**** BEGIN LOGGING AT Thu Sep 07 02:59:56 2006 Sep 07 07:37:10 RP: well if JustinPs problem shows on corgi Ill know as I have / on SD currentlu Sep 07 07:44:09 good morning all Sep 07 07:44:18 hi koen Sep 07 07:45:02 hey koen Sep 07 07:46:09 hi koen and all Sep 07 07:51:48 http://article.gmane.org/gmane.comp.version-control.monotone.devel/8169 Sep 07 07:58:47 Btw, i did not understand yet - how set right encoding for html pages in konq-e in Familiar-0.8.4 . I have problems with pages, that in any coding instead UTF-8. With UTF-8 all right. Maybe it is not encoding problem, but fonts problem? That problem according to mail too. Can anyone direct me to documentation ralated to this? I did not look in src yet. Sep 07 08:04:25 den-ros: ask #opie Sep 07 08:05:06 koen: thanks koen Sep 07 08:06:41 morning morning Sep 07 08:06:49 koen: I only test Familiar. I plan to use Angstrom. And does bug with no touchscreen in last Angstrom img was corrected already? Sep 07 08:07:00 yes Sep 07 08:07:52 good. does new imgs avalible? or it must be corrected under ssh? Sep 07 08:08:09 you can do both Sep 07 08:08:25 i prefer new imgs :) Sep 07 08:08:27 ipkg update ; ipkg upgrade will pull in the fixes Sep 07 08:08:39 hmm... Sep 07 08:09:00 i reflash Angstrom with Familiar for testing Sep 07 08:10:03 koen: and if i reflash Angstrom with that bug again - i can not ssh login here again to make upgrade Sep 07 08:10:37 the bug is gone in the newer images Sep 07 08:11:18 ok now i will search new images Sep 07 08:13:48 koen, with gcc4.1, are you using glibc2.4 ? Sep 07 08:13:59 i reflash Angstrom with Familiar for testing :) - i mean: reflash Familiar instead Angstrom :) Sep 07 08:14:03 tomas_f: yes Sep 07 08:14:19 hm Sep 07 08:14:34 tomas_f: OE doesn't have the patched CSL 2.3.x to make EABI work Sep 07 08:15:13 I have built an image over night without the iwmmxt optimatisations and it shows the same sigsegv issues as with Sep 07 08:15:15 tomas_f: doesn't POKYMODE=eabi2 pull that in for you? Sep 07 08:15:23 yes Sep 07 08:15:35 morning Sep 07 08:15:36 I have also tried the current csl release, same problems Sep 07 08:15:46 busybox 1.0.x will segv regardless of optimizations Sep 07 08:15:50 morning hrw :) Sep 07 08:15:52 hey hrw|work Sep 07 08:16:01 I have upgraded to 1.2.1, no difference Sep 07 08:16:09 hmmm Sep 07 08:16:11 hey hrw|work Sep 07 08:16:16 that solved most segv's for me Sep 07 08:16:30 except loadkeys, but that's useless on my ipaq Sep 07 08:16:44 loadkeys is one of the programs Sep 07 08:17:02 there are some gdk post-install scripts that sigsegv Sep 07 08:17:08 loadkeys works if you start it the second time Sep 07 08:17:09 and then the Xserver Sep 07 08:17:22 koen, yes, I found that out as well Sep 07 08:17:27 most of the time Sep 07 08:17:37 which puzzles me Sep 07 08:18:06 btw, I found a iwmmxt bug in gas Sep 07 08:18:39 but not a crasher, just would not build some stuff Sep 07 08:18:51 koen: if we could just find that bug, ANgstrom would be ready to roll Sep 07 08:19:19 tomas_f: I wonder if http://article.gmane.org/gmane.linux.ports.arm.kernel/27307 is responsible for that Sep 07 08:19:30 XorA: ready for "beta" :) Sep 07 08:19:43 koen: RC1 by microsoft standards Sep 07 08:19:45 * kergoth yawns Sep 07 08:19:58 hey kergoth Sep 07 08:20:52 koen, I will get RP to look at that ;) Sep 07 08:21:24 tomas_f: I'm not fond of CSL toolchains since the last one we had in OE was buggy and caused weird crashes Sep 07 08:22:40 * koen wasted to much time on that Sep 07 08:23:38 koen: might be worth dragging in 2006q1 as comparison though Sep 07 08:23:48 http://bugs.openembedded.org/show_bug.cgi?id=204 Sep 07 08:23:48 koen: if I get really bored at work :-) Sep 07 08:23:59 XorA: check o-hand svn Sep 07 08:24:13 XorA: http://svn.o-hand.com/view/poky?rev=697&view=rev :) Sep 07 08:24:27 good morning Sep 07 08:24:38 koen: cool, they are ahead Sep 07 08:26:27 XorA: it shows exactly the same problems for us as 4.1.1 does Sep 07 08:27:39 XorA: and I also want to try angstrom on PowerPC and mips, so we'll have to rethink toolchain selection anyway Sep 07 08:29:51 XorA: I suspect the CSL compiler might work on armv4t, since upstream gcc is broken for that Sep 07 08:30:13 where 'broken' read 'generates armv5 isns' Sep 07 08:41:14 Laibsch: did ztenv compile without the -DZAURUS flag? Sep 07 08:41:35 koen: no, unfortunately not. Sep 07 08:41:57 weird Sep 07 08:42:53 did you do -c c clean first? Sep 07 08:48:21 I updated the valid_sections list in the bittest content_checker according to http://lists.linuxtogo.org/pipermail/openembedded-devel/2006-August/000073.html and the answer of hrw|work Sep 07 08:50:07 we still need to know what to do with "x11/apps" -> "x11/applications" ? "x11/gnome" -> "x11" ? "x11/gnome/libs" -> " x11/libs" ? "x11-misc" -> "x11/apps" or "x11/misc" ? Sep 07 08:53:58 good morning Sep 07 09:12:48 morning all Sep 07 09:13:01 hey lrg Sep 07 09:13:06 hey mikearthur Sep 07 09:13:12 hey mickey|working Sep 07 09:13:24 good day koen Sep 07 09:14:05 koen: has OO finished building on nbp yet ? Sep 07 09:14:27 lrg: OE's toolchain seems to be missing c++ headers :( Sep 07 09:14:35 so it stopped at building stlport Sep 07 09:14:52 koen: any luck with q2 ? Sep 07 09:15:01 haven't tried that yet Sep 07 09:17:54 hey lrg Sep 07 09:18:05 lrg: static has mysteriously dissapeared on corgi Sep 07 09:18:25 morning folks Sep 07 09:18:29 XorA: cool, although we have a problem with spitz Sep 07 09:18:33 hey mickey|working Sep 07 09:18:40 lrg: yeah, that sounds nasty Sep 07 09:19:22 XorA: btw, what are you doing the week before you start ? Sep 07 09:19:31 lrg: working here Sep 07 09:19:55 hi, folks. I can't build bootstrap-image. :( Sep 07 09:20:33 but, I do bitbake nano . Sep 07 09:20:44 it report this error => ERROR: Cannot satisfy the following dependencies for task-bootstrap-unionroot: Sep 07 09:20:52 hmm, where is wrong ? Sep 07 09:21:04 XorA: If your available one night that week, I'd like to take you along for a meal with our US, Asia, Japan people when they visit Edinburgh Sep 07 09:21:24 lrg: I can make myself available Sep 07 09:21:29 morning all Sep 07 09:21:31 XorA: cool Sep 07 09:21:35 hey RP Sep 07 09:21:38 lrg: thursday I am nominally playing ADND, but I can change that Sep 07 09:22:13 RP: what gcc are you using on spitz ? Sep 07 09:22:26 folks. I am