[00:00] well, this was kubuntu, jaunty alpha 4 [00:05] gnome-panel isn't starting in jaunty here lately. I have to kill it to make it work(!). Any ideas how I could start to figure out the problem? [00:05] It's like it's there, just not showing because my tray items are minimized to it [00:41] How do i disable the new notifcations? [01:04] anyone else with this one https://bugs.edge.launchpad.net/bugs/337080 [01:04] Ubuntu bug 337080 in bluez "bluetoothd crashed with SIGSEGV in calloc()" [Medium,New] === drostie is now known as drostie|dreams === sarixe` is now known as sarixe === YDdraigGoch is now known as WelshDragon === WelshDragon is now known as YDdraigGoch === Andre_Gondim-afk is now known as Andre_Gondim [02:25] Sigh, who do I talk to about a universe package that is completely broken out of the box (to such an extent that I'd call it ftbfs)? uwsusp was synced + some additional patches, and those additional patches are bad, known to be broken. The bug has been filed, nothing's happened in quite a while. Notably, the debian version works properly, but even reverting to the old pre-jaunty version (which is quite a bit more out of date) w [02:25] ould be preferable [02:25] * cwillu stamps his feet [02:28] cwillu: I'd say look up the host package on packages.ubuntu.com then find who packages that [02:30] just lists "Ubuntu MOTU developers", no particular name [02:30] which package? [02:30] there's a name for original maintainer, but his packages works fine (from debian [02:31] uswsusp [02:31] oh, that might explain why i couldn't find it the first time [02:31] s2disk/s2both segfault out of the box, don't do anything [02:31] missed an s, sorry :p [02:31] -- Devid Antonio Filoni Fri, 13 Feb 2009 21:48:21 +0100 [02:31] his irc nick is probably in launchpad [02:32] * cwillu looks [02:32] http://patches.ubuntu.com/u/uswsusp/extracted/ [02:32] (aptitude changelog uswsusp ftw!) [02:32] Hobbsee: do you mind sponsoring the debdiff in bug 343254 for me, please? [02:32] Launchpad bug 343254 in linux "pulseaudio: alsa-util.c: snd_pcm_avail_update() returned a value that is exceptionally large" [High,In progress] https://launchpad.net/bugs/343254 [02:33] dtchen: preferably not, sorry. i'm coding an assignment [02:33] sure [02:33] dfiloni or devfil [02:33] (don't ask why i'm on irc talking ;) ) [02:33] don't see either of them [02:34] cwillu: only change the changelog mentions is splashy > usplash [02:34] bruce89, that patch is broken [02:34] bruce89, but debian's package works perfectly [02:34] "add usplash support, I hope the patch is good" [02:34] way to test things before committing them to universe, eh? :p [02:35] Hobbsee: just get yourself klined [02:36] heh [02:36] * Hobbsee is doing useful bits on irc - just not being everywhere [02:36] * Hobbsee is looking for a "find and delete this element in a vector in c++", with apparently no luck [02:36] stl :< [02:44] Hobbsee: assuming it's not a predefined type, you need to override find() [02:44] dtchen: darn, OK [02:45] Hobbsee: see also binary_search(). depending on whether you're doing insertions and deletions into the vector, you could also use a map instead of a vector. [02:45] * Hobbsee discovers that she already has the position of the element from before, thus doesn't have to find it again. woot! [02:45] dtchen: that's true [02:52] Hobbsee: that just gave me a good idea for how to handle something in my homework. thanks! [02:52] :) [02:53] instead of returning 1/0 for true/false when trying to find out if something's in a list, return its index if it's there and -1 if it's not [02:53] then i can refind it easier [02:54] that's usually the way to do it, yes :) [02:55] I tried the new 9.4-alpha fglrx... no Xorg 1.6 support. === Amaranth is now known as bluesmoke [02:56] so do i still need python 2.5 or 2.6 substitutes for it? [02:56] for some 2.5 didnt uninstall when i upgraded? [02:56] *some reason [02:57] 2.5 is still available for those who want to use it === bluesmoke is now known as Amaranth [02:58] ah i see [02:59] maybe some packages still need it (even python 2.4 is still around because of zope 2.x) [03:14] its just that i had some problems yesterday trying to install exaile. it was saying something like python version is not defined in header for a package mutagen or something similar. but python-central update fixed that, still i noticed i had python 2.5,2.6 and 3.0 installed :) [03:48] hmm. my python seems to be broken [03:52] i guess i'll just remove every broken package till i get a small enough set that upgrades [03:54] hi, getting errors when updating linux-headers-2.6.28-9, unable to create ./usr/src/linux-headers-2.6.28-9/arch/cris/kernel/Makefile: No space left on device... yet i have gigs of room on / ? [03:56] musikgoat|main: is either /tmp or /usr on a separate partition from / ? [03:58] hmm, i looked at df, and it does look like /usr only has 189M avail [03:58] perhaps the ext4 issue too [03:58] whats the ext4 issue bruce89? [03:58] i am running ext4 [03:59] i heard about a bug, but didn't read up on it [03:59] jldugger: in what way exactly? [04:00] musikgoat|main: sometimes there are issues with ext4 thinking it's full [04:00] usually a fsck fixes it [04:01] does somebody here know how I can disable those annoying deprecation-warnings for python that i get since 2.6 was set as default? [04:01] bruce89: do you know offhand how to force a fsck on boot? [04:01] bug #317781 [04:02] Launchpad bug 317781 in linux "Ext4 data loss" [High,Fix released] https://launchpad.net/bugs/317781 [04:02] musikgoat|main: can't remember, someone will I'm sure [04:02] musikgoat|main: sudo touch /forcefsck iirc [04:02] except you can't do that if the fs thinks it's full [04:03] yofel_: fix the code? :) [04:03] boot single. [04:03] yofel, -W ignore should do it [04:04] thanks yofel_ [04:05] bruce89, musikgoat|main I just linked the bug, and the fixes were just released a few hours ago [04:06] musikgoat|main, basically consisted of applications depending on specific behaviour of ext3 that isn't in common with any other filesystem, nor guarenteed by posix standards [04:06] all that debate about fsync [04:07] the 'fix' consists of a workaround to make ext4 behave as ext3 did for the common cases [04:07] cwillu: interesting, thanks for explaining... fix-released means the patch is in jaunty-proposed? [04:07] musikgoat|main, may not be synced with the repo's yet, but the changelog is up [04:07] is there really a proposed for jaunD[D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[D[Dty? [04:08] what the heck? [04:08] not yet [04:08] so i should expect it in the -10 kernel update coming along the pipes [04:08] nice [04:08] thanks cwillu [04:08] bruce89: they don't create that until release, right? [04:09] yes [04:11] i think DanaG just left the room to show off the unicode quit message :-P [04:11] Nope, was switching computers, actually. [04:12] I downgraded to the Intrepid X server, to get back fglrx, and thus compiz. [04:12] I was running pidgin on a second computer. [04:12] heh, i kid [04:12] I still have this issue, though: https://bugs.launchpad.net/ubuntu/+source/fglrx-installer/+bug/314600 [04:13] Ubuntu bug 314600 in fglrx-installer "fglrx versions newer than 8.543 cause system hang and panic" [Undecided,Confirmed] [04:13] Happens even on just plain Intrepid. [04:13] It's also just as broken with the 8.600 leaked driver. [04:17] leaked? where? [04:18] is it a beta for 9.3? [04:19] It's a beta for 9.4. No Xorg 1.6 support, though. [04:19] 9.4? [04:19] 9.3 isn't even out [04:19] so? [04:19] someone leaked a 9.1 beta before anyone had even seen a beta of 9.0 [04:19] well, for one, 9.4 is supposed to support xserver 1.6 [04:20] betas are supposed to be feature-complete :) [04:20] wouldn't be the first time someone has overpromised and underdelivered [04:20] and others are talking about a "9.4 beta" on phoronix too [04:20] yeah, I remember reading about ATI's ridiculous devcycle [04:21] I swear they must have two teams of developers there [04:21] And they go back and forth on releases so each one is actually two months of work, it's like a stepped advance in a battle or something :P [04:21] heh [04:21] * bruce89 SCO is waiting linux 2.7 [04:24] Amaranth: according to an old phoronix article, it's 11 weeks === yofel_ is now known as yofel [04:25] crdlb: 3 teams then [04:25] and beta starts 4 weeks before ship [04:25] goddamn metacity won't let go of compositing. [04:26] Well, it's also possible it's just suffering from the same brokenness as every other version has had for me. [04:26] for everyone [04:26] Even with the Intrepid X server... 8.600 didn't work. Kernel panic. [04:30] so ... now that fglrx's new versioning is > the old one, why don't they change the internal version? :/ [04:31] https://bugs.launchpad.net/ubuntu/+source/fglrx-installer/+bug/314600 [04:31] Ubuntu bug 314600 in fglrx-installer "fglrx versions newer than 8.543 cause system hang and panic" [Undecided,Confirmed] [04:32] ALL versions newer than 8.543 are just as broken for me. [04:32] It also seems to be related to this: https://bugs.launchpad.net/fglrx/+bug/288620 [04:32] how many times are you going to paste that? :/ [04:32] Ubuntu bug 288620 in fglrx "fglrx:fireglAsyncioIntEnableMsgHandler] *ERROR* interrupt source ff000066 is not supported on this hardware (return code = 1) [EPR#257840]" [Medium,Triaged] [04:32] there are -- afaik -- no fglrx developers in this channel [04:32] if there are any of you, please don't announce it [04:33] They're probably in the witness protection program by now :) [04:35] ugh, ccsm can't comprehend arbitrary keys. [04:35] 'arbitrary'? [04:36] lemme see... [04:36] state 0x0, keycode 199 (keysym 0x0, NoSymbol), same_screen YES, [04:36] yeah, arbitrary. [04:37] you mean a key with no known symbol? [04:37] ccsm needs keysyms, keycodes are different on every keyboard [04:39] e033:f21 [04:39] Should be F21, it seems. [04:40] I had to manually modify that file to add the "EliteBook" string to a few of them. [04:40] ah, it's been so long since I've used compiz. =þ === Andre_Gondim is now known as Andre_Gondim-afk [05:30] good: https://launchpad.net/ubuntu/+source/usb-creator/0.1.13/+changelog [05:31] now I can deploy a livecd to an sdhc card. [05:31] crdlb: i donno what exactly went wrong, but it's fixed now. probably related to a python transition [05:38] DanaG: nice [05:39] I can boot from sdhc card. Nice. [06:16] any reason why SDL was built without OSS support? does it really save that much disk space? [06:20] anyone know which package I need or do I need to rebuild SDL (sdl-mixer?) === Andre_Gondim is now known as Andre_Gondim-afk [06:39] I'm going to upgrade to jaunty now, wish me good luck :) [06:41] kane77: get ready for massive crash bugs in bluetoothd and gnome-panel and such.. oh wait, those were there in Hardy and Intrepid [06:41] :) [06:42] I was actually trying to switch to Affix due to bluetooth issues. Unfortunately it's not working for newer kernels [06:42] I want to see if I will be able to connect to channels above 11 with b43.. someone on forums said that it should either work or you might be able to get it working in jaunty.. [06:43] but hey, the repositories are really slooowww :) [06:55] anyone at all know what