**** BEGIN LOGGING AT Mon Apr 27 02:59:57 2009 Apr 27 06:17:53 Goodmorning Apr 27 06:19:30 Can it be, that an incorrect rootfs also causes a wrong rootfs (as in: only 6 dirs has been created instead of the normal 11 or 12) Apr 27 06:40:09 good morning Apr 27 06:43:02 ah, I'm not very awake yet, I meant to say that an kernel recipe kan cause an incorrect rootfs Apr 27 06:44:38 as I have the silly problem that I have 2 kernel recipes, the 1st gives me a working kernel, but ruins my rootfs Apr 27 06:44:57 and the seconds gives me a not-working kernel, but a correct rootfs Apr 27 06:57:20 good morning Apr 27 07:41:03 khem: boo! Apr 27 07:45:30 morning Apr 27 07:46:33 can someone help to resolve this issue "ERROR: Nothing PROVIDES quake after a bitbake quake Apr 27 08:05:13 and 'quake1' or 'quake2' ? Apr 27 08:05:53 I don't really care one which work on the EVM lol Apr 27 08:06:10 bitbake -v -b /home/fabroy01/oe/stuff/openembedded/recipes/quake/quake1_0.0.1.bb Apr 27 08:06:42 This is my log Apr 27 08:06:44 fabroy01@fabroy01-desktop:~/oe$ bitbake -v -b /home/fabroy01/oe/stuff/openembedded/recipes/quake/quake1_0.0.1.bb Apr 27 09:13:59 florian: good morning Apr 27 09:16:27 good morning Apr 27 09:25:52 morning Apr 27 09:28:15 hello Apr 27 09:30:03 I started building opie-image on my workplace's red hat 5, i586, but i am not root so I cannot continue building when locales are built (requires root to perform vm.mmap_min_addr = 0 for qemu). Can i simply copy paste all gigabytes on my laptop (gentoo, i586) and continue building ? Apr 27 09:33:05 03Jan Lübbe  07fso/milestone5.5 * rc8dfc97a2e 10openembedded.git/conf/distro/include/sane-srcrevs.inc: sane-srcrevs.inc: add webkit-efl revision Apr 27 09:33:16 sledgeas: no Apr 27 09:34:02 hrw: thanks for answer :( Apr 27 09:41:30 03Marcin Juszkiewicz  07org.openembedded.dev * rcc75c744a1 10openembedded.git/ (4 files in 3 dirs): genext2fs: added 1.4.1 version Apr 27 09:50:24 http://jmtd.net/log/dm9601/ Apr 27 10:03:12 I did a bitbake quake2, it seems to work out Apr 27 10:03:41 but in the local.conf I gave the path to the arm gcctoolcahin I want to use Apr 27 10:04:20 Oe is building it's own glibc, ... is there a way to use the prebuilt one ? Apr 27 10:07:55 hi mickeyl Apr 27 10:08:03 morning Apr 27 10:13:02 hi mickeyl Apr 27 10:13:52 yo florian Apr 27 10:52:58 guys some people have problems baking libtool-native-1.5.10-r4 in our dreambox branch Apr 27 10:53:12 they get a gzip: stdout: Broken pipe error Apr 27 10:53:48 any idea what could cause that? fetch is successful though, so download not corrupted or anything Apr 27 10:53:56 and it's independent cases Apr 27 10:56:46 mickeyl: hi, thx again for a good konference :) now i'm actually trying to use your project ;) Apr 27 11:13:59 does someone know if there a kind of eeprom (e2p) device emulation for linux? Apr 27 11:14:40 the smsc usb driver seems to need an eeprom for working properly Apr 27 11:15:13 i would like to get the driver without the eeprom working Apr 27 11:15:21 any idea? Apr 27 11:17:11 boto|work: without eeprom it should get random MAC address Apr 27 11:17:41 boto|work: I have dm9601 usb ethernet here. one has working eeprom and works as ethX, second has broken eeprom so gets random mac and works as usbX Apr 27 11:18:00 yep Apr 27 11:18:09 it gets usb1 then here Apr 27 11:18:27 but the driver shows that some usb specific parameters are hold in eeprom Apr 27 11:18:46 such as rx/tx modes and vendor id, etc. Apr 27 11:19:21 then hardcode them in driver? Apr 27 11:19:31 and having a predefined product specific mac is a must for us Apr 27 11:19:45 yes, this is the last option i wanted to do Apr 27 11:20:10 i thought there may be a better solution, e.g. an e2p emulator Apr 27 11:23:05 boto|work: 'ifconfig ethx hw ether 11:22:33:44:55:66' does not work? Apr 27 11:23:21 no Apr 27 11:23:41 if the eeprom is not present then you cannot change anything in smsc driver Apr 27 11:40:16 RP: 2.6.3'-rc2 resumes after suspend!!! Still recharging problems (that's very intricated code...as you know ;-) Apr 27 11:40:41 (on Zaurus spitz) Apr 27 11:40:59 http://www.omegamoon.com/blog/index.php?entry=entry090425-221741 Apr 27 11:41:30 nice Apr 27 11:41:40 Ill' update to 2.6.29+2.6.30-rc2 later Apr 27 11:41:51 * hrw upgrades bb to 2.6.29 Apr 27 11:41:55 usb1: register 'dm9601' at usb-musb_hdrc-1.4, Davicom DM9601 USB Ethernet, ff:ff:ff:ff:ff:ff Apr 27 11:41:57 (still waiting for the patches :-) Apr 27 11:41:59 thats ugly Apr 27 11:43:07 hrw: lot of requests about DM9601...strangely Apr 27 11:43:21 iirc one user asked for inclusion in linux-rp time ago Apr 27 11:43:28 does it work??? Apr 27 11:46:00 ant_work: dm9601 is cheap usb-ethernet donlt Apr 27 11:46:03 dongle Apr 27 11:46:28 he...I lack usb-host on c7x0 :/ Apr 27 11:48:57 I have two of them and use with devices which lacks onboard ethernet Apr 27 11:50:03 hrw: Andy Gospodarek (1): ..... dm9601: runtime mac address change support. Apr 27 11:50:03 ah, no..Peter Korsgaard (1): Apr 27 11:50:03 dm9601: runtime mac address change support Apr 27 11:50:03 seems the little bastard is not working out-of-the box with older kernel..users were right Apr 27 11:50:19 using with 2.6.24 / 26 ? Apr 27 11:52:32 ant_work: with 2.6.29 it gets random mac Apr 