T2 IRC Log: 2008-08-23
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 Aug 23 00:00:10 2008
01:48 -!- Netsplit calvino.freenode.net <-> irc.freenode.net quits: @ChanServ, felanha
01:49 -!- Netsplit over, joins: @ChanServ, felanha
01:50 -!- Netsplit calvino.freenode.net <-> irc.freenode.net quits: @ChanServ, felanha
01:51 -!- Netsplit over, joins: @ChanServ, felanha
01:58 < CIA-5> ragnar * r30175 /trunk/package/base/mdadm/mdadm.desc: * mdadm: mark as compatible with dietlibc
02:05 < CIA-5> ragnar * r30176 /trunk/package/base/mkinitrd/ (initrdinit.sh mkinitrd.sh): * mkinitrd: make initrd aware of md devices (only raid1 atm)
02:10 < CIA-5> ragnar * r30177 /trunk/package/base/mdadm/preconfig.in: * mdadm: have it build with dietlibc by default
06:14 -!- emte [n=emte@d154-5-45-238.bchsia.telus.net] has joined #T2
07:29 -!- emte [n=emte@d154-5-45-238.bchsia.telus.net] has quit [Read error: 104 (Connection reset by peer)]
09:02 < rxr> moin
09:02 < rxr> Ragnar|away: neat
09:04 < CIA-5> rene * r30178 /branches/7.0/package/ruby/ruby/ruby.conf:
09:04 < CIA-5> * added sparc* to the architectures where ruby segfaults if
09:04 < CIA-5> pthread support is enabled
09:06 < CIA-5> rene * r30179 /trunk/package/ruby/ruby/ruby.conf:
09:06 < CIA-5> * merged r30178 from branches/7.0: added sparc* to the architectures
09:06 < CIA-5> where ruby segfaults if pthread support is enabled
09:17 -!- LMJ [n=serwou@laf31-4-82-236-42-164.fbx.proxad.net] has quit [Connection timed out]
09:18 -!- LMJ [n=serwou@laf31-4-82-236-42-164.fbx.proxad.net] has joined #t2
09:33 < CIA-5> rene * r30180 /trunk/package/editors/ed/ed.desc: * updated ed (0.9 -> 1.0)
09:33 < CIA-5> rene * r30181 /trunk/package/network/neon/neon.desc: * updated neon (0.28.2 -> 0.28.3)
09:36 < CIA-5> rene * r30182 /trunk/package/network/ntp/ntp.desc: * updated ntp (4.2.4p4 -> 4.2.4p5)
09:39 < CIA-5> rene * r30183 /trunk/package/editors/moe/ (. moe.cache moe.desc): * added moe (1.0) - A powerful and user-friendly console text editor
09:39 < CIA-5> rene * r30184 /trunk/package/textproc/libmrss/libmrss.desc: * updated libmrss (0.19.1 -> 0.19.2)
09:39 < CIA-5> rene * r30185 /trunk/package/textproc/libnxml/libnxml.desc: * updated libnxml (0.18.2 -> 0.18.3)
09:40 -!- Baldzius [n=Baldzius@89.127.175.95] has joined #t2
09:40 < Baldzius> moin
09:40 < rxr> hi Baldzius
09:40 [Users #t2]
09:40 [@ChanServ] [ CIA-5 ] [ dsoul ] [ felanha] [ Ragnar|away] [ TobiX]
09:40 [ axionix ] [ Cyda ] [ DuckFault] [ LMJ ] [ rxr ]
09:40 [ Baldzius] [ Dallur] [ Enqlave ] [ mtr ] [ timlarson ]
09:40 -!- Irssi: #t2: Total of 16 nicks [1 ops, 0 halfops, 0 voices, 15 normal]
09:40 < Baldzius> hey rxr
09:40 < rxr> how do you do ?
09:40 < CIA-5> rene * r30186 /trunk/package/network/nss_ldap/nss_ldap.desc: * updated nss_ldap (261 -> 262)
09:41 < Baldzius> thanks, not bad, trying to rest after the night shift of whole week
09:41 < Baldzius> yourself?
09:45 < rxr> oh, night shift - does not sound fun
09:46 < rxr> I'm fine, too :-)
09:46 < rxr> thanks
09:48 < Baldzius> it's not so bad as it sounds but the worst thing is to go back to the normal routine
09:48 < Baldzius> cause now i got used to sleep on days rather nights :)
09:48 < Baldzius> need to switch back so it takes couple of days
09:49 < CIA-5> rene * r30187 /trunk/package/network/madwifi/kernel-change.patch: * svn rm madwifi/kernel-change.patch missed from ML patch in r30165
09:50 < rxr> Baldzius: noticed, we got new, pretty IRC logs: http://www.t2-project.org/irclogs/
09:50 < Baldzius> *click*
09:51 < rxr> and now 127 motivated enough to add themselv to the startpage: http://www.t2-project.org/startpage/
09:52 < Baldzius> nice
09:52 < rxr> and increased ML messaging :-)
09:53 < Baldzius> yeah, noticed that too
09:58 < CIA-5> rene * r30188 /trunk/package/editors/moe/moe.desc: * marked the "new" moe editor to cross build
09:58 < rxr> do you know something serious in the 7.0 branch? or can we go gold hopefully soonish ?
09:59 < Baldzius> no, haven't used 7.0 for a long time now
09:59 < Baldzius> sitting exclusively on trunk
10:00 < rxr> ah - ok as well
10:00 < rxr> despite sparc64 on the 360MHz box the x86, x86-64, ppc, ppc64 builds are sort of ready
10:01 < rxr> (mostly require a final test install ...)
10:01 < Baldzius> oh maybe one thing, try to install to is using ext3 as root
10:01 < Baldzius> the last time it didn;t work for me somehow
10:02 < Baldzius> so i had to stick with reiserfs
10:03 < CIA-5> rene * r30189 /branches/7.0/package/xorg/xorg-server/parse-config: * merged r30174 from trunk: silenced r30153 pkgprefix warning if libx11 was not yet build
10:15 < CIA-5> rene * r30190 /trunk/target/desktop/config.in: * removed "Exact" branding from desktop config entry
10:21 < CIA-5> rene * r30191 /branches/7.0/target/ (desktop/pkgsel generic/pkgsel/10-minimal-xorg.in):
10:21 < CIA-5> * removed freetype1 from generic minimal-xorg package selection
10:21 < CIA-5> template, removed duplicate selection from desktop package
10:21 < CIA-5> selection
10:22 < CIA-5> rene * r30192 /trunk/target/ (desktop/pkgsel generic/pkgsel/10-minimal-xorg.in):
10:22 < CIA-5> * merged r30191 from branches/7.0: removed freetype1 from generic
10:22 < CIA-5> minimal-xorg package selection template, removed duplicate selection
10:22 < CIA-5> from desktop package selection
10:50 -!- hwinkel [n=hwinkel@p54B69573.dip0.t-ipconnect.de] has joined #t2
10:55 < CIA-5> rene * r30193 /branches/7.0/package/security/gnupg2/gnupg2.conf: * work arounded inf. looping gnupg2 tests on powerpc64: skip them for now
10:56 < CIA-5> rene * r30194 /trunk/package/security/gnupg2/gnupg2.conf:
10:56 < CIA-5> * merged r30193 from branches/7.0: work arounded inf. looping gnupg2
10:56 < CIA-5> tests on powerpc64: skip them for now
10:59 -!- mpp [n=user@i53875BCA.versanet.de] has joined #t2
10:59 < mpp> guten tag allerseits
10:59 < rxr> moin moin mpp
11:00 < mpp> i built successfully the modified rescue target for pxe
11:00 < mpp> it works fine !
11:00 < mpp> no i am adding some network tools
11:00 < mpp> does anyone know where the ping resides in ?
11:00 < rxr> you tested network booting already ?
