**** BEGIN LOGGING AT Sat Nov 25 02:59:57 2006 Nov 25 07:46:50 morning all Nov 25 08:04:47 cyrilRomain : morning Nov 25 08:05:45 good morning all Nov 25 08:10:11 morning all Nov 25 08:11:39 hey RP Nov 25 08:11:57 RP: angstrom left PARALLEL_INSTALL_MODULES behind a while ago :) Nov 25 08:14:03 koen: Yes, I think I remember asking you. I just need to update OZ ;-) Nov 25 08:14:59 hi alle Nov 25 08:15:00 all Nov 25 08:15:10 hey greentux Nov 25 08:15:15 er gremlin[it] Nov 25 08:15:19 hi emte Nov 25 08:15:29 :) Nov 25 08:15:30 i'm trying to build gpe-image, but resulting image miss kernel package ;( Nov 25 08:15:36 hi emte Nov 25 08:16:09 sorry gremlin[it] , i cant help you Nov 25 08:16:28 i think is cause MACHINE_TASK_PROVIDER is empty :( Nov 25 08:17:33 i saw your h36 touchscreen stuff ... for 2.6? Nov 25 08:17:49 yes Nov 25 08:20:15 h3600 is about usable now :) Nov 25 08:20:52 with exception to all the other minor issues it's always had? Nov 25 08:21:23 i was talking about h3600 at all ... Nov 25 08:21:46 touchscreen behave quite good in conjunction with tslib filtering and dejitering ... Nov 25 08:22:27 i was curious about that Nov 25 08:22:54 03koen 07org.oe.dev * rbcc164e1... 10/ (1 conf/machine/h3600.conf): h3600.conf: add MACHINE_TASK_PROVIDER = "task-base" Nov 25 08:22:54 i was surprised too ! Nov 25 08:23:16 hahaha koen ... i just add this line at hand :) !!! Nov 25 08:23:50 i'll try it after i convince tslib to behave for me Nov 25 08:24:03 i'm starting thinking as oe :) Nov 25 08:24:15 what's your PDA emte ? Nov 25 08:25:47 same old 3760 Nov 25 08:26:05 Well, hello there Nov 25 08:26:14 is there any vim syntax highlighting file for .bb files ? Nov 25 08:26:30 I very much doubt it Nov 25 08:26:43 Unless .bb follows some previous convention Nov 25 08:27:03 Or unless someone has coded something specifically Nov 25 08:27:15 kergoth has one Nov 25 08:27:27 I found another bug Nov 25 08:27:36 koen: ah. Why not including it in the contrib directory ? Nov 25 08:27:48 Whilst configuring edb-native, i got this: Nov 25 08:27:49 /usr/local/oe/tmp/staging/i686-linux/bin/gtk-config: fork: Resource temporarily unavailable Nov 25 08:28:00 Which goes on forever Nov 25 08:28:14 cyrilRomain: because it's a bitbake file Nov 25 08:28:27 Two last lines are: Nov 25 08:28:28 checking for gtk_init in -lgtk... /usr/local/oe/tmp/staging/i686-linux/bin/gtk-config: fork: Resource temporarily unavailable Nov 25 08:28:28 /usr/local/oe/tmp/staging/i686-linux/bin/gtk-config: fork: Resource temporarily unavailable Nov 25 08:28:37 koen: ah yes Nov 25 08:28:46 It then goes into a loop, printing that last line' Nov 25 08:29:04 cyrilRomain: http://svn.berlios.de/viewcvs/bitbake/trunk/bitbake/contrib/vim/ Nov 25 08:29:37 koen: thanks, that what I was looking for :) Nov 25 08:31:25 I'm guessing this is because of my distribution or something, but it doesn't hurt to check, does it? Nov 25 08:38:04 emte should work out of the box. Nov 25 08:38:41 gremlin[it], having a few autotool issues Nov 25 08:39:07 u use gpe or opie ? Nov 25 08:39:23 gpe Nov 25 08:40:21 i can send my image :) Nov 25 08:40:32 sure Nov 25 08:40:53 i'd like to see how the h36 is behaving Nov 25 08:59:54 mhhh seem that also if i add kernel-module-xxx in MACHINE_EXTRA_RRECOMMENDS in h3600.conf it isn't satisfied :( Nov 25 09:08:50 Does "error: invalid 'asm': invalid operand for code 'w'" looks like a gcc bug for you ? Nov 25 09:17:57 cyrilRomain: looks like something is including headers from a wrong arch Nov 25 09:18:14 cyrilRomain: or some lazy coders assumes that every cpu can handle x86 opcodes Nov 25 09:19:05 koen: thanks. hmm .. Nov 25 09:20:44 ah maybe a clue: "NOTE: bitlbee's configure script is not an autotool creation", so maybe should I write my own do_configure instead of the autotools one Nov 25 09:23:47 right :) Nov 25 09:29:16 meh Nov 25 09:29:25 "I have just a DVD full of files from Guylhem" Nov 25 09:29:36 ~lart people who don't use SCMs Nov 25 09:29:37 * ibot puts on a hockey mask and jumps out at people who don't use SCMs Nov 25 09:31:28 * cyrilRomain think about Guylhem Aznar how has had a serious car accident ... :-( Nov 25 09:32:00 got hit by the cluetrain Nov 25 09:32:26 it took a serious accident before he wants to play with others Nov 25 09:34:09 koen: a 'cluetrain', what does that mean ? Nov 25 09:35:13 If you are clueless and suddenly get reach enlightenment Nov 25 09:35:20 analogous to the cluebat Nov 25 09:40:20 * cyrilRomain extends his vocabulary with the urban english dictionnary Nov 25 09:40:28 koen: i see Nov 25 09:50:51 ~hail zecke Nov 25 09:51:02 * ibot bows down to zecke and chants, "I'M NOT WORTHY!!" Nov 25 10:00:12 hi zecke Nov 25 10:04:15 03rpurdie 07org.oe.dev * r6e2050b8... 10/ (1 site/ix86-common site/x86_64-linux): site/: Correct db_cv_mutex for x86_64, add some entries to ix86-common for qemux86 (from poky) Nov 25 10:06:53 03rpurdie 07org.oe.dev * r252d58c4... 10/ (1 packages/linux/linux-nokia770_2.6.16-osso26.bb): linux-nokia770: Fix SRC_URI, add missing RPROVIDES Nov 25 10:08:49 03rpurdie 07org.oe.dev * r26bf6462... 10/ (1 packages/glib-2.0/glib-2.0.inc): glib-2.0.inc: Add missing import Nov 25 10:10:16 03rpurdie 07org.oe.dev * r0a128cf3... 10/ (1 packages/gtk+/gtk+-2.10.0 packages/gtk+/gtk+-2.10.1): gtk+ Drop empty directories Nov 25 10:13:25 03koen 07org.oe.dev * re01ca28e... 10/ (1 packages/linux/linux-nokia770_2.6.16-osso26.bb): linux-nokia770: bump PR Nov 25 10:15:04 mickey|away: Dr Schaller is mailing me the DVD with the n30 stuff, which is also s3c2410 Nov 25 10:15:24 mickey|away: I'll put the patches online so you can see if they are usefull Nov 25 10:15:31 Should I attempt to fix STAGING_BINDIR in OE? Nov 25 10:16:30 You could wait and see if more people comment on your mail Nov 25 10:16:38 03rpurdie 07org.oe.dev * ra2216f0f... 10/ (1 classes/base.bbclass): base.bbclass: Try to ensure the MACHINE variable isn't exported (needs bitbake support to complete) Nov 25 10:16:52 the USians are probably recovering for a turkey overdose this weekend Nov 25 10:17:09 s/for/from/ Nov 25 10:17:26 koen: excellent. thanks! Nov 25 10:17:29 morning guys Nov 25 10:17:39 hi mickeyl Nov 25 10:17:58 koen: ok, I'll wait :) Nov 25 10:17:58 hi mokol ;) Nov 25 10:18:08 mickeyl: the kernel(s) is(are) at http://www.openembedded.org/~koen/letux/ Nov 25 10:18:42 koen: Perhaps I'll prepare a patch for comments. I tested a build in poky after changing things and a mixed arm/x86 build worked Nov 25 10:19:31 a STAGING_BINDIR for the target could be usefull if we start using qemu more Nov 25 10:21:46 morning Nov 25 10:22:35 http://openzaurus.linuxtogo.org/wordpress/ - opinions? Nov 25 10:22:58 hi hrw Nov 25 10:23:36 i like that autumn theme. it looks a bit "stock", but nice. Nov 25 10:23:39 hey hrw Nov 25 