27 11:52:33 ant_work: as I said I have two dm9601 dongles. one has broken eeprom so 2.6.28 == ff:ff:ff mac, under 2.6.29 it gets random Apr 27 11:52:57 I'd check that patch ^^^ Apr 27 11:54:21 ant_work: I am moving to 2.6.29 on all machines Apr 27 11:56:28 <_cpo_> anyone who can supply me with a byte swapping solution in bash? like transfrom "31007889" into "00318978" Apr 27 12:10:36 hi all, a litte question: i have/had an running debian installation on an zaurus-c3200, when i am booting an openembedded kernel ( 2.26) i cannot chroot or execute anz binary .. some ideas ?? Apr 27 12:11:00 Debian arm or armel? Apr 27 12:11:57 i think debian is arm Apr 27 12:12:11 Use armel, at least. Apr 27 12:12:25 The arm is OABI and I'd be surprised if the OE kernels were built with support for it. Apr 27 12:12:40 (lenny has both architectures, arm will be dropped for squeeze) Apr 27 12:13:47 that means throw away my debian installation an install debian armel or get rid of oe kernel ? Apr 27 12:14:09 OE doesnt support OABI indeed, but you can enable it by yourself ofcourse Apr 27 12:14:22 pwgen: Yes. Apr 27 12:14:35 Or build your own kernel. Apr 27 12:15:32 whats the diffrent betwee OABI an non OABI ? some webpages about it ? Apr 27 12:15:37 pwgen: 2.6.26 kexecboot kernel has echo "CONFIG_AEABI=y" >> ${S}/.config Apr 27 12:15:46 (linux.inc) Apr 27 12:16:53 pwgen: same in linux-rp.inc (standard kernels) Apr 27 12:17:09 so your kernel is 'OABI-compat' Apr 27 12:17:28 http://en.wikipedia.org/wiki/Application_binary_interface (EABI is used by Angstrom / OE) Apr 27 12:17:52 see..echo "CONFIG_OABI_COMPAT=y" >> ${S}/.config Apr 27 12:18:19 broonie: we removed OABI emulation as it causes slowdown Apr 27 12:22:18 pwgen: still using u-boot? Apr 27 12:22:48 ant_work: yes latest ubeoot-kexeimage with 2.26 kernal as uboot Apr 27 12:22:57 I fear there is a mistake wrt LOAD_ADDRESS for akita/spitz... Apr 27 12:23:20 it seems addresses differs Apr 27 12:23:55 for corgi / c7x0 0xA0008000 Apr 27 12:24:12 the kexeboot process works fine for all images build with oe, but when i want to boot my deb installation or chroot to it fails .. Apr 27 12:25:52 pwgen: for akita / spitz should be 0xa1000000 Apr 27 12:26:19 pwgen: tried with Zubuntu? Apr 27 12:26:42 (armel) Apr 27 12:27:05 Debian has armel too, AFAICT he's just installed arm by mistake. Apr 27 12:27:22 possible... Apr 27 12:28:12 broonie: while you're there...about audio DSP Apr 27 12:28:40 are that Shark chips reallyso good? Apr 27 12:28:44 http://www.jrrshop.com/universal-audio-uad1e-express-p-6285.html Apr 27 12:29:03 can an Omap do similar audio dsp-processing? Apr 27 12:29:22 or custom silicon still has big advantage? Apr 27 12:29:42 Shark chips? Apr 27 12:29:43 ant_work ? Zubuntu ( has ubuntu something for arm ? ) Apr 27 12:29:55 pwgen: Yes, Ubuntu Jaunty has arm as standard now. Apr 27 12:30:46 (:-(( i dont want a new installation, i want my nice, stable and full featured debian back ... (:-(( Apr 27 12:30:46 ant_work: I suspect that's just a random DSP bolted onto a PCI card. Apr 27 12:31:05 well, one or four is the same? Apr 27 12:31:20 seems to be Analog Devices Apr 27 12:31:45 http://remixmag.com/gear/reviews/review-universal-audio-uad/ Apr 27 12:31:50 If it's a SHARC they're very old hat in the DSP world; $WORK-1 ditched them a few years ago. Apr 27 12:32:08 new Analog Devices SHARC 21369 DSP Apr 27 12:32:10 ? Apr 27 12:32:19 what i have to do too get my debian system back, compile a kernel without CONFIG_AEABI ? or with it ? Apr 27 12:32:21 They can do a moderate number of conference channels and will be cheap but they're not all that. Apr 27 12:32:52 a friend of mine says these keep up with 400channels zero latency... Apr 27 12:33:22 400 channels doing what? Apr 27 12:33:24 morning Apr 27 12:33:32 audio 96khz 24 bits Apr 27 12:33:34 ? Apr 27 12:33:35 Playback and record, yeah. Conferencing not so much. Apr 27 12:33:43 good afternoon Jay7 Apr 27 12:33:55 hi Jay7 Apr 27 12:35:50 broonie: I'm alsways dubious about DSP..and the quality of algs! Apr 27 12:36:51 Yes, though many of the companies doing that sort of stuff are actually software companies with hardware dongles they work to effectively. Apr 27 12:37:21 broonie: other fluff seems Holliwood Quality http://www.hqv.com/technology.cfm Apr 27 12:38:21 If you're critically dependant on that stuff you're often best coding the algorithms yourself. Apr 27 12:38:32 :-) Apr 27 12:38:47 nah, I'm just waiting to buy my next AV ampli Apr 27 12:38:51 since depending on external software is inherantly risky. Apr 27 12:39:09 Oh, in that case just treat it as a black box bit of software :) Apr 27 12:39:23 I've already had nice hardware with poor software Apr 27 12:39:57 Yes, it's the software that's interesting not the hardware. Apr 27 12:40:44 "one trillion ops sec Apr 27 12:40:50 " Apr 27 12:41:53 broonie: the UAD-2's Shark DSP deliver 2400Mflops per chip, a mainstream Intel E8500 delivers 21400Mflops Apr 27 12:42:04 now, two UAD-2 Quad's have +/- equal power - just have to pay 3000 instead of 200 USD for the "same" Apr 27 12:42:06 and then, some more money for plugs.... Apr 27 13:00:51 hm..help with HEX... Apr 27 13:01:06 # #0x00008000 est représenté par : 0x80 ror 24, codé par 0xC80 ; Apr 27 13:01:31 does it mean $ /oe/build/tmp/angstrom/cross/armv5te/bin/arm-angstrom-linux-gnueabi-objdump Apr 27 13:01:32 -p vmlinux | sed -n 's/LOAD.