--- Log opened Sat Nov 18 00:00:38 2006
00:03 -!- mtr [n=Michael@pD9E12692.dip0.t-ipconnect.de] has quit [Read error: 110 (Connection timed out)]
00:03 -!- mtr_ is now known as mtr
01:07 -!- bl4ckth0rn [i=lus2la@20151010141.user.veloxzone.com.br] has joined #t2
02:31 -!- emte [n=emte@d64-180-45-14.bchsia.telus.net] has quit [Connection timed out]
02:37 -!- emte [n=emte@d64-180-45-14.bchsia.telus.net] has joined #t2
03:52 -!- idealm [n=ideal@222.67.239.123] has joined #t2
04:04 -!- shevegen [n=adh4@chello080108103172.36.11.univie.teleweb.at] has joined #t2
04:05 < shevegen> for bootup, what does T2 use? sysvinit? runit? initng? something else? :-)
04:15 < emte> some use sysvinit
04:15 < emte> others dont
04:17 < _Ragnar_> you can select that
04:19 < shevegen> nice
06:55 -!- bl4ckth0rn [i=lus2la@20151010141.user.veloxzone.com.br] has quit []
07:32 -!- Netsplit leguin.freenode.net <-> irc.freenode.net quits: emte
07:35 -!- Netsplit over, joins: emte
09:06 < sepp> moin :)
09:07 < Stelz> moin sepp
09:07 < sepp> hi Stelz
10:06 -!- Netsplit leguin.freenode.net <-> irc.freenode.net quits: emte
10:06 -!- Netsplit over, joins: emte
10:17 -!- Netsplit leguin.freenode.net <-> irc.freenode.net quits: pamdirac, shevegen, @ChanServ, LMJ, emte
10:18 -!- Netsplit over, joins: @ChanServ, emte, shevegen, pamdirac, LMJ
10:30 < rxr> re
10:39 < capey> morning :)
10:44 < rxr> moin
10:44 [Users #t2]
10:44 [@ChanServ] [ CIA-9 ] [ LMJ ] [ rxr ] [ Stelz ]
10:44 [ _Ragnar_] [ emte ] [ mtr ] [ sepp ] [ valentin]
10:44 [ capey ] [ idealm] [ pamdirac] [ shevegen] [ zod ]
10:44 -!- Irssi: #t2: Total of 15 nicks [1 ops, 0 halfops, 0 voices, 14 normal]
10:44 < sepp> hello
10:44 < rxr> hi sepp
10:44 < rxr> LWE was quite nice, especially meting mtr :-)
10:44 < sepp> hi rxr, capey
10:45 < rxr> emte: do you start to feel comfortable with t2 and it is worth handing you write acess to SVN or do you hesitate to gain write access ?
11:37 < emte> i cant say that i am anywhere near comfortable with T2 yet, there is far too much that i dont know
11:39 < emte> ouch 1k to buy a dlue-ray drive
11:39 < emte> blue*
11:50 < CIA-9> rene * r20847 /trunk/package/database/metakit/metakit.desc: * updated metakit (2.4.9.5 -> 2.4.9.6), fixed [COPY] tag and cleaned spacing
11:51 < CIA-9> rene * r20848 /trunk/package/filesystem/cdrkit/cdrkit.desc: * updated cdrkit (1.0pre4 -> 1.0pre5)
11:52 < CIA-9> rene * r20849 /trunk/package/filesystem/parted/parted.desc: * updated parted (1.7.1 -> 1.8.0), cleaned spacing
11:53 < CIA-9> rene * r20850 /trunk/package/x11/ddcxinfo/ddcxinfo.desc: * updated ddcxinfo (0.6.18 -> 0.6.19)
11:54 < CIA-9> rene * r20851 /trunk/architecture/ (6 files in 2 dirs): * added work in progress avr32 and blackfin architecture defines
11:56 < CIA-9> rene * r20852 /trunk/package/base/units/units.desc: * updated units (1.85 -> 1.86), cleaned spacing
11:57 < CIA-9> rene * r20853 /trunk/package/develop/clearsilver/clearsilver.desc: * updated clearsilver (0.10.3 -> 0.10.4)
11:57 < CIA-9> rene * r20854 /trunk/package/develop/trac/trac.desc: * updated trac (0.10.1 -> 0.10.2)
12:06 < CIA-9> rene * r20855 /trunk/package/network/nginx/nginx.desc: * updated nginx (0.4.11 -> 0.4.13)
12:11 < CIA-9> rene * r20856 /trunk/package/python/itools/itools.desc: * updated itools (0.14.4 -> 0.14.5)
12:11 < CIA-9> rene * r20857 /trunk/package/develop/git/git.desc: * updated git (1.4.3.4 -> 1.4.4)
12:23 < CIA-9> rene * r20858 /trunk/package/develop/monotone/monotone.desc: * updated monotone (0.30 -> 0.31)