I'd need to do to get oss support in SDL? [06:56] hmm apparently libsdl1.2debian-oss exists [07:11] Hmm, I'm trying to update but Jaunty - Release - part is giving me a Key error? [07:44] A recent package update has brocken Synaptic and the package manager. apt-get dist-upgrade still working fine. Any idea aas how to fix this? [07:44] throw it out the window. [07:44] done that :) [07:45] Still it would be fine to understand what's wrong with the update manager [07:45] would help if you actually had an error message, too [07:45] any details? [07:45] where can I look in the logs? [07:45] crdlb: no, details not required. Just throw it out the window, and stomp on it ;) [07:46] arthur_: 'apt-get update && apt-get upgrade' is usually a good start [07:46] Hobbsee, pretty radical solution [07:46] arthur_: terminal output? [07:46] arthur_: *grin* [07:46] crdlb, I would be more tan happy to pass this on but, how? :) [07:47] I am actually getting crash messages from Gnome [07:47] But no hint as to what is actually crashing [07:48] try apt-get update && apt-get upgrade [07:48] run update-manager in a terminal [07:48] Hobbsee, done that, be there [07:48] no error messages? [07:49] no error messages from the update manager [07:49] it just poofs away? [07:50] I just did. btw it opens the gui and refreshes repos [07:50] nope [07:50] so ... it works? [07:51] well, i'ts not an error with apt / aptitude then [07:51] Yes update-manager works but does not report any updates [07:51] er, because there are none? [07:51] MAy be [07:51] But this morning I first tried that [07:51] * Hobbsee advocates defenestration [07:51] just wait 5 minutes [07:51] And subsequently did a apt-get update [07:52] apt-get dist-upgrade [07:52] and got 150 + Mb of updates [07:52] yup [07:52] there's a big difference between crashing and not displaying any updates [07:53] Hobbsee, Is there anyway I can easily look into my logs? [07:53] arthur_: cd /var/log, look for apt logs [07:53] A console.app, for instance? [07:53] I know that [07:53] A GUI to the logs [07:53] Same as on OSX for instance [07:54] there should be one [07:59] I just found this in my user.log [07:59] ubuntu-hp pulseaudio[4044]: module.c: Failed to load module "module-x11-xsmp" (argument: ""): initialization failed. [07:59] The only significant error message [07:59] Is it bad? :) [08:01] OK, I made some further progress [08:01] Invoking synaptic from shell, works just fine [08:02] It doesn't work from the Gnome menu [08:02] Perhaps time to update that [08:02] arthur_: what does "gksu id" on a terminal do? [08:03] uid=0(root) gid=0(root) gruppi=0(root) [08:03] Well, it's in Italian.. ;) [08:03] But it looks ok to me [08:04] mvo, agreed? [08:04] hm, that looks good, no idea what is wrong from launching it from the menus then (I had thought it might be gksu broken) [08:04] now here is a hint [08:05] starting synaptic from shel works [08:05] but reminds me about the missing privileges [08:05] arthur_, Are you starting synaptic from the shell without gksu ? [08:05] sudo synaptic doesn't work [08:05] arthur_: it won't open at all from the panel menu? [08:05] arthur_, You shouldn't use sudo with GUI programs [08:06] Ah [08:06] !gksudo | arthur_ [08:06] arthur_: If you need to run graphical applications as root, use « gksudo », as it will set up the environment more appropriately. Never just use "sudo"! (See http://psychocats.net/ubuntu/graphicalsudo to know why) [08:06] live and learn [08:06] arthur_, See if it starts properly with "gksu synaptic" [08:07] gksu synaptic is not working either [08:07] Any errors? [08:07] Gtk-WARNING **: cannot open display: :0.0 [08:08] That's interesting [08:08] As I said just typing "synaptic" works fine [08:08] IS there a way in Ubuntu to set all permissions in the filesystem to a default? [08:09] Same thing as "repair permissions" in OSX.. ;) [08:09] anyone having trouble with skype chat messages going out? [08:09] arthur_, Try: gksu zenity --entry [08:09] You will probably get the same error if I am right that this is not synaptic specific [08:09] bash: gksy: command not found [08:10] gksU [08:10] Jordan_U, something missing? [08:10] ah [08:10] command not found [08:10] lower case [08:10] :) [08:11] No protocol specified [08:11] (zenity:13575): Gtk-WARNING **: cannot open display: :0.0 [08:12] arthur_, Now to see if non Gtk apps have the same problem try "gksu xeyes" [08:12] same problem [08:13] in other words, it appears that something is broken with gksu [08:13] any ideas? [08:14] Jordan_U, I have another 10 minutes.. ;) [08:16] arthur_, What is the output of: echo $DISPLAY [08:16] Jordan_U, and btw, typing "xeyes" just works fine [08:17] :0.0 [08:17] looks OK [08:17] I have no idea what the problem is then [08:17] :D [08:18] Do you have another user set up? [08:18] thanks for trying [08:18] well, I can try yes [08:18] what do you have in mind? [08:18] Try running "gksu --su-mode xeyes" and choose to run it as your other user [08:19] same as before [08:19] :( [08:19] (gksu:13698): Gtk-WARNING **: cannot open display: :0.0 [08:22] I am leaving, thanks for the assistance so far [08:22] np, good luck [08:23] tia [08:23] Hi, has anybody noticed free and the system monitor program giving different memory usage? [08:24] And, if free is right, the base system using up a lot of memory (about 900mb for a pretty default version of xubuntu, last update about 10 hours ago) [08:24] ablomen: http://virtualthreads.blogspot.com/2006/02/understanding-memory-usage-on-linux.html [08:27] jpds, yeah well what the thing is, i know there is usualy a difference, but 350mb on system monitor vs almost 1gb on free seems a bit... huge [08:28] ablomen: That's... true. [08:29] ablomen: the right information about memory usage with free is the "-/+ buffers/cache:" line [08:29] I think gnome-system-monitor is just excluding cache [08:30] and i also did get "can not alocate memory" errors when trying to start some programs.. so im guessing free is about right, and all i was running is xfce, vlc and apache+mysql [08:30] canot allocate memory indicates some sort of corruption, usually [08:30] cannot* [08:30] oh nice, ok so memcheck it is [08:31] due to overcommit, the kernel will pretty much never refuse a reasonable memory allcation [08:31] bah, can't type [08:31] which is why glib has the policy of aborting on malloc failure :) [08:32] heh ok then, bwa i was hoping other people had this problem, seems i need to check my ram then.. well thanks for the tips jpds fargiolas and crdlb === Ronald is now known as Ronald|Work [08:33] for some weird reason, epiphany-browser 's menu icon doesn't appear unless you logout/login - can anyone confirm this? [08:33] ablomen: not that kind of corruption [08:33] crdlb, oh? [08:33] ablomen: I mean heap corruption, ie a bug [08:34] ah ok, so the best course of action may be using that pmap command to check the memory usage per process? [08:35] to see where to report a bug [08:35] oh, it's happening with multiple apps? [08:35] well no i dont know what app is doing it [08:35] its not an application i start myself [08:36] savvas: It's been reported on LP a few times. The binary and the icon are in different packages. If the icon gets installed before the binary, GNOME won't show it. Logging in and out makes GNOME look for the binary again --> it shows up. [08:36] i checked top but i didnt see anything weird there [08:36] so im guessing a daemon is bugged or something like that [08:37] rww: I tried update-menus and sudo update-menus after installing it and that wouldn't work either - thanks for shedding some light :) [08:37] rww: hmm, I guess karmic can fix that :) [08:37] (by only having one ephy) [08:37] crdlb: I guess karmic is dropping the gecko backend? [08:38] epiphany is [08:38] it was supposed to happen sooner, but ephy 2.28 should be webkit-only [08:39] yay :) [08:40] I'm actually using it now full-time, not that I really recommend it [09:23] do Notifications only show on Primary screen, and not where the mouse and intereacting are happening? [09:25] without looking at the code, I'm guessing it doesn't check the Xinerama outputs at all [09:26] so it just positions it at the top right of the screen, which could (in theory) be a place no output (monitor) is showing [09:30] bug 343576 [09:30] Launchpad bug 343576 in indicator-applet "Notifications only appear on the Defaul display" [Undecided,New] https://launchpad.net/bugs/343576 [09:31] crdlb: can you comment that there? [09:31] also bug 343575 [09:31] Launchpad bug 343575 in gnome-do "DO classic popup only show on Default Display " [Undecided,New] https://launchpad.net/bugs/343575 [09:31] ahh nice morning [09:31] event have read email and already filed two bugs === drostie is now known as drostie|dreams === drostie|dreams is now known as drostie [09:36] BUGabundo: are we talking about notify-osd notifications, still? [09:38] two diff bugs [09:38] one for notifications [09:38] and one for gnome do [09:38] sorry for the confusion [09:38] right, but the bug is filed against indicator-applet [09:39] bug 331369 [09:39] Launchpad bug 331369 in notify-osd "regression vs. notification-daemon: positioning when multiple screens are available" [Medium,Triaged] https://launchpad.net/bugs/331369 [09:44] yep === Eq|laptop is now known as EqUaTe [10:29] hi, I got an asus laptop X71SL-7S134C the network card is reconized like Ethernet controller: Silicon Integrated Systems [SiS] 191 Gigabit Ethernet Adapter (rev 02), I got the eth0 but I'm unable to use it, it refuse dhcp, even if I configure it static it doens't work, anybody now if it's will be fix ? or know how to fix that ? thanks [10:30] macsimlap: what error do you get? [10:30] macsimlap: SiS? good luck with that [10:31] slytherin, no error [10:31] spend 2 days around a laptop with board, wired and wifi SiS based [10:31] no luck! [10:31] BUGabundo, thanks ;) [10:31] would not conect to anything [10:31] to install it, I was forced to use noacpi too [10:31] it's strange everything works fine, except the NIC [10:32] the dhcp behaviour seems the same [10:32] as I suffered [10:32] yeah [10:32] seems you are out of luck [10:32] BUGabundo, you disable acpi and it solve your problem ? [10:32] many, many entries on forums with SiS hw support request [10:32] with zero feed back [10:32] macsimlap: no it didn't [10:32] but at least allowed me to install ubuntu on it [10:32] 1st time I got a computer on linux where everythings work except the NIC [10:33] wouldt even boot if not for that [10:33] but unlike you, everything on that laptop as SiS [10:33] you seem to only have the wired card [10:33] SiS is just bad for linux [10:33] BUGabundo, I got wireless too and it's works fine [10:33] send them an email.... [10:33] BUGabundo, I will ;) [10:33] macsimlap: I bet its not SiS, is it? [10:34] BUGabundo, you win Atheros Communications Inc. AR928X Wireless Network Adapter (PCI-Express) === elpargo__ is now known as elpargo [10:43] olá BUGabundo [10:47] anybody having trouble with playing DVDs on jaunty? I am wondering if I am the only person with problem. Please try totem (gstreamer), mplayer or VLC. [10:54] oias joaopinto [10:55] slytherin: there have been reports of css not compiling [10:55] what have you tried? [10:55] did you try to install the codec