*vaddr \([^ ]*\).*/\1/p' Apr 27 13:01:34 0x00000000 Apr 27 13:01:35 0xc0008000 Apr 27 13:01:49 found the right address (0xA0008000)? Apr 27 13:02:17 it's about u-boot automatic addresses retrieve Apr 27 13:19:57 03Koen Kooi  07org.openembedded.dev * r6259f5219f 10openembedded.git/recipes/linux/ (linux-kirkwood/defconfig linux-kirkwood_2.6.29-rc7+git.bb): linux-kirkwood: update to 2.6.29.1 and enable some goodies like mmc and ubifs Apr 27 13:19:57 03Koen Kooi  07org.openembedded.dev * reee57a831f 10openembedded.git/conf/machine/ (include/kirkwood.inc sheevaplug.conf): sheevaplug: add ubifs params Apr 27 13:27:03 some people have problems baking libtool-native-1.5.10-r4 in our dreambox branch in unpack step: "gzip: stdout: Broken pipe error"... does anybody have an idea what could be causing that? download is not corrupted... Apr 27 13:42:08 Does someone ever experienced this error before : ERROR: QA Issue: No GNU_HASH in the elf binary: Apr 27 13:42:14 ? Apr 27 13:43:11 yes Apr 27 13:43:21 we had many fixes for that recently Apr 27 13:43:32 I am on ubuntu 8.10 Apr 27 13:50:42 hrw: Where can I find this fixes Apr 27 13:51:33 git-pull Apr 27 13:57:14 I did the setup on Friday afternoon .. I doubt git pull will change anything Apr 27 14:06:14 so tell us which binary has this problem and which recipe it is Apr 27 14:07:23 Roman_Khimov, I think there's 2 votes at least for switching to coreutils-native, time for a patch :) Apr 27 14:09:20 Tartarus: Uh-oh Apr 27 14:09:29 Tartarus: there is another issue with /cross...if you flush /tmp (and .cross too) the cross-toolchain is rebuilt witha wrong path Apr 27 14:09:54 dunno if it's more for cross/toolchains or originated by packaged-staging Apr 27 14:11:04 result: [20090426 22:57:29] oh..sh@t..| configure: error: C compiler cannot create executables Apr 27 14:16:01 ant_work, hmm? I don't try and work w/ things outside of TMPDIR Apr 27 14:16:11 or for that matter, not blowing away TMPDIR frequently :) Apr 27 14:16:14 no, vanilla Apr 27 14:16:25 just wipe tmpdir but /pstage Apr 27 14:16:37 ok Apr 27 14:16:41 I also don't do that :) Apr 27 14:16:48 So no ideas, sorry Apr 27 14:17:00 strange is kergoth did it Apr 27 14:17:13 w/out issues... Apr 27 14:18:05 one is /cross/armv5te/bin other is just /cross/bin Apr 27 14:18:21 ^^^ from scratch ^^^ from pstage Apr 27 14:18:36 no 'armv5te' Apr 27 14:25:35 oh crap i will never ever get my debian istallation back ... without oabi -> no success , with abi no success ... CRAP ! Apr 27 15:01:21 03Jan Lübbe  07fso/milestone5.5 * r1daf1de20c 10openembedded.git/ (3 files in 2 dirs): fso-image: add midori (0.1.2) and bump frameworkd Apr 27 15:22:17 otavio: alive? Apr 27 15:22:25 hrw: yep Apr 27 15:23:56 otavio: genext2fs patch - you acked it before I sent my second mail (that it is for stable/2009) or you acked it for stable? Apr 27 15:31:41 hrw: I acked it before but it is OK to go to stable, no problem with that Apr 27 15:31:46 hrw: want me to ack it again? Apr 27 15:32:42 no - just wanted to be sure Apr 27 15:34:09 hrw: np :-) Apr 27 15:34:13 03Marcin Juszkiewicz  07stable/2009 * r7e1993c868 10openembedded.git/ (4 files in 3 dirs): Apr 27 15:34:13 genext2fs: added 1.4.1 version Apr 27 15:34:13 Signed-off-by: Marcin Juszkiewicz Apr 27 15:34:13 Acked-by: Otavio Salvador Apr 27 15:34:16 hrw: thx by the catches up :-) Apr 27 15:35:30 otavio: with that patch my buildbot should go green Apr 27 15:37:47 hrw: NICE :-) Apr 27 16:16:21 hi,canola 2 is now free Apr 27 16:16:56 http://en.wikipedia.org/wiki/Canola_%28software%29 Apr 27 16:17:04 maybe it could be interresting for mamona Apr 27 16:28:29 hrw, looking at / modifying your buildbot cfg now, I think I see a bug Apr 27 16:28:49 in your incremental build cleaning you rm tmp/deploy/images/*/*/*image* Apr 27 16:28:57 Should be tmp/deploy/*/images/*/*image* Apr 27 16:29:14 angstrom is ${TMPDIR}/deploy/${LIBC}/images and such Apr 27 16:31:46 Should actually use a find command to do the work; other distros (I'll let folks guess which one might be an example) do not use the $(LIBC) intermediate layer. Apr 27 16:32:12 Also true :) Apr 27 16:32:30 But it's a bug in angstrom_*_incremental functions at least :) Apr 27 16:32:49 Ah, yes, if it's angstrom-specific. Apr 27 16:41:18 Oh, and html.Waterfall is deprecated and wants to be html.WebStatus now Apr 27 16:41:24 I'll update once I see how this all looks, if you want Apr 27 16:43:56 who is working with uboot for zaurus ? Apr 27 16:47:06 OK, got buildbot started with a build that should start any second now Apr 27 17:00:16 bbl Apr 27 17:09:59 mmm...does the i.MX515 has the infamous powerVR? Apr 27 17:12:12 Hello all! Apr 27 17:13:39 Anyone here use Angstrom enough to be able to talk about it? Apr 27 17:13:43 03Roman I Khimov  07org.openembedded.dev * r3db9402b8a 10openembedded.git/conf/bitbake.conf: Apr 27 17:13:43 bitbake.conf: fix parallel build for gzip'd images Apr 27 17:13:43 Parallel build of gzip'd images is broken now because of the way Apr 27 17:13:43 ${DEPLOY_DIR_IMAGE}/tmp.gz dir is removed-created-removed for every image. Apr 27 17:13:45 Fix it by 