**** BEGIN LOGGING AT Sun Nov 26 02:59:57 2006 **** BEGIN LOGGING AT Sun Nov 26 05:45:37 2006 **** BEGIN LOGGING AT Sun Nov 26 06:19:27 2006 Nov 26 06:24:13 hey, hrw|gone Nov 26 06:38:19 I've shared out my openEmbedded dir via nfs... when something is being currently compiled is their a lock file created? as in if I tell another computer to compile task-opie will it skip over the already being compiled app and start on the next? Nov 26 07:24:44 is their a bitbake remove to remove a package? Nov 26 08:10:54 zezom: bitbake -c clean the_package Nov 26 08:53:24 cyrilRomain, thanks... is their a list of flags that -c can take? it seems to be quite uesful Nov 26 08:53:55 zezom: bitbake --help Nov 26 08:54:17 zezom: ah sorry, read too quickly Nov 26 08:54:36 zezom: I think it is in the bitbake manual Nov 26 08:56:17 zezom: you can use 'bitbake -i' for the interactive bitbake shell Nov 26 08:56:30 zezom: then 'help' give you a list of command Nov 26 09:10:56 ping zecke Nov 26 09:12:51 BACK Nov 26 09:15:02 zecke : As per the problem with bitbake i told you yesterday. The last version from the cvs that runs without the problem is 1.4.2 Nov 26 09:24:09 good morning all Nov 26 09:26:06 koen : Hi ! Nov 26 09:27:00 hey koen Nov 26 09:27:16 hey Ifaistos, cyrilRomain & pH5 Nov 26 09:27:27 good morning Nov 26 09:33:09 hey gremlin[it] , your image does not behave well on my h3760 Nov 26 09:33:58 hi emte :) Nov 26 09:34:02 what is not good ? Nov 26 09:34:04 gpe doesnt like to stay running Nov 26 09:34:21 the rest seems okay Nov 26 09:34:55 altho i am not sure what is causing hard power off Nov 26 09:35:45 but i could not test any of the touchscreen stuff Nov 26 09:35:49 mhh strange ... can you send me the dmesg output and /varr/log/message or other kind of log tounderstand what is wrong ? Nov 26 09:36:31 have not been able to keep it running long enough to capture the logs via serial Nov 26 09:36:52 ohhh really strange ... Nov 26 09:37:01 i'll test .. thanks :) Nov 26 09:37:13 was going to try to reinstall the image again to see if maybe it was messed up Nov 26 09:38:01 just thought i'd let you know i couldnt test your touchscreen stuff Nov 26 09:46:13 cyrilRomain, thanks again Nov 26 09:46:55 zezom: you welcome Nov 26 10:06:01 morning all Nov 26 10:06:55 morning Nov 26 10:07:14 I bought my first printer Nov 26 10:07:43 which is also scanner and copier Nov 26 10:07:43 ~hail usb 2.0 Nov 26 10:07:43 hey RP & hrw Nov 26 10:07:43 * ibot_ bows down to usb 2.0 and chants, "I'M NOT WORTHY!!" Nov 26 10:09:16 5096x7016 scan (a4 600dpi) = 60M png Nov 26 10:10:20 ~curse epson for manual in html Nov 26 10:10:20 May you be reincarnated as a Windows XP administrator, epson for manual in html ! Nov 26 10:15:01 hey zecke Nov 26 10:15:22 hrw: 600dpi is a bit bogus, since regular print is 300dpi Nov 26 10:15:48 600dpi is usefull for reading the fineprint on money :) Nov 26 10:17:19 koen: I know - wanted to check how long it will take to transfer 100M via usb Nov 26 10:18:12 anyway I have much better thing to test - my grandmother photo from 193x/194x Nov 26 10:18:16 interesting: http://hughsient.livejournal.com/9350.html Nov 26 10:20:09 how does it differ from current minilite (or how gpe one is called) Nov 26 10:20:38 this one queries HAL and uses gnome-power for policy Nov 26 10:20:57 and iirc hal has support for sysfs bl Nov 26 10:21:01 so should work anywhere without hardcoded paths? Nov 26 10:21:06 yes Nov 26 10:22:11 when I will have time I have to flash progear with older bios and learn acpi Nov 26 10:22:24 older bios? Nov 26 10:22:26 last ver of bios has broken DSDT Nov 26 10:22:31 heh Nov 26 10:22:33 no battery info Nov 26 10:23:05 "At the moment I would probably not recommend Lenovo to somebody that wants Linux compatible hardware, which is a shame considering how good most of the Thinkpads used to be." Nov 26 10:23:10 poor mickeyl ;D Nov 26 10:26:09 hrw: could you push that wpa-supplicant recipe? Nov 26 10:27:26 will do Nov 26 10:27:35 thanks Nov 26 10:27:40 * koen starts clean rebuild Nov 26 10:27:41 first have to finish packages/linux/linux_2.6.19-rc6.bb Nov 26 10:34:56 03koen 07org.oe.dev * r8aa660d3... 10/ (1 removal.txt): removal.txt: add bluez 2.x Nov 26 10:37:55 I need bigger desk Nov 26 10:38:05 atleast 90x130cm Nov 26 10:44:46 03hrw 07org.oe.dev * re70383b5... 10/ (4 files in 3 dirs): Nov 26 10:44:46 wpa-supplicant: add 0.5.5 with optional madwifi-ng support Nov 26 10:44:46 - machines with PCI feature will get madwifi-ng support and package will be Nov 26 10:44:46 MACHINE specific Nov 26 10:44:46 - this is result of my RFC [1] and Koen's solution [2] Nov 26 10:44:47 1. http://lists.linuxtogo.org/pipermail/openembedded-devel/2006-October/000698.html Nov 26 10:44:49 2. http://lists.linuxtogo.org/pipermail/openembedded-devel/2006-November/000848.html Nov 26 10:48:32 woooo Nov 26 10:48:39 I have a beer sponsor for FOSDEM 2007 Nov 26 10:48:56 the more pleasant kind of funding: beer Nov 26 10:49:11 ;) Nov 26 10:49:36 hm. fosdem will be occassion to check what you see in laffe Nov 26 10:50:23 http://bugs.treke.net/show_bug.cgi?id=1611 Nov 26 10:50:27 solved or not? Nov 26 10:50:48 I don't think it's a bug Nov 26 10:51:08 the exact same works for ep93xx-kernel, linux-omap, linux-handhelds and linux-rp Nov 26 10:51:25 I suspect /bin/sh isn't bash Nov 26 10:51:34 then comment Nov 26 10:51:46 bugs.treke.net isn't working for me Nov 26 10:52:12 I already commented on it with in a mail conversation with valentin and liam Nov 26 10:52:20 rm ~/screenlog.1;BB_NUMBER_THREADS=1 time bitbake -cfetchall world -k;grep 'Build of .