12:23 < CIA-9> rene * r20859 /trunk/package/multimedia/linphone/linphone.desc: * updated linphone (1.5.0 -> 1.5.1)
12:24 -!- kensai [n=kensai@206.248.86.8] has joined #t2
12:27 < CIA-9> rene * r20860 /trunk/package/security/gnutls/gnutls.desc: * updated gnutls (1.4.5 -> 1.6.0), and fixed spacing
12:28 < CIA-9> rene * r20861 /trunk/package/editors/tea/tea.desc: * updated tea (14.3.0 -> 14.3.1)
12:34 < CIA-9> rene * r20862 /trunk/package/kde/ksmoothdock/ (. ksmoothdock.cache ksmoothdock.desc): * added ksmoothdock (4.3)
12:40 -!- psycodad [n=obiwan@2001:4060:4419:b1:0:0:0:2] has joined #t2
12:40 < psycodad> good morning
12:41 < Stelz> moin psycodad
12:41 < Stelz> hehe.. T2 psy-lover? :)
12:41 < psycodad> i try to install t2 from cdrom and everything goes fine until reboot and then i get "Error 15: File not found" from grub...what may i be doin wrong ?
12:41 < psycodad> I have a 256MB /boot a 768 swap and ~20GB / partition
12:42 < psycodad> Stelz: no bundy fan :-)
12:42 < Stelz> :P
12:42 < psycodad> psycodad is al's tv hero
12:42 < sepp> maybe you forgot the grub setup in "stone"
12:43 < psycodad> no, i went thru it numerous times and it looks fine so far. I did the install twice already and it failed both times with this...
12:43 < sepp> if grub starts you can use the 'grub shell' to correct the wrong filename
12:44 < psycodad> i used 'e' to edit and now i see:
12:44 < psycodad> kernel (hd0,0)/boot/vmlinuz_2.6.17.14-dist root/dev/sda3 ro
12:45 < psycodad> sda3 is / while sda1 is /boot so looks fine to me
12:45 < sepp> no initrd line?
12:45 < psycodad> initrs (hd0,0)/boot/initrd-2.6.17.14-dist.img
12:46 < psycodad> s/initrs/initrd/
12:46 < psycodad> sorry i'm typing it from a vmw console
12:46 < sepp> np
12:47 < sepp> if (hd0,0) is right, that should work
12:47 < psycodad> i have only one disc and one cdrom in that (virtual) box
12:48 < psycodad> lets try again, maybe i get some hhs from you guys regarding the grub part of the setup
12:48 < sepp> hmmm
12:48 * psycodad rebooting box with cd in it
12:49 * psycodad entered stone
12:50 * psycodad selecting "sg latin1"
12:51 < emte> i though grub needed that stupid long form for partition access
12:51 * psycodad doing 256MB /boot as sda1, 768MB swap as sda2 and 19GB / as sda3
12:51 < sepp> the stupid long from?
12:52 < emte> last time i tried to use grub it didnt take "/dev/sda1" etc
12:52 < emte> yeah
12:52 < sepp> what is the stupid long from?
12:52 < rxr> psycodad: what t2 flavour did you install? a source build or a pre-built binary ?
12:53 < emte> something like: /bus/scsi/device/chain/blah/blah/blah/blah/2
12:53 < psycodad> rxr: from t2-minimal-6.0.1-x86-pentium-mmx_cd1.iso
12:53 < sepp> ahhh devfs
12:53 * psycodad starts gasgui
12:54 < rxr> psycodad: as far as I can see in a quick backscroll it looks like our stone grub setup does not cope with a seperate /boot partition
12:55 < rxr> your menu.lst should look like
12:55 < psycodad> rxr: oh...then how would i do best for an experimental setup, swap and / only ?
12:55 < rxr> I guess most people do not use a seperate /boot these days
12:55 < emte> lilo is my friend :) expecially v.22+ cuz i like framebuffer images
12:56 < rxr> this is a bug in our stone grub module that should be fixed ASAP
12:56 < rxr> psycodad: for a quick fix you can either edit the menu.lst or reinstall t2 without a /boot partition
12:56 < rxr> in the meantime I enter that bug in our bug manager and schedule to fix it after lunch
12:56 < psycodad> rxr: can it cope with a 20GB / ?
12:56 < rxr> jup
12:56 < emte> if your bios can, why wouldnt linux?
