'll take a look at that shortly. [21:36] bratsche - i'm not sure if this is coincidence or not, but i just rebuilt xsplash with a call to gtk_window_set_decorated to turn decorations off, and i'm not seeing the crash anymore [21:36] so, it seems definately related to something which happens when the WM loads [21:37] Well, fuck me. When it was doing set_fullscreen(true) then it wouldn't work right.. so when I stopped doing that and turned off decorations it fixed the other problems and now it does this. [21:39] chrisccoulson: This all seems like way more trouble than a stupid splash screen is worth. ;) [21:39] heh, yeah ;) [21:39] it does seem like a lot of effort currently === rickspencer3-afk is now known as rickspencer3 [21:40] i'm sure that the bug isn't really with xsplash - just that it does something different to expose the real bug [21:41] i'm not exactly sure how the WM works really - it reparents all the windows it decorates doesn't it? [21:41] I don't think compiz does.. but I might be wrong. [21:41] ah, i'm using metacity here [21:41] Yes, I think metacity does. [21:45] yeah, metacity does reparent - i just checked the xtrace output for xsplash with and without decorations from metacity [21:45] and i don't see a ReparentNotify for the non-decorated case [21:45] Is there some way to grep a file through bzr revisions? :) [21:45] brb, i have food to eat [21:45] does someone knows if karmic is going to ship with pulseaudio >= 0.9.16 [21:45] k [21:46] oh, it is already there it seems [22:16] chrisccoulson: I think I got rid of the flickering here. [22:16] cool, how did you manage that? [22:22] chrisccoulson: http://bazaar.launchpad.net/~bratsche/xsplash/composite-overlay-hack/revision/46 [22:23] Pull down that branch and play around with it if you have any time. Let me know if it seems to get rid of the flicker for you. [22:36] thanks bratsche - i'll have a look at that when i get the chance [22:36] probably not tonight though [22:40] that's cool. [23:02] yay! gnome-system-tools and system-tools-backends have been ported to polkit-1 [23:02] that will please pitti when he's back