* do_fetch failed' ~/screenlog.1 |sort -u|wc Nov 26 10:52:27 time to recheck fetching Nov 26 10:52:54 koen: Package ep93xx-kernel-2.6.17+2.6.18-rc1-r1 Nov 26 10:52:56 fetch fails Nov 26 10:53:36 ${@base_contains('COMBINED_FEATURES', 'pci', omap5912osk, arm, d)} Nov 26 10:53:36 ERROR: Error in executing: Nov 26 10:53:36 ERROR: Exception:exceptions.NameError Message:name 'omap5912osk' is not defined Nov 26 10:53:41 need to fix Nov 26 11:01:01 koen: angstrom distro will get tweaks Nov 26 11:06:05 03hrw 07org.oe.dev * r1d1e642d... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom: added progear support Nov 26 11:06:05 03hrw 07org.oe.dev * rabf23eb8... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom: orinoco 0.13e is 2.4 only - under 2.6 you want to use 0.15 from kernel Nov 26 11:06:10 03hrw 07org.oe.dev * r422a87f2... 10/ (1 conf/distro/include/angstrom.inc): angstrom: we want PCI support Nov 26 11:07:01 progear support will follow Nov 26 11:23:34 03hrw 07org.oe.dev * r5cccddf0... 10/ (1 conf/machine/progear.conf): progear: added machine config Nov 26 11:23:42 03hrw 07org.oe.dev * r74c74685... 10/ (1 packages/wpa-supplicant/wpa-supplicant_0.5.5.bb): wpa-supplicant: fix parse error Nov 26 11:23:51 03hrw 07org.oe.dev * r0db2e158... 10/ (1 conf/bitbake.conf packages/tasks/task-base.bb): Nov 26 11:23:51 moved COMBINED_FEATURES and rest of task-base calculations from task-base.bb into bitbake.conf Nov 26 11:23:52 - without this move we can not use COMBINED_FEATURES (and other task-base Nov 26 11:23:52 variables) in other recipes (like we do not in wpa-supplicant 0.5.5) Nov 26 11:24:01 03hrw 07org.oe.dev * r64360d71... 10/ (1 conf/machine/native.conf): native: set images to 'tar.gz' if not set Nov 26 11:24:08 03hrw 07org.oe.dev * r95d88825... 10/ (1 packages/tasks/task-opie.bb): task-opie: switched to task-base style Nov 26 11:40:36 does someone build x86 on x86-64? Nov 26 11:40:50 nope Nov 26 11:40:59 x86-64 userspace, you mean Nov 26 11:41:24 host: x86-64, target: x86 Nov 26 11:41:35 and by x86-64 I mean 64bit kernel+userspace Nov 26 11:41:49 nope Nov 26 11:42:05 I have been thinking about having OE build a x86_64 distro Nov 26 11:42:15 I could load that in a VM under OSX Nov 26 11:42:30 hrw: how much storage does the progear have? Nov 26 11:42:47 5GB hdd Nov 26 11:42:50 ,l Nov 26 11:43:03 its normal 2.5" hdd Nov 26 11:43:09 koen: got a core duo 2? Nov 26 11:43:16 I've also been thing about a task-fsck-busybox-I-have-space Nov 26 11:43:18 zecke: yes Nov 26 11:43:24 koen: bastard Nov 26 11:43:32 koen: nice idea Nov 26 11:44:11 coreutils, findutils, procps, psmisc, util-linux, bash, etc Nov 26 11:44:27 zecke : Hi ! Nov 26 11:44:50 zecke : As for the problem with bitbake i told you yesterday. The last version from the cvs that runs without the problem is 1.4.2 Nov 26 11:45:03 we don't have cvs Nov 26 11:45:15 zecke : svn Nov 26 11:45:39 Ifaistos: so 1.4.3 doesn't work anymore? 1.6.3 doesn't work either? Nov 26 11:45:53 Ifaistos: don't mix tabs and spaces in python code Nov 26 11:46:18 zecke : no they don't neither 1.7.2 (i think) dev brach Nov 26 11:46:38 I refuse to believe that (specially 1.4.2 vs. 1.4.3) Nov 26 11:46:50 specially as 1.4.2 and 1.4.3 are identical Nov 26 11:47:19 anyway I need to study now Nov 26 11:47:29 koen: slugos would use that task :-) Nov 26 11:47:30 Ifaistos: I have shown you the tools to debug, use them wisely Nov 26 11:47:41 zecke: did you get the patch from lennert I forwarded you? Nov 26 11:48:04 koen: to which branch should this be applied, I'm sure I have edited the MANIFEST somewhere already Nov 26 11:48:11 zecke : i tried 1.4.2, 1.4.4, 1.6.0, 1.6.2 and 1.6.4 Nov 26 11:48:14 good point Nov 26 11:48:20 Ifaistos: -DDD will show you which files got opened Nov 26 11:48:27 Ifaistos: take a look where you mix space and tabs Nov 26 11:48:53 zecke : will do that. thanks Nov 26 11:50:35 Ifaistos: and as I have said yesterday look at the output of python -vvv to see where it loads the bb modules from... Nov 26 11:53:36 | checking for memcpy... configure: error: Link tests are not allowed after GCC_NO_EXECUTABLES. Nov 26 11:53:39 | make[1]: *** [configure-zlib] Error 1 Nov 26 11:53:44 ok. looks like place to search for Nov 26 11:54:44 hrw: got the same error here Nov 26 11:55:29 hrw: are you trying to build on a 32bit cpu? Nov 26 11:57:26 nicolasfr: on 64bit for 32bit Nov 26 11:57:58 I solved this by removing "multidir = 64" in zlib/Makefile Nov 26 11:58:30 thx for hint Nov 26 11:58:37 I have a 32bits cpu and gcc-cross-4.1.1 tries to build with "gcc m64" flag, hence the problem. Nov 26 11:58:57 I haven't found how to solve the problem from the beginning Nov 26 12:00:04 are anonymous functions in .incs broken in bitbake trunk? Nov 26 12:01:40 in particular the binary locale funtion in glibc? Nov 26 12:08:03 koen: tabs. vs. spaces? got an error message? Nov 26 12:08:22 zecke|away: the glibc build acts like the function isn't there Nov 26 12:08:35 and no error messages I'm aware off Nov 26 12:08:59 if I type "jksdhjfjshgskj" inside the function I get an error during parsing Nov 26 12:09:38 I got suspicious when my hx4700 OOMed with localegen and the glibc build didn't take 2 hours Nov 26 12:13:30 koen: hmm Nov 26 12:14:07 koen: which file is this anonfunc in? Nov 26 12:16:19 zecke|away : Here is a full debug log -> http://pastebin.ca/258336 Nov 26 12:17:21 zecke|away : It complains that the base.bbclass