struggling with doing an up to date build for x86 on x86_64. build bootstrap-image when building glibc-initial fails with error: #error "TLS support is required.". anybody seen that? Sep 07 09:23:23 chris144: yes, and I never solved it Sep 07 09:24:06 XorA: did you use a workaround? Sep 07 09:24:20 chris144: no, I gave up Sep 07 09:24:35 XorA: so how are you building now? Sep 07 09:24:46 lrg: 3.4.4 Sep 07 09:24:56 chris144: I target arm, I had dreams of targeting my i586 server/router Sep 07 09:25:20 RP: got mail. Sep 07 09:25:28 lrg: any ideas? Sep 07 09:25:34 XorA: so this TLS thing is not an issue for arm? Sep 07 09:25:34 leoncamel, what's next line after the ERROR? Sep 07 09:25:54 chris144: did you start from epia.conf? Sep 07 09:26:06 chris144: there is some NPTL stuff you need to set for glibc Sep 07 09:27:05 XorA: yes, I know, I did that for glibc (--with-tls etc.). there was never an issue with glibc-initial until now. Sep 07 09:27:48 chris144: cool, as I said I never managed to get x86_64 to make x86 stuff Sep 07 09:28:24 RP: not yet, I'm going to turn on kernel debug and see if something stupid is happening. Iirc only the ac97 controller on pxa27x caused problems for mmc/sd. Sep 07 09:28:26 XorA: well, I was using gcc 3.4.4 before. Now i am doing a new build with all gcc and glibc preferences removed and up to date oe and bitbake 1.6. Sep 07 09:28:53 chris144: sounds like you have had more success than me n the past :-) Sep 07 09:29:11 O_Neil, base-files base-passwd busybox initscripts netbase sysvinit sysvinit-pidof tinylogin modutils-initscripts linux-hotplug ipkg-collateral ipkg ipkg-link modutils-depmod module-init-tools-depmod unionfs-modules unionfs-utils unionroot Sep 07 09:29:23 O_Neil, the next msg Sep 07 09:29:33 lrg: ok, I'm a bit lost with that one :-/ Sep 07 09:30:23 XorA: yes, I used to use an old oe version : OE-2006-0307.tar.bz2 Sep 07 09:31:07 leoncamel, you need define the RDEPENDS contains these packages Sep 07 09:31:30 XorA: i post my working (for x86 on x86_64) distro and machine files to pastbin, hold on Sep 07 09:34:54 XorA: here we go, maybe it helps you with your x86 target: http://pastebin.ca/163235 Sep 07 09:35:33 chris144: cheers Sep 07 09:39:20 O_Neil, ? what ? I am using standard bootstrap-image, and I change nonthing about this package. Sep 07 09:41:35 leoncamel, RDEPENDS defines packages that satisfy bootstrap-image. Sep 07 09:44:17 O_Neil, hmm, which file should I modify ? Sep 07 09:44:18 RP: worryingly I get the correct value for params_format() on mainstone with fake WM8750. Sep 07 09:44:55 RP: fwiw, the params_* code is very similar to the code the cse patch fixed Sep 07 09:45:49 leoncamel, "packages/images/bootstrap-image.bb" Sep 07 09:47:48 RP: I'll dig out the old memory barrier patch and send it to try on spitz Sep 07 09:50:32 XorA: ok, I found a workaround. when the build fails got to work/glibc-initial-2.4-r10/glibc-2.4/sysdeps/i386/elf directory. Sep 07 09:51:47 XorA: then patch both lines libc_cv_386_tls= in configure.in to "yes". then restart your build. It just worked for me. Sep 07 09:52:31 XorA: NOTE: package gcc-cross-initial-4.1.1-r6: task do_compile: started for x86 on x86_64 :-) Sep 07 09:52:54 chris144: maybe they are suitable for adding to site files? Sep 07 09:53:55 XorA: yes, possibly, but I am not very familar with site files. I think I put the problem and the workaround into bugzilla. somebody familar with site files may decide... Sep 07 09:58:29 all: I started my own business finally. Sep 07 09:58:41 ;) Sep 07 09:58:55 hrw|work: :) Sep 07 10:00:07 some paper work to do but it is started Sep 07 10:01:52 hrw|work: congrats. I did the same thing in february this year. A lot of work but freedom at last ! Sep 07 10:04:08 chris144: I still plan to work where I work as it takes most of payments from me Sep 07 10:04:28 hrw|work: did you found people who accept to pay you to work on their devices ? or ..? Sep 07 10:04:35 chris144: no need to pay ~200eur to goverment each month - current employer pay Sep 07 10:04:47 cyrilRomain: kind of Sep 07 10:08:04 hrw|work: I did the full jump into cold water, ok I have to admit that my ex employer gives me one contract after the other but that will end in january. There is a great lack of engineers in germany, so there should be contracts to be outsourced... Sep 07 10:11:11 chris144: I will look how will own work goes Sep 07 10:11:53 hi obergix[work] Sep 07 10:12:12 hi Sep 07 10:20:23 anyone know if MMCs should just work on tosa? Sep 07 10:21:05 mikearthur: should Sep 07 10:21:31 mikearthur: mmc is base of all mmc/sd/microsd/minisd/mmc4/mmc+ formats Sep 07 10:22:46 that rocks: NOTE: package glibc-2.3.5+cvs20050627-r12: task do_compile: started Sep 07 10:27:06 chris144: huh? Sep 07 10:33:00 mikearthur: I am just trying to build for x86 on x86_64 which does not seem to be so easy...but making progess Sep 07 10:34:25 brb Sep 07 10:47:25 Crofton: do you have an LCD or anything hooked to your OSK board ? Sep 07 10:47:47 Crofton: (via Expansion Module C, for example) Sep 07 10:50:27 I have an lcd Sep 07 10:50:38 but I am not fooling with it right now Sep 07 10:50:56 once upon a time I had opie running on it Sep 07 10:51:09 is it the NEC QVGA one, or a different one? Sep 07 10:52:24 nick chris144|away Sep 07 10:53:19 the reason i ask is because we're trying to find some sort of a reference that details what each of the 18 LCD-specific pins actually does. the pin out guide is of little help since things like LCD P0 is of little help to me. Sep 07 10:54:29 specifically LCD P0-P15, i guess. Sep 07 10:54:45 We have a mistral QVGA LCD panel Sep 07 10:54:58 we just plugged it in, built drivers etc and it worked Sep 07 10:54:58 damn. that's what i was afraid of Sep 07 10:55:10 sorry :( Sep 07 10:55:20 no worries. the search goes on :rawk: Sep 07 10:55:35 at least you are asking forward moving questions again :) Sep 07 10:56:03 true Sep 07 10:56:03 anyone here know much about getting video of your tivo, using Linux of course? Sep 07 10:56:35 if your TiVo has ethernet, it should show up as a network drive on your network Sep 07 10:56:43 probably accessible over samba Sep 07 10:56:47 hmm Sep 07 10:57:04 It should appear on the network as soon as my wireless brdige arrives :) Sep 07 10:57:36 bother, the bridge only just made it to Dallas Sep 