**** BEGIN LOGGING AT Thu Jul 24 02:59:56 2008 Jul 24 04:18:07 * * OE Bug 4453 has been created by linuxrulez(AT)land.ru Jul 24 04:18:09 * * do_install failed: opie-taskbar-images-1.2.3 Jul 24 04:18:11 * * http://bugs.openembedded.net/show_bug.cgi?id=4453 Jul 24 06:05:06 hi RP you there ? Jul 24 06:09:23 morning Jul 24 07:26:14 hi khem, here now :) Jul 24 07:26:58 khem: good news, qemuarm in poky seemed to build fine so that patch passes my first test :) Jul 24 07:32:10 hi Jul 24 07:33:59 RP: cool Jul 24 07:34:06 he khem Jul 24 07:34:14 still not at sleep? Jul 24 07:35:13 woglinde: no working on the toolchain patch Jul 24 07:35:18 hehe Jul 24 07:35:19 okay Jul 24 07:36:10 RP: Do you plan to do more testing on it ? Jul 24 07:36:24 khem: I have a qemux86 build running as we speak Jul 24 07:36:39 if that + meta-toolchain completes, I'm happy Jul 24 07:38:07 RP: ok cool. Jul 24 07:38:23 *crosses fingers* Jul 24 07:38:41 RP: if you could reply to my post to the newsgroup and give me ack Jul 24 07:38:50 then I can commit it in the morning Jul 24 07:38:57 khem: will do Jul 24 07:39:22 khem: One other question - whats the implication on anyone with an existing build? Jul 24 07:39:27 RP: the uclibc problem remember I was struggling yesterday turned out to be problem in uclibc Jul 24 07:39:30 I have fixed that too Jul 24 07:39:38 presumably it will just build gcc-cross-intermediate but that will break things? Jul 24 07:39:40 khem 0.9.29? Jul 24 07:39:46 did not compile Jul 24 07:39:56 RP: I would reccommend a clean build Jul 24 07:40:17 khem: From experience, people don't do that unless forced to Jul 24 07:40:18 woglinde: it was 0.9.27 Jul 24 07:40:50 RP: although I think it should not be required I guess Jul 24 07:41:14 because the changes are in the intermediate steps Jul 24 07:41:24 khem: In the final version perhaps you should bump the PR for gcc and glibc so then gcc, glibc and gcc-cross-intermediate will rebuild? Jul 24 07:41:34 but still I would recommend a clean build from scratch Jul 24 07:41:49 RP: yes I will do that before commit Jul 24 07:42:02 khem: ok, I think that will be enough Jul 24 07:42:26 and run a sanity meta-toolchain build after that Jul 24 07:43:22 RP: I also fixed gcc 3.4.4 for sdk build Jul 24 07:43:37 ported --with-build-sysroot patch Jul 24 07:43:45 khem: great :) Jul 24 07:45:34 khem: Do you agree with the way we build the -sdk version of gcc? Jul 24 07:45:57 RP: yes I think thats a good way Jul 24 07:46:18 ok, just curious :) Jul 24 07:46:23 so Jul 24 07:46:28 lets fight with boost again Jul 24 07:46:36 03koen 07org.oe.dev * r3cc0283a... 10/ (1 packages/linux/linux-omap2_git.bb): linux omap2 git: reduce i2c speed and enable 4bit for SD Jul 24 07:46:36 but unfortunately this needs lot of new features like --with-build* options Jul 24 07:47:11 khem: We did used to manage without that Jul 24 07:47:20 RP: the reason to have --with-build-time-tools and --with-build-sysroot were added to help build SDKs Jul 24 07:47:29 hm do we have nx in oe? Jul 24 07:47:34 khem: Using it is a lot safer though, right :) Jul 24 07:47:41 yep Jul 24 07:47:57 less surgeries are always better :) Jul 24 07:49:12 One thing I did in poky which hasn't happened in OE yet was trimming down the CFLAGS and LDFLAGS in bitbake.conf Jul 24 07:49:15 once done with this I should be able to test my arm uclibc nptl merge Jul 24 07:49:25 now we use sysroot we can do that Jul 24 07:49:57 RP: yeah and the option --with-local-prefix can be removed for newere compilers Jul 24 07:50:05 I think thats redundant Jul 24 07:50:33 sounds good Jul 24 07:51:30 * khem should catch some sleep Jul 24 07:51:34 good night Jul 24 07:51:54 'night khem Jul 24 07:53:11 sleep well khem Jul 24 07:54:35 ~curse python Jul 24 07:54:36 May you be reincarnated as a Windows XP administrator, python ! Jul 24 07:56:41 xora whats wrong? Jul 24 07:57:24 woglinde: setuptools package failing and it dumps no useful error messages Jul 24 07:57:38 woglinde: well not usefull to a non python dude like me Jul 24 08:05:01 NameError: global name 'log' is not defined Jul 24 08:05:16 I think thats the error, googling doesnt help though :-( Jul 24 08:08:46 xora mickeyl is awake he might help Jul 24 08:09:01 mickeyl: *POKE* Jul 24 08:09:10 nah Jul 24 08:09:12 no poking so early Jul 24 08:09:17 what'sup? Jul 24 08:09:21 early? Jul 24 08:09:24 mickeyl: its an hour less early for you than me :-D Jul 24 08:09:28 :D Jul 24 08:09:28 I am awake since 7 a.m. Jul 24 08:09:31 as always Jul 24 08:09:40 mickeyl: trying to build exposure in OM.dev and its failing in setup.py Jul 24 08:09:47 mickeyl: but I cant decode the error Jul 24 08:10:18 bonjour Jul 24 08:10:29 hi genesis Jul 24 08:11:01 mickeyl: http://rafb.net/p/RdmHNI99.html Jul 24 08:11:18 we should make a bitbake -c pastebin blah.bb :-D Jul 24 08:11:29 which dumps the latest logs to pastebin Jul 24 08:11:46 ya Jul 24 08:11:48 had that in bitbake -i Jul 24 08:11:54 hmm, 'log' not defined Jul 24 08:12:00 in setuptools-native Jul 24 08:12:02 fun Jul 24 08:12:11 do you have a recipe, then I try to reproduce that Jul 24 08:12:29 mickeyl: gimme a sec Ill push OM.dev Jul 24 08:14:43 mickeyl: packages/openmoko-projects/exposure_svn.bb in org.openmoko.dev Jul 24 08:16:58 okay, let me give that a try Jul 24 08:20:15 mickeyl: cheers, python is still alien to me Jul 24 08:20:44 didn't you want to brush up your python when you joined Om? :) Jul 24 08:22:41 ok, i can reproduce it Jul 24 08:22:44 now let me dig into Jul 24 08:23:07 mickeyl: Ive been stretching slowly, but the lack of usefull errors in that dump stumps me Jul 24 08:32:59 03tmbinc 07org.oe.dreambox * r12ae6213... 10/ (3 files in 3 dirs): linux-dm8000: add patch to support more than 4 demux devices Jul 24 08:33:02 03tmbinc 07org.oe.dreambox * rf614357b... 10/ (3 files in 3 dirs): perl: fix compile Jul 24 08:35:15 truly something strange going on there Jul 24 08:42:03 03koen 07org.oe.dev * r4e6e5c3a... 10/ (5 files in 2 dirs): neuros-public: clean up apps, add mainmenu Jul 24 08:50:58 NOTE: package exposure-0.0.1+svnr57: completed Jul 24 08:51:50 XorA: you want to backport the next commit Jul 24 08:52:01 exposure triggers a bug/leftover in setuptools Jul 24 08:52:54 mickeyl: cool, thanks for fixing it Jul 24 08:53:12 np Jul 24 08:53:37 * XorA wishes were were gitted up already so I could just cherry pick Jul 24 08:54:48 *nod* Jul 24 08:55:18 03mickeyl 07org.oe.dev * r16286a01... 