--- Log opened Wed May 23 00:00:51 2007
00:01 < CIA-9> sebastian * r24184 /trunk/package/graphic/epdfview/epdfview.desc: * updated epdfview (0.1.5 -> 0.1.6)
00:44 -!- emte [n=emte@d64-180-45-14.bchsia.telus.net] has quit [Read error: 110 (Connection timed out)]
01:37 -!- mtr_ [n=Michael@kobz-590cb8ce.pool.einsundeins.de] has joined #t2
01:53 -!- mtr [n=Michael@kobz-590cac08.pool.einsundeins.de] has quit [Read error: 113 (No route to host)]
01:53 -!- mtr_ is now known as mtr
02:46 -!- Stelz [i=stelz@unaffiliated/stelz] has quit ["There is intelligent life out there..."]
02:51 -!- Stelz [i=stelz@unaffiliated/stelz] has joined #t2
04:55 -!- keinek [n=chatzill@201.250.34.252] has quit [Remote closed the connection]
06:56 -!- LMJ [n=serwou@laf31-2-82-224-107-105.fbx.proxad.net] has joined #t2
07:48 -!- tri [n=richter@pd95b2df8.dip0.t-ipconnect.de] has joined #t2
07:48 < tri> moin
08:59 < rxr> re
09:15 < tri> wb
09:16 < tri> hm i get a error message with Create-Iso - genisoimage: Unexpected joliet directory length 546 expected: 550 ``
09:17 < rxr> .oO
09:17 < rxr> never saw something like that
09:21 < CIA-9> rene * r24185 /trunk/ (26 files in 26 dirs): (log message trimmed)
09:21 < CIA-9> * updated kdegames (3.5.6 -> 3.5.7)
09:21 < CIA-9> * updated kdebindings (3.5.6 -> 3.5.7)
09:21 < CIA-9> * updated kde-i18n-es (3.5.6 -> 3.5.7)
09:21 < CIA-9> * updated kde-i18n-it (3.5.6 -> 3.5.7)
09:21 < CIA-9> * updated kdeaddons (3.5.6 -> 3.5.7)
09:21 < CIA-9> * updated kde-i18n-bg (3.5.6 -> 3.5.7)
09:26 < jalsot> hi
09:26 [Users #t2]
09:26 [@ChanServ] [ Dallur] [ LMJ ] [ rxr ] [ TobiX ]
09:26 [ babygor ] [ idealm] [ mtr ] [ sepp_] [ tri ]
09:26 [ CIA-9 ] [ jalsot] [ Ragnarin] [ Stelz] [ valentin]
09:26 -!- Irssi: #t2: Total of 15 nicks [1 ops, 0 halfops, 0 voices, 14 normal]
09:26 [Users #t2]
09:26 [@ChanServ] [ Dallur] [ LMJ ] [ rxr ] [ TobiX ]
09:26 [ babygor ] [ idealm] [ mtr ] [ sepp_] [ tri ]
09:26 [ CIA-9 ] [ jalsot] [ Ragnarin] [ Stelz] [ valentin]
09:26 -!- Irssi: #t2: Total of 15 nicks [1 ops, 0 halfops, 0 voices, 14 normal]
09:27 < rxr> jalsot: it was you wanting bfin, right ?
09:27 < jalsot> yep :)
09:28 < rxr> mostly ready
09:30 < CIA-9> rene * r24186 /trunk/architecture/blackfin/ (config.hlp config.in gcc-options.in): * added blackfin cpu optimization options
09:31 < jalsot> wow, great!
09:31 < jalsot> can I test it now or let's wait for a bit?
09:31 < rxr> no that was just the tip of the iceberg
09:32 < rxr> that was just the cpu optimization config menu .-)
09:32 < jalsot> ok :)
09:33 < CIA-9> rene * r24187 /trunk/architecture/blackfin/package/ (. linux-header/ linux-header/linux-header.desc): * added blackfin linux-header overlay package
09:34 < CIA-9> rene * r24188 /trunk/architecture/blackfin/ (7 files in 2 dirs): * added blackfin binutils overlay package
09:34 < CIA-9> rene * r24189 /trunk/architecture/blackfin/ (8 files in 2 dirs): * added blackfin gcc overlay package
09:34 < CIA-9> rene * r24190 /trunk/architecture/blackfin/package/uclibc/ (. make_locale_after_headers.diff uclibc.desc): * added blackfin uclibc overlay package
09:35 < CIA-9> rene * r24191 /trunk/architecture/blackfin/ (kernel.conf.m4 package/linux26/ package/linux26/linux26.desc): * added linux26, kernel, blackfin overlay package and generic configuration
09:35 < tri> what is blackfin ?
09:35 < jalsot> tri: AnalogDevices's CPU+DSP
09:35 < rxr> http://www.t2-project.org/architectures/blackfin.html
09:37 < jalsot> rxr: wow, your bfin commits are pretty huge :)
09:37 < rxr> yes, and let me test them for an hour
09:37 < rxr> because I know some regressions from the last night test build
09:37 < jalsot> okay, I will
09:37 * jalsot is very happy :)
09:38 < tri> hm i now did another Build-Target and now Create-Iso works...
09:38 < jalsot> rxr: the bfin port is according to the current uclinux stuff or by on older one?
09:39 < rxr> we use the offical upstream toolchain, because the normal un-patched GNU flavours do not work
09:39 < rxr> that is gcc-4.1+blackfin-svn and such
09:40 < jalsot> :)
09:40 < rxr> all the blackfin package overlays are checkouts of their SVN repository
09:40 < jalsot> that's very nice
09:41 < tri> in what device is the blackfin cpu used for example ?
09:42 < rxr> satelites, digital cameras, audio products, ...
09:42 < rxr> http://en.wikipedia.org/wiki/Blackfin
09:44 < jalsot> it has pretty nice CPU power
09:46 < jalsot> rxr: you don't use the bfin support added into 2.6.22-rc1, right?
09:47 < rxr> nope
09:47 < jalsot> I'm still confused a bit with bfin wether it is MMU-less or not
09:47 < rxr> when we did our fist bfin project I tried to use the vanilla FSF GNU binutils + gcc and update the support
09:47 < rxr> however that is too cumerbsome and such
09:48 < rxr> we use their SVN repository until full blackfin supports get into gcc/binutils/etc.