07 10:58:45 ouch Sep 07 10:58:59 It is supposed to be here tomorrow (VA) Sep 07 10:58:59 have you looked up any of the TiVo hack guides? they probably have more detailed info Sep 07 10:59:07 just starting to Sep 07 10:59:29 do you know the series of your TiVo ? Sep 07 10:59:29 I wonder if anyone has tried using OE to build stuff for a tivo .... Sep 07 10:59:33 series 1 Sep 07 10:59:39 oh, then you're golden Sep 07 10:59:42 old sony Sep 07 10:59:48 directivo Sep 07 11:00:08 I just put a bigger drive in and one of these cache cards Sep 07 11:00:46 did you see Steve Jenkins's hack guide for series 1 TiVos ? Sep 07 11:01:07 not until now Sep 07 11:01:13 http://tivo.stevejenkins.com/network_cd.html Sep 07 11:01:21 it's probably what you're looking for... i think Sep 07 11:01:30 i could be wrong, but it's a start Sep 07 11:06:24 anyone seen this? It appears in my bootstrap-image step: ipkg: Cannot create directory `/usr/lib/ipkg': Permission denied Sep 07 11:06:32 ignore Sep 07 11:08:46 hrw|work: ok Sep 07 11:15:40 how should I satisfy recommendations? http://rafb.net/paste/results/ED83j953.html Sep 07 11:16:39 likewise: dropbear/portmap ones? Sep 07 11:16:50 equals bug #1172, yes dropbear portmap Sep 07 11:17:57 hrw|work: yes Sep 07 11:18:44 likewise: drop task-bootstrap-unionboot from tmp/deploy/ and build task-bootstrap Sep 07 11:19:18 and set a PREFERRED_PROVIDER for task-bootstrap Sep 07 11:20:01 someone should FIX this sick situation Sep 07 11:24:14 hrw|work: what is the fix? A default provider? or a working task-bootstrap-unionboot? Sep 07 11:24:52 likewise: provider Sep 07 11:25:08 likewise: I do not know why you got unionboot one Sep 07 11:26:22 hrw|work: ok. Dunno neither, I am doing a plain task-bootstrap for MACHINE = "ixp4xx" DISTRO = "openslug" Sep 07 11:28:32 koen: tnx Sep 07 11:28:46 someone made task-bootstrap-unionroot provide task-bootstrap Sep 07 11:29:01 and bitbake will pick one semi-randomly Sep 07 11:33:46 NOTE: package gcc-cross-4.1.1-r6: task do_compile: started on x86_64 :-) Sep 07 11:51:32 koen, are people actually working on doing chroot builds? Sep 07 11:56:03 Crofton: not yet Sep 07 11:56:06 ping hrw|work Sep 07 11:56:48 bummer Sep 07 11:57:04 omniorb and xerces bb files have a bad habit of finding stuff on the host Sep 07 11:58:25 I am hoping the problem "goes away" befiore I really need to fix it :) Sep 07 11:59:57 chris144: pông Sep 07 12:02:32 hrw|work: I got as far as glibc now building for x86 on x86_64 ! If I succed I will send my distro and machine files to the mailinglist, maybe you can get your x86 stuff done... Sep 07 12:04:44 later... Sep 07 12:06:21 chris144|away: nice Sep 07 12:11:38 More trouble, but I cannot find a matching bug reported for this one yet: http://rafb.net/paste/results/rl8fcn64.html Sep 07 12:12:49 likewise: how old metadata you are using? Sep 07 12:16:40 03nail 07org.oe.dev * r44322ee7... 10/ (3 files in 3 dirs): busybox: Update slugos-specific defconfig Sep 07 12:20:30 hum.. using 'date=' in SRC_URI with 'svn://' is wrong, isn't it ? Sep 07 12:20:55 hrw|work: yesterdays metadata I think Sep 07 12:24:28 03nail 07org.oe.dev * ra1795c18... 10/ (3 files in 3 dirs): openntpd: Fix initscript and reformat .bb to match OE style guidelines Sep 07 12:24:33 03nail 07org.oe.dev * r4a18ebe5... 10/ (1 packages/zd1211/zd1211_r83.bb): zd1211: Reformat .bb Sep 07 12:24:38 03nail 07org.oe.dev * rdc38d27c... 10/ (1 packages/litestream/litestream_1.3RC3.bb): litestream: Reformat .bb Sep 07 12:24:43 03nail 07org.oe.dev * r03a7d68c... 10/ (1 packages/ssmtp/ssmtp_2.61.bb): ssmtp: Reformat .bb Sep 07 12:25:02 hrw|work: when do you start a new biz? Sep 07 12:25:35 likewise: basically tomorrow Sep 07 12:25:51 hrw|work: let me know the details sometime. might be interested. Sep 07 12:26:42 likewise: ok Sep 07 12:26:42 how do I trick bitbake into remaking a cvs tarball? Sep 07 12:26:46 NAiL: openslug compiles for you? Sep 07 12:26:54 XorA: rm DL_DIR/cvstarball? Sep 07 12:27:07 hrw|work: doesnt seem to do it Sep 07 12:27:19 XorA: should.. Sep 07 12:27:21 bbl Sep 07 12:27:40 likewise: Not really. I'm trying to build on edgy eft x86_64, and it fails when it tries to do the rootfs. Sep 07 12:27:53 * XorA forget about the new fetch stamps Sep 07 12:28:20 NAiL: details? dapper x86_64 host here, fails during bootstrap task (libtool) Sep 07 12:28:54 likewise: log_check finds a non-existing match while grepping for "Cannot find package" in the logs. Sep 07 12:29:41 * NAiL restarts with a clean build Sep 07 12:29:57 NAiL: hmmmn. Did not see that. Restarted a clean build also a few minutes ago with up-to-date metadata... Sep 07 12:30:37 likewise: gimme a little while and I'll see how far I get. Sep 07 12:31:24 NAiL: yup, I will prob disappear for 30 minutes or so, soon. Sep 07 12:32:21 I'll be here a few more hours Sep 07 12:37:09 ok I checked into the bitbake code, 'date=' is ignored when used in SRC_URI using svn:// Sep 07 12:37:50 can someone please change in packages/konqueror/konqueror-embedded_20060404.bb: s/;date=${PV}// ? Sep 07 12:37:52 I think that should either use SRCDATE of 'rev' Sep 07 12:38:09 koen: SRCDATE is indeed used Sep 07 12:38:38 there is two occurence in this file Sep 07 12:39:33 * chouimat is back. Sep 07 12:40:01 03koen 07org.oe.dev * re0a90e17... 10/ (9 files in 2 dirs): gpe-calendar: reformat according to styleguide and drop obsolete versions Sep 07 12:40:10 morning Sep 07 12:40:10 cyrilRomain: 'GPE_TARBALL_SUFFIX' has to come before 'inherit' Sep 07 12:40:14 hey chouimat Sep 07 12:40:40 koen: ok, I update the wiki and script right now Sep 07 12:40:55 koen: hum at least the script Sep 07 12:41:06 cyrilRomain: I think it can be left oout of the wiki, since it is specific to gpe.bbclass Sep 07 12:41:19 koen: yep Sep 07 12:41:27 it's pretty obvious when you get it the wrong way around, since fetch will fail :) Sep 07 12:42:20 koen: btw I'm thinking that all unknown variables (including package specific ones) are put at the end of the file, i.e. _after_ the do_routines Sep 07 12:42:55 koen: which makes do_routines fail interpolating those variables, right ? Sep 07 12:43:13 I suspect so, but I'm not sure Sep 07 12:43:26 unless bitbake don't mind the order ... Sep 07 12:46:20 koen: I'm afraid that if variable order is important (for interpolation), it will be difficult to find the right order, especially when a variable use another one :-/ Sep 07 12:46:38 yes Sep 07 12:46:51 that's why sanitizing is requires thinking :) Sep 07 12:47:06 koen: :) Sep 07 12:47:28 s/is// Sep 07 12:50:01 * cyrilRomain think of the so great perl Template Toolkit interpolations ... Sep 07 12:50:03 RP: ping Sep 07 12:50:31 03koen 07org.oe.dev * r122cbbd5... 