**** BEGIN LOGGING AT Thu Jul 31 02:59:57 2008 Jul 31 03:38:03 I got iMate PPC,. HTC Himalaya,. what is the first step of deploying the OE to my machine? Jul 31 06:46:06 morning Jul 31 06:46:39 morning Jul 31 07:08:43 does someone have news from woglinde? Jul 31 07:08:47 oops hello Jul 31 07:10:49 03  07org.openembedded.dev * rc1b49c502a 10OE.dev/packages/swfdec/ (swfdec-mozilla_0.7.4.bb swfdec_0.7.4.bb): swfdec: update to 0.7.4 Jul 31 07:10:49 03  07org.openembedded.dev * r3e630fb7e7 10OE.dev/packages/ffmpeg/ffmpeg_git.bb: ffmpeg git: bump SRCREV Jul 31 07:10:49 03  07org.openembedded.dev * r5f061c2dc1 10OE.dev/packages/ffmpeg/omapfbplay_git.bb: omapfbplay: build shared version from now on Jul 31 07:10:56 03  07master * rc1b49c502a 10OE.dev/packages/swfdec/ (swfdec-mozilla_0.7.4.bb swfdec_0.7.4.bb): swfdec: update to 0.7.4 Jul 31 07:10:57 03  07master * r3e630fb7e7 10OE.dev/packages/ffmpeg/ffmpeg_git.bb: ffmpeg git: bump SRCREV Jul 31 07:10:57 03  07master * r5f061c2dc1 10OE.dev/packages/ffmpeg/omapfbplay_git.bb: omapfbplay: build shared version from now on Jul 31 07:20:37 good morning Jul 31 07:35:16 03  07org.openembedded.dev * rb2d13d9485 10OE.dev/packages/linux/ (4 files in 2 dirs): linux-omap2: add patches to fix serial hang Jul 31 07:35:30 03  07master * rb2d13d9485 10OE.dev/packages/linux/ (4 files in 2 dirs): linux-omap2: add patches to fix serial hang Jul 31 07:36:47 anybody there ? Jul 31 07:37:34 nope Jul 31 07:37:35 Unclejack, i'm there but i'm just a user Jul 31 07:38:36 Yep, I've got a problem during the building process of FSO Jul 31 07:38:42 any help is welcome Jul 31 07:45:06 Unclejack: so, er, what is your problem? Jul 31 07:45:12 morning Jul 31 07:57:57 morning all Jul 31 07:59:54 ok Jul 31 07:59:59 morning Jul 31 08:00:03 so Jul 31 08:00:55 I get ERROR: log data follows (/mnt/array/francesco/FSO/build/tmp/work/armv4t-angstrom-linux-gnueabi/libsdl-x11-1.2.11-r0/temp/log.do_stage.24166)| oe_libinstall: cd src| oe_libinstall: install -d /mnt/array/francesco/FSO/build/tmp/staging/arm-angstrom-linux-gnueabi/usr/lib/ | FATAL: oe_libinstall: unable to locate shared library Jul 31 08:01:13 while trying to build FSO image on OpenEmbedded Jul 31 08:02:47 sounds like libsdl.so hasn't been built for some reason Jul 31 08:03:07 log file does not help Jul 31 08:03:13 do you actually have a .so file in the libsdl build directory? Jul 31 08:03:21 wait Jul 31 08:05:48 apparently not Jul 31 08:06:05 okay, so you need to check the logs to find out why not Jul 31 08:07:16 log.do_stage.24166 reports just the error I've pasted b4 Jul 31 08:08:15 morning all Jul 31 08:09:44 any other logs where to look ? Jul 31 08:12:12 you need to look at do_configure and/or do_compile Jul 31 08:12:21 hi rp Jul 31 08:21:36 pb_: hey Jul 31 08:21:43 hi RP Jul 31 08:25:10 pb_: I digged down a little into who is asking for libgcc in the boot process and figured out that its the logger applet from busybox which is invoked during mountall that causes the message first "libgcc_s.so.1 must be installed for pthread_cancel to work" Jul 31 08:34:54 hi khem Jul 31 08:35:01 ah, right. can you tell what logger is doing to cause that message? Jul 31 08:36:15 khem: on another subject, some package seems to be installing a bogus libc.so into staging during the bootstrap process. do you have any idea what this is? Jul 31 08:36:29 it's a pretty small file: Jul 31 08:36:29 rwxr-xr-x 1 pb pb 1637 2008-07-30 21:31 /home/pb/oe/build-fso2/tmp/staging/arm-angstrom-linux-gnueabi/usr/lib/libc.so Jul 31 08:36:30 ok, libsdl relies on libSDL-1.2.so; that file is present but under image/usr/libs ... oe_libinstall opens "src" dir finding nothing, I don't know why Jul 31 08:37:11 Unclejack: sounds like it is (wrongly) being moved rather than copied into the image directory Jul 31 08:37:14 probably by do_install Jul 31 08:37:22 pb_: there is a dummy libc.so installed by glibc-initial Jul 31 08:37:32 khem: ah. it is broken. Jul 31 08:38:06 specifically, it doesn't have a SONAME, which means that anybody linking with it will generate a broken binary Jul 31 08:38:06 pb_: but it should be later overwritten by glibc Jul 31 08:38:25 yeah, it is, but its presence is causing the glibc build to go wrong Jul 31 08:38:31 what is the purpose of this dummy libc.so? Jul 31 08:38:44 pb_: but it is short lived only for bootstrap purpose it should not exist after glibc is built Jul 31 08:39:10 why is it needed at all? I don't think we used to have it. Jul 31 08:39:17 to generate proper libgcc and shared-lib support in cross-gcc-intermediate Jul 31 08:39:54 can you expand on that? what precisely does it do? Jul 31 08:40:07 if it doesn't have the right soname then I would have thought it would be doing no good for anything. Jul 31 08:40:41 pb_: I dont understand the problem that it it causing. Jul 31 08:40:54 bonjour Jul 31 08:41:07 pb_: is this the cause of the ISOXXXXX.so problem? Jul 31 08:42:00 XorA: yes Jul 31 08:42:24 khem: never mind the problem for the moment, can you explain exactly what its purpose is? Jul 31 08:42:42 it seems fundamentally wrong to install something as libc.so that doesn't expose the same abi as the real libc.so Jul 31 08:43:09 anybody who links against that is going to be disappointed when the real libc doesn't match the one they linked with. Jul 31 08:43:30 * XorA wonders sometimes how pb_ holds all this genius in his cranium Jul 31 08:43:56 pb_: It is installed when gcc-intermediate it built which generates shared libgcc and can compile with --enable-shared Jul 31 08:44:46 khem: okay, so it is basically serving the purpose that used to be done by glibc-intermediate? Jul 31 08:45:12 pb_: I can see a problem if you have compiling glibc because now glibc considers this as libc.so in that case it needs to be removed just after cross-gcc-intermediate is finished building Jul 31 08:45:21 pb_: yes Jul 31 08:45:34 why is it better than the old glibc-intermediate arrangement? Jul 31 08:45:41 pb_, btw, thanks for all the help yesterday, had to run :-) Jul 31 08:45:49 at least with glibc-intermediate, the installed libc.so did have the correct abi Jul 31 08:45:52 (was fighting with rrdtool) Jul 31 08:45:57 czr_: sure, no problem. glad you got it working. Jul 31 08:46:24 pb_: I thought that the