T2 IRC Log: 2008-04-12
This is the log as captured by an IRC bot in the channel. The statements are those of the individual people and might not neccessarily reflect the policy and legal rules as set forth by the T2 SDE Project.
--- Log opened Sat Apr 12 00:00:58 2008
01:22 -!- Baldzius [n=Baldzius@87.198.192.226] has quit ["Leaving"]
02:36 -!- Dallu1 [n=earl@fgfw.basis.is] has joined #t2
02:38 -!- Dallur [n=earl@fgfw.basis.is] has quit [Read error: 110 (Connection timed out)]
03:18 -!- xaw [n=xaw@unaffiliated/xaw] has quit ["leaving"]
04:20 -!- Ragnar|away [i=loki@D8FA4DE6.ptr.dia.nextlink.net] has quit [Remote closed the connection]
04:20 -!- Ragnar|away [i=loki@D8FA4DE6.ptr.dia.nextlink.net] has joined #t2
08:30 -!- Enqlave [n=stealth@sourcemage/guru/stealth] has quit [Read error: 104 (Connection reset by peer)]
08:50 -!- elias` [n=me@unaffiliated/elias/x-342423] has quit [Read error: 145 (Connection timed out)]
08:59 -!- elias` [n=me@unaffiliated/elias/x-342423] has joined #t2
10:17 -!- Netsplit kornbluth.freenode.net <-> irc.freenode.net quits: LMJ
10:18 -!- Netsplit over, joins: LMJ
10:30 < rxr> moin
10:38 < CIA-53> rene * r28681 /trunk/package/editors/ne/ne.desc: * updated ne (1.42 -> 1.43)
10:40 < CIA-53> rene * r28682 /trunk/package/x86/syslinux/syslinux.desc: * updated syslinux (3.62 -> 3.63)
10:43 -!- Netsplit kornbluth.freenode.net <-> irc.freenode.net quits: LMJ
10:43 -!- Netsplit over, joins: LMJ
11:37 -!- Netsplit kornbluth.freenode.net <-> irc.freenode.net quits: LMJ
11:38 -!- Netsplit over, joins: LMJ
11:53 < rxr> re
12:03 -!- Enqlave [n=stealth@sourcemage/guru/stealth] has joined #t2
12:15 -!- Enqlave [n=stealth@sourcemage/guru/stealth] has quit ["Dont wait to die to find paradise..."]
12:51 -!- Enqlave [n=stealth@sourcemage/guru/stealth] has joined #t2
12:58 -!- Enqlave [n=stealth@sourcemage/guru/stealth] has quit ["Dont wait to die to find paradise..."]
13:01 -!- Enqlave [n=stealth@sourcemage/guru/stealth] has joined #t2
14:23 < |beowulf|> rxr: Apr 11 23:51:11 vilya agetty[939]: /dev/ttyS0: No such file or directory <-- that's the last line in syslog on the cube before I turned it off because it didn't start agetty on the serial port
14:28 < rxr> hi
14:28 < rxr> |beowulf|: I just clicked on "send" regarding the article
14:28 < |beowulf|> hi :)
14:28 < rxr> so now turning to qemu and the initrd-less boot issue ...
15:47 -!- Enqlave [n=stealth@sourcemage/guru/stealth] has quit ["Dont wait to die to find paradise..."]
18:11 < CIA-53> rene * r28683 /trunk/package/xorg/xf86-video-xgi/ (xf86-video-xgi.cache xf86-video-xgi.desc): * updated xf86-video-xgi (2007-01-02 -> 1.5.0)
18:40 -!- dsoul [i=darksoul@vice.ii.uj.edu.pl] has quit [SendQ exceeded]
18:48 [Users #t2]
18:48 [@ChanServ] [ Dallu1 ] [ mqueiros_ ] [ sparc-kly_] [ |beowulf|]
18:48 [ __init__] [ dogmeat] [ mtr ] [ tfl0pz ]
18:48 [ axionix ] [ elias` ] [ Ragnar|away] [ TobiX ]
18:48 [ CIA-53 ] [ LMJ ] [ rxr ] [ valentin ]
18:48 -!- Irssi: #t2: Total of 17 nicks [1 ops, 0 halfops, 0 voices, 16 normal]
19:00 -!- Baldzius [n=Baldzius@89.127.178.76] has joined #t2
19:00 < Baldzius> moin
19:06 < rxr> hey Baldzius
19:07 < Baldzius> hi rxr
19:07 < Baldzius> trying libtool 2.2.2 locally
19:15 < rxr> so
19:16 < rxr> |beowulf|: my VM now boots without initrd ...
19:18 < rxr> http://people.exactcode.de/~rene/inittab -> etc/inittab
19:18 < rxr> http://people.exactcode.de/~rene/system -> sbin/init.d/system
19:21 < rxr> err sorry
19:21 < rxr> not inittab
19:22 < rxr> initscript
19:25 < |beowulf|> rxr: where should I put initscript?
19:25 < rxr> etc/initscript
19:28 < CIA-53> rene * r28684 /trunk/package/base/ (3 files in 2 dirs): * re-implemented support for booting without initrd
19:29 < |beowulf|> ah, very nice
19:29 < |beowulf|> I have a getty on the serial port
19:29 < |beowulf|> thank you! :)
19:29 < rxr> you're welcome !
19:30 < |beowulf|> there are still a few messages about missing /dev/console during boot though
19:31 < rxr> many ? or just one ?
