T2 IRC Log: 2005-01-20

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 Jan 20 00:00:47 2005
00:04 < _Ragnar_> rehi
00:05 < mnemoc_> hi _Ragnar_
00:05 < _Ragnar_> hi mnemoc_
00:29 < rxr> cu all
00:30 < _Ragnar_> cy
01:17 < _Ragnar_> mnemoc_: 2-boehm-bc fails for me
01:23 * _Ragnar_ disables for now
01:24 < mnemoc_> i need to do a full mono/boehm-gc reference build one of these days
01:58 < CIA-9> mnemoc * r5707 /trunk/package/base/uclibc/patches/man.conf: * fixing uclibc/man forcing DISABLE_NLS
02:10 < mnemoc_> CIA-9: ?
02:10 -!- ideal [~idealm@211.100.227.107] has joined #t2
02:44 < mnemoc_> == 01/20/05 01:44:16 =[3]=> Finished building package pam. :D
02:46 -!- sparc-kly [~sparc-kly@65-23-206-237.prtc.net] has joined #t2
02:46 < mnemoc_> but i'm not sure if it's operative :p
02:52 * _Ragnar_ has - pam in his pkgsel
02:53 < mnemoc_> my hack to remove nis is very weak
02:59 -!- rxr_ [~rene@p83.129.4.226.tisdip.tiscali.de] has joined #t2
02:59 -!- 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!
02:59 -!- Topic set by mnemoc [] [Wed Jan 12 03:04:43 2005]
02:59 [Users #t2]
02:59 [ _Ragnar_] [ ideal] [ mnemoc_] [ praenti] [ rxr_ ] [ valentin]
02:59 [ daja77 ] [ jsaw ] [ nzg ] [ rxr ] [ sparc-kly]
02:59 -!- Irssi: #t2: Total of 11 nicks [0 ops, 0 halfops, 0 voices, 11 normal]
02:59 -!- Channel #t2 created Sun Aug 8 21:15:33 2004
03:00 -!- Irssi: Join to #t2 was synced in 16 secs
03:00 < _Ragnar_> huh?
03:00 < _Ragnar_> I doubt it
03:01 < mnemoc_> i hate auto*
03:02 < _Ragnar_> *g* sorry gotta head off ... be back in an hour hopefully
03:02 < mnemoc_> ok
03:04 < mnemoc_> rxr: my postconf match to ncurses' Makefile don't apply at least to _Ragnar_ :(
03:05 -!- rxr [~rene@p83.129.13.82.tisdip.tiscali.de] has quit [Read error: 60 (Operation timed out)]
03:05 < mnemoc_> rxr: it failed to me on stage 1 too
03:07 < mnemoc_> rxr_: me hacking the hack
03:10 -!- CIA-temp448 [~CIA@flapjack.navi.cx] has joined #t2
03:10 -!- CIA-temp448 is now known as CIA-9
03:10 < CIA-9> mnemoc * r5711 /trunk/package/base/ncurses/ncurses.conf: * hacked ncurses not removing libs hack to not apply on stage1 (this shall be done on aclocal.m4)
03:11 < mnemoc_> wb CIA
03:12 < mnemoc_> i know, that's pathetic but i don't feel i can patch aclocal.m4 today and we need to keep this working
03:48 < CIA-9> mnemoc * r5712 /trunk/package/base/uclibc/patches/psmisc-add_libintl.patch: * fixed psmisc to build against uclibc adding -lintl dependency
03:48 < mnemoc_> Error logs from minimal-2.1.0-beta3-x86-pentium-32-mnemosyne-expert:
03:48 < mnemoc_> 144 builds total, 77 completed fine, 0 with errors.
03:48 < mnemoc_> using uclibc of course
04:02 < mnemoc_> aaaaaaaaaaaarg!
04:03 < mnemoc_> i got on 5-uclibc the same we had with 3-ncurses, install_wrapper and dead bash
04:04 < mnemoc_> 144 builds total, 79 completed fine, 1 with errors.
04:04 < mnemoc_> :'(
04:07 < _Ragnar_> back
05:17 < rxr_> re
05:18 < jsaw> hehe - shifted rythm
05:18 < _Ragnar_> re
05:18 < jsaw> hi rxr_ , _Ragnar_
05:18 < rxr_> hi jsaw
05:20 -!- You're now known as rxr
05:20 < jsaw> rxr: what do you think, if we handle downloaded patches in [D] tags generally with detect_patches?
05:23 < rxr> jsaw: you want to cleanup some packages ?
05:25 < jsaw> that's of course the draw back - cleaning up packages...
05:25 < jsaw> I just thought it, e.g. xpdf.
05:25 < jsaw> I add the patches up to patch level three. Instead of downloading the complete package.
05:25 < jsaw> ...adding this var_append patchfiles looks a bit upgly....
05:26 < jsaw> similiar to bash... and readline...
05:27 < jsaw> btw, how's sparc build going?
05:29 < rxr> just fixing it ...
05:30 < rxr> I chrooted into the first crossbuild and first had to build perl in it to get glibc going ...
05:30 < rxr> the bug from yesterday
05:30 < rxr> I addded perl to stage 1 now also - seems to have build - need to review if it cross build correctly ...
05:30 < rxr> now I'm hacking on getting gcc build with all languages
05:31 < rxr> there is some assembler complaint I'm currently try to investigate
05:31 < rxr> oh - i'm a moron ..
05:33 < _Ragnar_> rxr: perl didn't build for me for 64
05:33 < rxr> where did you built it?
05:33 < _Ragnar_> cross-build
05:33 < rxr> building it natively in my new 64bit chroot it built
05:33 < rxr> you mean cross build it from by adding it to the stage1 ?
05:34 < _Ragnar_> yeah
05:35 < _Ragnar_> lemme retry tho I think mnemoc fixed something on it
05:35 < rxr> did not try that yet
05:45 < rxr> Gnumeric 1.4.2 (Stable)
05:46 < rxr> AbiWord 2.2.3
05:47 < _Ragnar_> /export/t2-trunk/build/a64-2.1.0-beta3-x86-athlon64-64-cross-generic-expert/TOOLCHAIN/tools.cross/crosscc/../lib/gcc/x86_64-unknown-linux-gnu/3.4.3/../../../../x86_64-unknown-linux-gnu/bin/ld: warning: i386 architecture of input file `libperl.a(pp_sort.o)' is incompatible with i386:x86-64 output
05:47 < jsaw> There are 23 packages downloading a patch in the [D] tag. 16 have a name with ".patch." or ".diff." in it, others are named differently (e.g. bash: bash30-000)
05:48 < _Ragnar_> (that's when building miniperl)
05:49 < jsaw> rxr: I think it's to much of a hasle (at least for now for me) to sort this out correctly.