12:57 < psycodad> rxr: thx for the help! im off tryin that
12:57 < emte> btw you could skip the reinstall
12:57 < psycodad> emte: how?
12:57 < emte> just change the /boot name until you fix grub
12:57 < rxr> ouhm
12:57 < rxr> something is up with our uni box again ...
12:58 < emte> when you fix grub, change the name back and move your boot info into it
12:58 < psycodad> rxr: both mirrors seem to be unaivalable since this morning
12:58 < rxr> I hope not more than the usual network cable plugout problem
12:58 < emte> btw you dont need a /boot over 25-50M
12:58 < rxr> psycodad: I just noticed
12:58 < psycodad> emte: reinstalling is a matter of 5mins, its just experimental on vmware (a dual opteron i have for tests ;-))
12:59 < emte> yeah then reinstall is just as easy i suppose
12:59 < rxr> the mirrors and the bug tracker are on a univeristy machine - we had sporadig loose network cable problems in the mast - twice a year or so - I hope it is just that ...
12:59 < emte> seeing as its 4am i think its time for a nap
12:59 < emte> night
13:00 < psycodad> heh, noon here ;-)
13:00 < psycodad> but time for a nap too ;-)
13:02 < CIA-9> rene * r20863 /trunk/package/base/numactl/numactl.desc: * updated numactl (0.9.10 -> 0.9.11), and cleaned spacing
13:03 < rxr> time for lunch - cu
13:12 < psycodad> rxr: hah, now it boots fine, thx again
13:31 < psycodad> btw: some part of the build process should check the existence of gawk since everything just fails silently if its absent
13:56 < capey> i got the same "error" as psycodad on grub. it seems that grub generator fails for somereason when having diffrent boot and root partition
13:56 < sepp> :(
13:56 < capey> either have boot -> . link inside /boot or grub generator should remove /boot/ from directory def when having own boot partition
13:59 < capey> hmm,avr32 arch...
13:59 < capey> where did you get that rxr?
14:00 < capey> on grub, i suggest that we generate boot -> . symlink inside /boot, thats what most distros do
14:17 -!- misl [n=chatzill@84-104-172-187.cable.quicknet.nl] has joined #t2
14:17 < misl> hi folks :-)
14:18 < Stelz> rehi misl
14:18 < misl> hi Stelz
14:35 < rxr> re
14:35 < rxr> moin misl
14:36 < rxr> capey: yes, the grub generator should cope with it directly - no symlink tinkery needed
14:36 < rxr> in fact other boot modules in stone do this, I think yaboot and lilo properly handle /boot - maybe even the silo one ...
14:36 < misl> hi rxr
14:43 -!- tadao [n=tadao@254-199.adsl.pool.ew.hu] has joined #t2
14:49 < tadao> hi t2 , i'm stuck with postinstall of T2(setupcd from t2.org) on my build-system machine(duron800 -viachips) : on first boot grub loads kernel-kernel loads then freezes on some Splash or t2 logo - any hints?
14:50 < CIA-9> rene * r20864 /trunk/package/multimedia/gnash/ (gnash.conf gnash.desc): * updated gnash (20060312 -> 0.7.2)
14:51 [Users #t2]
14:51 [@ChanServ] [ emte ] [ misl ] [ rxr ] [ tadao ]
14:51 [ _Ragnar_] [ idealm] [ mtr ] [ sepp ] [ valentin]
14:51 [ capey ] [ kensai] [ pamdirac] [ shevegen] [ zod ]
14:51 [ CIA-9 ] [ LMJ ] [ psycodad] [ Stelz ]
14:51 -!- Irssi: #t2: Total of 19 nicks [1 ops, 0 halfops, 0 voices, 18 normal]
14:51 < rxr> tadao: is this inside grub? or is the kernel and initrd loaded already?
14:52 < tadao> hi rxr , no kernel loads and initrd too
14:52 < capey> what hook is called after make install but before flist gen... postmake?
14:53 < rxr> capey: yes
14:53 < rxr> tadao: ok - so the kernel starts to boot ?
14:53 < tadao> yes -during install i chnged default runlevel to from2 i guess - could that be it ?
14:54 < tadao> * to 3
14:54 < rxr> 3 should be fine
14:54 < rxr> what is the last message you have on the screen when it "freezes"?
14:56 < tadao> cough- its an old and ugly riva tnt with 15 " moni so i cant read it -yet kanotix boots with fb and no acpi and readable display
14:57 < rxr> maybe it is an issue with some FB module beeing loaded ?