19:32 < |beowulf|> http://www.hawo.stw.uni-erlangen.de/~siflfran/qube.8.log
19:33 < |beowulf|> and networking didn't come up
19:34 < |beowulf|> if I call dhclient manually I get an IP address
19:34 < rxr> have you updated the sbin/init.d/system ?
19:35 < |beowulf|> yes
19:36 < rxr> what does this output on that cube:
19:36 < rxr> grep "/dev " /proc/mounts
19:36 < |beowulf|> dev /dev tmpfs rw 0 0
19:36 < rxr> hm
19:37 < rxr> your log has these:
19:37 < rxr> ln: cannot remove `mouse': Read-only file system
19:37 < rxr> rm: cannot remove `fd': Read-only file system
19:37 < rxr> ln: cannot remove `core': Read-only file system
19:37 < rxr> ln: creating symbolic link `fd/fd': No such file or directory
19:37 < rxr> ln: cannot remove `stdin': Read-only file system
19:37 < rxr> ln: cannot remove `stdout': Read-only file system
19:37 < rxr> ln: cannot remove `stderr': Read-only file system
19:37 -!- teatime07 [n=jan@c-69-138-57-167.hsd1.tn.comcast.net] has joined #t2
19:37 < rxr> I wonder where it comes from
19:37 < |beowulf|> screen complains that it cannot find a pty
19:38 < rxr> there must still be something wrong or imperfect regarding the /dev setup ... :-(
19:38 < rxr> maybe remounted ?
19:42 < |beowulf|> which package do I have to install to get the host command?
19:42 < |beowulf|> ./scripts/Download: line 856: /dev/fd/62: No such file or directory
19:44 < rxr> that is the same issue as on boot
19:45 < rxr> something is not yet right with the /proc /dev and the /proc/self -> /dev/fd symlink
19:48 -!- teatime07 [n=jan@c-69-138-57-167.hsd1.tn.comcast.net] has quit ["weeeeee..."]
19:53 < rxr> |beowulf|: what does this output grep "/dev " /etc/fstab ?
19:54 < |beowulf|> dev /dev tmpfs defaults 0 0
19:55 < |beowulf|> http://www.hawo.stw.uni-erlangen.de/~siflfran/fstab
19:55 < rxr> can you add an "cat /proc/mounts"
19:55 < rxr> to /sbin/init.d/system
19:55 < rxr> before the line
19:55 < rxr> "Configuring the /dev filesystem ..."
19:58 < |beowulf|> http://www.hawo.stw.uni-erlangen.de/~siflfran/qube.9.log
20:03 < rxr> hm - there is no /dev mounted
20:18 < CIA-53> rene * r28685 /trunk/package/base/ (sysfiles/etc_initscript.txt sysvinit/system.init):
20:18 < CIA-53> * kill udevd after populating the initial /dev when booting
20:18 < CIA-53> without initrd as done inside the initrd/init, likewise
20:18 < CIA-53> * removed now phased out udevd killing in the
20:18 < CIA-53> sysvinit/system.init, the initrd does not yield a running
20:18 < CIA-53> udevd for some months, now
20:20 < rxr> |beowulf|: a workaround could be:
20:20 < rxr> --- etc/initscript.orig 2008-04-12 18:19:44.000000000 +0000
20:20 < rxr> +++ etc/initscript 2008-04-12 18:19:57.000000000 +0000
20:20 < rxr> - mount -n /dev
20:20 < rxr> + mount -n -t tmpfs none /dev
20:21 < Ragnar|away> shouldn't that be default anyway?
20:21 < rxr> Ragnar|away: depends ...
20:21 < rxr> yes, we could hardcode it at that place
20:21 < rxr> |beowulf|: wait with testing
20:21 < rxr> failed in my vm
20:21 < rxr> just a moment
20:21 < |beowulf|> too late...
20:22 < rxr> Ragnar|away: our initrd is setting it up that way anyway, but |beowulf| is booting without initrd and thus the initrd-less patchery right now
20:22 < Ragnar|away> yea
20:22 < Ragnar|away> well for some hosts I'd also like to be able to boot without initrd
20:23 < rxr> with the trunk patches it should be possible to boot with either a static (on-disc) device node setup or udev, just as needed
20:23 < rxr> module |beowulf|'s box's hickup
20:23 < rxr> modulo ...
20:23 < Ragnar|away> :)
20:24 < rxr> |beowulf|: I guess that change did not help on your box either ?
20:25 < |beowulf|> no, didn't change anything
20:25 < Ragnar|away> oh btw ... what I'd really like would be to be able to emerge a 32 bit package on a 64 bit system and a 64 bit package on a system that also has some 32 bit libs present without library-messyness ... can you think of a way to do that with not too big effort?
20:27 < rxr> for single packages that could work easier
20:27 < rxr> but for big packages with lots of deps that is trickier
20:27 < rxr> e.g. emerging the 32bit wine on a otherwise fully 64bit system ...
20:28 < rxr> what kind of packages would you want that for ?
20:28 < rxr> (wine ?)
20:28 < rxr> ah
20:28 < Ragnar|away> wine, firefox (flash without 32-64bit wrapper)
20:28 < Ragnar|away> firstclass client
20:28 < rxr> |beowulf|: can it be that you have some devices nodes from earlier tinkering in /dev?
20:29 < |beowulf|> yes, there are some
20:29 < rxr> |beowulf|: rm the on-disk nodes I remeber you put there (dev/console, dev/null etc?)
20:29 < |beowulf|> should I remove them?
20:29 < rxr> yeah - for that case the new code does not mount another /dev then
20:29 < rxr> as it expects you have a full device node collection in that case, like the good old pure static setup
20:29 < |beowulf|> should I delete everything in /dev? or just /dev/{console,null,zero}?