05:49 < rxr> jsaw: sort out generical patching ?
05:51 < jsaw> yep
05:52 < jsaw> correcting 16 packages takes time...
05:52 < rxr> jsaw: place it on the TODO
05:52 < rxr> we could get back reviewing if it cleans the packages up, later ...
05:52 < jsaw> meep, ... yes
05:55 < rxr> wiki todo I meant ...
06:02 < jsaw> (I got it, already put on the wiki todo)
06:03 < jsaw> _Ragnar_: stage?
06:06 < _Ragnar_> 1
06:08 < jsaw> hmmm.
06:08 < _Ragnar_> problem here is that miniperl should be linked with the host cc, not the cross cc
06:10 < _Ragnar_> and I thought mnemoc did fix that ... weird
06:13 < jsaw> obviously not. at least the 3-ncurses thing is work-arounded
06:13 < jsaw> == 01/20/05 06:14:55 =[3]=> Finished building package ncurses.
06:15 < _Ragnar_> ;)
06:47 < jsaw> gotta go, cu later
06:47 < _Ragnar_> cya
07:12 < rxr> cu jsaw
07:19 * rxr doing the abiword and gnumeric updates ...
07:19 < rxr> so - off, driving home ... - cu
07:19 < _Ragnar_> cu
07:58 < rxr> re
07:59 < CIA-9> rene * r5713 /trunk/package/gnome2/gnumeric/gnumeric.desc: * updated gnumeric (1.4.1 -> 1.4.2)
07:59 < CIA-9> rene * r5714 /trunk/package/gnome2/abiword/abiword.desc: * updated abiword (2.2.2 -> 2.2.3)
07:59 < rxr> ouhm - gcc/java still building on the U30 ...
08:00 < rxr> all the other stuff is long finished - maybe we should disable the java fluff in it by default ...
08:00 < rxr> ;-)
08:10 < rxr> http://www-106.ibm.com/developerworks/library/pa-emulation.html?ca=dgr-lnxw07EmulatePow
08:13 < rxr> still java ...
08:37 < rxr> finally - install running ...
08:50 < _Ragnar_> rehi
08:51 < _Ragnar_> okay stage 2 in the 32/uclibc build
08:52 < CIA-9> rene * r5715 /trunk/package/x86/grub/stone_mod_grub.sh:
08:52 < CIA-9> * improved the STONE grub module to only add memtest86 to the boot
08:52 < CIA-9> menu when it is installed (inspired from ROCK submaster)
08:52 < _Ragnar_> *g*
09:00 < rxr> here is another *g*:
09:00 < CIA-9> rene * r5716 /trunk/package/archiver/bzip2/bzip2.conf:
09:00 < CIA-9> * improved bzip2 to build and install a shared library
09:00 < CIA-9> ( inspired from ROCK submaster, but
09:00 < CIA-9> * cleanup on the way
09:00 < CIA-9> * modern / cleaner code
09:00 < CIA-9> * only if ROCKCFG_STATIC is not requested )
09:01 < _Ragnar_> so ... what do we do with ROCKCFG_STATIC anyway?
09:01 < _Ragnar_> since it seems to not work on certain important packages
09:01 < rxr> well - the long time solution should be to fix those packages, no?
09:02 < _Ragnar_> are they fixable?
09:02 < rxr> well - most stuff should be
09:02 < _Ragnar_> or is it a inherent problem to for a static binary to dlopen a .so ?
09:03 < rxr> AFAIR benjamin (aka. blindcoder ovar @ rock) used it for some movie player CD where he even built X static ...
09:03 < rxr> so the status should not be that bad ...
09:03 < _Ragnar_> ah okay
09:03 < rxr> hm - I think dl_open for some staic program or library should work
09:03 < rxr> at least I see no technical reason off hand why it should not work ...
09:03 < _Ragnar_> I should try a static build with glibc ... maybe this is a uclibc bug
09:04 < rxr> btw:
09:04 < rxr> This is perl, v5.8.6 built for sparc64-unknown-linux-gnu
09:07 < _Ragnar_> ?
09:07 < rxr> just the state of my sparc64 build ....
09:09 < _Ragnar_> ah okay
09:10 < rxr> (not that I'm a bot)
09:11 < rxr> I need to re-fix strace for sparc{32,64}
09:11 < rxr> currently on both it fails with tons of syscall2 or so errors ...
09:17 < rxr> oh - /me updating linux24
09:22 < CIA-9> rene * r5717 /trunk/package/mail/exim/exim.desc: * updated exim (4.43 -> 4.44)
09:26 < _Ragnar_> *grr*
09:26 < _Ragnar_> 3-ncurses still fails
09:29 < rxr> because bash is mangled
09:29 < rxr> ?
09:30 < CIA-9> rene * r5718 /trunk/package/base/linux24/linux24.desc: * updated linux24 (2.4.29-rc1 -> 2.4.29) - untested
09:32 < _Ragnar_> well no
09:32 < _Ragnar_> cd //lib && (ln -s -f libncurses.so.5.4 libncurses.so.5; ln -s -f libncurses.so.5 libncurses.so; )
09:33 < _Ragnar_> /bin/sh: can't resolve symbol '__guard'
09:33 < _Ragnar_> so now it does an atomic move, but the new ncurses has stackguard applied, which the previous seems to kinda not have;)
09:38 < rxr> huh
09:39 < rxr> wired - did s.th. change int he SSP region while I was asleep ?
09:40 < _Ragnar_> dunno:/
09:48 < CIA-9> rene * r5719 /trunk/scripts/functions: * next generation pkgchksum code
09:48 < CIA-9> jsaw * r5720 /trunk/package/base/tcl/tcl.desc: * update tcl (8.4.6 -> 8.4.9)
09:48 < rxr> cool - sparc64 build back on track
09:48 < CIA-9> jsaw * r5721 /trunk/package/x11/tk/tk.desc: * update tk (8.4.6 -> 8.4.9)
09:48 < rxr> after the perl build in the new 64bit root
09:49 < CIA-9> jsaw * r5722 /trunk/package/gnome2/atk/atk.desc:
09:49 < CIA-9> * update atk (1.8.0 -> 1.9.0)
09:49 < CIA-9> * use gtk ftk server
09:50 < rxr> jsaw: is this stable or devel ?
09:50 < jsaw> is in the ftp://ftp.gtk.org/pub/gtk/v2.6/ directory, so stable
09:50 < rxr> hm
09:50 < rxr> ok
09:51 < rxr> do you have gtk updates pending ?
09:51 < jsaw> yep
09:51 < rxr> ;-)
09:52 < CIA-9> jsaw * r5723 /trunk/package/gnome2/glib/glib.desc: * update glib (2.6.0 -> 2.6.1)
09:53 < rxr> our commit graph of the last days is a bit crazzy ...
