[00:01] anyone? [00:06] grr stupid delta codeshared flights can't do online checkin :\ [00:06] which means they are now going to weigh my carryon and ban it to checked baggage :\ [04:35] ‏ [04:55] Is there currently any project to make a ubuntu install on an external HD "highly portable" between machines? ie, copying the PortableSUSE project ? [04:56] if you dont need nvidia or fglrx, installs should already be very portable [04:59] superm1: was thinking framebuffer... [04:59] superm1: opinions on that? [04:59] what_if, poor performance likely [05:00] but really if you just have a vanilla install onto an external hard drive, not much should stop you from moving it around. you just wouldn't have whizz bang 3d effects on machines that would need -nvidia or -fglrx [05:01] For such an environment, you might want to use usb-creator to create a persistent live environment. That way you get HW detection, etc. cleanly. [05:01] No reason that the "USB stick" can't be a 1TB drive, after all. === williamd is now known as TURtl3 [05:02] HW detection is a bit of a misnomer though too. xorg.conf generation used to be the biggest point there, but now that it's a common xorg.conf for different graphics installs, less of an issue [05:03] superm1, Isn't there still libGL diversion though? [05:03] persia, that's why i'm saying to stay away from -nvidia and -fglrx [05:03] Ah :) [05:03] no real way to keep a portable install that supports any of the 3 possible libGL's [05:04] superm1: I was going to go the framebuffer route simply to keep a static xorg.conf and simplify my life [05:04] what_if, take a look at a vanilla 9.04 xorg.conf. you don't need to [05:04] hmm, I think it was kirkland who said he had such a thing, but blessing something that evil would be paradoxical :) [05:04] * what_if googles [05:05] slangasek, kirkland had a nice way to work with all 3 libGL's at once? [05:05] slangasek, That was an internal hard drive that switched between two machines (but yes, it included some diversion-managing scripts) [05:06] well there is registering libGL.so.1 with alternatives, which i started to look at what was involved last year [05:07] it could potentially make for a nicer experience for bulletproof X, and letting bulletproof X's failsafe session change the alternative to try to get the X session started back up [05:08] or just get noveau working and get rid of the two binary servers :) [05:11] so, from a feasibility standpoint, does everyone think this is viable? (minus 3D support) [05:11] what_if, Minus 3D support, it's trivial,as long as you only try to support a single architecture. [05:12] persia: yes, would be X86 only [05:12] Then, aside from 3D support, Wacom support, and a few other corner cases, every install is portable. [05:21] ‏ [05:22] brb, smoke break / reboot [05:46] ‏ [05:48] ‏ [05:49] ‏ [05:49] ‏ [05:50] !ops [05:50] Help! bhale, infinity, Hobbsee, jdub, thom, fooishbar, fabbione, mdz, lamont, or Keybuk === SEJeff is now known as SEJeff_home [06:36] good morning [06:37] good morning dholbach [06:37] hiya highvoltage === _Traxer is now known as Traxer [07:02] morning === tkamppeter_ is now known as tkamppeter [08:35] hi [08:35] anyone used Complete Fair Queueing in kernel development [08:37] hakkav, thast the default in ubuntu since a few releases already [08:37] ah [08:37] also do you guys use Broadcom b43 driver? [08:38] for item in "$(find /sys -name *scheduler*)";do cat $item;done [08:38] to find the currently used scheduler [08:39] sweet thanks [08:39] its been a few yrs [08:39] btw is the L7 firewall still being licenced? === yofel_ is now known as yofel [08:40] ogra without the quotes right? " "? [08:40] for i in $(find /sys -name *scheduler); do cat $i; done ? [08:40] no [08:40] just as it is there [08:40] Um.. No? [08:40] Oh, right. [08:40] soren, ?? [08:40] That happens to work :) [08:40] works here :) [08:40] i dont paste stuff i didnt check [08:41] i think you are both wrong. should be for banana in $(find /sys -name *scheduler); do cat $banana; done [08:41] LordKow, damned you got me :P [08:41] ;) [08:41] it may go with " " [08:41] no? [08:41] It happens to work, yes. [08:41] LordKow: thats what i did [08:41] for i in $(find /sys -name *scheduler); do cat $i; done ? [08:41] both will work [08:42] ah [08:42] Putting ogra's quotes there render the for-loop rather useless, though. [08:42] As it'll be treated as a single argument. [08:42] yea then you just got yourself a string [08:42] yeah [08:42] indeed [08:42] but im thiniing [08:42] if one is working with files that have spaces [08:43] both wont :) [08:43] like say blabla - blabla.info [08:43] that must be enclosed into quotes [08:43] no? [08:43] cat $(find /sys -name *scheduler*) works as well ... [08:43] <\sh> hakkav: if you have those filenames, kick the guy who created them... [08:43] for each item found in /sys where the name is scheduler, you must cat it [08:43] right? [08:43] find /sys -name "*scheduler*" -print0 | xargs -n1 -0 cat [08:43] hakkav: where the name includes scheduler [08:44] (note wildcards) [08:44] \sh why [08:44] here I have /sys/devices/pci0000:00/0000:00:1f.2/host0/target0:0:0/0:0:0:0/block/sda/queue/scheduler [08:45] LordKow: yeah i know [08:45] if anyone has any thoughts on this matter feel free to type them: which is better (1) /usr/bin/env or (2) /usr/bin/ [08:45] <\sh> hakkav: because those filenames gave me problems in the past...when it all was "new" ;) and it's too MS...and typically managment like... [08:45] ah [08:45] but there';s no issues [08:45] it makes your disk wear a tie ? [08:45] LordKow: neither, /usr/bin/env LD_LIBRARY_PATH=/home/geofft :) [08:46] hah :P [08:46] <\sh> ogra: yes ;) [08:46] LordKow: 1 [08:46] im just thinking of the person who puts python in his home directory, updates the path but wonders why a particular package keeps looking for non-existent python in /usr/bin ;) [08:47] i doubt its easily possible to have an ubuntu without python in /usr/bin [08:47] dont you think 1 is better? [08:47] you can define a new environment for the script you are running [08:48] indeed ogra, but it is possible [08:48] could someone accept gst0.10-python in karmic? [08:48] (indeed it theoretically is possible, but surely requires a lot of fiddling) [08:48] cant you just define a new env? [08:48] #!/usr/bin/env env python, so you can install your own version of env [08:49] heh yea [08:49] LordKow, you dont need to update the path btw, just create ~/bin, put python in there and re-login [08:49] is what i said [08:49] ogra: oh true indeed, never crossed my mind. [08:49] ~/bin is automatically in your path if it exists [08:49] but im pretending im john doe who does crazy things [08:49] poor john [08:49] BUT [08:49] it's not as safe as we wish, but for development and multiple versions, I find it useful ;-) [08:50] not safe though [08:50] exactly. and python is one where multiple versions is a very real possibility [08:50] as any other intrepreter ... [08:50] alternatives *should* be taking care of those symlinks so /usr/bin/python should still work so long as alternatives are updated accordingly [08:51] nothing that the user can modify is safe; if the library path points to rootkits kept by others... or so... [08:51] well, if you package the executable, you should make sure to use /usr/bin/python ... simply to make sure you use the python version the package was built against [08:52] else the world might break through incompatibilities [08:52] LordKow: yeah [08:52] or what i'd do is [08:52] change the executable name according to the env. [08:53] is that what ogra is saying? [08:53] (i did it with php) though [08:53] worked [08:53] if its for your own development and you want to use a newer python, use env ... and put the new python version into 7usr/local [08:53] or into ~ [08:53] (with LD_LIBRARY_PATH set for ~/lib though) [08:54] it's for packaging to the masses [08:54] then use the full path to the binary [08:54] so change the exe name to the env [08:54] that way you make sure john doesnt break it [08:55] yea [08:55] john owes me [08:55] well, the packaging policy :) [08:55] i think its written donw there somewhere [08:55] ogra [08:55] the library path has to be the lib you want to use with python [08:55] riht/ [08:55] right? [08:55] right [08:56] if you use ~ you might get into other weird issues though ... python will want ~/share too etc ... i would use /usr/local if i would want to use a different py version [08:57] if you keep python in /usr/share/backgrounds/python/blah... [08:57] LD_LIBRARY_PATH must be /usr/share/backgrounds/python/blah.../lib (or whatever you indicated when ./configureD ? [08:57] ogra: the user's home? [08:57] wouldnt the weird issues depend on the rest of the config and not just ~ [08:57] as i understand it some systems forbid to execute things in /home directories [08:58] and i think relative paths are icky for anything [08:58] ~ is a relative path [08:59] so folks [08:59] hakkav, ~ isn't relative to me $(echo ~) [08:59] im sorry? [09:00] ~ expands to an absolute path (by default) [09:00] it's kinda alias? [09:00] its like an alias? [09:01] that absolute path changes [09:01] i thought "relative path" always meant "relative [to the current directory] path" [09:01] maco: relative to the user that is currently executing the script [09:01] being careless [09:01] maco, Can be relative to some operating directory as well. [09:01] ~ doesn't change when you cd, only when you su [09:01] from one user to another [09:02] persia erm interesting point [09:03] everything is relative :P [09:03] ogra, You win :) [09:03] not me ... just quoting freaks with weird haircuts here ;) [09:04] easy i got a weird one [09:27] bigon: accepted [09:34] seb128: thx [09:35] you're welcome [09:39] seb128: could you also sync libchamplain and libchamplain-gtk which are in debian's incoming? [09:40] bigon: will sync when they reach the mirrors that's easier and there is no hurry [09:40] seb128: ok === Tonio__ is now known as Tonio_ === dpm_ is now known as dpm [11:12] odd question mayhaps, but why boost-library 1.35 still, when 1.39 is available since beginning of month, and 1.38 since a tad bit longer? === DaIRC28147 is now known as sea4ever [12:48] Riddell, ScottK: Debian has dropped boost1.35, and kde appears to still build-dep on it; is there a transition plan there for karmic? === azeem_ is now known as azeem [13:30] slangasek: We need to decide what the target is (and it's more than just KDE). It'll be 1.37 or 1.38. [13:30] * slangasek nods [13:31] boost-defaults seems to be pointing at 1.38, now [13:31] Last I looked 1.38 was still FTBFS in Ubuntu. [13:31] ah [13:32] I'd prefer 1.38 since that's what Debian is aiming for for Squeeze, but I don't know of that's Karmic or Karmic +1. [14:12] hi! [14:13] i am developing a daemon, is there any guidelines? i need mysql already running is there any standard way for that, like gentoo require? [14:15] iskywalker, Depends ought to do the latter. For the former, mostly try not to be root, and avoid non-local sockets by default. [14:18] persia: i was hoping a book or so, i must create a user and run the daemon on that useraccount, like www-data [14:20] iskywalker: have you seen the Debian policy manual? It's worth at least skimming [14:20] it is huge also :) === asac_ is now known as asac [14:20] iskywalker, Hrm. I don't know of one (but I'm not the best person to ask for book recommendations). I suspect you'll find more success from a more general forum, rather than something so Ubuntu focused (as I expect your daemon ought work for any linux, if not any unix) [14:21] :) that's why you skim it and know where things are, and later read the sections that apply to you [14:21] persia: sure, but i would maintain an ubuntu not an general daemon... [15:29] bryce: Bug 347569 http://ubuntu.pastebin.com/d58789f5b could I test this patch and if it works just post that to LP or would you prefer a debdiff? [15:29] Launchpad bug 347569 in xserver-xorg-video-ati "[RS482] glxgears low fps" [Undecided,Confirmed] https://launchpad.net/bugs/347569 [15:32] That was not ver gramatical of me... The question is would you prefer a .patch file or a debdiff? The .deb on this bug report made a significant differance in fps for me and I'd love to get this fix out to the greater good. [15:32] A debdiff ready to be uploaded is prefered. [15:32] alright [15:33] You might want to follow the original code’s indentation and bracket style. [15:35] I have a question though, how do you magical guys behind the curtian accully 'upload' to -updates? do you just upload teh debdiff and the build system applys it for you or do you need to apply is then submit the whole source deb to the build system? [15:35] The uploader will apply the debdiff to the extracted source package, build a new source package and upload that one. [15:36] I see, so me