**** BEGIN LOGGING AT Mon Feb 14 02:59:57 2011 Feb 14 03:42:17 03Mike Westerhof  07org.openembedded.dev * ra949dad178 10openembedded.git/conf/distro/ucslugc.conf: SlugOS: ucslugc -- switch to use standard opkg. Feb 14 05:17:13 morning Feb 14 06:15:26 Hey OE folks. I picked a project back up, nuked my OE sources, cleaned everything up, and began from scratch. Now, with a clone I did just minutes ago, I cannot get even the most simple build to complete. Lots of complaints about recipies in the obsolete directory and it then fails with "ParsingErrorsFound"... is this just the state of the repo or should I dig deeper for clues? Feb 14 06:44:06 my oe stucks at "task do_fetch: Started" when "bitbake nano"... Feb 14 06:57:50 hi Feb 14 06:58:13 coreyfro: you should remove the obsolete dir from your BB_PATH Feb 14 07:01:25 HeinervdmOff: Just delete it from the recipies or mask it? Feb 14 07:13:09 03Khem Raj  07master * rd808a0dc5c 10openembedded.git/classes/python-dir.bbclass: Feb 14 07:13:09 python-dir.bbclass: Add checks for 2.7 and 3.1 as well Feb 14 07:13:09 Signed-off-by: Khem Raj Feb 14 07:17:21 Ok, killing that directory and "nonworking" unstuck the process Feb 14 08:18:02 morning Feb 14 08:18:04 ka6sox: pong Feb 14 08:18:24 hi Jay7 Feb 14 08:18:44 lzma-native race was not fixed :( Feb 14 08:18:58 I and ant have seen it again Feb 14 08:20:08 Jay7, did your patch to slow down the number of http Puts by a builder get accepted and implemented? Feb 14 08:20:35 interesting build statistics: http://pastebin.com/0DQES7FL Feb 14 08:20:52 ka6sox: no, still only idea Feb 14 08:21:46 well, its abusive of the server infrastructure... Feb 14 08:22:07 I will benchmark builds with different PARALLEL_MAKE and BB_THREADS_NUMBER Feb 14 08:22:13 ka6sox: yeah.. Feb 14 08:22:51 with that many builders pounding on the servers this wasn't intended I hope to be a multiple builder single repo thing...because its poorly implemented from a resources standpoint. Feb 14 08:23:38 ka6sox: I'll try to speedup things... but I should time to kexecboot before.. Feb 14 08:24:33 okay, when other users of the resource start noticing that they are slowing down then its time to get it fixed. ASAP. Feb 14 08:25:49 hm.. I forgot word 'spend' in previous phrase Feb 14 08:26:08 ka6sox: we should fix it anyway :) Feb 14 08:27:15 Jay7, I understand that things take time...my issue is when you have a shared resource and one starts taking more than their share then it upsets the others. Feb 14 08:27:39 ka6sox: bother me then ;) Feb 14 08:28:41 okay, I need to stay on this till its fixed or we will have to turn it off till it is fixed. Feb 14 08:30:53 good morning Feb 14 08:32:41 Jay7, I'm not just picking on you...I had another project that was burning up 4Mb of traffic constantly polling a site and rewriting the feeds from that site. Feb 14 08:32:54 every 4 minutes Feb 14 08:33:01 hehe Feb 14 08:33:13 ya... Feb 14 08:33:41 now its every 4hrs. Feb 14 08:35:45 we need some python-skilled enthusiast Feb 14 08:41:24 bitbake fails : Error in compiling python function Feb 14 08:41:54 http://pastebin.com/yGaHX293 Feb 14 08:42:06 any clue? Feb 14 08:42:59 BitBake Build Tool Core version 1.10.2, bitbake version 1.10.2 Feb 14 08:43:06 bb.debug(1, "Fetched staging package %s" % (bn)) have wrong spacing? Feb 14 08:43:18 so it going out of block Feb 14 08:43:29 then exception have no opening try Feb 14 08:43:33 (just my ideas) Feb 14 08:49:44 Jay7: do you mean indentation? Feb 14 08:49:50 mckoan: yes Feb 14 08:50:03 Jay7: looks is not intentation issue Feb 14 09:03:54 Jay7: the problem is except Exception as ex: Feb 14 09:04:42 commit fe706bb9af34ce43e8ec55fbf929a85d3b6d612a Feb 14 09:04:52 Author: Bernhard Reutner-Fischer Feb 14 09:05:03 Signed-off-by: Tom Rini Feb 14 09:06:21 I modified and solved the prioblem Feb 14 09:06:48 * mckoan is preparing a patch to ML Feb 14 09:16:09 03Marco Cavallini  07master * r6a97f8c4f9 10openembedded.git/classes/packaged-staging.bbclass: Feb 14 09:16:09 packaged-staging.bbclass: partial revert of commit fe706bb9af34ce43e8ec55fbf929a85d3b6d612a due to build error Feb 14 09:16:09 * # ERROR: Error in compiling python function in: Feb 14 09:16:09 * # ERROR: The lines leading to this error were: Feb 14 09:16:09 * # ERROR: 22:SyntaxError:' except Exception as ex:' Feb 14 09:16:11 push is faster Feb 14 09:17:19 yep, in case of trivial fixes that's the way Feb 14 09:18:13 it's monday morning and I haven't time to prepare a ML patch ;-) Feb 14 09:41:58 any thoughts on the dependencies problem from the ML? Feb 14 09:56:13 *yawn* Feb 14 09:56:17 Morning all Feb 14 09:56:32 stefan_schmidt: moin Feb 14 09:57:13 hi mckoan Feb 14 09:57:15 hi florian Feb 14 10:15:23 good morning Feb 14 10:18:41 florian: moin^2 Feb 14 10:19:12 hi mckoan Feb 14 10:24:40 filip: sent another issue with dependencies/debian.bbclass to ML, check if it's somehow related to your Feb 14 10:49:59 JaMa|Off: do you add ASSUME_PROVIDED += "linux-libc-headers-native" building OE on Gentoo? Feb 14 10:53:00 ant_work: no Feb 14 10:53:23 ok, neither me but then I have random failures, now with curl-natie Feb 14 10:53:41 I'm reading now msgs in the ML (June 2010) Feb 14 10:53:46 it was 'debated' Feb 14 10:54:04 * ant_work fell in the trap Feb 14 10:54:59 ant_work: I can share my lxc container image for you :) Feb 14 11:04:49 ok, it seems indeed the sane default is assuming the buildhost headers are sane Feb 14 11:05:29 Tartarus fixed build RHEL4 which apparently has too old headers Feb 14 11:05:52 now, this is all fine but exposes to the risk of races...yes, headers can take long time to stage Feb 14 11:36:07 JaMa|Off: this looks crazy Feb 14 11:46:04 this is crazy Feb 14 12:11:40 I want to add my .vimrc to package, but I get: Feb 14 12:11:41 NOTE: the following files were installed but not shipped in any package: Feb 14 12:11:41 NOTE: /home/root/.vimrc Feb 14 12:12:07 it is in 'image', 'package' directory in correct place, but it is not included in 'packages-list' Feb 14 12:12:53 /s/packages-list/packages-split/ Feb 14 12:24:24 davidlt: why not /etc/vim/vimrc? Feb 14 12:24:38 davidlt: and how do you know that user's homedir will be /home/root? Feb 14 12:24:59 Location doesn't matter much there is only one user, root Feb 14 12:25:19 davidlt: then add right location to FILES_${PN} variable (and probably also CONFFILES_${PN} Feb 14 12:25:27 davidlt: and what about root in /root? Feb 14 12:26:16 I always putted .vrimrc after flashing to /home/root with all other files Feb 14 12:26:49 But how to write correct path to .vimrc ? Feb 14 12:27:19 Whatever I write, it is still not added Feb 14 12:27:58 ${D}/home/root/.vimrc , home/root/.vimrc Feb 14 12:32:59 JaMa|Off: Any suggestions? Feb 14 13:03:25 davidlt: FILES_${PN} += "{sysconfdir}/vim/vimrc" Feb 14 13:03:33 with $ Feb 14 13:03:37 davidlt: FILES_${PN} += "${sysconfdir}/vim/vimrc" Feb 14 13:03:41 JaMa|Off: Already fixed Feb 14 13:04:05 Yha, got wireless access point on my Gumstix running stable! Feb 14 13:08:55 nnnnnn1qqqqqwqqqqqqqqqqqqqqqqqqqwuuuuuuuuuuuuuuuwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu Feb 14 13:08:56 uuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu Feb 14 13:08:57 uuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu Feb 14 13:09:01 uuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu Feb 14 13:09:08 uuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuwwwwwwww Feb 14 13:11:36 * broonie wonders if someone has a cat. Feb 14 13:53:06 or a 2 year old Feb 14 13:56:10 tzanger: hmmmm... a 2 year old pushes more keys randomly :_D Feb 14 14:03:41 depends on if he's liking the result :-) Feb 14 14:07:18 hehe Feb 14 14:19:28 hi mickey|office Feb 14 14:21:30 hey pb_! Feb 14 14:21:34 how are things these days? Feb 14 14:40:04 mickey|office: good, thanks. building work nears completion which is nice, and I am enjoying my new job. Feb 14 14:44:06 new job? Feb 14 14:44:29 well, "new". I started about four weeks ago now. Feb 14 14:44:48 www.brightsign.biz Feb 14 14:47:42 ah, nice Feb 14 14:55:20 pb_, be sure to let us know what the backdoors are :) Feb 14 15:04:11 pb_: interesting, i thought you were quite a crucial part of reciva. may i ask for the reason to quit? Feb 14 15:05:12 mickey|office: just a general sense of ennui really. I'd been there for nearly seven years and it seemed like I was in a bit of a rut. Feb 14 15:05:51 pb_: right, after seven years i can imagine Feb 14 15:05:54 also, I was getting fed up with the administrative side of the job, having to deal with shareholders and lawyers and suchlike all the time Feb 14 15:06:33 *nod* Feb 14 15:24:55 pb_: still living around Cambridge? Feb 14 15:26:16 yes Feb 14 15:27:06 pb_: so this time you do rather remote work? Feb 14 15:27:14 no, we have an office in cambridge Feb 14 15:31:03 pb_: I will be in cambridge for this sun-sat Feb 14 15:31:18 ah, very good. what are you doing there? Feb 14 15:32:55 pb_: Emdebian sprint at ARM office Feb 14 16:02:51 shareholders and lawyers can be annoying .... Feb 14 16:11:04 guh, the bitbake manual doesn't even mention flags? Feb 14 16:11:13 what the hell. Feb 14 16:20:22 is there a uClibc (MMU-less) machine working in OE? I am trying ekiga but is broken Feb 14 16:25:58 Crofton|work, too true Feb 14 16:31:21 03Tom Rini  07master * r52145ecccc 10openembedded.git/recipes/linux/linux_2.6.28.bb: Feb 14 16:31:22 linux 2.6.28: Add name param to ronetixpatch URL. Feb 14 16:31:22 Signed-off-by: Tom Rini Feb 14 16:33:19 what's Tom Rini's nick Feb 14 16:33:37 Tartarus Feb 14 16:33:44 yes, hello Feb 14 16:33:55 qt4 issue or that push? Feb 14 16:35:16 Tartarus: that push, I tried to solve it this morning but I didn't find the problem Feb 14 16:35:21 Ah Feb 14 16:35:21 Tartarus: thank you Feb 14 16:35:24 yeah, np Feb 14 16:35:38 Tartarus: I saw your bot logs on tinderbox ;-) Feb 14 16:37:31 Tartarus: BTW, why is required ;name=ronetixpatch now? Feb 