14:59 < capey> hmm, my screen texts messes up during boot. but gets clear when fb is put to use
14:59 < capey> but i think its only this laptop, since the gfx card is quite beaten already
14:59 < tadao> maybe -I just gues to see the t2 logo (three screens on one overlayed ) -under live cd (slax kanotix ) i can see filesystem and everything ok -so how to switch of apm acpi and that -is there some safe mode ?
15:04 < tadao> btw i thought , there is no x on the box ...( no xorg at least in etc )so why would i need fb for prompt ? i just want to use it for building targets
15:06 < capey> can you disable fb on boot by passing kernel parameter?
15:07 < tadao> what would that be : "nofb" ?
15:10 < capey> dont know, but thats a good guess
15:11 < tadao> so ill go and try that -for the moment its a dual boot with ubuntu emc on hda2 in grub menu list i have only these 2 : emc +t2 -ubuntu has no problems there.
15:12 < rxr> regardless of nofb or so hotplug++/udev might most probably load the nvidia/rivafb module when they see the hardware
15:13 < rxr> the module would need to be placed in the /etc/conf/blacklist file to avoid it beeing loaded
15:14 < tadao> sounds good :nofb worked but delayed the problem only 30 seconds
15:15 < tadao> what would be the exact entry ?
15:15 < rxr> just the module name
15:15 < rxr> rivafb / nvidiafb I think they are
15:15 < rxr> just place them into that file
15:16 < capey> yep rivafb for old ati cards and radeonfb for new cards
15:17 < tadao> wasn't there a trick in grub shell : echo bla > /etc/conf/blacklist ?
15:18 < capey> echo rivafb >> /etc/conf/blacklist
15:18 < capey> not >, use >>, to preserve old lines
15:19 < tadao> thaaaaaaanks -so ill do that
15:19 < rxr> that works in grub ?
15:21 < capey> oh.
15:21 < capey> no
15:22 < capey> tadao: that might not work.... sorry i didnt read that you wanted to do that in grub shell
15:23 < capey> actually it might work, if prefixed it with hd info (hd0,0)/etc/......
15:25 < tadao> well anyway now i know too , cat works though
15:26 < tadao> how to use that ?
15:26 < tadao> cat (hd0,5)/etc/conf gives nme the file
15:27 < tadao> sonice that grub...
15:29 < tadao> anyway undergrub help seems only cat usable ... thought there was some editing too ...
15:29 < tadao> ill do it now with slax and mc
15:30 < tadao> would be a nice patch though for grub . to add echo ...
15:33 -!- Stelz [i=stelz@80.240.210.253] has quit ["There is intelligent life out there..."]
15:34 -!- Stelz [i=stelz@80.240.210.253] has joined #t2
15:37 < capey> what package provides dccprobe?
15:39 < rxr> x11/ddcxinfo
15:49 < CIA-9> rene * r20865 /trunk/package/develop/libtasn1/libtasn1.desc: * updated libtasn1 (0.3.6 -> 0.3.8)
16:01 < tadao> huahh burnt my grub menu.lst .. so anyway :what would be the second best distro to build t2 ? i have kanotix , slax , ubuntu and slackware 11 running around here
16:01 < rxr> any should be able to build t2
16:02 < rxr> if not we will (usually) quickly fix t2 to do so
16:02 < shevegen> kanotix :D
16:02 < shevegen> so it stays in the german-speaking area ;)
16:03 < tadao> so not important if gcc 3.4 (slack11) or 4.1 (Kanotix) ?
16:03 < rxr> nope - t2 should bootstrap on any of them cleanly
16:06 < tadao> aaahhh - so toolchain is downloaded too (not from basedistro ) ? i downloaded everything for target mini-livecd and desktop but i saw only gcc4.11
16:06 < tadao> yet qemu needs 3.x i read
16:09 < rxr> tadao: yes, our qemu is patched to build with gcc.4.1 again ...
16:09 < tadao> wow
16:10 < tadao> and no problems with kqemu ?
16:10 < rxr> guess it works - but I do not use binary only Linux software
16:10 < rxr> but this gcc4 patches are from the qemu "sort of co author" so they should be quite fin
16:10 < rxr> fine
16:19 < tadao> btw grub + stone : is there a way to install it to root and not mbr - or not at all and use existing one?
16:20 < rxr> not at all is possible
16:20 < rxr> but install to root was asked already - we really should implement it
16:20 < rxr> I'll do when I fix the /boot thing - hopefully today
16:21 < tadao> wow
16:21 < tadao> start to get bad habits here
16:21 < tadao> :)
16:21 < rxr> bad habits ?
16:22 < tadao> well the response time is quite impressive here ...