from archive? [10:55] using the build sh? [10:55] medibuntu codec repos? [10:55] BUGabundo: I have libdvdcss installed and all the codecs. [10:55] too many questions for a noob? [10:56] if you have it installed, can you check another dvd? [10:56] BUGabundo: I am not the noob. I am a MOTU. And I am responsible for sync/merge of libdvdread/libdvdnav from Debian. But now that I find the dvd playback broken, I am wondering why no one else has complained. :-) === Teiseii is now known as Teisei [10:59] slytherin: sorry... had no idea, once you came in... notice the detail on the reply so assumed you knew more [10:59] slytherin: to anwser your question: I've seen 2 users compalining [11:00] BUGabundo: Any pointers? [11:00] no... [11:00] bad memory on my side [11:00] check the IRC logs... [11:00] I know some one in here help one of the users debug it.... [11:01] he had luck with another movie... but not all [11:01] Ok. If it is related to libdvdcss then I don't think I can help. [11:03] slytherin: i didn't find any problems when i tried last time? [11:04] slytherin: and i hope you didn't drop my changes, like debian did... [11:05] just to be sure: what's libdvdread ? [11:05] Hobbsee: I didn't do the latest merge of libdvdread. Which changes are you talking about? [11:05] hrm. could not load location [11:05] slytherin: making install-css.sh executable, via debian/rules, mainly. [11:05] slytherin: and i thought you said you did ;) [11:05] Hobbsee: No. Those changes are there. And that is not the issue. [11:06] ah good ;) [11:06] I get permission error in totem and similar error in other players. I have found the solution but not able to understand the cause. [11:06] "Could not open location; you might not have permission to open the file." ? [11:06] Hobbsee: ^^ [11:07] Hobbsee: yes, same [11:08] hmmm [11:09] ubuntu-restricted-extras: "Drop libdvdread3, since libxine1 has its own private copy of libdvdread/libdvdnav". That's not related is it? [11:09] cwillu: nope [11:09] cwillu: xien works fine. [11:09] libdvdread3 isn't just xine though, no? [11:10] slytherin: hrm. i think it's a permissions problem with the actual cdrom drive? [11:10] crdlb: that bug, as been triaged and marked as dupe! [11:10] Hobbsee: No. Here is my analysis. There is a configure2 script and a Makefile shipped by upstream. I think that is causing problem. If you use the ./autogen.sh to generate configure and then use that then the resulting .so works fine. [11:11] cwillu: Currently all the packages that use libdvdread are broken. But since libxine has its own copy of libdvdread/libdvdnav, it is not broken. [11:11] slytherin: hit up seb128 about it? [11:11] ahhh [11:11] BUGabundo: well, it is, isn't it? [11:11] (or just fix it) ;) [11:12] just upgrade my kubuntu 8.10 to 9.04 [11:13] no problem at all [11:13] everything works fine [11:13] Hobbsee: I can fix it, but I want to understand the root cause. I am wondering if currently used Makefile is causing any problem in symbols being exported. [11:13] slytherin: https://bugs.edge.launchpad.net/ubuntu/+source/totem/+bug/342890 is the bug, fwiw [11:13] Ubuntu bug 342890 in totem "Cannot play DVDs" [Undecided,New] [11:13] hello, does anyone know how to disable ssh's feature to use any public key in my .ssh to login to a server even if it's not named as id_rsa.pub? [11:14] Hobbsee: Right. I will talk with seb128 provided I can not get hold of slomo. [11:14] slytherin: prodded it with the "OMGFIXIT" stick, too [11:15] gah. LP lost what I said. [11:15] Hobbsee: And once this is fixed, users are going to bless gstreamer developers. Because menus, subtitles work well. :-D [11:16] woot! [11:16] thanks for the report vistakiller [11:19] nelf, the .pub isn't what logs you in to the server, it's the id_rsa file [11:19] the .pub is the part you copy into a particular server so that it can give challenges for the private key [11:20] cwill: but if i rename the .pub to .pubb it won't use it anymore.. weird [11:21] nelf, also, be sure you're not using connection sharing/multiplexing or anything like that, that could confound issues [11:22] cwillu: i mean, it will ask me for then ask me for a password instead of using the id_rsa (which i renamed to something else so that it will not be used by default) [11:22] nelf, sorry, I didn't follow that [11:22] cwillu: nope, not using that.. [11:23] ssh-agent? [11:23] (ps aux|grep agent) [11:23] cwillu: before i upgraded to jaunty, i already had private keys in there, but they where renamed to something different (not id_rsa) so it won't be used by default [11:24] but now they're being used in jaunty, unless i rename their .pub to .pubb s or something else [11:24] ps aux | grep agent gives me pulse, and seahorse [11:31] nelf, I don't see anything relevant in the changelog [11:32] afaik, the .pub file isn't even used locally [11:32] i.e, I can move mine to a seperate folder, and still log into to any servers that have it in their authorized_keys files [11:32] you use the priv to login [11:33] cwillu: yes, that's what i knew [11:33] the remote site has the pub to match [11:35] i know, that's how it should be.. but i was just surprised why i when i login into my server it didn't ask me for a password [11:36] (it also did not allow me to type anything, as i added noptys in my remote authorized_keys)) hehe [11:36] oh, seahorse is doing more for this now, it runs by default and acts as an ssh-agent. [11:37] nelf: that's what ppl love about cert login [11:37] no user input necessary [11:37] a friend complain to me about beeing asked for the KeyChain key on on jaunty [11:37] and that I sucks [11:39] killed all seahorse and it's still trying to use my renamed id_rsa [11:40] it's still in memory, log out and back in and I bet it prompts for a password [11:41] ok, i'll try that [11:41] quick, everybody hide before he gets back! [11:43] * BUGabundo goes invisible [11:43] * BUGabundo he isssss backkkkkkkkkkkk [11:44] yup, it's seahorse alright [11:44] .. still, weird :) [11:58] Is the mythtv package broken? I get unmet dependencies for front-end and back-end packages when doing sudo apt-get install mythtv [11:58] BonezAU: let me guess: python? [11:58] mm omg i get koffice broken [11:58] wth [11:59] BUGabundo, please go on... I think I have python installed. Any specific version I should install first? [12:00] naa [12:00] recently we went undergona a HUGE python mygration [12:00] to 2.6 [12:00] many many many packages needed to be rebuild [12:00] some of course need a push [12:01] so should I install 2.6 ? [12:01] 2.6 is already installed according to aptitude [12:04] BonezAU, refresh your repos, you probably just caught things in the middle of a push [12:04] it should already be installed [12:04] BonezAU, dependencies are fine on my system [12:04] if you have the metapackage ubuntu-desktop [12:05] i will triple check that, back in 5 gotta hang washing out [12:06] How to upgrade to Kubuntu Jaunty from Kubuntu Intrepid? [12:07] cumulus007: do-release-upgrade -d [12:07] ah, thanks [12:07] Adept really needs a -d option [12:07] use update-manager... [12:07] that's GNOME stuff [12:07] its the recomened way to dist upgrade on Ubuntu [12:08] duh [12:08] use update-manager-kde [12:08] ROFL [13:00] Howdy! I just noticed that PREEMPT has been disabled in -generic kernels starting with Jaunty. is there a more desktop-oriented kernel that would have it? [13:02] Q-FUNK: I didn't think we ever enabled it [13:02] Amaranth: it used to be enabled and it showed in "uname -a" [13:02] Not having it seems to be our problem with pulseaudio, our latency is too high [13:03] exactly why I'm asking. it messes up operation with Skype [13:03] there is no kernel provided that has it [13:03] Q-FUNK, prempt-vountary is enabled, and has been the default for some time [13:04] can't type [13:04] preempt_voluntary [13:04] cwillu: ok, but that doesn't suit the needs of desktops. [13:04] :) [13:04] Q-FUNK, you might want to check into what that means before you say silly things like that :p [13:05] cwillu: what makes you think that I don't know the difference? [13:05] forced preemption is only required if you've got poorly written drivers that don't have the appropriate break points [13:05] and in some cases, forced preemption _breaks_ drivers [13:05] cwillu: Which obviously is our problem [13:06] fedora kernel: 5ms [13:06] our kernel: 130ms? [13:06] Amaranth, proprietary drivers are apparently a common culprit [13:06] I'm just trying to figure out how to get Skype working rock-solid again, the way it used to with Intrepid. [13:07] well, that's not the difference, the preempt setting hasn't changed [13:07] recent PA uploads just keep on providing messier an dmessier defaults that make Skype barf and cough. [13:09] i was thinking to do some bug report after the upgrade in kubuntu [13:09] but everything works fine!!! [13:09] almost as if the maintainer is trying different options to see what works across the majority of systems :p [13:09] :D [13:13] actually, recent PA changes even affect totemt and rhythmbox, which used to work fine [13:14] Amaranth: yup, that's some serious latency. [13:15] * Reenable 0030_set_tsched0.patch, which re-disables glitch-free; [13:15] too many users are reporting regressions and audio aberrations. [13:15] * Adjust 0003_change_resample_and_buffering.patch to use linear [13:15] resampler to work better with lack of PREEMPT in jaunty's [13:15] -generic kernel config (LP: #207135, #322250, #332761, #335955, [13:15] LP: #336965). [13:16] (changelog from latest pulseaudio) [13:17] precisely [13:17] !? [13:17] brb [13:17] you're upset that they tried something during an alpha, it didn't work, and so they backed it out? [13:19] * cwillu will learn some day to not take people with all-caps names seriously :p [13:20] * cwillu suggests people who want full preempt comment on bug #324133 [13:21] Launchpad bug 324133 in linux-rt "Wishlist - Provide PREEMPT kernel since RT is broken" [Undecided,New] https://launchpad.net/bugs/324133 [13:23] you're upset that they tried something during an alpha, it didn't work, and so they backed it out? [13:23] you just missed that :p [13:23] * cwillu suggests people who want full preempt comment on bug #324133 [13:23] Launchpad bug 324133 in linux-rt "Wishlist - Provide PREEMPT kernel since RT is broken" [Undecided,New] https://launchpad.net/bugs/324133 [13:24] Launchpad bug 324133 in linux-rt "Wishlist - Provide PREEMPT kernel since RT is broken" [Undecided,New] https://launchpad.net/bugs/324133 [13:24] Ubuntu bug 324133 in linux-rt "Wishlist - Provide PREEMPT kernel since RT is broken" [Undecided,New] [13:24] Launchpad bug 324133 in linux-rt "Wishlist - Provide PREEMPT kernel since RT is broken" [Undecided,New] https://launchpad.net/bugs/324133 [13:24] ugh [13:24] this was my five upgrade in this system [13:24] and for first time i dont have a problem [13:24] nice work guys!! [13:25] what's the rule on assigning or subscribing teams to bugs? [13:25] was thinking about assigning 324133 to ubuntu-kernel-team, but I don't know that that's appropriate [13:28] cwillu: normally a good idea not to but its really up to the person/team. that bug will not be fixed (id say about 75%) it would be a feature and FF is in place. talking with someone from the