09:53 < jsaw> hehe
09:53 < CIA-9> jsaw * r5724 /trunk/package/gnome2/gtk+/gtk+.desc: * update gtk+ (2.6.0 -> 2.6.1)
09:54 < _Ragnar_> *hrhr*
09:54 < rxr> wow - my sparc64 build is progressing smoothly ...
09:54 < rxr> I can't believe we seem to finally get a rocking sparc64 CD set ...
09:54 < jsaw> s/rock/driv/
09:54 < jsaw> ;)
09:55 < rxr> and soon a public 100MBit connected 375Mhz UltraSPARC ... at my uni
09:55 < _Ragnar_> ;p
09:56 < rxr> seems I can carry the box home tomorrow and clone the build state from my U30 onto it the weekend ...
09:56 < rxr> btw. my bzip2 cleanup should make it utiilize the makeopt injected "-jX" option
09:56 < rxr> formerly the package reset the makeopt ...
09:59 -!- mipe [~mipe@dsl10040.japo.fi] has joined #t2
09:59 < rxr> hi mipe
10:00 < mipe> hi rxr,brb
10:00 -!- mipe [~mipe@dsl10040.japo.fi] has quit [Client Quit]
10:00 < jsaw> ...?
10:00 -!- mipe [~mipe@dsl10040.japo.fi] has joined #t2
10:02 < rxr> jsaw: what do you think about splitting the 64 bit architectures out of the 32bit ones
10:02 < rxr> (like it was in ROCK in the past, before clifford merged them)
10:02 < rxr> e.g.
10:02 < rxr> - cleaner $arch casing
10:02 < rxr> - per architecture patches
10:03 < rxr> - useful [R] tag in .desc
10:03 < rxr> - easier kernel config and disable broken in architecture/
10:03 < rxr> - more sane Config user interface - currently each 64 bit arch has other options to be enabled
10:03 < rxr> and maybe more
10:04 < rxr> drawbacks (I see of hand)
10:04 < rxr> - ~20 1/2 lines in architecture/
10:04 < rxr> comments ?
10:05 < mipe> hmm,how to disable package for certain arch,f.e. x86_64
10:05 < mipe> any flags? or filelist to insert package to
10:05 < rxr> normally in the .desc file
10:05 < rxr> [R] or the long [ARCH] +/- arch
10:05 < rxr> e.g.
10:06 < rxr> [R] - powerpc
10:06 < rxr> [R] + x86
10:06 < mipe> ah,ok
10:06 < rxr> the first only removes powerpc
10:06 < rxr> the later enables onlx x86
10:06 < mipe> but theres no need to add + flags? defaulted to +?
10:06 < rxr> mipe: oehm - no idea if there is a default ;-)
10:06 < rxr> maybe not ...
10:07 < mipe> for example - x86_64,only disables build for that arch and builds on others
10:07 < rxr> but since the 64bit arches share the $CFG_ROCKARCH with the 32bit equivalent they can currently not enabled/disabled in the .desc ... :-(
10:07 < mipe> oh
10:07 < rxr> thats one point in my proposal above ...
10:07 < rxr> 10:03 < rxr> - useful [R] tag in .desc
10:09 < mipe> yeah.have to get something to eat. after that test if gcc installs both 32 and 64 libs correctly
10:12 < rxr> what 32bit libds ?
10:12 < rxr> libs ?
10:16 < mipe> yeah both 64bit and 32bit libs to install. i tested it manually and now it installs 64bit to /lib and 32bit to /lib32, now have to test it it builds on my target.
10:17 -!- mipe is now known as mipe|AFK
10:24 < jsaw> rxr: later,
10:24 < jsaw> have to leave. cu
10:32 < rxr> cu
10:34 < CIA-9> jsaw * r5725 /trunk/package/graphic/xpdf/ (3.00pl3.patch.xpdf xpdf.desc):
10:34 < CIA-9> * add patch level 3 patch for xpdf (SECURITY)
10:34 < CIA-9> * reflect patchlevel in version number (3.0 -> 3.0pl3)
10:49 -!- mipe|AFK is now known as mipe
10:50 < _Ragnar_> well, goodnight folks
10:50 < rxr> n8 _Ragnar_
10:51 < _Ragnar_> gn8 rxr
10:51 < _Ragnar_> or good day :)
10:51 < rxr> day here ;-)
10:51 < rxr> _Ragnar_: in which country are you ?
10:51 < _Ragnar_> calif.
10:52 < rxr> ah! - sleep well!
10:52 < _Ragnar_> thx :)
11:20 < rxr> so ! finally
11:20 < rxr> sparc64 crunching stage2 without manual intervention ...
11:36 < rxr> ln: `/t2-trunk/build/system-2.1.0-beta3-sparc-v9-64-desktop-expert/TOOLCHAIN/localtime': File exists
11:36 < rxr> ^- since when do we get this warning during the build?
11:36 < rxr> see it for the first time in my sparc64 build right now ...
12:33 -!- ideal [~idealm@211.100.227.107] has quit ["Leaving"]
12:37 -!- CIA-9 [~CIA@flapjack.navi.cx] has quit []
12:43 -!- CIA-temp825 [~CIA@flapjack.navi.cx] has joined #t2
12:43 -!- CIA-temp825 is now known as CIA-8
13:04 < rxr> oh - vi just coredumped on my rs6k ...
13:04 < rxr> maybe due to out of space on disk ...?
13:04 < rxr> damn program ...
--- Log closed Thu Jan 20 13:19:09 2005
--- Log opened Thu Jan 20 13:21:00 2005
13:22 -!- CIA-9 [~CIA@flapjack.navi.cx] has quit [Remote closed the connection]
13:30 -!- CIA-temp932 [~CIA@flapjack.navi.cx] has joined #t2
13:32 < CIA-temp932> jsaw * r5730 /trunk/package/archiver/bzip2/bzip2.conf: * makeinstopt also needs the PREFIX setting
13:56 < mnemoc_> moin!
13:58 < mnemoc_> jsaw: patch i [D] are almost all optional, using `match_source_file` may clean up packages is a much better way.
14:01 < mnemoc_> rxr: i haven't touched SSP in some days (since i updated it)
14:07 * rxr neither
14:07 < rxr> no build problem on my boxes so far either
14:09 < mnemoc_> lucky you
14:10 < mnemoc_> btw, the minimal you made me does "QM_MODULES function not implemented" everytime :\ then i can't load my network module :(
14:10 < mnemoc_> thanks anyway
14:11 < rxr> hm
14:11 < rxr> maybe it has no module-init-tools ?
14:12 < mnemoc_> oh
14:12 < mnemoc_> let me power on that box...
