T2 IRC Log: 2008-09-18

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.

« prev | next »

--- Log opened Thu Sep 18 00:00:42 2008
00:03 < CIA-8> mtr * r30474 /trunk/package/x11/blackbox/blackbox.conf:
00:03 < CIA-8> * fixed blackbox (cross) build for *64, guide libtool to pickup the
00:03 < CIA-8> target rootfs libraries
00:08 < CIA-8> mtr * r30475 /trunk/package/filesystem/aufs/compile.patch.cross:
00:08 < CIA-8> * fixed aufs cross compilation, build host tools with HOSTCC
00:08 < CIA-8> and set CFLAGS
01:05 -!- mpp_ [n=user@i538742DF.versanet.de] has joined #t2
01:08 -!- mpp [n=user@i53876282.versanet.de] has quit [Read error: 110 (Connection timed out)]
02:31 -!- mpp_ is now known as mpp
02:44 -!- mpp [n=user@i538742DF.versanet.de] has quit ["good night - good fight"]
02:54 -!- n6pfk [n=urk@c-76-104-40-104.hsd1.va.comcast.net] has joined #t2
02:59 -!- n6pfk [n=urk@c-76-104-40-104.hsd1.va.comcast.net] has quit [Remote closed the connection]
04:33 -!- Morglet [n=morgaine@cpc1-hem13-0-0-cust413.lutn.cable.ntl.com] has joined #t2
04:33 < Morglet> Just discovered T2, reading manual ;-)
04:39 -!- hikenboot [n=vmuser@c-75-67-11-49.hsd1.ma.comcast.net] has joined #t2
04:39 < hikenboot> hello all---can someone tell m e what the license is for T2 is it gnu or other?
04:41 < Morglet> I'm a newbie here, just discovered T2 and reading the T2 docs as wel speak, but it clearly says GPL.
04:41 < Morglet> http://www.t2-project.org/faq/
04:41 < hikenboot> how well is xen virtualization and uclibc supporteD?
04:42 < Morglet> No idea, haven't read about that yet. But I assume that T2's notion of targets allows virt guests as targets too. Just guessing though.
04:44 < Morglet> The manual says it supports compilation with alternative system libs, specifically for embedded and other cross, so I assume that covers your uclibc question.
04:44 < hikenboot> do you know how minimal with tinyx and the like i might be able to make somthing based on glibc?
04:46 < hikenboot> looks like there xen tools are based on glibc only
04:47 < Morglet> Don't know, haven't read about that yet.
04:47 < hikenboot> I am looking for people interested in a desktop project related to xen know anyone who might be interested?
04:49 < hikenboot> I think I am going to use T2 as the base
04:49 < Morglet> I thought Xen (nor any other virt system currently) doesn't support 3D graphics in guests, and hence no desktop.
04:50 < Morglet> You *can* have a no-3D desktop of course, I have several. But it's not particularly useful as a main desktop.
04:54 < hikenboot> what about exporting the 3d cards to domU?
04:56 < hikenboot> http://www.xen.org/files/xensummit_4/vmgl_Cavilla.pdf
04:57 < Morglet> Yeah, it would work in domU, but kind of defeats the object of virting, which is to keep the hypervisor's environment tiny and safe/stable and do everything dodgy in a guest.
04:58 < Morglet> I mean dom0, hehe
05:03 < Morglet> A few months ago I remember installing an nVidia driver in a Linux and the installer saying "Xen kernel detected, aborting".
05:03 < hikenboot> http://www.cs.toronto.edu/~andreslc/xen-gl/
05:04 < Morglet> Yeah, was reading that page afte Googling VMGL
05:04 < hikenboot> how small do you think i could make something similar to deli linux for x86_64 based on T2
05:05 < hikenboot> using gclibc
05:05 < hikenboot> the problems come in when trying uclibc
05:05 < Morglet> No idea, I have no experience of T2, just discovered it today and reading docs. Sounds good so far, maybe even better than Gentoo, so I'm likely to give it a go.
05:07 < hikenboot> no anybody interested in getting in on a project like I was saying looking for some dedicated people
05:09 < Morglet> There's been no other people talking on channel in the hour I've been here, so don't conclude "no interest" :-)
05:09 < hikenboot> there minimal x86_84 is pritty darn big is that just their build environment or a live cd OS do you know t2 8.0 trunk
05:11 < hikenboot> there minimal live desktop is http://dl.t2-project.org/binary/stable/6.0.3/minimal-desktop-live-x86_64 and its 369 mb
05:14 < hikenboot> only 4 meg for xfce desktop thats not bad
05:14 < hikenboot> I wonder what is taking all their space in the 369 megs
05:19 < Morglet> Use minix if you're making a tiny appliance. :-) Apparently T2 are working on minix kernel support.