10:24:11 mickeyl: I was thinking about replacing that gfx with zauri + oz dragon Nov 25 10:24:39 mickeyl: resize the window horizontally Nov 25 10:24:53 mickeyl: notice how the graphic wraps Nov 25 10:24:56 that's pretty neat Nov 25 10:26:50 koen: The STAGING_BINDIR changes I have in mind might remove the need for a lot of binconfig.bbclass Nov 25 10:27:35 STAGING_BINDIR changes? Nov 25 10:27:48 I will be very tough regarding these Nov 25 10:27:59 It took me ages to get bindir.bbclass to work :D Nov 25 10:28:14 binconfig, even Nov 25 10:28:59 mickeyl: I've only just realised binconfig solves a similar problem to the one I faced a different way Nov 25 10:29:13 ~lart linux for not allowing to set 'ondemand' cpufreq by default in kernel Nov 25 10:29:13 * ibot takes linux to the vet for a "special" visit for not allowing to set 'ondemand' cpufreq by default in kernel Nov 25 10:29:23 RP: the EFL stuff is a good testsuite. if that still builds after your changes, then it might be ok Nov 25 10:29:24 mickeyl: binconfig renames the files, I created two bindirectories and played with PATH :} Nov 25 10:29:58 RP: ya, there are a lot of possible solutions. pb_ once implemented the two bindirectories but that didn't work out afair. let's try your way Nov 25 10:30:38 mickeyl: http://sam.rpsys.net/poky-bindir.patch was what I applied to poky Nov 25 10:31:14 mickeyl: I vetted all the STAGING_BINDIR references in poky, I need to do the same for OE now... Nov 25 10:33:10 hmm Nov 25 10:33:24 why $STAGING_DIR/$BUILD_SYS/bin/$HOST_SYS and not just Nov 25 10:33:43 STAGING_DIR/$HOST_SYS/bin ? Nov 25 10:34:02 which is the place where all other target binaries go Nov 25 10:34:20 mickeyl: I mentioned this in the email to oe@. Basically, concerns about packages staging and putting arm binaries in PATH Nov 25 10:34:33 s/packages/packaged/ Nov 25 10:34:46 hmm, i got no mail. when did you send? Nov 25 10:35:03 2pm yesterday Nov 25 10:35:12 hmm Nov 25 10:36:16 mickeyl: http://lists.linuxtogo.org/pipermail/openembedded-devel/2006-November/000846.html Nov 25 10:36:39 thanks. Nov 25 10:36:40 * mickeyl reads Nov 25 10:38:02 well, i'm ok with that if E still builds Nov 25 10:38:26 I can leave binconfig alone just by s/STAGING_BINDIR/STAGING_BINDIR_NATIVE Nov 25 10:38:58 I can never get E to build in the first place - missing upstream sources atm :-( Nov 25 10:39:05 i don't consider binconfig as something elegant in its present stage. if your changes allow making it more simple, then I'm all for it Nov 25 10:39:28 mickeyl: I think it could mostly be removed entirely! Nov 25 10:39:35 RP: even better. Nov 25 10:40:18 mickeyl: I'll write a patch which we can experiment with... Nov 25 10:40:21 cool Nov 25 10:40:39 hmm thinking about it.. .wouldn't cross-dir be a more appropriate path? Nov 25 10:41:08 Yes, because they are a cross program Nov 25 10:41:12 *nod* Nov 25 10:41:43 It should be simple enough to work that into the patch just by changing bitbake.conf :} Nov 25 10:41:45 http://www.hrw.one.pl/2006/11/25/project-for-writing-gpl-driver-for-8385-marvell-wifi-chipset/ Nov 25 10:41:54 RP: yeah Nov 25 10:45:17 hrw: that's missing a link to the actual project Nov 25 10:45:45 03koen 07org.oe.dev * r42c12ae6... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom 2007.1: add another CVS_TARBALL_STASH Nov 25 10:46:26 2o yes Nov 25 10:47:35 I'll work on this later... Nov 25 10:47:38 * RP -> afk Nov 25 10:47:59 Another missing file Nov 25 10:48:18 dbus-python requires cross.patch, which is nowhere to be found Nov 25 10:49:12 file a bug Nov 25 10:50:10 Ok, sure Nov 25 10:50:20 * koen -> afk Nov 25 10:50:33 I'll do a 2-in-1 bug report Nov 25 10:53:51 angstrom-2007.1/progear started Nov 25 10:54:23 Is the bugtracker down? Nov 25 10:54:33 Can't seem to connect Nov 25 10:55:16 ip changed not yet propagated Nov 25 10:55:35 Ah, ok Nov 25 10:55:42 bitbake-trunk + fast machine == nice parsing speed Nov 25 10:55:44 I'll just file the bugs later Nov 25 10:57:14 How do I have bitbake not process a package? Nov 25 10:57:42 When building "world", I don't want bb to process dbus-python, how do I do that? Nov 25 11:00:13 BBMASK Nov 25 11:00:35 local.conf? Nov 25 11:01:01 found it, thank you hrw Nov 25 11:03:49 np Nov 25 11:35:41 I run out of space on my laptop so I copied all of open embedded onto my server and am sharing the files using a samba server... but now when I use bitbake I get the error bash: /data/openEmbedded/bitbake/bin/bitbake: /usr/bin/env: bad interpreter: Permission denied Nov 25 11:38:20 don't mount with 'noexec' flag Nov 25 11:39:34 noexec is useless Nov 25 11:39:57 | checking for kernel header at least 2.6.17... too old! Nov 25 11:40:07 koen: x86 need 2.4 intermediate? Nov 25 11:40:36 160 PREFERRED_VERSION_glibc-intermediate_guinness ?= "2.4" Nov 25 11:40:40 ~lart koen for it Nov 25 11:40:41 * ibot brandishes Excalibur! "With this sword, I vanquish thee, koen!" and lops off koen's head for it Nov 25 11:41:06 s/guiness/i686 Nov 25 11:42:32 MACHINE=guinness seems to need it Nov 25 11:43:10 koen: progear too so I think that this is rather i686 then guiness Nov 25 11:43:18 2.4 also fail Nov 25 11:43:24 | checking installed Linux kernel header files... 2.0.10 or later Nov 25 11:43:24 | checking for kernel header at least 2.6.17... too old! Nov 25 11:43:24 | configure: error: *** The available kernel headers are older than the requested Nov 25 11:44:19 ok. another guiness Nov 25 11:45:58 koen, this is what I use to mount my data dir //samurai/data /data cifs defaults,rw,user,username=guest,password="" 0 0 Nov 25 11:46:21 remove the 'user' Nov 25 11:46:30 will do thanks Nov 25 11:46:48 174 PREFERRED_VERSION_orinoco-modules_h3600 = "0.13e" Nov 25 11:46:49 175 PREFERRED_VERSION_orinoco-modules_h3900 = "0.13e" Nov 25 11:47:09 koen: we can drop it in angstrom as it is 2.4 only Nov 25 11:47:34 ah, right Nov 25 11:47:46 you can commit that i Nov 25 11:47:46 f you want Nov 25 11:48:58 ok. looks like I only need to add progear.conf to linux-rp Nov 25 11:49:18 and will need to add new feature: acpi Nov 25 11:49:55 cu for now - have a nice weekend Nov 25 11:56:01 03koen 07org.oe.dev * r506c1d6e... 10/ (4 files in 3 dirs): base-files: mv ipaq-pxa270 hx4700 Nov 25 12:28:41 hey gerwinin Nov 25 12:28:45 hey Nov 25 12:32:44 03koen 07org.oe.dev * r80d921d4... 