11:01 < mpp> i mean which packege
11:01 < mpp> yep
11:01 < mpp> works like a charm
11:01 < rxr> great
11:01 < mpp> i removed the x stuff from the target
11:01 < mpp> thanx to you
11:01 < mpp> !
11:01 < rxr> in t2 var/adm/* hols the installed package stuff in plain/text ASCII files, so you can find ping like this:
11:01 < rxr> # grep bin/ping /var/adm/flists/*
11:01 < rxr> /var/adm/flists/netkit-base:netkit-base: bin/ping
11:01 < mpp> cool
11:02 < rxr> to find ping related stuff as good as the maintainers filled the T2 .desc files in your t2 tree:
11:02 < mpp> these irc-logs are pretty precious
11:02 < rxr> # ./scripts/Find-Pkg ping
11:02 < Ragnar|away> that usually just gives me a 'line too long' error ;P
11:02 < mpp> nice
11:02 < mpp> i'll give it a shot right now
11:02 < rxr> Ragnar|away: only if you grep in /home/ragnar/src/t2-trunk-foo-bar/build/my-super-target-8.0-x86-64-core2-cross/var/adm/flists/* :-)
11:03 < rxr> Ragnar|away: recent kernels (after 2.6.23 or so) have dynamic command line allocation, now, FINALLY
11:03 < Ragnar|away> oh kewl :)
11:03 < rxr> so you should see this less often even in the above case
11:03 < rxr> though you can limit it with ulimits
11:03 < rxr> IIRC it defaults to 1/4 or so of the stack limit
11:04 < rxr> someone from google jumped over the quick brown fox and implemented this because it annoyed him too much :-)
11:04 < rxr> he came from plan9 ...
11:05 < Ragnar|away> you're right, it works now :) sweeeety
11:05 < mpp> by the way
11:05 * Ragnar|away dances
11:05 < mpp> the madwifi package has almost all the time trouble getting downloaded
11:05 * rxr wants to see a youtube video of that :-)
11:05 < mpp> so i had to exclude this
11:05 < Ragnar|away> LOL
11:05 < rxr> mpp: I guess that's the SSL cert of the server
11:06 < rxr> mpp: I hope the madwifi checkout is on the t2 mirrors by now so you should be able to crab it from somewhere
11:06 < rxr> maybe svn has some "--eat-the-cert" option to avoid this?
11:07 < Ragnar|away> madwifi-r3375.tar.bz2 ?
11:07 < rxr> the new one from yesterday, whatever the -r was .-)
11:08 < Ragnar|away> ooh ok
11:11 < Ragnar|away> I see what you mean
11:14 < rxr> mpp: btw. this flist program and other files search is on my TODO to integrate into the homepage package matrix
11:14 < Ragnar|away> hmm can we code the answer to the 'accept cert' into the Download ?
11:15 < rxr> so one can consult that to find where some obscure binary or header is located in
11:15 < Ragnar|away> oh btw when do you usually update the .cache files?
11:16 < rxr> Ragnar|away: no idea yet, mache echo -e "y\n" | would work - maybe not, svn has some --non-interactive switch, I just tried, it does not accept the cert then - maybe it has an hidden switch for that
11:16 < rxr> Ragnar|away: we should automate the .cache update more, indeed
11:16 < Ragnar|away> mpp: you can do the checkout once by hand, accept the cert permanently, then run the Download
11:16 < rxr> Ragnar|away: right now I update them when something majorly changes
11:16 < rxr> yes, I thought I had done so on the master server after I noticed this on my regression tester
11:19 < CIA-5> rene * r30195 /trunk/package/lua/luaffi/ (. lua-lib.patch luaffi.cache luaffi.conf luaffi.desc): * added luaffi (0.9.0) - FFI bindings to Lua
11:23 -!- Baldzius [n=Baldzius@89.127.175.95] has quit [Remote closed the connection]
11:25 < mpp> hey rxr
11:25 < mpp> the udhcpc should work in the rescue image ß
11:25 < mpp> ß
11:25 < mpp> ß
11:25 < mpp> damnit
11:25 < rxr> yeah the client scripts ?
11:25 < mpp> right
11:25 < rxr> it already on my todo ... :-(((
11:25 < rxr> patch welcome
11:25 < mpp> they give an error when booted
11:26 < mpp> a cool
11:26 < mpp> so ill dig into it
11:26 < rxr> yeah - not so cool that it does not yet just work ...
11:26 < rxr> yeah - patch most welcome
11:26 < rxr> maybe in conjunction with the queued bind one :-)
11:26 < mpp> im on it
11:26 < mpp> i have that already filed
11:26 < mpp> ill submit
11:29 -!- mpp [n=user@i53875BCA.versanet.de] has left #t2 []
11:32 -!- mpp [n=user@i53875BCA.versanet.de] has joined #t2
11:35 < mpp> ?
11:35 < mpp> ah
11:35 < mpp> keyboard working again
11:35 < mpp> nice :-)
11:37 < mpp> allright i user dhcpcd
11:37 < mpp> it works
11:37 < mpp> only thing is
11:38 < mpp> the directory /var/lib and /var/lib/dhcpcd were missing in the initramfs
11:38 < mpp> how to i patch the dhcpcd to include these directories ?
11:38 < rxr> are they empty or contain files ?
11:38 < mpp> they don't exist on the booted image
11:38 < mpp> so i created them manually
11:38 < rxr> and in the raw build ?
11:38 < mpp> and it worked
11:38 < rxr> build/...
11:38 < mpp> raw build ?
11:39 < mpp> you mean in the build/TOOLCHAIN ?
11:39 < rxr> the build before filtering and image compression
11:39 < rxr> I mean build/$configid-$uuid/
11:39 < mpp> ill check
11:39 < rxr> ok - my rescue build looks like it does not exist
11:39 < mpp> missing there
11:40 < rxr> if that's the case for you a hook_add postmake 5 "mkdir -p $root/var/lib/dhcpcd" should do
11:40 < rxr> --typos, etc.
11:40 < rxr> rebuild the package and rebuild the image
11:41 < mpp> where do i do this modification ?
11:41 < mpp> hook_add ?
11:42 < rxr> sorry in the .conf file
11:42 < mpp> i'll think ill read the dev docu first to catch up with this
11:42 < mpp> okay
11:42 < rxr> it's a t2 function to add code to some hook
11:42 < rxr> each package con influence it's build with conde in package/*/$pkg/$pkg.conf
11:42 < mpp> right
11:43 < mpp> but in this case we could also create them in the rootfs of the target definition
11:43 < mpp> which is bad as i see
11:43 < rxr> but then all the other targets would not be fixed
11:43 < mpp> vor the package is beter
11:43 < mpp> yep got it
11:43 < mpp> t2 rocks ;-)
11:43 < rxr> if the package needs some directory to work it should be in the package, not just a single target
11:43 < mpp> right
11:44 < mpp> looks like /var/lib is missing in this target in general
11:44 < mpp> which is not a good idea
11:45 < rxr> it probably just that no package is in the selection that created it
11:45 < mpp> ah i see
11:46 < mpp> so were do we patch for this
11:46 < rxr> with the hook above :-)
11:46 < mpp> i mean dhcpcd is very cool and udhcp gave lots of erros of diverse files and paths
11:46 < mpp> allright
11:46 < rxr> ah - you exchanged udhcpd with dhcp in your target ?
11:47 < mpp> no i kicked them both and used dhcpcd only
11:47 < mpp> works like a charm
11:47 < mpp> im rebuilding the target with the paths added
11:48 < mpp> wouldn it make sence to have a file hierarchy template used
11:48 < mpp> even if some paths were empty for certain targets
11:48 < rxr> not really
11:48 < mpp> okay - explain please
11:48 < rxr> t2 is about cleanlyness and sorted stuff
11:49 < mpp> capiche !