has tab/spaces problems Nov 26 12:18:24 Ifaistos: and you don't have any changes to it? Nov 26 12:18:40 but this doesn't explain 1.4 and 1.6 issues Nov 26 12:18:51 zecke|away : no its straight from the oe repository Nov 26 12:20:27 zecke|away : In all cases the metadata/conf files are exactly the same. i only change the bitbake version Nov 26 12:20:51 zecke|away : remove old, checkout another version from the rep Nov 26 12:21:09 any files not from OE/BitBake that gets loaded? Nov 26 12:21:16 e.g. local machines, distros... Nov 26 12:21:40 zecke|away : my distro definition files owmnr.conf Nov 26 12:22:05 zecke|away : these are not the oe rep Nov 26 12:22:37 sorry need to study but what about looking into these files? Nov 26 12:23:49 zecke|away : i'll re-write them from scratch (to be safe), but it looks like something changed after 1.4.2 Nov 26 12:27:16 Ifaistos: http://www.rafb.net/paste/results/DPsYn221.html this changed in the bitbake 1.4 branch Nov 26 12:27:24 03mwester 07org.oe.dev * r85bfa5e5... 10/ (4 files in 3 dirs): slugos-init: add support for the DSM-G600 Nov 26 12:27:33 if you say that 1.4.3 is broken already check the differences Nov 26 12:27:41 03mwester 07org.oe.dev * re2e9353b... 10/ (1 packages/linux/ixp4xx-kernel_2.6.19-rc5.bb): ixp4xx-kernel: add 77-dsmg600-cmdline patch Nov 26 12:27:47 03rwhitby 07org.oe.dev * r6e2f8d4e... 10/ (1 packages/slugos-init/files/functions): slugos-init: Added initial support for FSG-3 Nov 26 12:27:57 03rwhitby 07org.oe.dev * r101cc837... 10/ (4 files in 3 dirs): ixp4xx-kernel: Updated to 2.6.19-rc6 Nov 26 12:28:16 just eat some bran, CIA-4 ? Nov 26 12:28:19 Ifaistos: r533 could be a candidate which would be triggered (e.g. by my broken icecc class) Nov 26 12:29:08 I'm away now Nov 26 12:30:35 another rebuild attempt... Nov 26 12:32:23 Ifaistos: there is mixed tabs and space in base.bbclass but I don't have any problem building here ... Nov 26 12:32:30 s/is/are/ Nov 26 12:32:45 * cyrilRomain uses bitbake 1.6.3 Nov 26 12:33:09 cyrilRomain : From cvs ? Nov 26 12:33:13 svn Nov 26 12:33:25 bitbake 1.6.3 from svn yes Nov 26 12:33:50 and OE database from ... one or two days Nov 26 12:34:05 Ifaistos: any special INHERIT? Nov 26 12:35:02 zecke|away: you got it: base.bbclass:23 Nov 26 12:35:49 zacke|away : yes a INHERIT += "owmnr-mirrors" in local.conf Nov 26 12:36:16 Ifaistos: how many mirrors? one? Nov 26 12:36:34 Ifaistos: any special pythong magic in there? Nov 26 12:36:40 anyway, now I'm really away Nov 26 12:37:10 zecke|away: 2 mirros. no magic Nov 26 12:37:31 MIRRORS_prepend () { Nov 26 12:37:31 ftp://.*/.*/ http://www.ifaistos.awmn/source/current Nov 26 12:37:31 http://.*/.*/ http://www.ifaistos.awmn/source/current Nov 26 12:37:31 } Nov 26 12:37:33 hrw: hi Nov 26 12:37:43 hi schurig Nov 26 12:38:05 Ifaistos: hm. I need to push own-mirrors.bbclass Nov 26 12:38:13 hrw: I think it is easier (and saver) to continue with the GPL code from OLPC Nov 26 12:38:43 schurig: agreed Nov 26 12:39:16 I've made a directory with this in SVN, the wlan_* things compile cleanly against 2.. Nov 26 12:39:24 2.6.19-rc6 already Nov 26 12:39:37 great Nov 26 12:39:47 there's a GTODO Nov 26 12:39:59 (he battles with a murky laptop keyboard) Nov 26 12:40:21 there's a TODO file there that shows you all the errors/warnings I get. Not much :-) Nov 26 12:40:51 and of course they are all there becuase I don't have any pcmcia_cs support there :-) Nov 26 12:41:15 ok - I'll look Nov 26 12:41:24 from what I know so far the differences between 8388 (OLCP) and 8385 (us) are minimal Nov 26 12:41:34 schurig: my card is no name - checked Nov 26 12:41:59 in the patched directory, all they do is to load a different firmware blob, but AFAIK no different codepaths Nov 26 12:42:19 the only one signs are: "WLAN CF Card\n802.11b/g" and mac on reverse Nov 26 12:42:33 hrw: does your card give the same pccardctl info output as mine (as the one in the main README?) Nov 26 12:42:50 my card has an AmbiCom label Nov 26 12:42:55 let me see on my page Nov 26 12:43:10 Socket 1: product info: "Marvell", "802.11 CF", "ID: 04", "" manfid: 0x02df, 0x8103 function: 6 (network) Nov 26 12:43:45 same id Nov 26 12:43:46 hrw: yes, the same as mine Nov 26 12:43:57 hrw: I guess that Stephan simplyd Nov 26 12:44:05 didn't remove the label Nov 26 12:44:10 ;) Nov 26 12:44:28 00:16:41 is probably ambicom OUI Nov 26 12:44:56 nope Nov 26 12:45:16 where is this OUI from? And: what's an OUI? Nov 26 12:45:36 first 3 octets of MAC == manufacturer Nov 26 12:46:00 00-16-41 (hex) USI Nov 26 12:46:01 001641 (base 16) USI 141, LANE 351, TAIPING RD. SEC.1, TSAO-TUEN NAN-TOU 542 TAIWAN, REPUBLIC OF CHINA Nov 26 12:46:11 this is real manufacturer of card Nov 26 12:46:20 http://standards.ieee.org/regauth/oui/oui.txt Nov 26 12:46:34 ahh, an old friend Nov 26 12:46:46 our Hermes 2 cards where from USI Nov 26 12:47:20 ;) Nov 26 12:48:38 I imagined you with phone: 'hi man, I have on of yours 802.11g cf cards. dont ask where I got one but send me 2000 of them' ;D Nov 26 12:49:57 nicolasfr: disable java and you will get working gcc-cross Nov 26 12:51:58 hrw: how do I disable java? Nov 26 12:54:07 nicolasfr: look into gcc_4.1.1.bb Nov 26 12:54:46 hrw: thanks, I try right now Nov 26 12:55:50 koen: http://pastebin.ca/258365 Nov 26 12:55:55 hrw: --disable-java in extra_oeconf I guess? Nov 26 12:56:22 hrw, oh no I see the LANGUAGES variable Nov 26 12:57:29 FSCK! Nov 26 12:57:55 I bought CF Type II -> PCMCIA adapter and this does not accept Type I cards Nov 26 