10/ (2 files in 2 dirs): mb panel hacks: set package_arch to all Nov 25 13:02:09 koen: I have found the reason of gcc-cross-4.1.1 breaking. During configuration a "gcc -m64 flag" was passed and I do not have a 64bit cpu. This results in zlib/Makefile containing a "MULTIDIR = 64" directive. Commenting this line solves the problem. The question is: how the "-m 64" flag arrived here in the first place ... Nov 25 13:12:47 260 root 12 -4 1512 464 1352 S 0.0 0.2 0:00.20 udevd Nov 25 13:12:47 434 root 15 0 2848 1272 1608 S 0.0 0.6 0:00.10 devfsd Nov 25 13:12:52 * koen scratches head Nov 25 13:18:26 umh, "CROSS COMPILE Badness: /usr/include in INCLUDEPATH: /usr/include" in gcc-cross-4.1.1 Nov 25 13:19:31 Koen: I am trying to make a .bb file for the win32 codec Nov 25 13:20:15 Koen: I want that is puts the files in /usr/include/lib/codecs/win32 so that mplayer can see them on compilation Nov 25 13:20:38 Koen: you have any idea where to start reading docs to get an idea on how to do this Nov 25 13:21:01 Nicolas: did you install all the required files ? Nov 25 13:21:19 install -m 0644 foo.dll ${includedir}/lib/codecs/win32/ Nov 25 13:21:50 FILES_${PN} += "${includedir}/lib/codecs/win32/" Nov 25 13:22:01 gerwinin: what do you mean "install all the required files"? Nov 25 13:22:35 Koen so if I put that in a bb file for all dlls than it works ? Nov 25 13:23:20 yes Nov 25 13:23:47 the install part inside do_install(){} of course Nov 25 13:24:28 Koen: I will give it a try , I will try to put all my patches and stuff for x86 in bugs.openembedded.org tomorrow Nov 25 13:25:10 Koen: I got gpe running , mplayer, totem , a bunch of games running on my epia board with flash Nov 25 13:25:25 cool Nov 25 13:27:04 Koen I will publish some binaries including tutorial on how to flash Nov 25 13:27:27 Nicolasfr: for which board are you compiling ? Nov 25 13:28:08 i586 cpu (Tatung webpad) Nov 25 13:30:08 Nicolasfr : Is it for a epia cpu ? Nov 25 13:32:29 Nicolasfr: can you give me a link to this product Nov 25 13:32:56 gerwinin: no. But even when trying with other x86 I have the same problem. Fails when building gcc-cross-4.1.1 Nov 25 13:33:04 gotta go... Nov 25 13:34:01 Nicolasfr: I am busy on a release now for x86 I will keep you up to date, if I have finished this I have some more time to help Nov 25 14:25:49 gerwinin howdy :) Nov 25 14:54:38 BTW: anyone of you fixed an ipod? Nov 25 14:58:32 Genesis : hey Nov 25 14:58:50 Genesis : how did you include the win32 codecs ? Nov 25 14:58:54 zecke: harddisk troubles? Nov 25 14:59:26 Zecke: only the electronics , my display broke down changed it with a new one Nov 25 15:01:27 gerwinin : i donno Nov 25 15:02:26 Genesis: I think I managed to get that working now but I need some more testing Nov 25 15:02:56 gerwinin : i test my bb that works Nov 25 15:03:17 if you want to add configure option , it's easy Nov 25 15:06:58 Genesis : Your bb works with me as well Nov 25 15:07:53 Genesis but it doesn't want to see the win32 codecs you need to add an option include dir and you need to add those codecs Nov 25 15:07:53 koen: disk Nov 25 15:07:53 koen: it looks like al's disk broke Nov 25 15:08:17 zecke: most likely the connector is just loose Nov 25 15:08:23 Genesis: I missed the flash bb that I had to make Nov 25 15:08:38 koen: hehe, just use a screw driver to open the case? Nov 25 15:08:41 zecke: people have reported 'fixing' their ipods by dropping them Nov 25 15:09:33 koen: hmm, do you have any hints/tricks/manuals Nov 25 15:09:44 only the interblag Nov 25 15:10:17 gerwinin : to support .swf ? Nov 25 15:10:23 gerwinin : gnash , good luck :) Nov 25 15:10:34 Zecke: don't send it back to apple it cost you money and you need to wait for ages Nov 25 15:10:50 gerwinin: I wouldn't mind, it is the ipod of my ex :) Nov 25 15:11:07 gerwinin: I could declare it broken and use it myself, but I won't do it Nov 25 15:11:21 gerwinin: I'm more like either getting her a new disk, or just fixing the 'issue' Nov 25 15:13:48 zecke: biff Nov 25 15:14:02 biff? Nov 25 15:14:29 "you have mail" Nov 25 15:14:35 Zecke: oe seems to be bad for relations :) Nov 25 15:15:27 and for your mind :) Nov 25 15:16:38 Zecke: I have now girlfriend from the usa she is using oe as well Nov 25 15:18:25 drat Nov 25 15:18:29 Zecke: finding a girlfriend who uses linux as well saves a lot of time , especially with all the time you need to spent in removing spyware adware and tojans from your girlfriends computer Nov 25 15:18:36 the USians are shooting down the avian carriers Nov 25 15:18:54 ? Nov 25 15:19:02 * koen waits for bugs.oe.org to finish loading Nov 25 15:19:19 Crofton: http://www.faqs.org/rfcs/rfc2549.html Nov 25 15:19:41 ah Nov 25 15:21:01 koen: thanks for the hint with the connector Nov 25 15:21:12 koen: I opened the case, and pushed the connector a bit Nov 25 15:21:51 and it works now? Nov 25 15:22:00 looks like it Nov 25 15:22:12 cool Nov 25 15:22:28 thanks Nov 25 15:27:37 hail zecke Nov 25 15:33:11 master Nov 25 15:34:38 gerwinin : which ff works for you now ? Nov 25 15:36:07 pb_: how are you? Nov 25 15:36:34 Genesis : both Nov 25 15:36:42 Genesis I had to patch 1.0.7 Nov 25 15:37:04 Genesis : 1.5.0.6 works as well but I am working now on the codecs Nov 25 15:52:51 http://www.reghardware.co.uk/2006/11/23/wizpy_linux_media_player/ Nov 25 15:58:49 gerwinin : oki i'd some pb with the 1.0.7 , i'll test the 1.5 Nov 25 15:59:58 lkcl : it should support flac :) Nov 25 16:08:42 zecke: good, thanks Nov 25 16:08:44 * pb_ going to cinema now Nov 25 16:08:45 later all Nov 25 16:29:35 hey can i add a KEY_CAMERA to the input_pda.h file? Nov 25 16:29:47 oops sorry wrong irc channel :) Nov 25 16:59:52 gerwinin: I've got a Gentoo-running wife if that counts :p Nov 25 17:00:15 my mother runs gentoo ^^ Nov 25 17:15:27 * RP returns Nov 25 17:15:40 zecke: Did you see my email about readline? Nov 25 17:16:00 zecke: readline for x86? Nov 25 17:16:03 ehm Nov 25 17:16:08 s/zecke/RP/ Nov 25 17:16:35 koen: poky was failing to compile in tinderbox with a failure in readline (on arm) Nov 25 17:16:54 it fails for i686 over here Nov 25 17:17:07 upgrading readline + debian patch didn't help Nov 25 17:17:14 koen: Is acinclude.m4 being installed correctly? Nov 25 17:17:29 * koen has no idea Nov 25 17:17:53 koen: Do you have the configure log from it handy? Nov 25 17:18:05 sorry Nov 25 17:18:17 doing a build from scratch for arm right now Nov 25 17:18:37 I 'fixed' it by bypassing the check for posix signals in the .c file Nov 25 17:18:54 Ah, its was posix signals related... Nov 25 17:19:12 I suspect I broke the fetchers in bitbake Nov 25 17:19:49 I do have some weird issue with bitbake trunk Nov 25 17:19:57 koen: Were you rebuilding it with -b ? Nov 25 17:20:08 'bitbake angstrom-gpe-image' fails on perl-native do_configure Nov 25 17:20:25 bitbake -b perl-native -c rebuild works Nov 25 17:20:33 but not bitbake perl-native Nov 25 17:20:44 RP: I will check Nov 25 17:21:11 zecke: My theory is the acinclude.m4 will be from the wrong package Nov 25 17:21:29 RP: want a strace of the build? or how should I check it? Nov 25 17:21:47 zecke: See if acinclude.m4 exists, if so md5 it Nov 25 17:22:02 zecke: Then we'll see where it came from... Nov 25 17:27:37 RP: http://pastebin.ca/257702 Nov 