**** BEGIN LOGGING AT Tue Jan 30 02:59:57 2007 Jan 30 03:48:42 paulcox: well, you could just try disabling the patches to see if it works. If it doesn't, just go into the tmp/work/ARCH/PACKAGE/SOURCE dir, import the patch(s), fix the rejects, and rebuild the patches. I highly recommend using quilt (OE uses quilt already). Jan 30 04:23:55 03Tom 07org.oe.dev * r2680f2fb... 10/ (1 packages/xorg-xserver/xserver-kdrive_X11R7.1-1.1.0.bb): xserver-kdrive: more granular packaging - close #1830 Jan 30 06:36:44 re Jan 30 08:22:34 morning Jan 30 08:27:40 hi Jan 30 08:50:22 hi ! Jan 30 08:50:51 armeb-linux-strip: Unable to recognise the format of the input file Jan 30 08:50:58 what does that happens ? Jan 30 08:51:14 and what is strip ? Jan 30 08:52:02 vinvin: usually strip removes debugging symbols from a compiled binary executable or a library Jan 30 08:53:11 i have forgotten that, i'll remove them, but why does it fails ? Jan 30 08:53:50 well, guessing from your message - the file that you are feeding to the strip program is something that it is not expecting Jan 30 08:54:02 it has just built ir Jan 30 08:54:03 it Jan 30 08:55:32 use the file program to see what kind of file it is, that was produces, maybe it will give you a hint Jan 30 08:55:58 yes thanks Jan 30 09:00:29 it looks it has built it to i386 instead of arm... Jan 30 09:00:48 ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), for GNU/Linux 2.2.0, dynamically linked Jan 30 09:00:58 mickeyl: hey Jan 30 09:01:04 morning guys Jan 30 09:01:04 hi zecke Jan 30 09:01:10 yo mickeyl Jan 30 09:04:50 sorry to repeat, but its getting to me: Jan 30 09:04:51 [18:40:16] i still haven't been able to fix my rotation problem Jan 30 09:04:59 [18:40:37] the zaurus C3200 (on opie) is stuck in portrait Jan 30 09:06:12 probably better ask in #openzaurus, most people here have no idea what a Zaurus is Jan 30 09:06:13 I have put += instead of = for CFLAGS in my makefile, now i'm getting cc1: error: unrecognized option `-fweb' and `-funit-at-a-time' Jan 30 09:07:43 mickeyl: i have - no help :( - i think its a general opie problem.... isn't it? Jan 30 09:07:50 no idea Jan 30 09:08:16 * mickeyl didn't see Opie for more than 12 months Jan 30 09:11:11 what timezone is tmbinc in? any idea when it is the best time to catch him online? Jan 30 09:11:19 or is anyone else working on the dreambox branch? :> Jan 30 09:11:26 if he isn't on vacation he is CET Jan 30 09:11:51 thx Jan 30 09:13:20 btw, are there any plans on merging dreambox into dev? Jan 30 09:13:44 Jin^eLD: yes, but we postponed it after we have finished our clone efforts Jan 30 09:14:22 zecke: what are the "clone" efforts? :) Jan 30 09:14:24 Jin^eLD: it is not a question if we want to merge regulary, but simply a matter of when Jan 30 09:14:38 ok I see.. Jan 30 09:14:58 Jin^eLD: and this depends on man power Jan 30 09:15:06 hmm, bitbake crash: http://rafb.net/p/ubJPMR16.html Jan 30 09:15:23 It's from a bitbake gpe-image with bitbake 1.6.2 Jan 30 09:15:27 well, I am not quite sure if I should base my work upon dreambox or try to merge it into dev here, thats why I wanted to ask tmbinc about the status of the dreambox branch Jan 30 09:15:35 it does seem to build fine though Jan 30 09:16:12 stefan_schmidt: do you have something like rm_work enabled? Jan 30 09:16:44 this code has been touched recently by pb_ Jan 30 09:17:01 zecke: I should not. Don't even know how I would do it. :) Jan 30 09:17:25 zecke: Where should I take a look to be complete sure? Jan 30 09:18:07 morning all Jan 30 09:18:14 It happened after a complete rm -r tmp/ and trying to build the image over night. Jan 30 09:19:33 mickeyl: It'S the bitbake packaged for debian. No recent svn version. Jan 30 09:19:46 mickeyl: Perhaps already fixed in svn? Jan 30 09:19:47 the error comes from base.bbclass Jan 30 09:19:52 which has been touched by pb_ Jan 30 09:20:01 ah, ok. Misread Jan 30 09:20:20 try reverting mt diff -r4babe1caf820e719d27c7ed13d9518351ba729f3 -r77212e456a021803396d11bcad7616c51e5c6d25 Jan 30 09:20:25 that should fix it Jan 30 09:20:40 heh that revision has 'babe' in it Jan 30 09:20:42 * mickeyl chuckles Jan 30 09:21:06 ok Jan 30 09:21:15 * stefan_schmidt reads up some mtn doc about reverting Jan 30 09:21:28 just apply using patch Jan 30 09:21:33 and if it asks you to revert Jan 30 09:21:35 say yes Jan 30 09:21:56 or alternatively Jan 30 09:22:01 swap the arguments to diff Jan 30 09:22:04 and apply :) Jan 30 09:23:49 will do Jan 30 09:27:47 Patched, try building now. Jan 30 09:28:23 ~lart my old pentium m for being slow Jan 30 09:28:23 * ibot strangles my old pentium m with a doohicky mouse cord for being slow Jan 30 09:28:35 yeah, changing a class needs reparsing everything Jan 30 09:29:34 you need a better computer? i might part with my dual xeon 2.4 later this year (herbst). if you want it, drop me a note. Jan 30 09:29:56 morning Jan 30 09:30:25 mickeyl: hmm, sounds interesting. I always wanted a smp machine. I think about it. Jan 30 09:30:52 btw, anybody an idea about this: Jan 30 09:30:56 hi mickeyl Jan 30 09:31:01 morning pb_ Jan 30 09:31:05 !oebug #1826 Jan 30 09:31:07 * * Bug 1826, Status: NEW, Created: 2007-01-29 04:16 Jan 30 09:31:08 * * stefan(AT)datenfreihafen.org: CROSS COMPILE Badness in gnokii Jan 30 09:31:09 * * http://bugs.openembedded.org/show_bug.cgi?id=1826 Jan 30 09:32:33 btw., it's probably a bug in the gtk+ 1.x detection. i couldn't reproduce it here (probably because i don't have gtk+ 1.x installed), but it failed for me due to other things. updating it didn't work since they broke the buildsystem in their latest release. will send them a mail eventually Jan 30 09:32:35 ok, after reverting the changes the error log becomes smaller, but still exist. :) Jan 30 09:32:37 http://rafb.net/p/zIK7HD40.html Jan 30 09:33:26 then you may need to revert mt diff -r27b46e6b2f1768296232a9d2ce04946cd1d5514e -re03e452c09311f78ce177e8febeb9963f30bc1ef as well Jan 30 09:33:40 pb_: any idea ? Jan 30 09:33:42 * stefan_schmidt goes ahead Jan 30 09:44:03 mickeyl: sigh. Still the same error after reverting the second patch. Jan 30 09:44:11 hmm. bummer Jan 30 09:44:13 that's very odd Jan 30 09:44:28 let me see if the same problem occurs for me Jan 30 09:45:58 mm, yes, sure enough it does Jan 30 09:46:57 pb__: At least I'm not alone. :) Jan 30 09:52:09 is do_unpack_prepend() allowed? Jan 30 09:52:18 do_configure_prepend() works Jan 30 09:52:28 do_unpack_prepend() fails with a syntax error. Jan 30 09:52:30 do_unpack is generally python, not shell Jan 30 09:52:34 try do_fetch_append Jan 30 09:52:51 sweet, thx. Jan 30 09:53:07 I imagine prepending to do_unpack is ok if you match the languages. otherwise, you can introduce a new do_munge_files() method between fetch and unpack Jan 30 09:53:20 yeah, adding a new task is always possible Jan 30 09:53:37 ~dict munge Jan 30 09:53:38 do_fetch_append works for me Jan 30 09:53:39 Dictionary 'munge' (1 of 2): /muhnj/ 1. A derogatory term meaning to imperfectly transform information. 2. A comprehensive rewrite of a routine, data structure or the whole program. This term is often confused with {mung} and may derive from it, or possibly vice-versa. One correspondent believes it derives from the french "mange" /monzh/, eat. [{Jargon File}] ... Jan 30 09:53:55 sweet. /me learned a new word Jan 30 09:54:06 heh :-) Jan 30 09:55:13 hmm - actually do_fetch_append is not what I want, so I'll need to do a do_munge_files() task :-) Jan 30 09:55:45 is there a way to add a task *only* for a certain machine? or do I just check ${MACHINE} inside the task? Jan 30 09:56:08 I think you have to check inside the task. Jan 30 09:56:25 stefan_schmidt: I checked in a patch that fixes base-files for me. please see if it works for you. Jan 30 09:57:26 03pb 07org.oe.dev * rc061df3c... 10/ (1 classes/package.bbclass): package.bbclass: don't try to rmdir("./"), it won't work Jan 30 09:57:47 pb__: OK, I'll pull and try Jan 30 09:58:06 addtask do_remove_cmdline_hacks before do_patch after do_unpack Jan 30 09:58:59 heh - that turns into do_do_remove_cmdline_hacks :-) Jan 30 09:59:19 heh, yeah, the syntax is a bit inconsistent there Jan 30 10:03:35 morning Jan 30 10:03:48 i don't understand: i'm compiling with LDFLAGS=-lmylib, but it doesn't appear in the actual options so link does not success Jan 30 10:04:12 why does it, while CFLAGS=-Imyinc is of for compiling the .c to .o ? Jan 30 10:05:15 is there any way I can get the log of the operation performed in that new task? Jan 30 10:05:18 hey hrw Jan 30 10:05:22 vinvin: still light 1.0? Jan 30 10:05:29 no a new thing :) Jan 30 10:05:33 mickeyl: thx for answers Jan 30 10:05:36 np Jan 30 10:05:40 with a lib and an exe Jan 30 10:05:47 rwhitby: it shold emit log.do_yournewtask Jan 30 10:05:48 the lib is ok, the exe doesn't link Jan 30 10:05:59 rwhitby: does it not? Jan 30 10:06:02 no Jan 30 10:06:05 d'oh Jan 30 10:06:29 however, a subsequent new task (svnpatch) added between do_patch and do_configure does emit the log Jan 30 10:06:39 (and has a run.foo file as well) Jan 30 10:08:02 vinvin: EXTRA_LDFLAGS and do it like in irdadump? Jan 30 10:08:25 hrw|work http://vinvin.dyndns.org/~vinvin/cyrlink_1.0.bb Jan 30 10:08:28 yes Jan 30 10:09:15 pb__: That commit fixed it. Thanks. Jan 30 10:09:51 vinvin: show makefile Jan 30 10:10:39 mickeyl: even if I put remove_cmdline_hacks() in the same place as svnpatch, it still doesn't emit the run.* and log.* files. Jan 30 10:10:44 hrw|work http://vinvin.dyndns.org/~vinvin/Makefile Jan 30 10:10:55 rwhitby: odd. let's pester RP with that :) Jan 30 10:11:33 what causes those files to be emitted? i.e. do_fetch doesn't emit one, neither does do_unpack, but do_configure and do_compile do emit them Jan 30 10:12:41 here's another one: Jan 30 10:12:49 I have SRC_URI = "something" Jan 30 10:12:59 I then have SRC_URI += "something else" Jan 30 10:13:02 those work fine. Jan 30 10:13:34 if I then add SRC_URI_ixp4xx += "another thing", then only "another thing is unpacked - "something" and "something else" are ignored. what am I doing wrong? Jan 30 10:13:50 += doesn't support overrides or does it? Jan 30 10:13:52 rewrite that as Jan 30 10:14:00 SRC_URI_append_ixp4xx = " another thing" Jan 30 10:14:55 re. log, that is emitted in lib/bb/build.py:exec_shell_func. and it doesn't look to be conditionally to my eyes, so I wonder Jan 30 10:15:51 heh - that first space in " another thing" is significant :-) Jan 30 10:16:32 indeed Jan 30 10:16:37 mickeyl: I'm running bitbake 1.6 stable fyi Jan 30 10:18:11 Jan 30 10:18:13 ops Jan 30 10:19:38 hrw|work i don't get it. It's just the same than my light example, but -lupnp was properly added in that example Jan 30 10:22:19 mickeyl, rwhitby: There is some disparity between python function execution and shell function execution in the bitbake core. The log output is one side effect of that Jan 30 10:22:42 RP these are both shell functions - one emits and one doesn't Jan 30 10:23:12 rwhitby: It must emit a run.