12:57:57 ARGH Nov 26 12:59:29 hrw: doh! a couple of days ago I ordered a type1 cf card, and wanted to use it in CF type2->pcmcia adapter :/ Nov 26 13:00:10 nicolasfr: currently all my cf cards are type 1, but all hardware accept type 1/2 - now all except one... Nov 26 13:01:12 I ordered AmbiCom Compact Flash Wireless LAN 802.11b (WL1100C-CF). Do you know this card? Nov 26 13:01:32 normal prism2 card Nov 26 13:01:35 good choice Nov 26 13:03:02 good! these CF cards are so expensive... Nov 26 13:06:10 nicolasfr: can't comment - I never bought wifi card Nov 26 13:06:50 in whol life I bought 128M memorystick (sold later) and 16M cf card Nov 26 13:07:37 whol/whole Nov 26 13:16:10 ibot_: botmail for zecke: the anonfunc is in glibc-package.bbclass Nov 26 13:19:45 ERROR: log data follows (/data/build/koen/OE/build/tmp/angstrom/work/arm-angstrom-linux-gnueabi/wpa-supplicant-0.5.5-r0/temp/log.do_configure.27473) Nov 26 13:19:45 | /data/build/koen/OE/build/tmp/angstrom/work/arm-angstrom-linux-gnueabi/wpa-supplicant-0.5.5-r0/temp/run.do_configure.27473: line 272: syntax error near unexpected token `do' Nov 26 13:20:16 ~lart broken nslu2 mtn mirror Nov 26 13:20:16 * ibot_ does a little 'renice 20 -u broken nslu2 mtn mirror' Nov 26 13:22:24 http://www.hrw.one.pl/2006/11/26/epson-stylus-dx4000/ Nov 26 13:23:08 03koen 07org.oe.dev * r4bfb03b1... 10/ (1 packages/wpa-supplicant/wpa-supplicant_0.5.5.bb): wpa-supplicant 0.5.5: if [] ; then ; do isn't valid sh, remove ';do' Nov 26 13:23:59 I wonder if the hp laserjet duplex module works under linux Nov 26 13:25:03 koen: thx for fix Nov 26 13:25:04 ~lart gutenprint for running generated binaries Nov 26 13:25:04 * ibot_ strangles gutenprint with a 9-pole serial cable for running generated binaries Nov 26 13:25:17 hrw: disabling java in gcc-4.1.1.bb does indeed fix the problem thanks! Should that fix be pushed in OE? Nov 26 13:25:36 nicolasfr: no idea - can you report bug so we will remember and discuss? Nov 26 13:25:57 hrw: ok Nov 26 13:26:05 hi leoncamel Nov 26 13:26:33 leoncamel: got confirmation from ltg? Nov 26 13:27:09 hrw: just for my knowledge, how did you figure that disabling JAVA solves GCC compilation? Nov 26 13:28:05 nicolasfr: it failed inside of libjava Nov 26 13:28:17 hrw, well, not now. I will check it. Nov 26 13:29:01 hrw: ok thanks Nov 26 13:32:44 someone familiar with BT headsets? how many manufacturers produce shit which does not want to work with any device? Nov 26 13:40:59 almost all headsets Nov 26 13:42:25 koen: almost all works with any or vice versa? Nov 26 13:42:38 allmost all headsets are shit Nov 26 13:44:48 hey cool, I have python libiw bindings working on zaurus ;) Nov 26 13:44:54 Packaged contents of wpa-supplicant into /data/build/koen/OE/build/tmp/angstrom/deploy/glibc/ipk/wpa-supplicant_0.5.5-r0_arm.ipk Nov 26 13:44:57 WTF? Nov 26 13:45:17 'bitbake foo' creates armv5te ipkgs, 'bitbake -b foo.bb' creates arm ipkgs Nov 26 13:45:26 something is VERY broken in bitbake trunk Nov 26 13:45:28 simple wireless console scanner is working ;) Nov 26 13:46:43 ah, hmmm Nov 26 13:46:48 it's just wpa-supplicant Nov 26 13:48:04 koen: TARGET_ARCH instead of PACKAGE_ARCH probably... Nov 26 13:48:06 my fault Nov 26 13:52:23 have to go now Nov 26 13:53:28 03koen 07org.oe.dev * r2152f346... 10/ (1 packages/wpa-supplicant/wpa-supplicant_0.5.5.bb): wpa-supplicant 0.5.5: work around OE/bitbake arch bug Nov 26 13:58:07 bitbake sanity checker says I need 1.6.2. Can I upgrade from 1.6.0 or do I need to do a complete check out? "svn up" as usual did not do it. Nov 26 14:01:03 bitbake is not aware of .tbz2 extension? Nov 26 14:04:32 nicolasfr: only uclibc++/uclibc++_0.1.9.bb seems to use tbz2 Nov 26 14:05:00 03koen 07org.oe.dev * r9a4fb914... 10/ (3 files in 2 dirs): gcc-cross 4.1.1: fix libg* architecture bugs, closes #1265 Nov 26 14:07:37 * koen hugs immediate expansion Nov 26 14:07:53 i have big space problems with gpe on h3600 ... 16MiB of flash is really tight, Nov 26 14:08:00 Laibsch: http://www.openembedded.org/viewmtn/revision.psp?id=59e9ba26f46315792dd5f9c99d410cae393b9173 Nov 26 14:08:24 uclibc is full intterchangable with glibc ? Nov 26 14:08:25 gremlin[it]: no Nov 26 14:08:53 any doc about divverences / issues ? Nov 26 14:09:06 cyrilRomain: ok thanks. I am making recipes for packages that use .tbz2, I just used the uclibc++ method (add a new unpack method)... Nov 26 14:09:25 gremlin[it]: uclibc isn't even interchangeable with uclibc Nov 26 14:09:31 gremlin[it]: see uclibc .org Nov 26 14:11:54 mhh i notice uclibc is in oe.dev ... which architecture/distro use it ? Nov 26 14:12:57 gremlin[it]: I'd wait for uclibc .29 to get released Nov 26 14:13:04 .28 is over a year old Nov 26 14:13:15 ok thanks ! Nov 26 14:17:30 it really frustrating ... i build a gpe without any application and i have still a 15MiB image :( :( ... Nov 26 14:24:04 try using task-base with some less DISTRO_FEATURES Nov 26 14:25:16 mickeyl: guten morgen! Nov 26 14:25:19 hrw|gone: this is about the N100 (first Lenovo design), the thinkpads are very compatible Nov 26 14:25:22 koen: Guten Morgen! Nov 26 14:25:40 koen: btw, my HP 1320nw works pretty well on Linux with Duplex. Nov 26 14:25:53 ah, cool Nov 26 14:27:29 koen: nice, but I do not use poppler. Nov 26 14:27:38 bugs.openembedded.org down? Nov 26 14:27:52 no, its ip changed Nov 26 14:27:57 try bugs.treke.net Nov 26 14:31:57 neither treke.net work for me ... Nov 26 14:39:58 koen: thanks. Nov 26 14:43:19 how should I handle daemon files ? Gentoo uses /etc/init.d/ and /etc/conf.d/, but what is the policy in OE, or in particular in Angström ? Nov 26 14:43:41 the sintiax mount -at nonfs,nosmb isn't recognized ! Nov 26 14:43:52 neither mount -at jffs2,tmpfs ... Nov 26 14:43:57 * cyrilRomain guess it is more or less like debian Nov 26 14:44:06 only one filesystem at time can be specified :( Nov 26 14:44:07 cyrilRomain: correct Nov 26 14:44:29 gremlin[it]: use util-linux(u)mount Nov 26 14:44:33 koen: ok thanks. Should learn a bit then. I didn't used debian so much Nov 26 14:48:14 hello, were there any optimizations to build process of busybox? Like putting all sources in one big file and so... . Nov 26 14:48:36 * mmp just finished almost 5 hour long busybox compilation. Nov 26 14:49:00 most of time the laptop was doing swap-out, of course:) Nov 26 14:58:11 looks like bugs.treke.net is quite overloaded. Nov 26 15:11:59 In the receipe I'm trying to create, something creates a .debug directory in the do_install process. Do you know which command create it ? Nov 26 15:13:25 is it related to DISTRO_TYPE = "debug" ? Nov 26 15:17:48 it is 'install ${S}/utils/bitlbeed ${D}${sysconfdir}/init.d/bitlbeed' that create .debug directory into init.d Nov 26 15:22:08 does anyone know why .debug is created ? (if it is not a problem for now, I'd like in future to create .ipk without debug binaries) Nov 26 15:29:19 hmm are debug binaries always created, and not put into .ipk except when there is FILES_${PN}-dbg += "..." in the.bb ? Or am I wrong ? Nov 26 15:33:32 cyrilRomain: /etc/init.d/ only holds scripts Nov 26 15:36:21 koen: hmm why is .debug created then ? .. Nov 26 15:36:34 ah 'script' Nov 26 15:36:43 because bitbleed is a binary, not a script Nov 26 15:36:57 bitlbee comes with a dameon that is compiled Nov 26 15:37:05 system daemon usually live in /sbin Nov 26 15:37:21 where should bitlbeed be located then ? Nov 26 15:37:26 ok, thanks Nov 26 15:39:37 koen: I guess I should intensively use the variable set in bitbake.conf, like base_sbindir ? :) Nov 26 15:39:53 or sbindir even Nov 26 15:40:26 yes Nov 26 15:52:27 morning Nov 26 15:53:34 hi all Nov 26 15:53:41 hey florian Nov 26 15:54:06 If I do a mkdir in my bb file I am getting a permission denied how can I solve this ? Nov 26 15:54:57 gerwinin: using the 'install' command instead ? Nov 26 15:55:36 cyrilRomain if I do that I am getting the same :( Nov 26 15:56:15 gerwinin: ah.. :-/ Nov 26 15:59:11 gerwinin: check you don't try to create a directory on your host system. No idea alse Nov 26 16:00:04 http://www.dalprogrammers.com/pastebin/index.php?show=44 Nov 26 16:01:32 gerwinin: you probably have to use '${D}${includedir}' instead of '${includedir}' Nov 26 16:02:22 Okay I will try that Nov 26 16:06:00 cyril that seems to work Nov 26 16:07:32 gerwinin: to be fair if you asked me one our ago, I could not answer you ;) Nov 26 16:07:46 hi florian Nov 26 16:08:19 hey pb_ Nov 26 16:08:40 gerwinin: you can read the conf/bitbake.conf file which is a bible for the variable that should be used in .bb files Nov 26 16:09:37 gerwinin: you will see that using only ${includedir} was trying to write file on your host system Nov 26 17:10:07 ~seen zecke Nov 26 17:10:27 zecke was last seen on IRC in channel #oe, 5h 22m 1s ago, saying: 'Ifaistos: take a look where you mix space and tabs'. Nov 26 17:25:13 huhu Nov 26 17:26:41 koen: i didn't have a chance to say "i'm back" to dfoley. I'm really 200% loaded atm. and wouldn't have time to play with it anyway. Do you know someone who is interested? Nov 26 17:33:53 mickeyl: not from the top of my head Nov 26 17:34:36 koen: you use gprs from linux, right? have an URI for a good tutorial? (that is, gprs-over-usb here) Nov 26 17:34:54 gkdial Nov 26 17:35:00 it's just ppp Nov 26 17:35:09 with nasty carii Nov 26 17:35:12 carrier specific details Nov 26 17:35:51 hmm, i though everything i need is APN, login, pw Nov 26 17:35:59 i have it running on win ;) Nov 26 17:36:20 try putting 'debug; Nov 26 17:36:22 ' Nov 26 17:36:26 in the ppp options file Nov 26 17:36:29 grrr Nov 26 17:36:39 stupid different keyboard layouts Nov 26 17:36:46 mickeyl: I used wvdial, worked fine for me, I can give you my conf file if you want Nov 26 17:36:48 ~bon appetit Nov 26 17:36:57 from memory, bon appetit is smacznego. Guten Appetit. Eet Smakelijk. God Appetitt. Buon Appetito. Buen apetito Bom Apetite. buen apetito Nov 26 17:38:59 Hi! Nov 26 17:39:20 RP: Share the secret, how do you manage OZ kernel patches? ;-) Nov 26 17:41:01 psokolovsky: quilt Nov 26 17:41:11 psokolovsky: and actively pushing stuff upstream Nov 26 17:41:19 koen|away, only quilt? Nov 26 17:52:19 Well, that quilt thing appears to be a whole poor man's SCM ;-) Nov 26 17:52:56 It doesn't answer my usecase though - what if you already have normal (or subnormal) SCM to manage changes, Nov 26 17:53:13 you're screwed Nov 26 17:53:16 then how to manage patches against another branch of development? Nov 26 17:53:26 since you can't split out changes easily Nov 26 17:53:46 What I'm looking for is a way to do set arithmetic on patches Nov 26 17:54:10 koen|away, I exactly look for a way to split out changes - easily! ;-) Nov 26 17:54:16 you mean graph theory? Nov 26 17:55:30 psokolovsky: extract the big patch, split it up, run quilt, split it up further, run quilt, ... , profit Nov 26 17:55:58 koen, like substract one patch from another, or pull patch for file x, or pull hunk #y for file z Nov 26 17:57:01 yes, you have to pull it apart by hand or with diffsplit and tell