10/ (11 files in 3 dirs): gpe-conf: reformat according to styleguide and drop obsolete versions Sep 07 12:53:01 * cyrilRomain also wonders if the cheetah python template system could be used ... Sep 07 12:53:28 mickey|working: pöng Sep 07 12:53:39 hey hrw Sep 07 12:53:56 mickey|working: still want poodle? Sep 07 12:54:06 if we have one spare Sep 07 12:54:09 if not, no problem :D Sep 07 12:54:27 mickey|working: I have the project one running 3.5.4.2-rc2 now Sep 07 12:54:53 cool. bring that one to OEDEM Sep 07 12:55:35 mickey|working: want 110V power supply with it? Sep 07 12:55:41 I'll bring the netboook as well Sep 07 12:56:17 * koen adds some points to the OEDEM wiki Sep 07 12:56:32 I'll take poodle, tosa and probably c7x0 Sep 07 12:56:50 hrw|work: no, thanks. i have enough power supplies Sep 07 12:57:17 mickey|working: ok. then I will have another unused one Sep 07 12:57:29 "spares" Sep 07 12:58:19 4 Z, 1 Sharp charger, 1 noname universal one, 1 Delta 110V, 1 Delta 110-240V (both Delta with US connectors) Sep 07 13:02:01 lucky OEDEM isnt in the uk Sep 07 13:04:45 that's why I have a travel-adapter Sep 07 13:05:26 RP: looks like the MMC clock and the I2S sysclock is a shared GPIO on pxa27x. Will create patch Sep 07 13:06:39 heh Sep 07 13:06:55 "look we saved a gpio!" Sep 07 13:07:37 lrg: so SD breaks when you aren't playing music? Sep 07 13:08:28 koen: SD breaks when you do play music on pxa27x. Both drivers are using the same GPIO. Sep 07 13:08:35 koen: I only have us->europlug one Sep 07 13:08:42 koen: for sysclk Sep 07 13:24:08 any idea what i can do to make udev create /dev/ttyS0 ? Sep 07 13:24:30 yes, the serial drivers are in the kernel, i can perfectly log in via serial after manually creating ttyS0 Sep 07 13:26:16 mickey|working: write a udev rule? Sep 07 13:26:24 i was afraid of you saying that Sep 07 13:26:25 (assuming the serial stuff shows up in /sys) Sep 07 13:26:31 ah Sep 07 13:26:33 let me check /sys Sep 07 13:26:53 how would such a rule look like? Sep 07 13:26:57 if you're on ARM, you're probably have to use /dev/tt Sep 07 13:27:04 since RMK is still a slacker Sep 07 13:27:15 ttSAC + ttyS0 will do Sep 07 13:27:39 or.. on another note Sep 07 13:27:46 is the device_table still relevant when we are using udev? Sep 07 13:28:18 KERNEL=="uinput", NAME="input/%k" Sep 07 13:29:02 mickey|working: KERNEL=="ttyNAME" NAME="/dev/i-want/to/have/serials/in/weird/dir/%n" Sep 07 13:29:10 ah, cool, thanks Sep 07 13:29:18 when are the rules parsed? Sep 07 13:29:23 only on startup or regularly? Sep 07 13:29:26 mickey|working: s+"/dev/+" Sep 07 13:29:38 mickey|working: on startup Sep 07 13:29:48 mickey|working: you can force udev to reread Sep 07 13:29:53 and 'udevcontrol reload' iirc Sep 07 13:29:59 excellent. thanks Sep 07 13:30:16 /etc/init.d/udev reload Sep 07 13:30:19 udevcontrol reload_rules Sep 07 13:31:47 has anyone encountered bugs in udev 097? Sep 07 13:31:48 my speeded-up bittest source_checker starts to work :) Sep 07 13:34:11 does anyone know how can I check for a cvs repository avaiability without checkouting ? (for now I'm using 'cvs rlog' but I guess there is a better way to check that) Sep 07 13:34:19 cvs -c Sep 07 13:34:22 cvs co -c even Sep 07 13:34:30 mickey|working: thanks :) Sep 07 13:40:30 03koen 07org.oe.dev * r29567df6... 10/ (1 packages/openocd/openocd_svn.bb): openocd: unbreak build Sep 07 13:53:27 Can I disablee i18n support globally in a configuration? Sep 07 13:57:02 likewise: did your build break on libtool-native? Sep 07 13:57:32 NAiL: still running... Sep 07 13:57:55 mine too... Not gotten that far though Sep 07 13:58:02 NAiL: I am compiling kernels for some other target as well (testing SATA drive hotplug features)... Sep 07 13:58:06 binutils-cross right now Sep 07 13:58:27 NAiL: busybox here Sep 07 13:59:17 likewise: I think we have DISABLE_NLS or something Sep 07 14:00:51 likewise: 'USE_NLS' Sep 07 14:01:20 koen: ah yes, that rings a bell. Strange because my openslug-based build didn't compile locales earlier. I just took my 4 months old config and package overlay and bring it up-to-date with OE-current. Sep 07 14:07:10 mickey|working: do you know how I could 'cvs -c' for a module ? Sep 07 14:07:23 no, that's just too check whether a repository exists Sep 07 14:08:02 mickey|working: ok, so if I also want to check for the module, I'll have to use 'cvs rlog' or something else :-/ Sep 07 14:08:13 mickey|working: well, thanks anyway for your help Sep 07 14:13:10 03koen 07org.oe.dev * r2febf377... 10/ (1 packages/gimp/gimp_2.3.10.bb): gimp: add 2.3.10 Sep 07 14:14:35 NAiL: clean build worked ok... Sep 07 14:15:17 still on glibc here Sep 07 14:15:24 * NAiL goes out to play some pool Sep 07 14:18:36 hi all Sep 07 14:19:21 hi cedricb Sep 07 14:21:50 NAiL: had to set PREFERRED_PROVIDER_task-bootstrap = "task-bootstrap" though. Sep 07 14:23:22 I'm getting this in .dev -> http://pastebin.ca/163412 Sep 07 14:23:58 cedric: create that file (and the file it complains after that) and it will work Sep 07 14:24:52 koen: did you try that minimo recipe? Sep 07 14:28:13 XorA: I got distracted with flambe-ing pancakes yesterday Sep 07 14:30:09 koen: heh heh, I cant get it to work at all, it breaks in the first 5 lines of mozilla compile Sep 07 14:31:24 report it :) Sep 07 14:34:40 NOTE: Fetch cvs://anonymous@cvs-mirror.mozilla.org/cvsroot;module=mozilla;tag=MOZILLA_1_8_BRANCH;date=20060720 Sep 07 14:34:45 lets test.. Sep 07 14:35:50 thx koen Sep 07 14:42:20 koen: same thing with dbus-glib-bindings.h (missing) Sep 07 14:42:30 cedricb: yes Sep 07 14:42:38 a .xml and a .h file are missing Sep 07 14:47:17 NOTE: package minimo-20060720-r0: task do_configure: started Sep 07 14:47:26 still fetching... Sep 07 14:47:38 koen: ewi? get it from my dl_dir Sep 07 14:47:49 koen: /home/hrw/devel/sources/dl_dir/ Sep 07 14:48:27 | /usr/bin/ld: skipping incompatible /home/hrw/devel/build/3541/tmp/staging/arm-linux/lib/libIDL-2.a when searching for -lIDL-2 Sep 07 14:48:38 hmm. something