build sequence became consistent Jul 31 08:46:30 khem: yes, removing it after gcc-cross-intermediate is built would solve the particular glibc problem at hand, but I am a bit concerned that it will just be storing up more trouble for the future. Jul 31 08:46:31 yup. I'm wondering now whether anyone actually uses rrdtool via oe since it felt quite impossible to get working with cross Jul 31 08:46:49 earlier we had different build sequences for uclibc nptl and linuxthreads Jul 31 08:47:08 khem: it doesn't seem that linking with this dummy libc.so will produce any useful result. you might get a binary out, but it will be wrong. Jul 31 08:47:29 khem: well, yes, but we could just have made all of those use glibc-intermediate. Jul 31 08:47:50 the only reason it was different before is that linuxthreads didn't need the extra step, and there seemed no point in wasting time by building an extra pass of glibc just to make the process consistent. Jul 31 08:49:04 * pb_ -> office now Jul 31 08:49:05 bbiab Jul 31 08:49:33 pb_: I had seen issues when rebuilding uclibc produced different libgcc with previous sequence Jul 31 08:50:07 pb_: my email to the ml explains the problem I saw Jul 31 08:50:52 any hints how to build an ipkg-package native gcc for the target? (so that I could install it on the target). Jul 31 08:54:21 czr_: bitbake gcc? Jul 31 09:03:00 hrw, that easy? heh. /me goes away to cry in shame Jul 31 09:04:02 * XorA chuckles Jul 31 09:04:08 dont cry, bitbake Jul 31 09:04:53 glibc-dev would be the next package to build, right? Jul 31 09:04:58 or just glibc? Jul 31 09:08:19 khem: can you send me a link to that mail? Jul 31 09:08:26 or just forward a copy to me directly Jul 31 09:13:26 hmm. for some reason bitbake glibc doesn't result in glibc-dev package. I get a lot of other related packages, but not -dev. any hints? (.stable) Jul 31 09:19:36 frikker: check that the postinst script created/the symlinks were included in the tarball for liboctinterp.so -> liboctinterp.so.x(.x) Jul 31 09:19:53 czr_: are you sure it hasn't just been renamed to something else, e.g. libc6-dev? Jul 31 09:20:53 pb__, indeed! libc6-dev Jul 31 09:21:00 thanks :-) /me continues crying in shame Jul 31 09:21:08 heh Jul 31 09:28:52 hmh. gcc = no go: gcc: error trying to exec 'cc1': execvp: No such file or directory Jul 31 09:29:36 ah. missing cpp :-) Jul 31 09:42:42 that's odd, gcc should depend on cpp. Jul 31 09:44:12 hey folks, anybody has tried recently to build FSO image for GTA01 ? Jul 31 09:44:36 probably should depend on gcc-symlinks as well. but doesn't Jul 31 09:44:43 binutils is also not pulled in automatically Jul 31 09:44:55 but that's ok, it works now. thanks pb_ and hrw :-) Jul 31 09:49:07 czr_: bitbake task-native-sdk Jul 31 09:49:16 czr_: and then 'opkg install task-native-sdk' Jul 31 09:50:29 hrw, is there a package list of that somewhere? Jul 31 09:52:34 czr_: "find packages -name *.bb" etc? Jul 31 09:52:35 03  07org.openembedded.dev * rbab8c00529 10OE.dev/packages/dsplink/dsplink_1.50.bb: dsplink: bump PR for packaging change Jul 31 09:52:41 03  07master * rbab8c00529 10OE.dev/packages/dsplink/dsplink_1.50.bb: dsplink: bump PR for packaging change Jul 31 09:52:46 hrw, everything?? :-) Jul 31 09:52:59 I meant which packages are pulled by task-native-sdk :-) Jul 31 09:53:07 czr_: read recipe Jul 31 09:55:29 morning Jul 31 09:56:12 hi ant Jul 31 09:58:08 hi hrw, do you know perhaps if Poky has gitweb I can browse from here? Jul 31 09:58:20 ant|work: Poky use svn Jul 31 09:58:33 ant|work: http://svn.o-hand.com/ and look there Jul 31 09:58:37 I see, so no changelog... Jul 31 09:58:53 ant|work: you want changelog? Jul 31 09:58:56 I'm waiting to see your c7x0 patches (udev, kernel, snd) Jul 31 09:59:16 thx for reminder Jul 31 09:59:27 you know, http://gitweb.openembedded.net/?p=org.openembedded.dev.git;a=shortlog Jul 31 09:59:35 is very useful when I'm in office Jul 31 10:01:34 ant|work: http://svn.o-hand.com/view/poky/trunk/ Jul 31 10:02:10 ok, perfect Jul 31 10:05:56 hrw: what do you think about thesing work for booting mmcblckp1? Would we need these mods for c7x0 too I suppose...http://gitweb.openembedded.net/?p=org.openembedded.dev.git;a=commit;h=0bc0a3e646e5d0bf38c3dfe66903d45cf005c139 Jul 31 10:06:22 http://gitweb.openembedded.net/?p=org.openembedded.dev.git;a=commit;h=0bc0a3e646e5d0bf38c3dfe66903d45cf005c139 Jul 31 10:06:29 wtf? Jul 31 10:08:09 in particular 'files/device_table_add-mmc.txt' Jul 31 10:09:30 ant|work: it depends which way of booting you want to go on c7x0 Jul 31 10:09:44 ant|work: I still use standard sharp way Jul 31 10:09:53 hrw: but then we boot initramfs-bootmenu-image Jul 31 10:09:59 (soon) Jul 31 10:10:01 it works fine and 2.6.26 without oabi support still fit Jul 31 10:10:07 ok Jul 31 10:10:33 thesing already made kernel in /boot Jul 31 10:10:36 ant|work: c7x0 are fine to boot from flash for many users Jul 31 10:10:39 I know Jul 31 10:11:09 this russian guy (Jay) promised to finish the C menus for the image, after 4 aug Jul 31 10:11:25 let see...and hope they don't drink too much in Moskow :-) Jul 31 10:11:50 ant|work: Russians tolerate more alcohol then rest of world Jul 31 10:12:18 ah, not Poland ? Jul 31 10:12:23 ;-) Jul 31 10:12:46 ant|work: we do not have so strong winters as they Jul 31 10:13:11 yes, I was surprised by the spring/summer in Poland Jul 31 10:13:34 but not by the winter Jul 31 10:13:52 it's a bit colder than in South Europe, but not like in Rusiia ! Jul 31 10:15:01 (BTW at -19 C° the Zaurus was working nicely) Jul 31 10:17:50 ant|work: sound fix merged Jul 31 10:18:11 thxxx Jul 31 10:18:15 my first use of mtn in July... Jul 31 10:18:31 last July Jul 31 10:19:10 soon it would be "git format-patch -n -o to-be-merged-with-oe" + vim + "git am" Jul 31 10:29:46 03  07org.openembedded.dev * rb7c5b255d6 10OE.dev/: Jul 31 10:29:46 merge of '5f80d82b77d53ef77154af1dcd9298bf7c5f389d' Jul 31 10:29:46 and 'e10700938896fb54483d9c1a8c452585cdd764be' Jul 31 10:29:51 03  07org.openembedded.dev * r6351a8b953 10OE.dev/packages/linux/ (linux-rp-2.6.26/zaurus-i2c-init.patch linux-rp_2.6.26.bb): linux-rp: fix sound on c7x0 in 2.6.26 kernel Jul 31 10:29:51 03  07master * rb7c5b255d6 10OE.dev/: Jul 31 10:29:51 merge of '5f80d82b77d53ef77154af1dcd9298bf7c5f389d' Jul 