--- Log opened Sat Mar 19 00:00:27 2005 00:16 -!- praenti [~praenti@82.139.198.99] has joined #t2 00:17 < mnemoc> re 00:32 -!- jeru [~jeru@p54BFE721.dip.t-dialin.net] has quit [Read error: 110 (Connection timed out)] 01:24 -!- keinek [~Keinek@201.254.10.12] has quit [Read error: 60 (Operation timed out)] 01:28 -!- Postal [~edinatux1@ool-4356292a.dyn.optonline.net] has joined #t2 01:28 < Postal> Hey all 01:29 < Postal> Is anyone around? 01:36 < Postal> mnemoc, rxr ? 01:36 -!- keinek [~Keinek@201.254.8.37] has joined #t2 01:39 < mnemoc> re 01:39 < mnemoc> Postal: hi 01:39 < mnemoc> hi keinek-chan 02:11 -!- keinek [~Keinek@201.254.8.37] has quit [Remote closed the connection] 02:17 -!- Ken|JLime [~Ken@h72n2fls34o1105.telia.com] has left #t2 ["Lämnar"] 02:30 < CIA-9> amery * r7566 /trunk/package/network/openldap/openldap.desc: * updated openldap (2.2.23 -> 2.2.24) 04:04 < CIA-9> amery * r7567 /trunk/target/mnemosyne/ (pkg_mod_mono.conf pkgsel/Services/Web/mono.ask): * added ASP.NET plugin for mnemosyne 04:58 < Postal> mnemoc you still around? 05:07 < mnemoc> yes 05:12 < Postal> cool 05:12 < Postal> how is the 64 bit branch looking these days? 05:15 < mnemoc> fine afaik 06:00 -!- Postal [~edinatux1@ool-4356292a.dyn.optonline.net] has left #t2 [] 06:17 < rxr> re 06:18 < rxr> Error logs from system-2.1.0-rc1-x86-pentium-mmx-desktop: 06:18 < rxr> [5] mail/spamassassin [5] multimedia/dvb-apps [5] e17/efsd [5] e17/evidence [5] gnome2/libgda [5] gnome2/bug-buddy 06:18 < rxr> 683 builds total, 505 completed fine, 6 with errors. 06:29 -!- rxr_ [~rene@p213.54.202.211.tisdip.tiscali.de] has joined #t2 06:29 -!- Topic for #t2: T2 | 2.1.0-beta3 RELEASED | The next generation of System Development Environments (SDE) | http://www.t2-project.org/ | @CeBIT Hall 6 H30-532 06:29 -!- Topic set by ao [] [Wed Mar 16 09:21:34 2005] 06:29 [Users #t2] 06:29 [ _Ragnar_] [ jsaw ] [ nzg ] [ rxr ] [ sparc-kly] 06:29 [ CIA-9 ] [ mnemoc] [ praenti] [ rxr_] [ valentin ] 06:29 -!- Irssi: #t2: Total of 10 nicks [0 ops, 0 halfops, 0 voices, 10 normal] 06:29 -!- Channel #t2 created Sun Aug 8 21:15:33 2004 06:29 -!- [freenode-info] why register and identify? your IRC nick is how people know you. http://freenode.net/faq.shtml#nicksetup 06:29 -!- Irssi: Join to #t2 was synced in 11 secs 06:29 < mnemoc> :) 06:29 < mnemoc> 1477 builds total, 1269 completed fine, 35 with errors. 06:29 < mnemoc> libgda? 06:29 < mnemoc> it built on my reference 06:29 < mnemoc> http://www.geeks.cl/~amery/T2/regressions.html 06:30 < rxr_> ah - yesh - mysql missing 06:30 < mnemoc> bacterium's .conf is weird 06:31 < rxr_> indeed 06:32 < rxr_> ah - no - seems to use legal features ... 06:32 < rxr_> however I clean it up, ok ? 06:33 < mnemoc> sure 06:33 < mnemoc> i'll go to sleep :p 06:33 < rxr_> cu mnemoc - sleep well 06:33 < mnemoc> i hear about gnome-2.10.1 on #mono 06:33 -!- rxr [~rene@p213.54.201.213.tisdip.tiscali.de] has quit [Read error: 60 (Operation timed out)] 06:33 < rxr_> aha 06:33 < rxr_> soon? 06:33 < mnemoc> yes 06:34 < mnemoc> today or tomorrow 06:34 < rxr_> in terms of available - uploaded? 06:34 < mnemoc> not yet 06:38 < rxr_> http://osnews.com/story.php?news_id=10032 06:38 < mnemoc> *click* 06:41 < mnemoc> uhm 06:52 < mnemoc> http://www.geeks.cl/~amery/T2/regressions.html <-- if you want to fix some another failures :) 06:52 < mnemoc> updated hourly 06:53 < mnemoc> gn8 rxr_ 06:53 < mnemoc> good luck in your math exam 06:58 < rxr_> eBay erleidet Niederlage im Patentstreit 06:58 < rxr_> Die Berufungsinstanz hat im Streit zwischen MercExchange und dem Online-Auktionshaus entschieden, dass eBay zur Entfernung der "Buy it now"-Funktion veranlasst werden kann. 06:58 -!- You're now known as rxr 07:05 < CIA-9> rene * r7568 /trunk/package/editors/xemacs-beta/ (xemacs-beta.conf xemacs-beta.desc): * updated xemacs-beta (21.5.19 -> 21.5.20) 07:05 < CIA-9> rene * r7569 /trunk/package/audio/bacterium/ (4 files): 07:05 < CIA-9> * cleaned up bacterium to not use a very ugly custmain - but modern 07:05 < CIA-9> build features ... 07:09 < CIA-9> rene * r7570 /trunk/package/x11/mkttfontdir/mkttfontdir.conf: * removed custmain from mkttfontdir also 07:10 < rxr> # grep custmain */*/*.conf | wc -l 07:10 < rxr> 132 07:11 < rxr> # grep custmain= */*/*.conf | wc -l 07:11 < rxr> 123 07:41 < CIA-9> rene * r7571 /trunk/package/base/lsof/lsof.conf: * decustmained lsof 07:44 < rxr> Linux 2.4.30-rc1 07:44 < rxr> mnemoc: Linux 2.4.30-rc1 07:46 < rxr> mnemoc: -beta4 tag today - source only - jus for reference and PR ? 07:47 -!- Minto [~chatzilla@82-217-66-7.cable.quicknet.nl] has joined #t2 07:48 < rxr> valentin: oh - I missed to check in the archtest.sh for sh ... 07:48 < rxr> no wonder your result was different ... 07:49 < CIA-9> rene * r7572 /trunk/architecture/sh/archtest.sh: * added previously forgotten architecture/sh/archtest.sh ... 07:58 < CIA-9> rene * r7573 /trunk/target/desktop/config.in: * removed dosemu and bochs from the desktop target 08:03 < rxr> 15177 daemon 21 0 22228 16m 5080 R 25.6 3.3 0:33.69 rastertokm2430d 08:04 < rxr> ^ - ammount of Athlon CPU cycles wasted rastering a 7 pages document for my laser printer ... 08:07 -!- mnemoc_ [~amery@200.75.27.7] has joined #t2 08:07 -!- Minto [~chatzilla@82-217-66-7.cable.quicknet.nl] has quit [Read error: 60 (Operation timed out)] 08:16 -!- mipe [~mipe@dsl10040.japo.fi] has joined #t2 08:21 -!- mnemoc [~amery@200.75.27.34] has quit [Read error: 110 (Connection timed out)] 08:27 < CIA-9> rene * r7574 /trunk/package/graphic/xpdf/xpdf.conf: * removed debugging output from xpdf.conf and simplified it (slightly) 08:39 < CIA-9> rene * r7575 /trunk/package/ (10 files in 10 dirs): * substituted remaining kde_config.sh to kde-conf.in 08:53 < valentin> moin 08:53 < valentin> rxr: aha 09:00 < rxr> moin valentin 09:06 < valentin> schumacher once again screwed up in the qualifiing - this season could become interessting 09:17 < rxr> oh 09:17 < rxr> where is the racing going on this time ? 09:18 < valentin> sepang 09:18 < valentin> but i will most probably miss the end of the race tomorow, because we travel to dresden 09:20 < rxr> worth to switch on my 21" CRT ? 09:20 < mipe> i dont like the qualifying done this year. should revert to old one 09:20 < rxr> mipe: no ,-) the last years sucked ... - then better revert to the 90th style of doing two and taing the best lap ... 09:21 < rxr> but this year is fine with me, too ... 09:21 < mipe> ah yep,not last year. few years back... everyone can drive as many as lap they wants best lap wins 09:21 < rxr> ah Sepang 09:22 < rxr> Sepang International Circuit in Kuala Lumpur 09:23 < rxr> ok - have to wait until 13:00 to watch the retransmission on rtl - no other station available transmitting it sooner here in Berlin over DVB-t ... 09:26 < valentin> hi mipe 09:27 < mipe> hi valentin 09:27 < valentin> i agree about the new rules 09:27 < valentin> and they should alow changing the tires 09:27 < valentin> then the drivers will risk more 09:28 < valentin> and that motor rule is ridiculos 09:57 < valentin> just a stupid question: why do we have to rebuild the linux-headers in stage 1 ? 10:07 < valentin> my network connection sucks a bit today 10:12 < rxr> valentin: we do not need - it is just for fun ... 10:12 < rxr> ,-) 10:26 < valentin> damn, my dsl line realy lags today 10:27 < rxr> hm 10:28 < rxr> maybe svn up's ? 10:28 < valentin> grrrr 10:29 < valentin> ,) 10:29 < rxr> hm - nope 10:30 < valentin> rxr: now i get the same glibc error that you pasted yesterday 10:30 < rxr> have fun investigating ... 10:30 < valentin> hm - even svn.exactcode times out 10:30 < valentin> now it is loaded 10:30 < valentin> maybe a dns problem ? 10:31 < rxr> from my side I have no problems ... for *exactcode.de or other stuff on your server ... 10:38 < valentin> what exactly is glibc's rule for choosing implementations from sysdeps ? sth like : 10:38 < rxr> This month mirror cgi request log: 10:39 < rxr> 4 2.1.0-beta2 10:39 < rxr> 21 2.1.0-beta3 10:39 < rxr> 13 2.1.0-beta4 10:39 < rxr> 66 2.1.0-rc1 10:39 < rxr> valentin: configure time as far as I have seen ... 10:39 < valentin> if test ! -x sysdeps/$arch/$trg.c ; then return sysdeps/generic/$trg.c 10:39 < rxr> last month: 10:39 < rxr> 3 2.1.0-beta2 10:39 < rxr> 40 2.1.0-beta3 10:39 < rxr> 118 2.1.0-beta4 10:41 < valentin> what the f... 10:41 < valentin> #ifdef __STDC__ 10:41 < valentin> float __modff(float x, float *iptr) 10:41 < valentin> #else 10:41 < valentin> float __modff(x, iptr) 10:41 < valentin> float x,*iptr; 10:41 < valentin> #endif 10:41 < valentin> why can't he take the first alternative ? 10:41 < valentin> any compiler should compile that one 10:42 < rxr> ... 10:42 < rxr> glibc is well developed high quality code - Ullrich is doing his best do archive this ... 10:46 < valentin> rxr: maybe they messed sth up when fixing that issue: http://sources.redhat.com/ml/libc-alpha/2003-06/msg00027.html 10:48 < valentin> http://gcc.gnu.org/ml/gcc-patches/2002-12/msg01138.html 10:48 < valentin> >> The attached is for not to include __fpscr_values in libgcc.a if needed. 