11:49 < mpp> accepted
11:49 < rxr> if you have 100 extra dirs, empty then they are not only confusing and a mess, but also waste space on very minimal things
11:49 < rxr> for example we build < 2MB systems etc. ...
11:49 < rxr> you do not want to have 100 directories in them just in case ...
11:50 < rxr> if some package does rely on a not yet existing directory, then this single package needs to be fixed
11:50 < rxr> in this case it looks like a) this dhcpcd needs to be fixed to create this var/lib/dhcpd directory and the udhcpd needs to be fixed to find it's freaking config scripts
11:51 < rxr> the nearly only other thing would coudl do additionaly is to add /var/lib to the generic "FHS" 00-dirtree layout, but not sub-folders within it
11:52 < mpp> submited the patch via mail
11:52 < mpp> and other to follow for sure
11:52 < mpp> now that i know how and what to do
11:52 < mpp> so ill be tuning the pxe-rescue target since i need it professionally
11:53 < mpp> i called it pxe-rescue
11:53 < mpp> if its okay whith you
11:53 < rxr> sounds good, without the tinyX stuff it will probably be way smaller
11:53 < mpp> i kicked that out already
11:53 < mpp> was easy in the target definition
11:54 < mpp> even for a noob like me
11:58 < mpp> initrd is about 63mb for now
11:58 < mpp> i need ruby in the rescue image so ill add that to the target
12:10 -!- hwinkel [n=hwinkel@p54B69573.dip0.t-ipconnect.de] has quit [Client Quit]
12:13 -!- koan [n=koan@pec.ulyssis.student.kuleuven.be] has joined #t2
12:18 -!- koan [n=koan@unaffiliated/koan] has quit [Client Quit]
12:18 -!- koan [n=koan@unaffiliated/koan] has joined #t2
12:19 < CIA-5> rene * r30196 /branches/7.0/package/shells/mc-mp/umode_t.patch: * fixed mc-mp to build on powerpc64
12:22 < mpp> @rxr i did create a dhcpcd.conf inside the package/network/dhcpcd dir and added two lines for /var/lib and /var/lib/dhcpcd
12:22 < mpp> i rebuilt but the dirs did not get added to the initrd.img
12:22 < rxr> just a single mkdir -p ... inside the hook_add is enough
12:22 < rxr> what did you put into the dhcpd.conf ?
12:23 < CIA-5> rene * r30197 /trunk/package/shells/mc-mp/umode_t.patch: * merged r30196 from branches/7.0: fixed mc-mp to build on powerpc64
12:24 < mpp> these hook_add postmake 5 'mkdir -p $root/var/lib'
12:24 < mpp> hook_add postmake 5 'mkdir -p $root/var/lib/dhcpcd'
12:25 < rxr> you did rebuild the package ?
12:25 < mpp> yes
12:25 < mpp> and there should be a var/lib inside the build/...
12:25 < mpp> right ?
12:25 < rxr> yes, is there ?
12:26 < mpp> no it's not
12:26 < mpp> do i have to make a cleanup
12:28 < rxr> not really
12:28 < rxr> how did you rebuild the package ?
12:28 < mpp> scripts/Config -cfg pxe-rescue
12:28 < mpp> scripts/Download -cfg pxe-rescue -required
12:28 < mpp> scripts/Build-Target -cfg pxe-rescue
12:29 < rxr> ok - that did not rebuild the package
12:29 < rxr> you neither had to run Config nor Download
12:29 < mpp> okay..
12:29 < rxr> (btw. Download is optional today in any case, anyway - T2 auto-downloads when missing ... )
12:29 < rxr> for a package rebuild you can use two main ways (beside other manual hackery)
12:29 < mpp> with the scripts/Build-Target -job 5-dhcpcd then ?
12:29 < rxr> the first, explicitly one package:
12:29 < rxr> yes, like you said:
12:30 < rxr> scripts/Build-Target -job 5-$pkg
12:30 < rxr> or, to rebuild what changed:
12:30 < rxr> scripts/Create-ErrList -cfg rescue -newremove
12:30 < rxr> removed logfiles of package where anything changed
12:30 < rxr> for me this was now:
12:30 < rxr> Scheduling package 'mc-mp' for rebuild ...
12:30 < rxr> Scheduling package 'mkinitrd' for rebuild ...
12:30 < rxr> Scheduling package 'madwifi' for rebuild ...
12:31 < rxr> and then just re-run Build-Target, it will build the stuff Create-ErrList ... removed the log-files from
12:31 < rxr> btw. your hook works, I tested it on my side right now ...
12:31 < mpp> ah now rebuilding the package
12:31 < mpp> thanx !
12:31 < mpp> t2 is fun..
12:32 < mpp> i also want to remove the login
12:32 < mpp> in the inittab i think
12:32 < mpp> since it is rescue
12:32 < mpp> for my X-removed version here
12:37 < mpp> how can i add some post-processing to my target before the initrd.img is build
12:37 < mpp> ?
12:40 < rxr> either by patching packages
12:40 < rxr> or if you want to add tweak files via target/.../rootfs
12:40 < mpp> thats the one
12:40 < rxr> you probably already copied the rootfs "overlay" from the rescue target
12:40 < mpp> i did
12:40 < rxr> you can just add files, or even patch them
12:40 < mpp> so i just place a patched inittab there and rebuild ?
12:43 < rxr> yes, if you place a file there it will just be copied
12:43 < rxr> it's a little more elegant to patch a file
12:44 < rxr> you can place a patch file there starting with #!patch to indicate it's a patch
12:44 < rxr> but maybe start with replacing whole files
12:44 < mpp> can i get rid of these error logs on my build host :
12:44 < mpp> build/pxe-rescue-8.0-trunk-pxe-rescue-x86-i486/var/adm/cache/udhcp should not be present (now in src/invalid-files.lst)!
12:45 < mpp> since these packages are'nt needed anymore
12:46 < rxr> rm `cat src/invalid-files.lst `
12:58 < CIA-5> rene * r30198 /trunk/package/network/dhcpcd/dhcpcd.conf:
12:58 < CIA-5> Martin Papadopoulos:
12:58 < CIA-5> * fixed dhcpcd to create required var/lib/dhcpcd directory
13:06 < CIA-5> rene * r30199 /trunk/package/multimedia/dvbd/dvbd.init:
13:06 < CIA-5> * fixed dvbd.init to not yield errors if the dvbd socket directory
13:06 < CIA-5> already exists
13:10 < mpp> hey rxr
13:10 < mpp> can i submit my pxe-rescue tree to the t2-trunk ?
13:11 < rxr> sure if you like to
13:12 < mpp> how do i do that ?
13:12 < rxr> maybe create a tar and send it over
13:12 < mpp> okay thats cool
13:12 < mpp> send it via ? maillist :-)
13:12 < rxr> sure
13:22 < mpp> how do i deactivate the login ?
13:22 < mpp> in the inittab i think but how exactly
13:27 < rxr> replate the agetty lines with something that does not ask a password
13:27 < rxr> agetty -i 38400 vc/1 -n -l /sbin/login-shell
13:27 < rxr> ..
13:27 < rxr> might do
13:42 < mpp> does not :-)
13:43 < rxr> what does it do
13:43 < rxr> ?
13:43 < mpp> still login prompt but i gues is screwed the line in the inittab
13:43 < mpp> ill retry
14:07 -!- koan_ [n=koan@unaffiliated/koan] has joined #t2
14:07 -!- koan_ [n=koan@unaffiliated/koan] has quit [Client Quit]
14:20 < mpp> okay i tried but still the login prompt appears
14:29 < rxr> hm - then I would have to read man-pages and try as well ...
14:33 < mpp> okay never mind for now - ill dig around also
15:37 < mpp> @rxr: does the rescue target execute rcX.d/script ?