14:12 < rxr> yep
14:12 < rxr> I commit fixes
14:13 < rxr> fixed
14:14 < rxr> thanks for the report ;-)
14:14 < CIA-temp932> rene * r5731 /trunk/misc/pkgsel/ (minimal-xorg.in minimal.in):
14:14 < CIA-temp932> * fixed minimal and minimal-xorg package preselection templates to
14:14 < CIA-temp932> include module-init-tools - untested ;-)
14:14 < mnemoc_> :p
14:14 < mnemoc_> remember that's an option not a pkgsel
14:14 < mnemoc_> an option that needs to be reworked
14:15 < rxr> an option ?
14:15 < mnemoc_> yep
14:15 < mnemoc_> with pkgenanle and pkgdisable on config.in
14:16 < rxr> ouhjm
14:19 < rxr> ouhm - I think I factor out the main part of build_this_package
14:19 < rxr> into build_main or so ...
14:19 < mnemoc_> why?
14:19 < rxr> it could be handily reused
14:20 < rxr> e.g. I just tweak ncurses to build both - narrow and wide char support
14:20 < rxr> I need that ...
14:20 < rxr> you do this:
14:20 < rxr> cd narrowc
14:20 < rxr> ../configure -bla -blubs
14:20 < rxr> cd ..
14:20 < rxr> mkdir widec ; cd widec
14:20 < rxr> ../configure -bla -blubs --wide-char (or so)
14:20 < rxr> cd ..
14:20 < rxr> make isntall
14:20 < rxr> --typos
14:21 < rxr> and of course mkdir narrowc
14:21 < rxr> in the beginning ..
14:22 < rxr> this can currently not nicely be scripted with t2 ...
14:22 < rxr> having the main function handy simplifies such packags a bit ..
14:23 < mnemoc_> beta3 or beta4? ;)
14:24 < rxr> that should be an tiny change ...
14:24 < rxr> factor it out - and call it
14:24 < rxr> and use it in ncurses
14:24 < rxr> no regressions ...
14:24 < mnemoc_> promise?
14:25 < rxr> well - not if I'm not totally stupid ...
14:25 < rxr> ok - I hack it into a postmake hook of ncurses ...
14:25 < rxr> and clean it up later ...
14:25 < CIA-temp932> mnemoc * r5732 /trunk/package/base/fbset/ (fbset.conf fbset.desc): * flaged fbset as NOPARALLEL and .conf a bit modernized
14:28 < mnemoc_> make: Entering directory `/usr/src/t2-test/src/isomd5sum'
14:28 < mnemoc_> gcc -c -O -fPIC -D_FILE_OFFSET_BITS=64 -O -Wall -Werror -o implantisomd5.o implantisomd5.c
14:28 < mnemoc_> implantisomd5.c:8:18: popt.h: No such file or directory
14:28 < mnemoc_> nice, it needs popt on host
14:29 < rxr> ohm
14:29 < rxr> are you a friend of implantisomd5 anyway?
14:29 < rxr> --disable-leaks test: suppress permanent memory-leaks
14:29 < rxr> ^- what the hell ?#
14:30 < mnemoc_> o_O
14:30 < mnemoc_> where?
14:30 < rxr> guess what?
14:30 < rxr> ncurses ...
14:30 < mnemoc_> :D
14:30 < mnemoc_> [10:30:40 AM] Creating iso/mnemosyne-minimal-pentium-r5709_cd1.iso ...
14:30 < mnemoc_> [10:30:53 AM] Done. Have fun!
14:31 < rxr> maybe we should at least hack isomd5sum to not use popt
14:31 < mnemoc_> now i have to grab it from that 64 isdn line :\
14:31 < rxr> I consider it pretty deprecated
14:31 < mnemoc_> popt or isomd5sum?
14:31 < rxr> popt
14:32 < mnemoc_> i need it anyway because keepalived use it
14:32 < rxr> and I'm not such a implantmd5sum supporter anyway ... - are you ?
14:32 < mnemoc_> T2 is for users not for developers ;) and users may want that thing
14:32 < rxr> if you think so ..
14:33 < mnemoc_> developers are users too of course
14:33 < rxr> I thought this was only for use, to point them to: run this iso check - ha! - so reburn your ISO *g* ....
14:33 < rxr> s/use/us/
14:33 < mnemoc_> :p
14:41 < rxr> cliffords reply to your fl_wrapper submaster injection is cool
14:43 < mnemoc_> uhm?
14:43 < mnemoc_> the 'ack'?
14:43 < rxr> the quesion
14:44 < rxr> +t
14:44 < mnemoc_> hehe :)
14:46 -!- CIA-temp932 [~CIA@flapjack.navi.cx] has quit [Client Quit]
14:51 < jsaw> re
14:52 < mnemoc_> hi jsaw
14:52 < jsaw> hi mnemoc_
14:53 -!- CIA-temp934 [~CIA@flapjack.navi.cx] has joined #t2
14:53 -!- CIA-temp934 is now known as CIA-9
14:53 < jsaw> this thing can't decided if it wants to stay or not...
14:53 < jsaw> -d
14:54 < mnemoc_> scanline is moving the server and playing
14:55 < jsaw> scanline?
14:55 < rxr> cia master
14:59 < rxr> huh?:
14:59 < rxr> bin directory: ///bin
14:59 < rxr> lib directory: ///lib
14:59 < rxr> include directory: ///include
14:59 < rxr> man directory: ///man
14:59 < rxr> terminfo directory: ///share/terminfo
14:59 < rxr> many slashes ...
15:00 < mnemoc_> what is that?
15:00 < rxr> ncurses build log ...
15:01 < rxr> output after configure ...
15:02 < mnemoc_> uhm
15:09 < mipe> 1-perl fails:/bin/sh: line 1: ./miniperl: Accessing a corrupted shared library
15:09 < mipe> ldd ./miniperl
15:09 < mipe> /usr/bin/ldd: line 1: ./miniperl: Accessing a corrupted shared library
15:09 < rxr> hm
15:09 < rxr> which arch and which libc ?
15:10 < mipe> x86_64
15:10 < rxr> (I start to loose the overview who is building what here ...)
15:10 < rxr> and glibc or uclibc?
15:10 < mnemoc_> too many users? ;)
15:10 < mipe> glibc
15:10 < rxr> hm :-(
15:10 < rxr> mipe: in which stage? cross or native build?
15:11 < rxr> mnemoc_: no - too many libc hacking ...
15:11 < mipe> stage 1
15:11 < mipe> trying to build with multilib enabled
15:12 < mnemoc_> jsaw: ** Detect them automatically in detect_patchfiles <--- i dislike that
15:12 < jsaw> mnemoc_: why?
15:12 < rxr> mipe: disable perl for now ... in stage-1 locally
15:12 < mnemoc_> jsaw: i download optional patches on [D]S
15:12 < mipe> rxr:ok
15:12 < rxr> mipe: is your host already running 64bit or is that a cross build ?
