T2 IRC Log: 2005-02-13

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 Sun Feb 13 00:00:28 2005
00:32 -!- jeru [~jeru@p54BFDA9B.dip.t-dialin.net] has quit ["using sirc version 2.211+KSIRC/1.3.11"]
01:38 < mnemoc> rehi
02:48 < CIA-9> amery * r6405 /trunk/package/dictionary/ (4 files in 2 dirs): * changed deu-eng and eng-deu to use versioned files
03:07 < CIA-9> amery * r6406 /trunk/misc/archive/ (AuditBuild.sh IsBuilt.sh): * renamed IsBuilt.sh to AuditBuild.sh, and highly improved to give better output with better performance.
03:10 -!- mcdesr [mcdesr@dynamic-216-211-50-68.tbaytel.net] has joined #t2
03:10 -!- mcdesr [mcdesr@dynamic-216-211-50-68.tbaytel.net] has left #t2 []
03:11 < CIA-9> amery * r6407 /trunk/package/contrib/i8kutils/ (. i8kutils.desc): * added i8kutils (1.25)
03:15 < CIA-9> amery * r6408 /trunk/package/contrib/xpp/ (. xpp.desc): * added xpp (1.5)
03:17 < CIA-9> amery * r6409 /trunk/package/develop/mono-doc/mono-doc.desc: * flaged mono-doc as NOPARALLEL
03:18 < CIA-9> amery * r6410 /trunk/package/develop/log4cpp/log4cpp.desc: * fixed log4cpp download location
03:20 < CIA-9> amery * r6411 /trunk/package/graphic/skencil/skencil.desc: * fixed skencil download location
03:21 < CIA-9> amery * r6412 /trunk/package/graphic/aalib/aalib.desc: * fixed aalib download location
03:22 < CIA-9> amery * r6413 /trunk/package/gnome2/i8krellm/i8krellm.desc: * added CV-URL for i8krellm
05:50 -!- rxr_ [~rene@p213.54.200.111.tisdip.tiscali.de] has joined #t2
05:50 -!- Topic for #t2: T2 | 2.1.0-beta2 RELEASED | The next generation of System Development Enviroments (SDE) | http://www.t2-project.org/ | CIA, don't ever leave us again!
05:50 -!- Topic set by mnemoc [] [Wed Jan 12 03:04:43 2005]
05:50 [Users #t2]
05:50 [ _Ragnar_] [ daja77] [ mnemoc] [ Postal|Away] [ rxr ] [ valentin]
05:50 [ CIA-9 ] [ jsaw ] [ nzg ] [ praenti ] [ rxr_]
05:50 -!- Irssi: #t2: Total of 11 nicks [0 ops, 0 halfops, 0 voices, 11 normal]
05:50 -!- Channel #t2 created Sun Aug 8 21:15:33 2004
05:50 -!- Irssi: Join to #t2 was synced in 11 secs
06:06 -!- rxr [~rene@p213.54.195.112.tisdip.tiscali.de] has quit [Read error: 110 (Connection timed out)]
06:35 -!- Postal|Away is now known as Postal
06:35 < Postal> hey all
08:14 < rxr_> re
08:14 -!- You're now known as rxr
08:14 < rxr> hi Postal
08:16 < rxr> Electric Fence 2.4.13
10:16 -!- jeru [~jeru@p54BFE652.dip.t-dialin.net] has joined #t2
10:36 < rxr> re
10:36 < rxr> moin jeru
10:36 < jeru> moin moin - gut geschlafen?
10:37 < rxr> jep ,-) thanks - you 2?
10:37 < jeru> yep ...
10:39 < jeru> do you have some time looking over an error log of a failed build of e2fstools?
10:39 < rxr> yep
10:39 < jeru> I will send it to the mailing-linst , ok?
10:39 < jeru> list I mean
10:40 < rxr> ok
10:41 < jeru> talked about it with mnemoc yesterday, but the strange thing is, that he told me it builds on his machine ... hmmm
10:46 < jeru> seems that there a problems including header file 'ext2_err.h'
10:46 < jeru> btw building it manually from source seems to work
10:50 < rxr> hm
10:51 < rxr> hm - powerpc:
10:51 < rxr> Testing Electric Fence.
10:51 < rxr> After the last test, it should print that the test has PASSED.
10:51 < rxr> ./eftest
10:51 < rxr> Electric Fence: ptr=0x30045ff0 size=16 alloced from UNKNOWN (use #include "efenc
10:51 < rxr> e.h")(0) not freed
10:51 < rxr> Electric Fence: ptr=0x30047f80 size=128 alloced from UNKNOWN (use #include "efen
10:51 < rxr> ce.h")(0) not freed
10:51 < rxr> Electric Fence: ptr=0x30049fec size=20 alloced from UNKNOWN (use #include "efenc
10:51 < rxr> e.h")(0) not freed
10:51 < rxr> Electric Fence: ptr=0x3004bfe0 size=29 alloced from UNKNOWN (use #include "efenc
10:51 < rxr> e.h")(0) not freed
10:51 < rxr> ElectricFence Aborting:
10:51 < rxr> Electric Fence: EF_delFrame(): Found non free'd pointers.
10:51 < rxr> make: *** [all] Illegal instruction (core dumped)
10:51 < rxr> building manually worked ...
10:51 < rxr> flist wrapper ? *g*
10:54 < rxr> yes - ack
10:54 < rxr> removing the flist wrapper makes the tests pass ...
10:54 < rxr> jsaw: ?
10:54 < rxr> jsaw: what hardware are you building e2fsprogs on?
10:54 < rxr> some SMP box ?
10:55 < jeru> jsaw: p4 but i386 optimization
10:55 < rxr> oehm ...
10:55 < rxr> I mean jsaw ? flist wrapper and the next line to you jeru ...
10:56 < jeru> sorry ,-)
11:02 < jeru> rxr: don't building on a SMP box
11:03 < rxr> Linux 2.4.30-pre1
11:04 < rxr> Linux 2.6.11-rc4
11:05 < jeru> 2.6.10
11:07 < rxr> Re: Crosscompiling issues with glibc-2.3.4
11:07 < rxr> Cross-compiling is not supported.
11:07 < rxr> ➧ Ulrich Drepper ➧ Red Hat, Inc. ➧ 444 Castro St ➧ Mountain View, CA ❖
11:11 < rxr> have you tried if building with NOPARALLEL fixes e2fsprogs for you ?
11:13 < jeru> no
11:13 < rxr> could you try?
11:14 < rxr> I look into the makefiles in the meantime ...
11:14 < rxr> you just need to add NOPARALLEL to the [F] section of the .desc
11:14 < jeru> do it right now ;-)
11:16 < jeru> yep :-) it works ..
11:17 < rxr> ok - /me reviewing the makefiles
11:18 < jeru> ack
11:23 < rxr> strange
11:23 < rxr> your make is choosing to build:
11:23 < rxr> block.o
11:24 < rxr> before this one is preprocessed ...
11:24 < rxr> ex2_err.et
11:24 < rxr> although it is a dependency ...
11:24 < rxr> strange
11:24 < rxr> when I make clean and just build make block.o it correctly preprocessed this ext2_err.et fgirst
11:25 < jeru> hmmmm ...
11:25 < rxr> # Hack to parallel makes recognize dependencies correctly.
11:25 < rxr> #
11:25 < rxr> $(top_builddir)/lib/ext2fs/ext2_err.h: ext2_err.h
11:25 < rxr> ?
11:25 < rxr> seems someone already hacked around parallel makes ...
11:26 < rxr> what system are you building on?
11:26 < rxr> I guess some rock or t2 ?
11:26 < jeru> t2
11:29 < jeru> host is a t2-beta2
11:31 < rxr> ok - I have some hack
11:31 < rxr> I check it in so you have fun testing ...
11:31 < jeru> thank you ;-)
11:33 < jeru> should I remove the NOPARALLEL flag ?
11:33 < rxr> yep
11:34 < jeru> ack
11:35 < CIA-9> rene * r6414 /trunk/package/filesystem/e2fsprogs/parallel-make.patch: * work around e2fsprogs parallel make problem
11:35 < rxr> much fun testing
11:37 < rxr> == 02/13/05 11:37:26 =[9]=> Finished building package e2fsprogs.
11:37 < rxr> ^- on my iBook - but it built there before, too
11:55 < jeru> sorry ... was shortly off ... because my critter needed some devotion
11:56 < jeru> will now continue testing ,-)
12:16 < jeru> rxr: == 02/13/05 12:16:18 =[1]=> Finished building package e2fsprogs.
12:17 < rxr> good
12:17 < rxr> ,-)
12:18 < jeru> hehe ,-)
12:18 < jeru> but now having problems with iproute2
12:19 < jeru> trying to fix it
13:22 < mnemoc> moin!
13:22 < jeru> moin! mnemoc ,-)
13:23 < mnemoc> hi jeru
13:23 < mnemoc> you told me you were using HEAD
13:24 < jeru> do you mean in conjunction with the e2fsprogs problems?