14 16:38:19 So, that's how you specify a name for part of a URI Feb 14 16:38:30 it's the second half of SRC_URI[name.md5] = "md5sum" Feb 14 16:38:56 Tartarus: aha, for md5sum then Feb 14 16:39:08 Tartarus: thx Feb 14 16:47:39 mckoan, what version of python do you have? Feb 14 16:50:51 Tartarus: the one prvided by Debian lenny 5.0 is Python 2.5.2, why? Feb 14 16:57:02 trying to figure out your last commit :) Feb 14 16:59:33 I thought bitbake 1.10 required 2.6, but I guess that's still 2.5 Feb 14 17:01:37 Tartarus: bitbake-1.10.2 works smoothly with Python 2.5.2 Feb 14 17:01:51 Huh, I thought it was 2.6 as well.. Feb 14 17:02:45 have a nice rest of the day Feb 14 17:04:42 Hello. Did someone made any qt4 based sdk recently? Feb 14 17:05:05 Nope, and no one asked for me to add it to my list, either :) Feb 14 17:06:34 It seems it fails. Feb 14 17:06:39 howdy. trying to track down why nano won't build with prebuilt toolchain.... doesn't look quite as simple as I expected. the do_configure() step saw ncurses.h, but said no for get_wch and initscr Feb 14 17:06:42 It tried to install cdparania Feb 14 17:07:04 and there's no cdparanoia binary Feb 14 17:07:20 the libraries are in tmp/sysroots/* and I checked with file and they're the correct types (arm and x86 respectively) Feb 14 17:07:29 it seems a kind of uglyness due -dev packages Feb 14 17:08:20 oh... and nm says the symbol is actually in there. any ideas how to diagnose why automake didn't see it? Feb 14 17:09:37 otavio: I've answered your mail on the ML :) Feb 14 17:10:28 filip: which one? Feb 14 17:11:08 demigod2k, your external-toolchain recipe needs more work is the answer Feb 14 17:11:20 * Tartarus builds a lot of stuff, quite often, with external-toolchain-csl Feb 14 17:11:46 otavio: firmware splitting Feb 14 17:11:58 * otavio checks Feb 14 17:13:53 filip: saw it Feb 14 17:21:38 Tartarus: ahhh found it. under work in the source folder there's a config.log that has the exact source that caused it to fail. I can track it down using that now Feb 14 17:23:02 filip: I replied. Feb 14 17:23:11 filip: so .. do you see my argument now? Feb 14 17:23:36 filip, ping Feb 14 17:23:55 http://patches.openembedded.org/patch/569/ doesn't apply now Feb 14 17:24:00 otherwise I do like it :) Feb 14 17:24:58 03Andreas Müller  07master * r907bd09fef 10openembedded.git/recipes/clutter/clutter_1.4.2.bb: Feb 14 17:24:58 clutter: get dependency to json-glib to work Feb 14 17:24:58 * Error on fresh tmp build was Feb 14 17:24:58 * | checking for installed JSON-GLib... configure: error: not found Feb 14 17:24:58 * | ERROR: Function do_configure failed Feb 14 17:24:59 Signed-off-by: Andreas Mueller Feb 14 17:24:59 Signed-off-by: Tom Rini Feb 14 17:25:15 03Filip Zyzniewski  07master * r41e0fbf792 10openembedded.git/ (10 files in 5 dirs): (log message trimmed) Feb 14 17:25:15 deb packages support: switch from /var/dpkg to /var/lib/dpkg Feb 14 17:25:15 OE generally tries to use /var/dpkg, but the change is missing Feb 14 17:25:15 in some places, so it is needed to move completely to one of the Feb 14 17:25:16 options. Keeping the dpkg's default /var/lib/dpkg causes less Feb 14 17:25:16 maintenance overhead for OE. Feb 14 17:25:17 Signed-off-by: Filip Zyzniewski Feb 14 17:25:18 03Filip Zyzniewski  07master * rb32d23948f 10openembedded.git/recipes/makedevs/makedevs_1.0.0.bb: Feb 14 17:25:18 makedevs: increase alternatives priority Feb 14 17:25:18 makedevs from busybox is installed with priority 50, so the discrete Feb 14 17:25:19 version of it should be installed with a higher one. Feb 14 17:25:19 Signed-off-by: Filip Zyzniewski Feb 14 17:25:34 Signed-off-by: Tom Rini Feb 14 17:28:08 Tartarus: pong Feb 14 17:28:36 otavio: rather not... I don't know why would you like to force a specific vendor on a user Feb 14 17:29:26 filip: if my device has only a realtek NIC why I'd provide a 3com firmware on the system? Feb 14 17:29:47 filip, if you can re-submit 595 against current, I'll push it Feb 14 17:30:19 otavio: this firmwares are mostly for removable devices Feb 14 17:30:28 It seems qt4 recipes are broken in a way that bitbake doesn't know about the binaries Feb 14 17:30:40 otavio: so how do you know if it is a realtek? Feb 14 17:30:46 Tartarus: should I be worried if configure is saying --with-libtool-sysroot is an unrecognized option? Looks like it's failing the test because configure isn't looking under tmp/sysroots/ as my sysroot (because it didn't find the -lncurses) Feb 14 17:31:07 filip: sure, but that doesn't mean everybody wants to ship the drivers for every possible removeable device. Feb 14 17:31:11 it doesn't know that libqtgui-dev is built so when I ask for it to be installed it thinks it is not existent one Feb 14 17:31:27 I guess the ideal thing (for micro at least) would be to split it per device, so each firmware file gets its own package. Feb 14 17:31:31 otavio: I mean - next step could be splitting further by vendors (or rather devices), but I wasn't interested in it so didn't do it Feb 14 17:31:36 other distros might want less granularity than that I guess. Feb 14 17:31:46 filip: specially for embedded world most of time we want to be as specific as possible and not as general as possible Feb 14 17:31:49 pb_: true, but the current version of the recipe does exactly that Feb 14 17:32:10 pb_: a 24MB blob Feb 14 17:32:35 filip: fully agree but since you're on that makes sense to fix it in a way it solves both problems Feb 14 17:32:43 filip: your and others Feb 14 17:33:48 otavio: I need to focus - I am trying to fix deb support for jlime and it's taking a lot of time. So I don't do stuff I don't need for this ATM Feb 14 17:34:26 filip, otavio, ping on the ML from me about what you're talking about :) Feb 14 17:34:37 Tartarus: which ons is 595? I got confused :P Feb 14 17:34:50 Tartarus: pw 404s on me Feb 14 17:35:00 demigod2k, it's a little hard to say :) Feb 14 17:35:02 filip, sec Feb 14 17:35:07 filip: so for me it is a NACK. Feb 14 17:35:19 http://patches.openembedded.org/patch/569/ is what i meant Feb 14 17:35:53 filip: specially because it will add a upgrade path requirement that is broken imo Feb 14 17:36:20 ya fair enough :) appears that I have the same copying like -csl.bb like you suggested but it's still looking in the sysroot where the compiler was built, so I have to force it to pass some sort of sysroot telling it to look under tmp/sysroots/* Feb 14 17:36:41 demigod2k, yeah, sysroot stuff is fun... Feb 14 17:37:13 We should be passing in, when compiling --sysroot=/OE/... Feb 14 17:37:31 However, you might also need to make sure that the marvell compiler isn't doing something funny behind your back Feb 14 17:37:53 ie with CSL toolchains, --sysroot + -te500v2 (which is a convience flag they added) it mangles sysroot beyond the normal paths Feb 14 17:38:07 correct. I see a -isystem which would make it look there for includes, but I didn't see anything else on the configure step that would force it to look there for libraries Feb 14 17:38:26 You should see --sysroot= stuff too, but lemme make sure that's not a local mod Feb 14 17:38:28 I don't think it is... Feb 14 17:39:25 OK, yeah, it is, heh Feb 14 17:39:32 denix, ping Feb 14 17:39:45 ahhh ok so I'm not crazy. thats all the confirmation I needed :) Feb 14 17:40:24 I've got a local mod to do TOOLCHAIN_OPTIONS = " --sysroot=${STAGING_DIR_TARGET}" and then clearing TOOLCHAIN_OPTIONS in cross/native/etc.bbclass Feb 14 17:40:26 otavio: well for devices with buses available outside there will always be a reason for bus-specific packages to exist Feb 14 17:40:48 where did you put that... in the local.conf? or in the *.bb for each thing? Feb 14 17:41:02 otavio: even if linux-firmware was split into vendor or product pieces - I think bus metapackages are still needed Feb 14 17:41:05 filip: not certain. I think it is most more common to have a small set of vendors supported. Feb 14 17:41:21 local.conf Feb 14 17:41:28 filip: I never wanted to have those by bus in my case Feb 14 17:41:41 otavio: and you really want only a 3com card to work in a PCMCIA socket, and not SMC for example? Feb 14 17:41:41 filip: do you really test all the possible devices for a given bus? Feb 14 17:42:01 sweet. I will try it after lunch. I can at least confirm that mod is pretty critical for an external toolchain. Appreciate you checking on it! Feb 14 17:42:09 Someone with bitbake background can give me a hand on packaging "knownledge"? Feb 14 17:42:29 filip: if 3com is all my customers have requested and all I have tested, yes. there's no point shipping SMC drivers if nobody wants them and I don't know whether they work or not. Feb 14 17:43:09 I tried to make a task that rdepends on libqtgui-dev and it doesn't work Feb 14 17:43:17 it seems that it is missing something on qt4.inc Feb 14 17:43:21 but I fail to see what Feb 14 17:43:44 otavio: but this is your case, and OE is used for free distributions as well.. Feb 14 17:44:51 unless I got something wrong and OE is only for devices with a set limit of features Feb 14 17:44:58 filip: sure. That's why I didn't deny your use case. But for me to ack the patch it needs to fit my "requirements". If someone wants to ack it so be it. I won't ack it. Feb 14 17:45:11 ah I get it now :) Feb 14 17:45:28 pb_: we have a bit different usage scenario Feb 14 17:45:41 filip: if you push it I will have to work on that and fix it for our usage. Feb 14 17:45:50 filip: I prefer if you do that since you're working on that Feb 14 17:46:04 otavio: but I've already moved on to something else Feb 14 17:46:28 otavio: you can't use linux-firmware ATM, so I don't break anything for you, right? Feb 14 17:46:39 otavio: it's just that this patch is no use for you Feb 14 17:46:47 otavio: but no harm either Feb 14 17:48:31 Tartarus: ok, I'll resend 569 Feb 14 17:48:31 filip: partially. Feb 14 17:48:59 thanks Feb 14 17:49:14 otavio: the next step