25 17:29:08 ~fishslap zecke Nov 25 17:29:12 zecke: I guess the readline one is missing due to tinderbox's clean upon failure feature? Nov 25 17:29:15 * ibot slaps zecke up side the head with a wet fish. Nov 25 17:29:25 yeah, just recognized that :) Nov 25 17:31:39 wtf Nov 25 17:31:48 NOTE: package readline-4.3: completed Nov 25 17:31:48 Build statistics: Nov 25 17:32:05 zecke: clean readline, bitbake oh-image-pda Nov 25 17:32:12 zecke: I'd guess it will then fail Nov 25 17:32:31 might be a dash issue as well Nov 25 17:32:44 zecke: Its the fetcher url caching ;-) Nov 25 17:32:50 re Nov 25 17:32:59 Jin^eLD: wb Nov 25 17:33:04 hey :) Nov 25 17:33:05 zecke: At least I suspect it will be. I'd like to prove it Nov 25 17:33:17 RP: okay Nov 25 17:33:56 zecke: what are your thoughts on installing tinderserver on the box gerwinin has arranged? Nov 25 17:34:13 and bonsai and ikada? Nov 25 17:34:27 koen: I have a weird problem... I am not sure if it may be related to this: http://sourceware.org/ml/libc-alpha/2006-10/msg00056.html because I did not do any eabi settings... but it seems I can not run any threads, always returns out of memory. any idea what may have gone wrong? Nov 25 17:34:28 dunno about bonsai Nov 25 17:35:09 RP: it fails Nov 25 17:35:12 zecke: lets see how beefy the CPU is in that box Nov 25 17:35:27 koen: what did you mean by tinderboxserver? Nov 25 17:35:51 zecke: postgres db + webinterface Nov 25 17:37:30 RP: dline-4.3$ md5sum acinclude.m4 Nov 25 17:37:31 1ef285787c9590fa4bc85c32907ef336 acinclude.m4 Nov 25 17:38:27 zecke: Which is oprofile's acinclude.m4 :} Nov 25 17:38:58 could be true Nov 25 17:39:20 zecke: Both use the url file://acinclude.m4 and my fetcher's cache code is broken... Nov 25 17:40:08 meta/packages$ md5sum oprofile/oprofile/acinclude.m4 Nov 25 17:40:08 1ef285787c9590fa4bc85c32907ef336 Nov 25 17:44:40 is http://bugs.openembedded.org/ down @ the moment? Nov 25 17:45:13 see topic Nov 25 17:47:29 koen: can not reach it on 64.136.190.225/ either Nov 25 17:47:49 works over here Nov 25 17:48:46 koen: I do reach "treke.net" when I enter 64.136.190.225? ... Nov 25 17:49:32 argos-wlan:~ koen$ grep bugs /etc/hosts Nov 25 17:49:32 64.136.190.225 bugs.openembedded.org Nov 25 17:49:40 the wonders of vhosts Nov 25 17:52:19 koen: strange - do not understand this; however ... assume it will work again eventually Nov 25 17:53:31 pgfeller: The server at that IP hosts multiple sites. It knows which site to show you by the name you request it as. If you what koen showed you in /etc/hosts, the site will work Nov 25 17:53:52 or visit bugs.treke.net Nov 25 17:54:07 but my browser doesn't autocomplete my password there Nov 25 17:54:16 ~lart buggy brain cells Nov 25 17:54:16 * ibot says "boot to the head" and knocks buggy brain cells over Nov 25 17:54:36 koen & RP: thanks ... will adjust my hosts file accordingly Nov 25 17:57:08 RP: any fix? Nov 25 17:57:15 RP: why did you broke poky? Nov 25 17:57:30 zecke: Its a problem in bitbake trunk Nov 25 17:57:52 zecke: The easy solution - bin the url cache :-/ Nov 25 17:58:03 RP: As I'm tired I would relicense BitBake to GPLv2 now Nov 25 17:58:16 RP: any thing to take into account to ease merging with poky? Nov 25 17:58:23 zecke: The harder solution probably means adding a parameter to bb.fetch.localpath Nov 25 17:58:42 zecke: No, I think poky is pretty much in sync with bitbake, thanks :) Nov 25 17:59:12 zecke: The problem is base.bbclass calls bb.fetch.localpath so we can't just change bitbake :-( Nov 25 17:59:30 RP: how does your caching work? Nov 25 17:59:36 RP: how is it different to 'A'? Nov 25 18:00:16 zecke: The idea was to cache the return value of localpath for each url Nov 25 18:00:27 RP: like 'A' does? Nov 25 18:00:34 ;) Nov 25 18:01:07 zecke: No, as it also stores data internal to the fetcher backend, saving repeated calls to split_uri etc Nov 25 18:01:23 zecke: Do we actually use A? Nov 25 18:01:29 http://svn.berlios.de/viewcvs/bitbake/trunk/bitbake/lib/bb/parse/parse_py/BBHandler.py?rev=695&view=markup Nov 25 18:01:48 RP: yes in exactly one bbclass (the src_distributa...) Nov 25 18:02:59 zecke: The bug is basically the global urldata = {} in fetch/__init__.py Nov 25 18:03:39 zecke: Its obviously badly flawed but can we salvage any of the speedup it brought Nov 25 18:08:53 http://www.rpsys.net/openzaurus/temp/bitbake-fetcher-fix.patch should fix it I guess Nov 25 18:09:55 RP: hmm that is dangerous in regard to COW Nov 25 18:10:16 zecke: I wondered about that :-/ Nov 25 18:10:31 It doesn't work anyway... Nov 25 18:17:04 zecke: I don't really want to teach the data class about the urldata but I can't see any other way atm... Nov 25 18:18:09 RP: we can create a COW dict for private data Nov 25 18:18:18 RP: and then store urldata in this dict? Nov 25 18:18:25 dunno about the performance impact Nov 25 18:19:41 zecke: It shouldn't be COW Nov 25 18:20:30 RP: I doubt that. SRC_URI should not change between bb.data.createCopy but it can change :) Nov 25 18:20:30 zecke: The problem at the moment is this cache needs to be reset whenever we switch packages Nov 25 18:21:15 zecke: I worry about the calls to localpath when d = configurationData Nov 25 18:21:45 Although in theory, that shouldn't happen? Nov 25 18:23:24 not with the old parser Nov 25 18:23:48 and in theory INHERIT += "fooclass" should not change SRC_URI Nov 25 18:23:55 but it could happen Nov 25 18:29:30 zecke: We store the urls unexpanded. If the change or new ones are added, the url cache doesn't have a problem with that Nov 25 18:30:11 ah, but it does since we store localpath :-( Nov 25 18:31:56 how much speedup was the url cache? Nov 25 18:32:41 03mickeyl 07org.oe.dev * r8f3baeb5... 10/ (1 packages/dbus/dbus-python_0.71.bb): dbus-python: make it work Nov 25 18:32:53 zecke: Quite a bit. Saved a ton of os.path.join calls and all kinds of other things Nov 25 18:33:23 RP: I'm not sure where the current problem is Nov 25 18:33:33 But if we have a global cache with one (FILE) indirection Nov 25 18:33:36 would that work? Nov 25 18:34:44 zecke: it would work better than the existing method although we don't have file data available in that function :-/ Nov 25 18:35:08 bb.data.getVar('FILE') does not work? Nov 25 18:35:27 zecke: maybe :} Nov 25 18:36:19 I think I can gurantee that FILE points to the bb file Nov 25 18:36:36 but one might call update_data and change the SRC_URI Nov 25 18:38:13 zecke: By the time we're asked about SRC_URI's, all the update_data calls should have been made Nov 25 18:38:51 zecke: I'd say calling update_data after working with the SRC_URI would be a bug. I don't think its a problem Nov 25 18:38:55 RP: not in set_additional_vars from BBHandler where you will cache things first Nov 25 18:39:41 RP: hmm, I think most people should use