15:12 < mipe> already running
15:12 < rxr> hm
15:13 < jsaw> mnemoc_: diff is optional... (was always optional)
15:13 < mipe> is multilib build on stage 1 gcc?
15:13 < rxr> the gcc's built are the same - not depending on multilib ...
15:13 < mnemoc_> jsaw: can i force authors to publish their patches in this or that format?
15:13 < rxr> AFAICS my sparc64 built a 1-perl without error
15:14 < rxr> but there might be some more stuff in perl do be done different
15:14 < mnemoc_> jsaw: if you want to clean .conf use for x in `match_source_file -p ' .*\.patch '`; do
15:14 < rxr> mnemoc_: even
15:14 < mnemoc_> ?
15:15 < rxr> var_append patchfiles " " "`match_source_files ...`"
15:15 -!- mipe is now known as mipe|AFK
15:15 < rxr> no loop needed, no?
15:15 < mnemoc_> even easier :)
15:15 < mnemoc_> ack
15:15 < jsaw> mnemoc_: [D] clip-20234098.patch !url/clip-patch-20394820934.bz2
15:15 < jsaw> The point is that it is easy to forget patches applied in .conf while they are downloaded via .desc
15:16 < mnemoc_> jsaw: i do that becasue clip author keeps only one patch.tgz
15:16 < jsaw> mnemoc_: that was the answer to "can I force the author..."
15:16 < mnemoc_> match_source_files is there to help
15:16 < jsaw> who will match_source_file help here???
15:16 < mnemoc_> i can't force him
15:17 < mnemoc_> var_append patchfiles " " "`match_source_file -p ' .*\.patch '`"
15:17 < jsaw> mnemoc_: ping. Of course we can't. But we can choose whatever name that fits.
15:17 < mnemoc_> on packages you know you add stuff on desc
15:17 < jsaw> So what's the problem now with automatic detection?
15:18 < mnemoc_> what's the problem with var_append on .conf ?
15:18 < jsaw> A package that does not need a .conf file needs a .conf only for this var_append patchfiles line? hu?
15:19 < mnemoc_> [O] patchondesc=1 ?
15:20 < rxr> jsaw: put the var_append as [O] ... ;-)
15:20 < mnemoc_> look at rsbac
15:20 < rxr> well - I have mix feeling about the patchfiles
15:20 < jsaw> I could also ask, what is the problem of giving those [D] patches a meaningful name...
15:20 < mnemoc_> i even download patches that apply to different pacakges
15:20 < rxr> one would need to review the current downloaded patches and count how many are always applied and how many are options
15:20 < mnemoc_> jsaw: look at rsbac
15:21 < rxr> and thus decide whether automation simplifies, or needs more autopatch=0 lines than it saves ...
15:21 < jsaw> mnemoc_: what's with rsbac?
15:21 < jsaw> these are diff, and thus would be detected as optional...
15:22 < mnemoc_> they are not optional
15:22 < mnemoc_> some are for userspace thing
15:22 < mnemoc_> some are for kernel
15:22 < mnemoc_> and some are for kernel module
15:22 < jsaw> and?
15:22 < jsaw> where's the argument?
15:23 < mnemoc_> too much magic is harmfull
15:23 < rxr> jsaw: mnemoc_: you wold just put a autopatch=0 into the package and do it by hand, no
15:23 < jsaw> then may I ask why we detect patches automatically anyway?
15:23 < rxr> I mean such tripple category packages are not such a problem ...
15:24 < mnemoc_> patches on $confdir are ours, patches on [D]s are not
15:24 < mnemoc_> some are -p0, some -p1 and some -p2
15:25 < mnemoc_> treating them as any t2-controlled patch is senseless
15:25 < jsaw> why?
15:25 < mnemoc_> because they are not
15:26 < mnemoc_> is package A is friendly, with one line on [O] or .conf those patches are autoapplied
15:27 < jsaw> patch is patch. If we create it or they.
15:27 * rxr prepariing to leave ...
15:27 < mnemoc_> why mess the framework to 'simplify' this for that package
15:27 < mnemoc_> rxr: due to the discussion?
15:28 < rxr> nah
15:28 < rxr> just to trigger - "but i wanted to discuss this and that ..."
15:28 < mnemoc_> :)
15:28 < rxr> it's susans birthday ... so some relaxing ...
15:28 < rxr> mnemoc_: to the discussion: normally you are the one who wants to over generalize ... ;-)
15:28 < jsaw> mnemoc_: mess the framework.... I can't see any mess...
15:29 < jsaw> rxr: hehe
15:29 < mnemoc_> changing every [D] patch to be 'friendly' is a mess
15:29 < mnemoc_> yes, i am, but i like the cleanness of t2, and this solution is not clean imo
15:29 < jsaw> mnemoc_: leaving extra in the .conf files because we don't want to change to much... hmmm...
15:30 < rxr> mnemoc_: we also untar all the tarballs ...
15:30 < jsaw> s/extra/extra mess/
15:30 < rxr> mnemoc_: which can optionally be disabled by the pkg - so patching which can be disabled should not be that bad ....
15:30 < mnemoc_> jsaw: match_source_file cleans that mess
15:30 < jsaw> mnemoc_: nimo
15:31 < mnemoc_> what package are your trying to clean?
15:31 < rxr> ncurses get's on my nerves
15:31 < jsaw> okay, let's drag this onto ML...
15:31 < rxr> 20s rebuild ...
15:31 < jsaw> mnemoc_: 23 packages
15:31 < mnemoc_> jsaw: how many of those can not be cleaned with match_source_file?
15:31 < rxr> f*ck
15:31 < jsaw> I anyway only put this on the TODO so that we one day might think about it. Currently I'm sure it's too much of a hassle to do it.
15:32 < mnemoc_> :)
15:32 < rxr> mistypes : for ; - damn os x terminl font
15:32 < rxr> and I wonder about the crazzy bash errors my postmake hook yielded ...
15:32 < mnemoc_> we need to decide what to do with COPYRIGHT-NOTE too :)
15:32 < jsaw> mnemoc_: just for completeness, once again: I do not consider match_source_file "clean"
15:32 < jsaw> anyway...
15:32 < mnemoc_> jsaw: what is not clean there?
15:32 < jsaw> :)
15:33 < mnemoc_> the overfeatured is to be used on extension packages too
15:33 < mnemoc_> like vserver's parse-config seeing what patch to apply to linux26
15:34 < mnemoc_> but that overfeature doesn't afect performance
15:34 < jsaw> NO, leave it... Let's do it on ML... I just wanted to state that, if you ask me: "jsaw: how many of those can not b.*", I can't answer this.
15:34 < mnemoc_> ok :)
15:35 < jsaw> Because we have different viewpoints...