I fscked Sep 07 14:49:01 restarted Sep 07 14:49:17 * koen suspects '/usr/bin/ld' Sep 07 14:52:02 I will have to create new builddir soon ;( Sep 07 14:52:26 I skip minimo building for now Sep 07 15:01:47 koen: thx it works by creating empty files with 'touch' Sep 07 15:08:05 XorA: it fails patching over here Sep 07 15:09:37 hey, koen. which distro is stable now ? so, i could test building for it. Sep 07 15:10:30 koen, it seem that my distro and conf can't work now. :( Sep 07 15:10:56 koen, even the bootstrap-image can't build. Sep 07 15:13:17 koen: weird, gets to do_compile here Sep 07 15:16:18 hi, folks . Sep 07 15:17:08 which distro is OK now ? i mean which distro I could build now ? Sep 07 15:17:42 leoncamel: try generic-unstable, or don't build bootstrap-image Sep 07 15:18:44 koen, OK. but why the bootstrap-image can't build recently ? Sep 07 15:19:55 oz-unstable is pretty stable :-) Sep 07 15:25:34 * CosmicPenguin shakes the rust off his oe tree Sep 07 15:25:50 Do we have any closet OLPC developers in the house? Sep 07 15:26:15 send me one ;) Sep 07 15:26:41 http://wiki.laptop.org/go/Developers_program Sep 07 15:26:56 | /usr/bin/ld: skipping incompatible /data/build/koen/OE/build/tmp/angstrom/staging/arm-angstrom-linux-gnueabi/lib/libIDL-2.so when searching for -lIDL-2 Sep 07 15:27:06 ~curse minimo for using host ld Sep 07 15:27:08 May the fleas of a thousand camels infest your most sensitive regions, minimo for using host ld ! Sep 07 15:38:42 hrw|work, XorA: stick a 'inherit autotools' in the .bb and remove the EXTRA_OECONF Sep 07 15:39:18 hi, koen. which machine is stable now ? Sep 07 15:39:26 ? Sep 07 15:40:34 cu - time to home Sep 07 15:40:39 koen, i mean , which machine is stable , so that I could build with generic-unstable. ? Sep 07 15:41:59 ehm Sep 07 15:42:11 you can't just pick a machine, you need one to match your hardware Sep 07 15:42:16 thats a really bad queston Sep 07 15:46:05 XorA, sorry for that bad question. I am remote control my laoptop. so , it is very unconformtable here. :(. Sep 07 15:47:03 koen, , i just want test a typpical environment. :( Sep 07 15:49:13 heading home as well. cya later all. Sep 07 15:52:03 leoncamel: try c7x0 Sep 07 16:25:36 CosmicPenguin: has sugar even made it into oe yet ? Sep 07 16:26:03 ade|desk: not yet - but thats on my todo list Sep 07 16:26:09 it shouldn't be a problem Sep 07 16:27:21 i used a qemu'd build 59 i think, not the most pleasant UI Sep 07 16:27:46 and lacks https support, which is a bit naff Sep 07 16:28:06 its evolving Sep 07 16:28:10 build81 is the latest Sep 07 16:28:11 who's idea was it to give kids sugar? Sep 07 16:28:27 I need to get the all the driver stuff into OE Sep 07 16:28:32 better than fags i guess Sep 07 16:29:05 build81 is device only :( not qemu'ible Sep 07 16:29:42 looking at the build log , bit confused by the random scsi drivers Sep 07 16:30:06 302 meg just for what you get seems OTT Sep 07 16:30:18 XorA|gone: http://www.angstrom-distribution.org/repo/?action=details&pnm=minimo Sep 07 16:31:07 koen: nice.. push it Sep 07 16:31:10 ade|desk: thats why I roll my own Sep 07 16:31:23 hrw: It doesn't build anymore after my cleanup :( Sep 07 16:31:28 root@home:~# apt-get install --reinstall `cat lista-pakietow.text` Sep 07 16:32:19 now an hour to fetch it.. Sep 07 16:33:55 * mrz80 is away: Metaphysically I'm still here; physically, I'm elsewhere Sep 07 16:34:16 koen: pb is the maintainer .. groovy Sep 07 16:38:18 CosmicPenguin: whats the power drain on the cpu compared with say a PXA270 ? Sep 07 16:38:54 Whats the power drain on the PXA270? Sep 07 16:39:17 the pxa270 isn't what you'd call frugal Sep 07 16:40:20 but its sleep state is quite good Sep 07 16:42:20 which geode is it ? Sep 07 16:44:56 GX Sep 07 16:45:12 The total GX package is somewhere around 3W, I think Sep 07 16:45:23 The whole OLPC board budget is 10W Sep 07 16:48:10 active power is just over 4W so the AMD site says for the 1.1W ver but no figures for the 1W version Sep 07 16:48:14 sans LCD i take it? Sep 07 16:48:44 Xscale hits max 3 Sep 07 16:49:15 hmm no word on standby power for the GX ... will have to look harder Sep 07 16:50:30 Whats up with the wireless MMX thing - is that a real FPU? Sep 07 16:51:19 CosmicPenguin: no, it's powered my marketing Sep 07 16:52:20 lol AMD docs: This document clearly describes how the Geode GX Sep 07 16:52:21 533@1.1W and Geode LX 800@0.9W processors’ typical Sep 07 16:52:21 power ratings were obtained Sep 07 16:52:29 hmm i dont think so Sep 07 16:53:09 1.957W * 57% = 1.115W Sep 07 16:53:10 CosmicPenguin: when will amd start making ARM cores? Sep 07 16:53:18 when hell freezes over? Sep 07 16:53:27 i love the random 57% Sep 07 16:53:52 CosmicPenguin: shame AMD would make good arm cores i would have thought Sep 07 16:53:58 hrw: is your home machine feeling better now? Sep 07 16:54:08 ade|desk: why would you think that? Sep 07 16:54:27 MIPS cores were good Sep 07 16:54:35 We didn't make those Sep 07 16:54:53 that maybe why they were good then :( Sep 07 16:54:57 heh Sep 07 16:55:08 AMD has little or no expertise outside of x86 Sep 07 16:55:38 i liked the look of the au1500 stuff, why did amd offload it ? Sep 07 16:55:49 It was outside of the core business Sep 07 16:55:58 idiots Sep 07 16:56:13 and really, it had no business being in AMD - the company just isn't set up to market those Sep 07 16:56:14 eggs and baskets etc Sep 07 16:56:40 ade|desk: Intel is doing the same thing Sep 07 16:57:29 intel also invented netburst Sep 07 16:57:36 For two companies making money arm over fist with the desktop and server chips, why would they keep around small businesses with tiny margins? Sep 07 16:57:39 are they offloading the embedded stuff too .. grr Sep 07 16:57:54 ade|desk: intel sold pxa to marvel Sep 07 16:58:03 ade|desk: but they seem to have kept ixp and iop Sep 07 16:58:13 Rumor tells us that the 10k people fired wern't in the core processor business Sep 07 16:58:45 !!! Sep 07 16:58:46 ade|desk: Don't know, don't care Sep 07 16:59:01 sums it up nicely Sep 07 16:59:52 so who makes good low power cpus and will continue? samsung ? Sep 07 17:00:03 Raza took on Alchemy Sep 07 17:00:12 nec's VR series Sep 07 17:00:17 Marvell is going to continue with PXA Sep 07 17:00:23 the chips aren't dying - they're just going somewhere else Sep 07 17:00:34 Except