TypeWriters instead of Computers Nov 25 18:39:52 zecke: Arguably, the settings in A would then be wrong too :} Nov 25 18:39:54 I try to construct an example where SRC_URI changes, but I'm bad at these things Nov 25 18:40:03 RP: I would like to kill 'A' anyway Nov 25 18:40:09 zecke: Fine with me Nov 25 18:40:15 * RP was thinking this before Nov 25 18:40:18 RP: we need to patch one bbclass Nov 25 18:40:44 zecke: I noticed base.bbclass accesses the fetch class directly... Nov 25 18:42:18 zecke: Will FILE be valid when calling the confHandler supports method? Nov 25 18:43:11 confHandler supports methods? Nov 25 18:43:42 supports function then :} Nov 25 18:44:01 my python terminology is broken... Nov 25 18:44:15 s/terminology// :} Nov 25 18:44:41 I don't get it, but that might be due the smell of cookies in the bake oven Nov 25 18:45:02 def supports(fn, d): Nov 25 18:45:02 return localpath(fn, d)[-5:] == ".conf" Nov 25 18:45:14 cookies? Nov 25 18:45:27 COOKIES? Nov 25 18:45:29 in confHandler - localpath turns into a possible call to the fetcher. Is FILE valid at the point - I'd guess not? Nov 25 18:45:53 RP: ah, this weird stuff needs to be removed from the parser anyway... Nov 25 18:46:11 add BBFILES = "http://www.koen.org/weird/file.bb" is a disgusting idea Nov 25 18:46:40 RP: I think FILE is only set for true .bb/ files Nov 25 18:46:59 zecke: Dropping support for wierdness like that would make life simpler... Nov 25 18:47:46 zecke: yes, its set in the handler Nov 25 18:48:04 let us forget about that weirdnes, I will remove it Nov 25 18:48:18 people can write a wget/curl cript to download/update the file anyway Nov 25 18:49:30 zecke: I wonder what happens if you try an svn url in there? :) Nov 25 18:50:49 I guess scm archives wouldn't go down well with the parser... Nov 25 18:52:34 RP: just one file will work :) Nov 25 19:00:08 ${EXTRA_OECONF}="-without-fp" Nov 25 19:00:11 is that legal? Nov 25 19:01:38 -EXTRA_OECONF = "--disable-iso8859-2 --disable-iso8859-3 --disable-iso8859-4 --disable-iso8859-5 --disable-iso8859-6 --disable-iso8859-7 --disable-iso8859-8 --disable-iso8859-9 --disable-iso8859-10 --disable-iso8859-11 --disable-iso8859-12 --disable-iso8859-13 --disable-iso8859-14 --disable-iso8859-15 --disable-iso8859-16 --disable-jisx0201 --disable-koi8-r" Nov 25 19:01:45 -EXTRA_OECONF? Nov 25 19:02:36 heh Nov 25 19:04:41 03zecke123 * r700 10bitbake-parser/ (bitbakeparser.y standalone/bitbakeparser.y): Fix the parser to accept export FOO += "moo" Nov 25 19:05:03 zecke: the legality of '-without-fp' depends on the configure script used Nov 25 19:05:25 koen: ${EXTRA_OECONF} for a variable name is in question here Nov 25 19:05:59 ah, right Nov 25 19:06:20 not legal right? Nov 25 19:06:53 afaik not Nov 25 19:07:11 org.openembedded.dev/packages/libccaudio2/libccaudio2_0.9.0.bb Nov 25 19:07:31 -EXTRA_OECONF is wrong as well? Nov 25 19:10:22 I think so Nov 25 19:11:29 RP, you there? Nov 25 19:14:37 RP, drop me a note when your back, need some info on L3 bus Nov 25 19:23:29 oh my parser is even faster than I expected :} Nov 25 19:23:34 I parse more bb files :) Nov 25 19:25:24 not much though Nov 25 19:42:49 I've been trying to submit a patch to the OE bugtracker for several days now but the bug tracker website does not seem to be up. I'm trying to get to it from the link on the OE website, has its address changed? Nov 25 19:43:06 is there another way to submit a patch? Nov 25 19:44:20 check the log Nov 25 19:49:13 Gerrath: add 64.136.190.225 bugs.openembedded.org to hosts file Nov 25 19:49:35 or use bugs.treke.net Nov 25 19:51:56 hey XorA|notsogone Nov 25 19:53:48 koen: heh heh, I am here only in body :-) Nov 25 19:59:39 wow after 4 GB of stuff installed VS 2005 just decides to do "blub" (play this sound) when I try to create a project Nov 25 20:00:01 try XCode Nov 25 20:00:09 has a nicer blub sound Nov 25 20:00:15 it lacks Windows Mobile 5.0 plugins Nov 25 20:00:27 zecke: when did you move to the dark side? Nov 25 20:01:03 its quite good actually Neo1973 saves me buying VS 2005 Nov 25 20:01:32 XorA|gone: I'm a mercenary :) Nov 25 20:01:40 tick-for-hire Nov 25 20:02:05 Kristoffer: ping Nov 25 20:02:24 cygwin for wince would be an awesome beast Nov 25 20:02:56 * koen winces Nov 25 20:03:03 XorA|gone: that is called ssh client Nov 25 20:03:58 RP, :D Nov 25 20:04:10 RP, needing L3 driver for my jornada7xx sounddriver Nov 25 20:04:10 reminds me, need to get real SSH on my wince machine so I can port forward for the zaurus Nov 25 20:04:30 RP, but saw your message with russell, just asking why it didnt get implemented Nov 25 20:04:40 Kristoffer: search the arm kernel mailing list, I think RMK once posted a generic one Nov 25 20:04:57 * XorA|gone needs L3 driver for smdk2440 board as well Nov 25 20:05:07 XorA|gone, yeah, but why didnt it get implement? Just because i386 people would get pissed of getting i2c initialized later? Nov 25 20:05:11 earlier.. Nov 25 20:05:34 Can someone remind me what L3 is? :) Nov 25 20:05:52 some weird leetspeak Nov 25 20:05:57 RP: serial bus like i2c Nov 25 20:06:14 RP: it can run on same wires as i2c without interference Nov 25 20:07:05 yeah Nov 25 20:07:36 XorA|gone, so are we stuck with russells generic as a side patch? Nov 25 20:08:02 Kristoffer: You're sure it was me? I can't remember this at all :} Nov 25 20:08:02 Kristoffer: probably Nov 25 20:08:02 http://lkml.org/lkml/2005/11/2/320 Nov 25 20:08:03 RP, pretty sure Nov 25 20:08:24 Kristoffer: in a couple of weeks Ill probably add it to smdk2440 kernel in OE Nov 25 20:09:11 Kristoffer: didn't Chicken implement L3 as well? Nov 25 20:09:23 zecke, chicken? Nov 25 20:09:40 http://handhelds.org/~chicken/ Nov 25 20:10:04 Kristoffer: I did start that thread, yes. I've never had anything to do with L3 though, just i2c Nov 25 20:10:53 crappit another 1341 codec driver :-) Nov 25 20:11:09 zecke, looks like it :) Alex is a godsent as always Nov 25 20:11:32 Kristoffer: http://handhelds.org/cgi-bin/cvsweb.cgi/linux/kernel26/sound/oss/jornada56x-uda1341.c?rev=1.1&content-type=text/x-cvsweb-markup Nov 25 20:11:32 what cpu is in jornorda? Nov 25 20:11:39 sa1110 Nov 25 20:11:40 XorA|gone: which model :) Nov 25 20:11:54 zecke: you mean there is many? Nov 25 20:12:33 does Kristoffer one have pxa, s3c24XX or Atmel arms? Nov 25 20:12:35 XorA|gone: I think SH3 and ARM Nov 25 20:13:06 SH3 and ARM Nov 25 20:13:39 * koen has LEG Nov 25 20:14:39 * XorA|gone wondered if he could persuade Kristoffer to do an ASoC driver for it Nov 25 20:15:04 haha :D like spitz? Nov 25 20:15:23 hi Nov 25 20:15:29 Kristoffer: yeah Nov 25 20:15:34 hey Marcin Nov 25 20:15:46 Kristoffer: that way people only need to write the codec driver once, not once for each platform Nov 25 20:16:01 gerwinin: win32 codecs for mplayer need to be rather