09:48 < rxr> blackfin is MMU-less
09:51 < jalsot> the docs say: "The Blackfin processor contains a page based Memory Management Unit (MMU)."
09:51 < rxr> aha
09:51 < rxr> maybe of the latest greated ?
09:52 < jalsot> I see this in the "ADSP-BF53x/BF56x Blackfin Processor Programming Reference" guide
09:52 < rxr> the first ones only had a Memory Protection Unit
09:53 < jalsot> I also read: "The MMU does not support automatic address translation in hardware.:
09:54 < jalsot> maybe it is still about memory protection
09:55 < jalsot> http://www.analog.com/UploadedFiles/Associated_Docs/23439574Blackfin_PRM_1.1.pdf
09:56 < jalsot> [I read that in paper version - Revision 1.0 - June 2005]
09:56 < rxr> I look later
09:56 < jalsot> ok, thank you
09:56 < rxr> first I want to get trunk/blackfin completely building
09:56 < rxr> so you can have your fun
09:57 < jalsot> yep :)
09:57 < jalsot> which boot loader did you use in your bfin projects?
09:57 < rxr> u-boot
09:57 < jalsot> as I saw, the trunk u-boot does not have blackfin in supported architectures
09:58 < rxr> yes, it is the version + patch for avr32
09:58 < rxr> we either have to add the blackfin patch or use an overlay for the blackfin architecture
09:58 < rxr> in T2 architectures and targets can overlay packages - to use older, patched up version specifically for that case
09:58 < jalsot> how a wonderful environment, no limitations :)
09:59 < rxr> e.g. forward porting the gcc-4.1 blackfin bits to 4.2 is non-trivial and a months efford
09:59 < jalsot> and 'hacks' like with buildroot...
09:59 < rxr> effort ?
09:59 < rxr> well - I read thru the blackfin custom build shell script, not so nice either:
09:59 < rxr> http://blackfin.uclinux.org/gf/project/toolchain/scmsvn/?action=browse&path=%2Ftrunk%2Fbuildscript%2FBuildToolChain&revision=1635&view=markup
10:23 < CIA-9> rene * r24192 /trunk/package/x11/dwm/ (. dwm.cache dwm.conf dwm.desc): * added dwm (4.1) - A dynamic window manager for X11
10:24 < CIA-9> rene * r24193 /trunk/architecture/blackfin/pkg_binutils_post.conf:
10:24 < CIA-9> * wipe more blackfin/binutils overlay package sub-modules as most of
10:24 < CIA-9> the add-on stuff is not intended to be in our binutils package
10:29 -!- write_erase [n=olivier@81.80.156.36] has joined #t2
10:29 < write_erase> hi
10:30 < rxr> hi write_erase
10:30 [Users #t2]
10:30 [@ChanServ] [ Dallur] [ LMJ ] [ rxr ] [ TobiX ] [ write_erase]
10:30 [ babygor ] [ idealm] [ mtr ] [ sepp_] [ tri ]
10:30 [ CIA-9 ] [ jalsot] [ Ragnarin] [ Stelz] [ valentin]
10:30 -!- Irssi: #t2: Total of 16 nicks [1 ops, 0 halfops, 0 voices, 15 normal]
10:30 < write_erase> rxr, can I private msg ?
10:30 < write_erase> some questions about my journey
10:30 < rxr> sure
10:36 < sepp_> moin :)
10:36 < rxr> moin sepp_
10:36 < sepp_> hi rxr
10:41 < tri> hi sepp_
10:41 < tri> hi write_erase
10:41 < write_erase> hi tri
10:47 < CIA-9> rene * r24194 /trunk/architecture/blackfin/ (archtest.out pkg_gcc_post.conf): * updated blackfin target triplet and gcc config options for -mfdpic support
10:58 < rxr> jalsot: and no, I'm not yet thru ...
10:58 [Users #t2]
10:58 [@ChanServ] [ Dallur] [ LMJ ] [ rxr ] [ TobiX ] [ write_erase]
10:58 [ babygor ] [ idealm] [ mtr ] [ sepp_] [ tri ]
10:58 [ CIA-9 ] [ jalsot] [ Ragnarin] [ Stelz] [ valentin]
10:58 -!- Irssi: #t2: Total of 16 nicks [1 ops, 0 halfops, 0 voices, 15 normal]
11:00 < jalsot> :)
11:01 < CIA-9> susan * r24195 /trunk/package/mail/dcc/dcc.desc: * updated dcc (1.3.53 -> 1.3.56)
11:01 < CIA-9> susan * r24196 /trunk/package/python/dbutils/dbutils.desc: * updated dbutils (0.9.2 -> 0.9.3)
11:03 -!- emte [n=emte@d64-180-45-14.bchsia.telus.net] has joined #t2
11:20 -!- emte [n=emte@d64-180-45-14.bchsia.telus.net] has quit [Read error: 110 (Connection timed out)]
11:24 -!- emte [n=emte@d64-180-45-14.bchsia.telus.net] has joined #t2
11:31 < CIA-9> rene * r24197 /trunk/scripts/core-functions.in: * added arch2uname translation for blackfin <-> bfin
11:33 < CIA-9> rene * r24198 /trunk/package/base/uclibc/parse-config-9:
11:33 < CIA-9> * "hacked" uclibc/parse-config to not try to modify a libjava file as
11:33 < CIA-9> non-existing on the current blackfin gcc
11:40 -!- emte [n=emte@d64-180-45-14.bchsia.telus.net] has quit [Read error: 110 (Connection timed out)]
11:46 < rxr> jalsot: ok, I should be thru the bfin initial support submitting pass in some moments
11:59 -!- Netsplit calvino.freenode.net <-> irc.freenode.net quits: Ragnarin, sepp_
11:59 -!- Netsplit calvino.freenode.net <-> irc.freenode.net quits: CIA-8
11:59 -!- Netsplit over, joins: sepp_, Ragnarin
11:59 -!- Netsplit over, joins: CIA-8
12:05 < CIA-9> rene * r24199 /trunk/package/base/binutils/binutils.conf: * always build binutils in a seperated objdir
12:15 < CIA-9> rene * r24200 /trunk/architecture/blackfin/pkg_binutils_post.conf:
12:15 < CIA-9> * adapted the blackfin/binutils adaption to work in all build stages,
12:15 < CIA-9> requires the r24199 binutils objdir change
12:15 < rxr> ok - off for lunch
12:15 < jalsot> great, let me know when I can to make some tests
12:15 < rxr> blackfin is mostly there
12:15 < rxr> well - the basics just build now
12:15 < rxr> I'll continue to test with our in-house ExactMobile platform
12:16 < rxr> probably quite some packages will not yet build
12:16 < jalsot> good, I will try it now, thx
12:16 < rxr> but all the basics should now
12:16 < jalsot> so embedded target + bfin should work?