* file, else how does it run? Jan 30 10:23:30 well, that may explain why it's not running ... Jan 30 10:23:54 except it prints - task do_remove_cmdline_hacks: started and completed Jan 30 10:25:04 rwhitby: Can you show me a diff of how you're adding the task? Jan 30 10:25:47 http://blog.madduck.net/geek/2007.01.30_firefox-hate <- idiot, he's starting iceweasel and complains about firefox Jan 30 10:27:14 RP: on it's way - waiting for pastebin.ca to respond ... Jan 30 10:27:52 RP: http://pastebin.ca/332886 Jan 30 10:29:07 we have an svnpatch task, which downloads patches from our kernel svn area. If the machine is not an nslu2, then we want to remove the ugly cmdline patches from the series file before they are applied with quilt by the svnpatch task Jan 30 10:29:34 (these are the patches that compensate for linksys redboot not setting up ATAGs properly. Jan 30 10:29:35 rwhitby: s/remove_cmdline_hacks()/do_remove_cmdline_hacks()/ Jan 30 10:29:49 rwhitby: but only change that Jan 30 10:30:21 doh! Jan 30 10:30:23 thx Jan 30 10:33:11 RP: I'm going to use Marc Singer's arm-kernel-shim code to simulate ATAGs for boards which do not pass the right stuff to the kernel. Jan 30 10:33:43 rwhitby: Sounds good. I should look at that for the Zaurus :) Jan 30 10:34:18 Jin^eLD: hi Jan 30 10:34:39 RP: perhaps you can help me debug it :-) it's in OE now. Jan 30 10:35:04 Jin^eLD: you might also try to PRIVMSG me, as #oe usually scrolls out of my backscroll buffer. or email :) Jan 30 10:36:14 rwhitby: I'll try and have a look but I can't promise anything ;-) Jan 30 10:36:48 Marc says he tested it on nslu2 BE, but the kernel I prepended it to didn't get past the uncompress stage. Jan 30 10:37:04 I'm giving it another try tonight (that's why I'm removing the cmdline hacks here) Jan 30 10:40:04 why isn't the command irdadump in irda-utils ? Jan 30 10:45:48 vinvin: someone did not added building it years ago and now it was easier to add extra recipe then get irdadump built with irda-utils Jan 30 10:46:10 vinvin: thats why I made irdadump external recipe Jan 30 10:48:29 hrw|work: allright, so how can i build it for my oe system , Jan 30 10:48:32 ? Jan 30 10:48:50 vinvin: bitbake irdadump Jan 30 10:49:02 indeed Jan 30 10:49:06 sorry Jan 30 10:50:13 np Jan 30 10:50:19 you have any clues for my link problem ? Jan 30 10:50:45 vinvin: sorry but not now - some things to do at work Jan 30 10:51:11 :) Jan 30 10:51:43 np nevermind Jan 30 11:13:48 hi Cliff Jan 30 11:21:29 hrw|work: hello Marcin Jan 30 11:22:49 hrw|work: so you will be starting full time OE consulting soon? Jan 30 11:27:40 cbrake: I already signed fulltime contract for half year with OH. Jan 30 11:37:45 hrw|work: ahh, excellent -- did not realize you had already started. Jan 30 11:45:44 cbrake: 5.02.2007 is start (next monday) Jan 30 12:00:34 * XorA wonders if we can get mickeyl english lessons in Glasgow :-) Jan 30 12:03:50 i know i'll be in birmingham this year, but glasgow??? Jan 30 12:06:25 mickeyl: just thinking we could get you an interesting accent :-) Jan 30 12:07:17 hehe Jan 30 12:08:59 mickeyl: birmingham for guadec? Jan 30 12:09:04 yeah Jan 30 12:10:02 mickeyl: hope your not expecting some nice culture and architecural goodness :-) Jan 30 12:11:28 XorA: i won't :) Jan 30 12:11:50 i already heard some comments about it being a pretty dull place in general Jan 30 12:11:52 morning Jan 30 12:12:08 How on earth they decided Birmingham was a good idea I don't know... Jan 30 12:12:35 heh Jan 30 12:12:36 RP: cheapest bid? Jan 30 12:12:36 morning chouimat Jan 30 12:12:47 mickeyl: There are much more interesting places, lets put it that way Jan 30 12:13:09 RP: could be worse, could be in Southend Jan 30 12:13:36 RP: erm, I mean Eastbourne Jan 30 12:13:36 XorA: at least Southend is famous (thanks to Ali G) Jan 30 12:14:01 XorA: I guess it could, yes... Jan 30 12:16:00 friends of mine got married in Eastbourne, we all ended up in Games Workshop as its the only place not full of pensioners Jan 30 12:18:24 XorA: I won't ask why they got married there... Jan 30 12:20:50 Eastbourne is god's waiting room Jan 30 12:25:04 ade|desk: thats a bloody good description :-) Jan 30 12:28:52 I've only ever driven through it which sounds like it was good planning... Jan 30 12:29:38 RP: average age is probably 65, its like a blackpool with only pensioners Jan 30 12:33:04 Even blackpool would have been better