20:30 < rxr> yes, would be better if you like to use a new dynamic setup
20:30 < rxr> if you like to have static nodes, you would have to add the other missing ones :-)
20:30 < rxr> all or nothing :-)
20:32 < Ragnar|away> makes sense =)
20:36 < |beowulf|> rxr: the system I built had the following in /dev:
20:36 < |beowulf|> fd@ loop/ null random urandom zero
20:36 < |beowulf|> should I leave these or delete them?
20:42 < |beowulf|> ok
20:42 < |beowulf|> now it doesn't find /dev/md0 anymore
20:43 < |beowulf|> fsck.ext3: Unable to resolve 'UUID=e295b0a7-aa38-4c85-bfb5-7dcb5de2efba'
20:54 < rxr> hm
20:55 < rxr> can you undo the etc/initscript change ?
20:55 < rxr> 20:20 < rxr> --- etc/initscript.origI2008-04-12 18:19:44.000000000 +0000
20:55 < rxr> 20:20 < rxr> +++ etc/initscriptI2008-04-12 18:19:57.000000000 +0000
20:55 < rxr> 20:20 < rxr> -Imount -n /dev
20:55 < rxr> 20:20 < rxr> +Imount -n -t tmpfs none /dev
20:55 < rxr> ^- this one
20:55 < rxr> so just mount -n /dev
20:55 < rxr> and try again :-)
20:57 -!- Enqlave [n=stealth@sourcemage/guru/stealth] has joined #t2
20:59 < |beowulf|> it still doesn't find the root partition
21:00 < |beowulf|> sh-3.2# ls
21:00 < |beowulf|> console core fd mouse null pts shm stderr stdin stdout tty zero
21:00 < |beowulf|> that's all there is in dev at that time
21:00 < |beowulf|> and dev is a tmpfs
21:01 < |beowulf|> udevd is running
21:08 < rxr> anything intersting in a log ?
21:10 < |beowulf|> /var hasn't been mounted yet, so no log file
21:11 < |beowulf|> [ 55.837504] Warning: unable to open an initial console.
21:11 < |beowulf|> that's in dmesg
21:23 < rxr> yes, I have that Warning as well, that is exepcted and can not be avoid when booting without initrd for a dynamic /dev ...
21:23 [Users #t2]
21:23 [@ChanServ] [ CIA-53 ] [ Enqlave ] [ Ragnar|away] [ TobiX ]
21:23 [ __init__] [ Dallu1 ] [ LMJ ] [ rxr ] [ valentin ]
21:23 [ axionix ] [ dogmeat] [ mqueiros_] [ sparc-kly_ ] [ |beowulf|]
21:23 [ Baldzius] [ elias` ] [ mtr ] [ tfl0pz ]
21:23 -!- Irssi: #t2: Total of 19 nicks [1 ops, 0 halfops, 0 voices, 18 normal]
21:24 < rxr> Ragnar|away: btw. what you want regarding 32 vs. 64bit setups is on the T2 roadmap as more generic multilib support
21:25 < rxr> |beowulf|: oh - ic
21:25 < rxr> strange why the heck did that work for me ...
21:27 < rxr> |beowulf|: in etc/initscript also mount sys and proc where dev is mounted liks so:
21:27 < rxr> mount -n /proc
21:27 < rxr> mount -n /sys
21:27 < rxr> under the mount -n /dev line
21:35 < |beowulf|> http://www.hawo.stw.uni-erlangen.de/~siflfran/qube.10.log
21:36 < |beowulf|> networking didn't come up
21:37 < |beowulf|> root@vilya:t2-7.0 [0]> scripts/Emerge-Pkg lcdutils
21:37 < |beowulf|> misc/output/parse-config: line 47: /dev/fd/62: No such file or directory
21:37 < |beowulf|> still that error
22:00 < rxr> what's the content of your current /proc/mounts ?
22:01 < |beowulf|> root@vilya:linux-2.6.23-rc3 [130]> cat /proc/mounts
22:01 < |beowulf|> rootfs / rootfs rw 0 0
22:01 < |beowulf|> /dev/root / ext3 rw,noatime,data=ordered 0 0
22:01 < |beowulf|> dev /dev tmpfs rw 0 0
22:01 < |beowulf|> proc /proc proc rw 0 0
22:01 < |beowulf|> sys /sys sysfs rw 0 0
22:01 < |beowulf|> shm /dev/shm tmpfs rw,nosuid,nodev,noexec 0 0
22:01 < |beowulf|> devpts /dev/pts devpts rw 0 0
22:01 < |beowulf|> /dev/md2 /usr jfs rw,noatime 0 0
22:01 < |beowulf|> /dev/md3 /var jfs rw,noatime 0 0
22:01 < |beowulf|> /dev/md4 /tmp jfs rw,noatime 0 0
22:01 < |beowulf|> /dev/md5 /home jfs rw,noatime 0 0
22:01 < |beowulf|> dev /dev tmpfs rw 0 0
22:02 < rxr> remove the mount /dev after the start) in /sbin/init.d/system
22:02 < rxr> I think the issue might be that /dev is mounted twice, now
22:09 < rxr> if you have such a line there ...
22:25 -!- __init__ [i=stealth@unaffiliated/init/x-165853] has quit [Read error: 110 (Connection timed out)]
22:26 -!- Enqlave [n=stealth@sourcemage/guru/stealth] has quit [Read error: 113 (No route to host)]
22:34 < rxr> |beowulf|: I reproduced something simillar here
22:35 < rxr> you need to "fix" your fstab
22:36 < rxr> at least the tmpfs /dev entry needs to be
22:36 < rxr> none /dev tmpfs defaults 0 0
22:36 < rxr> with something else than none in the beginning dev is mounted a second time
22:36 < rxr> just happened here during testing
22:36 < rxr> I hope that helps
22:36 < rxr> btw. hi from an GCC_4.3.0 t2/trunk with X running ...