for Geode - Geode is dead Sep 07 17:01:00 and so olpc is using dying cpus .. shame Sep 07 17:01:14 there was no upgrade path anyway Sep 07 17:01:15 short term project then Sep 07 17:01:36 AMD still makes the GX Sep 07 17:01:50 olpc version2 will use ?? Sep 07 17:01:57 yeah - its up in the air right now Sep 07 17:02:13 In the interest of keeping my job, I would recommend the Geode LX, but thats selfish of me Sep 07 17:02:22 lol Sep 07 17:03:11 even the powerpc is dying too Sep 07 17:04:08 i think i'll go home before i get too depressed Sep 07 17:04:24 heh Sep 07 17:04:49 there is http://projects.ppczone.org/projects.php to revive powerpc Sep 07 17:05:11 but the genesis dudes have been delaying the efika for months now Sep 07 17:05:14 powerpc is alive and healthy ... in routers and switches Sep 07 17:06:02 cisco is probably the #1 buyer for ppc chips Sep 07 17:11:12 NOTE: package bootstrap-image-1.0: completed for x86 on x86_64. So it is possible despite a few problems. (using oe from yesterday and bitbake 1.6, all glibc 2.4, all gcc 4.1.1 Sep 07 17:13:38 ummm... did x86_64 builds get broken? Sep 07 17:24:07 re Sep 07 17:24:15 hrw: wb Sep 07 17:24:32 hvontres|poodle: this machine has to wait until weekend - I will check it then with other power supply Sep 07 17:28:57 CosmicPenguin: I had difficulties with glibc: error: #error "TLS support is required." XorA said he gave up on it some time ago. Sep 07 17:29:06 oh, heh Sep 07 17:29:11 yeah Sep 07 17:30:06 CosmicPenguin: although i didn't fix that problem, I worked around it manually... Sep 07 17:30:37 that problem has been haunting us for a while now Sep 07 17:31:13 CosmicPenguin: somebody suggested today that the workaround could be put in a site file... Sep 07 17:31:22 i am just not too familar with site files... Sep 07 17:32:38 workaround: when the glibc build fails with "TLS error" got to work/glibc-initial-2.4-r10/glibc-2.4/sysdeps/i386/elf directory. Sep 07 17:32:58 then patch both lines libc_cv_386_tls= in configure.in to "yes". then restart your build. It just worked for me. Sep 07 17:33:19 Does it run? Sep 07 17:33:54 hrw, XorA|gone: minimo segfaults Sep 07 17:34:14 CosmicPenguin: well, yes, at least build bootstrap-image just completed. I need to test the gcc though. Sep 07 17:34:21 add 'libc_cv_386_tls=yes' to the correct file in site/ Sep 07 17:34:28 well - it woudl be nice to know if the binaries ran Sep 07 17:34:36 just a minor detail Sep 07 17:35:08 CosmicPenguin: you mean the cross binaries gcc, as, etc? Sep 07 17:35:29 well - and since glibc is building with TLS support it thinks it has - does that negatively affect the library? Sep 07 17:35:41 * CosmicPenguin doesn't understand what the TLS thing is all about Sep 07 17:38:18 CosmicPenguin: well, TLS is needed by nptl somehow. I am not a TLS expert either. On x86_64 TLS support evaluates nagative in the config run. Sep 07 17:40:05 CosmicPenguin: so some define HAVE_TLS... gets set to no which then triggers the error. I think it is just a problem of the glibc configure script, which tries to compile i386 platform dependant code and fails on x86_64. Sep 07 17:40:16 * CosmicPenguin shrugs Sep 07 17:40:19 Ok - sounds good to me Sep 07 17:40:30 this won't be the first time we've hacked the site file for x86-64 Sep 07 17:41:29 koen: are you confident and will put the site file change in the repository or should I file a bug? Sep 07 18:10:57 I'm working on porting the latest version of sqlite3 (3.3.7) to OE. I have ported all of the patch files that were used for 3.2.7 but I'm getting a compile error that I'm not familiar with: http://pastebin.ca/163627 Sep 07 18:12:50 * mrz80 is back (gone 01:38:56) Sep 07 18:14:07 that one looks weird indeed Sep 07 18:20:09 koen, any suggestion were to start? I've been scowering the Makefile but all looks good from what I can see. Sep 07 18:20:23 * koen has no idea Sep 07 18:20:38 koen, there is only one Makefile so at least that makes it some what simple :-) Sep 07 18:29:03 lol, there were too many white spaces at the end of one of the Makefile lines :-) Sep 07 18:29:05 it now works.. Sep 07 18:33:14 ~lart wordpress plugins Sep 07 18:33:15 * ibot cuts off wordpress plugins's head with a halberd that could have been a little bit sharper Sep 07 19:00:04 good night Sep 07 19:14:13 * chouimat is away: playing wesnoth for a while Sep 07 19:16:23 #universal Sep 07 19:16:45 stefan_schmidt: what is qonsole, what makes it so special? Sep 07 19:17:21 zecke: http://gnumonks.org/~laforge/weblog/2006/08/29/#20060829-qonsole-author-sources Sep 07 19:17:42 koen: yes, I'm wondering what makes it special... Sep 07 19:18:02 koen: and I wonder if QtE+EZX Apps is defined as a system library for the EZX phones Sep 07 19:18:20 Im going to try installing linux on my htc universal, anyone here have any experience with this? Sep 07 19:18:21 cu Sep 07 19:18:22 koen: or if any GPL software running on the EZX phone is a GPL violation anyway Sep 07 19:19:02 zecke: using a OS from motorola violates my sensibility Sep 07 19:19:34 that is a different story Sep 07 19:40:18 so, the rumor is that the last bitbake is better then sex - true? Sep 07 19:42:36 CosmicPenguin: last meaning....? Sep 07 19:42:56 CosmicPenguin: 1.4 and 1.6 are very fast compared to the prior incarnations Sep 07 19:43:34 Well I kne wthat Sep 07 19:43:47 but I heard rumors of multi-threading and other such interesting concepts Sep 07 19:44:12 I don't know if that's done yet....there was some work to get it up to that, yes Sep 07 19:44:15 woul dbe in head, not 1.6 Sep 07 19:44:16 CosmicPenguin: no rumors Sep 07 19:44:28 CosmicPenguin: RP has implemented multithreading in BitBake trunk Sep 07 19:45:12 RP: I need to admit I do not get how a deterministic build system and multi-threading (non deterministic) can ever match :) Sep 07 19:47:06 JustinP: audio/mmc issue is solved. Turns out to be a GPIO that was changed between Cotulla and Bulverde. I've sent RP patch. Sep 07 19:49:08 okay good nite Sep 07 19:54:50 hmm this pub used to be the black hole of mobile Sep 07 19:55:36 * chouimat is back. Sep 07 19:56:07 xoragprs: I couldn't make it tonight :( Sep 07 19:56:42 lrg very busy tonight Sep 07 19:56:54 XorA|gprs: new members ? Sep 07 19:57:16 no idea, i barely know anyone Sep 07 19:57:59 XorA|gprs: heh, they're probably all a bit dodgy.... Sep 07 