than birmingham and thats saying something :-/ Jan 30 12:33:39 DNScon had many happy years in blackpool Jan 30 12:36:10 I guess I've never been there as an adult so I see it differently... Jan 30 12:36:46 RP: cheap alcohol and no-one minding a bunch of very drunk nerds Jan 30 12:37:02 RP: and strippers Jan 30 12:37:09 Birmingham is not as bad as it used to be Jan 30 12:37:33 i have to go there every 6 months for DR testing. i quite like the centre now Jan 30 12:38:03 lots of places to eat and things to do in the evening Jan 30 12:38:24 ade|desk: We'll see... :) Jan 30 12:39:16 and its cheap to stay for me, i can stay with my parents-in-law just south, Droitwich. Jan 30 12:39:18 ade|desk: strippers? Jan 30 12:39:35 If you want, yes along broadstreet Jan 30 12:40:19 ade|desk: sorted :-D Jan 30 12:40:41 XorA: i thought you were married Jan 30 12:40:53 I am Jan 30 12:43:32 um strippers Jan 30 12:43:35 XorA: hehe, fair play Jan 30 12:44:05 XorA: looking for a gay bar? Jan 30 12:44:13 RP: you are Newcastle way right ? Jan 30 12:46:09 zecke|food: are you Electric Six? Jan 30 12:46:45 XorA: I only know a music group with that name Jan 30 12:47:25 zecke|food: yeah, whos lyrics include "gonna take you to a gay bar, gay bar" :-) Jan 30 12:47:35 heh Jan 30 12:47:40 lol Jan 30 12:47:42 amusing song :) Jan 30 12:48:23 XorA: hmm, I thought I knew this from a peaches song/cover :) Jan 30 12:48:24 http://www.rathergood.com/gaybar/ Jan 30 12:49:14 Crofton: lol Jan 30 12:51:30 ade|desk: want to escape the desk http://www.edinburghrobotics.com/personal/joe/engineers.html Jan 30 12:53:03 are 25k pound worth anything? Jan 30 12:54:21 not enough to support a family on, sadly Jan 30 12:54:54 ~change 25000 gpb to eur Jan 30 12:55:00 25,000.00 Dollar (ZWN) makes 77.03985 Euro (EUR) (from http://www.xe.com/) Jan 30 12:55:13 ~change 25000 gbp to eur Jan 30 12:55:18 25,000.00 Pound (GBP) makes 37,856.09 Euro (EUR) (from http://www.xe.com/) Jan 30 12:56:21 seeing as my wife is wanting us to start breeding Jan 30 12:56:37 ~change 25 gbp to pln Jan 30 12:56:37 hehe Jan 30 12:56:46 25.00 Pound (GBP) makes 149.024 Zloty (PLN) (from http://www.xe.com/) Jan 30 12:57:36 hrw|work: what is a rough average salary in poland ? Jan 30 12:58:13 * ade|desk will back back later --> meeting Jan 30 12:58:52 ade|desk: depends on branch Jan 30 13:04:30 ade|desk: I shall drink beer for a fallen comrade then :-) Jan 30 13:28:30 hi zecke Jan 30 13:44:20 has anyone ever seen this? "!!! [Errno 16] Device or resource busy" during do_package? http://pastebin.ca/332920 Jan 30 13:45:29 are you working with nfs or something? :) just a wild guess Jan 30 13:45:51 hi all Jan 30 13:46:09 Jin^eLD: no, local ext3 filesys Jan 30 13:46:26 weird Jan 30 13:46:33 cbrake: only when I have suffered corruption which needs fsck to fix Jan 30 13:46:48 XorA: :-( Jan 30 13:46:49 cbrake: if you do "df", does it run through or does it hang somewhere? Jan 30 13:46:50 cbrake: might have been a problem due to the recent base.bbclass updates. update OE and rebuild Jan 30 13:47:14 oh.. there you go :> so my wild guesses were wrong :) Jan 30 13:47:49 mickeyl: ok, will do ... Jan 30 13:49:21 ibot, seen schurig Jan 30 13:49:38 schurig was last seen on IRC in channel #oe, 4d 2h 7m ago, saying: 'mickeyl: we're all coughing like mad, but otherwise good'. Jan 30 13:49:41 ade|desk: Yes, I'm near Newcastle Jan 30 13:49:56 oh ... D-A-N-G Jan 30 13:50:02 mickeyl: hmm, update just brought in a change: package.bbclass: don't try to rmdir("./"), it won't work Jan 30 13:52:39 mickeyl: that did not fix, back to debugging ... Jan 30 13:52:44 *fix it Jan 30 13:54:09 mickeyl: never mind, I had to clean it first. Everything seems to be working. Thanks! Jan 30 13:56:22 np Jan 30 14:01:29 XorA: I would drink beer to, but am not allowed until the happy news :( Jan 30 14:03:34 ade|desk: :( Jan 30 14:06:28 ade|desk: run while you still can :-D Jan 30 14:06:44 can't too drained ;) Jan 30 14:09:53 * florian wonders what ade|desk is waiting for Jan 30 14:12:12 have a nice day, see you some day :) thanks agains Jan 30 14:12:13 -s Jan 30 14:12:25 florian: could be blue or could be clear... Jan 30 14:13:04 there's a great advert over here for the most expensive thing a girl will ever pee on. its quite amusing Jan 30 14:13:28 *g* Jan 30 14:52:19 hmm.. in glibc, this "ports" directory, where is it coming from? Jan 30 14:55:36 Jin^eLD: look at SRC_URI: ftp://ftp.gnu.org/pub/gnu/glibc/glibc-ports-${PV}.tar.bz2 Jan 30 14:55:56 ooh, ok now, thanks Jan 30 14:56:16 what is it good for? I am trying to get glibc_2.3.5+cvs20051107.bb from the dreambox branch to work in dev Jan 30 14:58:01 btw, the ports stuff is not in 2.3.x, the line you pasted can be found in 2.4 and such... so I am sort of confused now :) Jan 30 14:59:22 hm.. at some point the arm (and other) ports were moved out of the main tarball, not sure about glibc 2.3.5 Jan 30 15:00:00 I see.. thanks Jan 30 15:29:57 koen|away, I was happy to see that there was a test image for gpe in Angstrom Jan 30 15:30:08 but it does not find the /dev/fb0 :( Jan 30 15:30:36 mccarthy: does on my zaurus Jan 30 