WebHosting Paid by #1Payday.Loans
[00:09] <th> "works"? [00:09] <th> it's meant to be copied [00:12] <fake> i thought it's copied by udevd [00:12] <fake> nevermind [00:25] <th> 334 builds total, 205 completed fine, 0 with errors. [00:26] <fake> th: is that the one from before? [00:26] <fake> th: quite fast... [00:26] <th> sure it is [00:26] <fake> now it's time for kde-* and koffice, hehehe [00:46] blindcod1r (n=blindcod@tor/session/x-d0e807e8e3daf440) joined #rocklinux. [00:46] blindcoder (i=id@tor/session/x-eac77c3f5c7ee0a0) left irc: Nick collision from services. [00:47] Nick change: blindcod1r -> blindcoder [00:51] BoS|afk (n=BoS@dslb-088-072-034-227.pools.arcor-ip.net) left irc: Remote closed the connection [00:51] BoS|afk (n=BoS@dslb-088-072-034-217.pools.arcor-ip.net) joined #rocklinux. [01:04] kasc_ (n=kasc@dslb-084-060-104-223.pools.arcor-ip.net) joined #rocklinux. [01:16] <th> uh - crystal is building qt33 and qt40? [01:16] <fake> o_O [01:18] kasc (n=kasc@dslb-084-060-102-179.pools.arcor-ip.net) left irc: Read error: 110 (Connection timed out) [01:18] Nick change: kasc_ -> kasc [01:58] <fake> th: [01:58] <fake> ping [02:16] <th> pong [02:16] <th> sorry for the delay [02:16] <th> still there, fake? [02:16] <fake> th: don't you thing syslogd/klogd belongs into minimal.set ? [02:16] <fake> err, basic.set ? [02:16] <th> no [02:16] <fake> okay [02:17] <th> i dont even think that it belongs into bbs [02:17] <th> (which is a superset thereof) [02:19] <th> 334 builds total, 229 completed fine, 0 with errors. [02:21] <fake> udev patch is coming along just fine. [02:22] <th> https://www.club-mate-shop.de [02:23] <fake> ARGH! [02:23] <fake> deutschlandweiter [02:23] <fake> argh argh argh! [02:23] <th> da zaehlt doch oesterreich bestimmt dazu ;) [02:24] <fake> die seite is ja urcool bitte [02:24] <th> https://www.club-mate-shop.de/folderblog/fb.php?p=image/micra_versand/ [02:25] <th> 334 builds total, 231 completed fine, 0 with errors. [02:25] <th> go go go [02:25] <fake> hihi, erinnert mich an was ;) [02:27] <fake> oops... export LANG=C [02:30] <fake> i just ordered to vienna... let's see... [02:57] <th> ok i better be off now [03:05] <fake> kay, n8 th [03:29] <fake> i guess we can remove the CORE flag from devfsd ... *sigh*... sad [05:12] netrunner (n=andreas@anvame.net) left irc: Read error: 110 (Connection timed out) [05:14] netrunner (n=andreas@anvame.net) joined #rocklinux. [07:30] <blindcoder> moin [07:37] <fake> moin blindcoder [07:37] <fake> and n8 [07:38] <blindcoder> heh [07:39] <fake> have a nice day ;) [07:49] Nick change: BoS|afk -> _BoS_ [08:42] daja77_ (n=daja77@dslb-088-072-035-235.pools.arcor-ip.net) joined #rocklinux. [09:00] daja77 (n=daja77@dslb-088-072-035-235.pools.arcor-ip.net) left irc: Read error: 110 (Connection timed out) [09:49] ija_ (n=ija@84.19.222.180) joined #rocklinux. [09:50] ija (n=ija@212.80.242.172) left irc: Nick collision from services. [09:50] Nick change: ija_ -> ija [10:45] <th> moin [10:45] <th> 334 builds total, 325 completed fine, 0 with errors. [10:45] <th> damn that qt+kde takes a good load of time [11:06] <th> build will finish any minute now. [11:08] <th> == 11:10:34 =[7]=> Building base/man-pages [2.17 0]. [11:08] <th> yea - go go go [11:12] <th> 334 builds total, 334 completed fine, 0 with errors. [11:33] <blindcoder> sweet [11:36] <th> ok - iso is transferring to my machine to be tested, [11:36] <th> next build is already started [11:36] <th> this time only boot+bbs [11:37] <th> which is fake's plus my journal merged. [11:47] <esden> moin [11:47] <esden> ping blindcoder [11:48] <blindcoder> esden: pong [11:48] <esden> blindcoder: I have problems with pallas [11:48] <blindcoder> define problems [11:48] <esden> see query [11:57] <owl> moin [12:23] <blindcoder> moin owl [12:32] <esden> moin owl [14:06] <owl> hi esden , blindcoder [14:23] [raphael] (n=raphael@raphael.netpark.at) left irc: "using sirc version 2.211+KSIRC/1.3.12" [14:29] <th> clifford: smng making progress? [14:36] <th> fake: ping [14:56] <clifford> th: yep. [14:56] <clifford> but not very fast.. ;-) [14:56] <th> clifford: i'm so thrilled ;) [14:57] <th> clifford: btw - do you have any builds running yourself? [14:57] <th> clifford: seems that fake and me are making good progress. [15:00] <daja77_> need a hand in running builds? [15:00] <blindcoder> yes, it finally feels like steady progress again \o/ [15:02] <th> daja77_: perhaps you could investigate the /opt/kde -> /usr change? [15:02] <th> daja77_: the desktop-files issue? [15:03] <daja77_> uhm have to read again what the issue was [15:03] <blindcoder> the way things go now I think I can rm -rf /home/svn/rock-community/ [15:03] <th> it's on rock-devel [15:03] <daja77_> th: i know, mom [15:08] <daja77_> i am blind i don't find it [15:10] <daja77_> ah now [15:15] <blindcoder> narf, I hate this proxy... [15:15] <blindcoder> it doesn't resolve domain names unless they have at least three dots [15:16] <blindcoder> so rocklinux.org. doesn't resolve [15:16] <blindcoder> but www.rocklinux.org. does [15:21] <owl> aren't you admin there? [15:21] <blindcoder> at siemens? [15:21] <daja77_> zfone beta for linux and macos ^^ [15:21] <owl> hrm. [15:21] <blindcoder> thank god I'm not! [15:31] <blindcoder> th: strange... I never do anything more than the X -configure [15:32] <th> it's a vmware [15:32] <th> i think it works with option 3.) [15:36] <blindcoder> vmware works in lvp [15:36] <blindcoder> though with a resolution 'jenseits von gut und boese' :) [15:39] <th> well [15:39] <th> so what's different there? [15:42] <blindcoder> X linked statically [15:43] <th> cant make that difference [15:43] <th> old vmware? [15:44] <blindcoder> mabye [15:56] <stf^rocklinux> hm, I get errors compiling the qemu 0.8.0 i386-user, arm-user and sparc-user emulators with gcc34 ... [15:57] <daja77_> +var_append confopt " " "--target-list=i386-softmmu" [15:57] <daja77_> my current qemu.conf diff .. [15:57] <stf^rocklinux> daja: wouldn't that only build the i386-softmmu emulator? [15:58] <daja77_> yes .. [15:58] <stf^rocklinux> daja: to work around the errors, I used the complete target list without the offending ones... [15:58] <daja77_> ok [15:58] <stf^rocklinux> which is: --target-list=armeb-user,ppc-user,mips-user,mipsel-user,i386-softmmu,ppc-softmmu,sparc-softmmu,x86_64-softmmu,mips-softmmu,arm-softmmu [15:59] <stf^rocklinux> didn't find any info about it in the QEMU forum, though... [16:34] nookie (n=nookie@85-124-51-55.dynamic.xdsl-line.inode.at) joined #rocklinux. [16:35] nookie (n=nookie@85-124-51-55.dynamic.xdsl-line.inode.at) left irc: Client Quit [16:43] daja77 (n=daja77@dslb-088-072-048-209.pools.arcor-ip.net) joined #rocklinux. [16:55] daja77_ (n=daja77@dslb-088-072-035-235.pools.arcor-ip.net) left irc: Read error: 110 (Connection timed out) [16:56] stf^rocklinux (n=user@heim-034-205.raab-heim.uni-linz.ac.at) left irc: Read error: 110 (Connection timed out) [17:07] stf^rocklinux (n=user@heim-034-205.raab-heim.uni-linz.ac.at) joined #rocklinux. [17:16] <daja77> *sigh* [17:19] <th> fake: ping! [17:46] stf^rocklinux (n=user@heim-034-205.raab-heim.uni-linz.ac.at) left irc: "using sirc version 2.211+KSIRC/1.3.12" [17:48] stf^rocklinux (n=user@heim-034-205.raab-heim.uni-linz.ac.at) joined #rocklinux. [17:56] <th> oh [17:56] <th> grub-1.92 [17:56] <th> that's pretty much different to me [17:56] <th> compared with 0.96 in stable [17:56] <th> is there no menu.list file any longer? [17:56] <th> do i have