22:37 < rxr> # cat /etc/VERSION ; gcc --version
22:37 < rxr> T2 SDE 8.0-trunk (2008/03/17)
22:37 < rxr> gcc (GCC) 4.3.0
22:38 < rxr> Baldzius: what was your resolution for the DHCP issue on trunk`?
22:38 < rxr> dhclient does not find it's helper scripts
22:41 < CIA-53> rene * r28686 /trunk/package/base/ (sysfiles/etc_initscript.txt sysvinit/system.init):
22:41 < CIA-53> * mount /proc and /sys for initrd-less boots, at least /sys
22:41 < CIA-53> is required for proper udev operation / event synthese
22:41 < CIA-53> * do not setup the /proc/self -> /dev/fd symlink early
22:41 < CIA-53> in sysvinit/system.init, it is currently done by
22:41 < CIA-53> /etc/conf/devfs, anyway
22:46 -!- __init__ [i=stealth@unaffiliated/init/x-165853] has joined #t2
22:51 [Users #t2]
22:51 [@ChanServ] [ Baldzius] [ dogmeat] [ mqueiros_ ] [ rxr ] [ TobiX ]
22:51 [ __init__] [ CIA-53 ] [ elias` ] [ mtr ] [ sparc-kly_] [ valentin ]
22:51 [ axionix ] [ Dallu1 ] [ LMJ ] [ Ragnar|away] [ tfl0pz ] [ |beowulf|]
22:51 -!- Irssi: #t2: Total of 18 nicks [1 ops, 0 halfops, 0 voices, 17 normal]
22:53 < |beowulf|> rxr: that did it
22:55 < rxr> :-)
22:56 < |beowulf|> is it possible to select/deselect optional dependencies for packages?
22:56 < rxr> during Emerge, not yet automatically, but manaully with Emerge-Pkg -deps=none I know what I want
22:57 < |beowulf|> vim for example has a gui that I don't want to build -- can I disable it that way?
22:57 < rxr> ok - vim is special in that it wires Config otions so you can choose some stuff there
22:58 < |beowulf|> how?
22:58 < rxr> scripts/Config
22:58 < rxr> expert options
22:58 < rxr> vim
22:58 < rxr> and there are some options
22:58 < rxr> like "two binaries, one for CLI and for GUI"
22:58 < rxr> it even has:
22:58 < rxr> choice SDECFG_PKG_VIM_GUI auto \
22:58 < rxr> no GUI_style:_No_GUI auto GUI_style:_Autoselect \
22:58 < rxr> gtk GUI_style:_GTK gtk2 GUI_style:_GTK-2 \
22:58 < rxr> gnome GUI_style:_Gnome gnome2 GUI_style:_Gnome_2 \
22:58 < rxr> motif GUI_style:_Motif_/_Lesstif
22:58 < rxr> interesting ...
23:00 < |beowulf|> 'Show expert and experimental options' is turned on but I don't see anything concerning vim
23:00 < rxr> it should be further down below ?
23:01 < rxr> (though only if vim is enabled in your current package selection)
23:03 < |beowulf|> 'Minimalistic package selection' is selected
23:03 < |beowulf|> do I have to create my own package selection?
23:05 < rxr> no, you can also just emerge stuff
23:05 < rxr> or add singles packages with this custom package selection rules
23:05 < |beowulf|> I emerged vim
23:06 < rxr> but honestly, I just emerge single packages, though for packages not perfectly "written" it can sometimes requires the package's .conf code to get fixed
23:06 < rxr> did it already fully built ?
23:06 < |beowulf|> yes
23:06 < rxr> :-)
23:07 < |beowulf|> scripts/Config is terribly slow
23:07 < |beowulf|> can I do anything to speed it up?
23:08 < rxr> yes, help rewriting it
23:08 < |beowulf|> :)
23:08 < rxr> the slowness is know, annoys us all and a rewrite is on our roadmap
23:08 < rxr> http://www.t2-project.org/releases/
23:08 < rxr> revisited configuration system, that is magnitudes (at least) faster
23:09 < rxr> not rocked sience, just avoid re-reading everything at every selected option and probably not in shell but Lua and you're done
23:09 < rxr> patches welcome :-)
23:09 < rxr> as most core people appear overly loaded with daily work these days ...