15:31:09 doh Jan 30 15:31:31 sorry I meant specifically for the h5000 series (which I have been bugging koen|away about for a while ...) Jan 30 15:31:51 mccarthy: oh feel free to give koen|away a shoeing then :-) Jan 30 15:31:57 good night Jan 30 15:40:40 does anyone here use icecc? Jan 30 15:56:03 how to hand in a patch? Jan 30 15:57:03 bugs.openembedded.org? Jan 30 15:58:25 zecke: I changed the simpad-2.6.conf file and added a new simpad kernel bb file + dir should I make a patch of it and c&p it to the bugtacker? Jan 30 15:58:48 create a bug, attach the files to the bug tracker Jan 30 15:59:04 zecke: which topic should I use then? Jan 30 15:59:06 in bb scripts, do patches have t o be provided by file: or can they be provided remotely? Or can file be used to retrieve a remote URL? Jan 30 16:00:25 nzg: to some degree you can use remote urls Jan 30 16:00:34 mr_nice: whatever you find appropriate Jan 30 16:01:13 This seems to get ignored: SRC_URI +=" http://arm.cirrus.com/files/linux/crunch/crunch-gcc/patch/binutils-2.15-crunch.patch.gz;patch=1" Jan 30 16:02:03 nzg: hmm, then you have a good question (well there are no such thing good vs. bad questions) Jan 30 16:02:23 nzg: I wonder if +=" irritates the old parser Jan 30 16:02:56 tried just adding it to the SRC_URI line, but it didn't seem to pay attention to that either Jan 30 16:03:07 I'm going to try doing it locally Jan 30 16:03:18 Also is it ok that it's a gzip? Jan 30 16:03:41 nzg: hmm, another question I don't know the answer immediately Jan 30 16:03:53 I'll do some experiments Jan 30 16:06:19 I don't see any example of patches using URL's though, I suspect they have to be local. Please correct me if you know otherwise Jan 30 16:06:28 yo koen|edge__ Jan 30 16:06:35 hey XorA Jan 30 16:06:59 XorA: thanks for adding that cairo patch Jan 30 16:07:16 koen|wifi: no probs, I was bored, and it looked funky Jan 30 16:08:20 koen|wifi: GTK is certainly getting faster and faster on my 3200 Jan 30 16:08:37 XorA: yeah, after 2 (3?) years.... Jan 30 16:08:43 XorA: how much is missing before you will call it ueberfast? Jan 30 16:08:56 zecke: integer pixop scaling Jan 30 16:09:06 zecke: and sapwood-like pixmap caching Jan 30 16:09:38 koen|wifi: integer pixmap scaling will be faster than softfloat pixmap scaling? Jan 30 16:09:54 zecke: it should be Jan 30 16:13:27 zecke: it looks as fast as 2.6.X ever did on mine me measuring by eye Jan 30 16:13:34 zecke: I dont beleive in benchmarks Jan 30 16:14:41 nzg: there are some examples of patch files fetched from remote URIs -- see packages/linux/linux-rp_2.6.18.bb Jan 30 16:15:46 nzg: I'm not sure what might be wrong with the SRC_URI you posted. Jan 30 16:16:12 nzg: you may want to start bitbake in interactive mode (bitbake -i), and then peek SRC_URI and make sure it is being set correctly. Jan 30 16:18:22 checking it out.... Jan 30 16:19:04 I am having a weird issue with building an image Jan 30 16:19:16 base_files is trying to do something odd Jan 30 16:20:04 sure are, compressed files too, hmmm Jan 30 16:20:27 expermenting Jan 30 16:20:33 "Failed to move ./ to /home/OE/tmp/work/BLAHBLAH/base-files/./ Device or Resource busy" Jan 30 16:23:04 hmm... when compiling glibc, who exactly defines the endiannes? I have TARGET_ARCH = "mipsel" in my machine.conf but somehow glibc-intermediate is not getting the __MIPSEL define Jan 30 16:23:08 any idea where to look for that? Jan 30 16:23:16 mccarthy: update OE and rebuild base-files Jan 30 16:23:40 mickeyl, you mean mtn pull; mtn update because I tried that ... Jan 30 16:23:51 (or am I missing something else) Jan 30 16:26:17 mickeyl: btw, eabi on armv4t is looking pretty solid from what lennert is telling me Jan 30 16:26:53 * koen|wifi_ needs to dust off his armv4t board soon Jan 30 16:28:16 koen: you should take a group picture of all your toys some time :) Jan 30 16:29:59 hvontres|poodle: like this: http://dominion.kabel.utwente.nl/koen/images/770+ipaqs.jpg ? Jan 30 16:30:08 that's a pretty old picture, btw :) Jan 30 16:34:34 koen|wifi_: yep.. just need to add in all those othre cool gizmos and baords you have laying around :) You might need to use that fisheye lens of yours :) Jan 30 16:35:57 what do I have to do to get a kernel.bb listed as a provider at virtual/kernel? Jan 30 16:36:39 PREFERRED_PROVIDER_virtual/kernel = "my-leet-kernel_2.10.5.bb" Jan 30 16:36:42 ehm Jan 30 16:36:50 PREFERRED_PROVIDER_virtual/kernel = "my-leet-kernel" Jan 30 16:37:57 hvontres|poodle: another old one: http://www.flickr.com/photos/koenkooi/374608756/ Jan 30 16:38:18 I made a simpad-2.6_2.6.19.1.bb file and a simpad-2.6_2.6.19.1 dir than i changed the simpad-2.6.conf to PREFERRED_PROVIDER_virtual/kernel = "simpad-2.6_2.6.19.1" Jan 30 16:38:19 efika + edb9315 + loft Jan 30 16:41:10 cbrake:my BBFILE_PATTERN was wrong, thx for the debug tip, it should work better now Jan 30 16:46:44 hi, all! Jan 30 16:47:22 Here (http://pastebin.ca/333057)is how it looks what did i wrong? the kernel.bb is based on xanadux-un-2.6_2.6.12.bb Jan 30 16:47:27 redefining pxa_pm_enter seems to delay 'entering' suspend, but not exiting from it :) Jan 30 16:47:34 any ideas? Jan 30 16:52:02 mr_nice: I don't know if this will solve your issues, but I thinkd the MACHINE_KERNEL_VERSION ?