team would be better than just assigning it. the team gets the bugs in email if its a kernel bug [13:29] gnomefreak, package linux-rt count as something they'd get notified on? [13:29] * Amaranth builds a 2.6.29 kernel with PREEMPT and modesetting, prepares for pain [13:30] cwillu: the bug is agiasnt the kernel right? [13:30] * cwillu suggests Amaranth throw that in a ppa :p [13:30] gnomefreak, it's against linux-rt [13:30] should it be against something else? [13:30] (pre-existing bug) [13:31] look on the right hand side of the bug it will tell you everyonre [13:31] cwillu: pfft, packages? [13:31] kernel team isn't listed [13:31] kernelcheck says this will take 2-4 hours [13:31] that's a bit...off [13:32] heh [13:33] unless it's going to compile the kernel 6 times [13:33] Amaranth, maybe it'll throw some random configs in for testing :p [13:35] gnomefreak, should I be looking for kernel-team in there somewhere? [13:35] cwillu: im looking at the bug atm [13:36] ah, k, thanks [13:36] Q-FUNK, thank the nice gnomefreak :p [13:36] :) [13:37] heh, it's already going through building the modules [13:38] cwillu: its done [13:39] * cwillu pokes Amaranth and Q-FUNK [13:39] thanks [13:40] the 1000Hz thing is a waste when we've got NO_HZ [13:41] gnomefreak: thanks [13:41] I tried to build a kernel on a Pentium machine once to get my modem working, after 6 hours I turned it off and installed Win2K [13:41] Q-FUNK: np [13:41] nice to be able to build one in the time it takes to eat breakfast now [13:42] * cwillu notes that there is an -rt kernel in universe, does it do anything useful? [13:42] yeah, but 1000Hz really produced a fast and responsive desktop, back when it was enabled. huge difference between 250 and 1000. of course, with zero hz enabled, it's a moot point. [13:42] cwillu: the -rt kernel doesn't do power management [13:42] because that would cause latency problems [13:43] so it always runs in C0 which causes heat and wastes battery [13:43] I understand that some hardware drivers barf if they run on a faster clock, but it felt sad when the default was brough back down, in the early days of 2.6 [13:43] of course 1000Hz is crap for a laptop too [13:51] hrm, looks like 20 minutes was not enough time [13:51] dang slow HD [14:11] ok [14:11] I have a fresh installon kubuntu64 [14:11] and firefox is crashing right and left [14:12] seamonkey is acting more stable [14:18] Hello why does my jaunty breakdown randomly [14:18] sometimes it goes right for a day [14:18] amortvigil: because it's not yet released [14:18] its alpha ware [14:18] Q-FUNK: i know but this is a strange things [14:18] thing [14:19] and not yet stable [14:19] and thus i want to ask how to find out whats wrong and how to report it\ [14:19] I have issues with firefox onit [14:19] oke [14:20] thats a good one [14:20] ill go check it bye [14:28] unixdawg: file a bug on it i'm doing email today [14:29] ok in a few filing office reports right now [14:29] gnomefreak: eheh [14:29] had server issues over the weekend [14:30] modeset fail [14:31] will have to build a 2.6.28 kernel later with just PREEMPT [14:33] why do I doubt that's going to happen? [14:33] kernel team has refused it [14:35] BUGabundo: I'm going to do it myself, not for others to use [14:37] ah ok [14:42] Hm... is audio completely destroyed for anyone else? [14:42] picklesworth: [14:42] picklesworth: me [14:42] (hoping this is the climax before some miracle fix to all the problems) [14:42] but using a daily LiveCD works [14:42] bug filed [14:42] oh, what's the bug #? [14:43] sec... [14:43] bug 343258 [14:43] Launchpad bug 343258 in pulseaudio "Logs from yesterday daily image" [Undecided,New] https://launchpad.net/bugs/343258 [14:43] its just a base line comparation [14:44] run http://www.alsa-project.org/alsa-info.sh [14:44] when filing bugs [14:44] Darn, I think ours are different problems. For me, audio just makes a horrifying, crackly popping noise for a few seconds then PA crashes [14:45] so it makes a sound. Just not a very nice one. [14:45] diff bug [14:46] Oh, now it works, so you're all going to think I'm crazy... [14:47] Meh, that probably stemmed from my failed attempt at getting a bluetooth headset to work [14:48] Hi, I still haven [14:50] tmeixner: then give it back [14:50] 't filed my bug for amarok. Music on the network (smb / ssh connection) doesn't play on kubuntu. Is it dophin who offers the filebrowser view with network/root/home ? [14:50] we need our haven, someday [14:51] hehe BUGabundo, I blame dell and their lousy keyboard layout for this :-) [14:53] I tried mapping a drive via sshfs and it works. I think that narrows it down to the filebrowser in amarok/kde. Is it called kio? [14:55] If Dolphin wont do what you need, try Konqueror [14:55] I think Dolphin is supposed to though [14:57] Eruaran: that's part of the problem. It's the filebrowser inside amarok which doesn't work correctly so I don't really have a choice. I think it's provided by KDE but I want to know which package. [14:57] nice [14:58] firefox just killed it self [14:58] tmeixner: ah i c [15:00] tmeixner: pretty sure Amarok would be a KIO slave [15:00] !kio [15:00] Sorry, I don't know anything about kio [15:01] neither does the bot :-) No problem, I think I have enough keywords for the bugreport. [15:02] launchpad doesn't really seem to distinguish between alpha and stable releases. You just throw it in one big pot? [15:03] surely not ? [15:03] tmeixner, they're seperated by the release name [15:04] there aren't seperate alpha and stable releases, there's just a release, and whether it's actually been released yet or not === jonathaN2 is now known as jonathaN [15:08] * cwillu patiently waits for a kernel update [15:08] cwillu: one new coming out? [15:08] * BUGabundo needs to put kernel ML up to date [15:09] BUGabundo, the ext4 workarounds are coming down the pipe [15:09] ahh [15:10] Just noticed all the little changes to GTk [15:10] I love the caps lock warning in password fields :) [15:17] cwillu: seems jens axboe (linux ata guy) took tytso to task about sync not being expensive [15:17] so there may still be some changes coming later [15:18] calc, that has little or nothing to do with the workaround though [15:23] cwillu: true but sync's can't be made to be cheap from jens told tytso so perhaps we will eventually end up with not needing to use sync with every write [15:24] heh [15:24] not sure if you read it, but even though sync on ext4 is for the file specifically in hardware it has to flush the entire disk buffer so it still writes loads of data to disk even for a 1 byte file sync [15:24] calc, they can be far cheaper than ext3's though, that's all that was at issue [15:24] xfs does it [15:25] cwillu: oh the issue is that tytso is advocating changing all code to use sync for every write [15:25] cwillu: and if that were to occur the expensiveness of sync would start to become more evident [15:27] I don't get it [15:27] I use XFS and never had any trouble [15:27] related to data loss on power failure [15:27] at least that I noticed [15:27] had way more with KDEPIM crashing and losing files [15:27] BUGabundo: heh i had data loss often with XFS, others i heard had data loss often with reiserfs [15:27] since I read the bug report I've been doing sync on occasion [15:27] BUGabundo: kdepim losing files... on XFS? [15:28] BUGabundo: if so that is due to XFS [15:28] and I notice that some times HUGE amount of stuff gets cached [15:28] calc, has there been some discussion outside of theo's blog post and the bug? jens doesn't really say much at all on the blog [15:28] and doing it before suspend and hibernate tends to make it MUCH faster [15:28] calc: I think I have a bug or two on kdepim [15:29] let me see if I find them [15:29] cwillu: probably on the fs list but what jens does say makes it clear that what tytso has been saying about needing fsync essentially for every posix file command would be much more expensive than he claims [15:29] can't [15:30] too old for FF Aweome bar [15:30] cwillu: ext4 may be somewhat faster currently wrt writing to disk but once every bit of code is changd to his utopian view it won't be any more [15:30] BUGabundo: ok [15:30] I've used them all [15:30] what jens is saying is that every time you call sync it flushes the entire disk buffer on hardware... which on modern hard drives is 32MB [15:30] ext2/3 reiser, xfs [15:30] not ext4 yet [15:31] so you force a 32MB write for a sync of a 1 byte file [15:31] but xfs is far faster to others [15:31] so sync is not nearly as cheap as tytso was claiming (or thinking?) [15:31] I only have 8MiBs of cache [15:31] dought you will find laptop with more then 16 in the next year [15:31] BUGabundo: older hard drives have smaller or no cache of course :) [15:31] mine is one year old!! [15:32] hmm yea actually my drive only has 16MB cache in my laptop but 32MB on my desktop [15:32] this sync performance issue with affect all of Linux though, not just laptops [15:32] calc, jens is talking about the physical drive cache though, wheras we've been talking about the os's page cache, which is completely seperate [15:32] cwillu: to completely separate it and say sync is really cheap without understanding underlying hardware though is very shortsighted which is what Jens post was about [15:33] i.e., in the case of dataloss, the data was never sent to the drive at all [15:33] sync isn't cheap even if the fs makes it cheaper at, just, that level [15:34] calc, flushing the hd's cache is never going to cause a 3 second synchronous delay on the desktop [15:34] it is cheaper than eg 500MB (100MB/s * 5s buffer) but it can still be in roughly the same order of magnitude write and as drive caches increase it will [15:34] yes, it will impact throughput, but that's not even close to what we're talking about. [15:34] guys this is angles Sex kinda discusion [15:34] what we need is better hw support to do it async [15:35] we ask disk to do it, and internally it should be handle [15:35] and report back when done! [15:35] BUGabundo, um, no [15:35] BUGabundo: what we need is better posix spec or barring that to do one on fdo or something to make the meta data atomic [15:35] the page cache isn't on the disk [15:35] that is not [15:36] BUGabundo: currently its not clearly defined so fs people have decided that meta data can be updated before the data in the file is even written [15:36] BUGabundo: which is the root cause of all of this mess [15:36] to get around it now they want to turn off sync()ing to disk with laptop_mode so nothing can safely ensure it's data is written to disk [15:36] ah [15:36] instead of just implementing proper meta data ordering [15:37] of course i don't know how hard the proper meta data ordering would be, but that should be the real goal, not hacking around the issue by turning off sync etc [15:37] the current 2.6.28 patch forces full sync of files to disk whenever meta data is changed to get around having to fix the issue [15:38] which is why there is a performance issue with the new patch [15:38] plus now we have SSDs [15:38] and sgould have into consideration [15:38] to optimize eat and warn out [15:39] SSD's don't have much seek delay but do have very low data io caps [15:39] eg the Intel one degrades > 20GB/day [15:39] and