15:37 < mpp> i got the impression that id doesn't
15:42 < CIA-5> rene * r30200 /trunk/package/lua/luacurl/ (. luacurl.cache luacurl.conf luacurl.desc): * added luacurl (1.0) - A Lua binding providing Internet capabilities based on CURL
15:45 < rxr> ah - heh :-)(
15:45 < rxr> the rescue target uses the new, and tiny and fast init called minit
15:45 < rxr> it's based on something entirely different
15:45 < rxr> ah !"
15:45 < rxr> that's also why your inittab stuff has no effect :-)!!!
15:46 < rxr> the minit config is in /etc/minit
15:46 < rxr> http://www.t2-project.org/packages/minit.html
15:48 < mpp> makes sence !
15:48 < rxr> I added "cleaning the sysfiles package to not provide the lingering inittab file but instead move it into the sysvinit package to my TODO)
15:53 < mpp> so what parmams should i pass to the fgetty besides vc/1 to avoid the prompt ?
15:53 < rxr> -l /sbin/login-shell
15:53 < rxr> ah - second
15:53 < rxr> not existing in my system
15:54 < rxr> maybe /bin/bash --login would do
15:54 < rxr> -l /bin/bash --login or so ...
15:54 < mpp> okay ill dig throug minit docu
15:54 < rxr> maybe with quotes
15:54 < rxr> it's not minit
15:54 < rxr> it's about the "getty" ...
15:54 < rxr> maybe try in a running system with vc/7 or vc/8, etc ...
15:55 < rxr> no need to rebuild image after image to give it a try
15:55 < rxr> just start the line as root ...
15:56 -!- dsoul [i=darksoul@vice.ii.uj.edu.pl] has quit [Read error: 110 (Connection timed out)]
16:26 -!- dsoul [i=darksoul@vice.ii.uj.edu.pl] has joined #t2
17:07 < koan> hm does anyone know how I could test T2 in a virtual machine on Debian/PPC?
17:22 < mpp> okay successfully added rclocal service to pxe-rescue minit
17:22 < mpp> dhcpcd eth0 comes up
17:22 < mpp> no on to loadkeys keymaps
18:04 < mpp> okay did some patching for ntp
18:04 < mpp> ntp needs /var/lib/ntp with an apropriate ntp.conf file
18:05 < mpp> i merged those into my local tree and into pxe-rescue
18:05 < mpp> so far dhcpcd , loadkeys and ntp work fine
18:05 < mpp> just wat i need for a pxe-rescue
18:05 < mpp> :-) t2 - just do it
18:06 < mpp> loadkeys also needed some files when rescue includes kbd package
18:46 < mpp> i kinda like minit
20:17 < mpp> .
20:19 < rxr> :_)
20:19 < mpp> hey rxr
20:19 < mpp> im still hacking the pxe-rescue
20:19 < rxr> you should try the original rescue target form a USB stick
20:20 < rxr> boots into X in some 7s or so
20:20 < rxr> (+ BIOS time ... of course)
20:20 < mpp> yes but i need it for server provisioning
20:20 < mpp> so pxe is the choice here :-)
20:21 < mpp> i recently also digged through puppy linux
20:21 < mpp> iam thinking about doing a parallel fork of the t2-trunk tree and commit back to the original :-)
20:22 < mpp> stuff i do and modify
20:22 < mpp> drivers for raid controllers and tools would be interesting for server farms and stuff
20:22 < mpp> my build server at intergenia is now running on t2-7.0
20:24 < mpp> i think i will have a close look at the distcc stuff , i could set up 3-4 pc's with dual core cpus to help speeding up the built process
20:24 < CIA-5> rene * r30201 /trunk/package/base/mkinitrd/ (initrdinit.sh mkinitrd.sh):
20:24 < CIA-5> * fixed new mkinitrd/init mdadm RAID support for systems without mdadm,
20:24 < CIA-5> by making it optional
20:24 < mpp> for that purpose a initramfs and machine with 4gb ram should suffice
20:24 < mpp> i could mount the t2-trunk via network
20:25 < mpp> and have building cluster with distcc that way
20:25 < mpp> ever run the build process over nfs ?
20:25 < rxr> for distcc you do not even need the mount
20:25 < mpp> that's true
20:25 < rxr> build over nfs is quite slow - we did that in the early ROCK Linux days
20:25 < mpp> as i read the ditscc docu
20:25 < rxr> distcc streams the preprocessed source over the LAN - so the cluster "node" only needs the same compiler and all just works
20:25 < mpp> so any network file system capable or do we stick with local drives
20:26 < mpp> maybe better SAN .-)
20:26 < mpp> :-)
20:26 < mpp> minit was pretty straightforward to understand
20:26 < mpp> i kinda like it
20:28 < mpp> im still struggling with minit and fgetty to bypass the login prompt
20:28 < mpp> this kinda sucks atm
20:28 < mpp> but with params set acording to the earlier mentioned - still the prompt :-(
20:31 < CIA-5> rene * r30202 /trunk/ (5 files in 5 dirs):
20:31 < CIA-5> * fixed (hopefully all) early user-space init scripts to define the
20:31 < CIA-5> /dev tmpfs mode to 755 - SECURITY
20:47 < CIA-5> rene * r30203 /trunk/scripts/Download:
20:47 < CIA-5> * fixed scripts/Download -all to also download all target overlay
20:47 < CIA-5> files, by reusing the existing list_dtags function (also 14 lines
20:47 < CIA-5> less code!)
20:52 < mpp> i diged throug fgetty
20:52 < mpp> looks like login always prompts
21:06 < mpp> hey rxr
21:06 < mpp> i read in the fgetty patch you made that the login is skipped if the user is marked with a x in shadow
21:07 < rxr> you do not need to use fgetty
21:07 < rxr> use the normal one or directly spin shells
21:08 < mpp> ahm i am on the issue with my pxe-rescue target
21:08 < mpp> it always prompts as minit uses fgetty
21:08 < mpp> which you have patched
21:08 < rxr> minit does not hard codedly use fgetty
21:08 < mpp> well fgetty is fine
21:09 < rxr> you can run anything, it's just I used it in the default config
21:09 < mpp> i just can't make sence of your patch
21:09 < rxr> you can change the etc/minit/fgetty or so to anything
21:09 < rxr> the patch is for passworless logins
21:09 < mpp> which means ?
21:09 < mpp> i just want to skip the login
21:10 < rxr> by default fgetty has the "bug" that it does not login if the user has a blank password
21:10 < mpp> nothing else
21:10 < mpp> ah.
21:10 < rxr> the normal util-linux agetty one does allow no passwd
21:10 < mpp> okay..
21:53 -!- mp1 [n=user@i5387690D.versanet.de] has joined #t2
21:55 -!- mpp [n=user@i53875BCA.versanet.de] has quit [Read error: 60 (Operation timed out)]
22:17 -!- mp1 [n=user@i5387690D.versanet.de] has left #t2 []
22:18 -!- mp1 [n=user@i5387690D.versanet.de] has joined #t2
22:36 < mp1> hey rxr
22:36 < mp1> oops change nick
22:36 -!- mp1 [n=user@i5387690D.versanet.de] has left #t2 []
22:38 -!- mpp [n=user@i5387690D.versanet.de] has joined #t2
22:38 < mpp> ge whats up with my i-net
22:38 < mpp> anyways
22:39 < mpp> @rxr another hint maybe on the rescue target
22:39 < mpp> fgetty uses login1
22:39 < mpp> is that right .
23:06 -!- mpp [n=user@i5387690D.versanet.de] has left #t2 []
23:10 -!- mpp [n=user@i5387690D.versanet.de] has joined #t2
23:19 -!- mpp [n=user@i5387690D.versanet.de] has quit ["Leaving."]