19:58:27 at least 2 zaurus owners Sep 07 19:58:42 XorA|gprs: btw what gcc do you use for corgi ? Sep 07 19:59:03 XorA|gprs: I think we may have a gcc issue with alsa. Sep 07 19:59:28 XorA|gprs: another alsa gcc issue Sep 07 19:59:34 3.4.4 Sep 07 20:00:24 XorA|gprs: same as RP, I'm seeing differences between some parameters on my setup (angstrom based) and spitz. Quite strange :-/ Sep 07 20:00:53 gcc ain't fun Sep 07 20:01:04 lrg well i hopey buy a terriefr nexgt week Sep 07 20:01:51 XorA|gprs: cool. btw, do you know if there is a roadmap for openzaurus to use EABI ? Sep 07 20:01:58 god typing on this keyboard sucks Sep 07 20:02:05 heh Sep 07 20:02:17 lrg no idea i guess when it works Sep 07 20:02:46 * XorA|gprs looks at koen Sep 07 20:03:06 * lrg looks at koen Sep 07 20:04:05 * koen looks at http://wiki.openzaurus.org/RoadMap Sep 07 20:04:24 "One of the rules of Angstrom is that "a machine must have a working 2.6 kernel" - so when the first release will appear, it will support c7x0 and cxx00, so I will drop those machines from OpenZaurus - this will force users to switch but I do not want to develop for two distros at the same time. After that OZ will support only the 2.4 powered machines (collie, poodle, tosa) and it will be less maintained." Sep 07 20:04:37 lrg: so is the probem gcc 4 only Sep 07 20:05:00 koen: you always have a good answer :) Sep 07 20:05:17 teflon shoulders Sep 07 20:05:24 :) Sep 07 20:05:51 XorA|gprs: I'm not sure, it works as expected on gcc 4.1.1 with cse patch. Sep 07 20:06:28 did anyone try it with the mb(); inserted? Sep 07 20:06:39 Lrg im sure the cse bug report said the bug affected all gcc versions Sep 07 20:06:47 koen: RP is going to try this Sep 07 20:07:03 XorA|gprs: ah, interesting Sep 07 20:07:11 i guess it would make sense if all modern had same bug Sep 07 20:07:23 XorA|gprs: true. Sep 07 20:08:32 I'm assuming with an embedded system you need to use ACPI for power managment correct? APM is only used if the BIOS handles the PM correct? Sep 07 20:08:53 Gerrath: apm and acpi are both x86-isms Sep 07 20:09:20 Gerrath or its emulated like on zaurus Sep 07 20:09:56 * chouimat is away: pub :D Sep 07 20:10:45 * lrg wishes he was in pub Sep 07 20:11:23 lrg still at work? Sep 07 20:13:10 XorA|gprs: sort of, at home cleaning up some patches. I'm pretty busy atm. It will be good when you start, I can get a day off :D Sep 07 20:13:36 Lrg erk Sep 07 20:14:13 XorA|gprs: I heard lrg is printing "The new guy will take care of it" posters Sep 07 20:14:35 lrg: I'll test now :) Sep 07 20:14:48 XorA|gprs: don't worry, it's not that bad. It's just this time of year (pre-xmas rush for oems to complete projects) Sep 07 20:14:55 RP: ok Sep 07 20:15:04 * lrg has his fingers crossed Sep 07 20:15:13 * XorA|gprs images the inquisition now Sep 07 20:15:19 koen, so how is the power management controled on an embedded system (ARM)? I've looked through the kernel source code and there is sleep and standby support for the ARM I"m using I'm just not sure how it gets hooked into everything. Sep 07 20:16:11 Gerrath: AFAIK mostly via apm emulation Sep 07 20:17:10 any python folks around? Sep 07 20:17:11 Sep 07 20:17:16 koen, so I would choose to enable APM emulation in the kernel then. Sep 07 20:17:19 what would cause a file.write() to emit that? Sep 07 20:17:20 heh Sep 07 20:17:28 looks like its encoding was ascii instead of utf8 or something Sep 07 20:17:29 hmm Sep 07 20:20:43 anyway i shall leave for train Sep 07 20:22:50 lrg: Still the same :-( Sep 07 20:23:25 RP: okay, thanks :-/ Sep 07 20:25:14 RP: I wonder if the mb() needs to be moved into another of the refines() or params(). I'll create another patch that forces only16bit as the param into alsa-core, it will be interesting if it still comes back with 32 Sep 07 20:25:50 lrg: ok. I'll also try and add some debugging further back and see where this value is coming from... Sep 07 20:26:47 RP: okay, thanks. Look at pcm_params.h, there are arrays of params here that are refined at runtime to create the pcm_format mask. Sep 07 20:33:18 RP: test patch sent Sep 07 20:41:36 lrg: Forcing that fixes it Sep 07 20:43:28 RP: very interesting ! Sep 07 20:44:10 RP: I'm thinking gcc issue (since it works with 4.1.1 +cse), what about you ? Sep 07 20:44:40 lrg: Its funny I've never seen any issue with gcc before though? :-/ Sep 07 20:46:08 RP: me too, although the arm gcc had difficulties with alsa param refinement. Btw, it might also be interesting to see what happens with a -O1 kernel. Sep 07 20:47:56 RP: afaik, we are the only guys using > 16 bit audio on ARM Sep 07 20:49:23 RP: thanks for this work, it now gives me some food for thought on this issue. Sep 07 20:49:33 RP: :) Sep 07 20:49:43 lrg: np :) Sep 07 20:50:47 XorA|gprs: train must have been running on time tonight ;) Sep 07 20:51:35 lrg mine is normally ok. just waiting to pull out of waverly Sep 07 20:52:15 XorA|gprs: btw free wifi at haymarket. unsecured access :) Sep 07 20:52:49 lrg cool, i shall have to scan that next time im bored Sep 07 20:53:18 lrg though with web n walk i normally too lazy to check these days Sep 07 20:53:29 XorA|gprs: bt voyager ap, works fine as long as no train stop in P1 Sep 07 20:53:46 * XorA|gprs chuckles Sep 07 20:55:03 reminds me, must remember to retain work wep key when i leave access on princes st always good Sep 07 20:55:22 XorA|gprs: heh, good idea. Sep 07 20:56:06 although aspects wont be there much longer Sep 07 20:57:16 now real test, tunnel Sep 07 20:57:46 XorA|gprs: it should stay connected, but with RTT > 10secs Sep 07 20:58:34 ping XorA|gprs Sep 07 20:58:35 and through Sep 07 20:58:50 heh Sep 07 20:59:05 never tried that before Sep 07 21:00:25 XorA|gprs: I've had gprs survive a 6 minutes 'middle of nowhere' outtage Sep 07 21:01:54 unfortuneately htc wizard likes to choose to hang up when idle Sep 07 21:02:46 htc lizard, only works when hot Sep 07 21:03:03 heh Sep 07 21:03:40 oh hurry up sharp with then gsm wifi z Sep 07 21:10:08 hmmm Sep 07 21:10:32 bitbake-trunk isn't exactly behaving like I thought it would Sep 07 21:10:50 CosmicPenguin: In what way? Sep 07 21:10:56 The cache isn't accelerated Sep 07 21:11:04 I just changed a .bb and its going through the whole thing again Sep 07 21:11:15 also, there are lots of blank NOTE: lines during a --dry-run Sep 07 21:11:27 nothing changed