'installing' the dir and let it stay in place. Apr 27 17:13:47 Acked-by: Otavio Salvador Apr 27 17:29:28 03Jeremy Lainé  07org.openembedded.dev * rcab24c255b 10openembedded.git/recipes/qt4/ (3 files): qt4: make Qt 4.5.1 the default version Apr 27 17:31:13 Tartarus: right.. I changed that from Poky configs Apr 27 17:33:08 figured Apr 27 17:33:23 almost got things working locally, heh Apr 27 17:33:46 I think my last failure was the 'git clone failed randomly, curses..' problem Apr 27 17:34:27 Tartarus: so I got 12GB back..' Apr 27 17:35:08 Tartarus: I know that waterfall is deprecated. but I prefer it ;D Apr 27 17:35:17 ok Apr 27 17:37:11 yay git clone finished Apr 27 17:37:18 and this time I have it pointing at scripts in the right spot Apr 27 17:37:57 and I see processors being used Apr 27 17:38:39 03Angus Ainslie  07fso/milestone5.5 * r8f10aa89b9 10openembedded.git/conf/machine/om-gta02.conf: om-gta02.conf : ship the corect kernel for this machine Apr 27 17:38:40 03Angus Ainslie  07fso/milestone5.5 * r9c5bdb238f 10openembedded.git/: Merge branch 'fso/milestone5.5' of git@git.openembedded.net:openembedded into fso/milestone5.5 Apr 27 17:38:44 03Angus Ainslie  07fso/milestone5.5 * rd4a641db2e 10openembedded.git/conf/machine/om-gta01.conf: om-gta01.conf : ship a newer kernel for this machine Apr 27 17:45:51 03Koen Kooi  07org.openembedded.dev * r755970ac7c 10openembedded.git/recipes/cherokee/ (cherokee_0.98.0.bb files/cherokee.init): Apr 27 17:45:51 cherokee: fix init script, daemonize is '-d', not '-b' Apr 27 17:45:51 * runtime tested on sheevaplug/angstrom Apr 27 17:48:38 when I do "bitbake console-image" I see X stuff downloading. Is this normal? Apr 27 17:48:49 yes Apr 27 17:48:57 if it ends up in the image itself, no Apr 27 17:49:22 hm Apr 27 17:49:35 it shouldn't, however Apr 27 17:49:49 We build "big" or comprehensive, and install specific Apr 27 17:50:14 Does it need intermediate stuff on the build machine? Apr 27 17:50:48 The build will build some X stuff, but no, it should not need host X stuff Apr 27 17:50:53 Tartarus: remember that this is python so you can do anything Apr 27 17:51:13 hrw, yeah Apr 27 17:51:28 I need to hack things a little bit for dev to work too, since bitbake isn't in there Apr 27 17:51:35 But for now, I just wanna start easy :) Apr 27 17:51:45 Tartarus: thats just edit to full-oe-autobuild ;) Apr 27 17:52:18 heh Apr 27 17:52:50 Yeah, I want to expand this to all covered stuff in stable, at least that'll be different for meta-toolchain and canadian-sdk when possible Apr 27 17:52:55 Then cover dev Apr 27 17:53:01 Then see if I have any spare "overnight" cycles :) Apr 27 17:53:28 At least since this is my work box I can kick it off sooner than 11pm Apr 27 17:53:38 Tartarus: ;) Apr 27 17:54:19 Tartarus: is your bbot available via http Apr 27 17:54:20 ? Apr 27 17:54:36 I haven't yet confirmed http is active in it, actually Apr 27 17:54:42 but I might setup a port forward for it Apr 27 17:55:37 Tartarus: in my setup 8010 port is only local, 8011 (without force build) is http://home.haerwu.biz:8010/ forwarded Apr 27 17:59:33 Tartarus: and remove 'touch conf/local.conf' and next touch lines to make use of cached parsing Apr 27 18:10:18 bye Apr 27 18:24:58 03Angus Ainslie  07fso/milestone5.5 * r7e89c06c2a 10openembedded.git/recipes/linphone/ (3 files in 2 dirs): linphone: add patch for preferences SIGSEGV Apr 27 18:26:02 03Koen Kooi  07stable/2009 * r2a2c9de89d 10openembedded.git/recipes/cherokee/ (cherokee_0.98.0.bb files/cherokee.init): Apr 27 18:26:02 cherokee: fix init script, daemonize is '-d', not '-b' Apr 27 18:26:02 Signed-off-by: Koen Kooi Apr 27 18:26:02 Acked-by: Marcin Juszkiewicz Apr 27 18:51:58 03Denis 'Gnutoo' Carikli  07org.openembedded.dev * r5fdfa25604 10openembedded.git/recipes/gtk+/gtk+_2.8.16.bb: Apr 27 18:51:58 gtk+-2.8.16 : fixed problematic " in SRC_URI: Apr 27 18:51:58 there was a suplementary " in SRC_URI that made it fail at: Apr 27 18:51:58 run.staging_helper.16868: line 869: syntax error near unexpected token `else' Apr 27 18:52:00 03Denis 'Gnutoo' Carikli  07org.openembedded.dev * r413cc98aa5 10openembedded.git/recipes/dvbtools/dvb-apps_1.1.1.bb: Apr 27 18:52:03 dvb-apps_1.1.1.bb : fixed LDFLAGS QA(GNU-hash) Apr 27 18:52:05 Thanks Tartarus for telling me what was the standard way to do it Apr 27 18:52:07 03Denis 'Gnutoo' Carikli  07org.openembedded.dev * r691fd364dc 10openembedded.git/: Merge branch 'org.openembedded.dev' of git://git.openembedded.net/openembedded into org.openembedded.dev Apr 27 18:52:10 03Denis 'Gnutoo' Carikli  07org.openembedded.dev * rcf47f64bfa 10openembedded.git/: Merge branch 'org.openembedded.dev' of git://git.openembedded.net/openembedded into org.openembedded.dev Apr 27 18:56:37 is there a way to find which recipe was used to build the cross tools? Apr 27 19:06:37 gcc-4.3.3 builds for me on Angstrom/beagle in .dev Apr 27 19:07:20 hoj, yes, look at your distro conf, they usually set it Apr 27 19:07:32 Crofton|work, target 'gcc' ? hmm Apr 27 19:08:14 One more reason I need to get my buildbot stuff going on dev, heh Apr 27 19:08:30 03Leon Woestenberg  07org.openembedded.dev * r2e90d22a16 10openembedded.git/recipes/linux/linux-rt_2.6.24.bb: Apr 27 19:08:30 linux-rt-2.6.24: Apply latest -rt patch