would be changing the dependency tree from linux-firmware -> linux-firmware-pci to linux-firmware -> linux-firmware-pci -> linux-firmware-rtl8168 Feb 14 17:50:45 * rick-work is very sorry about the long string of "U"s last night. My cat apparently decided she liked sleeping on the laptop. Feb 14 17:52:05 gm all Feb 14 17:52:14 rick-work, good for your cat Feb 14 17:52:40 03Chris Larson  07master * rc751c19129 10openembedded.git/classes/utils.bbclass: (log message trimmed) Feb 14 17:52:41 utils: drop exported var env caching for oe_popen Feb 14 17:52:41 This was dangerous, as an oe_run executed from the configuration metadata Feb 14 17:52:41 would result in caching the config metadata information, rather than the final Feb 14 17:52:41 metadata information. We could also cache it by the datasmart instance, but I Feb 14 17:52:41 don't think its worth the trouble at this time, particularly given the next Feb 14 17:52:41 bitbake release exports all exported vars into the process environment Feb 14 17:52:51 03Chris Larson  07master * r1d6bdc44ad 10openembedded.git/classes/packaged-staging.bbclass: Feb 14 17:52:51 packaged-staging: add debug info for oe_run failures Feb 14 17:52:51 Signed-off-by: Chris Larson Feb 14 17:53:03 filip: I understand your idea. Feb 14 17:53:24 python has QA_ERRORS hmmm Feb 14 17:53:27 ERROR: QA Issue with python-sqlite3: package python-sqlite3 contains bad RPATH /scratch/oe/sysroots/armv5te-oe-linux-uclibceabi/usr/lib in file /scratch/oe/work/armv5te-oe-linux-uclibceabi/python-2.7.1-ml1.2/packages-split/python-sqlite3/usr/lib/python2.7/lib-dynload/_sqlite3.so Feb 14 17:53:43 lets see Feb 14 17:54:47 ok I see Feb 14 17:55:15 mickey|bbl: is there a provision in python to differntiate cross or native build ? Feb 14 17:57:01 khem: no, that's why we have to patch it a bit here and there Feb 14 17:58:54 Tartarus: was too excited to wait for lunch. The TOOLCHAIN_OPTIONS worked. I'll post a mailing list followup after lunch so it's documented. Feb 14 17:59:16 mickey|bbl: how about introducing a env variable Feb 14 17:59:38 mickey|bbl: the problem I reported above is because setup.py is adding -R option to ld Feb 14 17:59:50 and it just adds the dir where it finds the shared object Feb 14 17:59:56 which will not work in cross env Feb 14 18:00:13 since the objects are in sysroot Feb 14 18:00:40 and the sysroot wont eisit on target as it is Feb 14 18:00:44 demigod2k, yay Feb 14 18:00:57 khem, re the libvpx patch on patches, had a chance to debug any further? Feb 14 18:01:00 thanks for the help :) Feb 14 18:01:10 Tartarus: nope Feb 14 18:01:16 k Feb 14 18:01:27 I'd like to have a verdict by the next testing branch if we can (thurs) Feb 14 18:01:55 you can take it in for release if you like Feb 14 18:02:04 but it wont go into master without finding the real cause Feb 14 18:04:56 Tartarus: I've given the wrong message-id and the second version of the insert_feeds patch will land here: http://patches.openembedded.org/patch/567/ Feb 14 18:05:04 Tartarus: should I resend with a proper message-id? Feb 14 18:06:24 no, that's fine Feb 14 18:06:35 * kergoth ponders Feb 14 18:06:38 Tartarus: hm, gmail put it in separate thread Feb 14 18:08:56 gm Feb 14 18:10:42 How about starting to default to python 2.7.1 Feb 14 18:10:50 any opinions Feb 14 18:11:19 I know 2.7 + psyco don't mix Feb 14 18:12:02 hmm Feb 14 18:12:06 2.7 has a couple of problems wrt. support for older extension modules Feb 14 18:12:22 so i would rather prefer to skip it completely and just start with py3 in parallel Feb 14 18:14:54 mickey|bbl: ok cool Feb 14 18:15:11 so let me fix this in 2.6.5 then Feb 14 18:15:16 or .6 Feb 14 18:15:23 which JaMa is about to push or already has Feb 14 18:17:07 err we have 2.7 Feb 14 18:18:40 hi, I've a problem with the bug 1.x kernel: | calibrate.c:(.text+0x10): undefined reference to `__gnu_mcount_nc' Feb 14 18:18:49 with angstrom 2010.x Feb 14 18:18:58 any ideas? Feb 14 18:19:02 or should I dig into it? Feb 14 18:21:15 03Filip Zyzniewski  07master * r3a11e3464d 10openembedded.git/classes/ (image.bbclass rootfs_ipk.bbclass): Feb 14 18:21:15 image.bbclass: move insert_feed_uris() to rootfs_ipk.bbclass Feb 14 18:21:15 This function generates feed configuration for opkg and as such Feb 14 18:21:15 should be used only for images using ipk, not rpm or deb. Feb 14 18:21:15 Signed-off-by: Filip Zyzniewski Feb 14 18:21:16 Signed-off-by: Tom Rini Feb 14 18:21:34 GNUtoo|laptop: are you using profiling Feb 14 18:22:15 no idea Feb 14 18:22:21 I use the stock angstrom 2010.x Feb 14 18:22:25 I'll check for the options Feb 14 18:22:32 that is to say Feb 14 18:22:39 I didn't add profiling myself Feb 14 18:22:45 maybe it's there by defualt Feb 14 18:22:48 maybe it's not Feb 14 18:23:55 03Tom Rini  07master * r62f40fb9dd 10openembedded.git/classes/packaged-staging.bbclass: Feb 14 18:23:55 packaged-staging: Don't pass a trailing ' ' to bb.mkdirhier Feb 14 18:23:55 Signed-off-by: Tom Rini Feb 14 18:24:00 GNUtoo|laptop: which package is it that complains Feb 14 18:24:07 linux-bug Feb 14 18:24:26 linux-bug_2.6.27.2.bb Feb 14 18:24:46 it's a kernel for the bug device 1.x Feb 14 18:24:56 which is a machine from buglabs Feb 14 18:25:40 hmmm strange Feb 14 18:25:46 calibrate.c is part of kernel ? Feb 14 18:25:53 I think so Feb 14 18:25:55 I'll check Feb 14 18:26:12 well if kernel is wanting some function from libc thats broken from onset Feb 14 18:26:24 indeed ouch Feb 14 18:26:32 I didn't realize that Feb 14 18:26:46 sounds strange Feb 14 18:26:50 I'll doubble check Feb 14 18:26:58 ./init/calibrate.c Feb 14 18:27:27 I'll tripple check Feb 14 18:27:32 because that sounds impossible Feb 14 18:27:39 maybe I didn't look well enough Feb 14 18:28:48 http://www.pastie.org/1563514 Feb 14 18:28:52 how is that possible? Feb 14 18:28:54 mickeyl: Í have found an issue with patches I sent today with python upgrade.. .pyc are not created somehow (after next rebuild) when I tested it in the morning I had those, checking why Feb 14 18:28:59 it compiled well before Feb 14 18:29:25 GNUtoo|laptop: can you run it with V=1 Feb 14 18:29:31 and see what options are being passed Feb 14 18:29:35 ok Feb 14 18:29:38 I'll do that Feb 14 18:29:58 maybe some LDFLAGS issues Feb 14 18:30:39 because oe_runmake include some userspace LDFLAGS by default Feb 14 18:30:41 or did in the past Feb 14 18:30:52 oe_runmake include/linux/version.h CC="arm-angstrom-linux-gnueabi-gcc -mno-thumb-interwork -mno-thumb" LD="arm-angstrom-linux-gnueabi-ld " Feb 14 18:30:58 and the kernel_do_compile does that Feb 14 18:31:00 but I'll try Feb 14 18:31:16 * oe_runmake uImage CC="arm-angstrom-linux-gnueabi-gcc -mno-thumb-interwork -mno-thumb" LD="arm-angstrom-linux-gnueabi-ld " Feb 14 18:31:53 mickeyl: http://pastebin.com/f7QLZ9pY Feb 14 18:32:06 mickeyl: this patch will fix the QA error that I am seeing Feb 14 18:32:09 in python Feb 14 18:32:44 http://www.pastie.org/1563532 Feb 14 18:33:15 khem: right, that looks good Feb 14 18:33:24 I'll try with LDFLAGS="" Feb 14 18:33:31 HOST_SYS and BUILD_SYS is exactly what i'm using in the distutils patches as well Feb 14 18:34:02 mickeyl: OK cool. I will push it Feb 14 18:34:32 export LDFLAGS="-Wl,-O1 -Wl,--hash-style=gnu" Feb 14 18:34:37 there is stuff like that in the script Feb 14 18:34:40 khem: some chages I did for 2.6.6 should be applied also to 2.7.1 (like manifest etc) Feb 14 18:34:55 JaMa|Off: 2.7 is out of question Feb 14 18:35:00 export TARGET_LDFLAGS="-Wl,-O1 -Wl,--hash-style=gnu" Feb 14 18:35:03 etc... Feb 14 18:35:05 we should just stay at 2.6.6 Feb 14 18:35:09 khem: but great that you found that RPATH issue (I tried and failed - prepared 2.6.6 only as side-effect) Feb 14 18:35:17 :) Feb 14 18:35:33 khem: ok I just noticed that patch in pastebin is against 2.7 Feb 14 18:35:36 JaMa|Off: it would be better if you moved to python 3.x Feb 14 18:35:49 JaMa|Off: I just tried it with 2.6.5 Feb 14 18:35:56 works Feb 14 18:36:09 GNUtoo|laptop: those options are harmless Feb 14 18:36:26 ok Feb 14 18:37:05 * GNUtoo|laptop always remove LDFLAGS when compiling uboot or the kenrel with the devshell else he has compilation errors Feb 14 18:37:07 GNUtoo|laptop: can you preprocess caliberate.c Feb 14 18:37:14 ok Feb 14 18:37:14 and pasteint Feb 14 18:37:34 I don't remember how to preprocess kenrel stuff Feb 14 18:38:20 I'll look Feb 14 18:39:17 khem: so should I push 2.6.6 and then you'll apply your RPATH fix on top? Feb 14 18:39:29 JaMa|Off: eitherway Feb 14 18:39:45 JaMa|Off: are you going to remove 2.6.5 ? Feb 14 18:39:53 and 2.7 Feb 14 18:40:38 only 2.6.5 Feb 14 18:40:54 hmmm ok Feb 14 18:40:55 go ahead Feb 14 18:40:59 there might be some folks who need 2.7, but lets not make it DP=1 Feb 14 18:41:07 yeah got it Feb 14 18:41:44 ok found it Feb 14 18:42:08 JaMa|Off: are you going to push those changes now or later Feb 14 18:42:50 you can add my acks I just looked at the patches on pw Feb 14 18:43:52 now Feb 14 18:44:21 ok I liked the cleanup wrt to packaging Feb 14 18:45:01 03Martin Jansa  07master * rbb99d26fb2 10openembedded.git/ (2 files in 2 dirs): Feb 14 18:45:01 python: generate-manifest-2.6.py, update few entries Feb 14 18:45:01 * config was in python-dev as well as python-distutils, keep in Feb 14 18:45:01 distutils Feb 14 18:45:01 * linecache.