15:35 < mnemoc_> what about COPYRIGHT-NOTE?
15:35 < jsaw> Sometimes one has to leave it at that... :)
15:35 < rxr> mnemoc_: just implant it the next days, no?
15:35 < jsaw> I wait for the final decision on the project leader
15:35 * jsaw hides
15:35 < rxr> I mean update the texts, and run it ...
15:36 < rxr> and inject the new checksummer before ,-)
15:36 < mnemoc_> rene wants one note over the other on the same file, and i want one note per file using different tag if it comes from rock or not
15:37 < jsaw> and jsaw want one t2 note + @@@ROCKCOPYRIGHT IF APPLICABLE@@@ ....
15:37 < jsaw> three views...
15:37 < jsaw> 15:37 < jsaw> I wait for the final decision on the project leader
15:37 < rxr> ok - ok - I reread the list and finally comment ...
15:39 < mnemoc_> jsaw: i'm woried about @@@ROCKCOPYRIGHT IF APPLICABLE@@@ detection, that's why i think using different TAG may help
15:39 < mnemoc_> T2-ROCK-COPYRIGHT-NOTE including @@@ROCKCOPYRIGHT IF APPLICABLE@@@ and T2-COPYRIGHT-NOTE without it
15:39 < jsaw> I can live with each of them. From my side it's actually only an astetic point... so it's that much of a problem to accept either. -> Therefore I wait for rxr's final decision.
15:40 < jsaw> so it's not that much ... even
15:40 < mnemoc_> :)
15:40 < mnemoc_> project leader has the final word :)
15:40 < rxr> .oO
15:40 -!- mipe|AFK is now known as mipe
15:40 < rxr> ok- /me away
15:40 < jsaw> that's the problem with the many cooks...
15:40 < rxr> cu all
15:40 < jsaw> cu rxr
15:40 < mipe> chroot: cannot run command `bin/bash': No such file or directory
15:41 < jsaw> hi mipe
15:41 < mipe> when entering to stage 2,maybe perl failure wasnt a bug in perl
15:41 < mipe> hi jsaw
15:45 < jsaw> <- does not know the previous history of this case
15:45 < jsaw> (or forgot it)
15:46 < mipe> during stage 1 perl build it gave error on miniperl,same error bash gives now:
15:46 < mipe> ldd bin/bash
15:46 < mipe> /usr/bin/ldd: line 1: bin/bash: Accessing a corrupted shared library
15:47 < mnemoc_> mipe: chroot . /lib/ld-linux.so.2
15:48 < mnemoc_> mipe: and chroot . /lib/ld-linux.so.2 --list /bin/bash after that
15:49 < mnemoc_> mipe: file bin/bash could be nice too
15:53 < mipe> spam coming:
15:53 < mipe> chroot . /lib/ld-linux.so.2
15:53 < mipe> chroot: cannot run command `/lib/ld-linux.so.2': No such file or directory
15:53 < mipe> chroot . /lib/ld-linux-x86-64.so.2 --list /bin/bash
15:53 < mipe> libncurses.so.5 => /lib/libncurses.so.5 (0x0000002a95557000)
15:53 < mipe> libdl.so.2 => /lib/libdl.so.2 (0x0000002a956ac000)
15:53 < mipe> libc.so.6 => /lib/libc.so.6 (0x0000002a957ae000)
15:53 < mipe> /lib/ld-linux.so.2 => /lib/ld-linux-x86-64.so.2 (0x000000552aaaa000)
15:53 < mipe> file bin/bash
15:53 < mipe> bin/bash: ELF 64-bit LSB executable, AMD x86-64, version 1 (SYSV), for GNU/Linux 2.4.0, dynamically linked (uses shared libs), stripped
15:54 < mipe> uhm,never mind that
15:54 < CIA-9> mnemoc * r5733 /trunk/scripts/Check-PkgVersion: * improved Check-PkgVersion to use download's proxy data
15:55 < mipe> chroot . lib/ld-linux-x86-64.so.2 --list bin/bash
15:55 < mipe> libncurses.so.5 => /lib/libncurses.so.5 (0x0000002a95557000)
15:55 < mipe> libdl.so.2 => /lib/libdl.so.2 (0x0000002a956ac000)
15:55 < mipe> libc.so.6 => /lib/libc.so.6 (0x0000002a957ae000)
15:55 < mipe> /lib/ld-linux.so.2 => lib/ld-linux-x86-64.so.2 (0x000000552aaaa000)
15:57 < mnemoc_> mipe: can you do the same to the libraries in which bash depends and see if there are errors?
15:58 < mnemoc_> jsaw: i have no idea neither
15:59 < jsaw> hmm... mipe you gotta wait for rxr, I'm afraid...
15:59 < mipe> :) no worries
16:00 < mipe> eh,solved :)
16:00 < jsaw> ?
16:00 < mnemoc_> how?
16:00 < mipe> ln -s ld-linux-x86-64.so.2 ld-linux.so.2
16:00 < mnemoc_> :p
16:00 < jsaw> ...
16:05 < mnemoc_> jsaw: do you think j{re,dk} should be "Restricted" instead of "Free-to-use" to get the warning?
16:06 < jsaw> Definitely. As far as I have read the comments, nobody is really sure... seems like we would need a lawyer... so before we have any problems...
16:07 < mnemoc_> ok
16:07 < mnemoc_> no gpl-java is really usable?
16:08 < jsaw> haven't heard so.
16:12 < CIA-9> mnemoc * r5734 /trunk/: * added iso and checkver dirs to svn:ignore
16:13 < mnemoc_> .oO( why is that damn closed language such popular )o
16:15 < jsaw> leaving, cu later!