23:21 -!- mpp [n=user@i5387690D.versanet.de] has joined #t2
--- Log closed Sun Aug 24 00:00:16 2008
01:48 -!- Netsplit calvino.freenode.net <-> irc.freenode.net quits: @ChanServ, felanha
01:49 -!- Netsplit over, joins: @ChanServ, felanha
01:50 -!- Netsplit calvino.freenode.net <-> irc.freenode.net quits: @ChanServ, felanha
01:51 -!- Netsplit over, joins: @ChanServ, felanha
01:58 < CIA-5> ragnar * r30175 /trunk/package/base/mdadm/mdadm.desc: * mdadm: mark as compatible with dietlibc
02:05 < CIA-5> ragnar * r30176 /trunk/package/base/mkinitrd/ (initrdinit.sh mkinitrd.sh): * mkinitrd: make initrd aware of md devices (only raid1 atm)
02:10 < CIA-5> ragnar * r30177 /trunk/package/base/mdadm/preconfig.in: * mdadm: have it build with dietlibc by default
06:14 -!- emte [n=emte@d154-5-45-238.bchsia.telus.net] has joined #T2
07:29 -!- emte [n=emte@d154-5-45-238.bchsia.telus.net] has quit [Read error: 104 (Connection reset by peer)]
09:02 < rxr> moin
09:02 < rxr> Ragnar|away: neat
09:04 < CIA-5> rene * r30178 /branches/7.0/package/ruby/ruby/ruby.conf:
09:04 < CIA-5> * added sparc* to the architectures where ruby segfaults if
09:04 < CIA-5> pthread support is enabled
09:06 < CIA-5> rene * r30179 /trunk/package/ruby/ruby/ruby.conf:
09:06 < CIA-5> * merged r30178 from branches/7.0: added sparc* to the architectures
09:06 < CIA-5> where ruby segfaults if pthread support is enabled
09:17 -!- LMJ [n=serwou@laf31-4-82-236-42-164.fbx.proxad.net] has quit [Connection timed out]
09:18 -!- LMJ [n=serwou@laf31-4-82-236-42-164.fbx.proxad.net] has joined #t2
09:33 < CIA-5> rene * r30180 /trunk/package/editors/ed/ed.desc: * updated ed (0.9 -> 1.0)
09:33 < CIA-5> rene * r30181 /trunk/package/network/neon/neon.desc: * updated neon (0.28.2 -> 0.28.3)
09:36 < CIA-5> rene * r30182 /trunk/package/network/ntp/ntp.desc: * updated ntp (4.2.4p4 -> 4.2.4p5)
09:39 < CIA-5> rene * r30183 /trunk/package/editors/moe/ (. moe.cache moe.desc): * added moe (1.0) - A powerful and user-friendly console text editor
09:39 < CIA-5> rene * r30184 /trunk/package/textproc/libmrss/libmrss.desc: * updated libmrss (0.19.1 -> 0.19.2)
09:39 < CIA-5> rene * r30185 /trunk/package/textproc/libnxml/libnxml.desc: * updated libnxml (0.18.2 -> 0.18.3)
09:40 -!- Baldzius [n=Baldzius@89.127.175.95] has joined #t2
09:40 < Baldzius> moin
09:40 < rxr> hi Baldzius
09:40 [Users #t2]
09:40 [@ChanServ] [ CIA-5 ] [ dsoul ] [ felanha] [ Ragnar|away] [ TobiX]
09:40 [ axionix ] [ Cyda ] [ DuckFault] [ LMJ ] [ rxr ]
09:40 [ Baldzius] [ Dallur] [ Enqlave ] [ mtr ] [ timlarson ]
09:40 -!- Irssi: #t2: Total of 16 nicks [1 ops, 0 halfops, 0 voices, 15 normal]
09:40 < Baldzius> hey rxr
09:40 < rxr> how do you do ?
09:40 < CIA-5> rene * r30186 /trunk/package/network/nss_ldap/nss_ldap.desc: * updated nss_ldap (261 -> 262)
09:41 < Baldzius> thanks, not bad, trying to rest after the night shift of whole week
09:41 < Baldzius> yourself?
09:45 < rxr> oh, night shift - does not sound fun
09:46 < rxr> I'm fine, too :-)
09:46 < rxr> thanks
09:48 < Baldzius> it's not so bad as it sounds but the worst thing is to go back to the normal routine
09:48 < Baldzius> cause now i got used to sleep on days rather nights :)
09:48 < Baldzius> need to switch back so it takes couple of days
09:49 < CIA-5> rene * r30187 /trunk/package/network/madwifi/kernel-change.patch: * svn rm madwifi/kernel-change.patch missed from ML patch in r30165
09:50 < rxr> Baldzius: noticed, we got new, pretty IRC logs: http://www.t2-project.org/irclogs/
09:50 < Baldzius> *click*
09:51 < rxr> and now 127 motivated enough to add themselv to the startpage: http://www.t2-project.org/startpage/
09:52 < Baldzius> nice
09:52 < rxr> and increased ML messaging :-)
09:53 < Baldzius> yeah, noticed that too
09:58 < CIA-5> rene * r30188 /trunk/package/editors/moe/moe.desc: * marked the "new" moe editor to cross build
09:58 < rxr> do you know something serious in the 7.0 branch? or can we go gold hopefully soonish ?
09:59 < Baldzius> no, haven't used 7.0 for a long time now
09:59 < Baldzius> sitting exclusively on trunk
10:00 < rxr> ah - ok as well
10:00 < rxr> despite sparc64 on the 360MHz box the x86, x86-64, ppc, ppc64 builds are sort of ready
10:01 < rxr> (mostly require a final test install ...)
10:01 < Baldzius> oh maybe one thing, try to install to is using ext3 as root
10:01 < Baldzius> the last time it didn;t work for me somehow
10:02 < Baldzius> so i had to stick with reiserfs
10:03 < CIA-5> rene * r30189 /branches/7.0/package/xorg/xorg-server/parse-config: * merged r30174 from trunk: silenced r30153 pkgprefix warning if libx11 was not yet build
10:15 < CIA-5> rene * r30190 /trunk/target/desktop/config.in: * removed "Exact" branding from desktop config entry
10:21 < CIA-5> rene * r30191 /branches/7.0/target/ (desktop/pkgsel generic/pkgsel/10-minimal-xorg.in):
10:21 < CIA-5> * removed freetype1 from generic minimal-xorg package selection
10:21 < CIA-5> template, removed duplicate selection from desktop package
10:21 < CIA-5> selection
10:22 < CIA-5> rene * r30192 /trunk/target/ (desktop/pkgsel generic/pkgsel/10-minimal-xorg.in):
10:22 < CIA-5> * merged r30191 from branches/7.0: removed freetype1 from generic
10:22 < CIA-5> minimal-xorg package selection template, removed duplicate selection
10:22 < CIA-5> from desktop package selection
10:50 -!- hwinkel [n=hwinkel@p54B69573.dip0.t-ipconnect.de] has joined #t2
10:55 < CIA-5> rene * r30193 /branches/7.0/package/security/gnupg2/gnupg2.conf: * work arounded inf. looping gnupg2 tests on powerpc64: skip them for now
10:56 < CIA-5> rene * r30194 /trunk/package/security/gnupg2/gnupg2.conf:
10:56 < CIA-5> * merged r30193 from branches/7.0: work arounded inf. looping gnupg2
10:56 < CIA-5> tests on powerpc64: skip them for now
10:59 -!- mpp [n=user@i53875BCA.versanet.de] has joined #t2
10:59 < mpp> guten tag allerseits
10:59 < rxr> moin moin mpp
11:00 < mpp> i built successfully the modified rescue target for pxe
11:00 < mpp> it works fine !
11:00 < mpp> no i am adding some network tools
11:00 < mpp> does anyone know where the ping resides in ?
11:00 < rxr> you tested network booting already ?