13:25 < mnemoc> yes
13:26 < jeru> right now my p4 is running beta2 but I want to upgrade as soon as possible
13:27 < jeru> but if I cannot build a target I've got some problems ,-)
13:28 < mnemoc> your running is machine is beta2 but are you building over a working copy of trunk:HEAD?
13:28 < jeru> but anyway, I'm sorry if I confused you - mea culpa ,)
13:28 < mnemoc> :)
13:31 < jeru> I 'screwed up' my last system and I just had the beta2 binary dist available :/
13:33 < jeru> mnemoc: do you have something agains it if I contribute some small things to mnemosyne ?
13:33 < mnemoc> sure
13:34 < mnemoc> that's why i have mnemosyne there :)
13:35 < jeru> you know I like it very much :)
13:35 < mnemoc> yes?
13:35 < mnemoc> :D
13:35 < jeru> first it will be some pkgsel related things
13:35 < jeru> I told you before (some month ago)
13:36 < mnemoc> but as i didn't hear you again i thought you got desapointed
13:37 < jeru> yep ... I know I wasn't so much on channel but I'm not disapointed with it ;)
13:38 < mnemoc> :D
13:39 < jeru> first I want to add bridging support to the network section
13:39 < mnemoc> add it as 'Service' please
13:39 < mnemoc> Service/Network/Bridging?
13:39 < jeru> ack
13:40 < mnemoc> i want to add a ipvsadm/keepalived module there too
13:40 < jeru> ok
13:41 < mnemoc> btw, is Config too slow?
13:41 < jeru> nope
13:41 < jeru> but it was the first time I used it
13:42 < jeru> RADIUS support is also on the way because I'll need it very soon
13:43 < mnemoc> add it freely :)
13:43 < jeru> intend to add freeradius package
13:43 < mnemoc> package/security/ ?
13:44 < jeru> hmm? or package/network?
13:44 < mnemoc> considering it is an 'authentication' package....
13:45 < mnemoc> but, follow your instints
13:45 < jeru> ok ... we'll place it in package/security
13:45 < jeru> :)
13:45 < mnemoc> :)
13:53 < jsaw> re
13:54 < mnemoc> re jsaw
13:54 < jsaw> hi mnemoc
13:55 < jeru> hi jsaw
13:55 < mnemoc> when is your daughter 'scheduled for release'?
13:55 < jeru> :)
13:55 < jsaw> hi jeru
13:55 < jsaw> 6 days to go, according to calc
13:57 < mnemoc> ansious?
13:57 < jsaw> have you been anxious with your second one?
13:58 < mnemoc> not really :p
13:58 < jsaw> <- no. Maybe a bit when it starts... until we arrive in hospital...
13:58 < mnemoc> it was much more 'natural' :p
13:58 < jsaw> yep
14:01 < rxr> oh! jsaw !
14:01 < jsaw> hi rxr
14:02 < mnemoc> hi rxr
14:02 < mnemoc> how finished your win98/scanner experience?
14:02 < jsaw> rxr, jeru: I'll commit a fl_wrapper that closes the libc_handle on exit, then you may test efence again
14:02 < rxr> with new knowledge how much window~1 sucks ...
14:03 < mnemoc> jsaw: 'on exit'?
14:03 < jsaw> atexit...
14:03 < rxr> jsaw: nice - I would do that, too - but am hacking my sane avision driver (as usual these days ...
14:03 < mnemoc> jsaw: how do you detect it?
14:03 < rxr> mnemoc: the libc does it ...
14:03 < jsaw> mnemoc: man atexit
14:03 < rxr> callint the atexit registed stuff ...
14:03 < mnemoc> oh
14:04 < mnemoc> great
14:05 < rxr> but at least in glibc there are more ctor dtor attribute hooks available
14:05 < CIA-9> jsaw * r6415 /trunk/misc/tools-source/fl_wrapper.c: * close libc_handle in fl_wrapper (hoping to make efence happy)
14:05 * rxr testing
14:06 < jsaw> mnemoc: runit fails
14:06 < jsaw> ctrlaltdel patch...
14:07 < jeru> mnemoc: runit fails during my tests too
14:07 < mnemoc> what?
14:07 < mnemoc> uhm
14:07 < mnemoc> rockify.patch fail to apply?
14:08 < jsaw> yep
14:08 < mnemoc> uhm
14:09 < rxr> rockify?
14:09 < mnemoc> :p
14:11 < mnemoc> rxr: can you review libgettex-$host thing today?
14:11 < rxr> ouhm ...
14:11 < rxr> yeah - ok
14:12 < mnemoc> please :)
14:13 < mnemoc> i'll try to dig into postlinux/custmain issue
14:15 < rxr> jsaw: your patch does not yet fix efence
14:15 < jsaw> same errors?
14:15 < rxr> yep
14:15 < jsaw> hmm...
14:15 < rxr> I could just remove to run the test in the newer efence for now ..
14:15 * mnemoc emerging runit directly on his 386 router :\
14:15 < jsaw> have to change atexit ordering...
14:16 < jsaw> rxr: remove it... if our atexit is called after efence's...
14:18 < rxr> do you want to track it - or should I patch the test away now ?
14:18 < jsaw> no time to track it more deeply... patch the test away.
14:18 < rxr> oki
14:18 < rxr> do you add it to the wiki todo ?
14:19 < rxr> (reviewing it)
14:21 < jsaw> hmmm. don't know if it is necessary. If only efence doesn't get it right...
14:22 < jsaw> you can test adding a printf statement near the dlclose. If efence complains first, we have no chance to correct this.
14:22 < mnemoc> runit bug was stupid :\ -p0 vs. -p1 ... i'll commit
14:23 < jsaw> If you add ./ to the diff paths, it wouldn't matter
14:23 < mnemoc> that's what i did :p
14:23 < jsaw> <- shutting up
14:23 < mnemoc> :D
14:24 < mnemoc> i don't understand how i commit it without doing such a simple test :|
14:24 < CIA-9> amery * r6416 /trunk/package/base/runit/rockify.patch: * fix runit rockify patch to apply with -p1 (stupid mnemoc)
14:25 < mnemoc> jeru: runit fixed (i hope)
14:26 < jeru> mnemoc: thank you .. will test it immediatly
14:27 < mnemoc> :)
14:30 < jsaw> works
14:33 < jeru> mnemoc: runit works for me too
14:33 < mnemoc> great
14:41 < CIA-9> rene * r6417 /trunk/package/base/electricfence/ (electricfence.desc no-tests.patch): * updated electricfence (2.1.14 -> 2.4.13)
14:41 < mnemoc> uh
14:46 < CIA-9> amery * r6418 /trunk/package/base/runit/rockify.patch: * fixed runit to clean state of 'reboot' too on boot up.
14:50 < rxr> mnemoc: yes ?
14:53 < mnemoc> i didn't noticed efence's new home
14:53 < CIA-9> amery * r6419 /trunk/package/base/electricfence/electricfence.desc: * added new author, url and CV-URL for electricfence
14:55 < mnemoc> .oO( Emerge-Pkg is not doing backups, it overwrites _and_ creates .new file )o
14:55 < rxr> it should
14:56 < rxr> per default it uses the Build-Pkg -update mode ...
14:56 < mnemoc> then Build-Pkg -update is broken
14:57 < mnemoc> i just emerged runit on my running beta4 router and etc/runit/* files got overwritten
14:59 < rxr> mnemoc: do you try to update any package, btw ?
15:00 < mnemoc> what do you mean?
15:02 < mnemoc> my sweep? i want 2.1 to have an updated enough base. but i don't update stuff i think will be harmfull or not release stable before $release21+1_month
15:03 < mnemoc> rxr: if you consider any of the updates i did or will dangerous same way _please_ tell me
15:04 < rxr> no - i just wanted to know what I asked - if you try any package ...
15:04 < rxr> e.g. if I should look for some myself ...
15:04 < rxr> but when you go over all - well I can continue typing this fluff ...
15:05 < mnemoc> i stoped right after network/
15:06 < mnemoc> there are many packages i would like to update after that but i prefered to stabilize this 157 pending packages i have here :)
15:06 < mnemoc> btw, misc/archive/AuditBuild.sg is getting nice
15:06 < mnemoc> .sh
15:09 < mnemoc> ext2fs.h:86:29: ext2fs/ext2_err.h: No such file or directory <--- what?! .... i wasn't having PARALLEL problems of e2fsprogs here
15:11 < jsaw> btw, the "Hack to parallel makes recognize dependencies correctly." that rxr posted 11:27 is in wrong order I would assume
15:12 < rxr> ?
15:12 < jsaw> $(top_builddir)/lib/ext2fs/ext2_err.h: ext2_err.h
15:12 < jsaw> should be:
15:12 < jsaw> ext2_err.h: $(top_builddir)/lib/ext2fs/ext2_err.h
15:13 * mnemoc just realized GCC_WRAPPER_APPEND and friends doesn't work on stage 9
15:15 < mnemoc> +a: ext2_err.h ext2_types.h all <--- this is the patch i have
15:21 < rxr> hm
15:23 < jsaw> & has precedence over == ?