quilt about each seperate group Nov 26 17:57:20 koen|away, well, I wrote what I needed in 2 hours, now I wonder if I should put another 5 hours in cleaning it up/documenting, or should just throw away because such a util(s) already exist ;-) Nov 26 17:57:40 koen|away, ok, thanks for hints Nov 26 17:57:45 quilt is pretty much the standard Nov 26 17:57:53 psokolovsky: and look into devmem2 as well Nov 26 17:58:29 psokolovsky: I'm getting sick of people reinventing the wheel at the kernel-discuss mailinglist Nov 26 17:59:38 koen|away, I know about devmem2, used to use it ;-) but if people use another crap during testing, it's better to just follow instead of starting discussion what should be used ;-) Nov 26 18:01:24 psokolovsky: As others have said, I simply use quilt Nov 26 18:01:36 psokolovsky: quilt+OE :) Nov 26 18:01:42 RP: ok, thanks Nov 26 18:04:33 psokolovsky: One of the reasons I gave up with hh.org kernels was that changes couldn't be kept as isolated entities within the control system. The other was all the bugs the hh.org tree had at the time compared to mainline Nov 26 18:06:43 RP, yes, I see. Btw, do you know about any SCM which has that ability - to define symbolic patchset? I.e. to say that change to file x and file y, etc. represent patchset "foo", with ability extract it regardless for further changes being done to those files (which would affect patchset content, obviously) Nov 26 18:08:24 psokolovsky: stgit might (its a quilt+git combo) but I've never tried it Nov 26 18:10:22 ok, will have a look in case - would like to know if there're solutions for such usecases. at least that would be something new in SCM evolution ;-) Nov 26 18:11:20 psokolovsky: The patch series approach has worked really well for openzaurus and works for the nslu2 people too. I'm taking the "if it isn't broken" approach atm... :) Nov 26 18:12:28 psokolovsky: I know some people had some issues with said approach when I started with it a few years ago but I think I've converted most into believers now ;-) Nov 26 18:14:05 RP, yes, I understand that, and experience from using it is really helpful. We still need to manage patches in HH.org CVS for mainline suibmission somehow, and I don't want to do that by manually converting to quilt ;-) Nov 26 18:14:54 psokolovsky: You could turn hh.org kernel cvs into a set of kernel patches ;-) Nov 26 18:15:28 psokolovsky: (I'm joking but only half joking ;-) Nov 26 18:16:06 RP: no, I can't ;-) But I can participate in submitting patches upstream ;-) Nov 26 18:21:17 psokolovsky: you can emulate quilt by putting every 'set' in its own branch Nov 26 18:22:36 koen, well, not with CVS ;-D. anyway, I have a solution here, seems to do what I need. Just want to be sure I don't reinvent the bike ;-) Nov 26 18:24:09 I was one of the people who rather have seen linux-rp being developed in a SCM Nov 26 18:24:18 I'm very happy now RP listened to me :D Nov 26 18:24:21 +not Nov 26 18:24:55 I'm convinced RP's workstyle is the best approach if you really are commited to pushing stuff upstream Nov 26 18:25:07 Jin^eLD: yes, please Nov 26 18:25:08 * koen coughs Nov 26 18:26:15 mickeyl: mail? Nov 26 18:26:27 Jin^eLD: mickeyl@linuxtogo.org please Nov 26 18:30:55 'morning' Nov 26 18:31:30 gremlin[it]: fitting in 16M flash is hardcore work Nov 26 18:31:52 gremlin[it]: if you have working sd or cf then I suggest going into root on external Nov 26 18:31:53 mickeyl: k, you have new mail :> Nov 26 18:31:56 hi hrw Nov 26 18:32:04 hrw: h36xx doesn't have SD/CF Nov 26 18:32:09 hrw: only via a sleeve Nov 26 18:33:03 suxx Nov 26 18:33:46 yep Nov 26 18:33:55 but h36xx is like poodle Nov 26 18:34:13 instead of pxa250/pxa255 it's 16MB/32MB Nov 26 18:34:13 koen: rather collie Nov 26 18:34:22 compaq ran out of small flash chips Nov 26 18:34:34 hi hrw ... h3600 don't have removable storage :( .. sich :( Nov 26 18:34:43 so sometimes you're lucky and get a 32MB h3600 Nov 26 18:34:46 gremlin[it]: so you are fscked Nov 26 18:35:00 gremlin[it]: stick with bootstrap rather Nov 26 18:35:51 maybe i can gain few hundreds KiB recompiling gpe with less library, sutting things like cairo or svg ... Nov 26 18:36:06 s/sutting/cutting/ Nov 26 18:36:15 gremlin[it]: remove wifi, bt support Nov 26 18:36:20 ipsec Nov 26 18:36:23 already done :) Nov 26 18:36:26 ppp Nov 26 18:36:52 at now only ppp/ipsec kernel modules ... no applications Nov 26 18:37:07 03koen 07org.oe.dev * r59e43ce3... 10/ (1 packages/bluez/bluez-gnome_0.6.bb): bluez gnome 0.6: add bluez-gnome, a config applet for bluez stuff Nov 26 18:37:38 gremlin[it]: ipv6 and ipsec takes over 0.5M Nov 26 18:39:51 ipsec and ipv6 .. neither kernel modules ... Nov 26 18:39:54 only ppp Nov 26 18:40:18 gremlin[it]: show me usr/lib/ipkg/status Nov 26 18:40:28 * koen needs to task-baseify task-angstrom* Nov 26 18:40:45 task-angstrom* has die Nov 26 18:42:50 Jin^eLD: arrived. thanks Nov 26 18:42:58 np Nov 26 18:44:43 hrw, i'll send via dcc or pastebin ? Nov 26 18:45:54 03koen 07org.oe.dev * rcf980daf... 10/ (1 packages/angstrom/task-angstrom-x11.bb): task-angstrom-x11: add rotate applet Nov 26 18:46:00 pastebin Nov 26 18:46:40 gremlin[it], so you're customer of uclibc, I guess ;-) Nov 26 18:47:25 psokolovsky yes i'm thinking about :) .. koen suggest to wait till next (soon) release will come Nov 26 18:47:32 ok Nov 26 18:47:42 .29 is a lot better as .28 Nov 26 18:48:59 if ill finish all other works befor 0.29 releas ei can start with a cvs snapshot ;) Nov 26 18:49:33 cvs? Nov 26 18:49:44 you mean