11:01 < mpp> i mean which packege
11:01 < mpp> yep
11:01 < mpp> works like a charm
11:01 < rxr> great
11:01 < mpp> i removed the x stuff from the target
11:01 < mpp> thanx to you
11:01 < mpp> !
11:01 < rxr> in t2 var/adm/* hols the installed package stuff in plain/text ASCII files, so you can find ping like this:
11:01 < rxr> # grep bin/ping /var/adm/flists/*
11:01 < rxr> /var/adm/flists/netkit-base:netkit-base: bin/ping
11:01 < mpp> cool
11:02 < rxr> to find ping related stuff as good as the maintainers filled the T2 .desc files in your t2 tree:
11:02 < mpp> these irc-logs are pretty precious
11:02 < rxr> # ./scripts/Find-Pkg ping
11:02 < Ragnar|away> that usually just gives me a 'line too long' error ;P
11:02 < mpp> nice
11:02 < mpp> i'll give it a shot right now
11:02 < rxr> Ragnar|away: only if you grep in /home/ragnar/src/t2-trunk-foo-bar/build/my-super-target-8.0-x86-64-core2-cross/var/adm/flists/* :-)
11:03 < rxr> Ragnar|away: recent kernels (after 2.6.23 or so) have dynamic command line allocation, now, FINALLY
11:03 < Ragnar|away> oh kewl :)
11:03 < rxr> so you should see this less often even in the above case
11:03 < rxr> though you can limit it with ulimits
11:03 < rxr> IIRC it defaults to 1/4 or so of the stack limit
11:04 < rxr> someone from google jumped over the quick brown fox and implemented this because it annoyed him too much :-)
11:04 < rxr> he came from plan9 ...
11:05 < Ragnar|away> you're right, it works now :) sweeeety
11:05 < mpp> by the way
11:05 * Ragnar|away dances
11:05 < mpp> the madwifi package has almost all the time trouble getting downloaded
11:05 * rxr wants to see a youtube video of that :-)
11:05 < mpp> so i had to exclude this
11:05 < Ragnar|away> LOL
11:05 < rxr> mpp: I guess that's the SSL cert of the server
11:06 < rxr> mpp: I hope the madwifi checkout is on the t2 mirrors by now so you should be able to crab it from somewhere
11:06 < rxr> maybe svn has some "--eat-the-cert" option to avoid this?
11:07 < Ragnar|away> madwifi-r3375.tar.bz2 ?
11:07 < rxr> the new one from yesterday, whatever the -r was .-)
11:08 < Ragnar|away> ooh ok
11:11 < Ragnar|away> I see what you mean
11:14 < rxr> mpp: btw. this flist program and other files search is on my TODO to integrate into the homepage package matrix
11:14 < Ragnar|away> hmm can we code the answer to the 'accept cert' into the Download ?
11:15 < rxr> so one can consult that to find where some obscure binary or header is located in
11:15 < Ragnar|away> oh btw when do you usually update the .cache files?
11:16 < rxr> Ragnar|away: no idea yet, mache echo -e "y\n" | would work - maybe not, svn has some --non-interactive switch, I just tried, it does not accept the cert then - maybe it has an hidden switch for that
11:16 < rxr> Ragnar|away: we should automate the .cache update more, indeed
11:16 < Ragnar|away> mpp: you can do the checkout once by hand, accept the cert permanently, then run the Download
11:16 < rxr> Ragnar|away: right now I update them when something majorly changes
11:16 < rxr> yes, I thought I had done so on the master server after I noticed this on my regression tester
11:19 < CIA-5> rene * r30195 /trunk/package/lua/luaffi/ (. lua-lib.patch luaffi.cache luaffi.conf luaffi.desc): * added luaffi (0.9.0) - FFI bindings to Lua
11:23 -!- Baldzius [n=Baldzius@89.127.175.95] has quit [Remote closed the connection]
11:25 < mpp> hey rxr
11:25 < mpp> the udhcpc should work in the rescue image ß
11:25 < mpp> ß
11:25 < mpp> ß
11:25 < mpp> damnit
11:25 < rxr> yeah the client scripts ?
11:25 < mpp> right
11:25 < rxr> it already on my todo ... :-(((
11:25 < rxr> patch welcome
11:25 < mpp> they give an error when booted
11:26 < mpp> a cool
11:26 < mpp> so ill dig into it
11:26 < rxr> yeah - not so cool that it does not yet just work ...
11:26 < rxr> yeah - patch most welcome
11:26 < rxr> maybe in conjunction with the queued bind one :-)
11:26 < mpp> im on it
11:26 < mpp> i have that already filed
11:26 < mpp> ill submit
11:29 -!- mpp [n=user@i53875BCA.versanet.de] has left #t2 []
11:32 -!- mpp [n=user@i53875BCA.versanet.de] has joined #t2
11:35 < mpp> ?
11:35 < mpp> ah
11:35 < mpp> keyboard working again
11:35 < mpp> nice :-)
11:37 < mpp> allright i user dhcpcd
11:37 < mpp> it works
11:37 < mpp> only thing is
11:38 < mpp> the directory /var/lib and /var/lib/dhcpcd were missing in the initramfs
11:38 < mpp> how to i patch the dhcpcd to include these directories ?
11:38 < rxr> are they empty or contain files ?
11:38 < mpp> they don't exist on the booted image
11:38 < mpp> so i created them manually
11:38 < rxr> and in the raw build ?
11:38 < mpp> and it worked
11:38 < rxr> build/...
11:38 < mpp> raw build ?
11:39 < mpp> you mean in the build/TOOLCHAIN ?
11:39 < rxr> the build before filtering and image compression
11:39 < rxr> I mean build/$configid-$uuid/
11:39 < mpp> ill check
11:39 < rxr> ok - my rescue build looks like it does not exist
11:39 < mpp> missing there
11:40 < rxr> if that's the case for you a hook_add postmake 5 "mkdir -p $root/var/lib/dhcpcd" should do
11:40 < rxr> --typos, etc.
11:40 < rxr> rebuild the package and rebuild the image
11:41 < mpp> where do i do this modification ?
11:41 < mpp> hook_add ?
11:42 < rxr> sorry in the .conf file
11:42 < mpp> i'll think ill read the dev docu first to catch up with this
11:42 < mpp> okay
11:42 < rxr> it's a t2 function to add code to some hook
11:42 < rxr> each package con influence it's build with conde in package/*/$pkg/$pkg.conf
11:42 < mpp> right
11:43 < mpp> but in this case we could also create them in the rootfs of the target definition
11:43 < mpp> which is bad as i see
11:43 < rxr> but then all the other targets would not be fixed
11:43 < mpp> vor the package is beter
11:43 < mpp> yep got it
11:43 < mpp> t2 rocks ;-)
11:43 < rxr> if the package needs some directory to work it should be in the package, not just a single target
11:43 < mpp> right
11:44 < mpp> looks like /var/lib is missing in this target in general
11:44 < mpp> which is not a good idea
11:45 < rxr> it probably just that no package is in the selection that created it
11:45 < mpp> ah i see
11:46 < mpp> so were do we patch for this
11:46 < rxr> with the hook above :-)
11:46 < mpp> i mean dhcpcd is very cool and udhcp gave lots of erros of diverse files and paths
11:46 < mpp> allright
11:46 < rxr> ah - you exchanged udhcpd with dhcp in your target ?
11:47 < mpp> no i kicked them both and used dhcpcd only
11:47 < mpp> works like a charm
11:47 < mpp> im rebuilding the target with the paths added
11:48 < mpp> wouldn it make sence to have a file hierarchy template used
11:48 < mpp> even if some paths were empty for certain targets
11:48 < rxr> not really
11:48 < mpp> okay - explain please
11:48 < rxr> t2 is about cleanlyness and sorted stuff
11:49 < mpp> capiche !