15:23 < mnemoc> i hope
15:23 < jsaw> sounds not logical to me...
15:26 < mnemoc> a & b == false
15:27 < jsaw> oh, my typo...
15:27 < Postal> Hey all
15:27 < mnemoc> hi Postal64
15:27 < jsaw> mnemoc: I did this, and it evaluated to a & (b == false)
15:27 < mnemoc> oh
15:27 < jsaw> I wanted to say it exactly the other way round...
15:28 < jsaw> hi Postal
15:28 < Postal> rxr, any news on glibc32?
15:28 < rxr> ehrm - well, ...
15:28 < rxr> mipe?
15:29 < mnemoc> hehe
15:29 < Postal> mipe = ?
15:30 < mnemoc> a nick
15:30 < Postal> oh
15:31 < Postal> rxr, ok, so when i build this generic with multilib, ill just disable glibc32..?
15:32 < rxr> you always need multilib for x86-64 - _and_ the glibc32 when you want to be able to generate 32bit binaries like grub wants
15:32 < rxr> Postal: maybe just build without glibc32 for now and use your current system grub
15:33 < mnemoc> dummy question, how can i add opt/*/{include,lib} to gcc defaults?
15:33 < rxr> mipe reported some more problems involved to get glibc32 built ...
15:33 < Postal> rxr, ok
15:33 < rxr> mnemoc: ohem - why ?
15:33 < mnemoc> samba, postgres headers
15:33 < mnemoc> heimdal headers
15:33 < rxr> you should not add all
15:34 < rxr> only that used by the packeg
15:34 < rxr> in a perfect workd the package should determine those itself
15:34 < rxr> using $pkg-config or pkgconfig -bla blubs ...
15:34 < mnemoc> they don't
15:34 < mnemoc> we do it using GCC_WRAPPER_APPEND
15:34 < rxr> take a look into other packages's pkg_prefix to determine the location of the packages you need to build against
15:34 < mnemoc> which doesn't work on stage 9
15:34 < rxr> it must work
15:35 < rxr> otherwise there is a bug
15:35 < rxr> never had a problem with GCC_WRAPPER* on any stage ...
15:35 < mnemoc> another :(
15:36 < rxr> why do you think it does not work
15:36 < rxr> maybe you mistyped s.th.
15:36 < rxr> if the WRAPPER would not work we would not have ccache and distcc on stage9
15:36 < mnemoc> on stage 5 smbc builds
15:36 < rxr> and it all works for me ...
15:36 < mnemoc> but not emerging it
15:37 * mnemoc checking deeper
15:39 < jsaw> <- taking a nap, cu later
15:40 < Postal> cu jsaw
15:40 < mnemoc> cu jsaw
15:40 < rxr> cu jsaw
15:42 < Postal> rxr, so any do you think that the build will chroot in this time?
15:43 < Postal> opps no "any"
15:43 < rxr> yes - it definetly should
15:43 < Postal> cause last time it did not..
15:44 < Postal> but im still downloading the packages so well see shortly
15:46 < mnemoc> Postal: Download is done in build order, so you don't need to wait to start building :)
15:49 < Postal> oh
15:51 < mnemoc> rxr: removing your patch e2fsprogs built again
15:51 < rxr> Postal: you _need_ multilib ...
15:51 < rxr> it did build here ...
15:51 < rxr> strange ...
15:51 < mnemoc> yep
15:51 < rxr> why did it fail ?
15:51 < Postal> rxr, i_have_multilib
15:52 < mnemoc> [11:09:18] ext2fs.h:86:29: ext2fs/ext2_err.h: No such file or directory
15:52 < rxr> oehm
15:52 < mnemoc> weird, isn't it?
15:52 < mnemoc> we have a big randomness problem here
15:53 < mnemoc> you don't have linux26/nvidia linux24/alsa-driver problem i do
16:03 < Postal> rxr, where is the package config to diable the building of a package?
16:04 < mnemoc> Config -> Expert -> Pkgsel -> 'O package'
16:06 < Postal> mnemoc, it wont let me edit the list..
16:06 < mnemoc> uhm?
16:06 * rxr shortly away
16:07 < Postal> to cancel out a package, i need to not build glibc32
16:07 < mnemoc> - glibc32
16:07 < mnemoc> O packages are re-enablable
16:07 < mnemoc> - packages are gone for ever
16:08 < Postal> right, im staring at the Package list right now, but config does not give me the option to edit it
16:09 < mnemoc> Additional Package Selection
16:09 < mnemoc> (none) No package preselection template
16:09 < mnemoc> [*] Custom package selection
16:09 < mnemoc> ===> Edit package selection rules <--- here
16:10 < Postal> ah
16:10 < Postal> so - glibc32]
16:10 < Postal> without the ]
16:10 < mnemoc> '- glibc32'
16:10 < mnemoc> without the quotes :)
16:12 < Postal> ok should be good now
16:32 < CIA-9> amery * r6420 /trunk/package/network/udhcp/ (. udhcp.desc): * added udhcp (0.9.x-9568)
16:57 < rxr> amery: and pump ? ;-)
16:59 < mnemoc> udhcp_d_ :)
16:59 < mnemoc> and a simple udhcpc too of course ;)
16:59 < Postal> building the kernel takes a goodly long time eh
17:00 < mnemoc> 1 hour here
17:00 < Postal> although it took me a little under 20 min to build glibc
17:00 < mnemoc> another hour here :\
17:01 < Postal> thats the beauty of 64 bit with 2gb of ram
17:01 < Postal> I had to work alota extra hours, but it was worth it
17:02 < mnemoc> my extra hours are to pay daughters kinder garden
17:03 < Postal> If I had a kid, id be in the same boat... im never having kids though, they might turn out like me... i could not deal with that ;-)
17:04 < mnemoc> :)
17:04 < Postal> i needed a new box though, my old one was an 800mhz p3 or 2 i cant remember.. but it was mighty slow
17:08 < mnemoc> 800? P3
17:13 < Postal> how big did yall make this kernel, i seem to remember being able to comile kernels by hand a bit faster than this..
17:14 < mnemoc> we build every module available on it
17:14 < Postal> that explains it..
17:14 < Postal> question
17:15 < Postal> what is the advantage of building something as a module as opposed to building it right into the kerne;?
17:15 < Postal> kernel*
17:17 < rxr> first when you build all into the kernel you could not even load it any more on most architectures because it is too fat
17:17 < rxr> then you would waste a lot or non-swapable kernel memory just to hold all that stuff ...
17:17 < Postal> ah
17:17 < rxr> then you can reload module when e.g. they have some problem or reload some modified module ...
17:18 < mnemoc> the question why to build 'foo' as built-in instead of module i think
17:18 < rxr> we will follow the other distributions way and make even more modular ...
17:18 < rxr> currently some stuff needed for booting is built in
17:18 < rxr> but in fact too much for many boxes - and too less for others ....
17:18 < Postal> ic
17:18 < rxr> so we make all modular and supply the stuff needed for the actual machine in a initrd (initial ramdisk) which is loaded in conjunction with the kernel ...
17:19 < Postal> ic
17:19 < rxr> it is a more complex process - but currently the kernel get's to fat for some architectures - notable sparc and powerpc
17:19 < rxr> and currently you can not boot a t2 system on some esoteric hardware for wich the drivers are not built in
17:19 * Postal nods
17:20 < rxr> with a modular initrd you can boot on any box just injecting the needed modules into it ...
17:35 < Postal> rxr, this multilib build seems to be going much smoother than the one without multi lib
17:36 < mnemoc> :)
17:36 < rxr> yes - Postal, I know I should finally automate enabling of that option
17:37 < rxr> sorry - too high load here ...
17:37 < Postal> rxr, on the website, it says it will take 3 days to build a full desktop build on a 3Ghz amd64, is that a correct time?
17:38 < rxr> on which page ?
17:38 < Postal> quick start
17:38 < rxr> ouhm - valentin ?
17:38 < rxr> Postal: maybe it lists ordinary athlon (xp)
17:38 < Postal> "A full build of the desktop target takes about three days on a 3 Ghz Athlon 64."
17:38 < rxr> but I think the time should match current trunk's HEAD
17:38 < rxr> oh - interesting ... ;-)
17:39 < rxr> yes this is right - maybe slightly faster
17:39 < Postal> well im building a generic, so i guess my machine is tied up for a few days..
17:39 < rxr> oh - yes, a generic will build longer
17:40 < Postal> well prehaps with the cachcc and distcc, with parallel build, it will build a little faster
17:40 < rxr> you enabled distcc ?
17:41 < Postal> prehaps
17:41 < Postal> i usually do
17:41 < rxr> you know that you need some other boxes to distribute the compile jobs then =
17:41 < rxr> ? even
17:43 < Postal> I was under the impression, that, if there where do other boxes, it would just not use it.. and actually, i did not enable it for this build, i just checked
17:43 < Postal> do = no*
17:44 < rxr> well - sure - but I think you loose a few milli seconds due to distcc starting and some conditoinals if you do not actually have the other boxes to distribute the jobs too
17:44 < rxr> -o
17:44 < Postal> well, im really worried about thos milli seconds.. lol
17:44 < Postal> they might add up to what 2 seconds over a whole build?