* was in python-core as well as python-lang, keep in core Feb 14 18:45:01 Signed-off-by: Martin Jansa Feb 14 18:45:11 03Martin Jansa  07master * r06be2fc468 10openembedded.git/ (2 files in 2 dirs): Feb 14 18:45:12 python: generate-manifest-2.6.py, sort packages when generating output file Feb 14 18:45:12 * iteritems and packages didn't have defined order, so every regeneration created huge patch, hard to review Feb 14 18:45:12 * alphabetical order should work fine here as long as we have disjunct FILES Feb 14 18:45:12 Signed-off-by: Martin Jansa Feb 14 18:45:13 03Martin Jansa  07master * rb980489fda 10openembedded.git/ (2 files in 2 dirs): Feb 14 18:45:13 python: generate-manifest-2.6.py, more s/python/${PN}/g and regenerate it Feb 14 18:45:13 Signed-off-by: Martin Jansa Feb 14 18:45:14 03Martin Jansa  07master * r3488225a01 10openembedded.git/recipes/python/python_2.6.6.bb: Feb 14 18:45:15 python: merge few s/python/${PN}/g from poky Feb 14 18:45:15 Signed-off-by: Martin Jansa Feb 14 18:45:16 03Martin Jansa  07master * re6a1fcb4ab 10openembedded.git/recipes/python/ (18 files in 2 dirs): Feb 14 18:45:54 python: upgrade from 2.6.5 to 2.6.6 Feb 14 18:45:54 * refresh patches from poky where applicable Feb 14 18:45:54 Signed-off-by: Martin Jansa Feb 14 18:46:28 arrg javascript.....is slowing down the paste Feb 14 18:46:35 pushed and now I gtg Feb 14 18:47:26 Tartarus: I have send the second version of the linux-firmware.patch Feb 14 18:47:47 it says the paste is larger than 100k, I'll try another pastebin Feb 14 18:48:51 heh Feb 14 18:49:01 http://pastebin.com/tNcPqz2T Feb 14 18:49:34 k, thanks Feb 14 18:49:51 it's from init/calibrate.i Feb 14 18:50:04 GNUtoo|laptop: and do you see the commandline options Feb 14 18:50:14 used to compile caliberate.i Feb 14 18:50:52 Tartarus: pong Feb 14 18:50:56 GNUtoo|laptop: can you also generate caliberate.s Feb 14 18:51:08 ok Feb 14 18:51:08 denix, you use external-toolchain-csl, do you do something to pass in --sysroot? Feb 14 18:51:27 oe_runmake uImage CC="arm-angstrom-linux-gnueabi-gcc -mno-thumb-interwork -mno-thumb" LD="arm-angstrom-linux-gnueabi-ld " init/calibrate.i V=1 Feb 14 18:51:36 http://pastebin.com/cmJUL5SB Feb 14 18:51:36 Tartarus: not yet Feb 14 18:51:56 GNUtoo|laptop: no I need all the options Feb 14 18:51:58 with V=1 Feb 14 18:52:07 oe_runmake V=1 Feb 14 18:52:08 that's in the pastebin Feb 14 18:52:49 Tartarus: you have a patch for that? Feb 14 18:52:58 not formally, no Feb 14 18:53:14 But you need one too then and we aren't crazy? OK, good :) Feb 14 18:53:25 oe_runmake uImage CC="arm-angstrom-linux-gnueabi-gcc -mno-thumb-interwork -mno-thumb" LD="arm-angstrom-linux-gnueabi-ld " init/calibrate.s V=1 => http://pastebin.com/3Y8Nhk6G Feb 14 18:53:27 :) Feb 14 18:53:53 GNUtoo|laptop: it says make[1]: `init/calibrate.i' is up to date. Feb 14 18:53:58 you need to delete it and redo Feb 14 18:54:05 then you will see options Feb 14 18:54:21 http://pastebin.com/0NVeLJ7D Feb 14 18:54:24 calibrate.s ^^^ Feb 14 18:54:30 ok thanks Feb 14 18:54:38 that calibrate.s has mcount Feb 14 18:54:54 yes i see Feb 14 18:55:00 redo caliberate.i Feb 14 18:55:52 http://pastebin.com/JbvQ1a5k Feb 14 18:57:20 * Tartarus kicks u-boot for insisting that the right way to build the fw env tools is to default to hostcc Feb 14 18:57:22 GNUtoo|laptop: ok so here it your problem you are passing -pg Feb 14 18:57:24 to compiler Feb 14 18:58:10 ah ok Feb 14 18:58:15 why is it passing -pg? Feb 14 18:58:18 I'll look Feb 14 18:58:20 GNUtoo|laptop: thats only useful if you will provide mcount and other profiling hook implementation Feb 14 18:58:25 compiler sort of expects it Feb 14 18:58:34 yes but I didn't change that myself Feb 14 18:58:36 I'll look Feb 14 18:58:42 its passing because you are asking for it :) Feb 14 18:58:48 machines are seldom wrong Feb 14 18:59:06 * khem needs tea Feb 14 18:59:13 I mean I didn't commit something, maybe someone else did Feb 14 18:59:22 another human Feb 14 19:05:29 where should I look for pg? Feb 14 19:05:30 hrmm now python is fixed lets move to the first problem I was trying to fix Feb 14 19:05:47 it's not in local.conf obviously Feb 14 19:05:53 GNUtoo|laptop: do bitbake -e > /tmp/xx Feb 14 19:05:58 ok Feb 14 19:06:02 and then search in /tmp/xx Feb 14 19:06:07 that might give you idea Feb 14 19:06:15 yes for getting the environment Feb 14 19:06:19 thanks for the idea Feb 14 19:06:25 filip, i'll push the package_deb.bbclass change shortly (just kicked a test build of something else) Feb 14 19:06:37 And I want someone else to make sure I really remember on device update paths right Feb 14 19:06:44 Pretty sure I do, heh Feb 14 19:06:57 Tartarus: yes deb changes are fine Feb 14 19:07:05 Tartarus: deb was broken anyway Feb 14 19:07:09 how about the linux-firmware one? Feb 14 19:07:17 did not look into that one Feb 14 19:07:42 http://patches.openembedded.org/patch/589/ Feb 14 19:07:44 this one ? Feb 14 19:07:53 yes Feb 14 19:09:00 03Filip Zyzniewski  07org.openembedded.dev * r0a4ef90cd6 10openembedded.git/classes/package_deb.bbclass: Feb 14 19:09:00 package_deb.bbclass: Use PKGV/PKGR variables where applicable. Feb 14 19:09:00 Using PV/PR causes problems like libncursesw5_5.7-r16 depending on Feb 14 19:09:00 libtinfo5 (>= 5.7+20110115). Feb 14 19:09:00 Signed-off-by: Filip Zyzniewski Feb 14 19:09:01 Signed-off-by: Tom Rini Feb 14 19:09:14 * Tartarus pushes the delegate button :) Feb 14 19:10:18 Try to defer http://patches.openembedded.org/patch/573/ Feb 14 19:10:27 as long as we fix the fallouts Feb 14 19:14:21 03Chris Larson  07master * rdcefe88bff 10openembedded.git/classes/packaged-staging.bbclass: Feb 14 19:14:21 Revert 6a7f8c in favor of a real fix Feb 14 19:14:21 Signed-off-by: Chris Larson Feb 14 19:14:21 Yeah Feb 14 19:15:42 Tartarus: looking at your build picture Feb 14 19:15:48 I guess we should add one of micro Feb 14 19:15:57 since it has flattened tree Feb 14 19:16:10 it might be a good one to test Feb 14 19:16:23 for atleast one machine/image Feb 14 19:16:43 I can try adding micro + qemu* Feb 14 19:16:50 and see what the fall out is in terms of time, too Feb 14 19:17:55 I know grg uses micro Feb 14 19:18:02 so you will have some help there Feb 14 19:27:35 filip: PKGR already contains DISTRO_PR Feb 14 19:30:13 03Tom Rini  07master * rd5157da6f3 10openembedded.git/recipes/u-boot/ (u-boot.inc u-boot_git.bb): (log message trimmed) Feb 14 19:30:13 u-boot: Make sure we use override HOSTCC for env tools Feb 14 19:30:13 Upstream says that the right way to build a target binary of the env Feb 14 19:30:13 tools is to pass in HOSTCC=your-cross-cc. Bump the PR in u-boot_git.bb Feb 14 19:30:13 to fix as many easily fixable ones as possible, but other platforms Feb 14 19:30:13 may have been shipping bad binaries (since we had skipped testing Feb 14 19:30:14 on this package due to GNU_HASH issues when built for the target). Feb 14 19:32:03 filip: afaik the control files of deb and ipk packages should be identical, except for the target architecture specification on some platforms Feb 14 19:34:16 obi: argh, so the patch is wrong? Feb 14 19:34:36 yes Feb 14 19:36:14 i think it would be best to use a common code path for both package types Feb 14 19:36:27 hm Feb 14 19:37:14 the idea overwhelms me ATM :) Feb 14 19:39:44 obi: do you suggest inheritance from a common file? Feb 14 19:40:01 yes Feb 14 19:40:33 eh, I am not the right person for this :P Feb 14 19:40:40 but I will try Feb 14 19:40:48 ok, it can still be done later Feb 14 19:44:59 khem, I see no pg in bitbake -e linux-bug Feb 14 19:45:08 only sutff like gnupg Feb 14 19:45:17 or gpg Feb 14 19:45:23 obi: will this do for now: http://pastebin.com/DKeDs3xL ? Feb 14 19:46:49 GNUtoo|laptop: should be possible to track it Feb 14 19:46:59 grep in your source tree Feb 14 19:47:29 ah ok Feb 14 19:47:36 source of linux-bug then Feb 14 19:49:31 yes Feb 14 19:49:54 there are a lot of -pg Feb 14 19:49:58 I'll investigate Feb 14 19:50:00 thanks a lot Feb 14 19:52:59 any recently tried to build rpm-native? for me it fails in do_compile, with file: could not find any magic files (after spitting out some warnings) Feb 14 19:53:07 see http://www.pastie.org/1563831 for log Feb 14 19:53:35 or can someone perhaps kick off a build for rpm-native to see if this only at my side ? Feb 14 19:54:30 evening Feb 14 19:54:55 eFfeM: I can do Feb 14 19:56:10 Jay7: cool Feb 14 19:56:26 gone for 15 mins or so but will stay in the channel Feb 14 19:58:06 filip: yes, that patch looks good Feb 14 19:58:41 * Jay7 vote for commit access for filip Feb 14 19:59:48 Hi, I am still leaning OE stuff. So far, I successfully carryed out "bitbake nano" as stated in "getting started". My questions are: what is the next I probably need to do? How can configure OE for MPC8313E RDB board? Feb 14 20:00:31 I'll try to change the defconfig Feb 14 20:00:36 there are some profiling stuff in it Feb 14 20:02:01 I just do'nt have a clear pictures how to get a "hello world" running in my mpc8313 board. Feb 14 20:04:30 bing, is your board supported by oe? Feb 14 20:04:34 for example: show "hello world" in the hyperterminal console which hook to the MPC8313 board Feb 14 20:04:37 if not it's really easy to add Feb 14 20:04:57 then you could bitbake helloworld Feb 14 20:05:02 and install it Feb 14 20:05:04 and run it Feb 14 20:05:54 I think OE support MPC8313 Feb 14 20:06:06 or nearly any oe built package Feb 14 20:06:07 ok Feb 14 20:06:13 I'll look Feb 14 20:06:18 what's in your local.conf? Feb 14 20:06:21 for MACHINE Feb 14 20:06:59 bing, there's an mpc8315e-rdb target at least, perhaps an mpc8313e-rdb too Feb 14 20:07:05 I know I build one of those... Feb 14 20:07:13 the key is how to "install"? Feb 14 20:07:29 on the machine? Feb 14 20:07:32 install depend on your board Feb 14 20:07:34 with opkg/rpm/dpkg :) Feb 14 20:07:46 first you need to install an image Feb 14 20:07:57 that depend on your board Feb 14 20:08:12 for instance can it boot on microsd? Feb 14 20:08:20 does it require to boot on NAND Feb 14 20:08:22 etc.... Feb 14 20:08:28 microsd is easier Feb 14 20:08:42 you just need to untar the rootfs in the microsd with the right command Feb 14 20:08:55 sorry, I do't really understand what is "install an image", and how? Feb 14 20:08:58 and then you use the usual way of booting on microsd Feb 14 20:09:13 bing, I'll explain Feb 14 20:09:20 at first he should build image :) Feb 14 20:09:20 imagine you've got a new laptop Feb 14 20:09:39 ok Feb 14 20:09:54 and you want to run a hello world application