31 10:29:52 and 'e10700938896fb54483d9c1a8c452585cdd764be' Jul 31 10:29:54 03  07master * r6351a8b953 10OE.dev/packages/linux/ (linux-rp-2.6.26/zaurus-i2c-init.patch linux-rp_2.6.26.bb): linux-rp: fix sound on c7x0 in 2.6.26 kernel Jul 31 10:31:59 morgen woglinde Jul 31 10:32:11 hi ant Jul 31 10:32:41 yesterday I got heavy storms in Germany! Jul 31 10:32:51 (Frankfurt - Hahn) Jul 31 10:32:56 hm Jul 31 10:33:07 we didnt even have rain Jul 31 10:33:14 had Jul 31 10:34:11 I think it was local, but for sure till Frankfurt-Flughafen (slot were closed too) Jul 31 10:34:30 Germany is getting tropical :-) Jul 31 10:35:45 maybee Jul 31 10:36:10 hi woglinde Jul 31 10:36:16 hi all Jul 31 10:36:18 hi pb :) Jul 31 10:36:22 hi florian Jul 31 10:36:35 pb is your bootime now faster? Jul 31 10:37:59 hi florian Jul 31 10:38:06 woglinde: not significantly, I got distracted by other projects Jul 31 10:38:17 one of my colleagues is working on boot time today Jul 31 10:38:19 pb oh Jul 31 10:38:48 but we have some ideas now :-) Jul 31 10:39:07 would be nice when you would share the ideas which worked Jul 31 10:39:12 about boot-time, I'm thinking about http://roy.marples.name/openrc for my workstation Jul 31 10:39:17 yah, will do Jul 31 10:39:24 is Gentoo but could be ported Jul 31 10:39:56 http://www.gentoo.org/doc/en/openrc-migration.xml Jul 31 10:40:01 hm I favor minit Jul 31 10:40:07 but didnt try it sofar Jul 31 10:40:16 but the ideas seems right to me Jul 31 10:40:23 they'll use it for embedded-gentoo Jul 31 10:40:25 * pb__ favours no init at all :-} Jul 31 10:40:38 if only our application didn't keep crashing, we could do without this supervising intelligence Jul 31 10:40:38 apparently Jul 31 10:40:55 right now, we plan to write a custom mini-init that just spawns our app, waits for it to crash, then spawns it again Jul 31 10:41:02 pb hm then fix the crashes Jul 31 10:41:05 *g* Jul 31 10:41:10 woglinde: ah, good plan Jul 31 10:41:11 didn't think of that Jul 31 10:41:19 pb sorry Jul 31 10:41:23 :-) Jul 31 10:41:26 I didnt want to bully you Jul 31 10:41:58 ant ??? Jul 31 10:41:59 It's fast - 29 seconds to boot my laptop running Gentoo/Linux Jul 31 10:42:05 29 seconds isnt fast Jul 31 10:42:12 read what tehy start for services.. Jul 31 10:42:33 http://roy.marples.name/openrc/bootchart-20080122.png Jul 31 10:42:41 yes Jul 31 10:42:44 I am looking at it Jul 31 10:43:12 why he is starting xdm and gdm Jul 31 10:43:18 this looks wrong Jul 31 10:43:59 the point is it's written in C (no bash) Jul 31 10:44:32 ant but you have udev which is shell scripts again Jul 31 10:44:52 oh, not in assembler? Jul 31 10:45:02 the eleet gentoo dudes are losing their edge :-} Jul 31 10:46:05 pb__: whic libc would it need? Jul 31 10:46:09 seriously though, it does still seem to be doing quite a lot of shell stuff Jul 31 10:46:20 (about your devasting issue) Jul 31 10:46:22 do you know what this "runscript.sh" thing is? Jul 31 10:46:39 ant|work: sorry, you lost me. what are we talking about? Jul 31 10:46:51 about these initscripts Jul 31 10:47:01 openrc Jul 31 10:48:05 ant|work: I didn't understand what you meant by "devasting issue" Jul 31 10:48:16 libc without SONAME Jul 31 10:48:22 read earlier Jul 31 10:49:02 oh, right, the glibc-intermediate thing Jul 31 10:49:25 yeah, I kind of feel that the right approach is to reinstate glibc-intermediate and fix up whatever problem khem was trying to solve by some other means. Jul 31 10:49:38 I don't think that installing a fake libc.so with some made up abi is a good idea. Jul 31 10:49:47 in Gentoo they simply decided to provide a compat lib libstdc++ when they changed abi Jul 31 10:49:55 right Jul 31 10:49:56 libstdc++v3 Jul 31 10:50:11 yeah, this is the correct way to handle an abi change Jul 31 10:50:26 unless it is small enough that you can do it with symbol versioning Jul 31 10:50:47 it seems to work here ...(sorry I don't know the internals as you do !) Jul 31 10:51:06 I'm just an old sysadm Jul 31 10:51:30 ant old? Jul 31 10:51:43 what was you first unix? Jul 31 10:51:55 he, xenix Jul 31 10:52:23 hm let me look on the time line Jul 31 10:52:25 was IBM AT at the time Jul 31 10:52:47 my first as admin was irix 5.x Jul 31 10:52:56 around 96 Jul 31 10:53:10 with the great 4dmw Jul 31 10:53:22 uh, no, her it was mid 80's Jul 31 10:53:49 256kb ram, 20mb hd Jul 31 10:53:56 my first unix was xenix in 93. but it was for fun not use Jul 31 10:54:04 informix sql was a bomb! Jul 31 10:54:15 (on i386) Jul 31 10:54:17 bomb? Jul 31 10:54:31 was rather fast, yes Jul 31 10:55:16 *g* Jul 31 10:55:26 today your cellphone is faster Jul 31 10:55:29 I remember the sysadm shell of xenix Jul 31 10:55:45 Midnight/Norton Commander are sooo similar.... Jul 31 10:55:53 he he Jul 31 10:57:25 you mean peter norton stole some stuff? Jul 31 10:57:32 or rhe xenix guys? Jul 31 10:58:11 IIRC the sysadm ASCII shell (ncurses-like) was precedent Jul 31 10:59:04 IIRC was called 'sysadmsh' Jul 31 10:59:08 hm Jul 31 10:59:20 seems I have to sing again for my sun Jul 31 10:59:27 so he takes his midsleep Jul 31 11:00:45 weather in 'realtime': http://www.ruwenzori.net/earth/Earth_EMEA+Atlantic_1280x1024.html Jul 31 11:01:18 woglinde: no sun today and tomorrow ;-( Jul 31 11:10:14 hm full diapaper Jul 31 11:10:29 thanks for the bulletin Jul 31 12:01:03 morning all Jul 31 12:15:47 slightly OT, any ideas on how to implement low_latency serial option into a platform serial driver? (pxa-serial in this case). the 32-byte FIFOs are killing latency atm. Jul 31 12:19:47 czr_: I guess the easiest way would be to define a new c_cflag bit in termios for "no buffering", call it CUNBUFFERED or some such, and implement it in your serial_pxa_set_termios() function. Jul 31 12:20:09 afaik, there isn't a pre-existing mechanism to turn the fifos on and off, though I don't know for sure Jul 31 12:20:13 * pb__ lunchtime now Jul 31 12:20:15 bbl Jul 31 12:20:23 indeed. however, the other option seems to be implementing something that is invoked via "low_latency" in setserial Jul 31 12:20:54 description implies that it trades CPU cycles for shorter latency, which implies (to me) at least tweaking the fifo settings or disabling them completely Jul 31 12:21:15 but I'll add this to longer term todo, I guess I'll live with current latencies for a while. Jul 31 13:13:25 woglinde, with the work you have committed i was able to build wesnoth: http://pastebin.com/m9c09c5c but i didn't try it yet...