23:10 < rxr> but in any case I think that will be done this year
23:10 < rxr> at least I'll sooner than later if no patch arrives :-)
23:10 < rxr> oh , and if you have some fre seconds a hardware page for your cute Cube would be nice, too:
23:11 < rxr> http://www.t2-project.org/hardware/server/
23:11 -!- teatime07 [n=jan@c-69-138-57-167.hsd1.tn.comcast.net] has joined #t2
23:11 [Users #t2]
23:11 [@ChanServ] [ CIA-53 ] [ LMJ ] [ rxr ] [ TobiX ]
23:11 [ __init__] [ Dallu1 ] [ mqueiros_ ] [ sparc-kly_] [ valentin ]
23:11 [ axionix ] [ dogmeat] [ mtr ] [ teatime07 ] [ |beowulf|]
23:11 [ Baldzius] [ elias` ] [ Ragnar|away] [ tfl0pz ]
23:11 -!- Irssi: #t2: Total of 19 nicks [1 ops, 0 halfops, 0 voices, 18 normal]
--- Log closed Sun Apr 13 00:00:58 2008
01:22 -!- Baldzius [n=Baldzius@87.198.192.226] has quit ["Leaving"]
02:36 -!- Dallu1 [n=earl@fgfw.basis.is] has joined #t2
02:38 -!- Dallur [n=earl@fgfw.basis.is] has quit [Read error: 110 (Connection timed out)]
03:18 -!- xaw [n=xaw@unaffiliated/xaw] has quit ["leaving"]
04:20 -!- Ragnar|away [i=loki@D8FA4DE6.ptr.dia.nextlink.net] has quit [Remote closed the connection]
04:20 -!- Ragnar|away [i=loki@D8FA4DE6.ptr.dia.nextlink.net] has joined #t2
08:30 -!- Enqlave [n=stealth@sourcemage/guru/stealth] has quit [Read error: 104 (Connection reset by peer)]
08:50 -!- elias` [n=me@unaffiliated/elias/x-342423] has quit [Read error: 145 (Connection timed out)]
08:59 -!- elias` [n=me@unaffiliated/elias/x-342423] has joined #t2
10:17 -!- Netsplit kornbluth.freenode.net <-> irc.freenode.net quits: LMJ
10:18 -!- Netsplit over, joins: LMJ
10:30 < rxr> moin
10:38 < CIA-53> rene * r28681 /trunk/package/editors/ne/ne.desc: * updated ne (1.42 -> 1.43)
10:40 < CIA-53> rene * r28682 /trunk/package/x86/syslinux/syslinux.desc: * updated syslinux (3.62 -> 3.63)
10:43 -!- Netsplit kornbluth.freenode.net <-> irc.freenode.net quits: LMJ
10:43 -!- Netsplit over, joins: LMJ
11:37 -!- Netsplit kornbluth.freenode.net <-> irc.freenode.net quits: LMJ
11:38 -!- Netsplit over, joins: LMJ
11:53 < rxr> re
12:03 -!- Enqlave [n=stealth@sourcemage/guru/stealth] has joined #t2
12:15 -!- Enqlave [n=stealth@sourcemage/guru/stealth] has quit ["Dont wait to die to find paradise..."]
12:51 -!- Enqlave [n=stealth@sourcemage/guru/stealth] has joined #t2
12:58 -!- Enqlave [n=stealth@sourcemage/guru/stealth] has quit ["Dont wait to die to find paradise..."]
13:01 -!- Enqlave [n=stealth@sourcemage/guru/stealth] has joined #t2
14:23 < |beowulf|> rxr: Apr 11 23:51:11 vilya agetty[939]: /dev/ttyS0: No such file or directory <-- that's the last line in syslog on the cube before I turned it off because it didn't start agetty on the serial port
14:28 < rxr> hi
14:28 < rxr> |beowulf|: I just clicked on "send" regarding the article
14:28 < |beowulf|> hi :)
14:28 < rxr> so now turning to qemu and the initrd-less boot issue ...
15:47 -!- Enqlave [n=stealth@sourcemage/guru/stealth] has quit ["Dont wait to die to find paradise..."]
18:11 < CIA-53> rene * r28683 /trunk/package/xorg/xf86-video-xgi/ (xf86-video-xgi.cache xf86-video-xgi.desc): * updated xf86-video-xgi (2007-01-02 -> 1.5.0)
18:40 -!- dsoul [i=darksoul@vice.ii.uj.edu.pl] has quit [SendQ exceeded]
18:48 [Users #t2]
18:48 [@ChanServ] [ Dallu1 ] [ mqueiros_ ] [ sparc-kly_] [ |beowulf|]
18:48 [ __init__] [ dogmeat] [ mtr ] [ tfl0pz ]
18:48 [ axionix ] [ elias` ] [ Ragnar|away] [ TobiX ]
18:48 [ CIA-53 ] [ LMJ ] [ rxr ] [ valentin ]
18:48 -!- Irssi: #t2: Total of 17 nicks [1 ops, 0 halfops, 0 voices, 16 normal]
19:00 -!- Baldzius [n=Baldzius@89.127.178.76] has joined #t2
19:00 < Baldzius> moin
19:06 < rxr> hey Baldzius
19:07 < Baldzius> hi rxr
19:07 < Baldzius> trying libtool 2.2.2 locally
19:15 < rxr> so
19:16 < rxr> |beowulf|: my VM now boots without initrd ...
19:18 < rxr> http://people.exactcode.de/~rene/inittab -> etc/inittab
19:18 < rxr> http://people.exactcode.de/~rene/system -> sbin/init.d/system
19:21 < rxr> err sorry
19:21 < rxr> not inittab
19:22 < rxr> initscript
19:25 < |beowulf|> rxr: where should I put initscript?
19:25 < rxr> etc/initscript
19:28 < CIA-53> rene * r28684 /trunk/package/base/ (3 files in 2 dirs): * re-implemented support for booting without initrd
19:29 < |beowulf|> ah, very nice
19:29 < |beowulf|> I have a getty on the serial port
19:29 < |beowulf|> thank you! :)
19:29 < rxr> you're welcome !
19:30 < |beowulf|> there are still a few messages about missing /dev/console during boot though
19:31 < rxr> many ? or just one ?
19:32 < |beowulf|> http://www.hawo.stw.uni-erlangen.de/~siflfran/qube.8.log
19:33 < |beowulf|> and networking didn't come up
19:34 < |beowulf|> if I call dhclient manually I get an IP address
19:34 < rxr> have you updated the sbin/init.d/system ?