10/ (5 files in 3 dirs): python-setuptools[-native] 0.6c8 fix log usage bug Jul 24 09:11:13 florian: good morning Jul 24 09:11:17 mickeyl: good morning Jul 24 09:13:16 mickeyl: you is a genius Jul 24 09:13:34 XorA: too much the honor ;) Jul 24 09:13:37 hey pb__ Jul 24 09:14:35 hi all Jul 24 09:14:58 hi florian Jul 24 09:15:01 h pb Jul 24 09:23:14 morning Jul 24 09:23:31 hi hrw Jul 24 09:58:54 helo Jul 24 09:59:35 auto-movind brightness control slider in gpe-settings module is ok? Jul 24 10:00:07 it decreases brightness automa[tg]ically Jul 24 10:00:23 tested on akita Jul 24 10:06:31 Are we likely going to see the git switch this week? Jul 24 10:06:48 How is the internal discussion progressing? Jul 24 10:06:54 Any stumbling blocks left? Jul 24 10:07:00 mickeyl: ? Jul 24 10:07:59 goodmorning Jul 24 10:08:42 Laibsch: last I saw there was concensus forming Jul 24 10:08:45 Laibsch: amazingly Jul 24 10:09:05 XorA: unbelieving indeed Jul 24 10:09:16 unbelievable Jul 24 10:09:49 but there still are a few stumbling blocks left? Jul 24 10:10:01 Laibsch: I think we are down to niggles and cleanups Jul 24 10:10:11 Laibsch: but Ive been busy with RL crap recently Jul 24 10:10:35 You guys are working collaboratively on the release text which includes the policy? Jul 24 10:10:51 Which brings me to an OT question I had Jul 24 10:10:58 Laibsch: shoot Jul 24 10:10:59 [OT] I anybody aware of a tool to collaborate on text creation using a simple browser interface? I know this sounds a lot like a wiki and it is very close. But the object of the collaboration would be a 1MB xml file. It being XML will likely lead to confusion, when the wiki engine and the browser try to interpret the XML, I think. Jul 24 10:11:19 I found gobby and it seems pretty nifty Jul 24 10:11:28 we used gobby at OEDEM Jul 24 10:11:35 nice Jul 24 10:11:39 abiword also has collabaration now Jul 24 10:12:14 I tried it out yesterday with the gobby devs and it immediately choked on the 1MB file Jul 24 10:13:33 collabarating by editing raw xml scares me though Jul 24 10:14:29 I have not been able to test abiword yet for lack of a partner. It is good that it has a Windows binary available, but I wonder if the barrier would be a little too high (Jabber setup plus installation of abiword) Jul 24 10:14:35 XorA: absolutely agreed Jul 24 10:14:49 But it is one of the reasons I want to get rid of doing it all by myself Jul 24 10:15:32 It is a gnucash chart of accounts and believe I have cursed the gnucash devs many times for their choices Jul 24 10:16:12 Luckily, python bindings were added recently and maybe I can manipulate the data via python from now on. That would make things a lot safer Jul 24 10:16:46 The XML file holds the definition of a standard set of accounts for German business Jul 24 10:20:21 morning Jul 24 10:22:38 ~curse python Jul 24 10:22:39 May you be reincarnated as a Windows XP administrator, python ! Jul 24 10:26:54 what happens with python? Jul 24 10:27:07 i see XorA and hrw cursing python :) Jul 24 10:29:25 does not want to build on my hacked system Jul 24 10:29:41 staging/arm-poky-linux-gnueabi/usr/lib/python2.5/config/Makefile (No such file or directory) Jul 24 10:33:58 something wrong with distutils I think Jul 24 10:36:18 ~curse lack of euro in europe Jul 24 10:36:19 May the fleas of a thousand camels infest your most sensitive regions, lack of euro in europe ! Jul 24 10:45:49 03koen 07org.oe.dev * rb1834de8... 10/ (1 contrib/angstrom/source-mirror.txt): contrib/angstrom: add source mirror description Jul 24 10:45:55 03koen 07org.oe.dev * r8ad08e33... 10/ (1 conf/distro/angstrom-2008.1.conf): angstrom 2008: gpe 2.8 is so last year, use latest versions from now on Jul 24 11:23:08 hrw: hello, I think you already pointed out this: install -m 0755 tools/mkimage ${STAGING_BINDIR_NATIVE}/ is both in u-boot.inc and u-boot-utils-native .Is there a specific reaon? Jul 24 11:33:17 it should be in native only rather but some people build uboot and kernel in one run so prefer that way Jul 24 11:35:07 ah, ok, I'm trying to automate the umage for Zaurus Jul 24 11:35:23 but in kernel.bbclass I see u-boot-mkimage-openmoko-native Jul 24 11:35:24 ? Jul 24 11:37:40 mickeyl seems to be the author Jul 24 11:40:23 # Add dependency on mkimage for kernels that build a uImage Jul 24 11:40:29 depends = "%s u-boot-mkimage-openmoko-native" % depends Jul 24 11:40:53 i used a custom distro derived from angstrom-2008.1.conf. Is there an automatic tool that help to update the version of package fix in a distro ? Jul 24 11:41:10 genesis??? Jul 24 11:41:40 ant|work: thats my change Jul 24 11:41:48 why not just calling u-boot-utils-native? Jul 24 11:42:05 because they are newer then this change? Jul 24 11:42:09 let imagine PREFERRED_VERSION_busybox = "1.7.2" in my distro Jul 24 11:42:13 ahh ;-) Jul 24 11:42:22 morning all Jul 24 11:42:31 hi thesing Jul 24 11:42:39 yo thesing Jul 24 11:42:39 thesing holidys now? Jul 24 11:42:50 genesis no I dont think so Jul 24 11:42:57 oki i look what i could do Jul 24 11:43:26 till later Jul 24 11:44:08 woglinde: only in theory. I work on my diploma thesis. Jul 24 11:45:11 * ant|work vaguely remembers when he was student... Jul 24 11:46:55 cd */*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/ Jul 24 11:47:21 heh.. two such ones worked.. Jul 24 11:48:13 hrw: you need GPS Jul 24 11:48:30 ant|work: I have 3 devices with gps here Jul 24 11:48:50 when you browse your tree :-) Jul 24 11:48:52 but two of them are neo1973 and third is n810... Jul 24 11:49:04 ach Jul 24 11:49:29 bad luck...then the issue of maps Jul 24 11:50:07 ah, you know what? The father ofmy wife is 'Geodeta' professor at university in Poland Jul 24 11:50:16 he could provide you some maps ;-] Jul 24 12:01:47 angstrom x11-images got really big. They don't even fit on a 32 MB mmc card. Jul 24 12:02:24 thesing, how did you find such a small card :) Jul 24 12:03:51 it's from hrw. He didn't need is for obvious reasons ;) Jul 24 12:04:05 Crofton|work: did you solve you nfs problem? Jul 24 12:04:18 yeah Jul 24 12:04:29 util-linux-ng has/had a problem Jul 24 12:04:53 koen has updated to a later version that appears to solve the problem Jul 24 12:05:00 I haven't had tie to check though Jul 24 12:05:12 I worked around the problem but installing util-linux :) Jul 24 12:06:19 so you didn't have a "real" mount in the end ;) Jul 24 12:11:32 is there a simple way to do build and use a program natively as part of the build in a bb description? Jul 24 12:12:14 i want to calc a simple checksum and put it in the code when building Jul 24 12:13:22 oliford: sure Jul 24 12:13:38 oliford: let it redepend on your checksum-native thing? Jul 24 12:14:29 dcordes_: no need Jul 24 12:14:42 oliford: you have to use proper cc Jul 24 12:15:19 some recipes do that Jul 24 12:15:29 oliford: one of them is mplayer Jul 24 12:19:00 oic, so i need to make a separate package which compiles entirely natively? Jul 24 12:19:24 ok, i'll look at mplayer and figure it out Jul 24 12:19:42 oliford: no need for separate Jul 24 12:20:08 in makefile you need to use proper CC for target code and proper one for host code Jul 24 12:22:41 ah yes, it uses $(HOST_CC), thats what i was looking for Jul 24 12:31:13 no, it uses BUILD_CC which doesnt exist for my package Jul 24 12:31:38 what is the correct var that gives you the host compiler? Jul 24 12:33:14 oliford: why don't you put a placeholder in the code and replac it with sed in do_configure_prepend? Jul 24 12:33:31 ant|work: because it is not needed usually? Jul 24 12:33:42 err, replace it with what? Jul 24 12:33:49 it seems to me enough for a md5sum ... Jul 24 12:34:03 how does oe give me the host's compiler? Jul 24 12:34:04 can use shell comands Jul 24 12:34:22 oliford: BUILD_CC which should be set by oe Jul 24 12:34:26 it isnt Jul 24 12:34:41 I be back in 20m Jul 24 12:35:47 oliford: perhaps I don't get the issue: how do you calc the checksum? Jul 24 12:36:09 just adding up in 32-bit Jul 24 12:36:32 at the time i did it, i couldn't find a util which did it simply, was quicker to write one Jul 24 12:36:52 ok, I see Jul 24 12:38:26 its a simple not-really-a-bootloader, which has the kernel image compiled into it. i copies the kernel to where it needs to be and does the cheksum as it goes Jul 24 12:39:48 the actual loading is done by the ipaq's proper bootloader, which is designed for windows so obviously doesnt create tags structure Jul 24 12:40:43 I'm battling with an insane bootloader too... Jul 24 12:42:10 but basically you rewrote smthg like mkimage (for uImages) Jul 24 12:42:51 I'm just looking at how to integrate this step in the u-boot deployment Jul 24 12:43:00 (for Zaurus) Jul 24 12:43:21 s/Zaurus/(some of) Zaurus/ Jul 24 12:44:19 c7x0/akita basically Jul 24 12:44:35 yep, lot of work done Jul 24 12:45:05 bb attached to oebug 3378 Jul 24 12:48:41 yea, i didn't want to get involved with u-boot. it wanted to keep the code incredibly simple so nothing goes wrong Jul 24 12:48:51 (i have no serial or otherwise connection into the thing) Jul 24 12:49:59 *i wanted... Jul 24 13:02:37 if I put patches in OE do I have to rename them every time I change them? (i.e. is SRC_URI = file://... different from SRC_URI= http://... ?) Jul 24 13:03:23 thesing???? Jul 24 13:04:10 thesing: you can keep http:// Jul 24 13:04:23 as long as they are fetchable Jul 24 13:04:45 if I have a patch foo.patch on webserver somewhere and I change it, I have to change the name of the patch (and in the recipe) in order to make bitbake fetch the changed version. Jul 24 13:05:11 my question is if I have to do the same if I store the patches in OE. Jul 24 13:05:40 not if you update patch in Oe Jul 24 13:06:21 great. I always mixed the different patch version up. Jul 24 13:17:47 hrw: it was oebug 3738 actually Jul 24 13:25:08 hrw: other subject: Attached to oebug 4118 are new bootlogs, some with udev-124 and poky's initscript (first boot and reboot) Jul 24 13:29:14 florian: ^^ in the logs there is a ton of alignment traps and one alignment exception too Jul 24 13:30:06 hi gnutoo Jul 24 13:30:49 woglinde, hi Jul 24 13:31:56 woglinde, at what point are you, should i test something? Jul 24 13:33:14 wesnoth? Jul 24 13:33:17 yes Jul 24 13:33:33 I am fighting with boost right now Jul 24 13:33:33 and sdl... Jul 24 13:33:38 ah ok Jul 24 13:33:42 but hats the last step Jul 24 13:33:54 ok...and thanks a lot Jul 24 13:33:55 thats Jul 24 13:34:20 you own me a beer for the boost building mess Jul 24 13:34:51 in what country are you> Jul 24 13:34:57 s/>/? Jul 24 13:35:05 ah germany Jul 24 13:35:32 i'm in italy and sometimes go in france... Jul 24 13:36:03 what's the exact problem with boost...with me it builds fine... Jul 24 13:36:23 yes Jul 24 13:37:41 wesnoth dont find it Jul 24 13:37:46 and 1.3.33 is really old Jul 24 13:37:48 ah ok Jul 24 13:38:50 good morning! Jul 24 13:38:57 hi frikker Jul 24 13:39:05 i am trying to build the bitbake package for octave - has anyone had experience with this? Jul 24 13:39:21 the problem is that i need a fortran compiler, like g77 or f77, but i'm not quite sure how to add this to my toolchain? Jul 24 13:39:37 hi woglinde :) Jul 24 13:41:25 octave.... Jul 24 13:41:36 you need to talk with lardman Jul 24 13:41:50 as octave is in his playground Jul 24 13:43:25 thank you very much hrw :) Jul 24 13:44:30 any general tips on how to add a fortran compiler to the build tools? Jul 24 13:45:05 ~seen lardman Jul 24 13:45:09 lardman was last seen on IRC in channel #maemo, 4d 17h 56m 43s ago, saying: 'bye chaps'. Jul 24 13:45:23 ~lardman Jul 24 13:45:38 oh cool Jul 24 13:45:46 ~lardman Jul 24 13:45:48 frikker: lardman is Simon Pickering - look in OE archives to get mail Jul 24 13:45:49 thanks Jul 24 13:46:49 hi chouimat Jul 24 13:48:20 03woglinde2 07org.oe.dev * rc678660c... 10/ (4 files in 3 dirs): libsdl-net: fix building with libtool2 Jul 24 13:48:24 03woglinde2 07org.oe.dev * r87f02842... 