in ${libdir}/win32 then any other place. Nov 25 20:16:20 XorA|gone, how compatible is that with current kernel though?.. Nov 25 20:16:52 koen: in 10 minutes 'the it crowd' in Polish TV - I wonder how bad translation is Nov 25 20:17:06 heh Nov 25 20:17:48 I remember the german tv had a hard time germanifying the puns Nov 25 20:18:15 XorA|gone, would be kewl to make something platform independant. But just wanna make sure it gets into git somewhere along the line Nov 25 20:18:30 the it crowd? german tv? what was the name? Nov 25 20:18:43 Kristoffer: ASoC is in ALSA upstream pending mainline inclusion Nov 25 20:18:55 RP, groovey Nov 25 20:19:32 RP: how much work is needed to move normal alsa driver into asoc one? Nov 25 20:20:33 hrw: They're rather different things. In theory you need a lot less code for ASoC, depending on whether the platform, codec or machine are already supported Nov 25 20:21:05 RP: this hardware has working alsa driver already Nov 25 20:21:24 hrw: Does ASoC make sense for it? Nov 25 20:22:12 RP: hard to tell - its battery powered device but cant tell how good PM support it has Nov 25 20:22:23 http://www.google.com/blogsearch?q=marvell - roxx Nov 25 20:23:04 http://www.google.com/search?q=marvell+cf8385 ;) Nov 25 20:23:12 XorA|gone, thanks. Nov 25 20:23:46 Kristoffer: I mostly try because I have a board with a uda1341 on it I need a codec driver for Nov 25 20:23:55 sometimes I wonder how does google works - so many searchs ends in my site Nov 25 20:25:07 koen, thanks. Nov 25 20:25:50 XorA|gone, :) Will look into it, need to implement sound asap anyways as its mostly that left Nov 25 20:26:25 hi hrw Nov 25 20:26:29 hi all Nov 25 20:26:33 Kristoffer: well if you have ASoC questions lrg and I work on it for pay Nov 25 20:27:02 hi nicolasfr Nov 25 20:27:10 XorA|gone, goodie Nov 25 20:29:47 hrw: have you ever played with wpa on zaurus? I just realised that resume/suspend on sharprom just works fine with wpa_supplicant but not on OE based distro. wpa_supplicant gets just stuck on resume. I guess this has to do with how apm is managed? Nov 25 20:31:33 nicolasfr: 2.6.17+ kernel + hostap == no card after resume Nov 25 20:31:54 hrw: right! this is a kernel problem? Nov 25 20:32:06 yes, not found yet Nov 25 20:32:55 btw, I saw you were working for a driver of a wifi CF card. Does the card have an original form factor (i mean smaller than other)? Nov 25 20:34:05 its smaller then all other wifi which I have Nov 25 20:34:27 it is shorter then metal part of collie stylus Nov 25 20:34:36 hrw: even for cards with firmware in flash? Nov 25 20:35:12 I don't know what size a collie stylus is, but if it's smaller than other ones, I'll buy one for sure. Nov 25 20:37:43 XorA|gone: ask coredump - I lent my prism card for friend Nov 25 20:38:02 i've checkout the latest bitbake (rev 700) and when i try to ryn it i get an error -> ERROR: Unable to open conf/bitbake.conf Nov 25 20:38:13 any idea why ? Nov 25 20:38:39 BBPATH not set Nov 25 20:38:53 its set. the older version works ok Nov 25 20:40:27 Ifaistos: do you have a version of BitBake installed? Nov 25 20:40:37 Ifaistos: PYTHONPATH set to a old version? Nov 25 20:41:04 zecke : hmmm not sure Nov 25 20:42:00 zecke : nop. python path shows the current bitbake Nov 25 20:42:20 unset $PYTHONPATH Nov 25 20:42:34 python -vvv bitbake YOUR Stuff Nov 25 20:42:39 where does it load the modules from? Nov 25 20:50:30 zecke : dlopen("/usr/lib/python2.4/lib-dynload/_locale.so", 2); Nov 25 20:50:30 import _locale # dynamically loaded from /usr/lib/python2.4/lib-dynload/_locale.so Nov 25 20:50:55 Ifaistos: more interestingly are stuff with bb/ in the path :) Nov 25 20:54:30 zecke : http://pastebin.com/832810 Nov 25 20:57:24 hmm, I'm not only a QA monkey, but logfile reader monkey :) Nov 25 20:57:58 * koen hands zecke some bananas Nov 25 20:58:10 Ifaistos: pastebin is slow Nov 25 20:58:48 zecke : Considering its Saturday... it seems that a lot of people are still coding :) Nov 25 20:59:26 zecke : this is where is mention the error http://pastebin.com/832818 Nov 25 21:00:09 Ifaistos: if you paste a log, then paste th complete log :) Nov 25 21:00:40 Ifaistos: pastebin.ca would be appreciated as well (or rafb.net) as it is faster today Nov 25 21:01:39 zecke : The complete log does not fit in pastebin. email it to you ? Nov 25 21:02:01 python -vvv bitbake -p foo does not fit? Nov 25 21:04:04 Ifaistos: : inconsistent use of tabs and spaces in indentation could be an error :) Nov 25 21:05:40 zecke : but where ? in bitbake.conf ? Nov 25 21:05:49 Ifaistos: or a bbclass Nov 25 21:06:32 zeche : bitbake is just from svn and org.openembedded.dev is just synced with OE Nov 25 21:07:34 hrw: I only have prism3's these days Nov 25 21:07:34 Ifaistos: yes, somewhere tabs and spaces are mixed Nov 25 21:07:57 Ifaistos: use -DDD after bitbake -DDD -p foo Nov 25 21:07:57 zecke : the only file i created is a distro file Nov 25 21:08:39 time for me to eat, cyas Nov 25 21:09:47 XorA|gone: I have prism2 broken and marvel8385 one Nov 25 21:12:53 sleep time Nov 25 21:12:54 cu Nov 25 21:13:04 'night hrw, XorA|gone Nov 25 21:13:12 03rpurdie * r701 10bitbake/lib/bb/fetch/__init__.py: fetch/__init__.py: Store urldata by 'FILE' to solve url contamination between .bb files Nov 25 21:13:17 ~seen offroadgeek Nov 25 21:13:35 offroadgeek was last seen on IRC in channel #openzaurus, 433d 4h 5m 35s ago, saying: 'hi lardman... I'm doing well.. you?'. Nov 25 21:13:37 koen: Could you see if that solves your strange bitbake problems? Nov 25 21:13:50 sure Nov 25 21:16:10 03rpurdie * r702 10bitbake/lib/bb/parse/parse_py/BBHandler.py: BBHandle: Remove unused __bbpath_found__ variable Nov 25 21:20:20 03rpurdie * r703 10bitbake/lib/bb/data_smart.py: data_smart: Micro optimisation - search for "${" instead of "$" Nov 25 21:21:36 03rpurdie * r704 10bitbake/lib/bb/runqueue.py: runqueue: Avoid running task sanity checks upon task failure Nov 25 21:22:35 string(r) ::= string(s) STRING_MULTI(t). Nov 25 21:22:37 eek :( Nov 25 21:23:06 zecke: ? Nov 25 21:23:32 03rpurdie * r705 10bitbake/lib/bb/data.py: data.py: Fix inherits_class() so it doesn't split lists. Add unexport handling to emit_env. Nov 25 21:24:32 wow Nov 25 21:24:35 error city Nov 25 21:24:56 ERROR: 'list' object has no attribute 'split' while parsing /home/koen/OE/monotone/org.openembedded.dev/packages/automake/automake_1.9.2.bb Nov 25 21:24:57 NOTE: exceptions.AttributeError:'list' object has no attribute 'split' while evaluating: Nov 25 21:24:57 ftp://ftp.gnu.org/gnu/automake/automake-1.9.3.tar.bz2 ${@['file://path_prog_fixes.patch;patch=1', ''][bb.data.inherits_class('native', d)]} Nov 25 21:24:57 ERROR: Error in executing: Nov 25 21:25:03 koen: svn up Nov 25 21:25:12 I'm at 705 Nov 25 21:26:12 koen: inherits_class had