11:49 < mpp> accepted
11:49 < rxr> if you have 100 extra dirs, empty then they are not only confusing and a mess, but also waste space on very minimal things
11:49 < rxr> for example we build < 2MB systems etc. ...
11:49 < rxr> you do not want to have 100 directories in them just in case ...
11:50 < rxr> if some package does rely on a not yet existing directory, then this single package needs to be fixed
11:50 < rxr> in this case it looks like a) this dhcpcd needs to be fixed to create this var/lib/dhcpd directory and the udhcpd needs to be fixed to find it's freaking config scripts
11:51 < rxr> the nearly only other thing would coudl do additionaly is to add /var/lib to the generic "FHS" 00-dirtree layout, but not sub-folders within it
11:52 < mpp> submited the patch via mail
11:52 < mpp> and other to follow for sure
11:52 < mpp> now that i know how and what to do
11:52 < mpp> so ill be tuning the pxe-rescue target since i need it professionally
11:53 < mpp> i called it pxe-rescue
11:53 < mpp> if its okay whith you
11:53 < rxr> sounds good, without the tinyX stuff it will probably be way smaller
11:53 < mpp> i kicked that out already
11:53 < mpp> was easy in the target definition
11:54 < mpp> even for a noob like me
11:58 < mpp> initrd is about 63mb for now
11:58 < mpp> i need ruby in the rescue image so ill add that to the target
12:10 -!- hwinkel [n=hwinkel@p54B69573.dip0.t-ipconnect.de] has quit [Client Quit]
12:13 -!- koan [n=koan@pec.ulyssis.student.kuleuven.be] has joined #t2
12:18 -!- koan [n=koan@unaffiliated/koan] has quit [Client Quit]
12:18 -!- koan [n=koan@unaffiliated/koan] has joined #t2
12:19 < CIA-5> rene * r30196 /branches/7.0/package/shells/mc-mp/umode_t.patch: * fixed mc-mp to build on powerpc64
12:22 < mpp> @rxr i did create a dhcpcd.conf inside the package/network/dhcpcd dir and added two lines for /var/lib and /var/lib/dhcpcd
12:22 < mpp> i rebuilt but the dirs did not get added to the initrd.img
12:22 < rxr> just a single mkdir -p ... inside the hook_add is enough
12:22 < rxr> what did you put into the dhcpd.conf ?
12:23 < CIA-5> rene * r30197 /trunk/package/shells/mc-mp/umode_t.patch: * merged r30196 from branches/7.0: fixed mc-mp to build on powerpc64
12:24 < mpp> these hook_add postmake 5 'mkdir -p $root/var/lib'
12:24 < mpp> hook_add postmake 5 'mkdir -p $root/var/lib/dhcpcd'
12:25 < rxr> you did rebuild the package ?
12:25 < mpp> yes
12:25 < mpp> and there should be a var/lib inside the build/...
12:25 < mpp> right ?
12:25 < rxr> yes, is there ?
12:26 < mpp> no it's not
12:26 < mpp> do i have to make a cleanup
12:28 < rxr> not really
12:28 < rxr> how did you rebuild the package ?
12:28 < mpp> scripts/Config -cfg pxe-rescue
12:28 < mpp> scripts/Download -cfg pxe-rescue -required
12:28 < mpp> scripts/Build-Target -cfg pxe-rescue
12:29 < rxr> ok - that did not rebuild the package
12:29 < rxr> you neither had to run Config nor Download
12:29 < mpp> okay..
12:29 < rxr> (btw. Download is optional today in any case, anyway - T2 auto-downloads when missing ... )
12:29 < rxr> for a package rebuild you can use two main ways (beside other manual hackery)
12:29 < mpp> with the scripts/Build-Target -job 5-dhcpcd then ?
12:29 < rxr> the first, explicitly one package:
12:29 < rxr> yes, like you said:
12:30 < rxr> scripts/Build-Target -job 5-$pkg
12:30 < rxr> or, to rebuild what changed:
12:30 < rxr> scripts/Create-ErrList -cfg rescue -newremove
12:30 < rxr> removed logfiles of package where anything changed
12:30 < rxr> for me this was now:
12:30 < rxr> Scheduling package 'mc-mp' for rebuild ...
12:30 < rxr> Scheduling package 'mkinitrd' for rebuild ...
12:30 < rxr> Scheduling package 'madwifi' for rebuild ...
12:31 < rxr> and then just re-run Build-Target, it will build the stuff Create-ErrList ... removed the log-files from
12:31 < rxr> btw. your hook works, I tested it on my side right now ...
12:31 < mpp> ah now rebuilding the package
12:31 < mpp> thanx !
12:31 < mpp> t2 is fun..
12:32 < mpp> i also want to remove the login
12:32 < mpp> in the inittab i think
12:32 < mpp> since it is rescue
12:32 < mpp> for my X-removed version here
12:37 < mpp> how can i add some post-processing to my target before the initrd.img is build
12:37 < mpp> ?
12:40 < rxr> either by patching packages
12:40 < rxr> or if you want to add tweak files via target/.../rootfs
12:40 < mpp> thats the one
12:40 < rxr> you probably already copied the rootfs "overlay" from the rescue target
12:40 < mpp> i did
12:40 < rxr> you can just add files, or even patch them
12:40 < mpp> so i just place a patched inittab there and rebuild ?
12:43 < rxr> yes, if you place a file there it will just be copied
12:43 < rxr> it's a little more elegant to patch a file
12:44 < rxr> you can place a patch file there starting with #!patch to indicate it's a patch
12:44 < rxr> but maybe start with replacing whole files
12:44 < mpp> can i get rid of these error logs on my build host :
12:44 < mpp> build/pxe-rescue-8.0-trunk-pxe-rescue-x86-i486/var/adm/cache/udhcp should not be present (now in src/invalid-files.lst)!
12:45 < mpp> since these packages are'nt needed anymore
12:46 < rxr> rm `cat src/invalid-files.lst `
12:58 < CIA-5> rene * r30198 /trunk/package/network/dhcpcd/dhcpcd.conf:
12:58 < CIA-5> Martin Papadopoulos
12:58 < CIA-5> * fixed dhcpcd to create required var/lib/dhcpcd directory
13:06 < CIA-5> rene * r30199 /trunk/package/multimedia/dvbd/dvbd.init:
13:06 < CIA-5> * fixed dvbd.init to not yield errors if the dvbd socket directory
13:06 < CIA-5> already exists
13:10 < mpp> hey rxr
13:10 < mpp> can i submit my pxe-rescue tree to the t2-trunk ?
13:11 < rxr> sure if you like to
13:12 < mpp> how do i do that ?
13:12 < rxr> maybe create a tar and send it over
13:12 < mpp> okay thats cool
13:12 < mpp> send it via ? maillist :-)
13:12 < rxr> sure
13:22 < mpp> how do i deactivate the login ?
13:22 < mpp> in the inittab i think but how exactly
13:27 < rxr> replate the agetty lines with something that does not ask a password
13:27 < rxr> agetty -i 38400 vc/1 -n -l /sbin/login-shell
13:27 < rxr> ..
13:27 < rxr> might do
13:42 < mpp> does not :-)
13:43 < rxr> what does it do
13:43 < rxr> ?
13:43 < mpp> still login prompt but i gues is screwed the line in the inittab
13:43 < mpp> ill retry
14:07 -!- koan_ [n=koan@unaffiliated/koan] has joined #t2
14:07 -!- koan_ [n=koan@unaffiliated/koan] has quit [Client Quit]
14:20 < mpp> okay i tried but still the login prompt appears
14:29 < rxr> hm - then I would have to read man-pages and try as well ...
14:33 < mpp> okay never mind for now - ill dig around also
15:37 < mpp> @rxr: does the rescue target execute rcX.d/script ?