10/ (4 files in 3 dirs): libsdl-image: fix building with libtool2 Jul 24 13:48:29 03woglinde2 07org.oe.dev * rbd0863ef... 10/ (1 packages/libsdl/libsdl-ttf_2.0.3.bb): libsdl-ttf: use autotools_stage_all Jul 24 14:22:01 * ant|work is astonished: there is u-boot for tosa and for spitz http://www.pdaxrom.org/?q=node/240 Jul 24 14:24:09 still 1.1.4? Jul 24 14:24:46 hopefully not, probably yes... Jul 24 14:51:49 Hi! The OE bug tracker shows no open compilation bugs, but a couple of us are having the same problems building angstrom-2007.1 (one of us on an h2200, the other on an hx4700). We get QA errors, which may or may not be part of the root cause, and the end result is that whatever we try to build won't build. In my case (hx4700) I end up with an unkillable qemu-arm process loading one of my cores at 100%. Should I file a bug report? Jul 24 14:57:45 hi mickeyl Jul 24 15:00:49 m1k3y Jul 24 15:01:37 awesome - i didnt end up using g77 or f77 for the octave isntall Jul 24 15:01:43 but theres a config option for --with-f2c Jul 24 15:01:55 which uses a fortran to C translator instead of a fortran compiler Jul 24 15:02:03 Humanl: the QA errors and other 'minor' issues are solved in 2008.1. Jul 24 15:02:24 now thats robust :) so i can install f2c on my host machine instead. Jul 24 15:03:55 I for one guess there are more build-issues in 2007.1 than in 2008.1 Jul 24 15:04:16 Humanl: but wait, toolchain changes are coming soon Jul 24 15:04:31 Human1: hold on one second Jul 24 15:05:08 OK, I was surprised about your "The OE bug tracker shows no open compilation bugs" Jul 24 15:05:16 I count 339 Jul 24 15:05:17 ant|work That sounds promising, but I'm just trying to build a single ipk. Jul 24 15:05:32 as defined by the search from the link on http://bugs.openembedded.net/ Jul 24 15:06:00 Humanl: I meant wait to switch to .dev Jul 24 15:06:01 If there is no bug yet, then please, open a new ticket Jul 24 15:06:03 Laibsch I must have clicked on a bad link. I tried from http://bugs.openembedded.net/ and yeah, I see a lot. Jul 24 15:06:47 Human1: what package are you trying to build? Jul 24 15:07:28 python-pexpect for Angstrom 2007.12 Jul 24 15:07:34 he he Laibsch, you did implement the queries for the lazy, nice :-) Jul 24 15:07:34 and the hx4700 Jul 24 15:09:12 ant|work: If it helps get more exposure to the well-triaged bugs, I have won a lot Jul 24 15:09:19 and saved myself some work Jul 24 15:09:46 sane words Jul 24 15:10:47 The build process lies to me, in that it says the build completed, but there's no ipk: Jul 24 15:10:47 bitbake python-pexpect Jul 24 15:10:47 [...] Jul 24 15:10:47 NOTE: package python-pexpect-2.1: started Jul 24 15:10:47 NOTE: package python-pexpect-2.1-r0: task do_qa_staging: started Jul 24 15:10:47 NOTE: QA checking staging Jul 24 15:10:49 [bunch of QA errors] Jul 24 15:10:51 NOTE: package python-pexpect-2.1-r0: task do_qa_staging: completed Jul 24 15:10:53 NOTE: package python-pexpect-2.1: completed Jul 24 15:12:48 where did you look for the ipk? Jul 24 15:12:58 everywhere under my build root Jul 24 15:13:03 task 511 of 1020 here Jul 24 15:13:32 have you found other ipk? Jul 24 15:13:37 any ipk at all? Jul 24 15:13:51 Let me check. Jul 24 15:14:54 I did successfully build stuff a couple of months ago using the same procedure, so something has changed. The Angstrom instructions (from 2006) say to use bitbake 1.8.8. Is that still the recommendation? Jul 24 15:16:40 No ipks found, but I was only trying to build python-pexpect, so I don't know if that would have generated other ipks as side-effects or not. Jul 24 15:17:31 well what do you think? Jul 24 15:18:36 ant|work what do I think? Jul 24 15:18:51 I use svn co svn://svn.berlios.de/bitbake/branches/bitbake-1.8/ bitbake Jul 24 15:19:34 I got 1.8.8 because http://www.angstrom-distribution.org/building-angstrom links to it. Jul 24 15:19:45 ah, angstrom ... Jul 24 15:19:50 And it worked a couple of months ago when I built another package. Jul 24 15:21:41 BTW, just so I'm not leeching, I did help fix the "stuck with UTC" timezone bug: http://bugs.openembedded.net/show_bug.cgi?id=3315 Jul 24 15:22:32 yes, I've read it...I noted the base files does not provide a valid /etc/localtime Jul 24 15:23:01 but just touching it won't work imho Jul 24 15:23:18 perhaps I'm wrong.... Jul 24 15:23:41 I don't have /etc/localtime on my device, and I get the right timezone and local time... Jul 24 15:23:54 cat /etc/localtime Jul 24 15:23:54 cat: /etc/localtime: No such file or directory Jul 24 15:23:54 root@hx4700:~$ date Jul 24 15:23:54 Thu Jul 24 11:19:04 EDT 2008 Jul 24 15:24:00 /home/laibsch/oe/tmp/angstrom/dev/deploy/glibc/ipk/armv5te/python-pexpect_2.1-r0_armv5te.ipk Jul 24 15:24:06 compiled just fine Jul 24 15:24:40 Human1: You need to find out what you set $TMPDIR to Jul 24 15:24:48 Laibsch 1) THANK YOU FOR THAT! :) 2) I'm not the only one having trouble building things for Angstrom lately. Are their build directions maybe broken? Jul 24 15:24:58 I don't know Jul 24 15:25:06 I don't use angstrom, really Jul 24 15:25:28 What did you set as your distro when you did the build? Jul 24 15:25:30 Human1: What distro are you compiling on? Jul 24 15:25:40 I'm on Ubuntu 7.10 x86 Jul 24 15:25:43 Human1: in that case it was angstrom Jul 24 15:25:49 Human1: I have something for you Jul 24 15:26:17 http://blog.leggewie.org/?p=39 Jul 24 15:26:27 That will get you a working setup for sure Jul 24 15:27:25 Very cool. Jul 24 15:28:02 If you glance at the instructions at http://www.angstrom-distribution.org/building-angstrom do they appear to be wrong/outdated to you? Jul 24 15:28:36 Everyone who tries to build Angstrom goes there, and as I said, it worked a couple of months ago, but the latest updates seem to have introduced badness. Jul 24 15:29:41 * Laibsch does not support angstrom Jul 24 15:29:46 I think L. would have DNS problems pointing to that URL right now...8-/ Jul 24 15:30:00 BTW, I know my $TMPDIR is /OE/angstrom-tmp, and there's no ipk in it. Jul 24 15:30:39 Laibsch, I understand that you don't support derivations of OE, but I was just wondering if anything stood out as being wrong in their build instructions, so I could help them fix it. Jul 24 15:30:43 sudo find / -name *.ipk Jul 24 15:31:13 Humanl: and check twice your local.conf Jul 24 15:31:16 Laibsch, I did an updatedb and a locate and no .ipk files came up that I didn't build months ago. Jul 24 