05:20 < Morglet> http://www.minix3.org/doc/screenies.html
05:20 < Morglet> :-)
06:09 < CIA-8> aldas * r30489 /trunk/package/security/openssl/openssl.desc: * updated openssl (0.9.8h -> 0.9.8i)
06:09 < CIA-8> aldas * r30490 /trunk/package/x11/fluxbox/fluxbox.desc: * updated fluxbox (1.1.0.1 -> 1.1.1)
07:05 -!- mpp [n=user@i538742DF.versanet.de] has joined #t2
08:44 < rxr> re
08:47 < rxr> hikenboot: linux, glibc, gcc, X.org and the usual user-space stuff takes the remaining 3xx MB ...
08:47 < rxr> hikenboot: but you can have "tinyX" + firefox and even mysql in 50MB ...
08:47 < rxr> been there, done that ...
09:02 < mtr> moin
09:02 < rxr> hi mtr
09:02 [Users #t2]
09:02 [@ChanServ] [ Dallur ] [ hikenboot] [ Morglet ] [ mtr ] [ tfl0pz]
09:02 [ axionix ] [ dsoul ] [ koan ] [ mpp ] [ Ragnarin] [ TobiX ]
09:02 [ CIA-8 ] [ Enqlave] [ LMJ ] [ mqueiros_] [ rxr ]
09:02 -!- Irssi: #t2: Total of 17 nicks [1 ops, 0 halfops, 0 voices, 16 normal]
09:04 < mtr> rxr: udev-128 builds and runs on the livecd target: 'udevtrigger' has to be replaced with 'udevadm trigger'
09:04 < mtr> I replaced this on all the target/share/... init scripts
09:04 < rxr> ok - so my first glace and blog entry is right
09:04 < rxr> well - thing is you have to update udev and initrd in lockstep or you have a broken system
09:05 < rxr> maybe we custom patch udev and add this compat stuff back?
09:05 < mtr> yes, and the old udev-120 had the compatibility symlinks to the old tools
09:05 < rxr> I hate, especially wasting my own time, with accidently broken systems
09:05 < mtr> ack
09:06 -!- mqueiros_ [n=mqueiros@217.70.75.226] has quit [Client Quit]
09:07 < mpp> moin
09:07 < rxr> maybe it's just the synmlinks to custom create
09:07 < rxr> though maybe it's adding back some arg[0] conditionals
09:07 * mtr wondering if the old udev symlinks to udevadm would still work
09:08 < mtr> moin moin mpp
09:09 < mpp> moin mtr moin rxr
09:10 * rxr wonders too
09:10 < rxr> have fun testing
09:10 < rxr> does anyone know if there is some extractor for Microsoft .msi installer files for extraction (unpacking) under Linux ?
09:12 < mtr> wine ?
09:12 < rxr> ah - cool
09:13 < rxr> cabextract works
09:13 < rxr> good random try :-)
09:13 < rxr> http://www.t2-project.org/packages/cabextract.html
09:14 < Morglet> Woohoo, my first t2 iso downloaded, now to burn it :-)
09:19 < mtr> rxr: for udev-128 it would be sufficient to add the missing symlinks for udevtrigger and udevsettle, they have the backwards-compatibility built-in
09:19 < mtr> the program '/bin/bash' called 'udevtrigger', it should use 'udevadm trigger ', this will stop working in a future release
09:20 < rxr> let's slicense the warning and react when the patch does no longer apply due to this compat removed
09:20 < rxr> even than we can just add the compat name matching back
09:21 < rxr> actually this ever changing linux madness really gets on my nuts ...
09:21 < rxr> and is one of the biggest showstoppers in further linux adaption ...
09:21 < rxr> because with this random changes linux is as fragile as other major operating systems
09:22 < rxr> just not a registry or dll hell, but ever changing kernel and user-space low-level "hell"
09:22 < koan> and why do they do this?
09:22 < rxr> too much time, nothing better to do, no idea ...
09:22 < koan> :-)
09:22 < rxr> they will probably argue it's cleaner this way
09:23 < rxr> but especially udev was already chagned 4 times ++ in incompatible ways to "be cleaner this way"
09:23 < rxr> and for something required to boot and access any device that is 3 times ++ too often for my taste
09:30 < Morglet> They "do this" because they love to tinker. The lack of stability is annoying yeah, but then, we all like to tinker too ;-)
09:31 < Morglet> The general answer to most such problems is "decouple". In fact that's a good answer to a lot of things :-)
09:33 < Morglet> It's perma hell over in Gentoo with things changing under us ... still works though.