15:37 < mpp> i got the impression that id doesn't
15:42 < CIA-5> rene * r30200 /trunk/package/lua/luacurl/ (. luacurl.cache luacurl.conf luacurl.desc): * added luacurl (1.0) - A Lua binding providing Internet capabilities based on CURL
15:45 < rxr> ah - heh :-)(
15:45 < rxr> the rescue target uses the new, and tiny and fast init called minit
15:45 < rxr> it's based on something entirely different
15:45 < rxr> ah !"
15:45 < rxr> that's also why your inittab stuff has no effect :-)!!!
15:46 < rxr> the minit config is in /etc/minit
15:46 < rxr> http://www.t2-project.org/packages/minit.html
15:48 < mpp> makes sence !
15:48 < rxr> I added "cleaning the sysfiles package to not provide the lingering inittab file but instead move it into the sysvinit package to my TODO)
15:53 < mpp> so what parmams should i pass to the fgetty besides vc/1 to avoid the prompt ?
15:53 < rxr> -l /sbin/login-shell
15:53 < rxr> ah - second
15:53 < rxr> not existing in my system
15:54 < rxr> maybe /bin/bash --login would do
15:54 < rxr> -l /bin/bash --login or so ...
15:54 < mpp> okay ill dig throug minit docu
15:54 < rxr> maybe with quotes
15:54 < rxr> it's not minit
15:54 < rxr> it's about the "getty" ...
15:54 < rxr> maybe try in a running system with vc/7 or vc/8, etc ...
15:55 < rxr> no need to rebuild image after image to give it a try
15:55 < rxr> just start the line as root ...
15:56 -!- dsoul [i=darksoul@vice.ii.uj.edu.pl] has quit [Read error: 110 (Connection timed out)]
16:26 -!- dsoul [i=darksoul@vice.ii.uj.edu.pl] has joined #t2
17:07 < koan> hm does anyone know how I could test T2 in a virtual machine on Debian/PPC?
17:22 < mpp> okay successfully added rclocal service to pxe-rescue minit
17:22 < mpp> dhcpcd eth0 comes up
17:22 < mpp> no on to loadkeys keymaps
18:04 < mpp> okay did some patching for ntp
18:04 < mpp> ntp needs /var/lib/ntp with an apropriate ntp.conf file
18:05 < mpp> i merged those into my local tree and into pxe-rescue
18:05 < mpp> so far dhcpcd , loadkeys and ntp work fine
18:05 < mpp> just wat i need for a pxe-rescue
18:05 < mpp> :-) t2 - just do it
18:06 < mpp> loadkeys also needed some files when rescue includes kbd package
18:46 < mpp> i kinda like minit
20:17 < mpp> .
20:19 < rxr> :_)
20:19 < mpp> hey rxr
20:19 < mpp> im still hacking the pxe-rescue
20:19 < rxr> you should try the original rescue target form a USB stick
20:20 < rxr> boots into X in some 7s or so
20:20 < rxr> (+ BIOS time ... of course)
20:20 < mpp> yes but i need it for server provisioning
20:20 < mpp> so pxe is the choice here :-)
20:21 < mpp> i recently also digged through puppy linux
20:21 < mpp> iam thinking about doing a parallel fork of the t2-trunk tree and commit back to the original :-)
20:22 < mpp> stuff i do and modify
20:22 < mpp> drivers for raid controllers and tools would be interesting for server farms and stuff
20:22 < mpp> my build server at intergenia is now running on t2-7.0
20:24 < mpp> i think i will have a close look at the distcc stuff , i could set up 3-4 pc's with dual core cpus to help speeding up the built process
20:24 < CIA-5> rene * r30201 /trunk/package/base/mkinitrd/ (initrdinit.sh mkinitrd.sh):
20:24 < CIA-5> * fixed new mkinitrd/init mdadm RAID support for systems without mdadm,
20:24 < CIA-5> by making it optional
20:24 < mpp> for that purpose a initramfs and machine with 4gb ram should suffice
20:24 < mpp> i could mount the t2-trunk via network
20:25 < mpp> and have building cluster with distcc that way
20:25 < mpp> ever run the build process over nfs ?
20:25 < rxr> for distcc you do not even need the mount
20:25 < mpp> that's true
20:25 < rxr> build over nfs is quite slow - we did that in the early ROCK Linux days
20:25 < mpp> as i read the ditscc docu
20:25 < rxr> distcc streams the preprocessed source over the LAN - so the cluster "node" only needs the same compiler and all just works
20:25 < mpp> so any network file system capable or do we stick with local drives
20:26 < mpp> maybe better SAN .-)
20:26 < mpp> :-)
20:26 < mpp> minit was pretty straightforward to understand
20:26 < mpp> i kinda like it
20:28 < mpp> im still struggling with minit and fgetty to bypass the login prompt
20:28 < mpp> this kinda sucks atm
20:28 < mpp> but with params set acording to the earlier mentioned - still the prompt :-(
20:31 < CIA-5> rene * r30202 /trunk/ (5 files in 5 dirs):
20:31 < CIA-5> * fixed (hopefully all) early user-space init scripts to define the
20:31 < CIA-5> /dev tmpfs mode to 755 - SECURITY
20:47 < CIA-5> rene * r30203 /trunk/scripts/Download:
20:47 < CIA-5> * fixed scripts/Download -all to also download all target overlay
20:47 < CIA-5> files, by reusing the existing list_dtags function (also 14 lines
20:47 < CIA-5> less code!)
20:52 < mpp> i diged throug fgetty
20:52 < mpp> looks like login always prompts
21:06 < mpp> hey rxr
21:06 < mpp> i read in the fgetty patch you made that the login is skipped if the user is marked with a x in shadow
21:07 < rxr> you do not need to use fgetty
21:07 < rxr> use the normal one or directly spin shells
21:08 < mpp> ahm i am on the issue with my pxe-rescue target
21:08 < mpp> it always prompts as minit uses fgetty
21:08 < mpp> which you have patched
21:08 < rxr> minit does not hard codedly use fgetty
21:08 < mpp> well fgetty is fine
21:09 < rxr> you can run anything, it's just I used it in the default config
21:09 < mpp> i just can't make sence of your patch
21:09 < rxr> you can change the etc/minit/fgetty or so to anything
21:09 < rxr> the patch is for passworless logins
21:09 < mpp> which means ?
21:09 < mpp> i just want to skip the login
21:10 < rxr> by default fgetty has the "bug" that it does not login if the user has a blank password
21:10 < mpp> nothing else
21:10 < mpp> ah.
21:10 < rxr> the normal util-linux agetty one does allow no passwd
21:10 < mpp> okay..
21:53 -!- mp1 [n=user@i5387690D.versanet.de] has joined #t2
21:55 -!- mpp [n=user@i53875BCA.versanet.de] has quit [Read error: 60 (Operation timed out)]
22:17 -!- mp1 [n=user@i5387690D.versanet.de] has left #t2 []
22:18 -!- mp1 [n=user@i5387690D.versanet.de] has joined #t2
22:36 < mp1> hey rxr
22:36 < mp1> oops change nick
22:36 -!- mp1 [n=user@i5387690D.versanet.de] has left #t2 []
22:38 -!- mpp [n=user@i5387690D.versanet.de] has joined #t2
22:38 < mpp> ge whats up with my i-net
22:38 < mpp> anyways
22:39 < mpp> @rxr another hint maybe on the rescue target
22:39 < mpp> fgetty uses login1
22:39 < mpp> is that right .
23:06 -!- mpp [n=user@i5387690D.versanet.de] has left #t2 []
23:10 -!- mpp [n=user@i5387690D.versanet.de] has joined #t2
23:19 -!- mpp [n=user@i5387690D.versanet.de] has quit ["Leaving."]
23:21 -!- mpp [n=user@i5387690D.versanet.de] has joined #t2
--- Log closed Sun Aug 24 00:00:16 2008