10:48 < valentin> In sh-linux, we used to do so and define __fpscr_values at crt1.o to 10:48 < valentin> avoid that each shared library has its copy. 10:49 < valentin> i take a look what is inside crt1.o 10:50 < valentin> 00000004 D __fpscr_values 10:50 < valentin> bingo 10:50 < valentin> so i need to tel 10:51 < valentin> l the linker he should include that object 10:55 < rxr> if you manage to get sh bootsrap today, we can announce -beta4 to include super h support ,-)( 10:56 -!- jeru [~jeru@p54BFFCD0.dip.t-dialin.net] has joined #t2 10:57 < valentin> you want to finish -beta4 today ? 10:57 < valentin> btw: an sed 's, , \\\n,g' over the glibc compiler log helps ,) 10:57 < rxr> I just wanted a reference tag and PR for the weekend ... 10:57 < rxr> -rc1 until end of next week ... 11:00 < CIA-9> rene * r7576 /trunk/scripts/Build-Pkg: * prettified "No xyz.log file." to a more informative message 11:00 < valentin> hm, i just cannot link crt1.o into libc.so 11:01 < valentin> if i do so i get an undefined reference to 'main' 11:01 < valentin> maybe i just should move the definition to some other place ? 11:02 < rxr> crt1.o is the start code injected by gcc into binaries ... 11:02 < rxr> s/binaries/executable/ with main function 11:02 < valentin> thought so 11:02 < valentin> but how can i fix that problem, then ? 11:03 < valentin> i do not want to move whole code sections around 11:04 < rxr> I have no real inside in this FP stuff on SH right now - maybe just browse those thread a bit more and review what's in glibc and gcc today ... 11:04 < rxr> should I try what sh3 does ? 11:04 < rxr> that is start a test build ... 11:06 < valentin> if you want 11:06 < valentin> i'll continue to investigate sh4 11:06 < rxr> sure 11:06 < rxr> should work some day ,-) 11:07 < rxr> # ./scripts/Build-Target -cfg sh-test 11:07 < rxr> running 11:14 < rxr> valentin: ah I remeber a case where linux-header needs a rebuild 11:14 < rxr> sparc IIRC ... 11:15 < rxr> some files are generated by the compiler ... 11:15 < rxr> and not needed in stage0 for the glibc bui.d - 11:15 < rxr> ild but some other user space stuff later on ... 11:15 < rxr> I add a comment to the prioritky of linux-headera 11:17 < CIA-9> rene * r7577 /trunk/package/base/linux-header/linux-header.desc: * added a comment to linux-header why we need them in stage 0 _and_ 1 11:39 < rxr> ok - so scrolled and commented thru a power point PDF for my electronic exam ... 11:39 < rxr> seems time to drive to susan for lunch ... ,-) 11:41 < rxr> # dvbsched dvb-t rtl "13.00 " 90 /var/tmp/formula1.mpg 11:41 < rxr> ,-) ok - /me off in some seconds ... 11:42 < valentin> cu 11:43 < rxr> btw. dvbsched is quite dump - I misstyped the station name yesterday evening (kabel instead of kabel1 or so) - and it accepted it ... 11:43 < rxr> and I need to hack it to be able to list the scheduled recordings with -l or so ... 11:44 < rxr> and it is also dump enough to not stop life stream obtained from dvbcat when it should start a recording on another frequency ... 11:44 < rxr> I would expect it to just stop the life stream in favour of the scheduled one ... 11:44 < rxr> nearly missed susans berlin, berlin due to that on Thursday ... 11:45 < rxr> because my enterprise live stream was running ... ,-) 11:45 < rxr> good that I took a look in time ... 11:46 < rxr> the sh build is at == 11:42:19 =[1]=> Building base/glibc [2.3.4 2.1.0-rc1]. 11:46 < rxr> but I'll leave before it hits the usual sh4 error ... 11:50 < rxr> if you do not mess with my mpg dumps you could log in (to haudrafinix) and see how /usr/src/t2-trunk/build/sh-test*sh4* goes ... 11:50 < rxr> cu 12:30 < valentin> ok 13:05 < valentin> rxr: haudraufinix cannot be resolved on obelix 13:06 < valentin> strange - reverse lookup of haudraufinix's ip works 13:08 < valentin> rxr: for sh3 the glibc compiled on your box 13:08 < valentin> the build failed later in 13:08 < valentin> == 12:25:00 =[1]=> Building base/linux26 [2.6.11 2.1.0-rc1]. 13:08 < valentin> but i still try to workaround my problem 13:09 < valentin> it is just that one - anything else build fine 13:16 < valentin> rxr: maybe we need to add no-z-defs=yes to glibcs config option 13:16 < valentin> because the symbol must be defined and only defined in crt1 13:17 < valentin> otherwise you get some math regressions 13:20 < valentin> see also http://www.m17n.org/linux-sh/ml/linux-sh/2003-08/msg00041.html 13:21 < valentin> >>> the compilation of glibc-2.3.2's libc.so failed with undefined 13:21 < valentin> >>> reference to `__fpscr_values'. 13:21 < valentin> > Adding the line 13:21 < valentin> > no-z-defs=yes 13:21 < valentin> > to the file configparms in my glibc build directory gets me past that 13:21 < valentin> > problem, thanks! 13:28 < valentin> ^- that file does not exist 13:28 < valentin> but there are makefile rules evaluating theese variables 13:30 < valentin> that file seems to be crosstool stuff 13:42 < valentin> i try to set that var directly over makeopts 13:43 < rxr> re 13:44 < rxr> no 13:44 < rxr> see glibc32 13:44 < rxr> it also messes with configparams ... 13:44 < rxr> have you attached to my screen ? 13:45 < rxr> root pts/9 192.168.2.1 Sat Mar 19 13:07 still logged in 13:45 < rxr> ? 13:46 < valentin> yes 13:46 < valentin> mom 13:46 < valentin> ok 13:46 < valentin> sorry, i killed the shell with the build output 13:46 < rxr> I noticed ... 13:46 < valentin> but as i pasted above .... 1-linux26 13:47 < valentin> rxr: what do you mean with 13:47 < valentin> 13:44 < rxr> no 13:47 < valentin> 13:44 < rxr> see glibc32 13:47 < valentin> 13:44 < rxr> it also messes with configparams ... 13:47 < rxr> # do not build the programs 13:47 < rxr> hook_add premake 5 "cat > configparms <<-EOT 13:47 < rxr> build-programs=no 13:47 < rxr> EOT 13:47 < rxr> " 13:47 < rxr> use this style on glibc.conf for sh4 ... 13:47 < rxr> and insert your fluff there ... 13:48 < valentin> ok 13:48 < valentin> but appending that varialbe to makeopt seems to work too 13:48 < valentin> which way do you prefer ? 13:51 < valentin> i greped a bit through the scripts - where is the configparams evaluated ? 13:51 < rxr> I think the well known configparams is nice ... 13:51 < rxr> it is included by one of the many self including makefiles ... 13:51 < rxr> IIRC 13:52 < valentin> i greped through makefiles, but .... 13:52 < valentin> maybe i am just to dump to find it 13:52 < valentin> ah - configparms 13:52 < valentin> without an 'a' .... 13:55 < rxr> ah - yes - I also continously mistype it ... 13:55 < rxr> those "shortcuts" are nice, no? ,-) 13:55 < valentin> like umount 13:56 < valentin> save one superfluos vovel 13:56 < rxr> we could translate it in the desktop target - back to unmount ,-) ,-) 13:56 < valentin> or consonant in case of the u 13:56 < valentin> or create a symlink 13:56 < valentin> but ok ... 13:57 < rxr> translate it and add a backward comp. link to umount ,-) 13:57 < rxr> you do the configparm injectoin conditionally based on $arch_machine = sh4, right ? 13:58 < rxr> (and violate the bosch patent on memory cell based comparision ... here in europe ... 13:58 < rxr> ) 13:59 < rxr> valentin: maybe we need a sh board option to get the kernel built, right ? 14:01 < rxr> "Bare CPU" aka "unknown" means an SH-based system which is not one 14:01 < rxr> of the specific ones mentioned above, which means you need to enter 14:01 < rxr> all sorts of stuff like CONFIG_MEMORY_START because the config 14:01 < rxr> system doesn't already know what it is. You get a machine vector 14:01 < rxr> without any platform-specific code in it, so things like the RTC may 14:01 < rxr> not work. 14:01 < rxr> This option is for the early stages of porting to a new machine. 14:03 < rxr> seems you need a SH_BOARD config option and one for the detailed CPU ... 14:03 < valentin> rxr: of course i do an if $arch_machine = sh4 14:03 < rxr> there are some to choose from in the SH linux config ... 14:03 < valentin> i take a look what board i need for my DC 14:03 < valentin> but how will we handle that in general ? 14:04 < rxr> in the kernel one board option is dreamcast 14:04 < rxr> for those embedded platforms we need a board and detalied CPU selection .. 14:04 < rxr> not just sh1, sh2, sh3, sh4 ... 14:04 < valentin> yep 14:05 < rxr> btw. we can remove sh1 - glibc does not build for it - (jump targets out of range) and the linux kernel selector starts at sh2, too ... 14:07 -!- c4y0 [~c4y0@cable200-116-158-18.epm.net.co] has joined #t2 14:07 < c4y0> hi! 14:07 < rxr> btw you need a endian switch, too ... : 14:07 < rxr> CONFIG_CPU_LITTLE_ENDIAN: 14:08 < rxr> Some SuperH machines can be configured for either little or big 14:08 < rxr> endian byte order. These modes require different kernels. Say Y if 14:08 < rxr> your machine is little endian, N if it's a big endian machine. 14:08 < valentin> hm - how comes we do not have to tell glibc ? 14:09 < valentin> or the compiler even ? 14:09 < rxr> there is a big / little endian switch in the compiler .. 14:09 < rxr> I bet there is also some hidden config tweak in glibc to select it ... 14:10 < rxr> SH Options: ???mb ???ml 14:10 < rxr> the ??? is a unicode - ... 14:11 < rxr> ???mb Compile code for the processor in big endian mode. 14:11 < rxr> ???ml Compile code for the processor in little endian mode. 14:12 < valentin> hi c4y0 14:12 < valentin> so what is the default ? 