= "2.4" in simpad.conf should be 2.6 Jan 30 16:53:53 hvontres|poodle: the simpad.conf file is unchanged and worked with the simpad 2.6 kernel I only changed the PREFERRED_PROVIDER variable in simpad-2.6.conf Jan 30 16:54:21 do I need to empty the tmp dir or the chache? Jan 30 16:56:47 mr_nice: what does your BBPATH look like? I didn't see a simpad-2.6 listed in the list of possible providers, so your .bb might not be getting parsed. Jan 30 16:57:13 hi, I'm trying to build directfb, but it looks like a new version is out and oe hasn't been updated. I'm trying learn what this might take to fix. If I change the bb file to look for the new tar ball, the build process pukes when it goes to patch the source. My question is: how are the patches in org.openembedded.dev/packages/directfb/files/ generated. If I can figure that out maybe I can generate new ones... Jan 30 16:57:46 paulcox: a new version and still not 1.0? Jan 30 16:57:53 rc3 Jan 30 16:58:42 paulcox: http://www.openembedded.org/bonsai/query/file/directfb/ Jan 30 16:58:47 rc2 vanished from their servers? Jan 30 16:58:57 as far as I can tell Jan 30 16:59:02 hvontres|poodle: the path is correct - i will kill the tmp dir and try a bitbake -n virtual/kernel Jan 30 17:00:13 paulcox: a) fix the SRC_URI and send a patch (http://directfb.org/downloads/Old/DirectFB-1.0.0-rc2.tar.gz) Jan 30 17:00:25 paulcox: b) update to rc3 and make sure the patches apply and send a patch Jan 30 17:02:46 can do a), may have more questions on doing b) Jan 30 17:03:03 ug, still broken Jan 30 17:03:12 peek binutils PR Jan 30 17:03:12 r3.1 Jan 30 17:03:20 peek binutils-cross PR Jan 30 17:03:20 r3 Jan 30 17:05:01 Which looks like BBFILE_COLLECTIONS is only working at the top level, but then requires are pulling are ignoring thepriority Jan 30 17:05:10 hopefully not, still digging Jan 30 17:08:05 If I add a .bb file at packages/linux is it automaticly parsed to the virtual/kernel target or have I to add it there? Jan 30 17:09:59 mr_nice: more is needed Jan 30 17:10:28 zecke: what I have to do? or where can I find information about it Jan 30 17:10:51 mr_nice: e.g. take a look at RP's great linux-rp example Jan 30 17:10:54 mr_nice, I think it will just appear Jan 30 17:14:23 zecke: wow it's a kind of compley Jan 30 17:14:31 zecke: -y +x Jan 30 17:15:05 zecke: I used the xanadux-un-2.6_2.6.12.bb as base Jan 30 17:15:45 zecke: my problem is that it is not listed as a possible provider Jan 30 17:17:09 zecke: so the main question is - is a .bb file which is putted in packages/linux automaticaly a possible provider for virtual/kernel or is there the need to edit a config file Jan 30 17:22:51 * koen foods Jan 30 18:01:46 mr_nice:the provider information typically comes from the machine.conf Jan 30 18:02:02 mr_nice:and the machine.conf to use is specified in local.conf Jan 30 18:03:15 nzg: yes, but he is asking the other way around Jan 30 18:11:22 zecke thx for botmail, I've dropped him a letter now Jan 30 18:12:26 something is wrong with my BBFILE_COLLECTIONS declaration or something is wrong with BBFILE_COLLECTIONS in general Jan 30 18:12:47 BBFILE_COLLECTIONS = "upstream IPAC9302" Jan 30 18:12:55 BBFILE_PATTERN_upstream = "${HOME}/lab/IPAC-9302/openembedded/org.openembedded.dev/" Jan 30 18:12:55 BBFILE_PATTERN_IPAC9302 = "${HOME}/lab/IPAC-9302/openembedded/IPAC9302/" Jan 30 18:12:55 BBFILE_PRIORITY_upstream = "4" Jan 30 18:12:56 BBFILE_PRIORITY_IPAC9302 = "10" Jan 30 18:13:05 doesn't work Jan 30 18:13:17 It works to a degree... Jan 30 18:13:36 The first reference to a file in the configurations pulls from IPAC9302 Jan 30 18:13:55 but require binutils_${PV}.bb Jan 30 18:14:09 pulls from upstream instead of IPAC9302 for some reason Jan 30 18:16:38 BB docs say "However, if the path specified on the include line is a relative path, BitBake will locate the first one it can find within BBPATH" Jan 30 18:16:53 but this does not appear to be true Jan 30 18:17:39 zecke, bug done:http://bugs.openembedded.org/show_bug.cgi?id=1832 Jan 30 18:18:17 Not sure how to get going w/rc3 Jan 30 18:31:29 ah, my first complete gpe-image compile from .dev. What a feeling. :) Jan 30 18:31:57 Time to take a deeper look why gnokii fails. Jan 30 18:37:43 nzg: what does your BBPATH env variable look like? I've found I need to put my custom directory first in order for *.conf files to be used. Maybe this also applies to the "require" statement. Jan 30 18:42:56 cbrake: if you're overriding .conf files then yes you'd need to put yours first. However, I'd say you should just be making changes in your local.conf or creating new conf files. Jan 30 18:43:03 then again, overriding is ok too Jan 30 18:47:11 JustinP: the problem I ran into was some of the variables in distro/machine conf are assigned with '='. So, in one case it was not possible to override a variable. Jan 30 18:47:50 JustinP: specifically, FEED_URIS in angstrom-2007.1.conf Jan 30 18:48:22 cbrake: right, some vars are to meant to be overridden by users :) Jan 30 18:55:51 03koen 07org.oe.dev * r81a8ce40... 