09:38 < rxr> yeah - decoupling is an idea that grows in me for some time now already :_)
09:42 < koan> I've been thinking about it too, for example wouldn't it be wonderful if you could just plug in a minix or bsd kernel instead of linux?
09:44 < Morglet> That would be nice. I'd hate to write the abstraction layer for that though ;-)
09:45 < rxr> yes, patches welcome!
09:45 < rxr> well - with t2 at least you would not need an abstraction layer you could build the dedicated environment
09:46 < rxr> if you want to exchange kernel without rebuilding the user-space you need some qemu userspace emulation a-like ABI translator
09:46 < Morglet> I'm about to discover the wonders of t2 for myself ... as soon as I find a burner that flaming works.
09:48 < rxr> (or run it in some VM)
09:48 < Morglet> Oh, I think someone asked about virt earlier, what the situation was with t2 and Xen or KVM. I've not seen it mentioned in the manual, but not read very far yet.
09:48 < rxr> xen is not so well supported, but could if someone sends patches
09:49 < rxr> mostly because xen was so mainline merge lazy and became less attractive due open sourcing of kqemu and then the appearance of kvm ...
09:49 < rxr> but patches for any topic very much welcome
09:49 < Morglet> And KVM doesn't really need support I guess, since every Linux is a KVM environment if compiled in.
09:49 < rxr> yeah
09:50 < rxr> of course xen needs equally zero support for the unpriviledges guests, and we have some xen optimization setting
09:50 < Morglet> Yeah, don't know what XenSource were playing at for years, went catatonic. I think KVM made them wake up a bit.
09:50 < rxr> however, automatically building the para-virtualization requires getting those patches somehow added to t2
10:01 < Morglet> At last, found a burner that works, on a P3 box ... glad I didn't chuck it out.
10:03 < CIA-8> rene * r30491 /trunk/package/security/strongswan/strongswan.desc: * updated strongswan (4.2.6 -> 4.2.7)
10:18 < tfl0pz> to build 32bit target on a 64bit system should i check 'this is a cross-build between architectures'?
10:19 < rxr> you can, but you do not need to if your 64bit kernel can run 32bit code
10:20 < rxr> usually that is the case unless you explicitly re-build your kernel from source disabling COMPAT_32BIT or what it was named ...
10:20 < tfl0pz> it's the standard kernel from the 8.0-trunk iso i've installed
10:20 < tfl0pz> okay
10:20 < tfl0pz> :)
10:20 < rxr> :-)
10:36 < CIA-8> rene * r30492 /trunk/package/develop/parrot/parrot.desc: * updated parrot (0.6.4 -> 0.7.1)
10:38 < CIA-8> rene * r30493 /trunk/package/x11/x11vnc/x11vnc.desc: * updated x11vnc (0.9.3 -> 0.9.4)
10:40 < CIA-8> rene * r30494 /trunk/package/multimedia/schroedinger/schroedinger.desc: * updated schroedinger (1.0.0 -> 1.0.5)
10:41 < mtr> I patched udev-128 to show an additional note, like:
10:41 < mtr> ... this will stop working in a future release (Note: T2-SDE generated backwards compatibility symlinks)
10:42 < mtr> now I'll recheck if the rescue livecd still works ...
10:42 < rxr> mtr: one gets many of those messages?
10:42 < rxr> actually I think I would patch it away completely ...
10:57 < mtr> usually udevtrigger is run only once, so only one message to remind of the annoying udev api changes
11:00 < Morglet> I'm installing my first t2 on a blank /dev/sdb in a system that has stuff on sda and sdc HD drives already ... do you reckon it's pretty foolproof, or should I be worried and disconnect those other two first? :-)
11:01 < rxr> t2 text install is not exactly foolproof
11:01 < rxr> maybe disconnect the HDs just to be sure
11:01 < Morglet> k, tnx
11:01 * Morglet heads back to the computer room
11:33 < CIA-8> rene * r30495 /trunk/package/graphic/cuneiform/ (a-html.patch hocr-output.patch html-hocr.patch): * imported conditional hOCR output patch for cuneiform
12:29 < Morglet> LOL @ the install date format, MM-DD hhmm YYYY .... novel ;P
12:32 < rxr> hm?
12:32 < mpp> is the gcc in any way built optimized for the target architecture / cpu ?
12:32 < mpp> with something like -mcpu=....
12:32 < rxr> anything being built by t2 should be build with the selection optimizations
12:33 < mpp> i mean the compiler itself is optimized ?
12:33 < rxr> it should
12:33 < rxr> do you have an indicator that it isn't ?