14:12 < valentin> sh4 gets complicated *g* 14:13 < rxr> don't know ... 14:13 < valentin> but if we manage to crosscompile for some archmachine combination then we might attract sh* board users from #linux-sh 14:13 < rxr> dreamcast seems to be little endian - according to the kernel config 14:14 < rxr> sh4 kernel does not compile ... 14:14 < rxr> arch/sh/mm/cache-sh4.c:40:31: macro "p3_cache_init" passed 1 arguments, but take 14:14 < rxr> s just 0 14:14 < rxr> arch/sh/mm/cache-sh4.c:41: error: syntax error before '{' token 14:14 < rxr> and so on ... 14:14 -!- c4y0 [~c4y0@cable200-116-158-18.epm.net.co] has quit ["c4y0 has no reason"] 14:15 < rxr> sh3 compiles further ... 14:16 < rxr> yep - sh3 compiling quite straight - but maybe more problems later on ... 14:17 < valentin> damn 14:17 < valentin> somehow the hook did not get registered 14:17 < rxr> maybe some sh4 code drop whatever upstream tree injection are required 14:17 < valentin> [ $archmachine = sh4 ] && hook_add premake 5 "cat > configparms <<-EOT 14:18 < rxr> arch_machine ... 14:18 < rxr> ^ - $ 14:18 < valentin> grrrr 14:18 < valentin> btw: why <<-EOT and not < does not matter 14:20 < rxr> one srips leading tabs ... 14:26 < rxr> valentin: sh3 kernel built - now modules running 14:26 < rxr> Error logs from system-2.1.0-rc1-x86-pentium-mmx-desktop: 14:26 < rxr> [5] mail/spamassassin [5] multimedia/dvb-apps [5] e17/efsd [5] e17/evidence [5] gnome2/libgda [5] gnome2/bug-buddy [5] base/daemon [5] audio/audacity [5] audio/ardour [5] gnome2/gnomemeeting [5] gnome2/control-center [5] gnome2/totem 14:26 < rxr> [5] gnome2libgnomedb [5] develop/mono-doc [5] develop/mono-develop [5] gnome2/gdesklets [5] audio/rezound 14:26 < rxr> 681 builds total, 664 completed fine, 17 with errors. 14:27 < valentin> RFC: what about renaming src.$pkg.$timestamp.$cryptic1.$cryptic2 to src.$pkg.$config.$date..... and autocreate a symlink from src.$pkg.$config.current 14:27 < valentin> ? 14:28 < rxr> no symlink - the rest is fine ,-() 14:28 < rxr> http://www.superh.com/ 14:28 < valentin> because this enter highest number enter next number <....> sucks 14:28 < valentin> why no symlink ? 14:29 < rxr> noone needs it 14:29 < valentin> ok ... 14:29 < rxr> ls sorty my inode create time - and the date sorts nicely, too .. 14:29 < valentin> i would appreciate it 14:30 < rxr> http://www.superh.com/products/computing.htm 14:30 < rxr> http://www.superh.com/products/consumer.htm 14:31 < rxr> and so on 14:33 < rxr> The SH-5 is part of the upward compatible SuperH family and there is a huge range of third party products already available, SuperH offers a C/C++ toolchain based on the open source GNU technology. 14:33 < valentin> of course T2 offers a C/C++ toolchain for SuperH 14:34 < rxr> SH-6 and SH-7 14:34 < rxr> SuperH has previously announced it intends to deliver two new families, the SH-6 and SH-7 that it is currently defining in conjunction with its licensees and partners. 14:34 < rxr> yes - of course ... 14:35 < rxr> make[2]: *** [drivers/char/genrtc.o] Error 1 14:35 < rxr> some modules error out ... 14:35 < rxr> you can just mark them as disable-broken for sh until you need them ... 14:36 < rxr> but of course this is sh3 - sh4 fails early ... 14:38 < rxr> hehe - in formula1 they have bottles labelled with H2O 14:39 < rxr> *lol* sh64 has this config option: 14:39 < rxr> Fixup misaligned loads/stores occurring in user mode 14:39 < rxr> from the superh overview it seems to be a nice architecture ... 14:40 < rxr> (to have handy ...= 14:40 < rxr> ) even 14:41 < rxr> ok - now that I know sh builds in theory (have fun doing the details ,) I start a sh64 test build ,-) 14:42 < mnemoc_> re 14:42 < rxr> hi mnemoc_ ,-) 14:43 < mnemoc_> hi rxr 14:43 < valentin> rxr: i theory i can do almost everything, just a matter of the details 14:43 < valentin> hi mnemoc_ 14:43 < rxr> mnemoc_: Linux 2.6.11.5 14:44 < mnemoc_> rxr: will we ship microversions? 14:45 < rxr> yes 14:45 < rxr> of course ... ! 14:45 < rxr> I did not include .1 since it contianed a fix from me and I did not care much about the others ... 14:45 < rxr> but now iwth .5 we should start pulling them ,-) 14:45 < rxr> I can pull it ... 14:46 < mnemoc_> :) 14:49 < mnemoc_> rxr: http://www.geeks.cl/~amery/T2/regressions.html 14:49 < rxr> I need to continue math and electronics soon ... 14:50 < mnemoc_> uhm, mplayer still broken 14:50 < CIA-9> rene * r7578 /trunk/package/base/linux24/lx_config.sh: 14:50 < CIA-9> * fixed linux builds for architectures where the Makefile does not 14:50 < CIA-9> start working without a .config (sh64) 14:51 < mnemoc_> cannot find -ldha 14:52 < rxr> hm 14:53 < rxr> valentin: glibc: checking sysdep dirs... configure: error: The sh64 is not supported. 14:53 < valentin> ouhm 14:53 < mnemoc_> you may need to patch config.sub-s 14:53 < valentin> what about ulibc ? 14:54 < rxr> mnemoc_: In the case of a c library a simple configure patch is not enough to get a new architecture ... ,-) 14:54 < mnemoc_> oh, glibc 14:56 < CIA-9> amery * r7579 /trunk/package/emulators/wine/ (wine.conf wine.desc): * updated wine (20050111 -> 20050310) 14:56 < CIA-9> amery * r7580 /trunk/package/gnome2/control-center/no_update_desktop_database.patch: * fixed control-center by not running update-desktop-database 14:56 < rxr> from a short look uclibc supports sh64 ... 14:57 < CIA-9> amery * r7581 /trunk/package/security/dropbear/dropbear.conf: * fixed dropbear by installing scp as dbscp 14:58 < rxr> http://sourceforge.net/mailarchive/forum.php?forum_id=9482&max_rows=25&style=nested&viewmonth=200304 14:58 < rxr> Latest version of SH5-Linux Toolchain Availible (2003-04-14) 15:00 < mnemoc_> good market 15:00 < rxr> market ? 15:00 < mnemoc_> sh* people is a market 15:00 < mnemoc_> to sell our product 15:01 < rxr> hm - strange 15:01 < rxr> I can find sh64 occurances in glibc ... 15:01 * rxr trackign ... 15:05 < rxr> valentin: there are sh and shle as well as shbe ... 15:07 < valentin> maybe super64 oder superh sth ? 15:08 < rxr> Support for the sh64 architecture has been in Linux 2.4 since 2.4.22 and in 2.6 since 2.6.8. 15:08 < rxr> valentin: no - mayb no 64bit libc right now .. 15:10 < rxr> http://www.rpcurnow.force9.co.uk/sh64/gallery.html 15:10 < rxr> ^ - want one .. 15:11 < valentin> rxr: maybe it is sh5 in glibc ? 15:11 < rxr> nope ... all negative .. 15:11 < valentin> ok - nope 15:12 < valentin> well, sh5 could be nice once it has a libc 15:12 < rxr> uclibc ,-) 15:12 < mnemoc_> it can be experimental and you have to define something to use it? 15:13 < mnemoc_> uclibc ,-) <--- oh 15:14 < rxr> mnemoc_: nope - no sh64 code AFAICS in glibc ... 15:14 < CIA-9> rene * r7582 /trunk/architecture/sh64/ (archtest.out config.in gcc-options): 15:14 < CIA-9> * fixed the sh64 archtest.out (to list sh64 not sh ,-) 15:14 < CIA-9> * implemented the #TODO optimizations option to include sh5 ,-) 15:14 < rxr> si the amigaone sh64 ??? 15:15 < rxr> ah - no powerpc - just dump google hits ... 15:16 < mnemoc_> hehe 15:18 < rxr> ok - let's see how sh64-uclibc goes ... 15:19 < mnemoc_> 1477 builds total, 1342 completed fine, 43 with errors. 15:20 < mnemoc_> how many brokes did we had on beta3? 15:21 < mnemoc_> you want to announce beta4 iso-less? 15:21 < mnemoc_> or at least some minimals? 15:23 < rxr> the usual uclibc error on my athlon: 15:23 < rxr> tar -chf - include/* | tar -xf - -C /usr/src/t2-trunk/build/sh-test-2.1.0-rc1-sh 15:23 < rxr> 64-sh5-cross-generic-expert/usr/; 15:23 < rxr> tar: This does not look like a tar archive 15:23 < rxr> mnemoc_: iso less 15:23 < rxr> -rc1 next week ... 15:23 < rxr> with full desktop iso ... 15:24 < mnemoc_> ia64? 15:24 < mnemoc_> brb 15:24 < rxr> ia? no sh64 ... 15:27 < valentin> == 03/19/05 15:28:57 =[1]=> Finished building package glibc. 15:28 < valentin> ^- SH4 :) 15:28 < rxr> ,-) 15:30 < CIA-9> valentin * r7583 /trunk/package/base/glibc/glibc.conf: * set no-z-defs=yes on sh4 machines to avoid undefined references to __fpscr_values 15:31 < valentin> rxr: i just did an svn up - do i need to rebuild/reconfigure anything before proceeding with stage 1 ? 15:32 < rxr> don't think so ... 15:32 < rxr> just go ahead ... 15:33 < rxr> mnemoc_: seems to be buggy tar - Emerge-Pkg tar cured the uclibc build error ... 15:36 < mnemoc_> re 15:37 < mnemoc_> ia64? <--- are we ready for ia64 desktop cds? 15:38 < rxr> if you find one with a relevant CPU to build them ... 15:38 < rxr> in theory yes ... 15:38 < rxr> 0-binutils-uclibc-sh64 just built ... 15:38 < rxr> now the 0-gcc-uclibc-sh64 is progressing ... 15:38 < mnemoc_> mipe? 15:39 < rxr> ia64 or x86-64 ? 15:39 < rxr> ia64 is this Intel Intanium (2) thing - you know ? 15:41 < mnemoc_> :) 15:41 < mnemoc_> x86-64 15:42 < rxr> should be quite polished ... 15:42 < rxr> I think mipe will have a ISO handy some time soon ... 15:43 < valentin> rxr: what target board is it in your kernel config for sh4 ? 15:43 < rxr> unknown 15:43 < rxr> 0-gcc-uclibc-sh64 built 15:43 < rxr> ,-) 15:43 < mnemoc_> :) 15:45 < valentin> Some Gentoo enthusiasts, including Karl Trygve Kalleberg, is again sniffing at making a Gentoo port to the Dreamcast. 