16:22 < tadao> so getting used to it
16:22 < rxr> :-)
16:22 < tadao> users will be spoiled for other distros
16:23 -!- kensai [n=kensai@206.248.86.8] has quit ["Leaving"]
16:25 < capey> hmm, stone xorg module doesnt set windowmanager probely
16:27 < rxr> it used to do so - but you need to reloading as it is sourced when you login
16:27 < capey> ah yes
16:27 < rxr> relogin ...
16:28 < capey> yep, worked as it should
16:28 < capey> just need to wait for kde to build. wont be working through twm here
16:30 < rxr> heh
16:31 < capey> so you got your self avr32?
16:31 < capey> some "hacker" ported doom for the atmel chip that is first avr32 based chip
16:32 < capey> 7000?
16:32 < capey> avr32 starter kit version... ships with QVGA lcd display
16:33 < rxr> I have no avr32 yet
16:33 < rxr> but I plan to get one
16:39 -!- idealm [n=ideal@222.67.239.123] has quit [Remote closed the connection]
16:40 < capey> me too, but havent noticed yet any distributors
16:41 < capey> avr32 or new mips64 chips for "poor-mans" grid, and one avr32 for driving solar cell pumps etc...
16:56 < rxr> the devel boards are way too expensive right now
16:57 < shevegen> i wonder if people are aware that php is not always the best solution :-) http://dbis.informatik.uni-freiburg.de/index.php?file=../../../../../../etc/passwd
16:58 < tadao> somewhere i read php=visualbasic for www bwoaharrharr ...btw target puppy : can i import/copy that simply into my trunk ?
16:59 < tadao> *puppy-linux 2.10
17:00 < rxr> sadly no, because puppyos just used t2 to build the binaries and then postprocessed it with existing scripts
17:01 < rxr> however they published the tarball of t2 with modifications on their development website
17:01 < rxr> maybe ask them if they wanna integreate as real t2 target and build their upcomming version with t2 postprocessing entirely :-)
17:03 < tadao> anyway if its not compatible -i have no interest -i ll stick with mini-live cd target - they use some non standard x too
17:04 < tadao> oooh last question best question ... do you have experience combining livecd + xen (host and guest) ?
17:05 < rxr> nope - there is no T2 Xen LiveCD, yet
17:05 < rxr> would be awesome if you want to create one
17:06 < tadao> it was just not asked before , or are there technical issues ?
17:06 < tadao> there is xenoppix for example ...
17:07 < tadao> so mybe xen doesnt bother if its running from livesystem
17:10 < tadao> qemu for example runs fine from livecd (grml) even with kqemu
17:11 < tadao> well anyway worst thing what can happen while playing around is creating an useless iso ...
17:11 < rxr> yes sure
17:11 < rxr> there is no technical contraint
17:12 < rxr> just noone build Xen into a LiveCD with assoiated preconfiguration
17:12 < rxr> (please do :-)
17:12 < tadao> cough cough
17:13 < Stelz> meow meow
17:13 < rxr> moin moin Stelz
17:14 < Stelz> evening evening rxr :)
17:14 * Stelz :: Date/Time is now [ Sat Nov 18 19:08:54 ]
17:16 -!- sepp_ [n=sepp@p83.129.202.97.tisdip.tiscali.de] has joined #t2
17:23 -!- sepp [n=sepp@p83.129.214.49.tisdip.tiscali.de] has quit [Read error: 60 (Operation timed out)]
17:24 < tadao> ok nice evening everybody and thanks for the answers ....