12:33 < mpp> gcc -v
12:33 < mpp> :-)
12:33 < rxr> and ?
12:33 < mpp> i thought there should be a switch related to cpu
12:34 < mpp> i might be wrong
12:34 < rxr> in new compilers there is, but as we have the full control about how anything is build we control this with our sandbox wrappers
12:34 < rxr> any file is build with the optimizations as selected in the config
12:35 < mpp> check
12:35 < mpp> thanx rxr
12:35 < mpp> i did some testing between 32bit and 64bit build with t2
12:36 < rxr> this is also the reason why there are random -cpu= ... lines in the T2 .log files
12:36 < rxr> any Makefile has whatever the maintainer thought would be nice
12:36 < mpp> seems like 64bit boosts openssl stuff like 4x up
12:36 < rxr> Gentoo and other try to set CFLAGS which often is ignored in the Makefile anyway and then they have to hunt that and patch any Makefile
12:36 < mpp> yeah i read the t2-handbook
12:37 < rxr> we on the other hand use our (invented) automations, such as the compiler wrapper to inject what we want
12:37 < rxr> so the text log lists whatever is in the Makefile and make prints
12:37 < rxr> not what is used finally under the hook
12:37 < rxr> hood
12:37 < mpp> t2 is still the man :-)
12:37 < rxr> though the .log should have a line on top showing the wrapper action by setting GCC_WRAPPER_DEBUG and invoking gcc with -v or so
12:37 < Morglet> man++
12:37 < Morglet> :P
12:38 < mpp> hey Morglet :-)
12:38 < rxr> Command Wrapper Debug: running 'gcc --version' ..
12:38 < rxr> Old Command: gcc --version
12:38 < rxr> ...
12:38 < rxr> New Command: gcc -pipe -O2 -march=nocona --version -s
12:38 < mpp> thats fun
12:38 < Morglet> It's one of my permanent bugbears, the fact that FOSS is ignoring man pages
12:38 < rxr> so T2 injected -O2 -march=nocona on my build here
12:38 < mpp> i see
12:39 < rxr> (among other automatic fixup, such as X11R6 pathes of aging packages and lib64 etc. pp
12:39 < rxr> )
12:39 < mpp> i had made a preliminary depth graph yesterday - for cluster build partitioning
12:39 < mpp> boy i have made foundings
12:39 < mpp> qt4 needs mysql ad dependency on the runtime system
12:39 < rxr> yes, some deps are wrong ...
12:40 < rxr> yes, qt has optional SQL support ..
12:40 < mpp> thats what you mentioned
12:40 < mpp> and the handbook
12:40 < rxr> oh - btw. we have on our TODO to manually mark options deps in the .desc file
12:40 < rxr> so T2 would know that mysql is option and qt can be build without it
12:40 < mpp> do you have graph for deps on you own
12:40 < mpp> im working on one to setup cluster build
12:40 < rxr> nope, had in the past, not done that the last 12 months
12:41 < mpp> from stage 2 onwards it should be reasonable to partition the build process safely
12:41 < mpp> pxe comes to mind :-)
12:41 < rxr> patches and dep visualization stuff welcome
12:42 < mpp> t2-pxe-build-cluster-nfs-jigga
12:42 < rxr> there should be some stuff already in the sources
12:42 < mpp> im on it
12:42 < rxr> note to self, score 1385
12:42 < mpp> what i have working so far is parent-child-sibling graph
12:42 < Morglet> Now the hardest pa
12:42 < mpp> 1385 ?
12:43 < Morglet> Now the hardest part of any install ... thinking up a hostname
12:43 < mpp> localhost ?
12:43 < mpp> :-) LOL
12:43 < Morglet> Hahaha
12:43 < rxr> comparing geekbench cpus, needed a pasteboard ...
12:43 < rxr> http://browse.geekbench.ca
12:43 < Morglet> mpp: I see you've optimized your install time :P
12:44 < mpp> like how ?
12:44 < mpp> do you think...
12:44 < mpp> wth is geekbench ?
12:44 < rxr> some cpu bench thing, I only look at it because they have quite some numbers to compare it with Intel Atom
12:45 < mpp> mpp dl's geekbench...tar.gz whatever
12:45 < rxr> do not run as root :-) I have never used it
12:46 < rxr> I only look at their numbers to get a feeling for the Intel Atom performance
12:46 < rxr> 1385
12:47 < rxr> 9xx
12:47 < mpp> vi sucks
12:48 < rxr> 561 <- VIA C7-M @ 1.60 GHz
12:49 < rxr> mpp: vi the editor? what's your prefrence ?