12:18 < jalsot> boune appetite
12:28 < jalsot> I got some problems during download: http://pastebin.ca/504915
12:39 < sepp_> svn checkout failed?
12:45 < jalsot> no, the ./scripts/Download -cfg bfin1 -required failed
12:45 < jalsot> svn co was fine for T2
12:47 < sepp_> i mean the svn checkout Download runs
12:48 < sepp_> "svn co -r 3174 svn://sources.blackfin.uclinux.org/ ..."
12:48 < jalsot> seems to
12:53 < sepp_> why?
12:55 < jalsot> did you check the pastebin?
12:56 < jalsot> what should I try
12:56 < jalsot> ?
12:56 < sepp_> run Download again
13:01 < tri> why isn't ephoto build if i do a emerge-pkg -repository e17 ?
13:02 < sepp_> it should build e17/*
13:03 < jalsot> sepp_: still doesn't work
13:04 < sepp_> can you manually run 'svn co -r 3174 svn://sources.blackfin.uclinux.org/linux-kernel/trunk trunk' ?
13:05 < jalsot> sepp_: yes, that works fine
13:05 < jalsot> isn't the issue related to curl?
13:05 < jalsot> curl: (22) The requested URL returned error: 404
13:05 < jalsot> ERROR: CURL Returned Error 22. Please read the curl manpage.
13:06 < sepp_> that just means that the file is not yet on the mirror
13:06 < jalsot> maybe it doesn't start svn, but gives svn ... command to curl argument
13:06 < jalsot> ah, ok
13:06 < sepp_> so it uses the original url (svn in this case) as fallback
13:10 < jalsot> what can be wrong than?
13:10 < sepp_> svn error, svn server down, interweb broken, i dont know
13:12 < jalsot> but if I run the svn co command by hand, it works
13:13 < sepp_> strange
13:14 < jalsot> I'm trying the blackfin arch. maybe this part is not well tested
13:15 < rxr> the download code is generic
13:15 < jalsot> does it work for you?
13:15 < rxr> we do not write top down all in one scripts like the official blackfin BuildToolchain.sh
13:15 < rxr> jalsot: sure
13:16 < rxr> I'll check downloading on our reference test
13:16 < rxr> but I doublt it will have problems
13:17 < jalsot> does this part do some logging?
13:17 < rxr> to the screen
13:21 < jalsot> can it be related to some locale settings?
13:21 < jalsot> I doubt, but...
13:21 < rxr> Downloading linux-header:download/mirror/l/linux-kernel-2.6.20.10-bfin-r3174.tar.bz2 ...
13:21 < rxr> svn co -r 3174 svn://sources.blackfin.uclinux.org/linux-kernel/trunk trunk
13:21 < rxr> 45M downloaded from archive so far ...
13:21 < rxr> which file failed for you ?
13:22 < jalsot> svn co -r 3174 svn://sources.blackfin.uclinux.org/linux-kernel/trunk trunk
13:22 < jalsot> svn co -r 1633 svn://sources.blackfin.uclinux.org/toolchain/trunk/uClibc uClibc
13:22 < jalsot> etc.
13:22 < rxr> do you get a specific error ?
13:22 < jalsot> no
13:23 < jalsot> http://pastebin.ca/504958
13:24 < rxr> you run T2 manually chroot'ed ?
13:24 < rxr> maybe the host is not found in the chroot jail ?
13:24 < jalsot> yes, manually chrooted
13:25 < jalsot> in chroot, manual svn co works fine
13:25 < rxr> well - we never had a complaint about that part
13:25 < rxr> maybe start debugging thru it
13:25 < rxr> Downloading linux-header:download/mirror/l/linux-kernel-2.6.20.10-bfin-r3174.tar.bz2 ...
13:25 < rxr> svn co -r 3174 svn://sources.blackfin.uclinux.org/linux-kernel/trunk trunk
13:25 < rxr> 105M downloaded from archive so far ...
13:26 < rxr> download code is in scripts/Download
13:26 < rxr> just read and instrument it
13:27 < CIA-9> rene * r24201 /trunk/architecture/blackfin/package/linux-header/scsi.patch: * avoid scsi headers in the blackfin linux-header overlay, likewise
13:29 < rxr> so - I spent the whole morning working on blackfin while a customer is waiting for some code delivery
13:30 < rxr> I finally have to start working now ...
13:33 < jalsot> rxr: thx for help,I will try to resolve this download issue - will let you know what was wrong
13:33 < rxr> ok
13:33 < rxr> svn up at least now once more
13:34 < rxr> as I fixed a blackfin/linux-header issue above
13:34 < jalsot> ok, I will - if I notice CIA-9 messages ;)
13:34 < rxr> I'll keep an eye on my testbuild here
13:34 < rxr> maybe I can fix one or another thing over the afternoon
13:35 < rxr> anything download related should be either a system problem or a generic t2 bug, nothing todo with blackfin
13:36 < jalsot> ok, thanks!