to create images? [17:57] <daja77> menu.lst is obsolete, it is called grib.conf iirc [17:57] <th> where will i find that file? [17:57] <daja77> /boot/grub [17:58] <th> ahh .cfg [17:58] <th> so what is grub-mkimage doing? [17:58] <th> "Make a bootable image of GRUB" [17:58] <th> gmm [17:58] <th> hm [17:58] <th> is it still all about stage 1 1.5 and 2? [17:59] <th> there are MODULES now. [17:59] <daja77> yep [17:59] <th> damn i need a crash-course [17:59] <daja77> and without them it can't do anything [17:59] <th> it reads them from filesystem? [17:59] <th> or must it be in some image? [17:59] <daja77> image [17:59] <th> only option? [18:00] <th> is that image the former stage2? [18:00] <daja77> ask blindy or clifford which modules you need [18:00] <th> that's not my concern right now. i try to understand the general architecture of it. [18:01] <daja77> ok [18:01] <daja77> i think so, but i am no expert in grub2 [18:01] <th> i saw the installer create an image [18:01] <daja77> yes [18:01] <th> that's propably stage2 [18:01] <th> that would sort of rock [18:02] <th> one had to rebuild grub for changes in 0.96 [18:02] <th> there were these 1.5 stages for different filesystems [18:02] <th> i guess that's where the modules come from [18:10] <daja77> argh [18:29] <SMP> you link a number of modules into a grub image that gets booted. provided you have the necessary modules linked in, you will also be able to load more modules at runtime from the filesystem [18:30] <SMP> it's all very flaky and undocumented, thus my resistance about adopting grub2 back when it was committed [18:31] <th> i see [18:37] <daja77> this is interesting stf's kernel update results in a cksum error [18:39] <stf^rocklinux> daja: that's probably because I created the download files manually with (old version & update patch) [18:39] <stf^rocklinux> daja: my internet connection is to slow atm for full downloads ... [18:39] <stf^rocklinux> daja: feel free to rediff my kernel update patch :) [18:39] <daja77> well but in that case the cksums are always different [18:39] <daja77> so better supply no cksum [18:41] <stf^rocklinux> daja: ok [18:41] <daja77> even when you unpack the archive and repack it, the cksum is different ;) [18:42] <daja77> don't worry [18:42] <stf^rocklinux> daja: with xdelta diffs it would have worked ^^ [18:42] <daja77> :) [18:45] <daja77> cat: package/base/wine/wine.desc [18:45] <daja77> no such file or directory [19:20] blindcoder (n=blindcod@tor/session/x-d0e807e8e3daf440) left irc: Read error: 113 (No route to host) [19:25] blindcoder (n=blindcod@tor/session/x-54f816fd3258942b) joined #rocklinux. [20:31] <th> !> Please create a man page for this tool. [20:31] <th> what's this? [20:33] <daja77> O_o [21:14] <th> ln: creating symbolic link `/lib/udev/devices/fd/fd' to `/proc/self/fd': No such file or directory [21:14] <blindcoder> -f [21:15] <blindcoder> hm, no [21:15] <th> yea [21:15] <th> you see that it's wrong [21:15] <blindcoder> yes [21:15] <th> ln -sv /proc/self/fd $root/lib/udev/devices/fd [21:15] <th> i bet $root/lib/udev/devices/fd already exists as directory [21:15] <blindcoder> yes [21:15] <th> so it should be checked for existence first [21:16] <th> but where does it come from? [21:28] <th> ahh [21:29] <th> propably because of prior stage [21:29] <th> cause i do stage9 [21:32] <th> perhaps it would be better to remove the whole devices directory before starting to fill it [21:32] <blindcoder> perhaps not [21:32] <th> why? [21:33] <blindcoder> there are devices that lack code for device node creation but provide the node anyway [21:33] <blindcoder> how would we create them sanely? [21:33] <th> haeh? [21:33] <th> you did not understand [21:33] <th> i'm talking of creating /lib/udev/devices/ [21:34] <blindcoder> yes [21:34] <blindcoder> so what would be a sane way of creating device nodes for faulty modules? [21:34] <th> mknod [21:34] <th> i dont see the problem you see [21:35] <blindcoder> where? manually via initrd plugins? [21:35] <th> are you talking of devices you need in early initrd stage? [21:35] <th> or later in productive stage? [21:36] <blindcoder> both [21:36] <blindcoder> especially my cm_4000 [21:36] <th> what's that? [21:36] <blindcoder> pcmcia cardreader [21:37] <th> ok. [21:37] <th> so you want a device for it in /dev [21:37] <th> so you want to have the device in /lib/udev/devices, right? [21:37] <blindcoder> yes [21:37] <blindcoder> well, that's what that directory is for, isn't it? [21:37] <th> so you need to modify your initrd, right? [21:37] <th> - The content of /lib/udev/devices directory which contains the nodes, [21:37] <th> symlinks and directories, which are always expected to be in /dev, should [21:37] <th> be copied over to the tmpfs mounted /dev, to provide the required nodes [21:37] <th> to initialize udev and continue booting. [21:37] <th> it's not for faulty devices (only) [21:38] <blindcoder> hmm, okay [21:38] <blindcoder> then I guess I can use a initrd plugin to call mknod and put a 'this is sane' tag on it :) [21:38] <th> but keep on playing this game. you need to modify your initrd in any case, cause you need it in stage1, right? [21:38] <blindcoder> yes [21:39] <blindcoder> for reading a gnupg card, reading the key from it and using it to decrypt the key holding the passphrase for my filesystem :) [21:39] <th> the point is: what is the problem if i rm -rf /lib/udev/devices/ in the udev.conf? [21:39] <th> it makes no difference [21:39] <blindcoder> yes [21:40] <blindcoder> I was thinking about hooking into that call to create device nodes for these cases from the udev package [21:40] <th> so please explain again to me: what's the problem with removing that directory before filling it? [21:40] <th> from the udev package [21:40] <th> hmmmmm [21:40] <th> so you create your hooks and then build udev? [21:40] <th> i think it should happen when you create your initrd [21:41] <blindcoder> okay, so much for initrd [21:41] <th> so like mkinitrd just copies devices in /l/u/devices [21:41] <blindcoder> now suppose i don't use it in initrd [21:41] <th> yea [21:41] <th> so what's our scenario now? [21:41] <blindcoder> or I have filled that directory with custom nodes and then rocket emerge udev [21:42] <th> obviously you dont need it in stage1 [21:42] <th> then there is this strange backup-before-build mechanism, no? [21:42] <blindcoder> I'm not familiar with that mechanism [21:43] <blindcoder> scenario: [21:43] <blindcoder> install system [21:43] <th> neither am i - but you always have this problem [21:43] <th> if you copy/create config files for example [21:43] <blindcoder> yse [21:43] <blindcoder> yes [21:43] <th> that's not specific to udev [21:43] <th> ok install system [21:43] <th> go ahead [21:43] <blindcoder> but I don't think we do rm -rf ${root}/important/directory somewhere [21:43] <blindcoder> install system [21:43] <blindcoder> go to /lib/udev/devices [21:43] <blindcoder> create a dozen-odd custom nodes [21:43] <blindcoder> rocket emerge udev [21:43] <blindcoder> rm -rf /lib/udev/devices [21:44] <th> ok - understood [21:44] <blindcoder> will the backup mechanism keep my nodes? [21:44] <th> i dont know. but wait [21:44] <th> - The content of /lib/udev/devices directory which contains the nodes, [21:44] <th> symlinks and directories, which are always expected to be in /dev, should [21:44] <th> be copied over to the tmpfs mounted /dev, to provide the required nodes [21:44] <th> to initialize udev and continue