(openmoko gta01 in qemu has the wrong keyboard so i can't have the network...previously i tried changing the keyboard in task-openmoko-base.bb but it didn't change it in the image...) Jul 31 13:13:30 oops hello Jul 31 13:36:44 woglinde: miracle! the clouds are disappearing! Jul 31 13:57:30 03  07org.openembedded.dev * r5faa6bad6a 10OE.dev/packages/dsplink/ (codec-engine_2.10.bb files/ticel-config): codec-engine: stop installing ticel-config Jul 31 13:57:45 03  07master * r5faa6bad6a 10OE.dev/packages/dsplink/ (codec-engine_2.10.bb files/ticel-config): codec-engine: stop installing ticel-config Jul 31 14:03:45 !oebug 1613 Jul 31 14:03:46 * * Bug 1613, Status: NEW, Created: 2006-11-23 09:14 Jul 31 14:03:47 * * ronan(AT)aimao.org: New recipe - mplayerplug-in_3.31.bb Jul 31 14:03:48 * * http://bugs.openembedded.net/show_bug.cgi?id=1613 Jul 31 14:20:06 * * OE Bug 4461 has been RESOLVED (FIXED) by dp(AT)xora.org.uk Jul 31 14:20:08 * * Add machine htcpolaris Jul 31 14:20:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4461 Jul 31 14:22:08 03  07org.openembedded.dev * r606f7f8dac 10OE.dev/ (2 files in 2 dirs): htcpolaris : add machine definition and expand kiaser kernel to work for this Jul 31 14:22:16 03  07master * r606f7f8dac 10OE.dev/ (2 files in 2 dirs): htcpolaris : add machine definition and expand kiaser kernel to work for this Jul 31 14:24:31 hrw: can you imagine what's causing 'oss_audio: failed to open audio device /dev/dsp' with 2.6.24? Jul 31 14:25:21 lack of /dev/dsp? Jul 31 14:25:30 no, there is Jul 31 14:25:45 and can you open it? Jul 31 14:26:00 I tried flite_time Jul 31 14:26:09 it was working previously Jul 31 14:26:51 it's about navit - gps program, which depends from flite and speech-dispatcher Jul 31 14:26:52 woglinde fixed some Jul 31 14:27:44 so, if you try to open the device by hand, what happens? Jul 31 14:32:08 pb__: sorry I don't have the device here Jul 31 14:34:07 pb_: I even chmodded 755 /dev/dsp Jul 31 14:35:47 ant|work: fwiw, you generally want /dev/dsp to have mode 600, or perhaps 660 Jul 31 14:36:06 but, anyway, I doubt that is relevant to the problem you're having Jul 31 14:36:31 it's hard to make any meaningful comment without knowing what the actual error is, unfortunately Jul 31 14:36:59 yeah, I was hoping some missing i2c patch for 2.6.24 perhaps Jul 31 14:37:10 like the one hrw added for 2.6.26 Jul 31 14:37:34 I'll have to test thoroughly later on Jul 31 14:37:50 ant|work: that i2c was needed to get sound *at all* Jul 31 14:37:50 okay Jul 31 14:37:57 RP: I want to commit sth. like http://www.pastebin.ca/1088466 should I leave it as is it or should I only make the RRECOMMENS line collie specific? Jul 31 14:38:12 hrw: I suspectI don't have sound at all! Jul 31 14:39:26 thesing: If doesn't need to be collie specific, just add it for all machines Jul 31 14:39:30 hrw: cat /proc/asound/card0/id -> Corgi Jul 31 14:39:38 hrw: cat /proc/asound/card0/oss_mixer -> all values are zero Jul 31 14:39:49 this was on 29th july on irc Jul 31 14:39:51 ant|work: on unpatched 2.6.26 there was no cards at all Jul 31 14:40:03 RP: the RRECOMMENDS line too? Jul 31 14:40:16 hrw: ok, good to know it's not totally broken Jul 31 14:40:21 thesing: Yes, its not like its a big file Jul 31 14:40:42 true. Jul 31 14:44:19 ant|work: use aplay to test do you have sound or not Jul 31 14:46:57 ant|work: and switch flite to using alsa (if it is possible) Jul 31 14:48:09 hrw: now I fear we enabled oss for flite.... Jul 31 14:48:21 there is flite-alsa Jul 31 14:58:10 hi all Jul 31 14:58:18 i need some simple help with my Bitbake recepie Jul 31 14:58:23 i have this:${CC} ${CFLAGS} ${LDFLAGS} -o hello hello.c Jul 31 14:58:28 how do i add the GTK libraries and headers to this ? Jul 31 15:06:05 anyone here? Jul 31 15:06:33 fgcity: Lots of people. Be patient Jul 31 15:06:51 ok Jul 31 15:09:41 fgcity: just like you would normally do it, with pkg-config. Jul 31 15:10:09 how would this look in my bitbake recepie? Jul 31 15:10:50 on my machine it looks like this:`pkg-config --cflags --libs gtk+-2.0` Jul 31 15:11:00 but how would it look in the bitbake recepie? Jul 31 15:11:04 and same will be in OE Jul 31 15:11:15 (i'm compiling for a different machine) Jul 31 15:11:23 so what? OE handle that Jul 31 15:11:26 http://sourceforge.net/community/forum/topic.php?id=2995 Jul 31 15:11:30 add pkg-config into makefile Jul 31 15:11:41 i don't have a make file Jul 31 15:11:48 i'm compiling without it Jul 31 15:12:36 to compile a simple file i have :${CC} ${CFLAGS} ${LDFLAGS} -o hello hello.c Jul 31 15:12:50 so to compile including the GTK libs how would i di it ? Jul 31 15:13:11 good morning Jul 31 15:13:31 fgcity: ${CC} ${CFLAGS} ${LDFLAGS} `pkg-config --cflags --libs gtk+-2.0` -o hello hello.c Jul 31 15:13:39 there really isn't any more to it than that Jul 31 15:13:46 khem: good morning Jul 31 15:14:01 tnx Jul 31 15:14:03 i'll try Jul 31 15:14:34 fgcity: Make sure you have the proper #include's in your hello.c of course Jul 31 15:14:44 pb__: logger from busybox uses pthread mutexes, I am suspecting that is calling unwinding code Jul 31 15:14:48 ok Jul 31 15:14:50 it worked Jul 31 15:14:53 tnx allot Jul 31 15:14:57 i really appriciate this Jul 31 15:15:25 re Jul 31 15:15:58 khem: hm, there is no sign of mutexes in my copy of logger.c. Jul 31 15:16:01 what version are you looking at? Jul 31 15:16:12 hi khem Jul 31 15:16:32 I can't think offhand why the mutex code should have any business calling the unwinder. Jul 31 15:16:46 maybe you could set a breakpoint on the unwind function and get a backtrace to find the call chain. Jul 31 15:17:23 it uses libpwdgrp Jul 31 15:17:30 which has some locking code Jul 31 15:17:30 fwiw, I tried "readelf -a busybox" | grep pthread on my machine and there was no output Jul 31 15:17:33 ah Jul 31 15:18:22 I think its some indirect call that forces glibc to go through that path where it needs to dlopen libgcc I guess Jul 31 15:18:32 yeah, could be. Jul 31 15:18:38 like I say, the debugger would be the easiest way to identify it Jul 31 