for all, except mpc8315e-rdb. Apr 27 19:08:30 mpc8315e-rdb is kept to the Freescale release, as it has numerous Apr 27 19:08:30 patches applied that don't mix with latest -rt. Apr 27 19:08:30 Signed-off-by: Leon Woestenberg Apr 27 19:08:32 03Leon Woestenberg  07org.openembedded.dev * rcc2046eec0 10openembedded.git/recipes/ltp/ (ltp-20090131/syscalls.patch ltp_20090131.bb): Apr 27 19:08:37 ltp-20090131: Added upstream fix for missing non-x86 syscalls. Apr 27 19:08:39 This fixes errors for powerpc targets (and likely other targets as well), Apr 27 19:08:41 such as: Apr 27 19:08:43 dup3_01.c:74:4: error: #error "need __NR_dup3" Apr 27 19:08:45 Signed-off-by: Leon Woestenberg Apr 27 19:08:47 03Leon Woestenberg  07org.openembedded.dev * rf67b2104da 10openembedded.git/recipes/madwifi/ (3 files in 2 dirs): madwifi-ng_r3837: Use LDFLAGS to fix QA GNU_HASH issue. Apr 27 19:08:50 03Leon Woestenberg  07org.openembedded.dev * r9e2951142f 10openembedded.git/recipes/madwifi/madwifi-ng_r.inc: madwifi-ng: Recipe whitespace fix. Apr 27 19:09:03 [balister@elephant oe]$ bitbake gcc Apr 27 19:10:11 I guess I stoppped, revert and restarted before 4.3.x got going Apr 27 19:10:22 I saw a ton of failures w/ 4.1.2 and 4.2.4 Apr 27 19:11:41 hello guys, just jumping in here. What's the status on GCC? Can I do a test run of some sorts? Apr 27 19:12:11 try building gcc for a ppc target, maybe Apr 27 19:12:20 pretty sure they aren't build 4.3.3 Apr 27 19:12:27 add ANGSTROM_GCC_VERSION to your EXTRAWHITE and pass it in, for any target :) Apr 27 19:12:47 I thought only armv7a was forcing 4.3.x in angstrom Apr 27 19:13:01 some ppcs are forcing lower which I don't have the time to check on being right or wrong Apr 27 19:13:10 Anybody using this stuff with a SheevaPlug? Apr 27 19:13:25 likewise, does some ppc work I think .... Apr 27 19:13:26 also, really, lunch, afk while i figure out where to get some today Apr 27 19:13:45 ah yes, he's in #mklinux, that's right :) Apr 27 19:13:53 really, i mean it this time, getting up Apr 27 19:15:39 so, gcc for 4.2.x first? I think my powerpc targets should work with gcc 4.3.x too, I built the OSELAS/ptxdist toolchain to verify which powerpc patches they use. Apr 27 19:16:33 SDuensin: yes, for those who have received their plugs Apr 27 19:16:44 SDuensin: timtimred, koen have plugs. I get mine tomorrow Apr 27 19:16:49 * SDuensin is bitbaking it now. Apr 27 19:16:52 * SDuensin has a plug. Apr 27 19:16:59 * likewise_ figured as much :-) Apr 27 19:17:39 ERROR: Nothing PROVIDES 'console-iamge' Apr 27 19:17:40 I was just going to stick with Ubuntu, but the distro that comes with it is kind of lacking. Apr 27 19:17:40 leon@witty:~/sandbox/mpc8315e/openembedded$ bitbake console-aamge Apr 27 19:18:06 sigh, need more coffee and/or more sleep Apr 27 19:18:07 my last powerpc platform i was using 4.2.x... sometime ago Apr 27 19:18:15 timtimred: on oE? Apr 27 19:18:17 yes plugged up here Apr 27 19:18:19 yep Apr 27 19:18:32 lsppchg/lsppchd/n1200 (all in OE) Apr 27 19:18:44 * SDuensin wishes he knew more about oE. Uphill work learning it all. Apr 27 19:19:03 SDuensin: it's worth the learning curve once you hit the base camp Apr 27 19:19:16 timtimred: see bottom of page: http://www.pengutronix.de/oselas/toolchain/index_en.html Apr 27 19:19:26 So far, I'm really impressed with it. Apr 27 19:19:40 Pretty sure it was timtimred who got me hooked on it. Apr 27 19:19:42 timtimred: gcc-4.3.2 glibc-2.8 linux-2.6.27 Apr 27 19:19:55 SDuensin: yeah he is a good ambassador Apr 27 19:20:38 woah. i think my last build was 4.2.x and glibc-2.6 and whatever version of linux is in OE for those platforms Apr 27 19:20:46 so it was a while Apr 27 19:21:15 maybe it was glibc 2.7 - but i remember having patches for that as the stock OE one didnt work at the time Apr 27 19:21:42 whats the powerpc platform just out of interest? :) Apr 27 19:24:52 timtimred: Freescale's mpc8313e-rdb and mpc8315e-rdb and AMCC's 460EX Canyonlands board. Before that, EFIKA, but it died on me. Still need a replacement. Apr 27 19:24:53 * timtimred notes to update all his NAS platform support in OE sometime soon... Apr 27 19:25:48 cool you have a newish kernel for that too? Apr 27 19:25:54 lol, we skipped glibc 2.8 in oe? Apr 27 19:35:44 re Apr 27 19:35:57 wb Apr 27 19:37:55 mickeyl: should libxml-native depend on python-native? Apr 27 19:38:07 it looks like Apr 27 19:38:18 let me check Apr 27 19:38:23 ah.. it depends Apr 27 19:38:35 but checks at parse phase Apr 27 19:39:02 at least in stable Apr 27 19:39:05 well, there's still this chicken-and-egg problem Apr 27 19:39:16 we just ignored it until now Apr 27 19:39:22 and it doesn't really hurt Apr 27 19:48:24 likewise, I think I asked the same question and got the answer, yes it is an abuse :) Apr 27 19:59:17 Crofton|work: do we have a better solution? Apr 27 19:59:53 I think you can patch the build system to respect LDFLAGS Apr 27 19:59:57 or use the hack :) Apr 27 19:59:57 likewise_: patch each makefile Apr 27 20:08:42 anyone have any pointers for building on jaunty? I have a gcc error configure: error: C preprocessor "i486-angstrom-linux-gcc -E" fails sanity check Apr 27 20:09:02 actually it's glibc.. Apr 27 20:10:31 What machine/distro? Apr 27 20:10:53 Just built