10/ (1 packages/gnokii/gnokii_0.6.14.bb): Jan 30 18:55:51 gnokii: update to 0.6.14 Jan 30 18:55:51 * this one seems to use gtk2, so goodbye gtk 1.x bugs :) Jan 30 19:07:06 hi ! Jan 30 19:07:32 koen: but Gtk2 is fluff? Jan 30 19:08:21 zecke: it is Jan 30 19:08:36 more is more? so more fluff is better? Jan 30 19:08:53 I guess so Jan 30 19:10:30 I haven't oe installed because on my mac doesn't work very well... but I need Xcircuit/ng-spice....and i tried to do native compiling but i had no results. Any1 can try to compile these two programs for openzaurus 3.5.4.1 please ? Jan 30 19:12:45 * koen heads towards the cinema Jan 30 19:17:10 alvise: native compile should work. Alternatively use parallels on your mac, or make OE work once again on mac Jan 30 19:21:21 03koen 07org.oe.dev * r8ee5fde1... 10/ (3 files in 2 dirs): gnokii: disable parallel make Jan 30 19:25:03 alvise: what is failing natively? Jan 30 19:29:04 JustinP, didn't find libxaw (-dev installed..) Jan 30 19:29:16 JustinP, i'm retrying.. Jan 30 19:29:56 alvise: well, then install the -dev Jan 30 19:29:56 zecke, thanks.. i'm installing now ;) Jan 30 19:32:07 JustinP, other libs ? Jan 30 19:38:32 cbrake|JustinP: BBPATH=$OE/IPAC9302/:$OE/org.openembedded.dev Jan 30 19:38:56 my local is first here Jan 30 19:39:20 alvise: what other libs? Jan 30 19:39:32 and also BBFILES = "${HOME}/lab/IPAC-9302/openembedded/${MACHINE}/packages/*/*.bb ${HOME}/lab/IPAC-9302/openembedded/org.openembedded.dev/packages/*/*.bb" Jan 30 19:39:34 here Jan 30 19:41:16 the require statement doesn't appear to be even looking at these path variables, just grabbing it locally Jan 30 19:42:17 nzg: sure some vars get prepended during parsing/evaluating Jan 30 19:43:02 Is that a bug? I'm not sure why that would be desired Jan 30 19:44:14 I don't see how someone can effectively use BBFILE_COLLECTIONS if this is the case Jan 30 19:45:00 also it's not just prepending, because if I remove the local file it doesn't find it at all Jan 30 19:47:19 * nzg steps off his high horse Jan 30 19:47:44 I suppose I'm just frustrated because it puts me back at square 1. Jan 30 19:48:01 JustinP, i mean ... to compile a x program...libx11-dev... Jan 30 19:48:22 the only way to add conditional patches is to either: Jan 30 19:48:36 1. modify the oe source directly and break everything else Jan 30 19:49:07 2. use BBFILE_COLLECTIONS and modify every bb script that references it Jan 30 19:49:26 I just want to find an easy way to do this Jan 30 19:50:10 SRC_URI_append_pn-package-name in local.conf is cheating Jan 30 19:50:43 nzg: the question is why you want to update every bbfile? Jan 30 19:50:55 zecke: I really don't Jan 30 19:51:12 zecke: what I want to do is add a conditional patch to a gcc version Jan 30 19:51:26 which is only applied for 1 specific build Jan 30 19:51:49 nzg: and you currently use BBFILE_COLLECTIONS, then copy this one version Jan 30 19:51:50 SRC_URI_append_pn-package-name sounds interesting Jan 30 19:52:00 zecke:thats right Jan 30 19:52:16 nzg: which includes/requires a couple others and then do the SRC_URI_append? Jan 30 19:52:53 zecke:I don't think that's quite right Jan 30 19:53:20 I'm copying over the gcc base with the SRC_URI in it gcc_3.4.3.bb Jan 30 19:53:24 and modifying that Jan 30 19:53:36 other scripts "require" the base gcc Jan 30 19:53:53 but they ignore the COLLECTION priority so they get the wrong file Jan 30 19:54:46 c what I mean? Jan 30 19:54:50 nzg: oh well, this is the current/'right' way. It looks at toplevel files and then evaluates the priority for these files Jan 30 19:55:16 nzg: so you need to copy the top level files, the ones you build and edit those Jan 30 19:55:25 right, did that Jan 30 19:55:28 o Jan 30 19:55:31 wait no I didn't Jan 30 19:55:35 I see what you man Jan 30 19:55:38 mean Jan 30 19:55:48 I did the bottom files Jan 30 19:56:28 but since SRC_URI is defined at the bottom that means I have to redo the whole tree to append to it? Jan 30 19:56:31 nzg: when I worked with collections, I had three of them. OE upstream, OE hacks and customer Jan 30 19:56:36 alvise: ok...so what's the problem? Jan 30 19:56:49 obvious bug fixes and improvements went to the OE part and were pushed Jan 30 19:57:13 for hacks I copied the whole dir (e.g. packages/xmms) over and edited these Jan 30 19:57:22 and customer stuff, well that is secret anyway :) Jan 30 19:58:05 sure, that would work, but how is that really different from modifying the upstream version? If you copy the whole thing, it's not even using the upstream version Jan 30 19:58:12 is it? Jan 30 19:58:43 nzg: you can always branch and make changes to your local copy then propagate changes from OE's branches.... Jan 30 19:59:17 nzg: We did merge with upstream every couple of month so the hack collection was always quite small Jan 30 20:00:06 well later guys Jan 30 20:00:11 later zecke thanks Jan 30 20:03:14 well, that all works, but it's sort of like using decapatiation as a cure for headaches Jan 30 20:03:21 just not worth it..... Jan 30 20:03:48 I think the best option is to go back to the original plan and create a new gcc version bb script Jan 30 20:04:49 Which seems simpler and cleaner than copying the whole tree and then corrupting it Jan 30 20:05:09 the new bb scripts could then be local Jan 30 20:05:25 and would not require the entire tree.....hopefully Jan 30 20:05:50 nzg: most packages are not as complex as gcc -- everything is contained one packages/<...