under ubuntu Feb 14 20:09:58 on your laptop Feb 14 20:10:06 first you need to install ubuntu Feb 14 20:10:14 and then to install a hello world program Feb 14 20:10:18 ok Feb 14 20:10:28 now.... Feb 14 20:10:38 what you have there is a bit different from a laptop Feb 14 20:10:53 you need to build or get an image that was built for your device Feb 14 20:11:01 if you want to build it yourself Feb 14 20:11:05 just look in Feb 14 20:11:09 recipes/images Feb 14 20:11:11 for an image Feb 14 20:11:17 simple image include console-image Feb 14 20:11:51 or console-base-image Feb 14 20:12:02 do u mean that "image" is the OS of the board? Feb 14 20:12:06 yes Feb 14 20:12:10 somehow Feb 14 20:12:28 it's the ubuntu you install on the laptop Feb 14 20:12:37 only that the installation method is different Feb 14 20:12:56 and that you don't need to go trough a GUI and so on Feb 14 20:12:58 to install it Feb 14 20:13:18 what are the typical way to install? Feb 14 20:13:22 there are simplier ways but from the command line unfortunately Feb 14 20:13:33 first let's talk about your board Feb 14 20:13:44 can your board boot on a microsd or an SD card? Feb 14 20:14:08 the board is new, do't know yet Feb 14 20:14:26 just tell me a example Feb 14 20:14:57 do you have an sd slot on it? Feb 14 20:15:04 there is a sd slot on it Feb 14 20:15:07 ok Feb 14 20:15:21 then to install an image on sd I usually do that: Feb 14 20:15:26 bitbake console-image Feb 14 20:15:31 wait some hours Feb 14 20:15:35 khem, hey, do you know why we pass --disable-debug to *glibc configure? Makes stepping into stuff harder than needed :) Feb 14 20:15:38 if you build from scratch Feb 14 20:15:43 else it can be very fast Feb 14 20:15:57 then I put the microsd in my workstation Feb 14 20:16:00 Tartarus: bing there is an mpc8313e-rdb target, I've build it quite a number of times Feb 14 20:16:03 the one who built the "image" Feb 14 20:16:22 in your case you will need to format the microsd Feb 14 20:16:30 to linux + ext3 Feb 14 20:16:34 it can be done with gparted Feb 14 20:16:39 but beware Feb 14 20:16:42 ok Feb 14 20:16:47 format the right device Feb 14 20:16:52 don't format your hard drive Feb 14 20:16:56 be sure to format the microsd Feb 14 20:16:59 then mount it Feb 14 20:17:08 to mount it go in the equivalent of my computer Feb 14 20:17:16 and click on the microsd card Feb 14 20:17:29 then look where it is Feb 14 20:17:38 for instance /media/some_long_hash Feb 14 20:17:43 then: Feb 14 20:17:44 sudo su Feb 14 20:17:49 cd /media/some_long_hash Feb 14 20:17:54 eFfeM: it fails: http://tinderbox.openembedded.net/packages/2138703/ Feb 14 20:18:06 tar xvjpf /path/to/the/image/tarball.tar.bz2 Feb 14 20:18:12 the image tarball is located here : Feb 14 20:18:53 ok Feb 14 20:19:13 in my case they are here: Feb 14 20:19:16 /home/gnutoo/embedded/oe/oetmps/nokia900/deploy/images/nokia900/ Feb 14 20:19:18 that is to say Feb 14 20:19:30 /home/gnutoo/embedded/oe/oetmps/nokia900 is my tmpdir Feb 14 20:19:32 Jay7: thanks, this is the same failure I get Feb 14 20:20:45 bing, once you have the image setup, go in the equivalent of my computer and umount/eject or something like that the card Feb 14 20:20:49 go on please! Feb 14 20:20:51 then put the card on your board Feb 14 20:21:00 then it depends on your card Feb 14 20:21:10 the card should be powered off Feb 14 20:21:15 when puting the card in Feb 14 20:21:17 but then.... Feb 14 20:21:21 hold on Feb 14 20:21:32 oops Feb 14 20:21:35 when I said: Feb 14 20:21:37 the card should be powered off Feb 14 20:21:44 after mounting , should I copy the image to the card? Feb 14 20:21:46 I mean the mpc8313e-rdb Feb 14 20:21:56 no Feb 14 20:22:02 not copy Feb 14 20:22:04 but extract Feb 14 20:22:08 like I said Feb 14 20:22:12 with tar xvjpf Feb 14 20:22:23 the j is for the tar.bz2 format Feb 14 20:22:29 if you have a tar.gz Feb 14 20:22:30 use: Feb 14 20:22:34 tar xvzpf Feb 14 20:22:38 Tartarus: disable-debug does not mean no debugging info Feb 14 20:22:51 Tartarus: enable-debugging would mean emit more debug info Feb 14 20:22:56 ok, extract to the sd card Feb 14 20:23:17 Tartarus: that may not fix the issue you might be having but bloat the packages Feb 14 20:23:19 khem, iirc, it was stripping out -ggdb2 from my flags, but i'll double check then Feb 14 20:23:31 Basically once you have your sdcard inside the mpc8313e-rdb Feb 14 20:23:39 Tartarus: yes it will strip -ggdb2 but add -g Feb 14 20:23:39 khem, we strip debug info out, recall Feb 14 20:23:39 you power the mpc8313 on Feb 14 20:23:43 and then I don't know Feb 14 20:23:50 Tartarus: yes we do Feb 14 20:23:51 you will see some messages on the serial port Feb 14 20:24:04 maybe it will take your microsd card and boot right on it Feb 14 20:24:19 maybe you will need to interact with the bootloader to make it boot on the microsd Feb 14 20:24:37 khem, what I'm getting at is that today we don't have useable debug info coming out of the libc, even when we're building the rest of the world with debug info spelled out Feb 14 20:24:45 the bootloader is the arm equivalent of the BIOS+x86 bootloader on x86 platforms Feb 14 20:24:49 partially since we override FULL_OPTIMIZATIONS Feb 14 20:25:03 * Tartarus will have to double check which change really did it for him then Feb 14 20:25:08 Tartarus: We are not debugging friendly with our global opptions anyway Feb 14 20:25:56 Tartarus: DEBUG_OPTIMIZATION_pn-eglibc="-Os -g" is what I use in my local.conf Feb 14 20:26:01 and it works well Feb 14 20:26:04 khem, agreed, but we should be Feb 14 20:26:17 I always wanted to get rid of those fancy options we use Feb 14 20:26:19 khem, I also really really hate the idea of changing to a debug vs release build to get that to work Feb 14 20:26:26 just using -O2 -g would be enough Feb 14 20:26:29 or -Os -g Feb 14 20:26:33 simplifies the things Feb 14 20:26:37 I'll dig a little more then, thanks Feb 14 20:27:01 -fexpensive-optimizations -fomit-frame-pointer -frename-registers Feb 14 20:27:13 you kill the debugging by -fomit-frame-pointer Feb 14 20:27:20 so no matter what you do Feb 14 20:27:37 you stack trace will look bad Feb 14 20:27:45 bing, by the way,do you work in some companies, maybe it would be more simple to ask the person who knows that stuff in your company? as he could show you how to do it better than me and would have the hardware accessible and would know the hardware better than me Feb 14 20:27:53 I only know the devices I have Feb 14 20:28:10 Hi, G..|L..., Thank you very much, I need review what you said here. Feb 14 20:28:18 ok Feb 14 20:29:31 Unforturnally, I am the 1st person digging into this kind stuff now. Feb 14 20:29:48 ah ok Feb 14 20:29:50 im my company Feb 14 20:30:36 then you need to start easy, and as you go, progress Feb 14 20:30:57 does your board comes with software already installed on it? Feb 14 20:32:14 and what operating system or distribution is there on it? Feb 14 20:32:41 in any case you'll need to find out how to boot something in it Feb 14 20:35:18 khem, ohhh I think some kenrel option in kenrel hacking is responsible of -pg Feb 14 20:35:22 hehe.. Feb 14 20:35:28 spammers are funny Feb 14 20:35:33 GNUtoo|laptop: yes if you enable profiling in kconfig Feb 14 20:35:34 My name is sonia, honest and nice looking girl, i saw your lovely contact at patchwork.openembedded.org decided to contact you for friend, please so that i will send you my pictures and more about me. Feb 14 20:35:44 * khem needs food Feb 14 20:37:15 Jay7, images would have been better, at least you could misunderstand (if you try really hard) and think it's an oe image Feb 14 20:37:34 khem, no I think it was ftrace Feb 14 20:37:38 GNUtoo|laptop: nice idea! :)) Feb 14 20:40:37 * Jay7 need a way to disable oestats-client from shell env Feb 14 20:40:50 RP: Your list address was wrong www.... instead of lists... Feb 14 20:41:07 I'm sure some anonymous python functions will help me.. but I don't know what to write there :) Feb 14 20:41:13 GNUtoo|laptop, the so called "image" file is the file with "tar.gz" or"tar.bz2", right? Feb 14 20:41:28 extension Feb 14 20:41:42 bing, yes Feb 14 20:41:47 bing: or ext2/ext3/jffs2/ubifs Feb 14 20:41:57 basically you should also have some flash Feb 14 20:42:14 on your device Feb 14 20:42:33 "falsh"? Feb 14 20:42:43 well.. another round of benchmarks Feb 14 20:42:46 flash is like an usb key, with the huge different that it's not mass storage nor block device Feb 14 20:42:48 flash Feb 14 20:43:05 you access the chip directly Feb 14 20:43:17 so you need a filesystem made specially for it Feb 14 20:43:29 installing an image to the flash is called flashing Feb 14 20:43:37 and the process is quite different Feb 14 20:43:47 can be easier or harder depending on your setup Feb 14 20:44:07 yes, you mean in the processor on the mpc8313 board, right? Feb 14 20:44:12 but your company has only to write an application that runs on your device right? Feb 14 20:44:29 I mean in the mpc8313 board Feb 14 20:44:39 in that board you have some chips Feb 14 20:45:06 ok, I know what you mean Feb 14 20:45:12 theses chips do the same function than a hard drive, but they're different in many ways: Feb 14 20:45:18 *they're smaller and look like chips Feb 14 20:45:49 *what I just tell before apply to them ( flash is like an usb key, with the huge different that it's not mass storage nor block device ) Feb 14 20:45:55 there one integrated with processor Feb 14 20:46:06 not inside the processor but on the board and soldered Feb 14 20:46:14 but that's a detail at that point Feb 14 20:46:40 basically it's a chip or more than one chip soldered on your mpc8313 board Feb 14 20:47:06 ok, what about if the board has a flash chip? Feb 14 20:47:31 then you could boot from flash too Feb 14 20:47:39 without a microsd Feb 14 20:47:44 can I burn the image into the flash chip using programmer Feb 14 20:47:48 you just need to flash an image Feb 14 