meta-toolchain for an arm target on jaunty Apr 27 20:10:58 MACHINE = "x86" Apr 27 20:11:07 DISTRO = "angstrom-2008.1" Apr 27 20:11:09 But it was an upgrade from 8.10 that I had all the needed dev stuff on Apr 27 20:11:21 x86 build is in the queu Apr 27 20:11:22 e Apr 27 20:11:27 Tartarus: this is a clean 9.04 Apr 27 20:11:42 Tartarus: oh, you're trying to build it now? Apr 27 20:11:49 toggles_w, lemme go check the buildbot Apr 27 20:11:58 toggles_w: share config.log please Apr 27 20:12:00 Tartarus: thanks mate, appreciated Apr 27 20:12:17 hrw: for glibc? Apr 27 20:12:36 toggles_w: for failed recipe Apr 27 20:12:47 vortex86sx / angstrom-2008.1 on stable is going right now Apr 27 20:13:10 atngw100/angstrom-2008.1/uclibc on my machine is going now Apr 27 20:13:18 hrw: http://pastebin.com/m896dd0a Apr 27 20:13:58 kdevelop with kexecboot code on my machine is going now :) Apr 27 20:14:01 toggles_w: share *config.log* please (not do_compile log) Apr 27 20:17:09 hrw: sorry, http://tommywatson.com/config.log.gz which is tmp/work/i486-angstrom-linux/glibc-2.6.1-r15/build-i486-angstrom-linux/config.log Apr 27 20:19:09 /hacking/resd/baobab/openembedded/build/tmp/staging/i486-angstrom-linux/usr/include/limits.h:125:26: error: no include path in which to search for limits.h Apr 27 20:20:14 no idea Apr 27 20:21:00 hrw: thanks Apr 27 20:23:21 toggles_w: try to rebuild glibc-initial and glibc Apr 27 20:25:41 kergoth: I hitorically keep BB_DEFAULT_TASK = "buildall"...nonsense today? Apr 27 20:25:49 Jay7: ok, thansk Apr 27 20:32:44 Are there any notes on how to add a new toolchain? 4.2.3+csl-arm-2008q1-126 is giving me assembler errors in some packages Apr 27 20:33:03 Trying to use that on x86? Apr 27 20:33:13 er, sorry Apr 27 20:33:14 nm :) Apr 27 20:33:25 hoj, what MACHINE again? Apr 27 20:33:39 omap3evm Apr 27 20:33:52 What packages? Apr 27 20:34:17 And can you try current .dev and gcc 4.4.0? It should fail as well Apr 27 20:34:19 I hope :) Apr 27 20:34:20 One failure I saw was in libxml Apr 27 20:35:56 Tartarus: You mean check out different branch of oe? Apr 27 20:36:16 What branch are you using? Apr 27 20:36:21 stable/2009 ? Apr 27 20:37:49 got source with "git clone git://git.openembedded.net/openembedded" Apr 27 20:38:04 Did you do a checkout of stable? Apr 27 20:38:08 If not, then you're in dev Apr 27 20:38:12 do a git pull Apr 27 20:38:15 and then try 4.4.0 Apr 27 20:38:30 Also, I'll try and kick off a build of that stuff soon Apr 27 20:38:49 How to specify 4.4.0 (in local.conf)? Apr 27 20:40:35 bye all Apr 27 20:43:54 Same way you did 4.2.3+... Apr 27 20:44:46 Tartarus: with PREFERRED_VERSION_gcc* = "4.4.0" or is there a suffix on the version number? Apr 27 20:44:57 just 4.4.0 since it's a normal upstream release Apr 27 20:45:01 not a wierd csl one Apr 27 20:45:14 got it. Will try... Apr 27 20:46:20 03Tom Rini  07trini/add-libc-overrides * r174880c8c8 10openembedded.git/conf/distro/ (minimal-uclibc.conf minimal.conf): minimal and minimal-uclibc: Rework order so both work now. Apr 27 21:00:07 crap.. i had a gcc segfault that it manages to get passed on a restart.. bad memory? Apr 27 21:00:23 Tartarus: ERROR: Error in executing: /home/leon/sandbox/sheevaplug/openembedded/openembedded.git/recipes/gcc/gcc_4.2.4.bb Apr 27 21:01:06 Tartarus: was a fresh build, started ~1 hour ago. Dunno if this gives any info. Apr 27 21:01:25 likewise_, trying to build on a ppc target? interesting :| Apr 27 21:01:41 Tartarus: you full time on OE maintenance jobwise? Apr 27 21:01:47 Tartarus: no, that was on x86 host Apr 27 21:02:07 no, not exactly :) Apr 27 21:02:41 Tartarus: your patch rate is high :-) Apr 27 21:12:34 03Andrea Adami  07org.openembedded.dev * r86d6684a9c 10openembedded.git/recipes/kexecboot/initramfs-kexecboot-image.bb: initramfs-kexecboot-image: just deploy cpio.gz Apr 27 21:12:35 03Andrea Adami  07org.openembedded.dev * ra30b7f8b66 10openembedded.git/conf/machine/include/zaurus-2.6.inc: zaurus-2.6.inc: deploy installkit only for kexecboot kernel Apr 27 21:12:36 03Andrea Adami  07org.openembedded.dev * r25aaf46886 10openembedded.git/conf/machine/include/zaurus-clamshell.inc: zaurus-clamshell.inc: fix u-boot addresses for akita/spitz Apr 27 21:14:45 Tartarus: likewise: we have a potential contributor wo was hit by bad ppc support in OE... Apr 27 21:14:59 flameman: ping :-) Apr 27 21:15:36 problems starting with m4-native iirc Apr 27 21:16:08 ant__: you mean, bad powerpc *build host* support? Apr 27 21:16:13 yep Apr 27 21:16:27 ah..target..I see Apr 27 21:16:30 I'm interested in powerpc targets. I know, bad excuse. Apr 27 21:17:47 Tartarus: building gcc for powerpc targets fails the same way, as expected. Apr 27 21:17:57 Odd, OK Apr 27 21:18:12 Should bug segher went I have time i guess :) Apr 27 21:18:19 s/went/when/ Apr 27 21:19:05 wtf 'segher'? italian dialect? Apr 27 21:19:17 ;-) Apr 27 21:37:04 Hmmm Apr 27 21:37:14 Is there a reason licenses.conf doesn't have 'GPLv3' and 'LGPLv3' ? Apr 27 21:41:55 no one addded them? Apr 27 21:49:37 * Jay7 -> sleep Apr 27 21:50:55 OK Apr 27 21:51:03 Now is there a policy on setting LICENSE? :) Apr 27 21:51:13 * Tartarus is checking out some packages for work atm Apr 27 21:57:08 ehm..for adding some packages to a machine, better