15:32:39 ant|work does the local.conf need to change as the build sources evolve? Because it's the same one I used a couple months ago when it worked. Jul 24 15:33:48 Human1: pastebin your local.conf Jul 24 15:33:53 and site.conf if applicable Jul 24 15:34:02 I just have local.conf, but here it is: Jul 24 15:34:39 (sorry, what do you mean by pastebin?) Jul 24 15:35:07 http://pastebin.com/ ? Jul 24 15:35:30 http://pastebin.com/m6cb55cea Jul 24 15:37:22 sudo find /OE/angstrom*/ -name *pexpect* > /tmp/pastebin_me.txt Jul 24 15:38:12 It'll take a while, since qemu-arm is eating up half my cores and won't be killed. Jul 24 15:38:26 take your time Jul 24 15:39:46 btw the script uses ${DISTRO} before DISTRO = Jul 24 15:40:17 DISTRO was set by 'source-me.txt' as per http://www.angstrom-distribution.org/building-angstrom Jul 24 15:40:47 http://pastebin.com/m16840dff Jul 24 15:41:02 Er. No it's not :) Jul 24 15:41:30 Yeah, that's strange. Jul 24 15:41:43 I don't know why that would work at all, unless the syntax curries it somehow. Jul 24 15:44:43 morning Jul 24 15:45:32 hi kergoth`work Jul 24 15:48:21 hi kergoth`work Jul 24 15:48:22 RP: When building sdk with gcc 3.4.4 I am seeing one problem. while building cross-gcc-sdk xgcc is using the native assembler 'as' instead of cross assembler. How did you set the links before ? in newer gcc we can use --build-time-tools config option but we need links setup manually for older gcc Jul 24 15:53:03 does anybody here have the eleet usb 5ki11z? Jul 24 15:53:08 * pb__ in some kind of interrupt related hell Jul 24 15:54:53 urks Jul 24 15:58:04 khem: I think in older versions we set something like AS_FOR_TARGET? Jul 24 15:58:44 seems gcc 3.4.4 is not honoring that Jul 24 16:01:01 i'm in networky hell if anyone happens to be good at writing/debugging network card drivers :) Jul 24 16:02:15 is this the right place to ask about packages in the open embedded repository? Jul 24 16:02:47 * RP is in binary only gles driver hell :) Jul 24 16:02:57 sw17ch: yes Jul 24 16:03:20 RP: great! it relates to the ruby package... i'm wondering how one gets socket support for ruby Jul 24 16:03:22 * chouimat|work is in printer hell :) Jul 24 16:04:54 * khem is alone in heaven no wonders Jul 24 16:05:30 khem: All my tests worked out so I have no objection to the toolchain changes being committed Jul 24 16:05:42 RP: very cool. Jul 24 16:06:12 chouimat|away: you too? Jul 24 16:06:35 hmm, is no-one building perl stuff, seems it doesnt have a clue how to handle sysroot Jul 24 16:07:14 XorA: it has no clue how to handle cross compiling full stop Jul 24 16:07:28 RP: configure says this hecking for arm-angstrom-linux-uclibc-as... /home/kraj/work/oe/build/uclibc/tmp-omap5912osk/cross/arm-angstrom-linux-uclibc/bin/as Jul 24 16:07:35 RP: its looking for /usr/lib/perl/5.8.8/site/Config.pm Jul 24 16:07:45 RP: I cant beleive its been broken this long Jul 24 16:07:50 XorA: perl is terribly broken Jul 24 16:10:10 hi guys Jul 24 16:10:12 RP: I will commit it after work. Now time to go behind the bars Jul 24 16:10:36 is fakeroot 1.9.4, the latest versiona available? Jul 24 16:10:56 because... there is no mirror to this version Jul 24 16:12:04 RP: would it be better to user opt-out instead of opt-in with the default preferences of linux-rp? Jul 24 16:14:23 XorA: I stumbled on this days ago: bug 4392 -> 3403 Jul 24 16:14:57 XorA: and hrw too said perl is broken Jul 24 16:15:13 there is a bug on perl Jul 24 16:16:23 hmm, those bugs are not dupes Jul 24 16:17:08 or at least comment #2 on 4392 is different bug than 3403 Jul 24 16:17:40 yes, tried other modules Jul 24 16:17:43 * XorA bets perl worked by luck until debin upgraded to 5.10 Jul 24 16:37:04 hi ph5 Jul 24 16:39:51 bye Jul 24 16:40:33 kergoth`work: heh. well, I know more about network cards than about usb. what's your issue? Jul 24 16:40:56 anybody who admits to knowing about USB is a dirty, dirty liar Jul 24 16:41:00 ~curse perl Jul 24 16:41:01 May the fleas of a thousand camels infest your most sensitive regions, perl ! Jul 24 16:41:49 CosmicPenguin: yah, true enough. in this instance though I guess my complaint is more with the crazy kernel dudes than with usb itself. Jul 24 16:42:20 somehow the pointer to the packet data (points to an address returned by dma alloc single) that's in the hardware descriptors is being corrupted, it ends up being an address outside of lowmem Jul 24 16:42:53 in the packet receipt functions, it yanks a new descriptor from the ring, and the address is invalid... it deosnt appear to have been when the ring was allocated, or when a descriptor is returned to the ring Jul 24 16:42:55 oh dear. and presumably you're sure it was a valid address when you poked it in there, right? Jul 24 16:43:21 i cant think of anything else it could be, unless its something up with the ethernet dma engine.. i dunno enough about the hardware either. Jul 24 16:43:29 * kergoth`work wonders if mv has a datasheet somewhere in the intranet Jul 24 16:43:36 what's your platform? Jul 24 16:44:56 Motorola PPMC-275 board, powerpc, its some marvell chipset that has the ethernet controller, which has no specs on the web at all Jul 24 16:45:04 naturally i dont know powerpc either :) such a lovely bug for me... Jul 24 16:45:17 heh, right Jul 24 16:45:23 I don't know powerpc either Jul 24 16:45:51 might be worth finding out if its caches are naturally dma coherent, or if it has arm-like requirements for explicit flushing around dma activity Jul 24 16:45:51 ah thats it, mv64360 chipset Jul 24 16:46:10 CONFIG_NOT_COHERENT_CACHE=y in the .config :) Jul 24 16:47:01 on an unrelated note, anyone done bash or zsh completion for bitbake, including providers? :) Jul 24 16:47:15 bitbake virtual/ would be nifty Jul 24 16:47:38 aaaaaaaaaaaaargh, who allowed perl to recurively run MakeMaker Jul 24 16:48:22 pb__: heh, think i need to brush up on my computer architecture, dma cache coherency issues always confuse me and give me a headache Jul 24 16:48:42 * kergoth`work waits for 6 technical books he ordered the other day to arrive, including one on that subject, and one on linux networky stuff Jul 24 16:48:51 hi woglinde, everyone Jul 24 16:49:41 03koen 07org.oe.dev * r5427f220... 