15:18:38 because it only happens for logger and not for sh Jul 31 15:18:55 yeah I will do that now Jul 31 15:19:09 ok, great Jul 31 15:19:15 woglinde: guten morgen, germany got rain news is tellin me :) Jul 31 15:19:16 I guess sh might not be using any of the libpwdgrp bits Jul 31 15:19:46 pb__: I want to understand the problem that dummy libc.so is creating. Jul 31 15:20:08 I am so far failing to understand the problem Jul 31 15:20:10 khem: any binary linked against it ends up with a bogus DT_NEEDED entry for "libc.so" Jul 31 15:20:34 which, obviously, doesn't exist on the target system Jul 31 15:20:47 pb__: I see but there should be a text version of /usr/lib/libc.so installed by glibc bb Jul 31 15:20:54 this is breaking glibc, specifically, because things like ISO8859-1.so end up with this brokenness and hence won't load Jul 31 15:21:15 I see Jul 31 15:21:16 khem: that is in the -dev package, and in any case it is not a valid dynamic object Jul 31 15:21:24 the text file won't help to resolve a DT_NEEDED Jul 31 15:21:43 but thats how it is even on other distros Jul 31 15:21:57 look at your desktop it should have /usr/lib/libc.so Jul 31 15:21:57 right, but other distros don't have a DT_NEEDED for libc.so in the first place Jul 31 15:22:08 hi khem Jul 31 15:22:15 HopsNBarley: hey Jul 31 15:22:22 sure, I understand that, but other distros do not link binaries against a broken dummy libc.so, and hence their binaries do not have the bogus dynamic record in them. Jul 31 15:22:31 pb__: thats because linker reads libc.so as linker script Jul 31 15:23:00 khem: well, it's more because other distros link against the real libc, not some bogus replacement. Jul 31 15:23:01 if glibc is broken that I can understand but no other package should be affected Jul 31 15:23:30 its easy to fix infact I forgot to do that fix Jul 31 15:23:31 any package that's built before glibc is staged might be affected. Jul 31 15:23:38 hm any python expert here Jul 31 15:23:44 I want to achieve Jul 31 15:23:45 ${@['','cp ${STAGING_DATADIR}/gettext/config.rpath ${S}'][bb.data.getVar( Jul 31 15:23:46 'PREFERRED_VERSION_libtool',d,1) => '2.2.4']} Jul 31 15:23:46 it should be removed just after gcc phase 2 is done Jul 31 15:24:01 right, but that will still leave you with potentially broken binaries in gcc itself Jul 31 15:24:03 but => dont seems right Jul 31 15:24:35 woglinde: there's a function in bitbake for doing version comparisons Jul 31 15:24:38 bb.vercmp() or something Jul 31 15:24:49 pb__: no it wont as fas as I have seen it only links with libc.so in configure Jul 31 15:25:05 pb__: tests Jul 31 15:25:07 ah, I see Jul 31 15:25:12 pb hm Jul 31 15:25:16 wouldn't it be easier just to patch the tests in that case? Jul 31 15:25:28 mickeyl wakeup! Jul 31 15:25:30 *g* Jul 31 15:25:35 but, if it's genuinely only needed in configure then I agree it should be fairly harmless Jul 31 15:25:45 maybe you could remove it at the end of do_configure() to ensure it doesn't affect the generated binaries Jul 31 15:25:50 pb__: I must look into that possibility but then these patching will carry on for real cross gcc Jul 31 15:26:30 pb__: yeah I will check it Jul 31 15:26:44 ok, great Jul 31 15:27:08 meanwhile I will prepare a fix to remove this before glibc builds Jul 31 15:27:09 woglinde: but, anyway, I think the python operator you are trying to use is ">=", not "=>" Jul 31 15:27:26 thanks Jul 31 15:27:47 pb yes Jul 31 15:27:51 I searched this Jul 31 15:27:57 but it isnt right Jul 31 15:28:00 in this case Jul 31 15:28:21 'PREFERRED_VERSION_libtool',d,1) >= '2.2.4']}: bad substitution Jul 31 15:29:08 that doesn't sound like a python error Jul 31 15:29:14 I think you are accidentally feeding your python to the shell Jul 31 15:29:45 hm Jul 31 15:29:47 okay Jul 31 15:29:52 this might bee Jul 31 15:30:00 I will try to declare it as python Jul 31 15:30:23 "bad substitution" is what bash will say if it sees a $-expression it doesn't understand. Jul 31 15:30:33 try typing something like "echo ${%}" for example Jul 31 15:30:46 pb I include it with require Jul 31 15:30:56 python errors are always much longer than that and involve multiple pages of backtrace :-) Jul 31 15:30:58 because its needed by more recipes Jul 31 15:31:02 yes Jul 31 15:31:03 sorry Jul 31 15:33:06 * * OE Bug 4462 has been created by stian(AT)nixia.no Jul 31 15:33:09 * * alsa-utils_0.1.15 fails to compile due to NLS problem (LC_ALL not defined) Jul 31 15:33:11 * * http://bugs.openembedded.net/show_bug.cgi?id=4462 Jul 31 15:33:29 damn alsa Jul 31 15:34:44 args Jul 31 15:35:03 pb is the damn cut and paste with linebreak error Jul 31 15:35:11 woglinde: doh! Jul 31 15:36:01 yes Jul 31 15:36:05 now it works Jul 31 15:36:10 very good Jul 31 15:36:12 thanks anyway Jul 31 15:36:24 any time :-) Jul 31 15:39:21 khem - you did see jordan - he was at indian oven Jul 31 15:40:08 HopsNBarley: yeah :) Jul 31 15:40:15 I Mezbaan Jul 31 15:40:37 ? Jul 31 15:40:44 HopsNBarley: I have no good experience there Jul 31 15:40:58 mezbaan is the name IIRC Jul 31 15:41:08 ah. Jul 31 15:41:32 i'll take your advice. Jul 31 15:41:51 hi genesis Jul 31 15:43:07 HopsNBarley: I have few good ones if you want indian food sometimes Jul 31 15:43:26 heck yeah! next time? Jul 31 15:44:01 sure Jul 31 15:50:20 pb__: so the syslog() and closelog() functions are the ones which are culprits Jul 31 15:50:29 * khem gets ready for work now Jul 31 15:52:36 khem hm? Jul 31 15:55:01 khem: okay, great Jul 31 15:55:14 do you have a full backtrace? I still don't quite understand how you get from pthread_mutex_xx into the unwinder. Jul 31 15:55:31 but, evidently this does happen somehow, so we will just have to deal with it. Jul 31 15:55:56 in that case, maybe the packaging rule should simply be that any use of a pthread function brings in libgcc. Jul 31 15:56:38 bbl, bye Jul 31 15:57:40 hi ph5 Jul 31 15:58:27 * pb__ go home Jul 31 15:58:30 bbl Jul 31 15:59:51 hi Jul 31 16:00:10 bye pb Jul 31 16:00:24 re florian Jul 31 16:00:56 re Jul 31 16:22:19 pb__: so these functions come from libc hence any application using these functions is needing libgcc_s.so.1 Jul 31 16:22:51 pb__: I will trace it in glibc how now it lands there Jul 31 16:23:12 * khem goes behind the walls Jul 31 16:59:30 khem: okay, excellent Jul 31 17:40:17 bye Jul 31 17:41:19 hm Jul 31 17:41:26 seems I will bye a beagle board too Jul 31 17:41:29 bye hrw Jul 31 17:43:17 woglinde, did u kill a beagle board? Jul 31 17:43:35 diego no Jul 31 17:43:47 but the specs are cool Jul 31 17:43:54 and the price is amazing Jul 31 17:44:03 yes Jul 31 17:44:23 hm only too few usb ports it seems Jul 31 17:44:57 1 otg and a non working 2.0 Jul 31 17:45:37 hm otg? Jul 31 17:45:39 it's very interesting, but atm i wait and see Jul 31 17:45:45 on the go Jul 31 17:45:58 it's just a 2.0 that work as host or client Jul 31 17:46:27 http://en.wikipedia.org/wiki/USB_On-The-Go Jul 31 17:46:32 ah yes now I remeber Jul 31 17:46:44 hm but 2 are to few Jul 31 17:46:50 mice keyboard Jul 31 17:46:53 then no harddisk Jul 31 17:47:04 not really, you can use a usb hub Jul 31 17:47:10 ah right Jul 31 17:47:16 I always forget about this Jul 31 17:47:22 but powered hub I would use Jul 31 17:47:47 for example 1 not powered usb for mice and keyboard is enough Jul 31 17:47:57 yes thats right Jul 31 17:48:24 hm Jul 31 17:48:30 and ethernet is missing Jul 31 17:48:56 thats not good Jul 31 17:49:09 usb net ^^ Jul 31 17:49:15 baahhhh Jul 31 17:49:18 thats crappy Jul 31 17:49:39 hm maybee there is something the second sd port Jul 31 17:49:48 let me google after sdio ethernet Jul 31 17:49:59 btw there are usb - ethernet adapter Jul 31 17:52:42 yes but thats ugly solution Jul 31 17:53:04 but you are right over sdio it will be slower Jul 31 17:53:48 diego but with ethernet connector you had streaming client out of the box Jul 31 17:55:21 usb2.0 should be fast nough for streaming Jul 31 17:55:28 yes Jul 31 17:55:38 in theory Jul 31 17:55:40 i have more or less 400kb/s over usb 1.1 Jul 31 17:56:27 and video are usually less the 1MB/s Jul 31 17:56:45 diego the tsp is hdtv compatible Jul 31 17:56:57 tsp? Jul 31 17:57:02 aeh dsp Jul 31 17:57:03 sorry Jul 31 17:57:45 so you're thinking at beagle as a video client? Jul 31 17:57:51 sure Jul 31 17:57:53 why not Jul 31 18:00:26 i see some works on ffmpeg for omap recently Jul 31 18:00:54 so someone else it's thinking at it too ^^ Jul 31 18:28:39 03  07org.openembedded.dev * rea8cf15ec3 10OE.dev/packages/linux/ (5 files in 2 dirs): linux omap2 git: add v2 of serial hang patches Jul 31 18:28:44 03  07org.openembedded.dev * r07fb5b7661 10OE.dev/: Jul 31 18:28:44 merge of '4161ac48e6849d2ce053bdbb00d74e709247fd18' Jul 31 18:28:44 and '6542fbe2d509990c5d4b56b4adaf0df31a29c221' Jul 31 18:28:45 03  07org.openembedded.dev * r336349e6dd 10OE.dev/: Jul 31 18:28:45 merge of 'f1c0af35653349c378da9671d2861f1aa08e4ecc' Jul 31 18:28:47 and 'f5b5a855a9ceef36d2e91eb4defd3cafafea854d' Jul 31 18:28:49 03  07org.openembedded.dev * r73ab16c787 10OE.dev/packages/gypsy/ (gypsy.inc gypsy_svn.bb): gypsy svn fix dependencies Jul 31 18:28:52 03  07master * r93d6963c94 10OE.dev/packages/linux/linux-kexecboot.inc: linux-kexecboot.inc: check kernelsize Jul 31 18:28:55 03  07master * r336349e6dd 10OE.dev/: Jul 31 18:28:57 merge of 'f1c0af35653349c378da9671d2861f1aa08e4ecc' Jul 31 18:28:59 and 'f5b5a855a9ceef36d2e91eb4defd3cafafea854d' Jul 31 18:29:01 03  07org.openembedded.dev * re87292f2f6 10OE.dev/: Jul 31 18:29:03 merge of '066c6fcbe98c8444c2fbff69204d0d8606d85077' Jul 31 18:29:05 and 'db68486a1e4128249ae5f54b7ac2b1278ba29b53' Jul 31 18:29:07 03  07master * r07fb5b7661 10OE.dev/: Jul 31 18:29:09 merge of '4161ac48e6849d2ce053bdbb00d74e709247fd18' Jul 31 18:29:13 and '6542fbe2d509990c5d4b56b4adaf0df31a29c221' Jul 31 18:29:15 03  07master * r73e368fb57 10OE.dev/packages/ffmpeg/ffmpeg_git.bb: ffmpeg git: make default for armv5+ on angstrom Jul 31 18:29:18 03  07master * re87292f2f6 10OE.dev/: Jul 31 18:29:20 merge of '066c6fcbe98c8444c2fbff69204d0d8606d85077' Jul 31 18:29:22 and 'db68486a1e4128249ae5f54b7ac2b1278ba29b53' Jul 31 18:29:24 03  07master * rea8cf15ec3 10OE.dev/packages/linux/ (5 files in 2 dirs): linux omap2 git: add v2 of serial hang patches Jul 31 18:29:29 03  07org.openembedded.dev * r93d6963c94 10OE.dev/packages/linux/linux-kexecboot.inc: linux-kexecboot.inc: check kernelsize Jul 31 18:29:32 03  07org.openembedded.dev * r73e368fb57 10OE.dev/packages/ffmpeg/ffmpeg_git.bb: ffmpeg git: make default for armv5+ on angstrom Jul 31 18:29:35 03  07master * r73ab16c787 10OE.dev/packages/gypsy/ (gypsy.inc gypsy_svn.bb): gypsy svn fix dependencies Jul 31 18:30:58 I followed the instructions on http://wiki.openembedded.net/index.php/Getting_Started -- where does it put the gpe-image file? Jul 31 18:32:09 tmp/deploy/images/..... Jul 31 18:32:41 thanks Jul 31 18:34:20 I'm trying to commit, but nothing is happening Jul 31 18:35:19 Crofton: what is I didn't specify a tmp dir in the local.conf Jul 31 18:35:34 03  07org.openembedded.dev * rceb807831c 10OE.dev/packages/images/fso-image-console.bb: remove bogus duplicated file Jul 31 18:35:34 03  07org.openembedded.dev * r91fbc2c313 10OE.dev/packages/freesmartphone/frameworkd_git.bb: frameworkd git catch up with buildsystem changes Jul 31 18:35:35 03  07org.openembedded.dev * rb37f3dd382 10OE.dev/conf/distro/include/sane-srcrevs.inc: sane-srcrevs.inc: bump frameworkd to catch up with buildsystem changes Jul 31 18:35:38 ls in tmp in the directory you ran bitbake from Jul 31 18:35:41 03  07master * rceb807831c 10OE.dev/packages/images/fso-image-console.bb: remove bogus duplicated file Jul 31 18:35:42 03  07master * r91fbc2c313 10OE.dev/packages/freesmartphone/frameworkd_git.bb: frameworkd git catch up with buildsystem changes Jul 31 18:35:42 03  07master * rb37f3dd382 10OE.dev/conf/distro/include/sane-srcrevs.inc: sane-srcrevs.inc: bump frameworkd to catch up with buildsystem changes Jul 31 18:35:48 and set a TMPDIR :) Jul 31 18:38:58 otherwise you rebuild a lot of stuff unless you run bb from the same cwd everytime Jul 31 18:44:25 Crofton: does bb creates the kernel as well? Jul 31 18:46:24 ls Jul 31 18:46:28 ? Jul 31 18:48:19 yes Jul 31 18:48:36 probably Jul 31 18:51:57 hi. the program I want to build in OE needs access to . I found out this header belongs to the kernel and the file can be found in the machine dependend staging directory. what is the expected way to the give the program access to it? Jul 31 18:53:40 re Jul 31 18:55:34 rschuster: it should Just Work(tm). Is