IMAGE_EXTRA_, MACHINE_EXTRA_DEPENDS or MACHINE_EXTRA_RDEPENDS + EXTRA_IMAGEDEPENDS (bd-neon.conf) ?? Apr 27 21:57:18 some packages = mtd-utils Apr 27 21:57:44 isn't there some feature or another for that already? Apr 27 21:57:51 another thing on my overly long todo list Apr 27 21:58:43 hey woglinde Apr 27 21:59:08 jo ant Apr 27 22:14:19 03Angus Ainslie  07fso/milestone5.5 * ref389e75d0 10openembedded.git/recipes/linphone/linphone-3.1.0/preferences-segv.patch~: linphone : remove backup file Apr 27 22:26:21 03Andrea Adami  07org.openembedded.dev * rb03c7b8bb9 10openembedded.git/conf/machine/include/zaurus-2.6.inc: zaurus-2.6.inc: fix typo in installkit Apr 27 23:07:56 Tartarus, in all seriousness,. if there is a question, make patch and send it to the list Apr 27 23:08:17 be very explicit that you will push if no one objects in a few days Apr 27 23:10:12 k Apr 27 23:10:20 I'll go push adding GPL/LGPLv3 now Apr 27 23:11:34 03Andrea Adami  07org.openembedded.dev * r41c855964b 10openembedded.git/ (conf/checksums.ini recipes/linux/linux_2.6.29+2.6.30-rc1.bb): linux_2.6.29+2.6.30-rc: move to rc3 Apr 27 23:12:49 03Tom Rini  07org.openembedded.dev * radfc2e343c 10openembedded.git/conf/licenses.conf: licenses.conf: Add strings for GPL/LGPLv3 and LGPLv2.1 Apr 27 23:16:59 03Andrea Adami  07org.openembedded.dev * r1b6bdbe0ee 10openembedded.git/recipes/kexecboot/kexecboot-cfg_0.1.bb: kexecboot-cfg: fix tabs and typo Apr 27 23:32:54 likewise, around still? Apr 27 23:34:43 likewise, looking for amcc460ex/stable eta, for my buildbot config Apr 27 23:39:31 And, yay, first run did fine, updating with bug to cover armv6 and adding canadian-sdk for armv5te/v6/ppc300c3 (as i'm just doing angstrom atm adn those are the 4.2.4-based targets) Apr 27 23:39:40 uclibc full build kicks off in a few hours :) Apr 27 23:55:34 Tartarus: still doing build w/4.4.0. I noticed that libx11 just built for target console-image Apr 27 23:56:49 hoj, yeah, but does it insall is the question Apr 27 23:57:04 and, whee, I'm gonna take me a risk and see if I can't just get a dev build going first time :) Apr 27 23:57:12 Will know in 2h or so, heh Apr 27 23:57:43 Hm Still seems like a waste of time to build Apr 27 23:58:17 Agreed Apr 27 23:58:26 OE doesn't care about wasting your host system disk or CPU Apr 27 23:58:36 It does care about wasting your target 'disk' space Apr 27 23:59:22 oe does the job it is designed to do Apr 28 00:00:18 Crofton|work, on my todo list is to try and argue that over-building is a distro choice / qa issue Apr 28 00:00:19 we really need to develop a good solution for people who want to do sw dev based on OE built imagees Apr 28 00:00:26 without them needing to learn OE Apr 28 00:00:46 meta-toolchain / sdk Apr 28 00:00:49 I guess any further grumbling will be answered with "patches accepted" :) Apr 28 00:00:54 my argument is we need to extract quality toolchains Apr 28 00:01:12 where wulaity includes ease of use by develoeprs Apr 28 00:01:18 hoj, no, disabling building X11 in some recipe or another will be replied to with "making this configurable is QA hell on packages" Apr 28 00:01:20 anyway dinner time Apr 28 00:08:54 hoj: part of the issue is what you are building; OE only *seems* to be Angstrom. Apr 28 00:09:42 SlugOS also builds in OE; it is what I would consider a console-image-like distro, builds in 30 mins, and fits in an 8MByte flash. Apr 28 00:10:09 So not everything in OE insists upon building X libs... Apr 28 00:10:38 (that's the Dark Lord of the Angstrom's choice, and I do not bow to that entity) Apr 28 00:11:26 :-) Apr 28 00:11:30 So if I really really wanted to skip X, I'd use another distro make a new distro Apr 28 00:11:59 Depends on your needs. Apr 28 00:11:59 hoj, or pick a different image target Apr 28 00:12:45 is there a way to list all available targets and their contents (w/o building)? Apr 28 00:13:09 Sadly no. Apr 28 00:13:32 You have to at least set up, and do the parsing phase Apr 28 00:13:52 03Rod Whitby  07org.openembedded.dev * r694436f224 10openembedded.git/recipes/ (110 files in 16 dirs): Apr 28 00:13:52 ixp4xx: Remove proprietary Intel ethernet driver. Apr 28 00:13:52 See Apr 28 00:13:52 http://lists.linuxtogo.org/pipermail/openembedded-devel/2007-January/001279.html Apr 28 00:14:00 So once you set up to build a particular image, you can then do "bitbake -g " Apr 28 00:14:00 03Rod Whitby  07org.openembedded.dev * r51733906e1 10openembedded.git/: Merge branch 'org.openembedded.dev' of git://git.openembedded.net/openembedded into org.openembedded.dev Apr 28 00:14:21 That will create a dependency graph (depends.dot) that you can use to see what all got pulled in. Apr 28 00:14:33 (it's sorta broken right now, but still usable). Apr 28 00:15:31 I started with the Useful_targets wiki page. mwester: thanks, I'll try that flag Apr 28 00:16:01 whatcha building? Apr 28 00:17:13 Trying gcc 4.4.0 toolchain for omap35xevm board. Eventually want a filesystem with GLX drivers Apr 28 00:17:41 cool. Apr 28 00:17:53 Hey gang. I'm missing something I feel is likely obvious... Apr 28 00:18:04 I've built angstrom and got an image... Apr 28 00:18:04 I mean SGX (OpenGL/OpenVG) Apr 28 00:18:20 Now I built "mediatomb" which has a lot of dependencies... Apr 28 00:18:36 How do I bundle up all