10/ (1 packages/vlc/vlc-davinci_0.8.6h.bb): vlc-davinci: fix depends Jul 24 16:49:50 kergoth`work: mickey's shell has that Jul 24 16:50:03 does it? nice Jul 24 16:50:05 mickeyl: boo Jul 24 16:50:15 XorA: Someone with severe drainbamage.. Jul 24 16:50:22 figure if i'm gonna start messing with oe again, anything to save time is a Good Thing Jul 24 16:50:56 indeed Jul 24 16:51:02 before i chase down other paths, has any one here managed to get ruby working on the gumstix hardware with socket support? Jul 24 16:51:25 kergoth`work: heh Jul 24 16:51:38 balls I just tried to climb a set of stairs I no longer have :-( Jul 24 16:51:44 i'm, unfortunately, quite new to how OE goes together and not quite sure how to dissect the packages to figure out what's ahppening Jul 24 16:51:49 khem: That as isn't a cross compiling as? I'll catch up when you get back from behind the bars... Jul 24 16:52:33 XorA: They were replaced with a wall or a void? Jul 24 16:52:53 RP: I moved out of my house to a flat, flat doesnt have stairs to bathroom :-D Jul 24 16:53:29 so nice Jul 24 16:53:33 wesnoth compiles Jul 24 16:54:06 XorA|dinner: Could be worse. I had visions of you falling down a hole ;-) Jul 24 16:54:48 thesing: Either can be used depending on the circumstances Jul 24 16:56:39 RP: at the moment we use opt-in which results in more lines as more machines get updated to use it. I thing the reverse would be better. Jul 24 16:59:26 * kergoth`work thinks maybe he'll email the netdev list for advice on this issue Jul 24 17:02:08 thesing: that doesn't clearly show which machines have been tested and work though :/ Jul 24 17:18:42 Laibsch, FYI, the link I followed to get the empty list of compilation bugs was at http://blog.leggewie.org/?p=40#comment-3366 - I just posted a comment so Rolf can update the link. Jul 24 17:19:27 Laibsch also, I followed the process at http://blog.leggewie.org/?p=39 to build python-pexpect for the hx4700 and Angstrom, but it failed :( Jul 24 17:20:32 http://pastebin.com/m46691756 Jul 24 17:20:38 I'll try troubleshooting. Jul 24 17:31:39 * piroko cries Jul 24 17:34:13 can anybody tell me how to control which static device nodes are create in images? Jul 24 17:37:48 Laibsch, it was a permissions issue. I posted on http://blog.leggewie.org/?p=39 so if others follow the instructions, they'll know to address it. Jul 24 17:38:07 Build is chugging along... Jul 24 17:51:28 thesing: hi, is the machine config using devfs or udev? Jul 24 17:52:20 just plain files. But I found it its in files/device_table_minimal.txt Jul 24 17:53:03 thesing: ok Jul 24 17:54:13 mr_nice: did you try colliets for simpad? Jul 24 17:54:35 thesing: is there a new one? Jul 24 17:55:17 thesing: you mean the touchscreen driver patches from collie or? Jul 24 17:55:46 yes. Jul 24 17:56:25 mr_data was confusing gpio to irq mapping so it didn't work in irq mode. Jul 24 17:56:26 thesing: I am still at kernel 2.6.24 and was not able to watch the changes to 2.6.26. are they now in mainline or do you have new patches? Jul 24 17:56:53 its still the same patch. (modified a bit) Jul 24 17:57:00 thesing: I can try them if you give me a link Jul 24 17:58:51 aloisiojr: I have 1.9.5 here. Are you on .dev? Jul 24 18:00:56 mr_nice: http://git.linuxtogo.org/?p=thesing/collie.git;a=history;f=drivers/mfd/collie-ts.c;h=ddde5fc3505dd5b6fae7ef1e35ba6a4588dc2f1e;hb=a58e147b07f7eb6b4fe8d8ed181d43839d62d15a Jul 24 18:01:11 he mr_nice Jul 24 18:01:53 woglinde: hi Jul 24 18:02:12 thesing: thx I will try them - could take a day or two Jul 24 18:02:33 mr_nice simpad patches for 2.6.26? Jul 24 18:02:50 woglinde: do you know how I can work on oetmp dir and get the changes compiled with bitbake Jul 24 18:03:14 no problem. I want to merge collie-ts and ucb-ts in the future. Jul 24 18:03:15 mr_nice? Jul 24 18:03:16 woglinde: not yet but I have some time and will start with 2.6.26 Jul 24 18:03:49 mr_nice :) Jul 24 18:04:02 woglinde: I currently try to do a bit opie stuff for the simpad. for that I want to recompile libopiecore2 Jul 24 18:04:20 mr_nice: if you want to push stuff upstream I recommend using git and bleeding edge kernel. Jul 24 18:04:35 thesing: good point Jul 24 18:04:49 mr_nice and where is the problem? Jul 24 18:05:32 woglinde: how can I do that? If I change the source and type bitbake opie-image the lib md5sum is still the same Jul 24 18:06:27 Laibsch: i've have pulled yesterday, from git repo Jul 24 18:06:35 Laibsch: i'm in master branch Jul 24 18:06:45 mr_nice hm???? Jul 24 18:07:04 mr_nice step for step what you are doing and whant to achieve Jul 24 18:08:43 woglinde: I am working on odevice_simpad.cpp and want to get libopiecore2 rebuild with the changes Jul 24 18:09:52 mr_nice hm okay Jul 24 18:09:59 you can bump the PR Jul 24 18:10:26 mr_nice okay make a patch Jul 24 18:10:39 put the patch with file:// in Jul 24 18:10:43 bump the PR Jul 24 18:11:16 woglinde: hm, ok I will try that Jul 24 18:11:16 then bitbake -c clean libqpe or libopie2 dont exactly know where odevice is used Jul 24 18:11:23 bitbake opie-image Jul 24 18:17:41 hi prioko Jul 24 18:19:40 03koen 07org.oe.dev * r3d20f1d1... 10/ (14 files in 7 dirs): mythtv 0.21: enable appropriate ARM optimization dependant on cpu and patch in NEON support for video Jul 24 18:21:37 What's the minimum recommended ARM CPU to run mythtv for standard-def video? Jul 24 18:27:25 Or do people only run backends on it? Jul 24 18:29:56 Human1: AFAIK koen only just added it for the beagleboard, I think there aren't many data points for lower powered arm platforms. Jul 24 18:31:31 pH5 Cool, thanks. I was having visions of people running the frontend on PDAs and wondering how this was possible :) Jul 24 18:31:33 brb, rebootong kc Jul 24 18:33:48 wogilnde: that worked thanks. btw I tried out poky with sato on the simpad and it could be a nice interface for the pad. Jul 24 18:34:51 mr_nice sure why not Jul 24 18:38:35 hms what was the variable for stagedir-arch Jul 24 18:39:04 ${STAGING_DIR} points only to oepath/staging Jul 24 18:40:33 okay handbook helps Jul 24 18:42:45 03koen 07org.oe.dev * r018282b0... 