flusing its buffer (if it has one?) every sync probably would hit that pretty fast [15:40] * calc doesn't think SSD's will be useful for real world for another 5yr+ [15:40] just upgraded to +1 and can't seem to get the right resolution out of my external monitor -- it's supposed to be 1680x1050 -- lspci says I have an intel mobile 4 chipset, integrated...any ideas how to fix this? [15:41] crankharder: GPU? [15:41] ah intel [15:41] crankharder, crt? [15:41] nvm, probably not with that resolution [15:42] lcd [15:42] * calc would guess a 19" lcd [15:43] 22 [15:43] linux display drivers need to do less thinking and more listening until they can autodect well [15:43] e.g., I shoudl be able to say, dammit, 1680x1050 i dont care what you think is attached [15:45] the problem usually that the panel is lying to the graphics card - there's nothing the drivers can do about that [15:45] crankharder: edid has been a defined spec since the mid 90s if your lcd isn't giving accurate data then it isn't linux's fault [15:46] maybe even the early 90s iirc my crt from 94 did edid and was a fairly cheap one at the time [15:46] calc: it JUST worked before I upgraded [15:46] there really is no excuse for there to be edid issues anymore [15:46] don't tell me it's the lcd's fault [15:46] crankharder: hmm if it did work before it might be a bug of some sort in the intel driver [15:46] exactly [15:47] see my point above [15:47] crankharder: what resolution is it running at? [15:47] crankharder: and do note font size may change between intrepid and jaunty [15:47] 1280x1024 on a widescreen display [15:47] read man intel specifically options "ModeDebug" and "DDC" [15:47] crankharder, in which case, you should be filing a bug on launchpad. Having said that, there already is a 'use this resolution damnit' feature in xrandr [15:47] which is the biggest it'll allow [15:48] crankharder: i would file a bug on the intel video driver in launchpad if it worked fine before [15:48] cwillu: oh really? [15:48] crankharder, check your /var/log/Xorg.0.log file for the modes line (it's probably listed, but not listed as the preferred resolution), and then you can use xrandr --newmode with that line to make it available [15:49] regardless of the solution, if it worked with intrepid fine but not jaunty its probably a bug and should be filed in launchpad [15:49] crankharder, in fact, you might be able to directly set it via xrandr --mode "1680x1050" [15:50] i've got 2 displays... [15:50] anyone having issues with wine? [15:51] beardbar: define issues? [15:51] mine worked fine last week [15:51] crankharder, the line I gave you woulnd't work anyway, for exactly that reason. I was presuming you'd check the docs for xrandr, and see that there's also an --output parameter related to this :p [15:51] This is my first bugreport, if s.o. has a second, is it written ok? https://bugs.launchpad.net/ubuntu/+source/amarok/+bug/343716 [15:51] Ubuntu bug 343716 in amarok "KIO filebrowser in amarok doesn't play files on a network share (smb/sftp)" [Undecided,New] [15:52] anyone reading the last few emails from Design team to devel-discuss ML ? [15:52] BUGabundo: every app i try bugs out but I think its because my soundcard [15:53] the logs all report sounddriver issues. [15:53] that's another prob [15:53] that should be fixed soon [15:53] dtchen already pactched it [15:53] cwillu: I got that far -- now I get cannot find mode 1680x1024 [15:53] ahh cool [15:56] 1050? [15:58] that one either :) [15:58] there is a 1680x1050 modeline in /var/log/Xorg.log [15:58] pastebin the output of xrandr, and your /var/log/Xorg.0.log file [15:59] ooh, Epiphany 2.27 is nice :) [16:04] cwillu: what are you looking for in Xorg.0.log -- 10K lines is too many for pastebin :/ [16:05] xrandr: http://pastie.org/417763 === drostie_ is now known as drostie [16:09] crankharder, 10k lines!? [16:10] crankharder: don't you log rotate? [16:10] i didn't write it :/ [16:11] crankharder, give me head -n 1000 /var/log/Xorg.0.log | pastebinit -i - [16:11] apt-get install pastebinit if you don't have it already [16:13] cwillu: that's one of the packages that should be in base metapackage [16:13] okay -- figure it out, in xorg.conf my Virtual screen was set too small [16:13] * BUGabundo consideres filing a bug on that wishlist [16:13] is it possible to grow the virtual limit? [16:13] was set to something like 28--x1024 when it needed to be (1200+1650)x1050 or 2930x1050 [16:14] isnt it hardcoded to X? [16:14] BUGabundo, it's specified in xorg.conf [16:14] which is working right now -- let's see if killing X blows it up [16:14] cwillu: I read some where that it was limited to some values [16:14] crankharder, this was a fresh install? [16:14] that is getting to low for current days [16:15] BUGabundo, drivers have maximum texture sizes that you typically run into, it's really only an issue for compiz though [16:16] ahh [16:16] and even then, only for windows which are larger than that max texture size [16:16] the problem is that the desktop is one of those windows [16:16] like me having 2x1440 screens? [16:16] having a seperate root window for each screen would fix that, but I think that would require x hacking to do [16:16] x server hacking, rather [16:18] Hi, I cannot connect to my works wifi, iwlist scanning shows http://pastebin.ca/1362321, what's this IE: unknow? [16:18] cwillu: no, this was an upgrade [16:18] calc, the second half of http://thunk.org/tytso/blog/2009/03/12/delayed-allocation-and-the-zero-length-file-problem/#comment-2032 may be useful for you [16:19] okay -- so the only problem now is that the icons on my laptop screen (1280x1024) when autoarranged are off the top of the screen [16:19] crankharder, open the desktop in nautilus, set it to icon mode, and move things around as required. It's the same layout [16:20] have you made the panels larger lately? [16:20] (or moved them, for that matter?) [16:20] if so, you'll need to restart gnome-panel. It has a nasty bug right now where it doesn't update the window struts [16:21] cwillu: no, the problem is autoarrange puts them off the top of the screen [16:21] oh, that [16:21] even though windows snap to the correct edge of the screen [16:21] that's the price you pay of having two monitors side by side with different vertical sizes [16:21] cwillu: yea i've read all that before, it doesn't address what jens axboe commented on earlier in the thread, i haven't had time to read the entire new thread (read all of the bug report thread though) [16:21] the desktop is the size of the total virtual size [16:21] deadspace isn't handled very elegantly [16:22] cwillu: he keep reiterating that sync is 'cheap' now when actually it is only cheap at the filesystem level not in reality :) [16:22] if I somehow get the smaller screen to the top of the virtual space will the icons scroll off the bottom then? [16:22] calc, it's as cheap as it is on any other fs, which is all he was talking about [16:22] calc, we're not talking about gratuitous fsyncs, we're talking about doing fsyncs when you need to make sure the data is written. It's the bare minimum === DBO3 is now known as DBO [16:24] cwillu: you need syncs for everything in the real world since you can't even expect meta data ordering to be correct currently, he hacked in a patch to 2.6.28 but doesn't consider it to be correct... but that every posix io call needs a sync to be correct... [16:24] cwillu: not sure if you read the entire bug thread to get that bit of information [16:24] cwillu: well thanks for the help :) [16:24] ... [16:24] cwillu: basically it seems that either properly tracking meta data on a per file basis in memory is terribly expensive or he just doesn't want to put in the effort to do it right (?) [16:25] cwillu: i haven't seen him fully address that part of the issue [16:25] ... [16:25] calc, he has, several times :p [16:25] you can't talk about the metadata of a file in isolation [16:26] if a block is allocated to a file, and a block is allocated to another file, you can't commit one and not the other [16:26] cwillu: yes in the current fs layer implementation... what he didn't address is whether that issue could be fixed to separately track per file [16:27] that's why delayed allocation is a huge win, because you _can_ commit the one set of metadata, because the other transaction hasn't happened yet [16:27] cwillu: if it were possible to be fixed then it would solve the problem in entirety [16:27] calc, you can't [16:27] why? [16:27] * calc notes this may be obvious but it has been several years since he hacked on fs stuff in kernel [16:27] if it is obvious i have forgotten why :) [16:27] calc, because the problem isn't about the file, it's about the metadata on the disk to say what blocks are allocated where [16:28] it's that structure that's being manipulated and commited and backed up by the journal and so forth [16:29] i still don't see why this couldn't be tracked on a per file basis for the meta data so it doesn't get flushed before the file without having to resort to flusing the file early as the 2.6.28 patch does now [16:29] if you've allocated blocks for file a, and then you allocate blocks for file b and fsync file b, then you _have_ to also commit the allocation for file a [16:29] can i install notify-osd in kubuntu? [16:30] cwillu: oh... so you don't run out of disk space and end up failing to allocate for a later on? [16:30] so this is essentially an overcommit related problem? [16:30] or end up allocating something that's already been allocated [16:30] not exactly [16:30] it's the things you have to do to avoid fsck'ing [16:31] basically, if this stuff didn't happen, then instead of a hard crash taking out a file, a hard crash could take out the entire file system [16:31] the same way that fat could hard-crash [16:31] or ext2 [16:31] allocate in your above example is really preallocate right? [16:31] it could be, the distinction doesn't really matter [16:32] currently the way it works (aiui) is that if you preallocate file a and then preallocate b then sync b it only writes b to disk (without the 2.6.28 patch) [16:32] aiui? [16:32] as i understand it [16:32] no, now we're into delayed allocation :) [16:32] from what tytso has said sync on a file only writes that file info to disk [16:32] with delayed allocation, the blocks haven't been reserved. It hasn't even been decided where they will be written yet [16:32] and sync only commits the data in the file as the meta data stuff all happens within 5 sec or something like that [16:34] so you: open() write() close() rename(), could happen in open() rename() write() close() order() as i understand it and if you crash after rename() then it has a 0 byte file [16:34] well, specifically, you have to fsync the containing folder to fsync the metadata [16:34] oops order() should not have had brackets [16:34] the 0 byte file comes from the truncate [16:34] cwillu: from my understanding of the problem the meta data is updated before the data is, the sync causes the data to actually be written [16:35] open('write_truncate') [16:35] at that point, you've dropped the blocks that were allocated in the filesystem. You asked for precisely that [16:35] cwillu: there was discussion that even the rename() can happen before