17:24 -!- tadao [n=tadao@254-199.adsl.pool.ew.hu] has quit ["Leaving"]
17:33 [Users #t2]
17:33 [@ChanServ] [ CIA-9] [ misl ] [ psycodad] [ shevegen] [ zod]
17:33 [ _Ragnar_] [ emte ] [ mtr ] [ rxr ] [ Stelz ]
17:33 [ capey ] [ LMJ ] [ pamdirac] [ sepp_ ] [ valentin]
17:33 -!- Irssi: #t2: Total of 16 nicks [1 ops, 0 halfops, 0 voices, 15 normal]
18:10 [Users #t2]
18:10 [@ChanServ] [ CIA-9] [ misl ] [ psycodad] [ shevegen] [ zod]
18:10 [ _Ragnar_] [ emte ] [ mtr ] [ rxr ] [ Stelz ]
18:10 [ capey ] [ LMJ ] [ pamdirac] [ sepp_ ] [ valentin]
18:10 -!- Irssi: #t2: Total of 16 nicks [1 ops, 0 halfops, 0 voices, 15 normal]
18:41 < CIA-9> rene * r20869 /trunk/package/zope/zopeinterface/ (. gcc41.patch zopeinterface.cache zopeinterface.desc):
18:41 < CIA-9> * added zopeinterface (3.0.1) - zope.interface package from Zope 3,
18:41 < CIA-9> needed for twisted, in turn needed for the new bittorrent
18:42 < CIA-9> rene * r20870 /trunk/package/network/twisted/ (. twisted.cache twisted.desc):
18:42 < CIA-9> * added twisted (2.4.0) - An event-based networking framework for
18:42 < CIA-9> Internet applications, needed for the new bittorrent
18:42 < CIA-9> rene * r20871 /trunk/package/network/bittorrent/ (bittorrent.cache bittorrent.desc): * updated bittorrent (4.20.2 -> 5.0.1)
18:45 -!- sepp_ is now known as sepp
19:29 < CIA-9> rene * r20872 /trunk/package/network/mod_security/mod_security.desc: * updated mod_security (2.0.3 -> 2.0.4)
19:39 < CIA-9> rene * r20873 /trunk/package/printing/cups/cups.desc: * updated cups (1.2.6 -> 1.2.7)
19:40 < CIA-9> rene * r20874 /trunk/package/ (x11/xli/xli.desc zope/ihotfix/ihotfix.desc): * cleaned superflous spacing, dots and comments in xi and ihotfix
19:42 [Users #t2]
19:42 [@ChanServ] [ CIA-9] [ misl ] [ psycodad] [ shevegen] [ zod]
19:42 [ _Ragnar_] [ emte ] [ mtr ] [ rxr ] [ Stelz ]
19:42 [ capey ] [ LMJ ] [ pamdirac] [ sepp ] [ valentin]
19:42 -!- Irssi: #t2: Total of 16 nicks [1 ops, 0 halfops, 0 voices, 15 normal]
19:53 -!- zod [i=zod@shodan.knid.net] has quit ["Terminated with extreme prejudice - dircproxy 1.0.5"]
20:09 -!- zod [i=zod@shodan.knid.net] has joined #t2
20:25 < capey> hmm, trunk admin awake?
20:28 < capey> rxr, would it be need big changes on current script system, for defining new variable that defines what script will be used to build pkg.
20:29 < capey> rxr, i would override the default one with the one that adds package build task to grid
20:29 < _Ragnar_> like custmain? o_O
20:29 < capey> thats for per package? i would need per target.
20:30 < _Ragnar_> oh
20:30 < _Ragnar_> hmm what do you actually wanna do?
20:31 < capey> well, Build-Target, builds packges through Build-Pkg script. on target build, i would send the package build task to grid system and let it build it there.
20:31 < _Ragnar_> what do you mean, grid system? like a cluster?
20:31 < capey> and fetch files after that or by using some networkfs
20:32 < _Ragnar_> I thought we had a distcc build already ... o_O
20:33 < capey> distcc distributes build load by compiled file
20:33 < _Ragnar_> ah yea
20:33 < capey> this would distribute load by compiled package
20:34 < capey> should be lot faster, since lower data rates and allows distribution to be on much large scale geographically
20:34 < _Ragnar_> hmmm but you'd have to have the build filesystem synchronized since packages tend to depend on each other
20:34 < capey> after package build yes
20:34 < _Ragnar_> like kde* all depends on kdebase, so you can'd build anything from that before kdebase
20:35 < _Ragnar_> also, you'd then need to send the files created by kdebase to all the other build systems when it's done
20:35 < capey> yep, thats true
20:36 < capey> that can be done using existing network fs or by custom one with sended data size in mind
20:37 < _Ragnar_> yea but you're sending 100s of megs around that way ... is that really gonna bring and improvement?
20:37 < _Ragnar_> any*
20:37 < _Ragnar_> cpu speed seems to grow faster then internet speed atm
20:38 < capey> thats the same principal on distcc. distcc just send files on smaller chunks in lesser time
20:38 < capey> time period
20:38 < _Ragnar_> yea but it's supposed to be on a local net in general
20:38 < capey> thats true
20:38 < _Ragnar_> and it should only send the needed files ... like for kde, it would only send the include files and libs, not the executables or the artwork
20:39 < capey> yep...
20:40 < _Ragnar_> well ok you could do the same for your approach with some network filesystem
20:40 < capey> well, generally speaking. some have "interrupts" on file opens and fetches file over network after that it lets the event pass through.