13:37 < rxr> if you need commercial support, point your browser or boss to: http://www.exactcode.de/
13:37 < rxr> (depending on whether you are the boss or no :-)
13:37 < jalsot> I think I found something
13:37 < jalsot> echo "$cmdline"
13:37 < jalsot> {
13:37 < jalsot> $cmdline || touch .cvs_error
13:37 < jalsot> } &> .cvs_output &
13:38 < jalsot> $cmdline is not run
13:38 < jalsot> echo works
13:38 < jalsot> but "$cmdline || touch .cvs_error" in brackets doesn't
13:38 < jalsot> GNU bash, version 3.1.17(1)-release (x86_64-pc-linux-gnu)
13:39 < rxr> works on 2.x and 3.0 bash here
13:40 < rxr> you can add a "set -x" in the { to log in detail what's going on there
13:41 < jalsot> ok
13:42 < jalsot> http://pastebin.ca/504976
13:42 < jalsot> so the part in brackets is not processed with my bash :(
13:54 < rxr> well - ok any output goes into .cvs_output anyway
13:55 < jalsot> this seems to fix the problem: http://pastebin.ca/504998
13:55 < rxr> add a cat .cvs_output
13:55 < rxr> this & is to run the checkout in the background while displaying the download size and waiting for completion
13:57 < jalsot> so that part does not work for me
13:57 < jalsot> now it is checking out svn - not displaying the download size - but I can live without :)
13:58 < rxr> hm, interseting
13:58 < rxr> can you fill some details about this code lines into bugs.t2-project.org
13:58 < jalsot> maybe a crap Ubuntu bash? :)
13:58 < jalsot> yes, sure
13:59 -!- idealm [n=ideal@c58-107-18-60.belrs2.nsw.optusnet.com.au] has quit [Remote closed the connection]
13:59 < rxr> jalsot: plese note your exact ubuntu and bash version output to be able to reproduce
14:01 < tri> hm inkscape cann't be build "no package gdkmm found" - i searched for this package with Find-Pkg but no success
14:01 < jalsot> ok
14:13 < jalsot> http://bugs.t2-project.org/ticket/208
14:19 -!- emte [n=emte@d64-180-45-14.bchsia.telus.net] has joined #t2
14:21 < rxr> oh - damn
14:21 < rxr> we again have chinese led trac spam
14:27 < tri> gtkmm does not build automaticly with emerg-pkg
14:28 < tri> i have to emerge-pkg cariomm to succesfully build gtkmm
14:28 < tri> i mean the dependcy was not solved autmaticly by emerge-pkg
14:33 < CIA-9> rene * r24202 /trunk/package/gnome2/gtkmm/gtkmm.cache: * updated gtkmm from the latest reference build so that cairomm is in the [DEPS]
14:33 < tri> same for gnome-vfs
14:33 < tri> it needs gnome-mime-data-2.0 and dbus-glib-1
14:36 < CIA-9> rene * r24203 /trunk/package/gnome2/gnome-vfs/gnome-vfs.cache: * updated gnome-vfs.cache from latest reference build so that dbus-glib is in the [DEP]s
14:36 < rxr> looks like gnome-mimedata was already in the deps
14:36 -!- emte [n=emte@d64-180-45-14.bchsia.telus.net] has quit [Read error: 110 (Connection timed out)]
14:40 < tri> t2 knows about deps from the cache files ?
14:40 < rxr> yes
14:41 < tri> why does inkscape need gnome-vfs
14:42 < rxr> no idea
14:42 < rxr> to load network files ?
14:43 < tri> it builds succesfully without gnome-vfs but yes load network files maybe :-)
14:43 < rxr> in my reference test build I clearly uses a lot of gnome-vfs
14:43 < rxr> maybe it's optional
14:43 < rxr> we planed to (manually) mark optional deps for some time already
14:44 < tri> attr can't be build now "libtoolize" command not found
14:48 < rxr> that's in libtool
14:49 < tri> thought this but i also thought that if install attr with emerge-pkg this dependcy is solved automaticly
14:50 < rxr> libools is in the deps
14:50 < rxr> is libtool installed ?
14:51 < tri> no libtool isn't installed
14:51 < tri> i installed it now manuelly
14:52 < rxr> is there no libtool output scrolling by in the Emerge dep resolution ?
14:52 < rxr> attr.cache:[DEP] libtool
14:54 < tri> yes is there
14:54 < tri> but i'm sure that wasn't installed
14:55 < tri> strange
14:55 < rxr> definetly
14:56 < tri> another problem if i do emerge-pkg -cfg $id -repository e17
14:56 < tri> scheduled to build: none
14:56 < tri> but if i do a emerge-pkg -cfg $id evas
14:57 < tri> checking evas ... Not installed
15:02 < rxr> yes
15:02 < rxr> -repository only emerges selected packages
15:02 < rxr> thi is because some stuff is optinal by default, like alternative ghostscript or other colliding stuff
15:03 < rxr> that would mess up the system otherwise
15:03 < rxr> this is also documented in the Emerge section in the handbook
15:03 < rxr> though I add another note, now
15:03 < tri> but why should i use repository then ?
15:06 < tri> if evas is optional for the e17 repository
15:07 < tri> hm i'm still not able to build attr
15:08 < tri> found shared files with other packages like fgetxattr.2: attr man-pages
15:08 < tri> hm who can i force attr to build or should i have to uninstall the man-pages package ?
15:09 -!- sepp [n=sepp@p85.212.66.210.tisdip.tiscali.de] has joined #t2
15:11 < rxr> tri: you can live with that, the shared files is merrily a warning note
15:11 < rxr> for a manual page it usually does not matter much
15:11 < rxr> however I wonder why you get it , as I do not get that
15:11 < rxr> maybe because usually attr and acl are installed first and man-pages do not overwrite it then ?
15:12 < tri> yes i think so
15:13 < tri> and another question: if i do a emerge-pkg attr it wants to build rpm also - but if i do a emerge-pkg gnome-vfs it wants to build attr but not rpm ?
15:14 < rxr> by default Emerge only emerges directy dependeicnes, not indirect ones as that took too much time in the past and was sufficient
15:14 < rxr> that is tweakable with -deps=indirect or so
15:14 < rxr> (seldom used that though=
15:14 < rxr> )
15:14 < rxr> yes, indirect it was
15:15 < rxr> probably should add that in the book when I'm at it now
15:16 < tri> without the indirect option some package may have problems to be build ?