19:35 < |beowulf|> yes
19:36 < rxr> what does this output on that cube:
19:36 < rxr> grep "/dev " /proc/mounts
19:36 < |beowulf|> dev /dev tmpfs rw 0 0
19:36 < rxr> hm
19:37 < rxr> your log has these:
19:37 < rxr> ln: cannot remove `mouse': Read-only file system
19:37 < rxr> rm: cannot remove `fd': Read-only file system
19:37 < rxr> ln: cannot remove `core': Read-only file system
19:37 < rxr> ln: creating symbolic link `fd/fd': No such file or directory
19:37 < rxr> ln: cannot remove `stdin': Read-only file system
19:37 < rxr> ln: cannot remove `stdout': Read-only file system
19:37 < rxr> ln: cannot remove `stderr': Read-only file system
19:37 -!- teatime07 [n=jan@c-69-138-57-167.hsd1.tn.comcast.net] has joined #t2
19:37 < rxr> I wonder where it comes from
19:37 < |beowulf|> screen complains that it cannot find a pty
19:38 < rxr> there must still be something wrong or imperfect regarding the /dev setup ... :-(
19:38 < rxr> maybe remounted ?
19:42 < |beowulf|> which package do I have to install to get the host command?
19:42 < |beowulf|> ./scripts/Download: line 856: /dev/fd/62: No such file or directory
19:44 < rxr> that is the same issue as on boot
19:45 < rxr> something is not yet right with the /proc /dev and the /proc/self -> /dev/fd symlink
19:48 -!- teatime07 [n=jan@c-69-138-57-167.hsd1.tn.comcast.net] has quit ["weeeeee..."]
19:53 < rxr> |beowulf|: what does this output grep "/dev " /etc/fstab ?
19:54 < |beowulf|> dev /dev tmpfs defaults 0 0
19:55 < |beowulf|> http://www.hawo.stw.uni-erlangen.de/~siflfran/fstab
19:55 < rxr> can you add an "cat /proc/mounts"
19:55 < rxr> to /sbin/init.d/system
19:55 < rxr> before the line
19:55 < rxr> "Configuring the /dev filesystem ..."
19:58 < |beowulf|> http://www.hawo.stw.uni-erlangen.de/~siflfran/qube.9.log
20:03 < rxr> hm - there is no /dev mounted
20:18 < CIA-53> rene * r28685 /trunk/package/base/ (sysfiles/etc_initscript.txt sysvinit/system.init):
20:18 < CIA-53> * kill udevd after populating the initial /dev when booting
20:18 < CIA-53> without initrd as done inside the initrd/init, likewise
20:18 < CIA-53> * removed now phased out udevd killing in the
20:18 < CIA-53> sysvinit/system.init, the initrd does not yield a running
20:18 < CIA-53> udevd for some months, now
20:20 < rxr> |beowulf|: a workaround could be:
20:20 < rxr> --- etc/initscript.orig 2008-04-12 18:19:44.000000000 +0000
20:20 < rxr> +++ etc/initscript 2008-04-12 18:19:57.000000000 +0000
20:20 < rxr> - mount -n /dev
20:20 < rxr> + mount -n -t tmpfs none /dev
20:21 < Ragnar|away> shouldn't that be default anyway?
20:21 < rxr> Ragnar|away: depends ...
20:21 < rxr> yes, we could hardcode it at that place
20:21 < rxr> |beowulf|: wait with testing
20:21 < rxr> failed in my vm
20:21 < rxr> just a moment
20:21 < |beowulf|> too late...
20:22 < rxr> Ragnar|away: our initrd is setting it up that way anyway, but |beowulf| is booting without initrd and thus the initrd-less patchery right now
20:22 < Ragnar|away> yea
20:22 < Ragnar|away> well for some hosts I'd also like to be able to boot without initrd
20:23 < rxr> with the trunk patches it should be possible to boot with either a static (on-disc) device node setup or udev, just as needed
20:23 < rxr> module |beowulf|'s box's hickup
20:23 < rxr> modulo ...
20:23 < Ragnar|away> :)
20:24 < rxr> |beowulf|: I guess that change did not help on your box either ?
20:25 < |beowulf|> no, didn't change anything
20:25 < Ragnar|away> oh btw ... what I'd really like would be to be able to emerge a 32 bit package on a 64 bit system and a 64 bit package on a system that also has some 32 bit libs present without library-messyness ... can you think of a way to do that with not too big effort?
20:27 < rxr> for single packages that could work easier
20:27 < rxr> but for big packages with lots of deps that is trickier
20:27 < rxr> e.g. emerging the 32bit wine on a otherwise fully 64bit system ...
20:28 < rxr> what kind of packages would you want that for ?
20:28 < rxr> (wine ?)
20:28 < rxr> ah
20:28 < Ragnar|away> wine, firefox (flash without 32-64bit wrapper)
20:28 < Ragnar|away> firstclass client
20:28 < rxr> |beowulf|: can it be that you have some devices nodes from earlier tinkering in /dev?
20:29 < |beowulf|> yes, there are some
20:29 < rxr> |beowulf|: rm the on-disk nodes I remeber you put there (dev/console, dev/null etc?)
20:29 < |beowulf|> should I remove them?
20:29 < rxr> yeah - for that case the new code does not mount another /dev then
20:29 < rxr> as it expects you have a full device node collection in that case, like the good old pure static setup
20:29 < |beowulf|> should I delete everything in /dev? or just /dev/{console,null,zero}?