12:49 < rxr> preferrence even
12:49 < koan> preference :-)
12:49 < Morglet> ed of course
12:49 < mpp> credits to the oldschool of course - they brought us here
12:50 < mpp> my pref is nano
12:50 < mpp> just my taste
12:52 < Morglet> It's all very Jurassic still. I'm stunned that after 30 years of Unix we're still hacking text files with text editors instead of using structured documents and structure editors.
12:53 < mpp> cheerz Morglet
12:53 < rxr> Unix was just the ripped of single user thing of multics anyway
12:53 < mpp> thank good there are stil long term geeks apreciating modern tools
12:54 < mpp> i remember back in the 80's
12:54 < mpp> digital alpha vax
12:54 < mpp> rsx11m
12:54 < mpp> i was about 8 :-)
12:55 < mpp> but take a look at the bottom geeks http://browse.geekbench.ca/geekbench2/view/76172
12:55 < mpp> score 95
12:55 < mpp> i can do lower than that with a paper and pencil :_) LOL
12:55 < Morglet> Today's "modern tools" are way inferior to what was available on LISP machines decades ago. Dunno what happened to get us into this rut, it's odd.
12:56 < mpp> it's what the mass think it's best
12:56 < Morglet> I think we may be stuck on a local minima.
12:57 < mpp> mpp does the integration calculus.... nope no optimum distance -> 80's found
12:57 < mpp> stay with vi
12:57 < mpp> its safe
12:57 < Morglet> Hahaha
12:57 < mpp> LOLOLOL :-)
12:58 < mpp> mpp is known for this kind uf stuff
12:59 < mpp> Hey Morglet - after all the mass has do stay within the normal distribution function - which is equaly low at all intervalls
12:59 < mpp> LOL :_)
12:59 < mpp> mpp kegelt sich wech....
13:01 < mpp> gee crosscompiling takes ages....
13:03 < Morglet> I'm not so sure that it is a "masses" issue. Geeks are pretty individualistic, tend to go off in random directions, so you'd think that at least one visionary would have headed in the direction of "Unix-NG".
13:04 < Morglet> (I don't know what Unix-NG is, except that it's not trad Unix :-))
13:04 < mpp> right
13:05 < mpp> mpp has geekbench 1502 on his 4year old laptop
13:05 < mpp> noted
13:06 < mpp> hey rxr
13:07 < mpp> my crossbuild stops on embutils in stage 1
13:07 < mpp> looks like they are not in the built directory ...
13:07 < mpp> what i can i do
13:08 < hikenboot> rxr, 50 MB doesnt seem too bad for size. only about twice what I could do with uclibc
13:10 < Morglet> I've just lost keyboard input in stone, even though I still have it on other VTs. If I kill off all processes with "stone" in the name from another VT, should that be enough to recover and restart?
13:13 < rxr> hm ? never heared of such an issue :-( ?
13:13 < rxr> in which dialog of stone have you been ?
13:15 < Morglet> I was configuring eth0 at the time, and switched my KVM to another port -- I think that might have caused it.
13:15 < Morglet> KVMs are the spawn of the devil, all of them
13:15 < rxr> strange when it still works in the other VTs ...
13:15 < Morglet> Yeah
13:16 < rxr> well - just try to kill away all stone stuff, and hope for the best
13:16 < rxr> or - as you apparently already installed
13:16 < rxr> and configured the boot loader? (right?)
13:16 < rxr> maybe just sync, and reboot ...
13:16 < Morglet> Anway, killing off stone failed, so I guess I need to reset vc/1 ... is killing the shell enough, will it respawn?
13:16 < rxr> should be ready already ß
13:16 < rxr> yes, the shells "respan" in the install environment
13:16 < Morglet> That's true, I think it's ready for a reboot already
13:17 < hikenboot> rxr you already built a mini distro based on somthing like deli linux for 64 bit machines?
13:18 < rxr> do not know deli linux, but I build and published our +/-50MB rescue for x86_64 some time ago
13:19 < hikenboot> got a link?
13:21 < mpp> dl.t2-project.org
13:21 < mpp> but it takes 2 hours to dl
13:23 < rxr> more mirrors welcome
13:23 < rxr> mpp you run some root server or so, maybe lots of traffic included ?
13:23 * rxr off to lunch
13:23 -!- mqueiros [n=mqueiros@217.70.75.226] has joined #t2
13:24 < Morglet> My first T2 system just booted fine :-)
13:24 < mpp> yep
13:24 < mpp> i have one
13:24 < mpp> ind im willing to share some bw for t2
13:24 < mpp> herzlichen glückwunsch Morglet
13:24 < mpp> hey hikenboot
13:25 < Morglet> I'm afraid my German is limited to Ja and zeit :-)
13:25 < mpp> got the link ?