17:45 < rxr> yes - or evevn 4
17:45 < Postal> hmm, im not to worried, i suppose that I could plug my laptop into my router, and distribute the jobs on that..
17:46 < Postal> its a 1.8ghz, with 512 megs of ram, it might speed things up
17:47 < Postal> rxr, where I to build t2 on my laptop, which is a 32 bit machine, do you susspect that it would build without many errors?
17:48 < rxr> you mean a nornal 32bit build natively - or some distcc involved ?
17:48 < Postal> nativley
17:49 < Postal> and its a P4 so yes 32 bit
17:50 < jeru> hmmm ... strange thing...on another machine e2fsprogs failed to build if NOPARALLEL flag isn't set
17:50 < Postal> yeah, i have the problem also
17:50 < rxr> well - generic always has some errors - e.g. due to toy packages added by people longer time ago which got not yet fixed for the actual gcc or kernel ...
17:51 < Postal> yeah, but i mean errors that actually matter lol
17:51 < rxr> well - normally trunk should not suffer from errors that actually matter
17:51 < Postal> oh and is gdesklets in the repository?
17:51 < rxr> except stuff like e.g. the current e2fschk that will soon be fixed ...
17:51 < rxr> yes
17:51 < rxr> gnome2 IIRC
17:52 < rxr> jeru: can you scroll back and see jsaw comment
17:52 < Postal> cool, cause i hate solving all the deps on that
17:52 < rxr> (if it is not yet committed)
17:52 < Postal> well, ill know if it was commited (e2fsprogs) in a sec, cause i think im about to build it
17:54 < jeru> rxr: i think it isn't commited already
17:54 < rxr> failed on your side ?
17:55 < jeru> yep
17:55 < rxr> what is the error?
17:55 < jeru> the same as before
17:55 < jeru> but the machine is in sync
17:55 < jeru> hmmm?
17:56 < jeru> added NOPARALLEL temporary
17:56 < rxr> is the error this: file not found thing ?
17:58 < Postal> well, ill just go add NOPARALLEL now then before i get to it and fail
17:58 < rxr> Postal: ehrm - what ?
17:59 < Postal> well, unless someone commited the change since last night, the build fails on e2fsprogs and a NOPARALLEL fixes it..
17:59 < Postal> ive had to inject it twice now..
18:00 < Postal> rxr, any objections?
18:01 < rxr> no - I just did not got all the details out of your last post ...
18:01 < rxr> I commited a fix today ...
18:01 < rxr> but it does not seem to fix it fully ...
18:01 < rxr> it would be helpful to know what error you just got
18:01 < rxr> some file not found error?
18:01 < Postal> well, NOPARALLEL is still not in the .desc..
18:01 < rxr> NOPARALLEL disabled parallel building - I actually patched to makefile to get it build in parallel ...
18:02 < Postal> ah so it should build without that option
18:02 < rxr> I mean - NOPARALLEL is just a workaround for packages we do not care enought about to fix properly ...
18:02 < Postal> well, i wont add the option then, and well see when we get to it
18:02 < Postal> ;-)
18:04 < rxr> 6623 avision.c
18:04 < rxr> well - in soem magic way it grew a lot lately ...
18:04 < jeru> rxr: have to check the e2fsprogs later ... will send you the log later
18:04 < rxr> that is the no of lines ...
18:04 < rxr> ok
18:05 < rxr> 195K Feb 13 18:04 avision.c
18:05 < rxr> a nice fat file *g*
18:05 < rxr> it was 39K some years ago ... and about 1469 lines ...
18:06 < Postal> rxr, what are you talking about?
18:06 < rxr> my scanner driver i/o core - as shipped licensed under the GPL within SANE
18:07 < rxr> but is also used in a commercial product now ...
18:08 < Postal> i dont think the download follows the build order as mnemoc suggested, cause i am now having to download packages as I go, so im going to wait for the download to finish
18:09 < rxr> well the problem is, that the download is run uncoupled asynchronly in the background ...
18:09 < rxr> my plan is, to link it more closely with the build - so missing files are fetched on the fly
18:09 < Postal> if i download packages as I go, will that mess up the current download if it runs into a package I already downloaded?
18:10 < rxr> well - when I implement it, I'll of course add propper locking
18:10 < Postal> i mean for right now... can i download the missing package and not screw up the whole other download process
18:11 < Postal> cause its in the middle of fetching 1331 or so files
18:12 < rxr> yes you can run multiple Download instances ...
18:12 < Postal> ok
18:12 < rxr> they lock the files and not corrupt them ...
18:12 < Postal> what do you mean by lock..?
18:13 < rxr> locking is some general term in computer science
18:14 < Postal> ok..
18:14 < Postal> e2fsprogs is going to build as soon as I dl it
18:14 < rxr> to lock some "primitive" is to prevent accessing it from more than "location" at a time
18:14 < Postal> ok that makes more sense
18:14 < Postal> ;-)
18:15 < Postal> so lets see if e2fsprogs builds shall we
18:15 < Postal> ohh error
18:16 < Postal> make[2]: *** [ind_block.o] Error 1
18:16 < Postal> make[2]: *** Waiting for unfinished jobs....
18:16 < Postal> CC cmp_bitmaps.c
18:16 < rxr> and a bit above ?
18:17 < Postal> ext2fs.h:992: error: `EXT2_ET_NO_MEMORY' undeclared (first use in this function)
18:17 < Postal> ext2fs.h: In function `ext2fs_resize_mem':
18:17 < Postal> ext2fs.h:1019: error: `EXT2_ET_NO_MEMORY' undeclared (first use in this function)
18:17 < Postal> COMPILE_ET ex2_err.et
18:17 < Postal> COMPILE_ET ex2_err.et
18:17 < Postal> CC bb_compat.c
18:17 < Postal> make[2]: *** [ind_block.o] Error 1
18:17 < Postal> make[2]: *** Waiting for unfinished jobs....
18:17 < rxr> and even higher *g* ?
18:17 < Postal> ../../lib/ext2fs/bitops.h:526: error: for each function it appears in.)
18:17 < Postal> ../../lib/ext2fs/bitops.h: In function `ext2fs_mark_block_bitmap_range':
18:17 < Postal> ../../lib/ext2fs/bitops.h:562: error: `EXT2_ET_BAD_BLOCK_MARK' undeclared (first use in this function)
18:17 < Postal> ../../lib/ext2fs/bitops.h: In function `ext2fs_unmark_block_bitmap_range':
18:17 < Postal> ../../lib/ext2fs/bitops.h:592: error: `EXT2_ET_BAD_BLOCK_UNMARK' undeclared (first use in this function)
18:17 < Postal> In file included from ind_block.c:20:
18:17 < Postal> ext2fs.h: In function `ext2fs_get_mem':
18:17 < Postal> ext2fs.h:992: error: `EXT2_ET_NO_MEMORY' undeclared (first use in this function)
18:17 < Postal> ext2fs.h: In function `ext2fs_resize_mem':
18:17 < Postal> ext2fs.h:1019: error: `EXT2_ET_NO_MEMORY' undeclared (first use in this function)
18:17 < Postal> COMPILE_ET ex2_err.et
18:17 < Postal> COMPILE_ET ex2_err.et
18:17 < Postal> CC bb_compat.c
18:17 < Postal> make[2]: *** [ind_block.o] Error 1
18:17 < Postal> make[2]: *** Waiting for unfinished jobs....
18:18 < rxr> hm - ic
18:18 < Postal> i can tail it higher if you need it
18:18 < Postal> do you need more lines?
18:18 < rxr> give me some secs - I commit an update
18:18 < Postal> ok
18:22 < rxr> so - now I @ e2fsprogs
18:22 < Postal> just lemme know when to svn up
18:26 < Postal> rxr, You know I have to give you some credit, I have thought about just throwing in the towel, and downloading some other distro, but every time I go and look, I just feel that this way is better, and so then keep on workin...
18:26 < Postal> ;-)
18:29 < rxr> Postal: go for test
18:29 < Postal> svn up?
18:29 < rxr> yeo
18:29 < rxr> yep, even
18:30 < Postal> you know what it is, i LOVE this package management system..
18:30 < jeru> rxr: have the same e2fsprogs error like Postal ... start testing ,)
18:30 < Postal> nope
18:31 < rxr> ,-) yes - this package parameter sets are really nice - a really cool huge knowledge base how to get the packages built ...
18:31 < rxr> Postal: nope == error persists ?
18:31 < Postal> linux:/mnt/t2/t2-trunk # tail -n 30 build/system-2.1.0-beta4-x86-64-generic-expert/var/adm/logs/1-e2fsprogs.err
18:31 < Postal> from ind_block.c:20:
18:31 < Postal> ../../lib/ext2fs/bitops.h: In function `ext2fs_test_block_bitmap_range':
18:31 < Postal> ../../lib/ext2fs/bitops.h:526: error: `EXT2_ET_BAD_BLOCK_TEST' undeclared (first use in this function)
18:31 < Postal> ../../lib/ext2fs/bitops.h:526: error: (Each undeclared identifier is reported only once
18:31 < Postal> ../../lib/ext2fs/bitops.h:526: error: for each function it appears in.)