20:30 < rxr> yes, would be better if you like to use a new dynamic setup
20:30 < rxr> if you like to have static nodes, you would have to add the other missing ones :-)
20:30 < rxr> all or nothing :-)
20:32 < Ragnar|away> makes sense =)
20:36 < |beowulf|> rxr: the system I built had the following in /dev:
20:36 < |beowulf|> fd@ loop/ null random urandom zero
20:36 < |beowulf|> should I leave these or delete them?
20:42 < |beowulf|> ok
20:42 < |beowulf|> now it doesn't find /dev/md0 anymore
20:43 < |beowulf|> fsck.ext3: Unable to resolve 'UUID=e295b0a7-aa38-4c85-bfb5-7dcb5de2efba'
20:54 < rxr> hm
20:55 < rxr> can you undo the etc/initscript change ?
20:55 < rxr> 20:20 < rxr> --- etc/initscript.origI2008-04-12 18:19:44.000000000 +0000
20:55 < rxr> 20:20 < rxr> +++ etc/initscriptI2008-04-12 18:19:57.000000000 +0000
20:55 < rxr> 20:20 < rxr> -Imount -n /dev
20:55 < rxr> 20:20 < rxr> +Imount -n -t tmpfs none /dev
20:55 < rxr> ^- this one
20:55 < rxr> so just mount -n /dev
20:55 < rxr> and try again :-)
20:57 -!- Enqlave [n=stealth@sourcemage/guru/stealth] has joined #t2
20:59 < |beowulf|> it still doesn't find the root partition
21:00 < |beowulf|> sh-3.2# ls
21:00 < |beowulf|> console core fd mouse null pts shm stderr stdin stdout tty zero
21:00 < |beowulf|> that's all there is in dev at that time
21:00 < |beowulf|> and dev is a tmpfs
21:01 < |beowulf|> udevd is running
21:08 < rxr> anything intersting in a log ?
21:10 < |beowulf|> /var hasn't been mounted yet, so no log file
21:11 < |beowulf|> [ 55.837504] Warning: unable to open an initial console.
21:11 < |beowulf|> that's in dmesg
21:23 < rxr> yes, I have that Warning as well, that is exepcted and can not be avoid when booting without initrd for a dynamic /dev ...
21:23 [Users #t2]
21:23 [@ChanServ] [ CIA-53 ] [ Enqlave ] [ Ragnar|away] [ TobiX ]
21:23 [ __init__] [ Dallu1 ] [ LMJ ] [ rxr ] [ valentin ]
21:23 [ axionix ] [ dogmeat] [ mqueiros_] [ sparc-kly_ ] [ |beowulf|]
21:23 [ Baldzius] [ elias` ] [ mtr ] [ tfl0pz ]
21:23 -!- Irssi: #t2: Total of 19 nicks [1 ops, 0 halfops, 0 voices, 18 normal]
21:24 < rxr> Ragnar|away: btw. what you want regarding 32 vs. 64bit setups is on the T2 roadmap as more generic multilib support
21:25 < rxr> |beowulf|: oh - ic
21:25 < rxr> strange why the heck did that work for me ...
21:27 < rxr> |beowulf|: in etc/initscript also mount sys and proc where dev is mounted liks so:
21:27 < rxr> mount -n /proc
21:27 < rxr> mount -n /sys
21:27 < rxr> under the mount -n /dev line
21:35 < |beowulf|> http://www.hawo.stw.uni-erlangen.de/~siflfran/qube.10.log
21:36 < |beowulf|> networking didn't come up
21:37 < |beowulf|> root@vilya:t2-7.0 [0]> scripts/Emerge-Pkg lcdutils
21:37 < |beowulf|> misc/output/parse-config: line 47: /dev/fd/62: No such file or directory
21:37 < |beowulf|> still that error
22:00 < rxr> what's the content of your current /proc/mounts ?
22:01 < |beowulf|> root@vilya:linux-2.6.23-rc3 [130]> cat /proc/mounts
22:01 < |beowulf|> rootfs / rootfs rw 0 0
22:01 < |beowulf|> /dev/root / ext3 rw,noatime,data=ordered 0 0
22:01 < |beowulf|> dev /dev tmpfs rw 0 0
22:01 < |beowulf|> proc /proc proc rw 0 0
22:01 < |beowulf|> sys /sys sysfs rw 0 0
22:01 < |beowulf|> shm /dev/shm tmpfs rw,nosuid,nodev,noexec 0 0
22:01 < |beowulf|> devpts /dev/pts devpts rw 0 0
22:01 < |beowulf|> /dev/md2 /usr jfs rw,noatime 0 0
22:01 < |beowulf|> /dev/md3 /var jfs rw,noatime 0 0
22:01 < |beowulf|> /dev/md4 /tmp jfs rw,noatime 0 0
22:01 < |beowulf|> /dev/md5 /home jfs rw,noatime 0 0
22:01 < |beowulf|> dev /dev tmpfs rw 0 0
22:02 < rxr> remove the mount /dev after the start) in /sbin/init.d/system
22:02 < rxr> I think the issue might be that /dev is mounted twice, now
22:09 < rxr> if you have such a line there ...
22:25 -!- __init__ [i=stealth@unaffiliated/init/x-165853] has quit [Read error: 110 (Connection timed out)]
22:26 -!- Enqlave [n=stealth@sourcemage/guru/stealth] has quit [Read error: 113 (No route to host)]
22:34 < rxr> |beowulf|: I reproduced something simillar here
22:35 < rxr> you need to "fix" your fstab
22:36 < rxr> at least the tmpfs /dev entry needs to be
22:36 < rxr> none /dev tmpfs defaults 0 0
22:36 < rxr> with something else than none in the beginning dev is mounted a second time
22:36 < rxr> just happened here during testing
22:36 < rxr> I hope that helps
22:36 < rxr> btw. hi from an GCC_4.3.0 t2/trunk with X running ...