15:17 < rxr> yes, rarely
15:17 < rxr> if you start from fresh some more
15:17 < rxr> but starting with a populated system some more
15:17 < rxr> we choose this way as in the past the auto-generated .cache DEPs had some more noise
15:18 < rxr> becauuse many configure and Makefiles do so freaky stuff like checking for rpm just for the fun or specific distro features
15:18 < rxr> we had way more "stay" DEPs than
15:19 < rxr> but these days they are way better and we soon can start removign this static buildf Priority tag and also let Emerge build up the full DEP graph
15:19 < rxr> in the past the whole graph just had too many cycles and such
15:22 < tri> ah ok
15:22 < tri> every day i learn a bit more about t2 - nice channel here and thank you for your good support :-)
15:24 -!- sepp_ [n=sepp@p213.54.31.246.tisdip.tiscali.de] has quit [Read error: 110 (Connection timed out)]
15:30 < CIA-9> rene * r24204 /trunk/package/base/automake/parse-config:
15:30 < CIA-9> * always apply the automake config.{guess,sub} fixup for avr32
15:30 < CIA-9> and blackfin builds, as for those architecture virtually no
15:30 < CIA-9> tarball includes the IDs, yet
15:31 < rxr> heh - nice, in the meantime I got half of our mobile platform built for blackfin
15:31 < rxr> quite some time since I did that the last time
15:36 < tri> hm i'm unable to build rpm
15:36 < tri> strange errors in the log file
15:44 < CIA-9> rene * r24205 /trunk/package/base/busybox/busybox.desc: * updated busybox (1.5.0 -> 1.5.1)
15:51 < tri> when do i have to use the -repository tag for emerge-pkg
15:58 < rxr> instead of ?
15:58 < rxr> you just use it to specify to emerge a whole repository
15:58 < rxr> especially as we have a few repositories that are named like packages, perl and python for example
16:00 < tri> hm - but how about the e17 repsoitory - if i want to emerge the whole repository emerge-pkg -repository e17 isn't enought ? or isn't evas included in the e17 repository ?
16:01 < rxr> depends on your current package selection
16:01 < rxr> Emerge -repository builds of a repository what is currently selected
16:01 < rxr> evas is in e17
16:01 < rxr> but when it is not selected in your Config right now it is not build
16:01 < tri> ah no i understand it
16:01 < sepp> add 'X e17/*' 1st
16:01 < tri> i have to select it in the pkg config
16:01 < rxr> or better '= e17/*'
16:02 < rxr> = is "assigned default"
16:02 < tri> sorry i was wrong - no it make sence
16:02 < rxr> some stuff in e17 is broken (2 or 3 packages) that are disabled by default
16:02 < rxr> X turns them on all (including the broken ones) = only re-eneabled the ones that are enabled by default
16:02 < tri> how can i do that with the Config script
16:02 < rxr> go in the config
16:02 < rxr> enable expert options
16:03 < rxr> enable custom package selection
16:03 < rxr> and then enter the rule:
16:03 < rxr> = e17/*
16:03 < rxr> or
16:03 < rxr> X e17/*
16:05 < CIA-9> rene * r24206 /trunk/architecture/avr32/pkg_busybox_post.conf:
16:05 < CIA-9> * use $arch instead of a hardcoded avr32 in architecture/avr32/pkg_busybox_post.conf,
16:05 < CIA-9> making it easier to copy to other architectures :-)
16:07 < CIA-9> rene * r24207 /trunk/package/powerpc/pbbuttonsd/pbbuttonsd.desc: * marked pbuttonsd FPIC-QUIRK, needed on x86-64
16:09 < tri> hm ok i add a package selection rule " 001: = e17/*"
16:09 < tri> but if i try to emerge-pkg -repository e17 there isn't installed anything ?
16:10 < rxr> still not ?
16:12 < tri> no
16:12 < sepp> -oldconfig ?
16:12 < tri> ah now it works
16:12 < tri> with = e17/* it doesn't
16:12 < tri> with X e17/* it does work
16:13 < rxr> interesting
16:13 < rxr> ah I think because the file is re-read for the custom pkgsel and then the packages are allready off
16:13 < rxr> yeah - well, when more parts of the Config got rewritten = will work as well
16:13 < rxr> thanks for the pointer
16:15 < CIA-9> rene * r24208 /trunk/ (2 files in 2 dirs):
16:15 < CIA-9> * only inject '--enable-runtime-cpudetection' in mplayer for
16:15 < CIA-9> architectures where mplayer supports it, removed custom removal
16:15 < CIA-9> from avr32/pkg_mplayer_post.conf
16:22 < CIA-9> rene * r24209 /trunk/architecture/blackfin/ (busybox.conf pkg_busybox_post.conf): * added blackfin busybox configuration to disabled stuff that currently fails
16:23 < CIA-9> rene * r24210 /trunk/architecture/blackfin/ (pkg_uclibc_post.conf uclibc.conf): * added blackfin uclibc configuration to force required options
16:25 < CIA-9> rene * r24211 /trunk/architecture/blackfin/package/gcc/uclibc-locale-hacks.patch:
16:25 < CIA-9> * added a quick'n dirty gcc/libstdc++/locale hack to get the libstdc++
16:25 < CIA-9> compiled for now
16:26 < Dallur> rxr: mind if I ask a question or two regarding the geode lx display driver / xorg with "strange" resolutions ?
16:26 < rxr> sure, start typing, I'll be back in some seconds
16:29 < Dallur> rxr: trying to use a custom Modeline "480x234" 25.175 480 664 760 800 234 491 493 525 to no success, I always get "Not using mode "480x234" (no mode of this name) even though it's correctly fonfigured and ddc/ignoreEDID in place
16:29 < jalsot> hmm, I got error with libtool [embedded-blackfin-537] checking how to run the C++ preprocessor... /lib/cpp
16:29 < jalsot> configure: error: C++ preprocessor "/lib/cpp" fails sanity check
16:30 < Dallur> rxr: It seems that the only resolutions I can use are those that match a "default mode", anything else fails
16:31 * jalsot trying now with current svn...