18:31 < Postal> ../../lib/ext2fs/bitops.h: In function `ext2fs_mark_block_bitmap_range':
18:31 < Postal> ../../lib/ext2fs/bitops.h:562: error: `EXT2_ET_BAD_BLOCK_MARK' undeclared (first use in this function)
18:31 < Postal> ../../lib/ext2fs/bitops.h: In function `ext2fs_unmark_block_bitmap_range':
18:31 < Postal> ../../lib/ext2fs/bitops.h:592: error: `EXT2_ET_BAD_BLOCK_UNMARK' undeclared (first use in this function)
18:31 < Postal> In file included from ind_block.c:20:
18:31 < Postal> ext2fs.h: In function `ext2fs_get_mem':
18:31 < Postal> ext2fs.h:992: error: `EXT2_ET_NO_MEMORY' undeclared (first use in this function)
18:31 < Postal> ext2fs.h: In function `ext2fs_resize_mem':
18:31 < Postal> ext2fs.h:1019: error: `EXT2_ET_NO_MEMORY' undeclared (first use in this function)
18:31 < Postal> make[2]: *** [ind_block.o] Error 1
18:31 < Postal> make[2]: *** Waiting for unfinished jobs....
18:31 < Postal> you could just inject NOPARALLEL for this one package..
18:32 < CIA-9> rene * r6421 /trunk/package/filesystem/e2fsprogs/parallel-make.patch: * next attempt to fix e2fsprogs parallel build
18:32 < Postal> it took CIA a long time to post that fix
18:33 < rxr> Postal: can you send me the whole log ?
18:34 < Postal> yeah email adress
18:35 < Postal> oh i have it
18:35 < rxr> ok - I see what I did "wrong"
18:37 < jeru> rxr: still same error
18:37 < Postal> ok well its sent
18:37 < Postal> so you should have the error log in a few seconds
18:40 < rxr> ok - go
18:40 < rxr> untested
18:41 < Postal> building now
18:41 < Postal> so well see in a few secs
18:44 < rxr> built on my iBook
18:44 < rxr> but it always built on it ...
18:44 < Postal> it build
18:44 < rxr> builds or built ?
18:44 < Postal> it is done finished, the patch works..
18:44 < Postal> ;-)
18:44 < jeru> nope
18:44 < rxr> good ,-) FINALLY!
18:44 < Postal> apparently not for jeru
18:45 < jeru> it doesn't build on my system :(
18:45 < rxr> jeru: ouhm
18:45 < rxr> still the same error ?
18:45 < Postal> jeru, you did svn up right
18:45 < jeru> yep
18:45 < Postal> silly question i know
18:45 < jeru> :9
18:45 < jeru> :)
18:45 < rxr> really really the latest change *g*
18:45 < rxr> jeru: still the same error - or did it change ?
18:46 < jeru> it's like the first error before you initialy wrote the patch
18:46 < jeru> ext2fs.h:86:29: ext2fs/ext2_err.h: No such file or directory
18:47 < jeru> send the log to the mailing list
18:48 < rxr> can you send me the new one in private?
18:48 < Postal> linux:/mnt/t2/t2-trunk # ./scripts/Download iproute2
18:48 < Postal> INFO: To force a new mirror auto-detection, remove download/Mirror.
18:48 < Postal> INFO: Found cached mirror URL in download/Mirror:
18:48 < Postal> INFO: http://gsmp.tfh-berlin.de/mirror/t2-source/2.1
18:48 < Postal> grrr it will not download iproute2
18:49 < rxr> maybe it is already in place
18:49 < Postal> i dont think so cause it asked me if i dl the package, i suppose it is possible that it happened JUST before the other dl go to it
18:50 < jeru> oh sh*t it's already send ... next time I'll send it to your private adress
18:50 < Postal> yup thats what happened
18:51 < rxr> jeru: are you really really sure you are at HEAD=
18:51 < rxr> svn up does not fetch anything =
18:52 < Postal> rxr, will ccache speed up the build much?
18:52 < jeru> erased e2fsprogs and synced again ... just a moment
18:52 < rxr> Postal: depends ...
18:53 < Postal> chroot: cannot run command `bin/bash': No such file or directory
18:53 < rxr> if you do a normal build wihtout stage9 it will not speed it up much
18:53 < rxr> but I think it should still speed it up my some percent ...
18:53 < Postal> yeah, well at this point, i would like to get to stage 2 lol
18:53 < rxr> ouhm ...
18:54 < Postal> you are sure that you commited the fix that we used last time correct?
18:54 < rxr> yes it sould be in ...
18:54 < rxr> but wait ...
18:54 < rxr> damn I did another change in the meantime ...
18:54 < rxr> do you have lib64 in your build/system-*.../ directory ?
18:55 < Postal> no just lib
18:55 < jeru> rxr: doesn't build
18:56 < rxr> jeru: do you have hyper threading ?
18:57 < jeru> it's a P4M so it should have HT
18:57 < rxr> you run a SMP kernel that takes advantage of HT then?
18:57 < jeru> flags : fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe cid xtpr
18:58 < rxr> and two CPUs in /etc/cpuinfo ?
18:58 < rxr> proc even
18:58 < jeru> no ... I had problems with it long time ago so it's not enabled right now ... should I do so?
18:58 < rxr> no - I was just curious ...
18:59 < rxr> an SMP kernel and SMP (or a like) system would be the reason why you see parallel issues more easily and reproducable ...
18:59 < jeru> no ... it doesn't display to cpu's
18:59 < jeru> I mean two
19:00 < jeru> :)
19:06 < Postal> rxr, any ideas on the chroot?
19:07 < rxr> what does this list:
19:07 < rxr> ls build/system-*/lib/ld* ?
19:08 < Postal> build/system-2.1.0-beta4-x86-64-generic-expert/lib/ld-2.3.4.so
19:08 < Postal> build/system-2.1.0-beta4-x86-64-generic-expert/lib/ld-linux-x86-64.so.2
19:08 < Postal> the first one is the 32bit and the second the 64bit..?
19:09 < rxr> what does this yield:
19:10 < rxr> strings build/system-*/bin/bash | grep /ld
19:11 < Postal> /lib/ld-linux.so.2
19:12 < rxr> so - this does not match that well ...
19:13 < Postal> yeah i suppose not
19:14 < Postal> would be best to link it to ld-linux-x86-64.so.2 no?
19:14 < rxr> Postal: I wanted to do some cleanups in this area anyway ...
19:14 < rxr> do you kill me if you commit an cleanup and tell you to rebuild ;-) ?
19:14 < Postal> I suppose not
19:14 < Postal> ;-)
19:15 < Postal> but you had better make damn sure it works
19:15 < rxr> haha
19:15 < Postal> cause at some point here i would like a t2 desktop system
19:15 < Postal> ;-)
19:17 < rxr> ok - run a scripts/Cleanup -ful ...
19:17 < rxr> +l
19:18 < Postal> running now
19:18 < Postal> ok
19:18 < Postal> done
19:19 < rxr> ok - svn up and give it a go
19:19 < Postal> ok
19:20 < Postal> this works right cause it will take about 3 hours to get to this point again
19:20 < rxr> you should get standard conforming lib64 now - and the toolchain is patched less
19:20 < rxr> I already wanted to do this - but wanted to wait until we have a working x86-64 system
19:20 < rxr> but since trunk:HEAD was broken again - it is a good time to do this transition
19:20 < Postal> mine is working.. for the moment..
19:21 < rxr> if I did not do a typo it should bulid well
19:21 < rxr> you are already in t2 ?
19:21 < Postal> no lol
19:21 < Postal> suse
19:21 < Postal> and im wishing i was in t2
19:22 < Postal> will any of the packages i dl before still be there?
19:22 < rxr> yes
19:23 < Postal> good
19:23 < Postal> otherwise i really might just reach through my computer screen and kill you lol
19:23 < rxr> jeru: can you retest e2fsprogs ?
19:23 < rxr> with HEAD of course ...
19:24 < rxr> CIA is kinda slow today ...
19:25 < rxr> mnemoc: what was the plan for KGB ?
19:25 < Postal> building now
19:25 < Postal> I think he said that we should code our own svn bot
19:25 < Postal> ;-)
19:31 < jeru> rxr: I'll test ... and of course with Head ;)
19:32 < rxr> yeah - ,-)
19:33 < Postal> rxr, its a good thing for you that i dont have much of a social life ;-)
19:34 < jeru> rxr aka mr. patch-o-matic: ... you're sure with the name 'biarch' not 'bitch' ;)
19:34 < Postal> lol
19:35 < jeru> rxr: what do you mean with social life ?