15:45 < valentin> guahrg 15:46 < valentin> they call it "porting" 15:46 < valentin> we don't port, we simply adapt 15:47 < valentin> as soon as we have some bootable stuff on the dc i'll push us to their newssite 15:47 < valentin> http://linuxdc.sourceforge.net/ 15:47 < rxr> as soon as ,-) 15:48 < valentin> rxr: now i stumbled over the kernel error 15:48 < valentin> need to set a board 15:49 < rxr> yep 15:50 < valentin> how can i do now ? 15:50 < rxr> as quick hack you could set it via a custom kernel config in the Config ... 15:50 < rxr> but I would rather just implement a board selection ... 15:51 < valentin> i can ask someone from linux-sh for a dreamcast ready config, so i have all the modules i need ... 15:51 < rxr> there is a dreamcast defconfig in the kernel dir ... 15:52 < valentin> ah - where ? 15:52 < rxr> arch/sh/defconfigs/ or so ... 15:52 < rxr> as usual ... 15:52 < valentin> do i have to copy it ? 15:53 < rxr> damn uclibc .. 15:53 < rxr> valentin: IMPLEMENT the board selection ... 15:53 < rxr> and then write code to copy the defconfig for the board selected ... 15:53 < valentin> there is no defconfigs dir 15:53 < rxr> if you write math for me on tu I could do that ... 15:53 < rxr> -def or -s or so .. 15:54 < rxr> just configs ? 15:54 < valentin> there is just one defconfig file 15:54 < rxr> arch/sh/configs/dreamcast_defconfig 15:54 < rxr> argh ... 15:54 < valentin> ah 15:54 < valentin> sorru 15:54 < valentin> /u/y 15:54 < valentin> go on and learn math ,) 15:55 < rxr> damn uclibc .. 16:11 < valentin> for the log 16:11 < valentin> http://kernel.xc.net/html/linux-2.6.0-test1/sh/ 16:13 < rxr> valentin: ? 16:14 < valentin> yep ? 16:14 < rxr> what is with that link ? 16:15 < valentin> just for copying the names of the boards 16:15 < rxr> make menuconfig ? 16:15 < rxr> or reading the arch/sh/Kconfig file ? 16:16 < rxr> did I mention how annoying uclibc make config is ... 16:16 < rxr> f*cking beast .. 16:17 < valentin> rxr: i know about make menuconfig 16:21 < mnemoc_> re 16:22 < mnemoc_> rxr: that's why i added the pkgsel-like config with globing 16:22 < rxr> that is what is not working right now ... 16:23 < mnemoc_> no? 16:23 < rxr> no 16:23 < mnemoc_> did you reviewed the .sed file? 16:24 < rxr> it is a bad time to ask this ... 16:24 < rxr> .o(if you do not wanna see me explode ...) 16:25 < mnemoc_> :) 16:28 < rxr> (this is not funny ...) 16:31 < mnemoc_> tar the files and i hunt it 16:31 < rxr> hunt a tar ? 16:32 < mnemoc_> hunt a bug 16:32 < mnemoc_> on the files you tar 16:37 < rxr> the problem is that you only substitute "not defined" for X options to =y ... 16:37 < rxr> and rely that they show up 16:37 < rxr> however if they are not yet in the file uclibc picks up the default for that value 16:37 < rxr> and in my case I would need 4 runs to be able to substitude it in the 4th run (when it first shows up) to be set ... 16:38 < mnemoc_> .oO 16:38 < valentin> rxr: how can i tell our build system to use a config that comes from inside the kernel tree ? 16:38 < rxr> I think X options should be in the file as early as possible ... 16:38 < rxr> not yet 16:38 < valentin> or should i mess in the kernel's .conf ? 16:38 < rxr> I would recomment that you inject code into linux24/lx_config.sh to choose the default one from the board selection on sh* 16:39 < rxr> there is already code to use the initial defconfig as base ... 16:39 < mnemoc_> why not on architecure/parse-config ? 16:39 < rxr> just that some embedded arches need one specific defconfig ... 16:39 < valentin> ok 16:39 < rxr> yes - that would also be possible and more modular - however I think one need to tweak the lx_config.sh anyway to make the config creation more modular ... 16:39 < mnemoc_> rxr: injecting the value without the loops may produce inconsistent info 16:40 < rxr> so a arch parse-config could hook in cleanly ... 16:40 < mnemoc_> rxr: because that option may enable/disable another things 16:40 < rxr> sure some looping to stabelize ... but injecting them would make life easier and the config stabelize more quickly ... 16:41 < mnemoc_> rxr: ok 16:41 < rxr> I try if rasiing from 3 (I already increased it) to 4 makes the config appear as I think it should ... 16:41 < rxr> n loop count ... 16:43 < mipe> oh hi all 16:43 < valentin> hi mipe 16:43 < rxr> welcome mipe 16:43 < mnemoc_> mipe! :) i was just asking for you :p 16:44 < valentin> some more stupid questions: 16:44 < mipe> mnemoc_: :) 16:44 < rxr> mnemoc_: oh - maybe 3 loops are enough as I initially thought - I overlooked that I edited onyl the stage0 auto_config - not the 2nd one 16:44 < mnemoc_> valentin: i think i hate you for patching glibc 16:44 < valentin> first either our t2 architecture kernel.conf or depconfig is copied into .config 16:44 < mnemoc_> rxr: nice 16:44 < valentin> so far so good 16:45 < valentin> then the depconfig is "merged" whatever that means 16:45 < valentin> mnemoc_: what ? i did not patch glibc 16:45 < mnemoc_> mipe: how far of a x86-64 desktop iso are we? 16:45 < mnemoc_> Author: valentin 16:45 < mnemoc_> Date: 2005-03-19 15:30:19 +0100 (Sat, 19 Mar 2005) 16:45 < mnemoc_> New Revision: 7583 16:45 < mnemoc_> trunk/package/base/glibc/glibc.conf 16:45 < mnemoc_> 8 ++++++++ 16:46 < mnemoc_> now i'll have to wait 3 hours to continue my reference build :\ 16:46 < valentin> mnemoc_: that is not 'patching glibc' 16:46 < mnemoc_> glibc package :) 16:46 < valentin> so what is the problem with that ? 16:46 < mnemoc_> i'll have to wait 3 hours to continue my reference build 16:46 < mipe> mnemoc_: i have built mnemosyne few days ago, system builds arent that big of concern right now,more about bootdisk 16:47 < rxr> yeah - and then we know valentin did not harm x86 ,-) 16:47 < mnemoc_> :) 16:47 < valentin> mnemoc_: there should be no difference when building for anything != sh4 16:47 < mnemoc_> it's just because you changed the checksum of the package and it will be auto-rebuilt 16:48 < mnemoc_> but don't worry 16:48 < mipe> mnemoc_: but sadly i dont have time to look at bootdisk(broken gcc) until next break(next week), atm currently swaped with school stuff(2 physics report+exams). 16:48 < valentin> mnemoc_: sorry :p 16:48 < mnemoc_> mipe: oh 16:48 < rxr> hm - damn - uclibc seems to be not 64 bit - but some stange combination or pure 32bit - I think ... 16:48 < mnemoc_> rxr: /join #uclibc and ask andersee 16:49 < rxr> no - in theory I need to learn for exams, too 16:49 < mnemoc_> valentin: try to do sh* specific stuff on architecute/sh*/ :) 16:49 < mnemoc_> me too :\ 16:49 < rxr> ouhm - well, ... 16:49 < valentin> mnemoc_: i do but for glibc there is to much stuff in there already 16:50 < mnemoc_> valentin: i know :\ 16:50 < valentin> but glibc.conf is simple compared to the linux config stuff 16:50 < mnemoc_> i guess i'll add support for architecture/*/pkg_${foo}_{pre,post}.conf 16:51 < rxr> grummel ... 16:51 < mnemoc_> uhm? 16:51 < rxr> mnemoc_: not you - uclibc - ... 16:51 < mnemoc_> :) 16:51 < rxr> mnemoc_: do the architecture thingi ... 16:51 < valentin> ah i just need to set ROCKCFG_PKG_LINUX_CONFIG_STYLE=none 16:51 < valentin> to skip that auto config mess ? 16:52 < rxr> valentin: yes - as I said - but that would be a first step thing - later on T2 should ahve SH board selection ... 16:52 < valentin> rxr: i have a selection 16:52 < rxr> valentin: mess ? - I call this fully automated kernel config generation ... 16:52 < rxr> ah - yes - you could also do this now without tweaking the linux packages ... 16:52 < valentin> the only thing to do is to copy arch/sh/configs/$ROCK_BOARDDEF-conf to .config 16:52 < rxr> set that config from withint a sh/parse-config and define the config to use ... 16:53 < rxr> yep ... 16:53 < valentin> i thought i do in lx_config() { ... } 16:54 < rxr> ok there for now until we factor more arch related hacks out of the package later 16:54 < mnemoc_> valentin: adding the exceptions directly on the packages turns them into the mess they are 16:55 < valentin> mnemoc_: i know .... but i am such a newbie to such things i have no idea what to do else 16:55 < mnemoc_> opt/gnome2/share/mime/*: gdesklets glabels :( 16:56 < mnemoc_> valentin: i'll do {pre,post}.conf think for architectures and then we can move arch exceptions there 16:57 < valentin> mnemoc_: do so and i'll remove the hack from the glibc.conf 16:57 < mnemoc_> :) 16:57 < mnemoc_> that's the idea :) 16:57 < valentin> and then you'll waste another 3h :P 16:58 < mipe> oh and latest changes i have made to gcc.conf are at http://koti.japo.fi/~pmika/tmp/gcc_tmp.patch 16:58 < rxr> .oO(la la la - I'm so crazzy ...) 16:58 < mipe> because of really ugly changes at gcc.conf for arch specific,i'd like some form of $arch/pkg_$pkg_post stuff 16:59 < mnemoc_> :) 16:59 < mnemoc_> i'm on it 17:00 < rxr> hehehe - we could finally remove the bare ugly hppa --enable-hacker-mode hack from glibc then ,-) 17:01 < mipe> but me back to write *something* on Einstein's general relativity :( 17:01 < valentin> mnemoc_: when will theese {pre,post}.conf stuffs be evaluated 17:01 < valentin> mipe: the general ? .oO 17:02 < valentin> the special is much easier 17:02 < mnemoc_> valentin: pre, just before $pkg.conf, and post right after 17:02 < mnemoc_> target independant 17:02 < valentin> ok, so i can add some hooks right there ? 