16:33 < jalsot> didn't help
16:33 < rxr> Dallur: hm, I would have to test and debug in detail on my side
16:33 < rxr> Dallur: I have no hint or answer from the top of my head
16:34 < Dallur> rxr: ok, but in theory custom modelines should work with non-standard resolutions right ?
16:34 < rxr> I never tried that
16:34 < rxr> in theory yes
16:34 < rxr> is there some output that the modeline is rejected from the modedb due to v/h sync constraints ?
16:34 < Dallur> rxr: ok, thanks, I think the problem might have to do with ddc being loaded before the ignore ddc lines are parsed
16:35 < rxr> does ddc work for you?
16:35 < Dallur> rxr: ddc works but my mode is not listed
16:35 < rxr> that is is there some mode matches from ddc ?
16:35 < Dallur> rxr: only standard modes are listed as suported, 640,800,1024,1600
16:35 < Dallur> rxr: I'm trying to get something with 16/9 aspect at least
16:36 < Dallur> rxr: I think ill dig around some more and see if I can find the cause
16:39 < tri> e17 installed into /opt/e17 with "emerge-pkg -repository e17" but if i do a "emerge-pkg evas" evas is installed into another prefix (which prefix?) and why or how can i config this ?
16:40 < rxr> with -repository or not e17 stuff installs into the Config configured E17 prefix
16:40 < rxr> that is indeed opt/e17 by default
16:40 < rxr> kde, gnome and others have equivalent options
16:40 < tri> yes saw this
16:40 < tri> but how about emerge-pkg evas why isn't that installed into /opt/e17/...
16:41 < rxr> it should be installed into the same e17 prefix
16:41 < rxr> yuche I did one of my 156 T2/trunk notes ...
16:42 < rxr> however, I know one but, those prefix options are only active when the packages are selected
16:42 < tri> hm ok where is this controlled ?
16:42 < rxr> otherwise if the base package is not selected they install into prefix "/" that should however be fixed
16:42 < rxr> in the Config is a E17 config section with the prefix
16:44 < jalsot> rxr: hmm, I got error with libtool [embedded-blackfin-537] "configure: error: C++ preprocessor "/lib/cpp" fails sanity check"
16:44 < tri> yes right - but is this e17 prefix in the config also there if there wasn't e17 in the package selection ?
16:45 < tri> so default for every package is the prefix = "/"
16:45 < rxr> no, default is /usr
16:46 < rxr> it only is a bug in those custom prefix config options that if the package is not selected the prefix is errourly empty and thus "/"
16:46 < tri> so e17 prefix is only in the config file if there was e17 in the package selection
16:46 < tri> ah ok
16:46 < tri> clear
16:47 < tri> so to have a "clean" system i do a mine -r evas and then emerge-pkg evas to get it into /opt/e17 after i changed the config
16:47 < rxr> yes
16:47 < jalsot> part of libtool's config.log: http://pastebin.ca/505327
16:47 < rxr> your evas is in / now ?
16:48 < tri> no allready did mine -r evas
16:48 < tri> just to understand how t2 works
16:48 < tri> i think it was in /usr
16:48 < tri> which make sence since this is the standard prefix you told me
16:48 < tri> but now it is in /opt/e17
16:48 < rxr> this config prefix options, however, should be fixed
16:49 < rxr> currently each big repo, like kde, gnome, e17 has such a custom option
16:49 < tri> but there are not so many big repos ?
16:49 < rxr> and as all of them where copied from the first, I think it was kde, config option all have the bug that if the base package is not selected they default to an empty prefix
16:49 < rxr> I think we should have a generic "package set config option" that all thos repos re-use
16:49 < rxr> and that has not this evil bug :-)
16:50 < rxr> tri: maybe you make this your first t2 patch :-)
16:50 < rxr> code for e17 prefix is:
16:50 < rxr> vi package/e17/imlib2/e17-conf.in package/e17/imlib2/config.in#
16:50 < rxr> package/e17/imlib2/e17-conf.in package/e17/imlib2/config.in
16:51 < tri> yes that would be nice
16:53 < tri> hm but there is allready a "else prefix="opt/e17" ... where is the $prefix_auto from ?
16:55 < tri> hm
16:55 < tri> ok it is allready fixed for imlib2 ?
16:55 < tri> but not for the other e libs and apps ?
16:57 < jalsot> rxr: I found the reason of my prev. Download issue, added comments to the tracker and closed it
16:58 < jalsot> [summary: fuser was missing]
16:58 < jalsot> ...back to the libtool issue
16:59 < jalsot> does libtool compile for you in stage 0 for bfin architecture?
16:59 < rxr> jalsot: ah, probably Download should spit out a "important tool missing" error instead of silently failing
16:59 < rxr> yes
16:59 < rxr> at -rw-r--r-- 1 root root 129821 May 23 10:24 build/bfin-7.0-trunk-mobility-blackfin-bf533-cross/var/adm/logs/0-libtool.log
17:00 < rxr> jalsot: if you like you can reopen the issue and add that the Download should error our with an indication, nicely
17:00 < jalsot> okay, I will
17:00 < tri> rxr: was that yes for me ?
17:00 < jalsot> for libtool I get: configure: error: C++ preprocessor "/lib/cpp" fails sanity check
17:00 < jalsot> maybe something is missing again from my box? :)
17:08 < tri> rxr: ah ok it understand the imlib2 prefix patch - should i post the *.in files to the mailing-list ?
17:09 < tri> ^/it/i
17:13 < tri> must leave now_ cu
17:13 -!- tri [n=richter@pd95b2df8.dip0.t-ipconnect.de] has left #t2 []
17:25 < jalsot> rxr: any idea what can be wrong with libtool? [I recompiled the whole again and got the same]
17:26 < jalsot> oops, I have an idea :)
17:26 < jalsot> checking...