that header file not staged into .../usr/include/asm/ucontext.h? If your recipe is sane, the default include path should find that. Jul 31 18:56:34 mwester: yeah, that is what I am expecting as well Jul 31 18:56:45 mwester: this file should be provided by linux-libc-headers, right? Jul 31 19:02:11 That's what I would expect. Jul 31 19:02:39 03  07org.openembedded.dev * r4bd0872ffa 10OE.dev/: Jul 31 19:02:39 merge of '4fe1dea49ed06e807afde9b963aafccb124edd4d' Jul 31 19:02:39 and 'aaffbdee48dde7a080e15c0f4d608f013c10a6ae' Jul 31 19:02:49 03  07org.openembedded.dev * r6762479d56 10OE.dev/packages/u-boot/u-boot_git.bb: u-boot_git : Update git rev to the tag for u-boot-1.3.4-rc2. Jul 31 19:02:56 03  07master * r6762479d56 10OE.dev/packages/u-boot/u-boot_git.bb: u-boot_git : Update git rev to the tag for u-boot-1.3.4-rc2. Jul 31 19:02:56 03  07master * r4bd0872ffa 10OE.dev/: Jul 31 19:02:56 merge of '4fe1dea49ed06e807afde9b963aafccb124edd4d' Jul 31 19:02:56 and 'aaffbdee48dde7a080e15c0f4d608f013c10a6ae' Jul 31 19:11:42 mwester: this is quite strange. I looked at the kernels' Kbuild files and they do not mention ucontext. as such it is not installed. however my debian and fedora boxen have their arch specific ucontext.h in /usr/include/asm. I assume we need to copy this file manually, then Jul 31 19:12:28 It's in the right places for my openmoko builds, so it may be something to do with your specific kernel? Jul 31 19:13:07 er, I take that back -- its in sys/ucontext.h but not in asm/ucontext.h Jul 31 19:13:38 mwester: which linux-libc-headers version is this? Jul 31 19:13:46 and as I poke about, I also find it in include itself, at the top level. Jul 31 19:14:18 linux-libc-headers-2.6.22-r2 Jul 31 19:14:26 mwester: yeah, right. I also have this file from glibc. but this header is not the right one. I need the definition for 'struct ucontext' Jul 31 19:15:12 * mwester vaguely remembers a change of this sort... Jul 31 19:17:02 mwester: by looking the recipes it seems that with 2.6.19 we use the kernel's install_headers target Jul 31 19:17:16 mwester: before that time we simply copied asm-$ARCH Jul 31 19:17:28 which of course included ucontext.h then Jul 31 19:17:50 rschuster which programm? Jul 31 19:18:30 woglinde: phoneme :) Jul 31 19:18:31 There were other recent problems; google "ucontext _GNU_SOURCE", I think. Jul 31 19:19:29 Basically a libc change removed some symbols or such, and required that one add a -DSOMETHING-OR-OTHER to get the right stuff pulled in. This won't fix your missing header file, but perhaps it's all tied up in the same restructuring? Jul 31 19:22:52 how do I specify in bb to compile only an image for my mobile? Jul 31 19:54:18 re Jul 31 19:57:57 re hrw Jul 31 20:01:32 have a nice rest of day guys Jul 31 20:37:03 Crofton: ping Jul 31 20:38:00 piong Jul 31 20:38:06 hi, I see you've committed u-boot-1.3.4-rc2 Jul 31 20:38:16 I have one question for you, pls Jul 31 20:38:20 http://bugs.openembedded.net/attachment.cgi?id=9860 Jul 31 20:38:50 no, I updated the git rev tags for two davinci boards Jul 31 20:38:51 I don't think will be committed, but howeever, what should be the right name and SRCREV? Jul 31 20:39:00 to build that for testing :) Jul 31 20:39:12 ok, just curious about namings here Jul 31 20:40:50 hmmm Jul 31 20:41:10 I'm not sure about the SRCREV Jul 31 20:41:18 I suppose +SRCREV_akita = "9bf86baaa3b35b25baa2d664e2f7f6cafad689ee" Jul 31 20:41:36 +SRCREV_c7x0 = "9bf86baaa3b35b25baa2d664e2f7f6cafad689ee" Jul 31 20:41:57 hmmm Jul 31 20:42:05 you are making me think :) Jul 31 20:42:23 :-] Jul 31 20:42:43 you are looking in u-boot_git.bb? Jul 31 20:42:57 yes Jul 31 20:43:23 what I see in your u-boot-zaurus_1.1.4 has the tag in the SRC_URI Jul 31 20:43:35 and it works... Jul 31 20:43:44 leave the tag out and use SRCREV_machine to set the actual rev Jul 31 20:43:55 on a per machine basisi Jul 31 20:44:15 maybe the SRCREV_ var overrides any tag in SRC_URI? Jul 31 20:44:58 I checked manually the tag Jul 31 20:46:42 ok, now system is rebuilding, I'll try in a few minutes integrating in u-boot_git.bb Jul 31 20:47:46 (it was meant to be the less intrusive way) Jul 31 20:48:51 lart mkdir for not reading from stdin in pipe Jul 31 20:49:35 ah, that's why there are piles of it... Jul 31 20:56:57 Crofton: yes, build completed Jul 31 20:57:04 does it work ? Jul 31 20:57:21 yes, as you expected Jul 31 20:58:13 great Jul 31 20:58:16 patches applied, it's the very same version! Jul 31 20:58:26 1.1.4 urgh Jul 31 21:00:04 Crofton: if I extend u-boot_git.bb, where should I put the patches? Jul 31 21:00:17 simply in /files ? Jul 31 21:02:47 yeah Jul 31 21:02:55 copy the beagleboard approach Jul 31 21:05:56 ok, thx Jul 31 21:13:43 Crofton: what's 'nicer' : SRC_URI_append_machine or SRC_URI_machine += ? Jul 31 21:14:04 I suspect both Jul 31 21:14:09 not sure which is best Jul 31 21:14:12 in fact I see both Jul 31 21:14:56 ant append is nicer I think Jul 31 21:22:01 and for do_configure_prepend() ? Jul 31 21:22:16 one for each machine (two atm)? Jul 31 21:25:07 hi again. Jul 31 21:35:08 he thesing Jul 31 21:35:33 it seems that my mainboard died again Jul 31 21:36:27 And this board was already repaired early this year. Jul 31 21:40:53 hi guys! Jul 31 21:41:09 can anybody tell me if the avr32 platform is supported with OE ? Jul 31 21:42:32 it is supported Jul 31 21:46:12 ant__: those two SRC_URI syntaxes do quite different things Jul 31 21:46:31 "SRC_URI_machine +=" will completely replace the old SRC_URI, not add to it Jul 31 21:46:33 great Jul 31 21:46:40 will it work with poky? Jul 31 21:46:49 does anybody have some experience with that? Jul 31 21:46:50 kwegner dont know in which shape it is now Jul 31 21:47:06 with poky or with avr32? Jul 31 21:47:28 both :-) Jul 31 21:50:59 pb_: yes, I see now that neurus has own SRC_URI_neuros-os2 Jul 31 21:55:52 kwegner: you should ask Crofton about avr32. I think he did some of the stuff. Jul 31 21:56:26 not me Jul 31 21:57:07 hihi Jul 31 21:57:12 well anybody? Jul 31 21:58:03 send something to the ml Jul 31 21:59:20 ok will do so Jul 31 22:04:34 kwegner: Poky does not contain avr32 support as standard since we don't have any hardware to test it on Jul 31 22:04:54 kwegner: OE does support avr32 and it