10/ (1 packages/mythtv/mythtv-0.21/ffmpeg-arm-update.diff): mythtv 0.21: fixup patch for armv7 Jul 24 19:16:45 03koen 07org.oe.dev * r1f4f0752... 10/ (3 files in 2 dirs): linux omap2: add missing patches Jul 24 19:22:33 03cbrake 07org.oe.dev * ra8931960... 10/ (9 files in 3 dirs): Jul 24 19:22:33 mono-1.9.1: first pass at adding mono 1.9.1 to OE Jul 24 19:22:33 still requires some manual intervention, but it now Jul 24 19:22:33 all builds. Not tested yet. Jul 24 19:25:07 * * OE Bug 4337 has been RESOLVED (INVALID) by linuxrulez(AT)land.ru Jul 24 19:25:09 * * linux-handhelds-2.6_2.6.21 fails do_configure Jul 24 19:25:11 * * http://bugs.openembedded.net/show_bug.cgi?id=4337 Jul 24 19:45:22 Where would I place code snippet to be sourced by a bash script? /usr/lib/$packagename? Jul 24 19:45:44 I want to carve out common functionality for my oe-setup scripts Jul 24 19:45:50 Laibsch: /usr/share? what does the FHS (filesystem hierachy standard) say? Jul 24 19:46:03 I don't know what the FHS says Jul 24 19:46:14 And it would take me days to understand it Jul 24 19:46:23 which is why I came here ;-) Jul 24 19:55:13 zecke: /usr/share scheint die korrekte Wahl zu sein Jul 24 19:56:14 wrong channel Jul 24 20:07:41 03bluelightning 07org.oe.dev * r5310e6e2... 10/ (1 packages/h2200-bootloader/h2200-bootloader.bb): Jul 24 20:07:41 h2200-bootloader: ensure mtdblock is available and loaded Jul 24 20:07:41 * Add kernel-module-mtdblock to RRECOMMENDS Jul 24 20:07:41 * modprobe mtdblock before accessing mtdblock device Jul 24 20:07:46 03bluelightning 07org.oe.dev * red4e5b42... 10/ (1 packages/h2200-bootloader/h2200-bootloader.bb): h2200-bootloader: slight correction to last commit Jul 24 20:25:45 khem: if you're there, sry I haven't been able to test-drive your work on the toolchain. Too busy with two short-term projects...(hussle hussle) Jul 24 20:33:49 likewise, did you ever figure out why linking c++ on ppc does not find the stdc++ library? Jul 24 20:34:20 Crofton|work: Cannot remember. It this a bug#? Jul 24 20:34:32 not sure Jul 24 20:34:38 I'll have to look Jul 24 20:36:56 Crofton|work: I see that on ARM as well... libstdc++ ends in a directory not in our -L libpath is the best guess Jul 24 20:37:10 hmm Jul 24 20:37:18 Crofton|work: so we link and see SO_NEEDED libstdc++ but gcc (the linker) does not know where g++ stuff is Jul 24 20:37:29 I have stuff that builds ok on arm, but not ppc Jul 24 20:37:30 weird Jul 24 20:37:33 Crofton|work: use g++ to link c++ apps Jul 24 20:37:42 Crofton|work: different gcc versions? what do you use to link? Jul 24 20:37:48 libtool is involved ... Jul 24 20:37:57 yes different gcc versions Jul 24 20:38:47 I'm only sort of playing with it atm, so won't make any progress Jul 24 20:39:07 Crofton|work: is it using gcc to link? Jul 24 20:39:14 http://git.openmoko.org/?p=openmoko.git;a=commit;h=4c05c474bf939d346fe0c92ce8c8ea3fc8f4c374 <- is an evil workaround... Jul 24 20:39:48 heh Jul 24 20:40:09 it is not critical for me atm, I was just doing a build to check some stuff Jul 24 20:40:18 I'll need to come back to it later :( Jul 24 20:40:39 Crofton|work: could you paste the error somewhere? not that I will have the time to fix it now... Jul 24 20:41:15 http://tinderbox.openembedded.net/public/logs/594420.txt Jul 24 20:41:31 http://tinderbox.openembedded.net/packages/78717/ Jul 24 20:45:08 Laibsch, did you get any "ERROR: QA Issue" badness when you built python-pexpect? I got less than I did when I tried Angstrom's build method, but even using the deb-based setup, it seems to fail in the same way: http://pastebin.com/m6806b039 Jul 24 20:45:35 Now I have two qemu-arm processes pegging both my cores... Jul 24 20:52:52 I think I'll bite the bullet and reboot to kill the unkillable qemu processes, then re-try the python-pexpect build. Jul 24 20:52:56 brb, I hope Jul 24 21:00:51 khem: I commited the toolchain changes into poky now so I could catch the next round of automated tests, I hope you don't mind Jul 24 21:01:14 In the OE build process, is there a way to force a single locale? I notice it doing work for a bunch of locales it doesn't need to. Jul 24 21:01:45 Human1: There is an option, yes but I forget its name in OE :/ Jul 24 21:02:44 # GLIBC_GENERATE_LOCALES = "en_GB.UTF-8 de_DE.UTF-8" ? Jul 24 21:02:46 ENABLE_BINARY_LOCALE_GENERATION and GLIBC_GENERATE_LOCALES ? Jul 24 21:02:52 https://neurosaudio.com/hackers/gremlin/trunk/OE/conf/local.conf Jul 24 21:02:55 aha Jul 24 21:03:13 You found the same document, or you knew it already? :) Jul 24 21:03:18 I was checking my local.conf right then Jul 24 21:03:22 Heh. Jul 24 21:03:22 ;-) Jul 24 21:03:53 ah, the DISTRO= thing, well...it's not php...will evaluated later Jul 24 21:05:17 Yeah, I see from the comments that bitbake will do variable expansion if you use ${} notation, which is why that Angstrom local.conf snippet was ok. Jul 24 21:05:57 hm for NLS I have reworked gettext.class a little bit Jul 24 21:06:45 and there are serval recipes which dont not include gettext class Jul 24 21:07:06 and therefor build NLS package even without USE_NLS=no Jul 24 21:09:19 Human1: compiles fine in .dev http://tinderbox.openembedded.net/packages/python-pexpect/ Jul 24 21:10:34 Hmm. Jul 24 21:11:20 And it only took 3 minutes? Jul 24 21:11:37 yes Jul 24 21:11:46 I retry Jul 24 21:11:49 Wow. I'm sitting here watching qemu-arm eat up CPU and not make any progress. Jul 24 21:12:23 well, I'm compiling over an existent layout Jul 24 21:12:32 not from scratch Jul 24 21:12:43 True. Jul 24 21:12:49 My build is sitting here: http://pastebin.com/m4f4bacc0 Jul 24 21:12:56 though, I'm building all dependencies (just to test) Jul 24 21:13:18 BB_DEFAULT_TASK = "buildall" Jul 24 21:14:25 Human1: in .dev the only nasty errmsg is: NOTE: preferred version 2.6.1 of glibc not available (for item virtual/arm-angstrom-linux-gnueabi-libc-for-gcc) Jul 24 21:14:49 harmful Jul 24 21:14:59 ? Jul 24 21:15:00 It complains that 2.5 of glibc is not available in stable... I guess it always complains about glibc... Jul 24 21:15:29 yes, I think khem and RP are working on this Jul 24 21:16:13 Human1: second build took less ! Jul 24 21:16:55 I wish I had your build environment :) Jul 24 21:17:03 no, it's not this Jul 24 21:17:24 is some