in trunk to do with the cache Sep 07 21:11:31 wierd Sep 07 21:11:52 The NOTE lines are from base.bbclass. I must remember to merge a fix for that... Sep 07 21:17:42 ah the bright lights oh home, bye Sep 07 21:17:52 hmm.... Sep 07 21:22:35 RP: I'm digging about in base.bbclass on an unrelated subject - whats the fix for the NOTE: thing? Sep 07 21:22:54 oh, wait - this is probably my bad Sep 07 21:29:46 --dry-run seems to have lost most of its usefulness Sep 07 21:30:46 CosmicPenguin: s/print msg/if msg: print msg/ Sep 07 21:30:55 RP: yeah, thats in there Sep 07 21:31:03 RP: we didn't have it in our copy of base.bbclass Sep 07 21:31:07 thanks Sep 07 21:31:43 I began to wonder if it had made it into OE :) Sep 07 21:33:12 Why has dry-run lost its usefulness? Sep 07 21:33:25 RP, Im getting my zaurus tommorow. Anything usefull I should know? Sep 07 21:33:26 because many of the packages get listed many multiple times Sep 07 21:34:34 CosmicPenguin: The multithreading changes a lot of assumptions the core makes. There are a few things like that I need to think about... Sep 07 21:34:45 nod Sep 07 21:35:12 Kristoffer: Not really, most info you need is on the OZ site Sep 07 21:35:36 CosmicPenguin: I want to totally change the message handling anyway... Sep 07 21:35:45 RP, oki thx Sep 07 21:35:56 RP: yeah - it seems like some new stuff is needed Sep 07 21:36:26 CosmicPenguin: lib/bb/msg.py is a start with different message domains Sep 07 21:36:42 then we can start to look at making the messages useful as well Sep 07 21:37:39 RP, an its no chance I can brick the zaurus right?...just wanna make sure, its an expensive toy Sep 07 21:38:08 Kristoffer: I'll never say never but it would be difficult to do Sep 07 21:38:24 RP: still - quite cool Sep 07 21:38:46 CosmicPenguin: Its working for you? How many cores? :) Sep 07 21:38:51 RP: 2 Sep 07 21:39:15 This is my regular old obsolete dual opteron Sep 07 21:39:16 :) Sep 07 21:39:20 CosmicPenguin: Could do better ;-) Sep 07 21:39:27 Indeed Sep 07 21:39:32 This box is my work-horse though Sep 07 21:39:44 its been building OE for neigh on two years now Sep 07 21:39:48 probably more Sep 07 21:40:07 My work horse keeps having disk failures due to OE :-/ Sep 07 21:40:54 RP: lrg says he fixed the audio/mmc issue. :-) Sep 07 21:41:54 JustinP: Yes, we have a simple patch for that. I also know how to hack ASoC to give sound... Sep 07 21:42:21 I should update the oz kernel Sep 07 21:42:48 RP, yes. Remember now that I got one of those I will bug you with kernel updates :D Sep 07 21:42:57 * CosmicPenguin brings up his dual dual core and gets it read for OE Sep 07 21:43:00 just for you, RP Sep 07 21:44:56 RP: :-) I appreciate your digilance Sep 07 21:45:08 CosmicPenguin: The sad fact is I wrote that code for my dual Xeon box and its yet to run on it :-/ Sep 07 21:45:31 Kristoffer: You mean give me updated patches or bug me for updates? Sep 07 21:45:53 RP, bug you for updates mostly :D But from what I've heard the kernel is in good shape Sep 07 21:46:05 RP: that is sad Sep 07 21:46:18 Kristoffer: bugging me for updates doesn't work ;-) Sep 07 21:46:38 CosmicPenguin: just lack of time. That box needs certain stability and is remote :-/ Sep 07 21:46:42 RP, :D Sep 07 21:51:23 It appears ASoC is only breaking for OSS emulation. Shame Liam has gone... Sep 07 21:52:12 RP: oss works on corgi Sep 07 21:53:11 XorA|gone: Corgi probably doesn't support 32 bit Sep 07 21:53:34 * RP is wildly guessing though :) Sep 07 21:53:49 RP: possibly, off to bed now so not gonna test it, cyas Sep 07 21:54:03 'night XorA|gone Sep 07 21:57:18 RP: yeah - it would be good to know which core was doing what Sep 07 21:58:12 CosmicPenguin: That leads into a whole branch of things to dowith a GUI... Sep 07 21:58:25 indeed Sep 07 22:01:17 This will be particuarlly nice when bulding X Sep 07 22:01:31 half the time for building all those dang protos Sep 07 22:02:18 Even on a single core, being able to fetch in advance will be a big bonus Sep 07 22:02:56 At present the number of thread handling code is just plain stupid but once we improve that to learn about different tasks... Sep 07 22:17:00 sigh Sep 07 22:28:40 sigh - why is it insisting on building glibc-intermmediate? Sep 07 22:29:46 Probably some very warped logic :-/ Sep 07 22:29:49 CosmicPenguin: probably because the PREFERRED_PROVIDER was never set because the default happened to be the right one for people Sep 07 22:30:03 CosmicPenguin: and your version of bitbake changed it :-P Sep 07 22:30:06 or what RP said Sep 07 22:30:14 he actually works on bb Sep 07 22:30:30 I removed the preferred provider - it seemed like nobody used it any more Sep 07 22:30:40 JustinP: The preferred_provider is a better idea ;-) Sep 07 22:30:41 and the 2.4 version was PREFERENCE="-1" anyay Sep 07 22:31:10 The sooner we make bitbake's output readable, the better :-/ Sep 07 22:31:27 I'm sure it has a reason it thinks its a good idea but sometimes its hard to read Sep 07 22:31:31 very hard to read Sep 07 22:49:49 03rpurdie 07org.oe.dev * r3f1cab36... 10/ (3 files in 3 dirs): linux-oz-2.6.17: Update to latest ASoC (and disable 32bit audio to get cxx00 working, include mmc conflict fix). Add cxx00 overlay support. Update poodle patches preparing patch for mainline. Sep 07 23:09:20 sigh - sugar is going to enforce a wholesale update to gnome 2.16 /me fears Sep 07 23:33:53 RP: :-) I'll try that kernel Sep 08 00:01:02 ~seen psokolovsky Sep 08 00:01:18 psokolovsky is currently on #oe (6h 46m 52s) #handhelds.org (6h 46m 52s), last said: 'O_Neil: RDEPENDS is a must, ipkg won't let you install package unless all RDEPENDS are satisfied. RRECOMMENDS just to note a relation to another package. There's even weaker form RSUGGESTS.'. Sep 08 00:53:26 <[g2]> is the OE/Bitbake evolution documented anywhere ? Sep 08 00:54:03 * [g2] is curious about the last few years from buildroot, to familiar/opie/gpe/bitbake Sep 08 00:54:19 <[g2]> I'm looking for a very rough overview Sep 08 00:57:18 [g2]: probably http://www.oesf.org/forums will give you such history , but I doubt you will find a rough overview :-/ Sep 08 00:58:01 [g2]: ah maybe the FOSDEM presentation by mickey|working Sep 08 00:58:44 [g2]: http://www.openembedded.org/papersandpresentations Sep 08 00:59:01 <[g2]> cyrilRomain THX **** ENDING LOGGING AT Fri Sep 08 02:59:56 2006