16:15 < mnemoc_> http://www.dwheeler.com/java-imp.html
16:16 < mnemoc_> cu jsaw
16:22 < CIA-9> mnemoc * r5735 /trunk/package/java/ (2 files in 2 dirs): * changed blackdown-{jre,jdk} to Restricted
17:18 < CIA-9> jeru * r5736 /trunk/package/network/pound/ (. config.hlp config.in postsysfiles.in pound.conf pound.desc): * added Pound package
17:20 < mnemoc_> i586-unknown-linux-gnu-gcc -L/usr/local/lib -o miniperl \
17:20 < mnemoc_> miniperlmain.o opmini.o libperl.a -lnsl -ldl -lm -lcrypt -lutil -lpthread -lc
17:20 < mnemoc_> /usr/src/t2-test/build/test-2.1.0-beta3-x86-pentium-32-mnemosyne-expert/TOOLCHAIN/tools.cross/crosscc/../lib/gcc/i586-unknown
17:20 < mnemoc_> -linux-gnu/3.4.3/../../../../i586-unknown-linux-gnu/bin/ld: crt1.o: No such file: No such file or directory
17:20 < mnemoc_> collect2: ld returned 1 exit status
17:20 < mnemoc_> native pentium build, 1-perl :\
18:07 < mipe> hmm,mnemosyne target looks interesting,more control on what packages to include without adding manually
18:27 < mipe> hmm,generic+minimal has only 2 errors others build :)
18:27 < mipe> on x86_64
19:03 -!- CIA-9 [~CIA@flapjack.navi.cx] has quit [Remote closed the connection]
19:10 -!- CIA-temp866 [~CIA@flapjack.navi.cx] has joined #t2
19:10 -!- CIA-temp866 is now known as CIA-7
19:25 < _Ragnar_> moin
19:27 < _Ragnar_> mnemoc_: did you see yesterday? it seems ncurses still fails is stack guard is enabled
19:33 -!- CIA-7 [~CIA@flapjack.navi.cx] has quit [Read error: 54 (Connection reset by peer)]
19:34 < _Ragnar_> and for ltrace, /bin/sh -c uname -m | sed -e s/i.86/i386/ -e s/sun4u/sparc64/ -e s/arm.*/arm/ -e s/sa110/arm/ <-- this loops endlessly
19:35 -!- CIA-temp893 [~CIA@flapjack.navi.cx] has joined #t2
19:36 -!- CIA-temp893 is now known as CIA-7
19:36 -!- mipe [~mipe@dsl10040.japo.fi] has quit ["study_mode"]
19:45 < _Ragnar_> which is really weird ;p
20:43 -!- mnemoc [~amery@200.75.27.59] has joined #t2
20:53 < mnemoc> _Ragnar_: uclibc?
20:53 -!- daja77 [[hfOsKdTkh@odoaker.hrz.tu-chemnitz.de] has quit [Remote closed the connection]
20:53 -!- daja77 [[b75FVEgGW@odoaker.hrz.tu-chemnitz.de] has joined #t2
20:59 -!- mnemoc_ [~amery@200.75.27.43] has quit [Read error: 110 (Connection timed out)]
21:04 < _Ragnar_> yeah
21:06 < mnemoc> i guess that uname -m | sed on ltrace might be hacked
21:07 < mnemoc> but i don't understand why could it be looping
21:12 < _Ragnar_> me neither
21:12 < _Ragnar_> it doesn't loop if I call it from the debug.sh env
21:13 < _Ragnar_> I've disabled ltrace for now
21:14 < _Ragnar_> oh, maybe this will help:
21:14 < _Ragnar_> first time it stopped looping after fl_wrapper.log got > 2gig with these lines:
21:14 < _Ragnar_> bash(Build-Pkg)*.make*.sh.open: /TOOLCHAIN/src.ltrace.1106247387.19570.fad8824d/ltrace-0.3.36/sysdeps/linux-gnu/..
21:15 < _Ragnar_> bash(Build-Pkg)*.make*.sh.open: /TOOLCHAIN/src.ltrace.1106247387.19570.fad8824d/ltrace-0.3.36/sysdeps/linux-gnu/../..
21:15 < _Ragnar_> usw up to:
21:15 < _Ragnar_> bash(Build-Pkg)*.make*.sh.open: /TOOLCHAIN/src.ltrace.1106247387.19570.fad8824d/ltrace-0.3.36/sysdeps/linux-gnu/../../../../..
21:16 < _Ragnar_> then starting again
21:17 < mnemoc> ../../../.......
21:17 < mnemoc> is the symlink between linux-gnu and linux-uclibc created?
21:17 < _Ragnar_> yes
21:18 < mnemoc> .oO
21:21 < _Ragnar_> weirdness
21:25 < _Ragnar_> anything I should try on that?
21:25 < mnemoc> not sure, i'll start a new uclibc build without NLS and with ltrace to see
21:26 < _Ragnar_> k
21:26 < _Ragnar_> without NLS? is that disabled by default?
21:27 < mnemoc> nope
21:28 < mnemoc> NLS is enabled by default
21:28 < mnemoc> i guess you don't want NLS neither ;)
21:29 < _Ragnar_> btw 3-uucp fails, too, due to a missing ftw.o (whatever that is)
21:30 < mnemoc> uucp... long time without building that
21:30 < _Ragnar_> *g*
21:31 < mnemoc> == 01/20/05 17:30:55 =[5]=> Finished building package jikes.
21:31 < _Ragnar_> ;p
21:32 < mnemoc> but against glibc :p
21:32 < mnemoc> == 01/20/05 17:32:42 =[5]=> Finished building package classpath.
21:32 < mnemoc> :D
21:32 < _Ragnar_> prelink fails with execstack.c:25:18: argp.h: No such file or directory
21:33 < mnemoc> :(
21:33 < mnemoc> what target are you building?
21:34 < _Ragnar_> 32 bit x86 uclibc
21:34 < _Ragnar_> generic
21:34 < mnemoc> o_O
21:34 < mnemoc> do you plan to build xorg and kde there too?
21:34 < _Ragnar_> no
21:35 < mnemoc> i guess you should do your own target
21:35 < _Ragnar_> - all the graphics related stuff
21:35 < _Ragnar_> probably ;)
21:35 < mnemoc> or play with mnemosyne :p
21:36 < daja77> a tooth paste?
21:36 < mnemoc> :p
21:36 < mnemoc> my WIP target
21:36 < _Ragnar_> *g*
21:36 < _Ragnar_> how do I create a new target?
21:36 < mnemoc> mkdir target/mytarget
21:37 < mnemoc> vi target/mytarget/{pre,}config.in
21:37 < mnemoc> vi target/mytarget/build.sh
21:37 < _Ragnar_> ;p;)
21:37 < mnemoc> uboot is a do-nothing target currently
21:37 < mnemoc> you can use it as example
21:38 < _Ragnar_> ah okay
21:38 < _Ragnar_> why's it called uboot?
21:38 < mnemoc> micro-boot
21:38 < _Ragnar_> oh okay
21:39 < mnemoc> i want to do a bootdisk replacement with uclibc and busybox
21:39 < mnemoc> that's why now you can play with uclibc :D
21:39 < _Ragnar_> ah nice
21:42 < CIA-7> mnemoc * r5737 /trunk/package/java/jikes/ (. jikes.desc): * added jikes (1.22) - IBM's Java Compiler
21:43 < CIA-7> mnemoc * r5738 /trunk/package/java/classpath/ (. classpath.conf classpath.desc): * added classpath (0.13)
21:44 < _Ragnar_> why does it sometimes build a package (like for 3-util-linux) and sometimes not (like for 3-shadow)
21:45 < mnemoc> not everything is uclibc-compatible
21:45 < mnemoc> not everything is static-compatible
21:45 < _Ragnar_> I'm not building static
21:46 < _Ragnar_> and shadow should be, otherwise this whole build would be pointless
21:46 < mnemoc> see the errlog
21:47 < _Ragnar_> no error for shadow
21:47 < mnemoc> do you have XTRACE enabled?