would be trivial to add that support to Poky Jul 31 22:05:44 Well, trivial as in just copying to files around and debugging any failures Jul 31 22:06:40 hmm Jul 31 22:06:49 sounds as if i should give it a shot :-) Jul 31 22:08:50 RP: I filed a report in the Angstrom-dev ML about QA issues (workdir in path). This happens when TMPDIR = "/OE/${DISTRO}-tmp/" Jul 31 22:09:19 dunno if it's a bug or a missing feature Jul 31 22:09:32 (from Poky :-) Jul 31 22:10:20 http://lists.linuxtogo.org/pipermail/angstrom-distro-devel/2008-July/002416.html Jul 31 22:10:58 but I suppose it's not distro-specific Jul 31 22:28:47 hmm, any reason why /machine/netbook-pro.conf is 755? Jul 31 22:29:06 accident Jul 31 22:29:11 and no one cares Jul 31 22:34:30 woglinde: I found it casually browsing for c7x0.conf Jul 31 22:38:36 ant__: I nearly have a new bootmenu ready Jul 31 22:39:08 thesing: wie weit? Jul 31 22:39:18 I planed to finish it today Jul 31 22:39:22 :-] Jul 31 22:39:54 Its nearly done. I just need to compile it against klibc and hope that it will work ;) Jul 31 22:40:09 great Jul 31 22:40:50 If you want to use it you need to change to defconfig of linux-kexecboot for your machine. You can take the collie defconfig as a basis. Jul 31 22:41:14 how much Kb would cost if you try to pack mtd-utils and nandlogical in intramfs? Jul 31 22:42:56 No idea. Try to build them with klibc. the kexec-static recipe is an example how oe can do this. Jul 31 22:43:49 ok Jul 31 22:44:56 Btw. why bother with a initramfs-rescue image? you can have one on some sd card. Jul 31 22:46:11 true, but I'd prefer it in flash Jul 31 22:46:50 I fear however, no NFS boot will be possible Jul 31 22:47:30 if we respect the partitions Jul 31 22:53:39 I'm not sure if one really needs nsf root on a pda. (except for development) Jul 31 22:54:32 and with sd cards so cheap you can have some life distro card. Jul 31 22:54:35 for dev and testing other distros Jul 31 22:56:06 * * OE Bug 4463 has been created by raj.khem(AT)gmail.com Jul 31 22:56:08 * * gdb-cross does not build on x86_64 hosts Jul 31 22:56:11 * * http://bugs.openembedded.net/show_bug.cgi?id=4463 Jul 31 22:56:15 thesing how can I test with sdcard? Jul 31 22:56:41 which machine? Jul 31 22:56:44 simpad Jul 31 22:57:05 before I struggle with nfsroot again Jul 31 22:57:15 just copy rootfs? Jul 31 22:57:58 you need to set the cmdline of the kernel and put the rootfs on the card. Jul 31 22:58:38 I just untar the oe-build tgz to a card Jul 31 22:58:47 ah simpadbootlodaer can only read vfat Jul 31 22:58:57 so I can but the kernel on vfat parttition Jul 31 22:59:06 and the rest under ext2 or whatever Jul 31 23:01:27 yes. Jul 31 23:03:30 hm maybee one problem occour Jul 31 23:03:39 bootloader rewrites cmd Jul 31 23:03:42 line Jul 31 23:07:10 can't you tell the bootloader which cmdline you want? Jul 31 23:21:26 so las try with my awk mkdir fixing stuff Jul 31 23:22:28 03  07org.openembedded.dev * r3ba61654f1 10OE.dev/packages/linux/linux-rp.inc: linux-rp.inc: package command line for a kernel Jul 31 23:22:36 03  07org.openembedded.dev * r522bcd0932 10OE.dev/: Jul 31 23:22:36 merge of '250b98d355e1a4b79caaec9ef041f42db95a2b7f' Jul 31 23:22:36 and '597f2d2c1d1ea251c244da83e5677a24b328d3f4' Jul 31 23:22:36 03  07master * r3ba61654f1 10OE.dev/packages/linux/linux-rp.inc: linux-rp.inc: package command line for a kernel Jul 31 23:22:46 03  07master * r522bcd0932 10OE.dev/: Jul 31 23:22:46 merge of '250b98d355e1a4b79caaec9ef041f42db95a2b7f' Jul 31 23:22:46 and '597f2d2c1d1ea251c244da83e5677a24b328d3f4' Jul 31 23:23:49 sweet now it works Jul 31 23:24:11 got your simpad to boot from sd? Jul 31 23:25:09 nope Jul 31 23:25:31 I fixed some buildsystem stuff Jul 31 23:26:01 when a java compiler dont make the dirs with -d options Jul 31 23:32:32 nite Jul 31 23:33:05 thesing maybee on weekend I have some time to sort things out with simpad Jul 31 23:41:14 03  07org.openembedded.dev * r565abf71a0 10OE.dev/packages/jamvm/jamvm-native.inc: jamvm-native.inc: Set RDEPENDS to empty to prevent 'classpath' dependency from jamvm.inc. Jul 31 23:41:15 03  07org.openembedded.dev * r251e31dfc9 10OE.dev/: Jul 31 23:41:15 merge of '42cb81c40550397bea69b71bcc8768f0b8b81476' Jul 31 23:41:15 and '8cb0572b0f44ef514300735efee6e6721ef539f1' Jul 31 23:41:21 03  07master * r565abf71a0 10OE.dev/packages/jamvm/jamvm-native.inc: jamvm-native.inc: Set RDEPENDS to empty to prevent 'classpath' dependency from jamvm.inc. Jul 31 23:41:21 03  07master * r251e31dfc9 10OE.dev/: Jul 31 23:41:21 merge of '42cb81c40550397bea69b71bcc8768f0b8b81476' Jul 31 23:41:23 and '8cb0572b0f44ef514300735efee6e6721ef539f1' Aug 01 00:03:12 night Aug 01 00:43:25 03  07org.openembedded.dev * r84bbd1fe80 10OE.dev/packages/classpath/classpath-native_0.97.2.bb: Aug 01 00:43:25 classpath-native 0.97.2: Aug 01 00:43:25 - Added patch to allow compilation of phoneme advanced Aug 01 00:43:26 03  07org.openembedded.dev * rfcf7431d2a 10OE.dev/: Aug 01 00:43:26 merge of '13530e1cbba5e3db86e8ea9713ffcf460f77b92c' Aug 01 00:43:27 and 'b70c18ddb17c1721c0c521628f96624b1a6bef2b' Aug 01 00:43:31 03  07master * rfcf7431d2a 10OE.dev/: Aug 01 00:43:33 merge of '13530e1cbba5e3db86e8ea9713ffcf460f77b92c' Aug 01 00:43:35 and 'b70c18ddb17c1721c0c521628f96624b1a6bef2b' Aug 01 00:43:37 03  07master * r84bbd1fe80 10OE.dev/packages/classpath/classpath-native_0.97.2.bb: Aug 01 00:43:40 classpath-native 0.97.2: Aug 01 00:43:42 - Added patch to allow compilation of phoneme advanced Aug 01 01:39:04 03  07org.openembedded.dev * rcf86070743 10OE.dev/packages/openmoko-panel-plugins/openmoko-panel-battery_svn.bb: openmoko-panel-battery: needs libnotify now Aug 01 01:39:05 03  07org.openembedded.dev * r8c37222306 10OE.dev/: Aug 01 01:39:05 merge of 'b99a6dc2eb39390c552fe0329acbb5ed94fa5fcf' Aug 01 01:39:05 and 'ef63b228877ae1c3c905d3921b256693b1914dea' Aug 01 01:39:10 03  07master * rcf86070743 10OE.dev/packages/openmoko-panel-plugins/openmoko-panel-battery_svn.bb: openmoko-panel-battery: needs libnotify now Aug 01 01:39:11 03  07master * r8c37222306 10OE.dev/: Aug 01 01:39:11 merge of 'b99a6dc2eb39390c552fe0329acbb5ed94fa5fcf' Aug 01 01:39:13 and 'ef63b228877ae1c3c905d3921b256693b1914dea' Aug 01 02:22:55 Hello. **** ENDING LOGGING AT Fri Aug 01 02:59:56 2008