22:37 < rxr> # cat /etc/VERSION ; gcc --version
22:37 < rxr> T2 SDE 8.0-trunk (2008/03/17)
22:37 < rxr> gcc (GCC) 4.3.0
22:38 < rxr> Baldzius: what was your resolution for the DHCP issue on trunk`?
22:38 < rxr> dhclient does not find it's helper scripts
22:41 < CIA-53> rene * r28686 /trunk/package/base/ (sysfiles/etc_initscript.txt sysvinit/system.init):
22:41 < CIA-53> * mount /proc and /sys for initrd-less boots, at least /sys
22:41 < CIA-53> is required for proper udev operation / event synthese
22:41 < CIA-53> * do not setup the /proc/self -> /dev/fd symlink early
22:41 < CIA-53> in sysvinit/system.init, it is currently done by
22:41 < CIA-53> /etc/conf/devfs, anyway
22:46 -!- __init__ [i=stealth@unaffiliated/init/x-165853] has joined #t2
22:51 [Users #t2]
22:51 [@ChanServ] [ Baldzius] [ dogmeat] [ mqueiros_ ] [ rxr ] [ TobiX ]
22:51 [ __init__] [ CIA-53 ] [ elias` ] [ mtr ] [ sparc-kly_] [ valentin ]
22:51 [ axionix ] [ Dallu1 ] [ LMJ ] [ Ragnar|away] [ tfl0pz ] [ |beowulf|]
22:51 -!- Irssi: #t2: Total of 18 nicks [1 ops, 0 halfops, 0 voices, 17 normal]
22:53 < |beowulf|> rxr: that did it
22:55 < rxr> :-)
22:56 < |beowulf|> is it possible to select/deselect optional dependencies for packages?
22:56 < rxr> during Emerge, not yet automatically, but manaully with Emerge-Pkg -deps=none I know what I want
22:57 < |beowulf|> vim for example has a gui that I don't want to build -- can I disable it that way?
22:57 < rxr> ok - vim is special in that it wires Config otions so you can choose some stuff there
22:58 < |beowulf|> how?
22:58 < rxr> scripts/Config
22:58 < rxr> expert options
22:58 < rxr> vim
22:58 < rxr> and there are some options
22:58 < rxr> like "two binaries, one for CLI and for GUI"
22:58 < rxr> it even has:
22:58 < rxr> choice SDECFG_PKG_VIM_GUI auto \
22:58 < rxr> no GUI_style:_No_GUI auto GUI_style:_Autoselect \
22:58 < rxr> gtk GUI_style:_GTK gtk2 GUI_style:_GTK-2 \
22:58 < rxr> gnome GUI_style:_Gnome gnome2 GUI_style:_Gnome_2 \
22:58 < rxr> motif GUI_style:_Motif_/_Lesstif
22:58 < rxr> interesting ...
23:00 < |beowulf|> 'Show expert and experimental options' is turned on but I don't see anything concerning vim
23:00 < rxr> it should be further down below ?
23:01 < rxr> (though only if vim is enabled in your current package selection)
23:03 < |beowulf|> 'Minimalistic package selection' is selected
23:03 < |beowulf|> do I have to create my own package selection?
23:05 < rxr> no, you can also just emerge stuff
23:05 < rxr> or add singles packages with this custom package selection rules
23:05 < |beowulf|> I emerged vim
23:06 < rxr> but honestly, I just emerge single packages, though for packages not perfectly "written" it can sometimes requires the package's .conf code to get fixed
23:06 < rxr> did it already fully built ?
23:06 < |beowulf|> yes
23:06 < rxr> :-)
23:07 < |beowulf|> scripts/Config is terribly slow
23:07 < |beowulf|> can I do anything to speed it up?
23:08 < rxr> yes, help rewriting it
23:08 < |beowulf|> :)
23:08 < rxr> the slowness is know, annoys us all and a rewrite is on our roadmap
23:08 < rxr> http://www.t2-project.org/releases/
23:08 < rxr> revisited configuration system, that is magnitudes (at least) faster
23:09 < rxr> not rocked sience, just avoid re-reading everything at every selected option and probably not in shell but Lua and you're done
23:09 < rxr> patches welcome :-)
23:09 < rxr> as most core people appear overly loaded with daily work these days ...
23:10 < rxr> but in any case I think that will be done this year
23:10 < rxr> at least I'll sooner than later if no patch arrives :-)
23:10 < rxr> oh , and if you have some fre seconds a hardware page for your cute Cube would be nice, too:
23:11 < rxr> http://www.t2-project.org/hardware/server/
23:11 -!- teatime07 [n=jan@c-69-138-57-167.hsd1.tn.comcast.net] has joined #t2
23:11 [Users #t2]
23:11 [@ChanServ] [ CIA-53 ] [ LMJ ] [ rxr ] [ TobiX ]
23:11 [ __init__] [ Dallu1 ] [ mqueiros_ ] [ sparc-kly_] [ valentin ]
23:11 [ axionix ] [ dogmeat] [ mtr ] [ teatime07 ] [ |beowulf|]
23:11 [ Baldzius] [ elias` ] [ Ragnar|away] [ tfl0pz ]
23:11 -!- Irssi: #t2: Total of 19 nicks [1 ops, 0 halfops, 0 voices, 18 normal]
--- Log closed Sun Apr 13 00:00:58 2008