21:47 < _Ragnar_> yes
21:48 < _Ragnar_> oh ... grep: lib/libshadow.la: No such file or directory <--
21:48 < _Ragnar_> like in:
21:48 < _Ragnar_> Creating file list and doing final adaptions ...
21:48 < _Ragnar_> Processing static lib corrections ...
21:48 < _Ragnar_> `/lib/libshadow.a' -> `/usr/lib/libshadow.a'
21:48 < _Ragnar_> Verifing the .la files ...
21:48 < _Ragnar_> grep: lib/libshadow.la: No such file or directory
21:48 < _Ragnar_> Found 349 files for this package.
21:49 < _Ragnar_> libshadow.la is in $root/usr/lib/
21:49 < mnemoc> `/lib/libshadow.a' -> `/usr/lib/libshadow.a' <--- it was moved
21:50 < _Ragnar_> that's .a, not .la
21:50 < mnemoc> oh
21:50 < mnemoc> true
21:52 < mnemoc> echo "Verifing the .la files ..."
21:52 < mnemoc> defect_la="`egrep 'lib/.*\.la$' $builddir/flist.txt |
21:52 < mnemoc> xargs egrep 'dependency_libs=.*-pthread.*' |
21:52 < mnemoc> cut -d : -f1 | sort -u | tr '\n' ' '`"
21:52 < mnemoc> can you see in your errlog if something happened to that file?
21:52 < _Ragnar_> which?
21:53 < mnemoc> libshadow.la
21:53 < mnemoc> $builddir/flist.txt should only have created and not deleted files
21:54 < _Ragnar_> /TOOLCHAIN/build/a32-2.1.0-beta3-x86-athlon-xp-32-generic-expert/TOOLCHAIN/tools.chroot/wrapper/install -c .libs/libshadow.lai /lib/libshadow.la
21:54 < _Ragnar_> that's the last reference I see
21:55 < mnemoc> and there is no /lib/libshadow.la?
21:56 < _Ragnar_> no, only in usr/lib
21:57 < mnemoc> uhm
21:57 < _Ragnar_> egrep '^lib/.*\.(a|la)$' $builddir/flist.txt <-- that also moves .la
21:59 < _Ragnar_> doesn't output anything for .la tho
21:59 < mnemoc> o_O
21:59 < mnemoc> how can a grep move a file?
21:59 < _Ragnar_> the grep selects the files to act upon
22:00 < mnemoc> .oO
22:00 < _Ragnar_> it also selects .la files from the file list, but the following loop doesn't print anything for .la, just for .a
22:01 < _Ragnar_> so the Verifing the .la files is pointlessly working on /lib/*.la ... there won't be any .la there, since they were just moved to /usr/lib/*.la one step prior
22:02 < mnemoc> else
22:02 < mnemoc> sed "s,\([ =']\)/lib\(.*\),\1/usr/lib\2,g" \
22:02 < mnemoc> $root/$fn > $root/usr/$fn
22:02 < mnemoc> rm $root/$fn
22:02 < mnemoc> fi
22:02 < mnemoc> add_flist $root/usr/$fn
22:02 < mnemoc> usr/$fn is added but $fn is not
22:02 < mnemoc> removed
22:03 < _Ragnar_> yeah
22:04 < mnemoc> but flist is not changed for *.a neither
22:05 < _Ragnar_> anyway, the error there shouldn't fail the build
22:06 < mnemoc> yes becasue those are sanity checks
22:06 < _Ragnar_> or?
22:06 < mnemoc> same for files written outside the sandbox
22:06 < _Ragnar_> then it should fail it completely, with a .err being generated, not just producing a .log, but not making any package files
22:07 < mnemoc> afaics a $del_flist $root/$fn is needed
22:07 < mnemoc> _Ragnar_: ack, that is a bug
22:11 < _Ragnar_> anyway, it still doesn't produce a *.gem when I remove that error
22:11 < mnemoc> *.gem are only produced on last stage of a package
22:11 < _Ragnar_> oh okay
22:12 < _Ragnar_> do it gets built again later?
22:12 < _Ragnar_> so*
22:12 < mnemoc> # grep '^\[P\]' package/base/shadow/*.desc
22:12 < mnemoc> [P] X ---3-5---9 179.100
22:12 < mnemoc> 3 and 5
22:13 < mnemoc> were you building 3 or 5?
22:13 < _Ragnar_> 3
22:13 < mnemoc> :)
22:13 < _Ragnar_> ah okay thx ;)
22:14 < mnemoc> about the error, try removing '^$fn' from flist.txt after that add_flist
22:14 < _Ragnar_> okay
22:15 < _Ragnar_> hmmm
22:15 < _Ragnar_> uhm
22:15 < _Ragnar_> I just checked the flist _without_ adding that del_flist $root/$fn ... and it doesn't have /lib/libshadow.a|la in it
22:17 < mnemoc> uhm?
22:18 < _Ragnar_> the only thing I had changed was: defect_la="`egrep 'usr/lib/.*\.la$' $builddir/flist.txt | <-- change lib/ to usr/lib/
22:18 < _Ragnar_> which avoids the error condition
22:19 < mnemoc> that way you don't test opt/$foo/lib neither
22:58 < _Ragnar_> mnemoc: you're invited to do a better fix :)
22:58 < mnemoc> :)
22:59 < _Ragnar_> 3-parted fails with:
22:59 < _Ragnar_> ../../libparted/.libs/libparted.so: undefined reference to `libintl_bindtextdomain'
22:59 < _Ragnar_> ../../libparted/.libs/libparted.so: undefined reference to `libintl_dgettext'
22:59 < _Ragnar_> ../../libparted/.libs/libparted.so: undefined reference to `libintl_gettext'
23:03 < mnemoc> add -lintl
23:05 < _Ragnar_> will do when I get back from lunch
23:05 < mnemoc> .oO( lunch )o
23:06 * _Ragnar_ nods
23:33 < CIA-7> mnemoc * r5739 /trunk/target/mnemosyne/pkgsel/Services/Web/00-apache.conf: * added config to enable suexec by default on mnemosyne
23:45 < CIA-7> mnemoc * r5740 /trunk/package/x86/lilo/etc-lilo.conf:
23:45 < CIA-7> * fixed default /etc/lilo.conf to name images linux instead of rock
23:45 < CIA-7> * fixed default /etc/lilo.conf to use lba32 instead of linear
23:56 < mnemoc> jsaw: on a pentium i tested rockplug took ages on "Scanning PCI sub-system", how the caching of this stuff works?
23:57 < mnemoc> jsaw: it seem it frozed :\
--- Log closed Fri Jan 21 00:00:13 2005