17:26 < jalsot> hmm, didn't help - I was checking wether I have cpp on my host system - I have, so this should not cause the issue
18:03 < jalsot> ooooops, my fault, g++ package was missing on host (I had cpp, but this is not enoug;)
18:14 < jalsot> now I got something with 00-dirtree: cp: `/DEVEL/t2-trunk/build/bfin1-7.0-trunk-embedded-blackfin-bf537-cross/TOOLCHAIN/tools.cross/usr/share/automake/config.guess' and `/DEVEL/t2-trunk/build/bfi
18:15 < jalsot> n1-7.0-trunk-embedded-blackfin-bf537-cross/./TOOLCHAIN/tools.cross/share/automake-1.9/config.guess' are the same file
18:17 < jalsot> the cp complaining is valid as automake is a soft link to automake-1.9
18:25 < rxr> .o?
18:39 < rxr> I can reproduce
18:43 < rxr> when I adapted the automake generic config.{sub,guess} update this regression was introduced
18:43 < rxr> and because my build was mostly finished that time I just see this now, with a full re-build test
18:43 < rxr> I fixed it locally
18:43 -!- emte [n=emte@d64-180-45-14.bchsia.telus.net] has joined #t2
18:43 < rxr> but let's wait some 30 minutes unil this dual-core rebuilt all of my embedded test-case fine
18:54 -!- misl [n=chatzill@84-104-172-187.cable.quicknet.nl] has joined #t2
18:54 < misl> hi folks
19:00 < rxr> hi misl
19:01 < misl> hi rxr
19:02 < CIA-9> rene * r24212 /trunk/package/develop/oprofile/ (oprofile.desc pulpstonec.pl pulpstoned.sh pulpstoner.pl): * added oprofiel Url tag and removed the pulpstone* scripts
19:07 < jalsot> hi, ok, I will wait :)
19:14 -!- write_erase [n=olivier@81.80.156.36] has quit ["Leaving"]
19:28 < rxr> jalsot: r24213 is for you
19:28 < CIA-9> rene * r24213 /trunk/package/base/automake/parse-config: * fixed automake/parse-config to only perform action for packages compiling source
19:29 < jalsot> rxr: many thanks! trying...
19:30 < jalsot> cool, went over :)
19:31 < jalsot> I saw there is asterisk in packages, is there any plan to add callweaver [formely openpbx.org]?
19:32 < rxr> if it compiles out of the box that is no problem at all
19:32 < rxr> we add everything that we find quickly
19:32 < rxr> you can even contribute a ready made patch for the new package
19:32 < jalsot> it needs spandsp only - which is missing atm
19:32 < rxr> did you read how to define a package in t2 ?
19:32 < jalsot> I will try :)
19:34 < jalsot> oops, still something wrong around 00-dirtree
19:34 < jalsot> Found shared files with other packages:
19:34 < jalsot> usr/include: 00-dirtree uclibc
19:34 < jalsot> uclibc reported
19:35 < jalsot> can I somehow remove a package [and than recompile]?
19:35 < rxr> hm - we had this usr/include report once already
19:35 < rxr> was it you?
19:35 < jalsot> no
19:36 < jalsot> that was somebody else :)
19:36 < rxr> I can not imagine how that one happens now
19:36 < rxr> for now rename the .err to .log
19:36 < rxr> the build/.../var/adm/log/ uclic.err
19:37 < jalsot> ok
19:37 < jalsot> maybe previouse 00-dirtree kept some garbage there?
19:37 < rxr> hm - no
19:38 < rxr> you got the error on uclibc ?
19:38 < jalsot> yes
19:38 < rxr> maybe the debian tar has some directory creation difference
19:38 < rxr> just rename the logfile for now
19:38 < jalsot> ok
19:39 < jalsot> thx
19:46 < jalsot> another issue - with perl - "/usr/bin/ld: cannot find -lgcc_s", I have libgcc1 installed on host
19:58 < jalsot> /usr/bin/ld: skipping incompatible /usr/lib/gcc/x86_64-linux-gnu/4.0.3/libgcc_s.so when searching for -lgcc_s
20:07 -!- d-marc [n=d-marc@HSI-KBW-082-212-033-229.hsi.kabelbw.de] has joined #t2
20:07 < rxr> perl in which stage ?
20:08 < jalsot> stage 1
20:09 < rxr> then that is built for your blackfin cpu
20:09 < jalsot> I use the embedded target, without modifications
20:09 < rxr> built here
20:09 < rxr> but if you likely do not need perl then you can deselect perl for now
20:09 < jalsot> right trying that :)
20:10 < jalsot> to add "- perl" to rules?
20:10 < rxr> yes
20:10 < rxr> O
20:10 < rxr> O perl
20:12 < jalsot> thx
20:17 < jalsot> unfortunately patches for linux26 are rejected so it failed
20:17 < jalsot> 1 out of 1 hunk FAILED -- saving rejects to file Documentation/filesystems/00-INDEX.rej
20:17 < jalsot> 1 out of 2 hunks FAILED -- saving rejects to file fs/Kconfig.rej
20:17 < jalsot> 1 out of 1 hunk FAILED -- saving rejects to file fs/Makefile.rej
20:18 < jalsot> did you try with the embedded target or something else?
20:23 < rxr> I used an in-house mobile target
20:24 < rxr> but I skipped the kernel
20:24 < d-marc> hello
20:25 < jalsot> hi
20:25 < jalsot> I doubt I could step forward without the kernel ;)
20:26 < rxr> sure
20:26 < rxr> but for a custom board you usually need a custom kernel anyway
20:27 < rxr> but I can get the vanilla kernel built for now
20:27 < jalsot> yes, probably there will be some patches after some time:)
20:30 < jalsot> where are thos patches located which are failing?
20:31 < rxr> loook into the build/.../var/adm/logs/1-kernel26.err
20:31 < jalsot> this seems to be related to unionfs
20:32 < jalsot> ok, found: linux-2.6.21-rc7-u1.diff.bz2
20:32 < rxr> oh, yeah
20:32 < rxr> ok - disable unionfs :-)
20:33 < rxr> in general with this per-architecture custom kernels non-upstream architectures the third party patches usually do not apply too well
20:36 < jalsot> what would you suggest? how to proceed?
20:36 < rxr> disable unionfs in your package selection as you did for perl
20:36 < rxr> here the bfin kernel starts to compile so far
20:37 < rxr> maybe with the kernel patching packages that do not patch this custom bfin kernel disabled the kernel will already build
20:37 < jalsot> yes, it could build, but the process stops at patching
20:38 < jalsot> how will the patch part know - it has to ignore unionfs?