a bug in which should be fixed in 705... Nov 25 21:27:47 wow Nov 25 21:27:52 KeyError: 'CACHETIMESTAMP' Nov 25 21:28:56 touch conf/local.conf didn't work out so well Nov 25 21:29:40 koen: Instead of removing an error from trunk, I added it :-( Nov 25 21:30:31 03rpurdie * r706 10bitbake/lib/bb/data.py: data.py: Revert inherits_class change - merge error :-( Nov 25 21:30:44 koen: try that one ;-) Nov 25 21:30:57 Four local copies of bitbake make my head hurt Nov 25 21:31:01 much better Nov 25 21:31:34 Finally, I have a clean eye diagram Nov 25 21:32:32 after whacking the scope? Nov 25 21:32:39 itbakescanner.l:217: warning, Regel kann nicht passen Nov 25 21:32:43 fixing my code :) Nov 25 21:32:54 nice way of flex saying me I suck... Nov 25 21:32:59 I was was only sending half the symbols Nov 25 21:33:56 now it is time for beer Nov 25 21:36:13 RP: r706 doesn't fix my perl issue Nov 25 21:36:42 ERROR: function do_configure failedERROR: log data follows (/data/build/koen/OE/build/tmp/angstrom/work/armv5te-angstrom-linux-gnueabi/perl-5.8.7-r21/temp/log.do_configure.18406)| rm: remove write-protected regular file `Makefile.SH.patch'? cp: cannot create regular file `./Makefile.SH.patch': Permission denied Nov 25 21:37:26 koen: bitbake perl works though? Nov 25 21:37:33 psokolovsky: ping Nov 25 21:38:12 koen: I've seen that error with perl too thinking about it :-/ Nov 25 21:38:13 RP: no Nov 25 21:38:21 RP: only with -b Nov 25 21:38:30 Ah, -b works? :-/ Nov 25 21:38:33 yes Nov 25 21:38:45 which is why I always forget to report it Nov 25 21:39:33 That makes no sense... Nov 25 21:39:57 03zecke123 * r707 10bitbake-parser/ (4 files): (log message trimmed) Nov 25 21:39:57 Add special MultiLine handling to the lexer and scanner Nov 25 21:39:57 To support stuff like SRC_URI = "foo \ Nov 25 21:39:57 # not yet ready Nov 25 21:39:57 more stuff" Nov 25 21:39:58 We will need to add another state to parse each line Nov 25 21:40:00 of a multiline statement separately. Nov 25 21:40:52 koen: That issue where bitbake scrolled a ton of failed tasks should be fixed now btw Nov 25 21:41:34 * koen wonders why bluez-libs 2.25 *and* 3.7 get built Nov 25 21:42:14 koen: The -DDD output from that might be interesting Nov 25 21:44:28 RP: http://ewi546.ewi.utwente.nl/tmp/perl-borkage.log Nov 25 21:45:17 koen: Was that for the bluez problem or the perl one? :) Nov 25 21:45:23 perl Nov 25 21:45:38 I think bluez is cause by versioned depends Nov 25 21:45:46 ok :). I've just been able to reproduce the perl one locally... Nov 25 21:46:21 since we have bluez_2.25 and bluez-utils_3.7, which probable depend on bluez-libs_{PV} Nov 25 22:28:17 03rpurdie 07org.oe.dev * r488e80d2... 10/ (1 packages/perl/perl.inc packages/perl/perl_5.8.7.bb): perl: Add -f option to rm, fixing read only file issues Nov 26 01:11:26 when ever I compile a package I get a line saying NOTE: preferred version 2.5 of glibc not available is it refuring to glibc on my desktop? or an arm version that has already been compiled? I would like to update it to the version that it requests Nov 26 01:14:26 zezom, I get that also, but I seem to build glibc OK Nov 26 01:14:46 it is talking about the one it builds for the tool chain Nov 26 01:15:00 hey Crofton Nov 26 01:15:09 hi Nov 26 01:15:28 I have another weirdness and I am not sure to what it is related Nov 26 01:15:37 the weird things never stop I guess :) Nov 26 01:15:37 hmm Nov 26 01:16:21 I can not start threads - reports out of memory Nov 26 01:16:30 it's with my newly built distro Nov 26 01:16:39 from the feeling of things it sounds like that http://sourceware.org/ml/libc-alpha/2006-10/msg00056.html Nov 26 01:16:46 but I think I do have those patches Nov 26 01:16:48 Crofton, thanks I'll have a closer look at the arm version, thanks Nov 26 01:16:51 and I am not doing eabi Nov 26 01:17:03 what do you think? Nov 26 01:17:03 zezom, I get the messge also, but things work for me Nov 26 01:17:19 yeah everything compiles fine Nov 26 01:17:34 just thought I might be able to get a speed upgrade? Nov 26 01:17:59 zezom, I do not think so :) Nov 26 01:18:18 ok, no worries then :) Nov 26 01:18:18 Jin^eLD, what version of glibc are you using? Nov 26 01:18:23 thanks for the info Nov 26 01:18:44 Crofton: from what I can say it must be 2.5, allthough when building stuff I get "preferred provider for glibc 2.5 not available" Nov 26 01:18:51 heh Nov 26 01:18:56 don't we all Nov 26 01:19:01 :) Nov 26 01:19:11 Jin^eLD, have you tried eabi? Nov 26 01:19:15 no Nov 26 01:19:35 should i? Nov 26 01:19:38 when I run things for real, it is on eabi Nov 26 01:19:51 what do you mean? Nov 26 01:19:51 I think I would, or go back a few versions of things Nov 26 01:20:02 but I think must of the active testing is eabi now Nov 26 01:20:27 when I run things on the OSK that use a lot of functions, it is in the eabi images Nov 26 01:20:44 I only try to start dropbear on oabi to check patch for that problem Nov 26 01:20:49 I see Nov 26 01:20:50 hm Nov 26 01:20:51 well Nov 26 01:21:21 something I've ben wondersing for a little while now... I've been compiling all of this stuff but where do I find the image that I put onto my palm and where are all of the packages located? and is their an easy way to bundle hem all together for instalation? Nov 26 01:21:25 yu can start processes, but f they try to sytart threads it fails? Nov 26 01:21:44 zezom, look in tmp/deploy Nov 26 01:21:53 Crofton, thanks :) Nov 26 01:22:36 Crofton: it looks like it... I straced my program and it craps out on mmap2 with OUT_OF_MEMORY Nov 26 01:22:40 Jin^eLD, what is your machine Nov 26 01:22:52 the error messages from my app say - could not start task bla, so pthread_create fails Nov 26 01:22:57 a custom one, arm9 Nov 26 01:23:07 Crofton, that's exactly what I wanted, just didn't see it mentuned in the wiki Nov 26 01:23:40 Jin^eLD, you are working out of my league :) Nov 26 01:23:47 :) Nov 26 01:23:55 so far I had no hints Nov 26 01:24:08 how much RAM? Nov 26 01:24:20 not much, 8MB, but my app runs with the other image Nov 26 01:24:31 what other image? Nov 26 01:24:49 the one that came from the vendor, and I think I built one myself too, but last week or something Nov 26 01:24:56 hmmm Nov 26 01:25:02 so I have two images, one mine and one from the vendor, app works there Nov 26 01:25:06 you need more RAM :) Nov 26 01:25:11 and I have this new image - where it does not work anymore Nov 26 01:25:20 I know, but right now its not an option :) Nov 26 01:25:24 :) Nov 26 01:26:01 from what I can see via "top" - memory usage/availability is the same for the new image and for my old and vendor image Nov 26 01:26:10 yeah Nov 26 01:26:16 so it must be something else Nov 26 01:26:23 what kernel came from vendor? Nov 26 01:26:34 2.6.16.19 Nov 26 01:26:49 I now build