20:47:52 not exactly Feb 14 20:48:02 you could but that's strongly not the advised thing to do Feb 14 20:48:06 you have several ways Feb 14 20:48:20 if you don't already have an operating system on it..... Feb 14 20:48:36 you should still have something called a bootloader Feb 14 20:49:06 that's something that is like a BIOS and a bootloader on normal computers like PC Feb 14 20:49:22 basically that is loaded when you power on the board Feb 14 20:49:22 I think OE provide a bootloader Feb 14 20:49:37 right? Feb 14 20:49:39 it's very HW-dependent Feb 14 20:49:40 yes but that's a chicken and egg problem Feb 14 20:49:47 don't replace the one you have Feb 14 20:49:54 hm, can we provide a recipe which needs wine and propietary software to build? Feb 14 20:50:04 unless you've very good reason to do so and huge knowledge Feb 14 20:50:11 which you don't seem to have Feb 14 20:50:15 if you do replace it.... Feb 14 20:50:20 I mean linexec (buildable with microsoft compiler which is downloadable from the net Feb 14 20:50:22 you could render your board un-bootable Feb 14 20:50:42 filip, what's that? Feb 14 20:50:51 somehting like haret? Feb 14 20:51:01 GNUtoo|laptop: tell me how to flash image into flash memory Feb 14 20:51:07 bing, so back to what I was saying Feb 14 20:51:23 03Filip Zyzniewski  07org.openembedded.dev * r48752ac441 10openembedded.git/classes/package_deb.bbclass: Feb 14 20:51:24 package_deb.bbclass: remove double DISTRO_PR from the package version. Feb 14 20:51:24 PKGR already contains DISTRO_PR, so no need to add it twice. Feb 14 20:51:24 Signed-off-by: Filip Zyzniewski Feb 14 20:51:24 Signed-off-by: Tom Rini Feb 14 20:51:24 GNUtoo|laptop: yup Feb 14 20:51:26 bing, when you power the board, do you see something like u-boot written on the serial console Feb 14 20:51:27 ? Feb 14 20:51:37 03Filip Zyzniewski  07org.openembedded.dev * rcf1790d483 10openembedded.git/classes/package_deb.bbclass: (log message trimmed) Feb 14 20:51:37 package_deb.bbclass: depend on apt, not dpkg. Feb 14 20:51:37 Both package_deb.bbclass and rootfs_deb.bbclass use apt-get. Feb 14 20:51:37 Without apt system cannot be bootstrapped because there is no optimal Feb 14 20:51:38 way to fetch new packages. Apt depends on dpkg, so switching from dpkg Feb 14 20:51:38 to apt in EXTRA_DEPENDS provides both. Feb 14 20:51:39 Signed-off-by: Filip Zyzniewski Feb 14 20:52:11 sorry GNUtoo|laptop, have not boot it yet scien I bought it Feb 14 20:52:20 ok Feb 14 20:52:25 maybe that's the first step to do Feb 14 20:52:29 try to boot it Feb 14 20:52:33 and look at what you see Feb 14 20:52:39 ok Feb 14 20:52:42 look if it's GNU/Linux (also called linux) Feb 14 20:52:47 .. and google about it :) Feb 14 20:52:56 be familiar with it Feb 14 20:53:51 Jay7, I've only a touchpad, so bitbaking stuff and talking in some other channel at the same time....prevents me from using a browser and a search engine Feb 14 20:53:59 maybe I'll try to find a mouse Feb 14 20:54:18 Hi, GNUtoo|laptop, Jay7 and Tartarus, thank you guys replies. I got to leave now. I'll try boot the board and see what I am getting. Feb 14 20:54:29 ok Feb 14 20:54:34 that's the first thing to do Feb 14 20:54:35 GNUtoo|laptop: I just mean that you have said later - be familiar with it :) Feb 14 20:54:41 ok Feb 14 20:55:09 thanks, bye Feb 14 20:55:25 qt4-tools-sdk is with wrong values for qmake Feb 14 20:55:26 :-( Feb 14 20:55:32 looking into it Feb 14 20:55:37 btw....is there some users of freeciv? Feb 14 20:55:49 do they really want the GTK version? Feb 14 20:56:00 because the sdl version is better for small resolutions Feb 14 20:56:06 iirc, it have x11-only version as well Feb 14 20:56:18 ah, sdl Feb 14 20:56:32 because apart beagleboard and bug20 I wonder who has such a big screen Feb 14 20:57:52 03Khem Raj  07master * r827af469bd 10openembedded.git/recipes/python/ (python-2.6.6/python-module-rpath-fix.patch python_2.6.6.bb): Feb 14 20:57:52 python_2.6.6.bb: Fix QA ERRORS due to staging libdir in rpath Feb 14 20:57:52 Signed-off-by: Khem Raj Feb 14 20:58:04 03Khem Raj  07master * r393c0d5bc5 10openembedded.git/recipes/gstreamer/gst-rtsp.inc: Feb 14 20:58:04 gst-rtsp.inc: Add check to DEPENDS Feb 14 20:58:04 Signed-off-by: Khem Raj Feb 14 20:58:05 03Khem Raj  07master * r71514ed606 10openembedded.git/recipes/check/ (2 files in 2 dirs): Feb 14 20:58:05 check_0.9.5.bb: Add patch to use quadrigraphs in macro names Feb 14 20:58:05 Signed-off-by: Khem Raj Feb 14 20:59:01 what is the proper way to build a feed for multiple machines? so that not platform-specific packages are shared Feb 14 20:59:22 there's nothing magic about it. just build for one machine, then build for the next.. Feb 14 20:59:30 its all already handled under the hood Feb 14 21:00:24 kergoth: but the config allows specifying only one, right? Feb 14 21:00:34 kergoth: should I change the config on each iteration? Feb 14 21:00:59 yes, and yes Feb 14 21:01:22 hm, nice, I'll have to try that Feb 14 21:01:32 kergoth: how can I make INHERIT += "oestats-client" be switched on/off by some bb variable? Feb 14 21:01:46 Jay7: anonymous python or ${@} Feb 14 21:02:24 it is possible to do INHERIT+="" ? Feb 14 21:04:06 what do you mean? Feb 14 21:04:08 ah.. hehe.. if OESTATS_BUILDER is empty oestats-client does nothing Feb 14 21:04:29 of course you can append an empty string, why wouldn't you be able to? Feb 14 21:04:38 also, INHERIT isn't special, its just a variable that bitbake looks at.. Feb 14 21:05:36 that was right Q - is INHERIT just variable or is it something special :) Feb 14 21:05:44 now I know answer, thanks :) Feb 14 21:07:43 NOTE: package freeciv-2.2.4-r0: task do_build: Succeeded Feb 14 21:07:47 let's try it.... Feb 14 21:12:06 Jay7: hi for a similar thing I have INHERIT += "${@['', 'rm_work'][bb.data.getVar('EUKREA_NORMWORK', d, 1) == '0']}" Feb 14 21:12:26 ericben: ah, thanks Feb 14 21:12:37 I think we should place this somewhere on wiki Feb 14 21:12:43 'tips and tricks' Feb 14 21:18:57 re Feb 14 21:19:13 hi florian Feb 14 21:22:07 ka6sox: you see i'm alive... well... sort of Feb 14 21:22:48 ouch...thats not fun. Feb 14 21:28:11 florian, still ill? Feb 14 21:30:01 Philippe_: not really... almost back to normal Feb 14 21:32:08 florian, good to hear. Too bad you did not make Fosdem Feb 14 21:32:13 Dimitry is around? Feb 14 21:32:52 Philippe_: I made it but one day only Feb 14 21:32:55 I am having trouble to fix qt4-tools-sdk ... qmake has wrong internal values for QT_INSTALL_* Feb 14 21:33:18 florian, damn then I missed you. Only had some 10 min to run around quickly on sunday Feb 14 21:35:59 hello Feb 14 21:37:22 maybe a stupid question, but, can we use openembedded to, lets say, a minimalized linux to make it run on a old winmobile2003se platform like on xscale processors with limited ram (32gb or so), i own an old panasonic toughbook cf-p2 mobile now, and wm2003se s... a lot Feb 14 21:39:29 03Khem Raj  07master * r6262bfce3e 10openembedded.git/recipes/gstreamer/gst-rtsp.inc: Feb 14 21:39:30 Revert "gst-rtsp.inc: Add check to DEPENDS" Feb 14 21:39:30 check is already in the DEPENDS list plus it had typo in it Feb 14 21:39:30 This reverts commit 393c0d5bc5850592b51013d0918d0fe623d40de8. Feb 14 21:48:18 vizzy: 32gb is quite a lot of RAM actually :p Feb 14 21:48:39 ynezz, yes, ... i know kernel 1.2.13 ;-) Feb 14 21:48:55 linux fits on one 1,44MB disk, with X and all that Feb 14 21:49:00 and just 4mb of ram Feb 14 21:49:16 eh Feb 14 21:49:21 but actually i am a debian legacy (lenny) user , and neeed lots of ram Feb 14 21:50:04 vizzy: there are some alike machines Feb 14 21:50:14 so, just in case i find a suitable distribution... would it be possible to flash one of those wince xscale arm phones with an embedded linux? Feb 14 21:50:26 but w/o cellphone functions Feb 14 21:50:39 hmm, but thats the meaning of a cellphone :)) Feb 14 21:51:00 i got mine only because it is rugged... no stylish iphone that breaks well Feb 14 21:51:53 check conf/* files for similar machines Feb 14 21:51:59 cool :) Feb 14 21:52:00 conf/machine even Feb 14 21:52:16 just an idea, but i thought of it too often in the last days Feb 14 21:52:17 then try to google about distro they are using Feb 14 21:52:32 I know there was some work around HTC machines Feb 14 21:53:12 maybe i could talk to some of your developers on the cebit :) Feb 14 21:53:55 do i need to be a c coder to port a distro to my phone? or just have experiences in setting up linux and compiling, maybe patching programs/sources? Feb 14 21:54:23 vizzy: second mostly Feb 14 21:54:43 thats ok for me, so its worth trying to digg into oe too :) Feb 14 21:54:52 but I'll repeat that I'm unsure about cellphone functions ;) Feb 14 21:54:56 will take months maybe, but its worth it i think Feb 14 21:55:11 if the phone chip is supported in some way..... ;) Feb 14 21:55:24 yeah.. drivers are main question Feb 14 21:55:43 phone.exe is also just a programm in wm2003 Feb 14 21:56:17 and the first i installed on my phone wa spocketblender3d, hahaha Feb 14 21:56:23 so, everything is possible i guess Feb 14 21:57:52 well.. you can try at least Feb 14 21:57:57 if there's some FPGA, it might be quite harder... Feb 14 21:59:01 hmmm. i know there are some people working on android for old phones, but they fight with memory, ... and last but not least i want to avoid google os Feb 14 21:59:24 isnt android a linux too? Feb 14 21:59:46 it's not "GNU" Linux, but "Android" Linux.. Feb 14 22:00:07 oh, oooh, "android linux" , wowowow :)) Feb 14 22:00:28 (I hate the term GNU/Linux... but the reality is Android isn't "GNU/Linux") Feb 14 22:00:36 must be very special :) Feb 14 22:00:53 Android is the Bionic Libc + modified Linux kernel + a dalvik.. Feb 14 22:00:57 kernel + java vm Feb 14 22:01:19 if it were glibc based.. then I would say it's "just another Linux".. but it's not Feb 14 22:01:27 modified linux kernel? ah, ok, so only a few parts are under