setting on your side Jul 24 21:17:45 I've tried it two different ways, and it fails almost identically. Jul 24 21:17:48 but remember, I'm building 2008.1 Jul 24 21:18:18 My problem is that I need a package that'll work with 2007.1 (Angstrom 2007.12) Jul 24 21:18:54 I'm sorry I don't have any 2007.1 checkouts Jul 24 21:19:10 did you simply cherry-pick? Jul 24 21:19:24 I'm not sure what you're asking. Jul 24 21:19:43 is the same version of the package in angstrom-stable ? Jul 24 21:19:54 or did you pick it from .dev? Jul 24 21:20:04 I checked out stable Jul 24 21:20:44 and you're building python-pexpect 2.1 r0 Jul 24 21:21:08 Trying to :) Jul 24 21:21:34 just to be sure you did not make any change Jul 24 21:22:01 xjqian: hey Jul 24 21:24:11 ant_ what version of qemu-arm are you using? Mine is misbehaving by pegging the CPU and not letting itself be killed. Mine is 0.9.0. Jul 24 21:24:49 ah Jul 24 21:24:55 0.9.1 iirc Jul 24 21:25:28 Is that version tied to the branch? Or can I get 0.9.1 and use it to do a build of stable? Jul 24 21:26:29 http://thread.gmane.org/gmane.comp.handhelds.openembedded.stable Jul 24 21:26:48 1st thread Jul 24 21:27:39 does anybody know where (0,0) is supposed to be on a touchscreen? Jul 24 21:29:42 Seems that 0,0 is upper left corner for X Windows Jul 24 21:31:09 thesing : http://mirror.centos.org/centos/4/docs/html/rhel-rg-en-4/s1-x-server-configuration.html : The numbers to the right of "Screen0" give the X and Y absolute coordinates for the upper-left corner of the screen (0 0 by default). Jul 24 21:31:13 HTH Jul 24 21:31:24 ok. thanks. Jul 24 21:36:28 *sigh* my CPU is pegged again by the runaway qemu-arm :( Going to reboot again. Jul 24 21:42:53 Human1: what distro you said you have on host? Jul 24 23:38:37 03thesing 07org.oe.dev * rf87808d3... 10/ (8 files in 3 dirs): Jul 24 23:38:37 linux-rp_2.6.24: -fix build Jul 24 23:38:37 linux-rp_2.6.26: -update patches for collie.The kernel won't build atm Jul 24 23:38:37 because of too fat defconfig. But this is intended because kernel Jul 24 23:38:37 will be in rootfs in the future. Jul 24 23:38:38 linux-rp.inc: -change commandline for collie to root on mmc Jul 24 23:38:43 03thesing 07org.oe.dev * rd2f1a9f3... 10/ (1 classes/packaged-staging.bbclass): Jul 24 23:38:44 packed-staging.bblcass: change DEPLOY_DIR_PSTAGE from "=" to "?=" to be able Jul 24 23:38:46 to override it in local.conf Jul 24 23:50:44 thesing: Thanks Jul 24 23:51:01 What do you think how for we are from rootfs? Jul 24 23:51:16 Laibsch: wait 10 min and read oe-devel ;) Jul 24 23:51:32 Oh, *that* far ;-) Jul 24 23:51:47 I am not sure if I can be patient any longer Jul 24 23:51:49 Hehe Jul 24 23:56:26 Mail sent. Jul 24 23:57:42 Laibsch: The kernel doesn't support pm atm. It broke and I thought it's more important to work on the mmc stuff. Now you can start on the userspace stuff. ;) Jul 24 23:58:07 OK Jul 24 23:58:09 Nice Jul 24 23:58:18 Anything in particular you have in mind? Jul 24 23:59:21 No. I have no idea how well images that are not console-image work. Jul 25 00:01:25 I'll try console-image first, too Jul 25 00:01:33 I wonder if my SD cards will work Jul 25 00:02:02 All card that I have here worked flawlessly. Jul 25 00:02:52 But it's always been like that hasn't it? Jul 25 00:03:10 No I had some cards that didn't always work. Jul 25 00:04:04 I wonder if sdhc cards work. Jul 25 00:10:56 OK, let's see Jul 25 00:14:25 thesing: I have applied the patch here Jul 25 00:14:33 Do I need to clean out tmp or anything Jul 25 00:14:44 Or can I just build away "bitbake opie-image"? Jul 25 00:14:57 It should just work (tm) Jul 25 00:16:00 cool Jul 25 00:16:03 good night Jul 25 00:18:32 night. Jul 25 00:35:07 * * OE Bug 4433 has been RESOLVED (FIXED) by Jul 25 00:35:09 * * Update Chicken Scheme to 3.3.0 (bb included) Jul 25 00:35:11 * * http://bugs.openembedded.net/show_bug.cgi?id=4433 Jul 25 01:01:48 03Laibsch 07org.oe.dev * re6021195... 10/ (3 files in 2 dirs): chicken: update to 3.3.0. Closes 4433. Jul 25 01:01:53 03Laibsch 07org.oe.dev * rae2c3da1... 10/ (4 files in 2 dirs): chicken: make sure there are no mistakes we have different inc files for 2.x and 3.x Jul 25 01:14:22 03khem 07org.oe.dev * r49b1d83e... 10/ (157 files in 17 dirs): Jul 25 01:14:22 Change the toolchain build sequence. Helps in reproducable toolchains components Jul 25 01:14:22 when recompiled. Also same sequence for all different combinations Jul 25 01:14:22 (uclibc/nptl/eglibc/glibc/linuxthreads). Also fixes and patches that I worked to Jul 25 01:14:22 get various ancient compilers working. Jul 25 01:16:12 RP: I finally committed the big patch :) Jul 25 01:18:33 * mwester makes a note to record the "n-1" revision, just in case... :D Jul 25 01:20:06 mwester: I hope it breaks less Jul 25 01:22:25 I have a SlugOS build queued up to make sure all is well; I've been waiting for this commit actually :) Jul 25 01:22:36 mwester: cool Jul 25 01:22:46 mwester: will you do a clean build ? Jul 25 01:23:37 Yes. I renamed the tmp directory that I built just last night, and I'll do a clean build. That makes it easy to compare, in the event that there are important differences. Jul 25 01:24:13 mwester: that would be great. Let me know if something is odd. I will try to fix it. Jul 25 01:39:20 can i tell oe/bitbake to not build but only download everything that is needed for a target/package? Jul 25 01:40:28 (so i can build it later offline) Jul 25 01:40:53 pigeon: -c fetchall (or so) Jul 25 01:54:13 thesing: thanks, that seems to be doing it. Jul 25 02:21:37 03thesing 07org.oe.dev * rbdbc9f1d... 10/ (1 packages/initrdscripts/initramfs-kexec_1.0.bb): initramfs-kexec: search for zImage on mmc card Jul 25 02:21:42 03thesing 07org.oe.dev * r91420cde... 10/ (1 packages/initrdscripts/initramfs-kexec_1.0.bb): disapproval of revision 'bdbc9f1de56a38029ef1396ead1cd220bf2c07b2' Jul 25 02:21:47 03thesing 07org.oe.dev * r3421ff70... 10/ (1 packages/initrdscripts/initramfs-kexec_1.0.bb): initramfs-kexec: load zImage from mmc card Jul 25 02:21:52 03Laibsch 07org.oe.dev * rd8d258ba... 10/ (4 files in 2 dirs): lcdproc: unify (in preparation of closing 2488) **** ENDING LOGGING AT Fri Jul 25 02:59:57 2008