17:02 < mnemoc_> yep 17:03 < mnemoc_> generaly on pre you alter ROCKCFG stuff, and on post you add hooks 17:03 < mnemoc_> but there is no policy about 17:03 < valentin> mnemoc_: ok 17:05 < CIA-9> rene * r7584 /trunk/package/base/linux-header/linux-header.conf: 17:05 < CIA-9> * added dual header generation for sh64 to linux-header as done for 17:05 < CIA-9> sparc64 and x86-64 17:05 < valentin> rxr: for arch/sh the defconfig and the configs/* are unmergable 17:05 < CIA-9> amery * r7585 /trunk/scripts/Build-Tools: * added architecture/$arch/pkg_*_{pre,post}.conf support (NOT TESTED!) 17:06 < rxr> valentin: ? 17:06 < CIA-9> rene * r7586 /trunk/package/base/uclibc/uclibc.conf: 17:06 < CIA-9> * added config tweaks for the ugly uclibc sh64 sub arch config (bah) 17:06 < CIA-9> * increased looping over the config from 2 to 3 to let the ARCH 17:06 < CIA-9> config settle for sh64 17:07 < rxr> Seit gestern abend sind X-Org packages f? 1/4 r Debian GNU/Linux auf PowerPC verf? 1 1/4 gbar. 17:07 < mnemoc_> rxr: some day i would like to discuss about porting uclibc/busybox configurator to kernel 17:07 < mnemoc_> rxr: ? 17:07 < rxr> ^ - *lol* in T2 we have - them - since - since - well - since the release of 6.8.2 ... 17:08 < mnemoc_> Debian isn't famous due to their speed doing things 17:08 < rxr> sure not ... that is why we have t2 for ,-) 17:08 < mnemoc_> :) 17:08 < mnemoc_> .oO( what's slower, debian or r*ck? ) 17:09 < rxr> no idea ... 17:09 < rxr> in former time debian - but nowadays ... 17:09 < rxr> [V] 6.8.1 17:09 < rxr> ROCK ;-) 17:09 < mnemoc_> :) 17:10 < rxr> http://www.rocklinux.net/submaster/smadm.cgi?i=2005021223461604959 17:10 < rxr> they have it since 2005-02-12 lingering in sumpf-master ? 17:10 < rxr> *lol* 17:11 < mnemoc_> *g* 17:12 < mnemoc_> they have a tiny bottleneck 17:12 < rxr> yeah - tiny + the lack of people 17:13 < mnemoc_> we have lack of people too 17:13 < rxr> somehow they manage to chase away all the known people ... 17:13 < mnemoc_> yep 17:13 < rxr> we have nearly 15 write accessors ,-) increasing every week ... 17:13 < rxr> looks quite fine .-) 17:13 < mnemoc_> where is 1.4 people? 17:14 < mnemoc_> gentoo? 17:14 < mipe> hmm,when is next relase planned? 17:14 < mipe> release 17:14 * mipe back in few mins 17:14 < mnemoc_> beta4? today :p 17:14 < rxr> mipe: ? 17:14 < rxr> mnemoc_: ? even - what gentoo people ? 17:14 < mnemoc_> rc1 next week i hope 17:14 < mnemoc_> rock-1.4 people 17:15 < mnemoc_> are they gone to gentoo? 17:15 < rxr> hehe - no idea ... 17:15 < rxr> one is at gentoo ... 17:15 < rxr> and retired there also ... 17:15 < rxr> the others ... 17:15 < mnemoc_> .oO( why the heck gnome-media doesn't find orbit2 )o 17:15 < rxr> well - doing other stuff I think ... 17:16 < mnemoc_> r*ck has a big tradition loosing people 17:17 < rxr> hm - in the past ucibc did cross built, right ? 17:17 < rxr> me testing with arm ... 17:17 < rxr> on sh64 the kernel headers are not found ... 17:17 < mnemoc_> i haven't touched uclibc in monthes 17:17 < rxr> and I'm off learning again in some secs ... 17:18 < rxr> arm build running 17:18 * rxr back to electronics and math ... 17:19 * mnemoc_ back to paid work 17:19 < rxr> t2 has 12 write comitters ... 17:19 < rxr> just wc -l'ed them 17:20 < mnemoc_> considering login renames? 17:20 < valentin> is there anything that sets ROCKCFG_PKG_LINUX_CONFIG_STYLE ? 17:22 < mnemoc_> linux24/config.in ? 17:22 < valentin> oh 17:23 < valentin> 17:05 < valentin> rxr: for arch/sh the defconfig and the configs/* are unmergable 17:23 < mnemoc_> valentin: can you see if $arch/pkg_*_{pre,post}.conf is working? 17:24 < valentin> because they contain contradictionary choices 17:24 < valentin> mnemoc_: i did not try yet 17:24 < mipe> uhm rc1 next week? system(mnemosyne) i can build without any errors but on bootdisk, theres something wrong with dietlibc and gcc 17:24 < mnemoc_> :\ 17:25 < valentin> mnemoc_: i can re-write my glibc thing, if you do not mind that checksum changes 17:25 < mnemoc_> mipe: we wanted last week but men/hour were not enough 17:25 < rxr> mipe: I can take a look next week 17:25 < mnemoc_> valentin: go for it 17:25 < mipe> rxr:ok 17:25 < valentin> ok 17:25 < mnemoc_> i'm still at 7577 :p 17:25 < valentin> so i just add architecture/pkg_glibc_post.conf, right ? 17:26 < mnemoc_> architecture/sh*/pkg_glibc_post.conf 17:26 < valentin> yep :D 17:26 < mnemoc_> :) 17:26 < mnemoc_> valentin: tools need to be rebuilt 17:27 < mnemoc_> valentin: if they are not auto-rebuilt, force it 17:27 < rxr> mnemoc_: if they do not auto rebulid there is a bug, no ? 17:31 < mnemoc_> yes 17:31 < mnemoc_> newer files on architecture/$arch must be detected 17:31 < valentin> mnemoc_: tools beeing rebuilt 17:31 < mnemoc_> :) 17:31 < valentin> -> Creating tools.native/lib/pkg_*_{pre,post}.conf 17:31 < valentin> but nothing happens after that 17:32 < valentin> mom 17:32 < mnemoc_> see if tools.native/lib/pkg_glibc_post.conf has your file 17:33 < valentin> you mean tools.cross ? 17:33 < valentin> mom 17:33 < valentin> sorry 17:33 < valentin> of course native 17:33 < mnemoc_> :) 17:34 < mnemoc_> i just copy&paste what you pasted :p 17:34 < valentin> but there are only the three normal files in there 17:34 < mnemoc_> uhm? 17:34 < valentin> -rw-r--r-- 1 root root 952 Mar 19 17:32 parse-config 17:34 < valentin> -rw-r--r-- 1 root root 45 Mar 19 17:32 parse-config-fixarch 17:34 < valentin> -rw-r--r-- 1 root root 212 Mar 19 17:32 wrapper-config 17:35 < mnemoc_> find build/$foo/TOOLCHAIN -name pkg_glibc_post.conf 17:35 < mnemoc_> uhm 17:36 < mnemoc_> how can you have no {pre,post}.conf file?? 17:36 < valentin> im investigating 17:36 < valentin> there is no pkg_glibc_post.conf in build/ at all 17:36 < mnemoc_> fl_wrapper.so 17:36 < mnemoc_> parse-config 17:36 < mnemoc_> parse-config-fixarch 17:36 < mnemoc_> pkg_coreutils_post.conf 17:36 < mnemoc_> pkg_e2fsprogs_post.conf 17:36 < mnemoc_> pkg_gcc2_pre.conf 17:36 < mnemoc_> .... 17:37 < valentin> what dir ? 17:37 < mnemoc_> build/ref-2.1.0-rc1-x86-pentium4-generic-expert/TOOLCHAIN/tools.chroot/lib/ 17:37 < mnemoc_> TOOLCHAIN/$toolsdir/lib 17:37 < mnemoc_> depending on your stage 17:38 < valentin> i am at stage 1 17:38 < valentin> but i only got tools.croos in the cfgs dir 17:39 < mnemoc_> Creating tools.native/lib/pkg_*_{pre,post}.conf <-- that doesn't lie 17:39 < mnemoc_> valentin: can you add set -x on Build-Tools:285 17:40 < mnemoc_> and set +x on 297? 17:40 < mnemoc_> set -x; pwd; on 285 bettwe 17:40 < mnemoc_> better 17:43 < valentin> mnemoc_: my only tools.native is in TOOLCHAIN/tools.native 17:43 < mnemoc_> yes 17:44 < mnemoc_> tools are all inside $xroot/TOOLCHAIN/ 17:44 < mnemoc_> where are those three files you showed? 17:45 < valentin> in build/TOOLCHAIN/tools.native/lib 17:45 < valentin> if i understand your script it puts those stuff to 17:45 < valentin> build/$ROCKCFG_ID/TOOLCHAIN/$toolsdir/lib/ 17:45 < mnemoc_> yes 17:46 < mnemoc_> xroot=build/$ROCKCFG_ID 17:47 < mnemoc_> .oO( is ROCKCFG_ID somehow empty? )o 17:47 < valentin> and toolsdir == ? 17:47 < mnemoc_> tools.{native,cross,chroot} 17:48 < valentin> but why did it print Creating tools.native/lib/pkg_*_{pre,post}.conf 17:48 < valentin> ? 17:49 < mnemoc_> echo_status "Creating $toolsdir/lib/pkg_*_{pre,post}.conf" 17:49 < valentin> yes - but it should be tools.cross ?! 17:49 < mnemoc_> build/$ROCKCFG_ID/TOOLCHAIN/$toolsdir/lib/ is a bit long 17:49 < mnemoc_> .oO 17:49 < valentin> i know 17:50 < valentin> there is no tools.native inside that TOOLCHAIN, so that is why the script failed 17:51 < mnemoc_> if [ $stagelevel -le 1 -a "$ROCKCFG_USE_CROSSCC" = 1 ] 17:51 < mnemoc_> uhm 17:51 < mnemoc_> scripts/parse-config:62 17:53 < valentin> moment 17:53 < valentin> sth is quite broken here 17:53 < valentin> i inserted an echo_status "build/$ROCKCFG_ID/TOOLCHAIN/$toolsdir/" 17:53 < valentin> -> build//TOOLCHAIN/tools.native/ 17:53 < valentin> but i did a -cfg dream 17:54 < mipe> tools.native? what stage is that created? 17:54 < valentin> mipe: the output is wrong anyway in this case 17:55 < mipe> yep,i was wondering i never had tools.native 17:55 < valentin> you have tools.native when merging into your systems for example 17:55 < valentin> but not in build/TOOLCHAIN 17:55 < mipe> ah ic 17:55 < valentin> seems like it did not parse the config 17:56 -!- martin_ [~martin@pD9E73B14.dip.t-dialin.net] has joined #t2 17:56 < valentin> hi martin_ 17:56 < martin_> hi 17:56 < martin_> hi valentin 17:57 < mnemoc_> i'll add some harden to be sure $ROCKCFG_ID has content 17:58 < valentin> mnemoc_: but what went wrong in my case ? everything worked fine up to the last svn up 17:58 < mnemoc_> ./scripts/Config -oldconfig ? 18:00 < valentin> mnemoc_: why ? 18:00 < mnemoc_> to regenerate your config :) 18:01 < valentin> but if i run ./scripts/Config -cfg dream everything looks as it should 18:01 < CIA-9> valentin * r7587 /trunk/ (2 files in 2 dirs): * moved sh4 specific hook into architecture using the new UNTESTED pkg_*_post feature 18:03 < valentin> hm, now it works ????? 