the kernel too, I extracted patches from the vendor kernel and created a recipe Nov 26 01:27:04 my newly compiled kernel works with the old image - can run my app on old image + new kernel Nov 26 01:27:04 good idea Nov 26 01:27:33 so now I can build my whole distro including kernel via OE Nov 26 01:27:34 can you try and recreate old image in oe Nov 26 01:27:43 that was the goal.. but it does not work anymore :) Nov 26 01:27:44 this is the goal :) Nov 26 01:27:51 hmmm Nov 26 01:27:53 well hehe.. I do not htink I can reproduce it Nov 26 01:28:10 the old image was more a testing run, I was trying to figure out how to create a distro, etc. Nov 26 01:28:19 yeah Nov 26 01:28:19 you guys helped me with it, the new image s derived from the old one Nov 26 01:28:22 this is importnat Nov 26 01:28:35 so the config files are almost the same, but of course I did an update since then Nov 26 01:28:42 it is nice gettitng all the info stoed in the meta data Nov 26 01:28:59 yep Nov 26 01:29:02 Jin^eLD, I'm sure you will work through the problem Nov 26 01:29:19 I spent the past 1.5 weeks banging my head against a wall with some other code Nov 26 01:29:27 and what was it? Nov 26 01:29:44 the final solution revovled around me being a moron :) Nov 26 01:29:50 :))) Nov 26 01:29:58 that's the usual case, same happens with me all the time ;) Nov 26 01:30:02 yeah Nov 26 01:30:10 the path to enlightment is not easy hehe Nov 26 01:30:21 I passed an array of shorts and the length in elements Nov 26 01:30:40 to a routine expecitng an array of char and length in bytes Nov 26 01:30:54 so I would only transmit half the data Nov 26 01:31:06 heh Nov 26 01:31:09 so there would be "glitches" at teh endpoints Nov 26 01:31:25 in the cases where the state change was different Nov 26 01:31:46 well, at least your problem was in your code... I prefer having crap going on in my stuff where I at least have the overview Nov 26 01:31:58 so, the answr to your problem is likely very simple Nov 26 01:32:07 the path is long and winding :) Nov 26 01:32:34 with this memory/threads problem - it's for sure not in my code, the thread stuff is stable and has been checked/valgrinded/etc. Nov 26 01:32:51 so right now I do not even know where to start looking :) Nov 26 01:33:05 btw, what is this eabi stuff? I heard a lot of talk about it lately Nov 26 01:33:27 I read that its something that improves performance on arm? Nov 26 01:33:54 heh Nov 26 01:34:00 I do not know the details Nov 26 01:34:24 I suspect it is the syscall interface Nov 26 01:34:37 well, why did you chose to use it? Nov 26 01:34:44 I think you did build for it, right? Nov 26 01:35:39 well, because I was operating under the principle the newer the better ... Nov 26 01:35:47 and it solved my dropbear problem Nov 26 01:35:52 hmm Nov 26 01:35:55 ok I will try it then Nov 26 01:36:16 since you are just starting dev for a pplatform, I waould try it Nov 26 01:36:47 setting just the TARGET_OS is not enough, right? Nov 26 01:37:16 set TARGET_OS to linux-gueabi in local.conf Nov 26 01:37:25 with an = Nov 26 01:37:36 that should do the trick Nov 26 01:37:51 the tools will ahve names like arm-linux-gnuebi-gcc Nov 26 01:39:11 bother I cant log into my build machine to check .... Nov 26 01:51:03 if I were stupid enough to change the location of where my openEmbedded dir was what would be all of the locations that I need to change in order to get it to work again? I just did a grep for my old path in build/tmp/staging/ and I got a whole heap of lines like tmp/staging/arm-linux/lib/libvorbisidec.la:dependency_libs=' -L/usr/src/openEmbedded/build/tmp/staging/arm-linux/lib' Nov 26 01:52:21 as far as I know you just need to have your environment set up right, variables like BBPATH and so on Nov 26 01:52:28 (or maybe only BBPATH?) Nov 26 01:52:46 laso when I do a bitbake nmap I get | Can't locate Automake/Struct.pm in @INC (@INC contains: /usr/src/openEmbedded/build/tmp/staging/i686-linux/share/automake-1.9 /etc/perl /usr/lib/perl5/vendor_perl/5.8.8/i686-linux /usr/lib/perl5/vendor_perl/5.8.8 /usr/lib/perl5/vendor_perl /usr/lib/perl5/site_perl/5.8.8/i686-linux /usr/lib/perl5/site_perl/5.8.8 /usr/lib/perl5/site_perl /usr/lib/perl5/5.8.8/i686-linux /usr/lib/perl5/5.8.8 /usr/local/lib/site_perl .) at /ope Nov 26 01:52:47 nEmbedded/build/tmp/staging/i686-linux/bin/automake line 47. Nov 26 01:52:50 the rest should not matter... I think :) I never tried moving it Nov 26 01:53:02 yeah I though I had set all of that up Nov 26 01:53:14 it's probably a good idea not to Nov 26 01:53:27 I ran out of space on my laptop so I had to Nov 26 01:53:31 I spent quite some time trying to figure out a good setup, I am still not finished :) Nov 26 01:53:45 it's all good :) Nov 26 01:54:17 it offers a lot of settings/possibilities, especially if you have a lot of different hardware Nov 26 01:54:40 can you see what would be still remembering the /usr/src/openEmbedded location instead of the /openEmbedded location? Nov 26 01:55:06 no idea.. if you did not yet compile too much stuff you could try rebuilding.. Nov 26 01:55:23 but I'm really not an expert... Nov 26 01:55:44 yeah, I've compiled quite a bit but then again I won't be installing it onto my palm for a while... how would I do a make clean world? Nov 26 01:56:12 I also have a few different hardware targets Nov 26 01:56:26 I would probably just wipe the tmp directory, the one with staging, work, etc., maybe keep the sources Nov 26 01:56:30 well, let me find a link Nov 26 01:56:40 hrw posted his OE setup Nov 26 01:56:45 yeah I will deffently keep the sources :) Nov 26 01:56:51 you should probably take a look Nov 26 01:56:55 cool Nov 26 01:56:58 thanks for that Nov 26 01:57:21 http://hrw.one.pl/2005/10/31/my-openembedded-enviroment/ Nov 26 01:57:37 thanks again Nov 26 01:57:41 I'll read it right now Nov 26 01:58:17 and there should be an update from 2006 Nov 26 01:58:37 http://www.hrw.one.pl/2006/11/22/my-openembedded-environment-ii/ Nov 26 01:58:58 I'll read them both... Nov 26 01:59:39 I think the key information for me was, that you can set more than one path for BBPATH, you can split them with : just as you would with normal PATH variable Nov 26 02:00:03 then you could have your own build directory and your own local conf and maybe a sitelconf there, but you would not mess anything up in the org.openembedded.dev dir Nov 26 02:02:18 that's good to know. it seems like a very powerfull tool Nov 26 02:02:26 it is indeed Nov 26 02:02:50 it just takes time to learn to use power :) Nov 26 02:03:37 yep Nov 26 02:03:44 <- still learning Nov 26 02:30:30 Crofton: my eabi build is failing, it tries to build glibc-2.3.5 and bails out with "Error: cannot represent SWI relocation in this object file format" **** ENDING LOGGING AT Sun Nov 26 02:59:57 2006