doing that extra work making the debdiff really does make a differance then? :) === cr3_ is now known as cr3 [16:32] hi all [16:58] If there's a buildd admin about I'd appreciate having qt4-x11 on armel and kdebase-workspace on ia64 and sparc rescored so I'll know in less than several days if I've fixed problems there. [17:12] I decided to test radeon rewrite and would love to help out a bit if possible. Who should I ping? === CarlFK1 is now known as CarlK [17:54] If I want to integrate my own app to the new notification system in Ubuntu 9.04, which library should I use? [18:03] henux: #ayatana is probably a better channel for that question. [18:07] superm1: slangasek: hmm, not sure if i know what you guys are talking about ... ? [18:08] superm1: i have a little hack that generates a md5sum "signature" of my hardware configuration, and the associated xorg [18:08] superm1: hardware configuration as shown by lshw [18:09] superm1: i do this nasty thing when i travel ... i physically move my hard drive from my 14" thinpad to my 11" thinkpad, for portability purposes [18:09] one of which has an nvidia card, and the other an intel card [18:10] superm1: and I have an init script that computes my current hw config, and looks for a "signed" xorg.conf that works with it [18:10] Bug #126774 would someone please check my statement and provide some guideance [18:10] Launchpad bug 126774 in apt "package update proceeds without adequate disk space" [Medium,Triaged] https://launchpad.net/bugs/126774 [18:20] kirkland, what do you do about libGL though? [18:23] superm1: that's the sore point ... install/uninstall [18:24] kirkland, so there are two solutions that i've been pondering about with that kind of scenario [18:24] 1) alternatives system for libGL.so.1 [18:24] 2) automatically try to use "nvidia" driver if it's available. there was a patch proposed for this a long time ago, but it didn't make it upstream [18:25] and doesn't work as it should currently [19:24] Any dpkg wizards around? I'm having a hard time working out the interactions between python-xen-3.3 and xen-utils-3.3 on upgrades === jdong_ is now known as jdong [20:16] hmm i started using ext4 and now i see 4 new bugs fixed in ext4 for 30-rc6 [20:16] * calc wonders if it was a good idea for him to switch to ext4 already [20:24] calc, "no". next question? [20:28] directhex: heh [20:28] directhex: i guess to late for me at least until after UDS === crd1b is now known as crdlb [21:01] Would there perchance be any cdbs experts around? I'm looking at the current Ubuntu delta in the python-distutils cdbs class, and it looks wrong to me [21:02] maxb: What specifically? [21:05] It replaces --root=$(cdbs_python_destdir) with --root=$(CURDIR)/debian/$(cdbs_curpkg) which seems a bit wrong, since the first is derived from potentially user-specified variables [21:06] What is the default for cdbs_python_destdir? [21:06] maxb: bug 374892 [21:06] Launchpad bug 374892 in cdbs "Use correct root path when converting dist- to site- in arch packages" [High,Fix released] https://launchpad.net/bugs/374892 [21:06] that's why the change was made it appears [21:07] I see the bug, but I'm not really convinced that the fix is corretc [21:08] I'm trying to get mercurial to build with the newer cdbs - which I have, by way of a slightly dodgy hack, but this Ubuntu change has broken it [21:10] So what would you recommend instead? [21:12] Well, the "obvious" solution would be to have fixed the disparity between the upstream rules and the added ubuntu-specific fragment by making the ubuntu-specific fragment conform [21:12] I'm just not sure if there are nonobvious problems with that [21:24] greetings [21:24] I am in the process of creating a front-end UI for espeak and other text-to-speech engines [21:25] Is there anywhere that has guidelines / requirements for what a package needs to support [21:26] in order for it to be included in Ubuntu? [21:27] reece, How do you mean? As in which engines, or something else? The main requirement (aside from packaging details) is that it compile and work on the current development release. [21:28] I mean things like being localised (which it isn't at the moment, but I am working on this) [21:29] Ok, so you just need to provide the details on how to build the package