18:03 < valentin> damn 18:03 < valentin> ok 18:03 < valentin> but the hook is not listed inside debug.hooks 18:04 < mnemoc_> .oO 18:04 < valentin> and the build fails 18:04 < valentin> mom 18:05 < valentin> the file is not there 18:06 < mnemoc_> ? 18:06 < valentin> and why the hell do i have both build/cfg-dream*{,-expert} ?? 18:07 < mnemoc_> you enabled expert :) 18:07 < valentin> no 18:07 < valentin> yes 18:07 < valentin> but why did it change ? 18:08 < valentin> so i had -expert 18:08 < valentin> but now it is 18:08 < mnemoc_> uhm? 18:08 < valentin> drwx------ 6 root root 152 Mar 19 18:05 dream-2.1.0-rc1-sh-sh4-cross-desktop 18:08 < valentin> drwx------ 18 root root 440 Mar 19 09:10 dream-2.1.0-rc1-sh-sh4-cross-desktop-expert 18:08 < valentin> take a look at the times 18:09 < valentin> 18:05 was localtime - 4 minutes 18:10 < valentin> ok, i re-enabled expert 18:11 < mnemoc_> your config got corrupted i guess 18:12 < valentin> root@localhost:/usr/src/t2-trunk# ls build/dream-2.1.0-rc1-sh-sh4-cross-desktop-expert/TOOLCHAIN/tools.cross/lib 18:12 < valentin> . .. fl_wrapper.so gcc libiberty.a parse-config parse-config-fixarch wrapper-config 18:12 < valentin> so still no glibc stuff 18:13 < valentin> what is ${y##*/} ? 18:14 < valentin> ah !!! 18:14 < valentin> nope 18:14 < mnemoc_> s,^.*/\([^/]*\),\1, 18:16 < valentin> ouhm 18:16 < valentin> why y ? 18:16 < valentin> you mean x ? 18:16 < mnemoc_> oh 18:17 < mnemoc_> wrong copy&paste 18:17 < valentin> for x in `echo architecture/$arch/pkg_*_{pre,post}.conf`; do 18:17 < valentin> if [ -f $y ]; then 18:17 < valentin> echo " . $y" >> build/$ROCKCFG_ID/TOOLCHAIN/$toolsdir/lib/${y##*/} 18:17 < valentin> fi 18:17 < valentin> hehe 18:17 < valentin> done 18:17 < mnemoc_> x :) 18:17 < valentin> just change the first x into y 18:17 < mnemoc_> easier 18:17 < valentin> should i do it now ? or do you want to commit ? 18:18 < mnemoc_> test and commit :) 18:19 < valentin> mom 18:21 < valentin> now the file is inside lib 18:22 < valentin> ok, the hook is there 18:24 < CIA-9> valentin * r7588 /trunk/scripts/Build-Tools: * changed one x to y in the creation of $toolsdir/lib/pkg_*_{pre,post}.conf 18:24 < valentin> so your new implementation passed the first test 18:24 < mnemoc_> :) 18:25 < valentin> i made more changes in t2 today than in the last month :/ 18:30 < valentin> cool - with your new preconf i can easily work around the merging problem 18:31 < CIA-9> amery * r7589 /trunk/package/gnue/gnue-forms/fix_curses.patch: * fixed gnue-forms to run curses ui properly 18:32 < mnemoc_> :) 18:35 -!- jeru [~jeru@p54BFFCD0.dip.t-dialin.net] has quit [Read error: 110 (Connection timed out)] 18:36 < valentin> where is $base/config/$config/linux.cfg pointing to - i.e where is base ? 18:37 < valentin> my t2-trunk dir ? 18:38 < mnemoc_> yes 18:38 < mnemoc_> base is your tree-root 18:50 < valentin> damn, if there is a custmain, can i add a hook to be executed before that ? 18:50 < mnemoc_> hook_eval hookname 18:51 < mnemoc_> preconf if runconf!=0 18:51 < mnemoc_> {pre,post}patch if autoextract!00 18:55 < valentin> mnemoc: if i do a hook_add preconf, it will be evaluated even if there is a custmain ? 18:55 < mnemoc_> yes 18:55 < mnemoc_> if runconf is not set to 0 18:55 < mnemoc_> mom 18:55 < mnemoc_> no sorry 18:55 < valentin> it is not set to 0 18:55 < mnemoc_> postpatch it is 18:56 < mnemoc_> but the proper way is to add preconf, premake, inmake, postmake, where correspond inside the custmain 18:56 < valentin> yep 18:57 < valentin> i do not need the hook, found another way 18:57 < rxr> and continue to remove unneeded custmains in general .. 18:57 < mnemoc_> :) 18:57 < valentin> but i need to tweak two lines in lx_config.sh 18:57 < rxr> what package are you add ? 18:58 < rxr> linux should have proper hooks - doesn't it ? 18:58 < mnemoc_> ye 18:58 < mnemoc_> s 18:58 < rxr> linux24.conf: hook_eval premake 18:58 < rxr> linux24.conf: hook_eval inmake 18:58 < rxr> linux24.conf: hook_eval postmake 18:58 < rxr> lx_config.sh: hook_eval prepatch 18:58 < rxr> lx_config.sh: hook_eval postpatch 18:58 < rxr> lx_config.sh: hook_eval preconf 18:58 < rxr> valentin: what do you wanna patch ? 18:58 < valentin> rxr: i need the hock before the ROCKCFG_PKG_LINUX_CONFIG_STYLE is evaluated 18:58 < valentin> but i do not need it anymore 18:58 < valentin> mom 18:59 < valentin> rxr: i paste what i do in here before i commit, ok ? 18:59 < valentin> so you can take a look if you like that 19:00 < rxr> sure 19:01 < rxr> does anyone here know how to calculate the error on the output voltage due to input voltage offset on an operational amplified? 19:01 < rxr> opamp ... 19:01 < rxr> jsaw: you maybe ? 19:02 < rxr> should be trivial - but I'm unsure if the tiny calculation I did is correct ... 19:02 < valentin> rxr: 19:02 < valentin> =================================================================== 19:02 < valentin> --- package/base/linux24/lx_config.sh (revision 7586) 19:02 < valentin> +++ package/base/linux24/lx_config.sh (working copy) 19:02 < valentin> @@ -210,9 +210,11 @@ 19:02 < valentin> echo "Generic linux configuration ..." 19:02 < valentin> hook_eval preconf 19:02 < valentin> if [ "$ROCKCFG_PKG_LINUX_CONFIG_STYLE" = none ] ; then 19:02 < valentin> - echo "Using \$base/config/\$config/linux.cfg." 19:02 < valentin> + [ -z "$ROCKCFG_PKG_LINUX_CUSTOMCONF" ] && \ 19:02 < valentin> + ROCKCFG_PKG_LINUX_CUSTOMCONF=$base/config/$config/linux.cfg 19:02 < valentin> + echo "Using $ROCKCFG_PKG_LINUX_CUSTOMCONF." 19:02 < valentin> echo "Since automatic generation is disabled ..." 19:02 < valentin> - cp -v $base/config/$config/linux.cfg .config || true 19:02 < valentin> + cp -v $ROCKCFG_PKG_LINUX_CUSTOMCONF .config || true 19:02 < valentin> else 19:02 < valentin> echo "Automatically creating default configuration ...." 19:02 < valentin> auto_config 19:03 < valentin> so i just take what is inside ROCKCFG_PKG_LINUX_CUSTOMCONF which i set to $base/config/$config/linux.cfg by default 19:03 < mnemoc_> that doesn't need to be a ROCKCFG_* 19:03 < mnemoc_> lx_customconf is good enought 19:03 < rxr> but fine so, too ... 19:03 < mnemoc_> set it on top of lx_config 19:04 < mnemoc_> and alter it on _post.conf 19:04 < mnemoc_> two lines :) 19:04 < valentin> mnemoc_: but the variable is set in _pre.conf 19:04 < rxr> mnemoc_: we need to sed ROCKCFG to CONFIG or CFG or so at some point ... 19:04 < rxr> valentin: you could also set it on post ... 19:04 < rxr> and just set it on top of lx_config.in ... 19:04 < valentin> rxr: you are right 19:05 < valentin> rxr: why on top ? 19:05 < mnemoc_> valentin: instead of seeing if it's set, you set it always and overwrite 19:05 < rxr> if I would be so sure about damn operational amplifiers ... 19:05 < mnemoc_> i don't remember anything of electronics 19:05 < rxr> valentin: if you do it in the function you will overwrite your arch/*/_post set value during functional evaluation 19:05 < mnemoc_> but there is a handbook somewhere 19:06 < valentin> rxr: ? 19:06 < mnemoc_> lx_config:1 lx_customconf=$base/config/$config/linux.cfg 19:07 < mnemoc_> arch/*/_post.conf:1 lx_customconf=blah 19:07 < valentin> ok 19:07 < mnemoc_> and on the function you just use lx_customconf 19:07 < valentin> ok 19:07 < mnemoc_> _but_ i consider illegal to override a config value on target 19:07 < mnemoc_> on arch 19:08 < mnemoc_> config -> target -> arch -> extenders -> package 19:08 < rxr> ok - valentin do as 19:08 < rxr> above 19:08 < rxr> but do not use the custom config thing .. 19:08 < mnemoc_> .oO( what is above? )o 19:08 < rxr> but tweak the defconfig to be used as base ... 19:08 < rxr> mnemoc_: your proposal 19:08 < rxr> like so: 19:09 < rxr> mom 19:09 < valentin> 19:08 < rxr> but tweak the defconfig to be used as base ... 19:09 < valentin> what do you mean ? 19:09 < rxr> second 19:11 < rxr> --- lx_config.sh (revision 7578) 19:11 < rxr> +++ lx_config.sh (working copy) 19:11 < rxr> @@ -19,6 +19,7 @@ 19:11 < rxr> +lx_defconfig="arch/$lx_cpu/defconfig arch/$lx_cpu/configs/common_defconfig" 19:11 < rxr> - echo " merging (system default): 'arch/$lx_cpu/defconfig'" 19:11 < rxr> - grep -h '^CONF.*=y' arch/$lx_cpu/defconfig arch/$lx_cpu/configs/common_defconfig | 19:11 < rxr> - cut -f1 -d= | 19:11 < rxr> - while read tag ; do 19:11 < rxr> + echo " merging (system default(s)): $lx_defconfig" 19:11 < rxr> + grep -h '^CONF.*=y' $lx_defconfig | cut -f1 -d= | while read tag 19:11 < rxr> + do 19:11 < mnemoc_> lx_defconfig has to be outside to be modified 19:12 < rxr> yes - I forgot the paste the 2nd chunks @@ ... 19:12 < mnemoc_> :) 19:12 < rxr> the lx_defconfig at the top - the rest in the config function ... 19:12 < rxr> I can commit ... 19:12 < valentin> rxr: this will conflict my change 19:13 < rxr> but it will allow cleanly what you want - define the defconfig for your board ... 19:13 < mnemoc_> rxr: i like it 19:13 < rxr> just that other T2 arch defined configs are merged in as usual ... 19:14 < rxr> if you do not want that - (normally you want that) we can add a 4th option to the config style: "Plain defconfig" ,-) 19:14 < CIA-9> rene * r7590 /trunk/package/base/linux24/lx_config.sh: 19:14 < CIA-9> * added lx_defconfig variable to lx_config.sh to be able to overwrite 19:14 < CIA-9> the defconfig to merge e.g. on a per architecture or target base 19:18 < valentin> rxr: so now i just overwrite the lx_defconfig variable in my _post and leave CONFIG_STYLE set to auto ? 