20:41 < _Ragnar_> yea
20:41 < _Ragnar_> hmmm ... might be worth looking at ... but you'd need to write a FAQ on how to set it up too :)
20:41 < capey> yep :)
20:42 < _Ragnar_> so ... if we have consistent storage ... we can substitute the call to Build-Pkg with an ssh call to build remotely
20:42 < capey> still needs lot of work to get the support for T2 grid building. my current framework wasnt designed to do that. it was designed to distribute math calculations
20:43 < emte> umm
20:43 < emte> it sounds liek you guys are talking about how icecream works ...
20:43 < emte> like*
20:43 < capey> yes
20:43 < capey> :)
20:43 < _Ragnar_> icecream?
20:43 < emte> simular to distcc
20:43 < _Ragnar_> ah ok
20:44 < emte> but more along what you guys are talking about
20:44 < emte> btw i have no performance issues with distcc
20:44 < capey> distcc works
20:44 < emte> all my links are over the net
20:44 < capey> thats the beauty of it
20:44 < emte> icream works too, just not fully for T2 yet
20:45 < capey> have to check that.... you know the url?
20:45 < emte> have to check what?
20:45 < emte> that icecream works?
20:45 < _Ragnar_> well if you get the fs right, you would save a lot of redundant transfers
20:46 < emte> a network fs isnt the right way to go ... imo
20:46 < _Ragnar_> why not?
20:46 < emte> they are far slower than tranfering chunks
20:46 < _Ragnar_> only until the file is cached on the other side
20:46 < _Ragnar_> then it should be a lot faster
20:47 < emte> if the remote host has ccache the data is cached anyway
20:47 < _Ragnar_> also, if the grid scheduler is smart, it could have all the kde packages build on one host, and all gnome on the other, in parallel
20:47 < emte> you can allocate things ...
20:47 < emte> atleast in distcc
20:47 < _Ragnar_> ccache never gave any big improvements for me
20:48 < emte> ccache only shows it's use on recompiles
20:48 < emte> and cuts compile time in 25-75%
20:48 < _Ragnar_> yea but it still will spread out all the load between the nodes, so all the files have to be on all the nodes in the end
20:48 < emte> not exactly
20:48 < _Ragnar_> how not?
20:49 < emte> the only thing the remote hosts need is the cross compiler
20:49 < emte> they dont need the full build enviroment
20:50 < emte> atleast that is my experiance so far
20:50 < emte> i just export my cross compilers to my distcc links
20:50 < emte> and then they love me
20:50 < emte> :)\
20:51 < emte> which reminds me
20:51 < emte> i need to contact one of my links
20:52 -!- F-117 [i=stelz@80.240.210.253] has joined #t2
20:52 < _Ragnar_> *g*
20:52 -!- F-117 [i=stelz@80.240.210.253] has quit [Client Quit]
20:53 < _Ragnar_> yea but you still have to transfer all the data for every build call to the nodes
20:53 < _Ragnar_> include files, libs, ...
20:54 < _Ragnar_> if you have a dedicated build host for one package, you only have to transfer the data to one host
20:55 < emte> i have yet to notice transfer of any libs/includes
20:55 < emte> just small chunks
20:55 < _Ragnar_> well it has to happen, otherwise there'll be no working object file o_O
20:55 < emte> not really
20:56 < emte> binary data is binary data
20:56 < _Ragnar_> you saying the node magically guesses what it needs to compile? :)
20:56 < emte> you send what you need to compile
20:56 < _Ragnar_> exactly.
20:56 < _Ragnar_> this includes the .c file, all the header files, and if it's a link, all the libraries
20:57 < _Ragnar_> and you transfer that for every distcc call
20:57 < emte> i belive i set the chunk size tho
20:59 < _Ragnar_> how does that matter? :)
21:01 < emte> interesting bug
21:01 < emte> Writing object files from distcc to an NFS directory can cause corrupt output: object files will be full of zeros.
21:01 < emte> looking now
21:02 < _Ragnar_> anyway, gotta go, bbl :)
21:02 < emte> hmm kde is funny
21:02 < emte> Using the --enable-final configure option of KDE makes distcc almost useless.
21:02 < emte> --enable-final makes the build system concatenate all sourcefiles in a directory (say, Konqueror's sourcefiles) into one big file.
21:05 < emte> nice script ... i should modify it for heavy shoutcast use http://lists.samba.org/archive/distcc/2006q1/003271.html
21:06 < emte> okay time to fight with tslib
21:07 < emte> http://www.rafb.net/paste/results/7ukLIh42.html
21:07 < emte> any ideas?
21:08 < emte> is there a way to suppress autogen?
21:20 < capey> emte: working only on console, what can you paste the most important lines?
21:21 < capey> suppress?
21:21 < capey> as in not run?