then? [21:38] reece, Being localised is better, of course :) But yeah, just have a well-written INSTALL file that explains how to build and install, and make sure your licensing is all correct. [21:39] ok, cool [21:41] does the standard autotool INSTALL file count? I am using an autogen.sh script to generate the configure, et. al. files [21:43] Adam Conrad here? [21:43] * Elbrus doesn't know his nickname [21:49] infinity: ping [21:50] bugging about bug 2253 [21:50] Launchpad bug 2253 in fpc "fpc needs bootstrapping on buildds" [Medium,Fix released] https://launchpad.net/bugs/2253 [21:50] or actually bug 67544 [21:50] Launchpad bug 67544 in fpc "PPC build of fpc fails" [Undecided,Confirmed] https://launchpad.net/bugs/67544 === fta_ is now known as fta === williamd is now known as TurtL3 [21:53] reece, As long as the file correctly describes how to do it, it doesn't matter if it's lovingly hand-crafted, or boilerplate. [21:54] ok, thanks [22:00] * maxb determines that the cdbs change is in fact a regression bug, and starts to put together a bug report [22:10] can someone explain something about uds? [22:10] if you don't ask probably not [22:11] hi, does anyone use xmonad? [22:11] mbana: hi, try #ubuntu [22:11] that won't help. there's something with the package i believe [22:12] so the question you want to ask is not if somebody uses it [22:12] i've even tried the official xmonad channel, they don't seem to have a solution either. i've been told it's packing issue or something wrong with gnome [22:12] what about asking your question directly? [22:12] because if no one responds, there's no point in me asking typing a long a detailed problem ;)? [22:13] mbana: perhaps someone reads log and give you the answer later? [22:13] and you expect people to reply if you don't ask the question? [22:13] those things happen (; [22:13] ok i'll ask on the ML. which one would you guys recommend? [22:13] you could probably have described the bug in the same time you are asking those question and explaining why [22:14] what about opening a bug on launchpad? [22:16] mbamford, i'm using xmonad [22:16] bah [22:16] mbana, ^ [22:17] maco: hi there, would you prefer that i file a bug report? [22:18] i think whomever packages it would prefer a bug report if its actually a bug [22:18] or even possible bug [22:18] what's going on that you think's a bug or that you'd like to have reproduced? [22:26] maco: https://bugs.launchpad.net/ubuntu/+source/xmonad/+bug/378111 [22:27] Ubuntu bug 378111 in xmonad "Xmonad refuses to start automatically on 9.04" [Undecided,New] [22:27] which desktop environment are you using? [22:27] it starts automatically in kubuntu just fine [22:27] and i believe it starts automatically for dtchen in ubuntu [22:27] maco: gnome. it was working perfectly in intrepid [22:28] well i had it working in 9.04 w/ gnome then reinstalled and switched to kde and it starts automatically just fine in kde too. dtchen's using it in gnome and it starts automatically for him [22:29] something changed in gnome so you have to set it up differently, but it does work [22:29] i think there are directions on xmonad.org for how to make it work with current gnome [22:29] it's interesting to note that i used the jaunty instructions whilst on intrepid and it workedd, here's the link http://www.haskell.org/haskellwiki/Xmonad/Using_xmonad_in_Gnome#Ubuntu_Jaunty [22:30] did you blast away .gconf? [22:30] hell no [22:31] did you read the paragraph you linked that says intrepid's .gconf isn't compatible with jaunty? [22:32] i'm not using .gnomerc. it seems rather odd that i have to clear out the entire .gconf [22:32] ohok [22:32] i didnt have to get rid of it...i think i just put the "killall metacity ; xmonad &" in the autostart thing [22:32] autostart thing? [22:37] does this apply to all window managers set in gconf, not just xmonad? [22:38] (that upgrading from Intrepid to Jaunty causes you to lose your WM?) [22:42] i can't confirm, i only use xmonad === ripps_ is now known as ripps [23:03] mbana, system -> preferences -> session -> autostart [23:08] maco: i'm afraid i don't have that. do you mean startup applications? [23:08] yes [23:08] and i'm AFK now === zorael is now known as Zorael === Arand is now known as ienorand === ienorand is now known as arand