svn Nov 26 18:49:54 ok a snapshot :) ... Nov 26 18:50:33 i didn't look what tool they use :) Nov 26 18:51:03 hrw, http://pastebin.com/833373 Nov 26 18:51:27 koen@bitbake:~/OE/monotone/org.openembedded.dev$ ls packages/uclibc Nov 26 18:51:27 angstrom-xscale.config files uclibc-0.9.26 uclibc-0.9.27 uclibc-0.9.28 uclibc-cvs uclibc-initial_svn.bb uclibc.inc uclibc_svn.bb Nov 26 18:51:35 gremlin[it]: that makes it pretty clear ;) Nov 26 18:52:14 i just don't look at :) Nov 26 18:54:31 psokolovsky: #1265 has been fixed, btw Nov 26 18:54:55 hrw: wpa-supplicant also has the right arch right now Nov 26 18:58:14 gremlin[it]: drop ALL locales Nov 26 18:59:29 gremlin[it]: alsa modules, ppp modules, irda modules Nov 26 19:01:11 drop procps - it use ncurses which take space Nov 26 19:01:25 koen: great, thx Nov 26 19:01:43 my progear build is at gcc-cross... Nov 26 19:05:07 also glibc-localedata-transit-* Nov 26 19:06:38 koen, seen notification, nice! now to rebuild gcc ;-) Nov 26 19:13:23 gremlin[it]: yes Nov 26 19:16:55 * koen hugs psplash Nov 26 19:17:22 * koen stabs non-working binary localegen Nov 26 19:21:41 koen, I heard opie has issues with psplash. just btw Nov 26 19:21:51 psokolovsky: so? Nov 26 19:21:58 psokolovsky: opie isn't supported in angstrom Nov 26 19:22:22 psokolovsky, yep, nothing ;-) just one of them has something to fix Nov 26 19:26:37 koen: did u see my previous post about qemu-native? Nov 26 19:27:20 cedricb: yes, the error message is abundantly clear Nov 26 19:27:50 qemu-*native* + "you need gcc3" error message Nov 26 19:28:38 but angstrom is gcc 4 Nov 26 19:28:46 or not? Nov 26 19:29:06 cedricb: qemu doesn't build with gcc 4.x Nov 26 19:29:39 ok so I need to build with gcc 3 and switch to angstrom? Nov 26 19:29:46 after Nov 26 19:31:09 cedricb: I suggest you to switch your native compiler just for qemu-native Nov 26 19:31:18 no need to switch Nov 26 19:31:36 'apt-get install gcc-3.4' will install it in parallel to gcc4 Nov 26 19:31:43 qemu will pick the right one Nov 26 19:32:04 koen: I had to ... at least on Gentoo I switchedwith gcc-config Nov 26 19:32:19 ~lart g3nt00 Nov 26 19:32:19 * ibot_ puts on a hockey mask and jumps out at g3nt00 Nov 26 19:32:50 koen: grrr ;) Nov 26 19:38:32 koen debian SUCKS! Nov 26 19:38:34 :) Nov 26 19:38:55 I like hrw's quote Nov 26 19:39:12 ? Nov 26 19:39:24 ubuntu: african for 'can't configure debian' Nov 26 19:39:31 koen hehe Nov 26 19:39:43 koen: LOL Nov 26 19:40:05 which reminds me of: http://xkcd.com/c178.html Nov 26 19:40:09 * koen hugs xkcd Nov 26 19:41:53 koen this is a good one ... I noticed that since 1996 all Linux distributions went from good to better and they are getting suckier (probably because they're run by Marketing drones now) Nov 26 20:10:41 re Nov 26 20:10:47 koen: its not mine Nov 26 20:10:50 its from #oe Nov 26 20:18:19 hrw: I got it from your sig :) Nov 26 20:18:36 I know Nov 26 20:19:28 06-06/#oe.log:07 20:02< woglinde> it's african word for "cant configure debian" Nov 26 20:19:28 06-06/#oe.log:08 10:40< hrw> it's african word for "can't configure debian" Nov 26 20:19:28 06-08/#oe.log:02 13:16< woglinde> ubuntu is an african word for "can't configure debian" *g* Nov 26 20:19:55 chouimat|bored: the rest of the dexter episodes is a bit better as the first one Nov 26 20:20:06 chouimat|bored: so if you get bored again :) Nov 26 20:20:20 07 20:02< woglinde> I heard a cool joke Nov 26 20:20:24 07 20:02< woglinde> whats mean ubuntu? Nov 26 20:20:29 07 20:02< woglinde> it's african word for "cant configure debian" Nov 26 20:20:36 first occuranc Nov 26 20:20:37 e Nov 26 20:23:51 hrw: so true =) Nov 26 20:32:28 koen hehe Nov 26 20:33:16 inkjets are slow Nov 26 20:34:40 compared to a laser printer, yes. But they are amazingly fast compared to a 24-needle printer =) Nov 26 20:34:59 koen was reading this ... http://www.popsci.com/popsci/flat/bown/2006/innovator_5.html Nov 26 20:42:19 question Nov 26 20:42:37 does it was decided where OE team will book hotels for fosdem? Nov 26 20:43:05 I just spent 318 pln on tickets so want to finally book hotels too Nov 26 20:43:16 NOTE: package linux-2.6.18+2.6.19-rc6-r0: task do_compile: started Nov 26 20:49:24 hrw: no idea Nov 26 20:49:29 no responses to my email Nov 26 20:49:31 mickey|tv: still around? Nov 26 20:58:16 aha Nov 26 20:58:56 bumped Nov 26 20:59:20 I've updated the STAGING_BINDIR proposal. Now I'm suggesting we need STAGING_BINDIR, STAGING_BINDIR_CROSS and STAGING_BINDIR_NATIVE Nov 26 21:05:43 RP: what do you think about packages/linux/linux_2.6.18+2.6.19-rc6.bb? Nov 26 21:06:18 hrw: Creating a new one or using something as a base? Nov 26 21:06:32 linux-x86 one Nov 26 21:06:38 simplest possible Nov 26 21:06:47 hrw: I'm ok with that as a standard one Nov 26 21:07:01 I use it for my progear Nov 26 21:07:22 It would nice to have it OZ comaptivle eventually :} Nov 26 21:07:59 RP: SRC_URI_zaurus_append =? Nov 26 21:08:25 hrw: Makes it tricky for people to upgrade the .bb :-/ Nov 26 21:08:25 koen: why angstrom does not use wpasupplicant 0.5.5? Nov 26 21:08:27 RP: yep Nov 26 21:08:47 RP: thats why I decided to create linux_ one Nov 26 21:08:54 RP: simple, no !mainline patches at all Nov 26 21:09:00 hrw: no idea Nov 26 21:09:04 one defconfig for machine (called defconfig) Nov 26 21:09:09 koen: DEF_PREF-2 Nov 26 21:09:19 hrw: Add progear support to the OZ kernel - it should just work with a defconfig :} Nov 26 21:09:55 RP: I know that it should. but from time to time people are asking about adding their kernels Nov 26 21:10:18 koen: any plans on moving angstrom as is to pcmciautils? Nov 26 21:10:35 hrw: everything should be using pcmcia-utils Nov 26 21:10:59 koen: so we have to track why progear got pcmcia-cs built at all Nov 26 21:11:07 hrw: gpe-conf Nov 26 21:11:16 koen: for bootstrap-image too? Nov 26 21:13:11 conf/bitbake.conf:PCMCIA_MANAGER ?