19:35 < Postal> jeru, that was me, and it means i dont get out much lol
19:35 < jeru> ahhhh ... my eyes ;)
19:36 < Postal> yeah its kinda like that
19:36 < jeru> now I understand why you're laughing about 'bitch' ;)
19:36 < Postal> well, were in stage 1 with no errors rxr
19:37 < Postal> should be about 1 to 2 hours to complete stage 1
19:37 < jeru> hmm ... still the same error ...
19:38 < Postal> hmm
19:39 < jeru> postal: does it work on your machine?
19:39 < Postal> well i should be building it in about 45 min to an hour, but it worked fine after rxr patched it the last tine
19:39 < Postal> time*
19:39 < jeru> that's strange ...
19:44 < Postal> well i have to build glibc and the two kernels that will take about 1 hour, then its all quick from there... so ill let you know if e2fsprogs builds in about 1:30 min
19:48 < rxr> Postal: did you svn up to get my latest e2fsprogs patch?
19:48 < rxr> jeru: did it built for you?
19:49 < Postal> rxr, i svn up about 25 min ago
19:49 < Postal> give or take
19:49 < Postal> cause you said Postal, svn up and go
19:50 < jeru> I did svn up ... but it doesn't build
19:51 < rxr> ouch
19:52 < rxr> flag it NOPARALLEL remove the patch and go on
19:52 < rxr> damn autoconfiscated makefile ...
19:52 < Postal> LOL
19:52 < jeru> hmmm ... sorry rxr ... for this xtra work :/
19:54 < Postal> i just built glibc in 15min 37 sec
19:54 < Postal> ;-)
19:57 < _Ragnar_> moin
20:00 < _Ragnar_> hmmm
20:00 < rxr> re
20:00 < jeru> moin
20:00 < _Ragnar_> nvidia seems to fail 2.6 builds with the current HEAD
20:00 < rxr> ih _Ragnar_ !
20:01 < _Ragnar_> Running main build function 'build_this_package' ...
20:01 < _Ragnar_> Creating /usr/<..> if required ...
20:01 < _Ragnar_> Running make ARCH=i386 CROSS_COMPILE= KCC=kcc CC=gcc CPP=cpp CXX=g++ prefix=/usr.
20:01 < _Ragnar_> make: *** No targets specified and no makefile found. Stop.
20:01 < _Ragnar_> hey rene:)
20:01 < rxr> _Ragnar_: alejandro also reported this ...
20:01 < _Ragnar_> ah okay
20:01 < rxr> mnemoc said custmain is ignored ...
20:01 < rxr> it was not when I packaged it ...
20:03 < Postal> rxr, we have gnupod in the repository, thats quite nice
20:03 < _Ragnar_> seems to not unpack the nvidia package at all
20:04 < rxr> as mnemoc said - custmain seems to be completely ignored ...
20:04 < _Ragnar_> yea
20:04 < rxr> are other 3rd party kernel packaged build and work - or is nvidia just the first beeing built anyway ?
20:05 < rxr> maybe s.th. general in the post pkg build iteration got damanged ...
20:05 < _Ragnar_> negative, fuse works
20:06 < rxr> ah
20:06 < rxr> what is the build order? fuse -> nvidia? or s.th. in between =
20:06 < _Ragnar_> (whatever that is) :)
20:06 < _Ragnar_> it aborts with nvidia, so that's the last
20:10 < _Ragnar_> fuse has no custmain tho
20:14 < _Ragnar_> rxr: I see no custmain setting in postlinux.conf for nvidia
20:14 < _Ragnar_> just autoextract=0
20:15 < rxr> good point ...
20:15 < rxr> let me review the pkg history and rever that intrusive delete
20:15 < _Ragnar_> :D
20:18 < rxr> well - committed
20:21 < jeru> I'm out for some time ... critter and me need some food ... see you later :-)
20:21 < _Ragnar_> =) have a good one
20:24 < Postal> cu jeru
20:42 < Postal> rxr, do you think I can get up to build level 4 today?
20:43 < rxr> where are you know ?
20:43 < Postal> 1-kiss
20:43 < Postal> but again, it only took me 15 min to build glibc
20:43 < Postal> and about 22 for the 2.6 kernel
20:43 < rxr> yeah - nice box you have
20:44 < Postal> i build it myself
20:44 < Postal> built*
20:44 < Postal> good for testing LOTS of packages
20:44 < rxr> yep
20:44 < Postal> ;-)
20:44 < rxr> takes my U30 Ultra SPARC @ 290 Mhz or so 3+ hours to build glibc ...
20:44 < Postal> wow, i dont know what i would do
20:45 < rxr> sleep ;-)
20:45 < Postal> how many days does it take you to build generic
20:45 < rxr> start the glibc test in the night - review the result in the morning ...
20:45 < rxr> weeks++
20:45 < Postal> OMG
20:45 < rxr> but I would not build a generic on it - only my lighter desktop / workstation subset ...
20:45 < Postal> ic
20:45 < rxr> windows is touchy - my o my ...
20:46 < Postal> yup, i dont use it
20:46 < rxr> 3 reboot and 5 scanner power cycle and still no scan done ...
20:46 < Postal> wow
20:46 < Postal> I thought it was for MAC
20:46 < rxr> well - I have to get some usb i/o bits out of it for my driver ...
20:46 < rxr> yep - but since the scanner SPEC is so fuzzy I reverse engeneer some bits out of the windows bianry ...
20:47 < Postal> ic
20:47 < Postal> sound like fun...
20:47 < rxr> except windows and mac it is ...
20:47 < Postal> osx is not that bad
20:48 < Postal> its better than windows
20:49 < Postal> I will be sooo happy when this build finishes
20:50 * rxr too
20:50 < Postal> youll be happy when my build finishes?
20:51 < rxr> yes - because I would know t2 builds for x86-64 ...
20:51 < Postal> well, we already know it builds...
20:51 < Postal> we just dont know if it will build a full generic
20:53 < Postal> 1-mine has a NOPARALLEL on it?
20:54 < Postal> by the way, whatever you did to e2fsprogs, its broken now..
20:54 < rxr> it only build accidently for you last itme ...
20:55 < Postal> i am just going to inject a NOPARALLEL
20:55 < rxr> wait
20:55 < rxr> I commit it officially right now
20:55 < Postal> ok
20:55 < Postal> commit what the NOPARALLEL?
20:56 < rxr> yep
20:56 < rxr> so - done
20:56 < Postal> lol
20:57 < Postal> ok building
20:57 < Postal> ok it built fine
21:00 < Postal> i think its about to chroot
21:00 < Postal> its on sysfiles
21:01 < Postal> chroot: cannot run command `bin/bash': No such file or directory
21:02 < rxr> ok - do you have lib64 in the build/* ?
21:02 * jeru is back
21:02 < Postal> let me check
21:02 < Postal> no
21:02 < Postal> and i know i svn up
21:02 < rxr> ...
21:03 < rxr> ok - forgot s.th.
21:03 < Postal> yeah
21:03 < Postal> :-\
21:03 < rxr> remove build/*/var/adm/logs/*glibc*
21:04 < jeru> still in e2fsprogs loop?
21:04 < Postal> rxr, ok done
21:04 < rxr> svn up
21:04 < rxr> and run Build-Target
21:04 < jeru> oh yeah
21:04 < Postal> ok
21:05 < Postal> its building
21:05 < Postal> it should take 30 min to build both
21:05 < Postal> cause i assume it builds twice once in 0 and again in 1
21:06 < rxr> 0 it fast - just header extraction
21:07 < Postal> i know it took less than a minute
21:12 < jeru> rxr: you're the man with the master plan ... e2fsprogs builds now succesfully :-)
21:13 < rxr> jeru: I just disabled parallel builds ...
21:13 < rxr> because properly fixing it proofed to be too difficult with the messed up Makefile ...
21:13 < rxr> so I threw the NOPARALLEL sledgehammer on it ...
21:14 < jeru> ok ... "retraction" ... ;)
21:14 < jeru> hadn`t recognized the changes you did
21:16 < Postal> rxr, i sure hope this fixes the problem
21:16 < jeru> don't hope ;)
21:17 < Postal> diffrent problem..
21:17 < jeru> just jokin
21:17 < Postal> i know
21:18 < Postal> chroot: cannot run command `bin/bash': No such file or directory
21:18 < Postal> still no lib64
21:19 < jeru> yep my syne target is almost complete
21:19 < jeru> :)
21:19 < Postal> i wish I could say the same for mine
21:19 < jeru> don't worry
21:20 < Postal> im not worried, just frustrated lol
21:20 < jeru> building the destop target takes a lot of cpu cycles
21:20 < Postal> yup
21:20 < _Ragnar_> yeah
21:20 < Postal> 3 days on a 3ghz amd athlon64
21:20 < Postal> according to the website
21:21 < jeru> but with p4m 2GHz it took also 3 days
21:21 < Postal> will then ther you go
21:22 < Postal> i thought the numbers were off but
21:22 < Postal> i cannot prove it until we get this chroot problem fixed
21:22 < Postal> rxr, any ideas?