20:40 < rxr> when unionfs is selected the unionfs package injects those patches
20:40 < rxr> so simple disableding unionfs will make those unionfs patches stop interfering your build
20:41 < rxr> jalsot: do you already have a bfin cpu / board ?
20:41 < jalsot> yes, I have a BF537-STAMP for playing
20:41 < rxr> btw. advice: when tinkering with board that come with a linux kernel installed we usually just exchange the user-space for testing until it runs as intended
20:41 < rxr> and tinker with the kernel last
20:43 < jalsot> the kernel in the board is pretty old, AFAIK
20:43 < jalsot> it has image made with 2005R3 toolchain
20:43 < jalsot> which is pretty old, don't know if except 'hello world' there would run anything
20:44 < jalsot> btw, unionfs is now ignored and compiling kernel :)
20:46 < jalsot> but compilation failed :(
20:47 < rxr> it might still not build completely
20:48 < rxr> I'll hack a blackfin board selection menu entry over the night
20:48 < rxr> gotta go
20:48 < rxr> btw. you do this bfin stuff for a company ?
20:48 < jalsot> http://pastebin.ca/505748
20:49 < jalsot> yes, we try to make something out from bfin before makeing useful things
20:49 < rxr> are you a freelancer or have a job position ?
20:50 < jalsot> have a job position, but this part is a hobby at the moment
20:50 < jalsot> but we are open to talk about some funding
20:51 < rxr> if you want to test the t2 build binaries, skip the kernel and go on
20:51 < jalsot> I mean, fund any project which could help with reaching the goals :)
20:51 < rxr> if you want to have some t2 hacking fun get the current blackfin kernel tree to link ;-)
20:51 < jalsot> not decided yet where to go - there are also promises to get bfin into openwrt
20:52 -!- d-marc [n=d-marc@HSI-KBW-082-212-033-229.hsi.kabelbw.de] has quit ["KVIrc 3.2.4 Anomalies http://www.kvirc.net/"]
20:52 < rxr> and if you need professinoal support you know the exactcode homepage, my mail is rene @ exactcode . de :-)
20:52 < jalsot> "get the current blackfin kernel tree to link" - good idea, but don't know how to do that yet
20:52 < jalsot> yes, I know :)
20:53 < rxr> you'r t2 blackfin kernel will fail to link with size and section mismatches
20:53 < rxr> at least it did just here
20:53 < rxr> I can take a look sometime tomorrow
20:53 < rxr> first I have to continue please our customers ...
20:53 < jalsot> maybe we could talk about some commercial side openly? just to know what I should "show" in the company
20:53 < jalsot> okay
20:53 < rxr> take a look on the homepage, we do this kind of low-level design stuff, including this brand new avr32 and blackfin
20:54 < rxr> like kernel driver porting, development assembly optimization and normal application development
20:54 < jalsot> yes, I was looking for your website already
20:54 < rxr> for the usual reasonable pricings
20:54 < jalsot> what does reasonable price mean? ;)
20:54 < rxr> but as you can see we also do our best to support you using T2 for free here in the IRC
20:54 < rxr> but that is of course just as time permit
20:55 < rxr> jalsot: which country are you located ?
20:56 < rxr> but pricing depends on project size and working area, low-level fiddeling is obviously higher priced than "end-user application coding"
20:56 < rxr> I really have to go now
20:56 < jalsot> we are in Hungary
20:56 < rxr> if you like you could drop some project summariy into my inbox and I can come back with some numbers
20:56 < rxr> ah hungary, yes you wrote that already .-)
20:57 < jalsot> okay, I will write the summary in email :)
20:57 < jalsot> thank you for your help!
20:57 < rxr> btw. just to make this clear, T2 is Open Source and GPL you can just use it and also rely on interim help here
20:57 < jalsot> I will play a bit with kernel
20:58 < rxr> and we also do qutie a lot just for the fun, though seldomly I can do so much I did the last 30h here, that was mostly because I had it already around from a former project and I long wanted it cleanly sorted into T2
20:58 < rxr> but with some contracted work you obviously get exactly what you need, in a garuanteed time-frame
20:58 < jalsot> yes, that's true
20:59 < jalsot> I will write an e-mail soon ;)
20:59 < rxr> nontheless T2 is becoming one of the best scaling (supported CPUs and package wise) and well documented systems these days
20:59 < jalsot> yep, I read that OpenEmbedded is a nightmare :)
20:59 < rxr> I even saw company inquries for embedded boards where T2 was exactly listed as must criteria to base the system on
20:59 < jalsot> buildroot is not so good
21:00 < jalsot> openwrt is on a good level, 16 platforms
21:00 < rxr> but not 16 cpu platforms, more like 16 differnt ARM or MIPS router boards, no ?
21:01 < jalsot> yes, true
21:01 < rxr> ok - my girl is kicking me - gotta go
21:01 < rxr> have a good night
21:02 < jalsot> :)
21:02 < jalsot> have a nice evening
21:02 < jalsot> gutte nacht
21:02 < sepp> 'gute' ;)
21:02 < jalsot> [probably I wrote wrong as my German is far-far...]
21:03 < jalsot> :)
21:03 < jalsot> thx
21:03 < jalsot> danke
21:03 < jalsot> and here my German knowledge nearly stops
21:04 < sepp> well, soon they will reform it again and then noone will know howto write any word, so that does not matter :p
21:52 < CIA-9> sebastian * r24214 /trunk/package/mail/claws-mail/claws-mail.desc: * updated claws-mail (2.9.1 -> 2.9.2)
22:02 -!- keinek [n=chatzill@201.250.34.252] has joined #t2
22:07 -!- emte [n=emte@d64-180-45-14.bchsia.telus.net] has quit [Read error: 110 (Connection timed out)]
23:21 -!- misl [n=chatzill@84-104-172-187.cable.quicknet.nl] has quit [Read error: 110 (Connection timed out)]
--- Log closed Thu May 24 00:00:54 2007