13:25 < hikenboot> whats up mpp!
13:25 < hikenboot> got link to what?
13:25 < mpp> i could push it to my rootserver
13:25 < mpp> if you have some time
13:25 < mpp> cause i also have low bw up
13:25 < mpp> link to rescue image
13:25 < mpp> i understood you where looking for it...
13:26 < mpp> i could be wrong
13:26 < hikenboot> yes I would be interested to see whats included in the package list
13:26 < mpp> mom
13:26 < mpp> i l give u the package list in a moment
13:26 < hikenboot> I think based on what I see with deli linux that I could get an image down to 25 meg still using glibc
13:26 < Morglet> mpp: Bablefish translation: "cordial glà ¼ ckwunsch Morglet"
13:27 * Morglet chuckles
13:27 < Morglet> It's life, but not as we know it
13:27 < mpp> of course
13:27 < mpp> hikenboot
13:27 < hikenboot> yo
13:28 < mpp> i stripped down the rescue of t2 to 40mb
13:28 < mpp> and i was lazy to be honest
13:29 < hikenboot> is it somthing i could use as a starting base...I think it might save a lot of work. does it use tinyx?
13:29 < mpp> hold on a se
13:30 < mpp> http://pastebin.com/m3f29ad6e
13:31 < mpp> there is the package list
13:31 < mpp> who need clamav - dozens of filesystems editors etcetc.
13:31 < mpp> i think you can handle that one
13:38 < hikenboot> is there directions for remastering with t2?
13:38 < hikenboot> or do i just stream a package list as input to somthing fresh?
13:39 < hikenboot> kind of like buildroot or somthig?
13:41 < mpp> you build it with t2
13:41 < mpp> its part of the t2
13:41 < mpp> easiest way:
13:41 < mpp> grab sources via svn and adjust to your needs - the build
13:42 < hikenboot> so this is how it works, I get the t2 disk 32 bit image (cause thats my build machine ) and it helps to make a build root? got a link?
13:42 < mpp> t2-project.org
13:42 < mpp> download section
13:43 < hikenboot> one more question...does t2 allow you to feed it a list of packages or is it a little more work than that....right now i have a very small list of necessary packages...somewhere on this machine....even less than deli linux
13:44 < Morglet> Neat, new systems seems to be fine. I notice that no default runlevels are suggested when you add a service (eg. I added gpm and sshd, but their run-level list is empty) -- do we just take a best guess at it?
13:47 < mpp> there is adefault package list within the rescue target
13:48 < mpp> have a look at the pastebin link earlier
13:48 < mpp> @Morglet - t2 is a distro build kit . there are quite some things need to handled manually
13:49 < mpp> in stone there is however runvelel config
13:49 -!- hwinkel [n=hwinkel@p54A7718B.dip.t-dialin.net] has joined #t2
13:50 < Morglet> Yeah, and I just poked in 2 and 3. But was wondering where that info would come from in the general case.
13:52 < hikenboot> thanks mpp...see you around later...I am off to harvard
13:52 < mpp> from within the admin
13:52 < mpp> bye hikenboot
13:52 < mpp> :-)
14:04 -!- hwinkel [n=hwinkel@p54A7718B.dip.t-dialin.net] has quit [Client Quit]
14:05 < rxr> re
14:07 < rxr> Morglet: t2 default is start as few services as possible (openbsd like)
14:07 < Morglet> Yeah, I read that in manual, good policy
14:07 < rxr> Morglet: the runlevels are documented (AFAIRC) and like most linux systems do them: 5 normal + GUI, 3, normal without GUI, 2, normal without network and UI
14:08 < rxr> so you probably want to change the default runlevel to 3(with network) or 5(with net and UI) and add gpm and what was the other?) the the 2-3-5 runevel as needed
14:09 < Morglet> Yep, did that, and now gpm and sshd work fine, I'm already setting up users
14:09 < Morglet> It's all gone smoothly
14:09 < mpp> @rxr im preparing a mirror dl site
14:10 < mpp> what name do you want it to have ?
14:12 < Morglet> World works in mysterious ways: I only discovered T2 because someone said they wanted an X11 proxy so I googled for it, and gave them the link. Then I saw it was on t2-project, and I said "WTF is that?" ... :P
14:12 < rxr> mpp: well - I would push all t2 files via rsync
14:12 < mpp> well said Morglet, well said
14:12 < rxr> mpp: do you have reasonable amounts of disk space and outgoing bandwidth ?