the data hits the disk, not just the truncate part [16:35] yes, because the rename is metadata [16:36] if you don't fsync [16:36] cwillu: so even presumably correct code of open(tmp) write() close() rename(to orig) will fail with a 0 bytes file [16:36] that's not correct code though [16:36] open(truncate), write(), fsync(), close(), rename() is the correct sequence [16:37] bah, that's wrong [16:37] open('new'), write, fsync, close, rename [16:37] there is no truncate needed in the above [16:37] heh [16:37] that's the right sequence [16:37] yep :p [16:37] that just killed your battery as well [16:37] eh? [16:37] no [16:37] which is why people want the spec extended to have meta data tracked properly or at least in implementation [16:37] what fsync actually does is left to the os [16:37] the solution tytso came up with to this is just to disable fsync entirely in kernel [16:38] which then requires you change code that you want to really hit disk for critical reasons not just for consistency purposes to use a new syscall [16:38] so we need to change all code in existence to use fsync, then change all critical code to use this new currently non-existant syscall [16:39] calc, that's not what he said [16:39] but regardless :p [16:39] cwillu: he said he created a patch for 2.6.28 to work around this issue so that all data is syncd automatically with metadata since he doesn't think the above is actually achievable [16:39] when I hit save in my editor, that needs to do an fsync [16:40] but the feature that laptop-mode is supposed to provide is that it still doesn't spin up the drive [16:40] the need to use a new syscall is needed and he even said himself so that data that actually needs to hit disk will in power management cases [16:40] since if you have laptop-mode disable fsync then even critical writes will no longer happen [16:40] define 'critical writes' [16:41] a cache write isn't using fsync [16:41] and writing to a database is [16:41] as is saving an editor file [16:41] as is writing to a log file [16:41] soon almost everything will be using fsync [16:41] as they need to be [16:42] yes... but fsync will no longer ensure something is on disk (which is actually broken posix behavior aiui?) so we have to extend syscall to have real_fsync for apps that need it [16:42] any app that predates ext3 is already using fsync, because that's how you get stuff written [16:43] no, the entire point of laptop-mode is that you're holding back changes that otherwise would be considered to be vital [16:43] you can potentially lose 5 minutes of data [16:43] if you don't write to the drive for 5 minutes (including fsyncs) [16:43] but you don't corrupt anything by doing that [16:43] you just straight up lose the data [16:43] time machine style :) [16:43] that's what laptop-mode is giving you [16:44] so it's completely reasonable to delay fsyncs until that point [16:45] perhaps that will work ok, it sounds good enough for my personal use, but not sure if for everyone... but it still ends up causing hw buffer flush on every write which is not very good for performance (i think?) [16:45] * cwillu starts muttering :p [16:45] heh, i'll shut up, i'm sure tytso and jens can come to some conclusion on how to fix that :) [16:45] Sounds like all we need is a COW ramfs overlay :) [16:45] ext3 is already doing that, and far worse, because it has to allocate _every_ pending transaction [16:46] cwillu: calc with so many fsync I wonder if everything will be written to disk on time! [16:46] although i think by definition sync() would require that not to happen [16:46] IntuitiveNipple, almost like a page cache :p [16:46] BUGabundo, it's still better than ext3, because ext3's semantics already required worse performance :p [16:46] the whole idea of optimizing everything for battery life at any costs troubles me [16:47] i mean what is the point of an extra 5 minutes of battery when the machine is unusable and loses data in the event of a crash? [16:47] cwillu: almost but not quite :) [16:47] IntuitiveNipple, shush you :p [16:47] (and a crash is pretty likely to happen if you are using every last bit of power in the battery) [16:47] * IntuitiveNipple hums to himself and rocks back n forth :) [16:47] ali1234, well, laptop-mode is a special case. The last 15 lines are so are really just about demonstrating that laptop-mode still has a sensible meaning in an fsync'ing world :) [16:47] ali1234, laptop-mode-tools turns off laptop mode when the battery live gets below a threshold for exactly that reason [16:47] hmm yes fsync() requires the data to be on disk after it returns (so besides breaking posix for laptop-mode) this would make physical disk buffers useless for writes at least [16:48] BUGabundo: I've responed to your bug #335507 update - we are making progress :) [16:48] Launchpad bug 335507 in netspeed "netspeed applet will not measure wired" [Undecided,New] https://launchpad.net/bugs/335507 [16:48] calc, that's not actually what posix says :p [16:48] IntuitiveNipple: I read it [16:48] cwillu: ah well the man page, i suppose it should be updated :) [16:48] ah fsync directory [16:48] I don't want to say best-effort, but it's basically that [16:48] misread that bit (distracting kid running around room) [16:49] but regardless, laptop-mode works exactly as advertised even if it disables fsync's globally [16:49] cwillu: well the kernel can't completely disregard that to allow the physical disk buffer to be useful again [16:50] IntuitiveNipple: 1. My interpretation of "...it didn't detect..." means the applet does not offer the wired connection in the list of devices in the preferences dialog. [16:50] means its there and even enabled manually [16:50] * calc gets back to rewriting OOo's debian/rules file [16:50] 800+ lines done and millions to go ;-) [16:50] * cwillu has a suspicion that a delicious fruit blended beverage is about to appear at his desk [16:50] I'm like a cat: I can hear the can opener running :) [16:51] BUGabundo: really?! But it *did* offer the WiFi? Please add that observation to the bug report - that is significant [16:51] BUGabundo: and probably easy to figure out from the source code [16:51] 2) even when manually selected it will not mesuare anything [16:51] the icon is the same as not plugged [16:51] calc, when you're done that, you wanna start converting all our sysvinit rc jobs to upstart? [16:51] BUGabundo: Yes - that's that running = 0 flag again [16:51] I thought the report already stated that! [16:51] on the OP [16:51] we've had upstart for 2 years now, and we still don't use it for anything but tty's and ctrl-alt-delete :p [16:52] IntuitiveNipple: I'm confused with your questions there [16:52] lol [16:52] so I think you are too [16:52] BUGabundo: Sometimes it helps to be clear, the entire thing can be confusing when remote debugging and seeing different results that don't reproduce the problem [16:56] IntuitiveNipple: so let me rephrase things and tell me what you don't understand [16:56] so I clear on the ticket [16:56] I have 2 netspeed applets [16:57] one for Lo, one Auto [16:57] a few days before I filed the bug [16:57] it stopped detecting what interface was plugged [16:57] Yes... that'll be the "running = 0" flag [16:57] after a few tries it started to auto detect wifi, 3G and vpns [16:57] which is libgtop [16:57] but wired NEVER gets autodetect [16:58] manually selecting ANY will work other then wired [16:58] wired NEVER works [16:58] I'm adding some more debug code so you can capture more information [16:58] ok [16:58] thanks [16:58] is this clear? [16:58] No :D but wel'll muddle through :) [17:00] There have been no changes to netspeed so it has to be the changes to the underlying system [17:01] IntuitiveNipple: I'm a bit way [17:01] on the job... [17:01] cwillu: oh, i thought keybuk already had that stuff done :\ [17:01] so im is on another windows [17:01] I may delay to reply [17:05] BUGabundo: I'll reply to the bug report when there's anything to know [17:08] ok [17:08] do you need me to add anything there? [17:08] IntuitiveNipple: ^^^^^ [17:13] BUGabundo: not right now... the next thing will be reporting the output of the next debug version [17:15] calc, had a set up upstart jobs? [17:18] http://cache03.stormap.sapo.pt/fotostore02/fotos//10/c4/f7/2872313_J3YfW.jpeg NSFW [17:18] IntuitiveNipple: let me know when it's the ppa [17:19] BUGabundo, in what sense do you think nsfw links are appropriate for ubuntu channels? [17:19] sorry cwillu [17:19] just trying to make your afternoon funnier [17:19] its a broswer comparation [17:20] I know what it is, but I'm just wondering what you were thinking :p [17:21] to make you guys laugh a bit [17:21] seems I overdid it [17:21] BUGabundo: interesting, for some reason here the auto mode has grabbed the 'tun0' device (openvpn connection) [17:21] yes [17:21] here too [17:22] VPN takes precedence over any other [17:22] What's happening with XUL? [17:22] It deleted suddenly after an update [17:22] so I installed it back , because without it I can't run Firefox [17:22] but Firefox still doesn't work [17:22] when running xulrunner, it gives a: [17:23] update-manager check for updates again, installing any that show up [17:23] *** buffer overflow detected ***: /usr/lib/xulrunner/xulrunner-bin terminated [17:23] you probably just caught something mid-update [17:23] and a lot of bactracing stuff [17:23] BUGabundo: It's not that - it just takes the last device name reported and generally that's tun/tap devices and similar [17:23] cwillu: did you say that to me? [17:23] lol skpye wont install on kubunu64 [17:23] skpye/skype [17:23] yes [17:24] unixdawg: I have (had?) it running on 64biys [17:24] ubuntu (gnome with lots of KDE libs) [17:24] What negative side effects should I expect in Kubuntu Jaunty if I pin KDE at version 3.5.x? [17:25] well I do dpkg --install and it says wrong arch [17:25] jaunty uses kde 4.2 [17:25] sparr: mm [17:25] sparr: probably have trouble upgrading [17:25] i am on kubuntu 9.04 [17:26] and some features there aren't [17:26] but now you can use it at least [17:27] back up your data and start fresh [17:27] * calc hopes KDE doesn't decide to go break the world again in 2011 [17:27] eg after KDE 4.5 [17:27] kde5 will be the end of all os's as we know them [17:28] * calc has been using gnome since late 2004 after he stopped maintaining kde 3.x [17:28] hi [17:29] ok [17:29] i've got Intel video card... how can i use applications with 4:3 window in full screen, with correct aspect ratio? [17:29] my screen is 16:10 [17:30] Glacies_Hu, applications as in movie players? [17:30] they should already be displaying in the correct aspect ratio [17:30] BUGabundo: this netspeed issue - are you running on 32-bit or 64-bit OS ? [17:31] cwillu, no, video players are good. I thought about maybe games. [17:32] games should be handling it properly too [17:32] file a bug against the game in question is the best answer [17:32] as a work around, you could apply a window rule via compiz to size it [17:33] can i add windows rule to full screen applications too? [17:33] yep [17:33] window* [17:34] oh, thats good, thx for help. [17:34] IntuitiveNipple: 64 bits [17:34] but... i turn of compiz when im playing. [17:35] BUGabundo: ok... same here. I was speculating