gpl etc Feb 14 22:01:46 so not everything is free and open i guess Feb 14 22:01:48 the only GPL bit is the Linux kernel itself.. everything else [generalization] is not GPL) Feb 14 22:02:06 i believe in oss :) Feb 14 22:02:11 vizzy: wm2003 device and 32mb ram == hard to run anything modern on it Feb 14 22:02:24 hrw, but legacy :) Feb 14 22:02:41 legacy linux for a legacy phone Feb 14 22:03:06 vizzy: 4-5 years ago it was starting to be hard when most of our devices had at least 64mb. Feb 14 22:03:14 hmmm, maybe running dosbox (its installed) , launch linloader for dos and execute and old umtsfs linux :)) Feb 14 22:03:58 hrw, not mine, i have my brand new 6 year old phone just a few days now :P Feb 14 22:06:09 in theory...i had once slackware 3 running on my old 486 with 32mb and 50MB SCSI disk including sound and enlightenment and so on, now i want that on my phone ;) Feb 14 22:06:40 shr distro have enlightenment-based iface iirc Feb 14 22:07:07 vizzy: I had debian on 386sx with 5mb ram Feb 14 22:07:16 sx, ouch :)) Feb 14 22:07:25 in 2001 Feb 14 22:07:35 wow Feb 14 22:07:42 this was an old computer in 2001 Feb 14 22:08:03 I had to write some software for such antics Feb 14 22:08:48 so I got this hyundai from uni. it had 386sx, fpu, 5mb ram, 120mb hdd, ati gfx onboard and 2-3 ethernet cards Feb 14 22:09:14 my first one was a used 286 12MHZ with 4MB!!! ram, and 10GB ESDI drive, fullpage monitor b/w... running quarkexpress and oracle for dos :)) but i mostly played monkey island on it Feb 14 22:09:28 and a copro! Feb 14 22:09:34 ok, going OT :) Feb 14 22:09:56 vizzy: your phone is ~8 years old now. if no one worked on kernel at that time then no one will do that now. Feb 14 22:10:13 hmmm, true Feb 14 22:10:31 vizzy: so first you need to check did anyone booted linux on it. if not then use winmo on it or put it into trashcan Feb 14 22:11:01 vizzy: if yes, then check how old it was and if there never was 2.6 for it then again - winmo or trashcan Feb 14 22:11:26 hm.. Feb 14 22:11:36 very like to zaurus'es situation.. Feb 14 22:11:50 Jay7: worse - zauri had linux from start and have serial port Feb 14 22:11:54 trashcan may be good idea :) Feb 14 22:12:20 yes, but we are still have no power management in current 2.6 Feb 14 22:12:22 Jay7: in 2003 zaurus was very hacker friendly device I think Feb 14 22:12:37 Jay7: no one sane works on zauruses... Feb 14 22:12:43 sure.. Feb 14 22:13:27 only tosa have some progress Feb 14 22:14:23 i had a zaurus for several years too but i sold it , accidently :) also had qt on it, but was not so useable Feb 14 22:14:41 03Andreas Müller  07master * r1b1543c9f6 10openembedded.git/contrib/qa/collect-qa-errors.sh: Feb 14 22:14:41 Add a small helper script listing all recipes with QA errors. Feb 14 22:14:41 Signed-off-by: Andreas Mueller Feb 14 22:14:41 Signed-off-by: Khem Raj Feb 14 22:14:41 Acked-by: Otavio Salvador Feb 14 22:16:32 hrw: how insane I must be to work on a jornada 720 ;) Feb 14 22:17:01 32MB ram, masked ROM, DSTN Feb 14 22:17:37 hmmm Feb 14 22:18:00 it's a bit different than zaurus though - no successor Feb 14 22:18:06 could that http://scap.linuxtogo.org/files/0fafa1dfaebc710020e929434aec7bf9.png be compiler obtimizations again? Feb 14 22:19:38 gnutoo: looks like wrong font of wrong codepage Feb 14 22:19:39 filip: zaurus had successor? Feb 14 22:19:54 no real successor Feb 14 22:20:01 hrw: well IMO jornada is more unique than zaurus Feb 14 22:20:16 hrw: there are newer devices with keyboard of quality not worse than zaurus Feb 14 22:20:20 there was two devices - Netwalker and another one.. can't remember Feb 14 22:20:43 hrw: there are no newer devices with keyboard comparable to the jornada (and form factor) Feb 14 22:20:47 Jay7: sigmarion 3 Feb 14 22:20:48 Jay7: netwalker was interesting. 5" with 1024x600 iirc Feb 14 22:21:07 hrw: and windows inside :( Feb 14 22:21:10 yeah, bu netwalker is expensive and jornada is like disposable Feb 14 22:21:28 Jay7: nope, netwalker runs ubuntu Feb 14 22:21:34 filip: jornadas use strongarm or something newer? Feb 14 22:21:35 hm.. Feb 14 22:21:45 then I've messed things Feb 14 22:22:03 hrw: sa1110 @ 206MHz Feb 14 22:22:27 like my first linuxpda - collie Feb 14 22:22:36 :) Feb 14 22:22:45 but it was 7 years ago Feb 14 22:22:47 my frind has one, doesn't use it anymore though Feb 14 22:23:08 * Jay7 looked over collie on desk.. Feb 14 22:23:22 my current linuxpda has 1GHz and runs android Feb 14 22:23:39 hrw: some *pad? Feb 14 22:23:51 filip: nexus s Feb 14 22:24:21 hrw: did you returned your N900? Feb 14 22:24:30 I've remember some posts about it in your blog :) Feb 14 22:24:47 Jay7: return? it is my own property Feb 14 22:24:57 ah.. I've messed things again :) Feb 14 22:25:15 well I just enjoy tinkering and improving OE. age of this machine doesn't matter that much for it Feb 14 22:26:57 I cheated a bit though - I've soldered 128MB of ram into mine :P Feb 14 22:27:01 filip: as long as you use this device Feb 14 22:27:21 filip: 128MB ram change a lot in such device Feb 14 22:27:36 yup Feb 14 22:27:54 filip: add bt on serial and you will have easy networking ;D Feb 14 22:28:33 i once ordered an openpandora box, but i give up waiting ;) Feb 14 22:28:41 hehe.. Feb 14 22:29:02 my friend is selling his akita now to get money to order replacement items for openpandora Feb 14 22:29:04 vizzy: I played with one recently Feb 14 22:29:16 oh, how was it? Feb 14 22:29:24 I have to sell my n900 until it has any value Feb 14 22:29:44 hrw: it has PCMCIA, I put orinoco or 3G modem in it, and I need no stinkin' bluetooth :) Feb 14 22:29:45 I've played with OP too.. interesting device Feb 14 22:29:53 it was promising, but i got my money back and bought a used thinkpad for it Feb 14 22:30:08 but case quality is worst Feb 14 22:30:37 yes, a lot of trouble with the cases... and nubs Feb 14 22:30:57 strange.. ben nanonote is cheaper but have excellent case Feb 14 22:31:16 but the keyboard was a bit weird i mentioned, i wanted it for coding, ssh'ing not gaming Feb 14 22:31:49 heh, ben nanonote - when I first saw it it almost killed me with its cuteness Feb 14 22:32:11 is this the one from alwaysinnovative or how they are called? Feb 14 22:32:37 vizzy: no, it's qi hardware's Feb 14 22:32:37 vizzy: http://en.qi-hardware.com/wiki/Ben_NanoNote Feb 14 22:32:49 this one i mean alwaysinnovating.com Feb 14 22:32:51 very strange but nice device :) Feb 14 22:32:52 ah, ok Feb 14 22:33:14 so my wm2003 sucks a bit, i think i will need a new toy soon Feb 14 22:33:46 vizzy: it's just like running convert -scale 25% on a regular notebook Feb 14 22:33:58 hehe Feb 14 22:34:25 down to the screen latch Feb 14 22:34:32 ~15-20% of my thinkpad x61t ;) Feb 14 22:34:34 vizzy: if you want to play with linux on pocketable device then it is hard to susggest something for me now Feb 14 22:35:11 nanonote is ok to play with linux.. Feb 14 22:35:23 hrw: netwalker seems nice, except for the price Feb 14 22:35:27 but if you wish to do something useful with it.. :) Feb 14 22:35:29 its just strange device Feb 14 22:35:52 just buy used thinkpad then :) Feb 14 22:36:08 filip: efika mx smartbook also looks nice. but if I would have to pay for it I would not buy it Feb 14 22:37:25 hrw: haven't you played with alwaysinnovating's devices? Feb 14 22:37:30 * khem has a efika smarttop Feb 14 22:37:35 I mean touchbook Feb 14 22:37:52 there is a mips64 based netbook Feb 14 22:37:57 I was searching for it Feb 14 22:38:00 but never got around Feb 14 22:38:25 khem: btw, what kernel do we have for efikamx? some from powerdeveloper.org? Feb 14 22:38:46 Jay7: no, never played with AI ones Feb 14 22:39:12 khem: and second Q - did you tried to run OE-based images on efika? Feb 14 22:39:13 Jay7: for efikamx/sb there is only powerdeveloper.org one which works Feb 14 22:42:12 Jay7, ok, any idea on how to fix? Feb 14 22:42:33 gnutoo: can you try to use other font? Feb 14 22:45:09 khem: since I added ASSUME_PROVIDED += "linux-libc-headers-native" curl-native always build (well, 5/5 builds) Feb 14 22:45:29 I tracked back the origin of the issue to http://lists.linuxtogo.org/pipermail/openembedded-devel/2010-June/020603.html Feb 14 22:45:42 [oe] [PATCH, RFC] Add linux-libc-headers-native, make it default dep for native Feb 14 22:46:18 we got so far? Feb 14 22:46:50 linux-libc-headers-native are needed to uild curl-native to be able to fetch archives? or to be able to build git? Feb 14 22:46:52 imho that code can lead to unpredictable results, depending on how quickly the headers are staged Feb 14 22:47:18 hrw: on Gentoo curl-native fails do_configure very often Feb 14 22:47:25 second run is just fine Feb 14 22:47:43 sometimes fails do_compile Feb 14 22:48:00 since l-libc-h-native are a dep for everything, it can't be adding a race :) Feb 14 22:48:16 hrw: it's too big, I use portege r200 for my daily computing and it's not much bigger/heavier Feb 14 22:48:29 now, that was discussed on the ML but the default behavior implies we need to add the ASSUME_PROVIDED += "linux-libc-headers-native" Feb 14 22:48:39 we = Gentoo users Feb 14 22:49:06 but you really really shouldn't Feb 14 22:49:17 ant__, any chance you can find me a VMware container w/ gentoo already installed? Feb 14 22:49:31 Jay7, didn't work Feb 14 22:49:38 basically thie font is at: Feb 14 22:50:05 Tartarus: no idea, I can give you ssh overnight Feb 14 22:50:29 /usr/share/freeciv/themes/gui-sdl/human/DejaVuSans.ttf Feb 14 22:50:38 I copied the one from /usr/share/fonts Feb 14 22:50:49 the DejaVuSans one Feb 14 22:50:57 and still the same result Feb 14 22:51:01 maybe I'll tr on htcdream Feb 14 22:51:06 khem: Tartarus btw gnutoo saw the same issue yesterday Feb 14 22:51:17 gnutoo: no idea then.. Feb 14 22:51:23 ant__, which issue? Feb 14 22:51:30 curl-native vs. Gentoo Feb 14 22:51:37 yes Feb 14 22:51:42 I remember vaguely Feb 14 22:51:44 http://www.dynamism.com/top-notebooks/sony-vaio-p.shtml