14:12 < mpp> yeah thats okay - i need a fqdn
14:12 < mpp> yep traffic no prob
14:12 < mpp> disk space like how much ?
14:15 < mpp> do you want packets +/ isos hosted
14:15 < mpp> ?
14:16 < rxr> 36G /home/public-t2-dl/binary/stable
14:16 < rxr> 2.0G /home/public-t2-dl/binary/unstable
14:16 < rxr> 38G total
14:16 < mpp> puhh
14:16 < mpp> checking
14:16 < mpp> okay
14:16 < mpp> done
14:17 < mpp> because i have this machine for building images and stuff i have no users on it
14:17 < mpp> i would like to leave it that way
14:17 < mpp> so i would pull the stuff over if you're okay with that
14:17 < mpp> i also need a fqdn
14:17 < Morglet> This machine I just installed T2 on has 2 TB, and I have a 2nd identical box waiting. They're intended for iSCSI nodes, when I get around to it.
14:18 < mpp> cool !
14:18 < rxr> mpp: the box has a static IP and the images are probly to last there for a while?
14:19 < mpp> yep
14:19 < rxr> ok - great :_)
14:19 < mpp> i have also a second root server - maybe later on
14:19 < mpp> so just pass the fqdn and rsync url or whatever and ill get it done
14:20 < mpp> i can place a cronjob for rsyncing
14:20 < mpp> with the "master" server
14:20 < Morglet> Right, after that success, think I'll go back to reading the manual a bit, see how you guys do things.
14:21 < mpp> ...said Morglet and went back to his desk studying the holy t2 book :-)
14:22 < Morglet> Hehehe ... it came down from a local minima on tablets of stone? :P
14:23 < mpp> hail to the king baby....
15:07 < tfl0pz> haha i read this 'mail to the king baby' the first time
15:15 < Morglet> T2 was written by Elvis? :P
15:16 < rxr> hm ?
15:19 < CIA-8> rene * r30496 /trunk/package/network/thttpd/ (. no-extras.patch thttpd.cache thttpd.desc): * added thttpd (2.23b1) - A simple, small, portable, fast, and secure HTTP server
15:24 < Morglet> How do you search for all packages containing (say) "usb" in the name (not counting mine -q | grep usb)
15:25 < rxr> scripts/Find-Pkg runs some grep invocations on the package repositories
15:27 < Morglet> K. And they're grabbed with svn, or is that part of mine?
15:27 < rxr> grabbed ?
15:28 < Morglet> The scripts. I don't seem to have that directory in the install
15:28 < Morglet> So I guess there's an equivalent to Gentoo's "emerge --sync" somewhere.
15:29 < rxr> ah - yes scripts/* is inside the t2 source - either on the install media in t2-src package, or better checkout into, let's say /usr/src/t2-trunk
15:29 < rxr> actually I do not really know gentoo's commands
15:29 < Morglet> Heh, ignore Gentoo, I want to figure out how you guys do it.
15:31 < Morglet> I've been using gentoo for many years, but was sold on T2 when I read "automatic dependency management". A lot of Gentoo ebuilds have their deps totally messed up.
15:39 -!- hwinkel [n=hwinkel@p54A7718B.dip.t-dialin.net] has joined #t2
15:58 < CIA-8> mtr * r30497 /trunk/package/base/man-pages/man-pages.conf: * fixed cross-build of man-pages
16:01 [Users #t2]
16:01 [@ChanServ] [ Dallur ] [ hikenboot] [ LMJ ] [ mqueiros] [ rxr ]
16:01 [ axionix ] [ dsoul ] [ hwinkel ] [ Morglet] [ mtr ] [ tfl0pz]
16:01 [ CIA-8 ] [ Enqlave] [ koan ] [ mpp ] [ Ragnarin] [ TobiX ]
16:01 -!- Irssi: #t2: Total of 18 nicks [1 ops, 0 halfops, 0 voices, 17 normal]
16:06 < mpp> yes they where indeed broken the man-pages
16:07 < mpp> @rxr maybe you can have short look on my cross-build issue
16:07 < mpp> above
16:07 < rxr> how far above ?
16:08 < rxr> the embutils one ?
16:10 < mpp> yep that one exactly
16:10 < mpp> i have this on all machines
16:10 < mpp> sadly
16:11 < rxr> did you posted a log link ?
16:12 < mpp> nope
16:12 < rxr> ah - feared I twas blinded
16:12 < mpp> it says that the embutils cannot be find inside the built/.....