booting. [21:44] <th> let's read it again [21:44] <th> and again [21:44] <th> i think it is meant as a bare necessaty skeleton [21:44] <th> AND... [21:45] <th> we require a mechanism of copying this while booting [21:45] <th> so we do mount /dev [21:45] <th> cp /lib/udev/devices /dev [21:45] <th> why not hook something after that? [21:45] <th> /lib/udev/user-devices [21:45] <blindcoder> hmm [21:45] <blindcoder> either that [21:46] <th> i always want to have a SANE skeleton [21:46] <blindcoder> or put in a file called 'WARNING' which explains how to set up devices correctly without having to fear that a package update will kill them [21:46] <th> yea [21:46] <blindcoder> ie: /etc/conf/initrd and /etc/rcS.d [21:46] <th> so a WARNING should be placed when doing rm -rf [21:46] <th> i agree [21:46] <th> well i'm not too sure that rm -rf would be a good idea [21:47] <th> neither am i sure where to put user nodes [21:47] <th> perhaps /etc/conf/udev.d [21:47] <blindcoder> [ -e /lib/udev/devices/$device ] || mknod ? [21:47] <th> (scripts for bootup) [21:47] <th> blindcoder: that's (test -e) is how i currently fixed it [22:07] <th> btw... [22:07] <th> why are we creating links here? and not nodes? [22:08] <th> oh [22:08] <th> it's links to /proc [22:08] <th> no matter [22:08] <th> stupid me [22:14] <stf^rocklinux> th, blindcoder: /lib/udev/devices should contain entries for devices that never generate a hotplug event (like initial console, but also buggy drivers) [22:14] <stf^rocklinux> s/should/is meant to/ [22:18] <th> and proc symlinks? [22:21] <stf^rocklinux> it's for symlinks too [22:22] <th> ok. then everything is fine [22:26] <stf^rocklinux> IMO /lib/udev/devices files should also be copied to /dev during udev build [22:26] <stf^rocklinux> so you can e.g. boot with init=/bin/bash [22:27] <blindcoder> during build? [22:27] <blindcoder> that won't help you much after a reboot [22:29] <stf^rocklinux> yes so there is always a minimal static /dev directory, even when udev is not started in init [22:30] <th> uhhhh hwscan did not detect BusLogic [22:37] <th> ok - and our installer is broken. i've IDE+SCSI disk in my system and i only see the IDE disk in the installer [22:40] <blindcoder> that'd be still the udev scripts? [22:43] <th> ugh i was wrong [22:43] <th> it has nothing to do with only one of two is displayed [22:43] <th> the scsi disk is not displayed even if alone [22:44] <th> yea propably it's searched somewhere, where we have no symlink [22:44] <blindcoder> anyway [22:44] <th> i dont know how the installer searches [22:44] <blindcoder> sleep happens [22:44] <th> to whom? [22:44] <blindcoder> it searches in /dev/discs/ only [22:44] <th> to the weak? [22:44] <th> /dev/discs? [22:44] <th> that must die [22:44] <th> (not the /dev/discs/) [22:45] <th> (only the fact that that's the place for the search) [22:45] <blindcoder> why? it should contain all discs [22:45] <blindcoder> aaanyway [22:45] <blindcoder> sleep happens to me now [22:45] <blindcoder> bye [22:45] <th> /dev/disks is old [22:45] <th> and obsolete [22:45] <th> not that i dislike it [22:46] <th> but it should not be our primary source of information [22:46] <th> perhaps something in /sys [22:46] <th> or /proc/partitions [22:50] <th> 22:47:10 < blindcoder> it searches in /dev/discs/ only [22:50] <th> that's plain wrong. [22:50] <th> i've no /dev/discs/ at all - and it found the IDE [22:52] <th> i bet it is /dev/{ide,scsi} [22:52] <th> and our scsi is messed cause of our bad udev helper scripts [23:03] <th> ok - i think the disk-offering part needs a complete rewrite. [23:05] <stf^rocklinux> hm... [23:07] <th> it's some sort of /dev/discs/* and black magic. [23:08] <th> i think it could be based on /proc/partitions [23:08] <th> or something in sys [23:08] <th> /sys that is [23:09] <stf^rocklinux> look in target/bootdisk/linuxrc.sh, getcdromdevice () [23:09] <stf^rocklinux> devicelists="/dev/cdroms/* /dev/floppy/*" [23:09] <stf^rocklinux> these devices are scanned and added to the list in stage 1 [23:10] <th> there must be more [23:10] <th> somewhere [23:10] <th> my ide disk was in without any /dev/discs [23:10] <th> (i'm talking of the very first dialog, where you define the partitions and mount poitns) [23:11] <stf^rocklinux> ah, in stone [23:11] <th> yea [23:13] <stf^rocklinux> you mean here: Partitioning your discs [23:13] <stf^rocklinux> This dialog allows you to modify your discs partition layout and to create filesystems and swap space - as well as mounting [23:13] <stf^rocklinux> / activating it. Everything you can do using this tool can also be done manually on the command line. [23:13] <th> yea [23:13] <stf^rocklinux> that's in /etc/stone.d/mod_install.sh [23:13] <th> yes [23:13] <th> i know. [23:13] <th> i had a look [23:14] <th> found the place and it looked like it's only using /dev/discs/ [23:14] <th> which cant be true [23:15] <th> i tried to ignore all this. accepted the "No hard-disk found" and went over to install [23:15] <th> fdisking by hand and mounting at /mnt/target [23:15] <th> but guess what. [23:15] <th> it refuses somehow to mount my cdrom [23:15] <stf^rocklinux> ? [23:15] <th> i did that by hand. [23:15] <th> and installer always umounts it [23:15] <stf^rocklinux> so what device node is created for your discs? [23:16] <th> /dev/sda2 [23:17] <th> /dev/sda{,1,2} [23:17] <th> to be exact [23:17] <th> and even /dev/scsi/.../{generic,part1,part2} (no disc) [23:17] <th> uhh and the installer fails to install tar.bz2... only knows about gems i assume [23:22] <th> ok i did the foreach-tar thingie... [23:23] blindcod1r (n=blindcod@tor/session/x-3f8baf35a8f82303) joined #rocklinux. [23:24] blindcoder (n=blindcod@tor/session/x-54f816fd3258942b) left irc: Nick collision from services. [23:24] Nick change: blindcod1r -> blindcoder [23:27] <th> " ** dumping you into an emergency shell" THANKS that's a beginning [23:30] <th> where am i? [23:30] <th> some sort of stage1 [23:30] <th> there is no "modprobe" to load BusLogic [23:31] <fake> th: one cigarette and i'm all yours [23:32] <th> fake: great! [23:32] <fake> th: that 'dumping you' is from the initrd. [23:32] <fake> th: if you need to load a module, you have to either select it in the stone dialog, or add it to /etc/conf/kernel [23:32] <fake> th: after that run 'mkinitrd' again. [23:32] <fake> th: then the module will be included in the initrd.... [23:33] <fake> th: (selecting it in the stone hardware dialog adds it to /etc/conf/kernel, so...) [23:33] <th> i should do that after extracting my tars which the installer could not? [23:33] <th> i had to SIGINT stone [23:33] <fake> th: the gasgui installer is for .gem packages [23:33] <th> ok [23:33] <fake> th: the bize installer is for .tar.bz2 packages [23:34] <th> bize was not offered to me [23:34] <fake> th: you don't have it in your package selection i assume [23:34] <th> might be [23:34] <th> is. [23:34] <fake> (though i never tried bize...) [23:35] <fake> th: lemme just have cigarette, then i'll go over my scrollback [23:35] <th> yea [23:36] <th> ok - so it seems i'm stuck in my initrd cause i forgot to add BusLogic module [23:42] <fake> th: teamspeak? [23:42] <fake> nope [23:42] <fake> argh [23:42] <fake> wrong window [23:42] <stf^rocklinux> th: can you access the CD-ROM? maybe the module is on the CD? [23:44] <th> stf^rocklinux: well i'll just boot from the boot-target and add the module [23:45] <fake> brb [23:46] <fake> si [23:46] <fake> so [00:00] --- Thu Mar 16 2006