21:21 < emte> yes
21:21 < capey> set autogen=0 in .conf
21:22 < emte> !>configure.ac:25: error: possibly undefined macro: AC_DISABLE_STATIC
21:22 < emte> !> configure.ac:26: error: possibly undefined macro: AC_ENABLE_SHARED
21:22 < emte> !> configure.ac:27: error: possibly undefined macro: AC_LIBTOOL_DLOPEN
21:22 < emte> !> configure.ac:28: error: possibly undefined macro: AC_PROG_LIBTOOL
21:22 < capey> define autogen=0 in .conf file
21:22 < emte> doing so now
21:24 < emte> i am thinking its an autotool incompatability
21:26 < emte> nope that didnt seem to do it
21:26 < emte> hmm its still running autogen
21:30 < capey> i have gui now, can you paste new results?
21:30 < emte> same
21:30 < capey> from src.tslib*/ERROR-LOG
21:30 < emte> yup
21:31 < emte> http://www.rafb.net/paste/results/WbINNf46.html
21:33 < emte> i am hoping kergoth wakes up at some point so i can ask him whats up
21:35 < capey> can you paste .conf?
21:35 < emte> only has what you stated
21:35 < capey> also you could try to generate it self, by running libtoolize --force --automake;aclocal
21:36 < capey> not expert on auto* stuff...
21:36 < emte> i suspect that tslib's autotool stuff is from an old gen3 setup
21:36 < emte> which makes it incompat with newer
21:37 < emte> which is why i am trying to suppres regen
21:37 < capey> hook_add preconf 9 "libtoolize --force --automake;aclocal;autoconf"
21:38 < emte> same issue
21:38 -!- sireficuL [n=CAM@89.240.84.130] has joined #t2
21:38 -!- sireficuL [n=CAM@89.240.84.130] has left #t2 []
21:38 < misl> capey: mipe in disguise ?
21:39 < misl> haven't seen you in while :)
21:39 < capey> yep, mipe was already registerd one. so i registered this
21:39 < capey> :)
21:39 < capey> hi misl
21:39 < misl> hi capey :)
21:39 < misl> where have you been all this time ?
21:40 < capey> finished up my studies, moved to other city. and waited to get new network connection to home
21:40 < misl> does it take that long in .fi to get wired?
21:41 < capey> 4 weeks
21:41 < misl> guess you finished studies with success :)
21:41 < capey> yep :)
21:41 < misl> 4 weeks, I guess that's in winter when everything is coverd in snow ;-)
21:42 < capey> heh
21:42 < misl> good to see you again :-)
21:43 < misl> I already wondered where you had gone, so welcome back :)
21:43 < capey> thanks :)
21:43 < emte> maybe its some weird caching issue ... wiping everything and rebuilding
21:44 < capey> emte: theres some messages about same stuff at http://lists.arm.linux.org.uk/pipermail/linux-arm-kernel/2002-February/007288.html
21:44 < capey> heh
21:44 < capey> now i noticed that its dated 2002
21:45 < emte> kergoth hasnt updated it in a while
21:45 < emte> no real need
21:46 < emte> touchscreens havent changed
21:47 < capey> ah... at stage 1? is this cross
21:47 < capey> build
21:47 < capey> make sure you have libtool before
21:47 < emte> i just need to suppress regenerating
21:47 < emte> since the autotool versions dont recog the same macros
21:48 < capey> IIRC, autogen=0 in config
21:48 < capey> lemme check
21:48 < emte> i have that and its still regenerating
21:48 < emte> but it might be somefubf that Clean -cache isnt getting
21:48 < emte> something*
21:50 < emte> and yes libtool is built
21:53 < capey> well, autogen=1, might work
21:54 -!- Stelz [i=stelz@80.240.210.253] has quit ["There is intelligent life out there..."]
21:56 -!- Stelz [i=stelz@80.240.210.253] has joined #t2
22:02 < emte> ... these really need doccumentation
22:03 -!- kensai [n=kensai@206.248.86.8] has joined #t2
22:06 < emte> will change that as soon as it errors out
23:06 -!- misl [n=chatzill@84-104-172-187.cable.quicknet.nl] has quit ["Chatzilla 0.9.75 [Firefox 1.5.0.8/2006102516]"]
23:44 -!- mtr_ [n=Michael@pD9E11685.dip0.t-ipconnect.de] has joined #t2
23:53 -!- mtr [n=Michael@pD9E11B36.dip0.t-ipconnect.de] has quit [Read error: 145 (Connection timed out)]
23:53 -!- mtr_ is now known as mtr
--- Log closed Sun Nov 19 00:00:39 2006