16:12 * rxr waiting
16:12 < mpp> gee.. ill have to redo the built
16:12 < mpp> but ill post it probably tomorrow
16:13 < mpp> right now machine is a bit busy building my custom rescue target
16:13 < rxr> ok
16:13 < mpp> thx
16:13 < rxr> just let me know
16:13 < mpp> ill do - promised
16:13 < rxr> btw. if I'm not too responsive in the channel one day - just drop a note to the mailing list for asynchron processing
16:14 < mpp> copy that !
16:42 < mpp> to take off the load of your office server you could already put the d2-link to your webiste
16:42 < mpp> its still syncing but people will notice and try the new location
16:43 < mpp> are these 2.x iso even necessary today ?
16:43 < mpp> i mean even the 6.x is somewhat old...
16:43 < mpp> anyways ...
16:45 < mpp> oops wrong channel
16:45 < rxr> well - they worked quite well :-)
16:45 < rxr> I know I need to get new ISOs tested and out of the door
17:17 -!- hwinkel [n=hwinkel@p54A7718B.dip.t-dialin.net] has quit [Client Quit]
17:27 < mpp> hey rxr might be worth it signin the isos with a key
17:42 < mpp> why is the gcc at stage 5 compiled with jc1 - it's for embedded systems afaikt
17:42 < mpp> it takes hours to build
17:46 < rxr> jc1 is the java compiler, no ?
17:46 < mpp> yes
17:46 < rxr> you can disable the build of java in gcc, just for the SDECFG_GCC_HAVA or so config option to 0
17:46 < mpp> whats it needed for gcc on a t2 generic target
17:47 < rxr> generic, you mean with most of the 3000 packages selected?
17:48 < mpp> ahm even minmal target does it
17:48 < mpp> thats why i am asking
17:49 < mpp> it takes many hours for this step
17:49 < mpp> unless vital for the build process i could kick it out
18:11 -!- mich1985 [n=mich1985@stgt-5d825d0c.pool.einsundeins.de] has joined #t2
18:12 < mich1985> hi all
18:13 < rxr> hi
18:13 < mpp> hey mich1985
19:02 < CIA-8> rene * r30498 /trunk/package/graphic/libpng/libpng.desc: * updated libpng (1.2.31 -> 1.2.32)
19:07 -!- Ragnarin [i=loki@planeshift/developer/Ragnarin] has quit ["brb"]
19:07 [Users #t2]
19:07 [@ChanServ] [ Dallur ] [ hikenboot] [ mich1985] [ mqueiros] [ tfl0pz]
19:07 [ axionix ] [ dsoul ] [ koan ] [ Morglet ] [ mtr ] [ TobiX ]
19:07 [ CIA-8 ] [ Enqlave] [ LMJ ] [ mpp ] [ rxr ]
19:07 -!- Irssi: #t2: Total of 17 nicks [1 ops, 0 halfops, 0 voices, 16 normal]
19:08 < CIA-8> rene * r30499 /trunk/package/gnome2/gtk+/gtk+.desc: * updated gtk+ (2.12.12 -> 2.14.2)
19:08 < CIA-8> rene * r30501 /trunk/package/filesystem/ntfs-3g/ntfs-3g.desc: * updated ntfs-3g (1.2812 -> 1.2918)
19:08 < CIA-8> rene * r30500 /trunk/package/gnome2/glib/glib.desc: * updated glib (2.16.6 -> 2.18.1)
19:12 -!- Ragnarin [i=loki@D8FA4DE6.ptr.dia.nextlink.net] has joined #t2
19:15 < CIA-8> rene * r30502 /trunk/package/base/rsyslog/rsyslog.desc: * updated rsyslog (3.18.3 -> 3.18.4)
19:34 -!- mich [n=mich1985@stgt-5d836b1e.pool.einsundeins.de] has joined #t2
19:40 -!- mich1985 [n=mich1985@stgt-5d825d0c.pool.einsundeins.de] has quit [Read error: 110 (Connection timed out)]
19:44 < mpp> i am just making a test-run on the dependency graph
19:45 < mpp> theres quite an amount of mutual dependencies
20:15 -!- felanha [i=volker@66-146-166-62.skyriver.net] has joined #t2
20:49 -!- mich [n=mich1985@stgt-5d836b1e.pool.einsundeins.de] has quit []
21:24 < mpp> hey rxr
21:24 < mpp> still at work ?
21:26 < mpp> just wanted to let u know that i finished a dep-analyzer to isolate packages with single dependency
21:26 < mpp> a the end of a leave in the depth-tree ( more like a net )
21:26 < mpp> anyways
22:46 -!- mpp [n=user@i538742DF.versanet.de] has quit [Read error: 110 (Connection timed out)]
--- Log closed Fri Sep 19 00:00:48 2008