that a bit-width and flags issue might be to blame. [17:35] IntuitiveNipple: why just wired? [17:36] BUGabundo: it was just a theory. [17:36] hey compiz is running slow on my computer [17:36] why [17:36] on intrepid was perfect [17:36] peace: disable vsync [17:36] BUGabundo: it isn't just wired... there's an issue with the 'running' flag returned by libgtop [17:37] peace: and Indirect [17:37] peace, because stuff changes in major releases, and alpha's are when we fix the stuff that breaks :) [17:38] cwillu: humm we break stuff that works too [17:38] BUGabundo, yes, I just said that [17:38] nobody can consistently write stuff that works the first time, there needs to be testing [17:39] peace: i am also on kubuntu 9.04 [17:39] peace, search the bug reports in launchpad is a good place to start (for things like your current video card, etc) [17:39] cwillu: updated, but it doesn't help :( [17:39] calc: after KDE 4.5 I would expect it to be at least as usable as 3.5, so it won't be such a concern then. [17:39] cwillu: mm i have serched and i don't see so much stuff [17:39] calc: im more concerned with the next 12-18 months when i would rather be using 3.5 than 4.2/4.3 [17:39] peace, what chipset, whats the exact symptoms? [17:40] cwillu: intel 945gm , slow effects [17:40] but on debian it works fine [17:40] on kubuntu no xD [17:40] peace, basically, just remember that nobody owes you a trouble free alpha experience. If you want support and the like, wait for the release :p [17:40] peace, sec, I think I know the bug [17:41] cwillu: i know i know it's an alpha but xD [17:41] bug #303011 [17:41] Launchpad bug 303011 in xserver-xorg-video-intel "[i945] 2.5.1 driver poor performance" [Unknown,In progress] https://launchpad.net/bugs/303011 [17:41] peace, I've been able to workaround it by running the -server kernel instead of -generic, but that hasn't worked for many people [17:42] i think around 4.4 KDE will be back to where it was with 3.5, in terms of features and usability [17:42] yes now kde4 it's fine here [17:43] cwillu: mm thank y [17:44] is anyone here running the Deskbar applet successfully? [17:44] for me it's been crashing as soon as I load it since upgrading to Jaunty [17:44] sparr: now you can use kde4 maybe something has to come but now youi can enjoy [17:44] picklesworth: no [17:45] its dead for one or two weesk [17:45] cwillu: you mean on the boot? [17:45] feel free to open a bug and let me know the id [17:45] peace: im using it right now, hate it [17:45] I suppose I should... [17:45] peace, yes [17:45] sparr: mm ? i love it xD hihihi [17:46] picklesworth: I'll save you the work [17:46] peace, install linux-server, (and probably linux-restricted-modules-server as well), and then reboot into the server kernel [17:46] my apport just fired on it [17:46] peace, worked for me :p [17:46] uploading logs NOW [17:46] Righto [17:46] cwillu: well i will try i have to create a lots of screencast now [17:46] :P [17:47] oh, I did that one. Never really trusted apport, though. It just kind of disappears when it's finished, leaving me in the dark :/ [17:47] maybe apport doesn't like me [17:50] ah i have another issue with my wacom tablet [17:50] xD [17:50] eraser dissapear [17:50] i can't use it [17:51] i have my xorg.conf configured [17:52] Yar, the upgrade tool seemed quite convinced that HAL would be able handle autoconfiguring Wacom tablets for us and proceeded to comment out the lines in xorg.conf [17:52] which is nice; an empty xorg.conf is awesome; except that HAL absolutely does not manage to do it [17:52] the lines should still be there, though [17:56] Oooh, interesting! [17:56] I just ran deskbar successfully this way in the terminal [17:57] python2.6-dbg /usr/lib/deskbar-applet/deskbar-applet [17:57] expecting to debug the segfault, but instead it worked [17:57] super weirdly, running it with python2.6 (not debug) still crashed [17:58] BUGabundo: Did that create a bug report on Launchpad, or should I create one? [18:05] cwillu: it's working faster xD [18:05] peace_, yay :p [18:06] peace_, hopefully they get it fixed properly, but ya [18:06] only helps on the 32bit server kernel [18:06] haha thank you man [18:07] -server enabled pae on 32bit machines, which isn't compatible with the new gem memory manager, which is the source of the grief [18:07] i have tried kwin and it' fastest [18:07] so gem gets disabled, and we're happy :p [18:07] fantastic [18:08] picklesworth: got on the phone! [18:08] need to find the upload window [18:08] let me check [18:09] BUGabundo: There's a new libgtop package, can you test with that once it hits the archives? [18:10] upgrading === EQP[AwNg is now known as LjL [18:10] BUGabundo: also, upstream seem to be reworking the struct glibtop_netload - that could be part of the issue although why it doesn't affect me I don't know. [18:13] IntuitiveNipple: my system is trash [18:13] need to reinstall [18:13] eheh [18:13] Eeek! Weird... NetworkManager disconnects then reconnects when I make a saved connection global [18:13] * BUGabundo misses UpdateManager number of updates count :(( 53MiBs doesn't say enough [18:13] picklesworth: expected... it'll drop the per-user connection and then do a system connection (I'd expect) [18:13] picklesworth: it makes some sense [18:14] indeed. I wish it told me, though :/ [18:14] BUGabundo: At least you can enable it showing the version information in gconf [18:15] can I? [18:15] mhmm [18:15] * BUGabundo looks at gconf [18:15] /apps/update-manager/show_versions [18:15] humm [18:15] it looks like they've completely abandoned counting updates, though. That's odd. [18:15] are we talking the same ting? [18:17] yeah [18:17] I miss it [18:17] I wonder if I should file a bug against the Design team [18:17] like bruce89 did === PcYpVWD\ is now known as LjL [18:28] Oops, here's that bug report by the way: https://bugs.edge.launchpad.net/ubuntu/+source/deskbar-applet/+bug/343799 [18:28] Ubuntu bug 343799 in deskbar-applet "deskbar applet crashes when starting" [Undecided,New] [18:28] bug 304871 [18:28] Launchpad bug 304871 in xserver-xorg-video-intel "[i845G] Fatal server error: Couldn't bind memory for BO front buffer (Jaunty)" [Unknown,Confirmed] https://launchpad.net/bugs/304871 [18:29] stupid question: using usbcreator to place ubuntu on a flashpen, will work with ext2/3 and not fat/fat32 ? [18:41] gym time [18:41] see you guys tomorrow [18:41] every time krunner is too slow to load i remember why i am hating 4.x [18:44] ok is there no flash for kubuntu64 ? [18:46] unixdawg: if you mean adobe flash player then that would be flashplugin-nonfree [18:56] unixdawg: pretty crashy though on 64 bit. [19:01] ok flash works [19:02] on 32 bit all was fine but 64bit flash often crashes on me; already on 8.10 [19:02] since I upgraded to the abode 64bit alpha release, I have no flash crashes.. with 8.10 [19:03] SwedeMike: npview.bin has something to do with flash don't it? [19:04] hmm well i AMON JAUTY [19:04] we willsee [19:04] Hi, my PulseAudio is crappy, sometimes it works, sometimes it doesn't [19:04] and it uasually takes 50% of my cpu [19:04] my syslog is overwelmed by this [19:04] Mar 16 20:03:15 sander-desktop pulseaudio[12285]: alsa-util.c: snd_pcm_avail_update() returned a value that is exceptionally large: 13835058054713115660 bytes (418290290514 ms) Most likely this is a Linux bug. Please report this issue to the ALSA developers. [19:05] this notification is being put there ~10 times a second [19:07] perhaps report it to the alsa developers [19:11] ribo: I'm going to do that [19:11] ribo: but what I don't understand, is why PulseAudio is being used on my KDE system [19:12] I thought KDE uses Phonon? [19:17] cumulus007: it is using the pulseaudio backend - maybe you can configure it to use the xine backend [19:18] SandGorgon: but how? I'm sure the Xine backend is selected, and it's the only backend available in the list [19:19] i found some nice tools for ubuntu [19:19] cumulus007: in that case, you need to kill the pulseaudio daemon [19:19] http://fnords.ch/ubuntu [19:20] SandGorgon: I'm doing that all the time [19:22] cumulus007: in that case, did you install KDE from Gnome, there might be some problems because of that. try uninstalling pulse [19:23] SandGorgon: GNOME was pre-installed [19:23] the Kubuntu alpha downloaded much slower that time [19:23] I will install a plain Kubuntu system in the future, I think ;) [19:24] cumulus007: try uninstalling pulse... [19:38] cumulus007, could you file a bug on that? [19:43] anyone here get a bcm4318 working [19:43] wifi card [19:43] why is there no news, no anouncement about 9.04 whatsoever? its not even mentioned on the ubuntu website [19:43] ? [19:44] i thought it was supposed to conquer the home user market [19:44] shiny_people: Its not released yet. [19:44] yeah but I mean windows 7 is not released yet either, but I already know everything about it because its advertised [19:47] shiny_people: https://wiki.ubuntu.com/JauntyJackalope [19:51] Hello.. anybody else have NVidia proprietary driver breakage with recent Jaunty update? [19:51] It'll be advertised with a countdown once Jaunty is in beta [19:51] I need to get my bcm4312 working [19:51] is it just me or has the flash plugin for firefox always been so laggy? [19:54] so I take it here no one has broadcm mini pci wifi cards [19:56] unixdawg: what kind of problem you having? [19:57] it shows there is a wlan but when I try to bring it up it fails [19:57] root@unixdawg-laptop:~# ifconfig wlan0 up [19:57] SIOCSIFFLAGS: No such file or directory [19:57] seems that something is missing [19:58] the drivers for that are proprietary arent they? [19:58] no there are kernel drivers [19:58] i have a similar card [19:58] ok [19:59] just woundering whats missing [19:59] mine actually works fine though [19:59] if theyre kernel that makes less sense [20:00] yes I agree [20:01] I wounder if thy have issues on thier kubuntu64 ver ? [20:05] seems the firmware is missing [20:06] dont they have a download on their site? [20:06] ? [20:06] for the firmware? [20:06] who bcm [20:06] yeah [20:06] orkubuntu [20:06] like intel have one for their cards [20:06] thats for the windows driver [20:06] thats not firmware [20:06] hmmm [20:07] intel have a firmware for their wireless cards for linux [20:22] got it [20:22] I had to fallow a webpage and get a firmware cutter [20:23] ok [20:24] stupid question, did you have a look at the restricted hardware drivers app? [20:27] is fglrx working with the current xorg version ? [20:27] not yet [20:28] is IntuitiveNipple around? [20:28] cause I just hit the same seg fault [20:36] we have wireless [20:39] the firmware work [20:40] but now pidgin is not conneting [20:40] hehe [20:40] one problem after another [20:41] well the fact they dont include the firmware and dont really tell you is a issue [20:41] unixdawg_: join #pidgin and read topic [20:41] it was working before i moved to my wireless [20:50] hi all, had anybody had any luck installing hellanzb on jaunty? [20:53] so you guys don't get update notifications automatically? [20:53] just wondering why it doesnt show up in synaptics anymore... is the only way to install it is terminal? [20:55] yao, since i moved to jaunty I didn't get notifications [20:56] antoranz: same