= "pcmcia-cs" Nov 26 21:13:25 doesn't progear.conf select that? Nov 26 21:13:30 koen: no Nov 26 21:13:45 if the oldest kernel is 2.6.19 it shouldn't have the pcmcia-cs ioctl anymore Nov 26 21:13:49 koen: I think that pcmcia-cs/pcmciautils is rather distro machine Nov 26 21:19:18 cu - time for sleep Nov 26 21:19:42 hrw, what does BROKEN= 1 in a bbfile mean> Nov 26 21:20:09 that it's broken, but the person was too lazy to move it to packages/non-working Nov 26 21:20:16 thanks Nov 26 21:20:18 :) Nov 26 21:20:30 broken because they won't fetch? Nov 26 21:20:57 broken as 'I cant build it nevermind which reason' Nov 26 21:21:06 ah Nov 26 21:21:12 koen: I drop packages, not move to nonworking Nov 26 21:21:40 where would I find an example for using a username /pw with the svn fetcher? Nov 26 21:21:43 we have SCM Nov 26 21:21:53 Crofton|oe: svn://user:pass@svnserver? Nov 26 21:21:58 cu for now Nov 26 21:22:20 l8r Nov 26 21:24:35 03koen 07org.oe.dev * r3f62b62c... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom 2007.1: prefer cairo 1.3.4 (for fpu speed ups) and wpa-supplicant 0.5.5 (<0.5.x breaks with networkmanager) Nov 26 21:58:15 RP: I agree (STAGING_BINUTILS_*) Nov 26 21:58:36 binutils? Nov 26 21:58:39 hey mickeyl Nov 26 21:58:40 urghs Nov 26 21:58:42 i'd better go to bed Nov 26 21:58:46 STAGING_BINDIR_* even Nov 26 21:58:51 hi Nov 26 21:58:53 that would be my suggestion as well :) Nov 26 21:59:09 hi mickeyl Nov 26 21:59:15 i must stay awake until 0:00. have a phone call scheduled with taipeh Nov 26 21:59:20 *sigh* Nov 26 21:59:48 mickeyl: Do you agree we vetting all the existing entires, changing to the three different ones, then prepare a smaller patch to switchover at some future date? Nov 26 22:00:04 RP: i do. Nov 26 22:00:33 mickeyl: I'm tempted to proceed with it as I don't think its going to get many further comments on the mailing list :} Nov 26 22:00:40 bitesized upgrade steps are neat Nov 26 22:01:06 RP: yeah, looks like not a lot care about binconfig :) Nov 26 22:01:26 pkgconfig is the future Nov 26 22:01:30 mickeyl: binconfig is still needed but will perhaps change role slightly Nov 26 22:01:57 who volunteers for refreshing the glib patches for pkgconfig 0.2x? Nov 26 22:01:59 We will need to find some way of indicating the presence of -config cross files for packages staging... Nov 26 22:02:23 binconfig + packaged staging = headache :-( Nov 26 22:02:45 I'm still not convinced we need a seperate cross/ dir Nov 26 22:03:28 koen: We do need the cross binaries in a different dir or at least symlinked from a dir in path Nov 26 22:04:22 We could install them into STAGING_BINDIR and symlink to STAGING_BINDIR_CROSS but in theory the installed versions should have different paths so I don't like this idea Nov 26 22:06:33 right Nov 26 22:06:52 the symlink could be provided by the foo-symlinks package Nov 26 22:07:49 03koen 07org.oe.dev * r2cc7d534... 10/ (1 packages/gcc/gcc-cross-sdk_4.1.1.bb): gcc-cross-sdk: change include order to make it fail in a different place Nov 26 22:08:24 anyway Nov 26 22:08:24 time to sleep Nov 26 22:18:27 mickeyl: With these changes I'm optimistic that we can lose a lot of these "export XMMS_CONFIG="${STAGING_BINDIR}/xmms-config" style variables entirely which would be really nice to do :) Nov 26 22:19:21 RP: i agree. that was a bit ugly and non-deterministic all the time Nov 26 22:19:55 mickeyl: ~100 .bb's left to review :} Nov 26 22:20:14 *cough Nov 26 22:36:10 Hello; I just noticed that the most recent angstrom gpe image for h5000 builds cups as well... I know there are bluetooth-enabled printers, but is it really neccessary for gpe image? :) Nov 26 22:46:52 Is there a way to specify a niceness level with which to run OE/bitbake? Nov 26 22:52:48 you mean other than "nice -n bitbake " ? Nov 26 22:53:11 * mickeyl going to bed Nov 26 22:53:13 g'night guys Nov 26 22:53:21 'night mickey|zzZZzz Nov 26 23:08:46 Nov 26 23:12:30 yes, I meant not having to issue "nice -n $number bitbake $pkg" but just "bitbake $pkg" and bitbake got the niceness level from some config file. Nov 26 23:12:38 But maybe this is not possible. Nov 26 23:12:43 yet Nov 26 23:14:56 Laibsch: Have it handle BB_NICE_LEVEL and submit a patch? :) Nov 26 23:15:49 sort of that thing. ;-) Nov 26 23:17:25 are those huge numbers of different variables documented anywhere? :) Nov 26 23:18:06 Jin^eLD: The one above doesn't exist yet Nov 26 23:18:46 yeah I know that :) I meant the ones that do exist Nov 26 23:19:22 Jin^eLD: There is the OE users manual under development. There is also the bitbake manual for the actual core bitbake functionality Nov 26 23:19:41 is it the one under "Documentation" on the OE website? Nov 26 23:20:02 a reference list would be kind of cool Nov 26 23:20:40 Jin^eLD: The OE manual is the one there, yes Nov 26 23:21:06 Jin^eLD: Some variables are documented in conf/documentation.conf as well Nov 26 23:21:22 aah, documentation.conf! I guess I missed that one, thanks Nov 27 00:19:55 koen|away, http://likerabbits.blogspot.com/2006/11/developers-projects-plans.html Nov 27 00:25:34 good night Nov 27 00:39:20 'night all **** ENDING LOGGING AT Mon Nov 27 02:59:57 2006