21:23 < jeru> unfortunately I cannot test it on an AMD64 :(
21:23 < Postal> I can ;-D
21:23 < jeru> you lucky one
21:23 < Postal> had to work alot of extra hours, but it was worth it
21:24 < rxr> Postal: hm ...
21:24 < rxr> cd build/...
21:24 < rxr> find -name ld.*
21:25 < rxr> sorry people /me a bit in the background, departing and fixing an automtic document feeder unit ...
21:25 < Postal> ./system-2.1.0-beta4-x86-64-generic-expert/etc/ld.so.conf
21:25 < Postal> ./system-2.1.0-beta4-x86-64-generic-expert/usr/share/info/ld.info
21:25 < Postal> ./system-2.1.0-beta4-x86-64-generic-expert/usr/share/man/man1/ld.1
21:27 < rxr> find -name ld-*
21:27 < rxr> :-)
21:28 < Postal> ./system-2.1.0-beta4-x86-64-generic-expert/lib/ld-2.3.4.so
21:28 < Postal> ./system-2.1.0-beta4-x86-64-generic-expert/lib/ld-linux-x86-64.so.2
21:28 < Postal> ./system-2.1.0-beta4-x86-64-generic-expert/usr/lib/ld-2.3.4.so
21:28 < Postal> ./system-2.1.0-beta4-x86-64-generic-expert/usr/lib/ld-linux-x86-64.so.2
21:28 < rxr> oehm ?
21:29 < Postal> I feel a little bit like im bannging my head into a brick wall
21:32 < rxr> ok - I fix it manually for now - and crawl in the glibc makefiles later ...
21:32 < Postal> i like this idea
21:32 < Postal> i really dont want to build it again
21:32 < Postal> ;-)
21:32 < rxr> well - you need to rebuild glibc - but not more
21:32 < rxr> and only remove the 1-glibc.log ...
21:32 < rxr> not the 0 one ...
21:33 < Postal> ok
21:37 < rxr> ok - give it a try
21:37 < rxr> I hope it contains no syntax errors ...
21:37 < rxr> untested ...
21:37 < CIA-9> rene * r6428 /trunk/package/base/glibc/glibc.conf:
21:37 < CIA-9> * teached glibc.conf about slibdir what glibc should manage alone
21:37 < CIA-9> (review later)
21:39 < Postal> ok building
21:40 < mnemoc> re
21:40 < Postal> rxr, so you think this should work?
21:46 < rxr> yes why?
21:46 < Postal> no reason
21:46 < Postal> its taking a long time to build
21:46 < Postal> longer than usual
21:47 < jeru> hi mnemoc ... back on track ;)
21:47 * mnemoc hates rene
21:48 < mnemoc> Scheduling package 'gcc' for rebuild ...
21:48 < mnemoc> Scheduling package 'glibc' for rebuild ...
21:48 < Postal> ..?
21:48 < mnemoc> hi jeru
21:48 < Postal> hi mnemoc
21:49 < mnemoc> hi Postal64
21:49 < Postal> just Postal is fine lol
21:49 < jeru> mnemoc: my syne target is almost complete :) preparing UML for testing
21:50 < mnemoc> :D
21:51 < _Ragnar_> rxr: still no go for nvidia
21:52 < _Ragnar_> Running main build function 'build_this_package' ...
21:52 < _Ragnar_> Creating /usr/<..> if required ...
21:52 < _Ragnar_> /TOOLCHAIN/src.linux26.1108326587.8873.fad8824d/archdir/: /TOOLCHAIN/src.linux26.1108326587.8873.fad8824d/archdir/: is a directory
21:53 < mnemoc> linux26 runs build_this_package but nvidia should use build_package
21:54 < Postal> bin/bash: error while loading shared libraries: libncurses.so.5: cannot open shared object file: No such file or directory
21:54 < Postal> grrr
21:54 < rxr> ok - that might be an generic error introduced lately ...
21:55 < rxr> what does find tell you?
21:55 < rxr> find libncurses.*
21:55 < rxr> in build/...
21:58 < Postal> find: libncurses.*: No such file or directory
21:58 < mnemoc> Postal: was it building ncurses?
21:59 < Postal> no
21:59 < Postal> glibx
21:59 < Postal> glibc*
21:59 < mnemoc> uhm
21:59 < rxr> Postal: do you have build/.../var/adm/logs/1-ncurses.* ?
22:00 < Postal> yes
22:00 < Postal> 1-ncurses.log
22:00 < rxr> cd build/...
22:01 < rxr> find -name libncurses*
22:01 < _Ragnar_> & escape the *
22:01 < rxr> nah - no need
22:01 < Postal> ./system-2.1.0-beta4-x86-64-generic-expert/lib/libncurses.so.5.4
22:01 < Postal> ./system-2.1.0-beta4-x86-64-generic-expert/lib/libncurses.so.5
22:01 < Postal> ./system-2.1.0-beta4-x86-64-generic-expert/lib/libncurses.so
22:01 < Postal> ./system-2.1.0-beta4-x86-64-generic-expert/lib/libncursesw.so.5.4
22:01 < Postal> ./system-2.1.0-beta4-x86-64-generic-expert/lib/libncursesw.so.5
22:01 < Postal> ./system-2.1.0-beta4-x86-64-generic-expert/lib/libncursesw.so
22:01 < Postal> ./system-2.1.0-beta4-x86-64-generic-expert/usr/lib/libncurses.a
22:01 < Postal> ./system-2.1.0-beta4-x86-64-generic-expert/usr/lib/libncurses_g.a
22:01 < Postal> ./system-2.1.0-beta4-x86-64-generic-expert/usr/lib/libncursesw.a
22:01 < Postal> ./system-2.1.0-beta4-x86-64-generic-expert/usr/lib/libncursesw_g.a
22:02 < rxr> are ou in build/system-bla-blubs?
22:02 < Postal> no sorry just build.. let me cd into system and do it again
22:03 < Postal> ./lib/libncurses.so.5.4
22:03 < Postal> ./lib/libncurses.so.5
22:03 < Postal> ./lib/libncurses.so
22:03 < Postal> ./lib/libncursesw.so.5.4
22:03 < Postal> ./lib/libncursesw.so.5
22:03 < Postal> ./lib/libncursesw.so
22:03 < Postal> ./usr/lib/libncurses.a
22:03 < Postal> ./usr/lib/libncurses_g.a
22:03 < Postal> ./usr/lib/libncursesw.a
22:03 < Postal> ./usr/lib/libncursesw_g.a
22:03 < rxr> ok
22:03 < rxr> cat etc/ld.so.conf
22:04 < Postal> /usr/lib64
22:04 < Postal> /usr/lib
22:04 < Postal> /usr/X11/lib
22:04 < Postal> /usr/local/lib
22:04 < Postal> /opt/*/lib
22:04 < rxr> file ./lib/libncurses.so.5.4
22:05 < mnemoc> plain/text
22:05 < rxr> ?
22:05 < _Ragnar_> ^_^
22:05 < Postal> ./lib/libncurses.so.5.4: ELF 64-bit LSB shared object, AMD x86-64, version 1 (SYSV), stripped
22:07 < rxr> does it work when you ad /lib to ld.so.conf ?
22:07 < Postal> ...
22:07 < mnemoc> rxr: ===[build_package:992 (last $?=0)> . /TOOLCHAIN/package/x11/nvidia/postlinux.conf
22:08 < mnemoc> ====[source:33 (last $?=0)> autoextract=0
22:08 < mnemoc> ===[build_package:997 (last $?=0)> '[' '' = true ']'
22:08 < mnemoc> rxr: that's why i said it _ignores_ custmain
22:08 < Postal> rxr, you lost me
22:08 < mnemoc> Postal: echo /lib >> etc/ld.so.conf
22:09 < rxr> mnemoc: it had no custmain at all
22:09 < rxr> I accidently removed it ...
22:09 < rxr> I just readded it minutes ago ...
22:09 < mnemoc> o_O
22:09 < Postal> nothing was outputted by the command
22:09 < mnemoc> Postal: nothing to output, now try again
22:10 < mnemoc> try again your build
22:10 < rxr> Postal: the echo was redirecto into etc/ld.so.conf ;-)
22:10 < Postal> ah i c
22:10 < rxr> Postal: does chroot .
22:10 < rxr> work now ?
22:10 < Postal> mom
22:10 < Postal> no
22:10 < Postal> same error
22:11 < mnemoc> Postal: chroot /sbin/ldconfig
22:12 < mnemoc> Postal: chroot . /sbin/ldconfig
22:12 < mnemoc> :p
22:13 < Postal> it wont let me do it says no such file or directory..
22:13 < Postal> and ive typed it tree times now so i know its correct
22:14 < rxr> Postal: lib64/libc.so.6 does exist?
22:14 < Postal> dont know let me find out
22:15 < Postal> yes
22:16 < rxr> ok: chroot . lib64/libc.so.6
22:17 < mnemoc> isn't ldconfig static??