here [20:56] that was till KDE broke..... I'm waiting for a fix to come out so I can see how it's going lately [20:56] kde4 is working wifi is workign but now pidgin is not working === linux1 is now known as sentient-kernel [20:57] antoranz: right. someone in #kubuntu-devel says update-notifier has a bug [20:57] I'm in this bunch: https://bugs.launchpad.net/bugs/338205 [20:57] Ubuntu bug 338205 in kdebase-workspace "[jaunty] plasma crashes on start with SIGSEGV" [Medium,Fix released] [20:58] no hellanzb love in here? [20:59] shit [20:59] whats up unixdawg [21:00] wekk irc works on wifi [21:00] but pidgin is not [21:00] I just got my wifi on this laptop working [21:00] !hellanzb [21:00] Sorry, I don't know anything about hellanzb [21:00] nor do i :P [21:00] thenmin I plugin the etho and bring it up pidgin works [21:01] hmmmm... i had to do a fresh install of JJ... the upgrade process wasn't 100% for me [21:01] !ohmy [21:01] Please remember that all Ubuntu IRC channels share the same attitude of providing friendly and polite interaction with all users of all ages and cultures. Basically, this means no foul language and no abuse towards others. [21:01] sudo bash [21:01] thn do apt-get update [21:01] the apt-get dist-upgrade [21:01] The right way to get a root login is: sudo -i [21:02] sudo su [21:02] !sudo [21:02] sudo is a command to run programs with superuser privileges ("root"). Look at https://help.ubuntu.com/community/RootSudo for more information. For graphical applications see !gksu (Gnome, XFCE), or !kdesudo (KDE) [21:02] :-) [21:04] just nice to have wifi again === a3Dman_ is now known as a3Dman [21:21] anybody know why the graphics tend to be a bit slower on JJ vs. II? [21:21] dell intel GMA [21:21] mostly alpha blending slowdown, and some tearing on videos [21:22] I keep getting timeout exceptions from the ata subsystem in Jaunty, never happened on Intrepid [21:22] could it be a kernel bug or will everyone blame my hardware? [21:22] http://paste.ubuntu.com/132196/ [21:22] happens with different bios settings, different sata ports [21:23] only my ext4 root filesystem during big dist-upgrades, but that's probably because it's such a heavy load [21:26] and all questions fall on deaf ears [21:26] deafbot [21:26] !slow video [21:26] Sorry, I don't know anything about slow video [21:26] !slow graphics [21:26] Sorry, I don't know anything about slow graphics [21:26] !ubuntu [21:26] Ubuntu is a complete Linux-based operating system, freely available with both community and professional support. It is developed by a large community and we invite you to participate too! - Also see http://www.ubuntu.com [21:29] smak: you probably need graphics drivers from System > Administration > Hardware drivers [21:29] didnt need it with Intrepid... but i suppose i can try it [21:29] oh yeah, tormodvolden's PPA has newer ati and intel open-source drivers. [21:29] no, the intel drivers come with X - but they're slow here as well. thanks X 1.6 :/ [21:30] hmmmmm, so it's likely nothing I can really 'fix' per say? [21:32] operative word: NEWER drivers. [21:32] Not the same as the ones in Jaunty out-of-the-box. [21:33] ok, OOB drivers are bundled with X, which are dogshit slow. so try new drivers via S/A/Hardware Drivers [21:34] smak: using compiz? [21:34] just the default install with JJ, let me see what settings i have going [21:35] i'm on my 2nd month with linux, so i'm n00b status bigtime, but luckily i'm not too retarded [21:37] yes, because computers hate retards [21:37] smak: there is a known problem with some intel cards AFAIK [21:39] https://launchpad.net/~tormodvolden/+archive/ppa [21:39] oh wait, it doesn't have new intel driver -- just new drm modules. [21:40] no, wait, it does have newer intel. [21:41] It's just not shown on that page, for some reason. [21:42] i'm looking for an intel i815 driver for jaunty [21:44] I just wish radeon had power management. [21:45] seems like there is comming a new intel driver that needs some kernel changes first [21:45] thanks DanaG [21:45] hmmm... interesting JanC... makes me wonder if i should just wait it out [21:45] argh, battery life for me currently sucks. [21:46] thiebaude: try the -intel driver (it might work) [21:46] ok its half connecting now [21:47] JanC: the ppa one? [21:47] thiebaude: the xserver-xorg-video-intel driver supports i815 [21:47] thiebaude: the one in the normal repository [21:48] is it in synaptic in 9.04? [21:48] I think it should be installed by default? [21:48] It is installed by default [21:48] bug 304871 [21:48] Launchpad bug 304871 in xserver-xorg-video-intel "[i845G] Fatal server error: Couldn't bind memory for BO front buffer (Jaunty)" [Unknown,Confirmed] https://launchpad.net/bugs/304871 [21:49] apt-cache policy xserver-xorg-video-intel [21:49] that the problem i've been having [21:49] That's i845 [21:50] it's also confirmed for i815 [21:50] thiebaude: what's lspci -nn report for your video device? [21:51] let me check [21:52] Also, did you try adding the Legacy3D option ? [21:52] can i paste it from the terminal [21:52] just need the one line for your video card [21:52] ok [21:53] 00:02.0 VGA compatible controller [0300]: Intel Corporation 82815 Chipset Graphics Controller (CGC) [8086:1132] (rev 04) [21:53] im running 8.10 [21:54] thiebaude: interesting. I've got a notebook sitting next to me with that same exact chipset and the intel driver, with no problems [21:54] how do i add Legacy3d option, IntuitiveNipple [21:54] maybe i can get it to work so i can upgrade to 9.04 [21:54] thiebaude: According to that bug: https://bugs.edge.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/304871/comments/83 [21:54] Ubuntu bug 304871 in xserver-xorg-video-intel "[i845G] Fatal server error: Couldn't bind memory for BO front buffer (Jaunty)" [Unknown,Confirmed] [21:56] what is edit command to edit xorg? === linux1 is now known as sentient-kernel [21:58] wish me luck im going to upgrade to 9.04 === DanaG1 is now known as DanaG [22:12] oh yeahzz, another stupid thing: the damned volume control slider is STILL BACKWARDS! [22:12] Scroll up.... volume goes down! [22:13] that's standard gtk behavior [22:14] it makes perfect sense from the perspective of conveying left-to-right [22:14] it makes a terrible volume slider though [22:14] Yeah. [22:14] Why not use a friggin' VERTICAL slider? [22:15] well, i solved the hellanzb [22:15] mystery [22:16] anybody know how to keep the desktop from infinite-switching on mousewheel? [22:16] it used to stop on the 2nd desktop, but now it will loop... [22:17] I assigned it so horizontal-scroll switches, instead of vertical-scroll. [22:17] Use buttons 6 and 7. [22:18] in II, you could scroll down on the wheel, and it would stop at the last window... now it loops back to the leftmost, and keeps scrolling through. Useful, but i'd like to have the stop-action back. [22:32] my network device is listed as unmanaged on network manager, any ideas ? [22:39] joaopinto: Is it mentioned in /etc/network/interfaces or did you do a net-boot or have netconsole enabled? [22:40] IntuitiveNipple, it's mentioned in /etc/network/interfaces [22:41] joaopinto: That'll be why. NetworkManager doesn't try to take over anything mentioned in the manual configuration, or that is active during boot as a result of the kernel actions [22:41] joaopinto: The solution is to remove the mention of it from the interfaces file [22:41] hum, so removing it from there and restarting nm should work ? [22:42] Indeed it should. [22:43] restarting NetworkManager did not change [22:44] sometimes the entire network stack needs restarting... easiest is to restart to be sure. [22:44] ok tks :) [22:47] hey anyone having an issue with desktop effects needing to be renabled every boot and desktop cube effects not working just desktop wall [22:48] working now :) === a3Dman_ is now known as a3Dman [22:58] greetings [22:59] today I installed my SoundBlaster Audigy 2 ZS, entered Bios to disable onboard sound rebooted and i have no more available soundcards at all. [23:00] lspci: 01:05.0 Multimedia audio controller: Creative Labs SB Audigy (rev 04) [23:00] any idea how to enabnle the driver or somewhat? [23:02] is someone awake? [23:13] CosmiChaos: [23:13] Check "Audigy Analog/Digital Output Jack" switch. [23:13] This must be *off* to output from the analog jacks. [23:24] IntuitiveNipple: As i repeat i have NO SOUNDCARD DETECTED i have NO AUDIO JACK MIXER [23:25] IntuitiveNipple: i have the bulk version withoput Audio Jack its a AUdigy 2 ZS bulk [23:25] So have you modprobed the driver manually? checked the kern.log for any errors? [23:25] if i would know which module i would have tried [23:25] but its not modprodbe audigy or audigy2zs [23:26] what's the PCI vendor:product ID? [23:27] 3 [23:27] 3 [23:27] 3 [23:27] 01:05.0 Multimedia audio controller: Creative Labs SB Audigy (rev 04) [23:27] no, the PCI ID (lspci -nn will show it) [23:29] [1102:0004] [23:29] thanks [23:29] IntuitiveNipple: thanks to you for helping [23:30] /lib/modules/2.6.28-9-generic/modules.pcimap:snd-emu10k1 0x00001102 0x00000004 [23:30] IntuitiveNipple: i guess its emu10k1 but modprobe emu10k1 does not work and there is no package [23:30] so, snd-emu10k1 [23:30] :) [23:30] lets try [23:31] should it work instantly or after sed >> /etc/modules.conf ? [23:32] Is the module now loaded? (lsmod | grep snd) [23:32] much lines seems like [23:33] but still no soundcards [23:33] restart alsa [23:33] k [23:33] sudo /etc/init.d/alsa-utils restart [23:34] still no soundcards [23:35] for dev in /sys/class/sound/card?; do readlink $dev; done [23:35] ? [23:35] run that command [23:35] from where to where [23:35] hole line [23:35] just run it in a shell and show me the result [23:36] no result [23:36] okay... lets go back a bit then [23:36] You did "sudo modprobe snd-emu10k1" ? [23:37] yes [23:37] When you did "lsmod | grep snd", was snd-emu10k1 listed? [23:37] cypherdelic@HaeckFlaisch:~$ sudo modprobe snd-emu10k1 [23:37] cypherdelic@HaeckFlaisch:~$ cat /proc/asound/cards--- no soundcards --- [23:38] snd_pcm 98952 3 snd_emu10k1,snd_ac97_codec,snd_pcm_oss <--- one line of plenty for example [23:38] okay, so the module is loaded. [23:39] IntuitiveNipple: will try adding it to modprobe.conf reboot and check that oout should be more adequate [23:39] let's try unloading it and reloading it and checking the log as you do.... have two terminals going. In one, start a log monitor: tail -f /var/log/kern.log [23:39] logging is off :) [23:39] now, in the other terminal, do sudo modprobe -r snd-emu10k1 && sudo modprobe snd-emu10k1 [23:39] Why? [23:40] Can't debug without logging [23:40] FATAL: Module snd_emu10k1 is in use. [23:40] okay, so, something is using it [23:41] iw will reboot shortly brb [23:41] Why is logging off? if you've got that in a non-standard configuration, it makes me wonder if you've changed other things that are causing this issue. [23:41] ok [23:47] hi all...what is the best raid tool in jaunty? [23:48] mdadm [23:48] what's the difference to dmraid? [23:51] dmraid is a fakeRAID aka Promise FasTrak compatible module that supports the BIOS data structures of the FasTrak and other compatible controllers. Linux kernel's md drivers are kernel-only software RAID [23:51] I see [23:51] thanks :)