19:19 < mnemoc_> rxr: http://www2.ing.puc.cl/dirdoc/index.phtml?url=normasproc/alumnos/examen_comun/FE-ReferenceHandbook.pdf 19:19 < mnemoc_> valentin: yes 19:20 < valentin> rxr: good, but ... if i read the patch you posted, that variable is set inside the function and not top of script 19:20 < rxr> valentin: svn up - it is at the top 19:21 < rxr> I forgot pasting the 2nd @@...@@ mark .. 19:22 < mnemoc_> rxr: that handbook shall have your answers 19:22 < rxr> mnemoc_: thanks ... - s.th. I could hand to that shiny laser printer ... 19:24 < CIA-9> valentin * r7591 /trunk/package/base/linux24/lx_config.sh: 19:24 < CIA-9> * allow to overwrite the configs kernel location by setting 19:24 < CIA-9> lx_customconf 19:24 < mnemoc_> uhm? 19:25 < valentin> ouh - kernel location <-- kernel config 19:26 < mnemoc_> i thought you would use lx_defconfig 19:26 * rxr too 19:26 < valentin> mnemoc_: rxr : i do now 19:26 < valentin> but i wanted to commit that anyway 19:26 < rxr> lx_customconf just for fun ? 19:26 < rxr> ic ... 19:27 < mnemoc_> :) 19:27 < valentin> raise the commit graph line for 1% 19:28 < mnemoc_> :D 19:28 < rxr> and here goes the next 19:28 < valentin> ouhm mom 19:28 < CIA-9> rene * r7592 /trunk/package/base/linux24/lx_config.sh: * removed stay newline and prettified the status output 19:28 < valentin> rxr: did i remove the comment above or did you ? 19:28 < rxr> I did - sorry 19:28 < valentin> you did 19:28 < rxr> fixed 19:29 < valentin> hehe 19:29 < CIA-9> rene * r7593 /trunk/package/base/linux24/lx_config.sh: * fixed comment defect I caused 19:29 < rxr> mnemoc_: your book has not the direct answer - does not even talk about offset or bias errors ... 19:29 < mnemoc_> :( 19:30 < rxr> of course one could solder the equation together based on the equation there ... 19:30 < mnemoc_> damn 19:30 < rxr> anyway - do not care - do you work and I do electronic tinkering ... 19:30 < mnemoc_> :) 19:30 < rxr> but thanks anyway - maybe I need s.th. of the other topics covered in the book 19:30 < mnemoc_> i'll do lunch :p 19:33 < rxr> and I evening dinner some time soon ... 19:33 < valentin> me too 19:33 < rxr> valentin: when you use the defconfig as base as I proposed you get all modules enabled 19:33 < valentin> rxr: i just found out 19:34 < valentin> but that sucks in my case ... anyway .... 19:34 < rxr> you need to list the defect ones in the arch.../sh/kernel26-disable.lst 19:34 < valentin> rxr: ack 19:34 < rxr> no - then we know which ones built on sh ... 19:34 < valentin> first error 19:34 < rxr> it is always nice to have them handy - otherwise you will miss tun there, ppp next and so on ... 19:35 < rxr> sh4 errored out quite quick here ... 19:35 -!- mnemoc_ is now known as mnemoc 19:35 < rxr> sh built further ... 19:35 < valentin> rch/sh/kernel/cpu/sh4/sq.c:348: error: unknown field `unmap' specified in initializer 19:35 < rxr> sh4 has some deep defects and seems to be a bit rotten in the vanilla tree 19:35 < valentin> arch/sh/kernel/cpu/sh4/sq.c:348: warning: initialization from incompatible pointer type 19:35 < valentin> arch/sh/kernel/cpu/sh4/sq.c:349: error: unknown field `sync' specified in initializer 19:35 < valentin> arch/sh/kernel/cpu/sh4/sq.c:349: warning: initialization from incompatible pointer type 19:35 < valentin> hm 19:35 < rxr> I think I had something like that, too 19:36 < rxr> either try to pull fixes from the up-stream tree or patch it yourself 19:36 < mnemoc> 2.6.11.5? 19:36 < rxr> sh3 built - at least the whole kernel binary - only modules did not ... 19:36 < rxr> I doublt that contains sh fixes ... 19:36 < mnemoc> bbl 19:36 < mnemoc> me too :p 19:37 < valentin> what exactly is bbl ? 19:37 < mnemoc> be back later 19:37 < valentin> ah 19:38 < valentin> i thought something like be back (after) lunch 19:38 < mnemoc> :p 19:40 < rxr> back before lunch 19:45 < rxr> oh - I just notice w/ ngem 19:46 < rxr> ngerman Hoch-/Tifpa becomes Hoch-/Tiefpass ,-) 19:46 < rxr> first is sz of course - if it does not show up on your side *g* 19:49 < rxr> hehehe - I have and old exam here ... 19:50 < rxr> the question with some a/d convertion is - what is the final bit combination on b3 b2 b1 b0 at the end of the conversion - the student wrote 6,5 19:50 < rxr> and the comment from the corrector: That is not a bit combinator for a digital value ! 19:51 < mipe> :) 19:51 < rxr> I want those questions on monday .. 19:51 < rxr> damn - I bet I get more difficult ones - as usual .. 19:51 < rxr> well - that was only one point out of 45 anyway - but ... 19:53 < martin_> Hi rxr, it is EN? 19:53 < rxr> hi martin_ yes - don't tell me it is your exam I'm at ,-) 19:54 < rxr> martin_: how are you - how is the new ini raum going ... ? 19:54 < rxr> room even ... 19:54 < martin_> Don't know. 19:55 < rxr> you don't know about the ini room ? 19:55 < martin_> we are in the old room. Don't know, what's about the new room is 19:55 < martin_> I must lern vor exam, too. PDV on mo 19:55 < rxr> was someone in it already ? 19:55 < rxr> yeah - damn early this time - bad for those mates just arrived back from a cebit booth ... 19:56 < rxr> I think there should be special regulations for those ... *g* 19:56 < martin_> We don't have buy 'holz' know ... 19:56 < martin_> But I think, we must go into the new room next week 19:56 < rxr> holz == wood 19:57 < rxr> is there anything done - or plain empty so far ? 19:57 < martin_> tell me the root password for gsmp and I can shutdown gsmp in this time :-P 19:57 < rxr> I could help a bit after dienstag ... 19:57 < rxr> *lol* 19:57 < rxr> I'm here - I can shut down it any time ... 19:58 < rxr> martin_: was there anything done int he 19:59 < rxr> new room - or is it a untouched virgin so far ? 20:07 < valentin> rxr: we'll be eating now and afterwards i have to clean up the kitchen and so 20:08 < CIA-9> valentin * r7594 /trunk/architecture/sh/ (config.hlp config.in pkg_linux26_post.conf): 20:08 < CIA-9> * added an option to choose the board 20:08 < CIA-9> * created a post.conf for linux that takes the predefine config for the choosen board 20:09 < rxr> always add a reference to the working area to the log - like added an board selection option to sh 20:09 < rxr> svn cp pkg_linux26_post.conf pkg_linux24_post.conf ? 20:13 < valentin> rxr: no 20:13 < valentin> better a . pkg_linux....conf ? 20:13 < rxr> well - that does not fit in the per pkg pre conf file automation 20:14 < valentin> ok 20:14 < valentin> then svn cpy ,) 20:14 < valentin> but i have no idea if the all the configs are in the same location for the 2.4 kernel 20:14 < valentin> bbl 20:15 < rxr> just rely on that for now ... 21:11 < mnemoc> re 21:12 < mnemoc> no, svn cp is a maintainance problem 21:12 < mnemoc> do a common pkg_linux_post.conf and source is from pkg_linux{24,26,-header}_post.con 21:12 < rxr> well - source bla blugs ... ,-) 21:12 < mnemoc> yes 21:18 < valentin> short re 21:18 < mnemoc> valentin: svn mv pkg_linux{26,}_post.conf 21:19 < mnemoc> valentin: svn commit 21:19 < valentin> mnemoc: ? 21:19 < valentin> i can do that, but ... why will that work ? 21:19 < mnemoc> valentin: echo '. $base/architecture/blah/pkg_linux_post.conf' > pkg_linux26_post.conf 21:19 < mnemoc> valentin: echo '. $base/architecture/blah/pkg_linux_post.conf' > pkg_linux24_post.conf 21:20 < mnemoc> valentin: echo '. $base/architecture/blah/pkg_linux_post.conf' > pkg_linux-header_post.conf 21:20 < valentin> special cases ,) 21:20 < mnemoc> with only one source of code to maintain 21:20 < valentin> :) 21:22 < mnemoc> :) 21:23 < CIA-9> valentin * r7595 /trunk/architecture/sh/ (pkg_linux26_post.conf pkg_linux_post.conf): * moved pkg_linux26_post.conf to pkg_linux_post.conf 21:23 < mnemoc> *g* 21:23 < mnemoc> have you seen jeru? 21:23 < CIA-9> amery * r7596 /trunk/package/gnome2/bluefish/bluefish.conf: * fixed bluefish by not updating desktop database 21:24 < valentin> nope 21:25 < mnemoc> jsaw: clisp is frozen again :\ 21:28 < mnemoc> jsaw: uhm... iirc when i tested it i had a local NOPARALLEL.... i'll try again with that 21:28 -!- mipe [~mipe@dsl10040.japo.fi] has quit ["leaving"] 21:41 < valentin> mnemoc: did you commit your special cases ? 21:43 < mnemoc> uhm? 21:43 < valentin> the linux conf thing 21:44 < mnemoc> you need to do it, now you are not extending linux26 21:44 < valentin> ?? 21:44 < mnemoc> 7595 21:45 < mnemoc> pkg_linux_post.conf extends nothing 21:45 < mnemoc> well, an unexistant package called 'linux' 21:45 < valentin> 21:20 < mnemoc> valentin: echo '. $base/architecture/blah/pkg_linux_post.conf' > pkg_linux26_post.conf 21:45 < valentin> 21:20 < mnemoc> valentin: echo '. $base/architecture/blah/pkg_linux_post.conf' > pkg_linux24_post.conf 21:45 < valentin> 21:20 < mnemoc> valentin: echo '. $base/architecture/blah/pkg_linux_post.conf' > pkg_linux-header_post.conf 21:45 < valentin> i thought you implemented that somewhere 21:46 < mnemoc> nope, those are the three trailing lines of my suggestion 21:46 < valentin> *g* 21:47 -!