this one is cool Feb 14 22:51:51 but I don't run gentoo anymore Feb 14 22:51:53 gnutoo: is in the irclogs ;) Feb 14 22:51:55 Tartarus: the issue is that linux-libc-headers-native is not build early enough Feb 14 22:52:07 khem, I don't see how Feb 14 22:52:09 I just cleaned it Feb 14 22:52:11 lemme re-read Feb 14 22:52:12 and it causes some packages to mix headers from your build machine and linux-libc-headers-native Feb 14 22:52:16 and I bitbaken the image again Feb 14 22:52:19 filip: yes, but vaio is very windows-wired :) Feb 14 22:52:20 and it went on Feb 14 22:52:23 if I remember well Feb 14 22:52:30 I had a brief chat with RP and he says those newer headers are evil... Feb 14 22:52:39 only quilt-native goes before linux-libc-headers now Feb 14 22:52:39 let me look Feb 14 22:52:46 Tartarus: where do we create the dep for linux-libc-headers Feb 14 22:52:50 native Feb 14 22:52:55 Jay7: I just mean the form factor. For example in netwalker I hate that they wasted so much space around the screen Feb 14 22:52:56 base.bbclass Feb 14 22:53:08 filip: ah Feb 14 22:53:11 Everything but quilt-native, unifdef-native and shasum-native get it Feb 14 22:53:19 yes, formfactor is pretty Feb 14 22:53:22 Oh and INHIBIT_DEFAULT_DEPS Feb 14 22:53:41 I had udev issues too Feb 14 22:53:52 that made me switch distro from angstrom 2008.1 to 2010.x Feb 14 22:54:04 Tartarus: btw. are linux-libc-headers-native needed on all build distros ? Feb 14 22:54:11 They're always safe, yes Feb 14 22:54:13 or it is just ot make old RHEL happy Feb 14 22:54:30 Tartarus: as it seems they are not safe Feb 14 22:54:39 They are safe Feb 14 22:54:44 :/ Feb 14 22:55:08 with 8 bb threads curl-native failed again Feb 14 22:55:12 hmmm I don't find libcurl in my recent logs Feb 14 22:55:20 I'll grep for curl Feb 14 22:55:25 ant__: is that when you disabled linux-libc-headers-native Feb 14 22:55:37 Show me how curl-native is starting before linux-libc-headers-native is done Feb 14 22:55:46 It's either a bitbake bug is a curl recipe bug Feb 14 22:55:52 Feb 13 20:47:29 | checking for send... no| configure: error: Unable to link function send NOTE: package curl-native-7.21.2-r4.0: task do_configure: Failed Feb 14 22:56:00 I think I just cleaned it Feb 14 22:56:01 since curl does not set INHIBIT_DEFAULT_DEPS Feb 14 22:56:01 khem, yes, is provided thus not built Feb 14 22:56:27 ant__: ok then it could be something else in curl thats needed Feb 14 22:56:40 is it same function its failing at always Feb 14 22:56:45 http://tinderbox.openembedded.org/packages/2139023/ Feb 14 22:57:10 heh google to welcome the 1800 employees that nokia will fire :) Feb 14 22:57:23 nokia is really in bad spot Feb 14 22:57:46 indeed Feb 14 22:57:55 specially for FSO+SHR Feb 14 22:58:18 and I believe with meego and now with this MS story Feb 14 22:58:33 yes Feb 14 22:58:50 there is a old indian saying Two turkeys can not make an eagle Feb 14 22:59:04 khem, well that is good news for me then :D Feb 14 22:59:15 how come Feb 14 22:59:26 are you in the firing line Feb 14 22:59:35 khem, I am one of those poor sods working in Meego... Feb 14 22:59:37 khem: you mean microsoft + nokia :)? Feb 14 22:59:39 Jay7: it seems a wrong link is created for tinderbox build errors Feb 14 22:59:49 I mean both deals Feb 14 23:00:18 I heard nokia is going to release one product from meego and only one to _sense_ the market Feb 14 23:00:34 so in plain words that means meego is dead after that Feb 14 23:00:44 well I understood it like that too. Feb 14 23:00:47 and they go in bed with M$ Feb 14 23:01:03 indeed the smartphone distros are too crowded Feb 14 23:01:07 khem: this time configure failed for other reasons (contamination ?) Feb 14 23:01:11 but meego guys will do the mobile optimizations and enhancements to the new win7 platform Feb 14 23:01:15 | checking run-time libs availability... failed Feb 14 23:01:15 | configure: error: one or more libs available at link-time are not available run-time. Libs used at link-time: -lldap -lrt -lz Feb 14 23:01:17 so Philippe_ you still might be needed Feb 14 23:01:39 ant__: dude no idea Feb 14 23:01:46 khem, needed maybe. Rather I do not want to be in that spot ;) Feb 14 23:01:47 ant__, and what does config.log show? Feb 14 23:01:48 you always have weirdest of errors Feb 14 23:02:12 hehe Feb 14 23:02:42 ant__: what is wrong exactly in link? Feb 14 23:03:18 Philippe_: so what stuff do u do on meego Feb 14 23:03:40 khem, I have done tons of things. Atm been working on the USB control mostly. Feb 14 23:03:40 * Jay7 remember his ALT Linux' times.. Feb 14 23:03:41 someone wanted me to work on meego some 6months back but I did not believe it Feb 14 23:03:51 well.. now ant__ have weirdest errors :) Feb 14 23:03:51 filip: friend is using one of those small vaio with ubuntu just fine Feb 14 23:04:17 hrw: I am going to apply recent patches from linaro in OE Feb 14 23:04:23 another week of bugs :) Feb 14 23:05:02 moment, I'm uploading the file..is very long Feb 14 23:05:11 khem, you escaped a disaster there. Feb 14 23:05:43 We had a strategy meet today and one guy complained he had just been hired and felt pretty bad about the current situation Feb 14 23:07:03 Philippe_: get some HW guys and found own company ;) Feb 14 23:07:17 that would be cool..... Feb 14 23:07:40 I think that nokia will release one meego device just to show that they not abandoned platform. thats all Feb 14 23:07:56 hrw: and then abandon platform :) Feb 14 23:08:20 and it will be something like tablet with 3g modem to not have to say 'it is a phone' Feb 14 23:08:21 Philippe_: back than I called it offspring of two dinasaurs :) Feb 14 23:08:25 Jay7: yes Feb 14 23:08:36 Well Nokia strategy with Maemo and Meego has pissed me off very often. Feb 14 23:08:51 Philippe_: they had strategy for those platforms? Feb 14 23:09:08 Philippe_: maybe two years before releasing hw... Feb 14 23:09:11 place it on wikileaks! ;) Feb 14 23:09:11 I think nokia is never OSS savvy Feb 14 23:09:18 hrw, no idea. IMHO management screwed up at least two years ago already Feb 14 23:09:43 I remember trying to convince them to use lnx for next gen phones and they were simply reluctant Feb 14 23:09:51 hmmm same issue on htcdream Feb 14 23:09:52 strange Feb 14 23:09:55 for freeciv Feb 14 23:10:01 Philippe_: as owner of 770/n810/n900 I feel that nokia tends to forget about device when first ones goes from production line to boxes. Feb 14 23:10:16 from the outside, it looks like Nokia has done a poor job converting research into products Feb 14 23:10:21 hrw, one of the things I complained about internally a lot. Feb 14 23:10:38 Philippe_: n800 guys were happiest ones cause they got nice software update when n810 was released. Feb 14 23:10:39 Not to mention the habit of making the bloody same mistakes with each new product. Feb 14 23:11:35 n900 was said to be forgotten just after release. ok, there were few sw updates but nothing special came with it. just minimal maintaince Feb 14 23:11:51 I would personally have liked to see a N900+ or something Feb 14 23:11:51 anyway it is time for me to finally get to bed Feb 14 23:12:06 Philippe_: my n900+ was made by samsung Feb 14 23:12:13 We could have had and shipped something like that 9 months ago... Feb 14 23:12:27 hrw, which one? Feb 14 23:12:36 Philippe_: nexus s Feb 14 23:12:44 need to check that one. Feb 14 23:12:49 hrw, interesting Feb 14 23:12:54 what's the modem protocol? Feb 14 23:12:58 phonet? Feb 14 23:13:00 Until then I hope you have a good night. Feb 14 23:13:07 Philippe_: lacks hw keyboard but android works nice Feb 14 23:13:07 I saw some phonet stuff in the nexus s defconfig Feb 14 23:13:09 hi j Feb 14 23:13:10 oops Feb 14 23:13:12 hi jconnolly Feb 14 23:13:14 have a nice rest of day Feb 14 23:13:31 night here too :D Should go to bed also Feb 14 23:21:24 Tartarus: you are cherry picking stuff onto release branch right ? Feb 14 23:21:33 not at this time, no Feb 14 23:21:34 so if I commit something radical to master you wont be hosed Feb 14 23:21:38 still a few more weeks of testing Feb 14 23:21:38 oh Feb 14 23:21:48 ok Feb 14 23:21:58 I want to commit gcc updates Feb 14 23:22:06 but I will wait for release to happen Feb 14 23:24:35 Tartarus: khem : in config.log I see /usr/lib/libgnutls.so.26: undefined reference to `gcry_cipher_setiv@GCRYPT_1.2' Feb 14 23:25:17 I've seen that one before Feb 14 23:25:20 you need gnutls-native Feb 14 23:26:34 there are so many issues in that file :/ Feb 14 23:26:55 Tartarus: I think it would not be wise to change compile at this moment so we have tested the existing 4.5 well Feb 14 23:27:06 and once we release out then we can have the update Feb 14 23:27:07 I think Feb 14 23:27:41 Tartarus: ...--without-gnutls --without-ssl --disable-ares Feb 14 23:31:42 * Jay7 -> sleep Feb 14 23:31:58 'nite Feb 14 23:35:23 is marvell armada armv7 or armv6 Feb 14 23:47:28 Tartarus: INHIBIT_DEFAULT_DEPS = "1" doesn't help Feb 14 23:48:26 Right, I was saying that would be the only way it could be a race was if it did do that Feb 14 23:48:44 morning Feb 14 23:49:03 it even fails a bit later Feb 14 23:49:23 | checking for sys/socket.h... (cached) yes Feb 14 23:49:23 | checking for send... no Feb 14 23:49:35 always in the same way Feb 14 23:49:51 (btw oestats doesn't report the whole log) Feb 14 23:51:34 Check what config.log says Feb 14 23:51:40 for what it's really failing there Feb 14 23:52:40 03Tom Rini  07org.openembedded.dev * r91afc3ec0e 10openembedded.git/recipes/linux/linux.inc: Feb 14 23:52:40 linux.inc: Add lzma-native for nokia900 Feb 14 23:52:40 Add a check for nokia900 to the existing DEPENDS mangler to add Feb 14 23:52:40 lzma-native, otherwise: Feb 14 23:52:40 | LZMA arch/arm/boot/compressed/piggy.lzma Feb 14 23:52:41 | /bin/sh: lzma: command not found Feb 14 23:52:41 Signed-off-by: Tom Rini Feb 15 00:27:08 XorA|gone, gm **** ENDING LOGGING AT Tue Feb 15 02:59:56 2011