22:17 < Postal> it wont let me chroot it says its not a directory
22:17 < rxr> yep it is
22:19 < rxr> strings lib64/libc.so.6 | grep lib/
22:19 < Postal> usr/lib/gconv
22:19 < Postal> /usr/lib/locale
22:19 < Postal> /usr/lib/locale/locale-archive
22:19 < Postal> /usr/lib/gconv/gconv-modules.cache
22:20 * Postal bashes head into wall (repeatedly)
22:20 < rxr> strings lib64/libc.so.6 | grep lib64/
22:20 < rxr> Postal: hey - just relax
22:20 < rxr> you have nearly all the time of the world - a lot of fun - and learn awesome amounts of linux internals ...
22:20 < Postal> I am quitting smoking... im tense because of that not this...
22:20 < Postal> I enjoy this
22:21 < Postal> hmm
22:21 < Postal> "/lib64/ld-linux-x86-64.so.2
22:21 < Postal> "
22:21 < Postal> there
22:22 < rxr> that file exists ?
22:22 < Postal> apparently
22:22 < Postal> well should it not?
22:22 < rxr> and when you run the same strings ... grep /lib64 over the bash binary?
22:23 < rxr> sure it should
22:23 < mnemoc> echo /lib64 >> etc/ld.so.conf
22:23 < Postal> ok
22:23 < Postal> thanks ;-)
22:23 < Postal> done
22:24 < rxr> does it change anything ?
22:24 < mnemoc> Postal: chroot . /sbin/ldconfig
22:24 < Postal> let me check mom
22:25 < rxr> the path is hardcoded into the binary it shoud not matter ...
22:25 < mnemoc> true, but i'm puzzled
22:25 < rxr> AFAIR it complained about ncurses
22:25 < Postal> well it did not output anything
22:25 < rxr> maybe something wrong with ncurses ...
22:25 < rxr> Postal: I think no output is ok ...
22:25 < Postal> prehaps it chrooted, how do I check?
22:25 < rxr> what does this now:
22:26 < rxr> chroot . /lib64/ld-linux-x86-64.so.2
22:26 < Postal> it gives a usage error
22:26 < mnemoc> rxr: previous 'chroot . /sbin/ldconfig' gave him 'no such file or directory' that's why i wanted to check
22:27 < CIA-9> jeru * r6429 /trunk/target/mnemosyne/pkgsel/Services/Network/ (. bridge.ask): * added bridging firewall support
22:27 < mnemoc> chroot . /lib64/ld-linux-x86-64.so.2 /lib/libncurses.so.5.4
22:28 < mnemoc> errr
22:28 < Postal> chroot: cannot run command `/lib64/ld-linux-x86-64.so.2/lib/libncurses.so.5.4': Not a directory
22:28 < Postal> whoops
22:28 < Postal> forgot a space
22:29 < mnemoc> :)
22:29 < mnemoc> iirc you may need a --list in between
22:29 < Postal> Segmentation fault :-\
22:29 < mnemoc> or --verify
22:30 < mnemoc> chroot . /lib64/ld-linux-x86-64.so.2 --list /lib/libncurses.so.5.4
22:30 < Postal> libc.so.6 => /lib64/libc.so.6 (0x0000002a956ae000)
22:30 < Postal> /lib64/ld-linux-x86-64.so.2 (0x000000552aaaa000)
22:30 < mnemoc> :D
22:31 < mnemoc> rxr: see? :p /lib64 was needed
22:31 < Postal> this means.?
22:31 < mnemoc> you have a working libncurses.so.5.4 now
22:31 < rxr> does chroot . work now ?
22:32 < Postal> can i just run Build-Target?
22:32 < mnemoc> chroot . is a faster test
22:32 < Postal> well, i did it with Build-Target, and its building stage 2
22:32 < Postal> ;-)
22:33 < mnemoc> :D
22:33 < Postal> I dont know what i would do without you guys THANKS
22:33 < mnemoc> you are welcome
22:34 < Postal> Thats nifty, it tells me what percent Cache has been hit so far in chroot
22:35 < mnemoc> 0%?
22:35 < Postal> 5.62
22:35 < mnemoc> :p
22:35 < Postal> seems like an odd # dosent it 5.62
22:36 < Postal> Man other distros WISH they had this kind of tech support ;-)
22:36 < _Ragnar_> *lol*
22:37 < mnemoc> :)
22:40 < rxr> so - _Ragnar_ what is the acutal nvidia postlinux.conf error ?
22:41 < _Ragnar_> ooh
22:42 < _Ragnar_> 1 mom
22:49 < Postal> well things are going quite nicley in stage 2 :D
22:56 < mnemoc> :)
22:57 < Postal> it is not doing parallel make anymore...
22:58 < mnemoc> uhm?
22:58 < rxr> sure it should
22:58 < rxr> (if you not disable it)
22:58 < Postal> well, its not
22:58 < rxr> some package shave this disabled ...
22:58 < Postal> i have not touched it
22:58 < Postal> it hasnt run for any in stage 2 or 3
22:58 < Postal> oh wait thats a lie
22:59 < Postal> it ran for libelf
22:59 < Postal> never mind
23:20 < CIA-9> amery * r6430 /trunk/package/network/udhcp/ (udhcp.desc udhcpd-foreground.patch): * added foreground option to udhcpd, and fixed udhcp [D] to not use checksum
23:21 < jeru> mnemoc: can you try to build ebtables? ...post-install adaptions doesn't function anymore on my machine
23:21 < jeru> Making post-install adaptions.
23:21 < jeru> !> Found files in /usr/local:
23:21 < jeru> !> /usr/local/man
23:21 < jeru> !> /usr/local/man/man8
23:21 < jeru> !> /usr/local/man/man8/ebtables.8
23:21 < rxr> Postal: you build continues fine ?
23:22 < jeru> mnemoc: or can you tell me what's the problem, that I can try to fix it
23:25 < mnemoc> jeru: the packages is not using the right prefix
23:26 < mnemoc> jeru: read the makefile, you may need to var_append makeinstopt ' ' 'DESTDIR=...' or PREFIX or stuff like that
23:27 < jeru> I think I've found the problem
23:27 < mnemoc> what was it?
23:28 < Postal> kdosfs.c:107:3: #error _llseek system call not present
23:28 < Postal> mkdosfs.c: In function `_llseek':
23:28 < Postal> mkdosfs.c:109: error: `__NR__llseek' undeclared (first use in this function)
23:28 < Postal> mkdosfs.c:109: error: (Each undeclared identifier is reported only once
23:28 < Postal> mkdosfs.c:109: error: for each function it appears in.)
23:28 < Postal> make[1]: *** [mkdosfs.o] Error 1
23:28 < Postal> make[1]: Leaving directory `/TOOLCHAIN/src.dosfstools.1108333556.17576.0x7f0200/dosfstools-2.9/mkdosfs'
23:28 < Postal> make: *** [all] Error 2
23:28 < jeru> mnemoc: I had to add prefix=$xroot/usr
23:28 < mnemoc> where?
23:28 < _Ragnar_> rxr:
23:28 < _Ragnar_> =====[nvidia_main:13 (last $?=0)> sh /TOOLCHAIN/src.linux26.1108331695.25718.fad8824d/archdir/ --extract-only
23:28 < jeru> I think ebtables.conf
23:28 < _Ragnar_> /TOOLCHAIN/src.linux26.1108331695.25718.fad8824d/archdir/: /TOOLCHAIN/src.linux26.1108331695.25718.fad8824d/archdir/: is a directory
23:28 < mnemoc> jeru, no never
23:28 < jeru> I will commit it
23:29 < jeru> why?
23:29 < mnemoc> prefix has a differnt meaning on .conf
23:34 < rxr> _Ragnar_:
23:34 < rxr> - sh $archdir/$1 --extract-only
23:34 < rxr> + sh $archdir/$f --extract-only
23:35 < rxr> I commit it - it is obvious
23:35 < mnemoc> :)
23:36 < rxr> _Ragnar_: could you retest ;-)
23:36 < CIA-9> rene * r6431 /trunk/package/x11/nvidia/postlinux.conf: * the next fix for nvidia to build further
23:39 < Postal> rxr, any ideas on the dosfstools error?
23:40 < mnemoc> Postal: that's due to kernel's insanity
23:40 < Postal> so...
23:40 < mnemoc> you have to patch linux-header
23:40 < Postal> how?
23:40 < mnemoc> mainly adding #ifdef __KERNEL__
23:41 < mnemoc> on proper locations
23:41 < mnemoc> do you need dosfstools?
23:41 < Postal> no\
23:42 < mnemoc> pkgsel:- dosfstools
23:42 < _Ragnar_> rxr: okay
23:45 < _Ragnar_> that'll take a while ... I'll go shopping in between
23:48 < CIA-9> jeru * r6432 /trunk/package/security/ebtables/ebtables.conf: * cleaned up / adjusted package conf file
23:56 < Postal> stage 5 ;-)
23:56 < mnemoc> :D
23:57 < Postal> This is the big one right?
23:57 < mnemoc> yep
23:58 < Postal> how long do we estimate?
23:58 < Postal> 6 hours?
23:58 < mnemoc> no idea, 2 days here for a full generic :p
--- Log closed Mon Feb 14 00:00:01 2005