- martin_ [~martin@pD9E73B14.dip.t-dialin.net] has quit ["Leaving"] 21:47 < valentin> mnemoc: rxr told me that will not work 21:48 < valentin> but i can try 21:48 < mnemoc> don't forget to change the 'blah' with the right arch 21:49 < valentin> :) 21:49 < mnemoc> :) 21:51 < CIA-9> valentin * r7597 /trunk/architecture/sh/ (3 files): * created pseudo pkg_linux*_post.conf files including the generic file pkg_linux_post.conf 21:57 < mnemoc> valentin: does it work? 22:02 < valentin> i am not sure 22:03 < mnemoc> you may use if [[ $ver = 2.4* ]]; then to handle exceptions on your common _post.conf 22:04 < valentin> yep 22:04 < valentin> i wonder if i ever try to compile a linnux24 for sh .... 22:05 < mnemoc> [[ $ver = 2.4* ]] && abort "don't try to compile linux24 for sh dammit!" 22:05 < valentin> works.... 22:05 < valentin> at least for the kernel 22:07 < rxr> maybe 2.4 builds for sh ... 22:08 < rxr> it includes support and maybe due to less intrusive changes sh4 builds there ... 22:08 < mnemoc> oh 22:08 < valentin> hi rxr 22:08 < valentin> i googled a bit but i found no information about any regressions 22:10 < rxr> 2.6 obviously has regressions 22:10 < rxr> maybe all the others use the drop-in tree ? 22:10 < rxr> (aside that there do not seem to be that many sh developers ...) 22:10 < rxr> as I wrote for sh3 the kernel builds ... 22:12 < mnemoc> 2.6 has regressions?? where? 22:12 < mnemoc> it's the most stable kernel ever 22:12 < valentin> mnemoc: but not for weakly maintained archs ,) 22:13 < valentin> due to its abi stableness 22:13 < mnemoc> :D 22:13 < valentin> api even 22:14 < valentin> rxr: what does .identifier mean in C ? 22:19 < valentin> rxr: how far did your sh3 build get ? 22:19 < valentin> do we have gnat sth ? 22:21 < valentin> ada compiler 22:21 < mnemoc> ada?? 22:23 < valentin> someone on linux-sh asked for a static version 22:32 < valentin> i do not understand this struct syntax at all 22:32 < valentin> what about the last comma - isn't that too much ? 22:33 < mnemoc> a comma? 22:33 < mnemoc> o c? 22:33 < mnemoc> on C? 22:33 < valentin> static struct file_operations sq_fops = { 22:33 < valentin> .owner = THIS_MODULE, 22:33 < valentin> .mmap = sq_mmap, 22:33 < valentin> }; 22:33 < valentin> what is the '.' and what about the last comma ? 22:33 < valentin> the code is full of such structs 22:34 < mnemoc> that last comma is a typo 22:34 < rxr> valentin: field initializer ... ? 22:34 < valentin> aha 22:34 < valentin> and why a comma ? 22:34 < rxr> never used those ... ? 22:34 < rxr> the comma is a defect - is it the only thing the compiler complains about ? 22:35 < valentin> no - i just used struct x = { ..... } 22:35 < valentin> rxr: the compiler does not complain about the commas 22:35 * rxr crawling with electronics and math into the bed - cu 22:35 < valentin> cu 22:35 < valentin> arch/sh/kernel/cpu/sh4/sq.c:348: error: unknown field `unmap' specified in initializer 22:35 < valentin> arch/sh/kernel/cpu/sh4/sq.c:348: warning: initialization from incompatible pointer type 22:35 < valentin> arch/sh/kernel/cpu/sh4/sq.c:349: error: unknown field `sync' specified in initializer 22:36 < valentin> arch/sh/kernel/cpu/sh4/sq.c:349: warning: initialization from incompatible pointer type 22:36 < rxr> btw - we have no gnats 22:36 < rxr> it is on my TODO with low priority ... 22:36 < rxr> gnats is ugly to bootsrap as hell 22:37 < rxr> since it was rewritten from C to ADA it needs ADA for the initial bootsrap - quite dump ... 22:37 < valentin> ?? 22:37 < valentin> that truly sucks 22:38 < rxr> guess why we have no ada up to now ... 22:38 < valentin> rxr: lars has an ada 22:38 < valentin> ,) 22:39 < rxr> lars != we - and ada != gnats ... 22:39 < mnemoc> http://www.geeks.cl/~amery/T2/regressions.html <--- doesn't look good :/ 22:39 < valentin> rxr: found the problem 22:39 < valentin> but no idea how to fix 22:40 < valentin> once again a structure changed between kernel versions 22:40 < mnemoc> i love 2.6 stability 22:42 < rxr> stability wher ? 22:42 < rxr> +e 22:42 < rxr> ah - ic - you mean between the new subversions ... 22:42 < rxr> 2.6.11.{1,2,3,4,5} 22:42 < rxr> n8 all 22:43 < valentin> that regression must be older 22:43 < valentin> it is due to the memory management api if i understand that code 22:44 < rxr> mm changed between .10 and .11 22:44 < rxr> it is 4-level instead of 3-level IIRC ... 22:44 < rxr> to support the full address range on 64bit arches - and some other features in general IIRC ... 22:44 < rxr> so - n8 - finally - off - cu 22:45 < valentin> cu rxr 22:45 < valentin> damn, that is not obviosly to fix 22:45 < valentin> the vmm structures completly changed 22:49 < rxr> you are in working in between major linux release series - that is _normal_ between 2.6 and 2.7 or 2.8 - eerrr oh - they are named 2.6 - our bad ... 22:49 < rxr> *g* 22:49 < rxr> cu 22:49 < mnemoc> cu rene :) 23:05 < valentin> rxr: the linux-sh ppl told me i need their dropin tree - they claim that one is adapted to the new mm 23:06 < valentin> hi highly recommended to use the drop in in general - whatever that means 23:06 < valentin> s/hi/he 23:06 < valentin> i'll take a look when i return from my holidays 23:06 < valentin> in 5 days or so 23:07 < valentin> 22:55 < Lethal> I wouldnt use stock 2.6.11 anyways. 2.6.12-rc1 has a number of fixes for sh. but you are better off with 23:07 < valentin> the linuxsh tree regardless. 23:09 < mnemoc> on 1.7 times i suggested arch dependent downloads 23:09 < mnemoc> on a similar way that target dependant's 23:09 < mnemoc> but master said that wasn't necesary 23:09 < valentin> mnemoc: there is no such thing like a separate sh kernel - tsh folks only provide drop-ins 23:10 < mnemoc> drop-in=patch? 23:10 -!- jeru [~jeru@p54BFDB6F.dip.t-dialin.net] has joined #t2 23:10 < valentin> we can use that pkg-post stuff or so to download the drop-in 23:10 < valentin> hi jeru 23:10 < jeru> hi valentin 23:10 < jeru> hi all :) 23:10 < mnemoc> jeru! 23:10 < jeru> mnemoc! 23:10 < jeru> ;) 23:11 < mnemoc> valentin: you can't use pkg-post to download 23:11 < mnemoc> valentin: you can use it to append an already downloaded patch 23:11 < valentin> mnemoc: that is true - but there are not even patches 23:11 < mnemoc> what are the drop-ins? 23:11 < valentin> you need a full kernel tree and then use gnu-arch to "fill it up" 23:12 < mnemoc> i guess exactcode.de/valentin will provide the patches 23:12 < valentin> http://linuxsh.sourceforge.net/cgi-bin/moin.cgi/Getting_20source 23:12 < mnemoc> we have tla on T2 23:12 < valentin> mnemoc: hehe 23:12 < valentin> nice 23:12 < valentin> i am in dresden for the next days 23:13 < valentin> our hotel claims to have a good wlan, but i am not sure if i'll have some time to hack that 23:13 < mnemoc> if you phone rxr ask him about architecture/$arch/download.txt 23:14 < mnemoc> i can do that part for you 23:14 < valentin> nice 23:14 < valentin> i guess a dreamcast target will draw some more attention to T2 23:14 < mnemoc> i hate the idea on linux26-benh, blah blah blah 23:15 < valentin> maybe you can convince our new master in chief ,) 23:16 < mnemoc> new master in chief owns linux26-benh 23:17 < _Ragnar_> ^_^ 23:17 < valentin> but how would you realize arch dependent downloads ? 23:17 < valentin> esp how will you resolve conflicts when doing a ./scripts/download linux for different configs ? 23:17 < valentin> hi _Ragnar_ 23:18 < mnemoc> download -cfg foo -required or -all 23:18 < mnemoc> includes target specific 23:18 < mnemoc> it may include arch specific too 23:19 < valentin> ok 23:20 < valentin> mnemoc: write a detailed proposal to the ml - the core architect will answer eventually 23:20 < jsaw> .oO(linux-%) 23:20 < mnemoc> jsaw: yes 23:20 < valentin> hi jsaw 23:20 < mnemoc> jsaw: btw, clisp is broken again :| 23:20 < jsaw> hi all 23:20 < jeru> mnemoc: btw ... I was looking into the udhcpd code ... it should be quite easy to integrate sth like a LDAP plugin ... so I decided to start writing one ;) 23:21 < jsaw> mnemoc: :(((( 23:21 < jeru> hi jsaw 23:21 < jsaw> hi there! 23:21 < mnemoc> jeru: don't forget to commit dnsldap :) 23:22 < jeru> ah ... of course I will not forget .. ;) 23:27 < valentin> mnemoc, rxr: maybe the vanilla kernel compiles for sh4 when we turn off store queues 23:27 < valentin> but i will not try today - need to get up early in the morning 23:27 < mnemoc> :) 23:34 < valentin> so cu 23:34 < valentin> gn8 23:34 < jeru> good night valentin 23:34 < jsaw> gn8 valentin 23:41 -!- mnemoc [~amery@200.75.27.7] has quit [Remote closed the connection] 23:42 -!- mnemoc [~amery@200.75.27.7] has joined #t2 23:44 < mnemoc> :| 23:44 < jeru> mnemoc: I'm struggeling with my DSL line too :| 23:45 < mnemoc> i must switch connection provider 23:45 < jeru> do you use a wireless link ? 23:46 < mnemoc> not here 23:46 < mnemoc> there are two companies that provide the 'DSL' thing 23:46 < mnemoc> and multiple ISPs 23:46 < jsaw> .oO(should dig tunnel to next de back-bone cable) 23:46 < jsaw> .oO(guess 200m) 23:47 < mnemoc> i can pay for changing DSL from telefonica to IFX keeping GTD as ISP 23:47 < mnemoc> jsaw: :) 23:47 < jsaw> hehe --- Log closed Sun Mar 20 00:00:27 2005