the stuff it built to install it onto my Angstrom drive? Apr 28 00:19:41 turn the deploy/.../ipkg/ dir into a feed, and use opkg to install mediatomb on the target' Apr 28 00:19:59 Whoa. Was that in English? :-) Apr 28 00:20:13 Did I miss a page in the wiki? Apr 28 00:20:21 Or you can copy the entire ipkg dir to the target, or mount it via nfs, and opkg install all the packages. Apr 28 00:20:27 * SDuensin doesn't mind reading to figure it out, but he didn't see anything. Apr 28 00:25:00 Tartarus: my gcc4.4.0 build stopped, but I hit the same assembler errors in libxml "Error: offset out of range" Apr 28 00:31:06 good :) Apr 28 00:31:09 libxml bug Apr 28 00:31:14 or compiler bug Apr 28 00:31:29 i bet if you change the optimization for libxml it'll go away Apr 28 00:31:32 talk w/ khem :) Apr 28 00:34:40 * SDuensin is just amazed at the number of files this thing generates. Apr 28 00:38:29 Tartarus: I don't see cflags in libxml2.inc If I were to change them, that's where it would be added, right? Apr 28 01:30:11 mwester, slugos also does not support lots of machines Apr 28 01:53:49 hoj, added there, yes Apr 28 01:54:05 Crofton|work: that is true, but console-image does not really require X11 on _any_ machines, does it? Apr 28 01:54:18 hoj, cortex-a8 based, right? If so: Apr 28 01:54:19 TARGET_CC_ARCH_armv7a = "" Apr 28 01:54:40 Crofton|work, but how much does a distro _need_ to do to support a machine? Apr 28 01:54:40 I'm fairly certain that I could add support for a whole raft of other machines, without requiring X11. :) Apr 28 01:55:02 (And I'll do that soon -- will be supporting the Sheeva Plug in a week or two) Apr 28 01:55:11 Aside from "bleeding edge" targets that need specific compiler, there's not really any, so long as the distro in question obeys the "normal" hooks for adding machine specific stuff Apr 28 01:58:10 hrw|gone, mckoan|away, is vortex86sx + uclibc a valid combo? Fails in dev, haven't run stable yet, and it's not in hrw's buildbot Apr 28 01:58:11 I suspect that the problem for console-image is that includes something that uses dbus, and dbus pulls in X11. That can be fixed --- if someone cares to. Apr 28 01:58:35 mwester, the question is, is it build time only or run-time pulled in too? Apr 28 01:58:58 According the Koen, it is run-time, if it is the dbus/x11 issue. Apr 28 01:59:08 mwester, I really need to find the time to see if there's a case where we pull in extra crap at image time and then argue we need a way at the distro level to turn off some deps Apr 28 01:59:36 ie for Angstrom maybe it's a big deal to have to warn users's DON'T TOUCH THIS VARIABLE AND PUBLISH YOUR BUILDS Apr 28 01:59:53 but for another distro it's a bigger deal to disable X/Y/Z random stuff Apr 28 02:01:13 And I can see his point -- if one assumes that console-image is just a stepping stone to a full image, then to alter recipes to remove the runtime dependency, or even create slimmed-down versions without build-time dependencies is not a good use of time. But for those devices that do not ever need those dependencies --- eg, headless servers or such -- then Apr 28 02:01:39 mwester, right it's a distro choice Apr 28 02:01:43 it is a problem to build needless runtime deps -- somebody has to maintain that part of the build. Apr 28 02:01:49 Angstrom needs to have packages that can work everywhere Apr 28 02:01:58 But not all distros want that Apr 28 02:02:06 :) Precisely. Apr 28 02:02:14 ie if I know my stuff will never be used on bluetooth-enabled HW, I don't want that pulled in Apr 28 02:02:27 Just need to find the time to propose it all in a sane way Apr 28 02:02:36 And I want PREFERRED_LIBC and a few others in first :) Apr 28 02:03:11 I suspect the point of resistance will be that any solution would smell of "use" flags (I think that's the term). Apr 28 02:03:22 Right Apr 28 02:03:29 And fundamentally, that's the correct thing to have Apr 28 02:03:36 IOW, people that don't want it are wrong :) Apr 28 02:03:51 and if we have multiple recipes (look at opkg for an example of how horrible that can get, a problem I am responsible for creating, in fact). Apr 28 02:04:07 Exactly Apr 28 02:04:28 We do want a use flag, at the distro level with big fat warnings to not mess with this unless you're the distro maintainer Apr 28 02:05:18 I have a feeling that virtual/ will end up fitting in here somehow? Apr 28 02:05:48 I don't think so atm Apr 28 02:06:09 That's just giving us the 50 permutations of the recipes people want, or don't want, random stuff enabled in Apr 28 02:06:40 We want USE="bluetooth X11 directfb" and a big warning so that people don't put up random angstrom packages with non-standard USE flags Apr 28 02:07:21 And in the recipes, we then need all sorts of anonymous python stuff, unless we get expanded support for conditionals from bitbake. Apr 28 02:07:22 And then maybe argue that no, people shouldn't start out with Angstrom for their weird custom project, if they want to fiddle with USE flag at least, then they want to start with minimal Apr 28 02:32:09 Hmmm Apr 28 02:32:16 PPC isn't skipping out on the GNU_HASH thing is it? Apr 28 02:32:21 That's just MIPS that can't, right? Apr 28 02:36:51 yeah, there's my failure :) **** ENDING LOGGING AT Tue Apr 28 03:00:02 2009