> directory. Jan 30 20:06:00 nzg: so bbcollections works pretty well for that. Jan 30 20:06:20 cbrake:I don't think it does Jan 30 20:06:46 cbrake:if all requires are local I can't append to the source without copying the whole tree Jan 30 20:06:56 cbrake:which defeats the purpose Jan 30 20:07:11 cbrake:doesn't it? Jan 30 20:07:12 nzg: very few packages use requires, except for *.inc files that are in the same dir as the bb Jan 30 20:07:45 cbrake:hmm, if thats the case Jan 30 20:07:59 nzg: for instance, if I need to modify initscripts for a build, I copy the entire initscripts directly into my oe.custom tree and bbcollections uses it. Jan 30 20:08:08 cbrake:I suppose I could just copy all the gcc_3.4.3 scripts Jan 30 20:08:34 cbrake:you mean the entire package directory, that seems clean, ok Jan 30 20:08:36 nzg: occasionally, I diff -urN the two directories to see what is changing and if there is anything I want to merge into my custom copy. Jan 30 20:08:58 nzg: right -- oe/packages/initscripts Jan 30 20:09:43 nzg: I can then check the custom OE tree into my source control system for the project. Basically any custom packages for the project are contained in one place. Jan 30 20:09:54 cbrake:that sounds like a plan Jan 30 20:10:00 thank you Jan 30 20:10:16 nzg: the custom tree usually only contains a few packages that need to be modified for a typical project, so it is pretty manageable. Jan 30 20:10:50 very nice, back to it, thank you again cbrake Jan 30 20:11:03 nzg: welcome -- good luck! Jan 30 20:11:19 JustinP, checking for main in -lXaw... no Jan 30 20:11:19 configure: error: Couldn't find Xaw library Jan 30 20:11:38 JustinP, ls /usr/lib/libXaw.so Jan 30 20:11:41 /usr/lib/libXaw.so Jan 30 20:12:55 alvise: ok, well, look into the configure script (I mentioned this before) and the config.log and see what is causing it to fail Jan 30 20:13:22 alvise: you probably need another -dev package. If you can't figure out which package has the file(s) you need just ask. Jan 30 20:14:04 JustinP, ok...thx ;) Jan 30 20:14:40 alvise: there may be a Packages.files or simila rin the feed dir as well Jan 30 20:15:36 JustinP, may i give you configure.log and/or configure script ? Because i don't understant what happens Jan 30 20:17:18 alvise: I don't have time for that Jan 30 20:17:34 alvise: look in the config.log for Xaw and see what it's fialing on Jan 30 20:18:58 ok Jan 30 20:21:59 JustinP, when you have time .. http://pastebin.ca/333299 Jan 30 20:25:42 alvise: looks like you have diet-x11 installed instead of the real x11 Jan 30 20:25:54 just a guess... Jan 30 20:25:58 JustinP, is possible to install real x11 ? Jan 30 20:26:11 yes Jan 30 20:26:17 JustinP, yea :D Jan 30 20:26:31 JustinP, ipkg ? Jan 30 20:26:40 yeo Jan 30 20:26:43 yep Jan 30 20:26:48 JustinP, thx ... Jan 30 20:26:49 you just need to remove diet-x11 and install x11 Jan 30 20:27:02 JustinP, thanks... I love you ghghhg ;) Jan 30 20:28:55 JustinP, mmm .. scuse for disturbing another time.... there aren't *diet* and x11 packages in the feed :( Openzaurus feed i mean.. Jan 30 20:29:53 no diet package? Jan 30 20:30:22 sure you looked in the right feed? Jan 30 20:30:29 JustinP, on openzaurus not... i have tried to see on openzaurus feed.. but nothing (and no "x11" ) package.. only libx11 Jan 30 20:31:00 JustinP, openzaurus .. not angstrom.. Jan 30 20:31:35 you don't see any diet-x11 or libx11? Jan 30 20:31:55 JustinP, libx11 yes .. diet-x11 no Jan 30 20:32:08 hmph Jan 30 20:32:11 libx11-6 Jan 30 20:32:21 looks like they may be packaging diet-x11 as libx11.. Jan 30 20:32:21 the fools Jan 30 20:32:52 JustinP, well... only diet available then.. Jan 30 20:32:57 quite possibly Jan 30 20:33:16 you can try patching your program to remove those function calls... Jan 30 20:33:21 or perhaps it's just a linking error Jan 30 20:33:30 mayb eyou need to add another lib to the linking Jan 30 20:34:14 JustinP, ok.. thx :) Jan 30 20:36:02 JustinP, ... may(only if it doesn't disturb you) regive me gpe-angstrom image ? :).. i got only e-image that works not very well.. and bootstrap doesn't start x Jan 30 20:36:38 well..... same for Jan 30 20:36:50 angstrom.. there isn't diet-x11.. Jan 30 20:37:58 not right now Jan 30 20:38:21 JustinP, ok np .. thank you and scuse me for disturb Jan 30 20:49:09 has angstrom full X11 ? Jan 30 20:52:13 generally diet-x11 is shipped for embedded devices Jan 30 20:52:30 I'm surprised there isn't a full x11 in the feeds, though... Jan 30 20:58:17 koen: I have tested you gnokii changes. This fixes the bug for me. Should I close #1826 or just add a comment? Jan 30 21:08:14 JustinP:I think that's because people are moving to the Trolltech stuff which bypasses X11 completely Jan 30 21:10:14 nzg: you mean Qt/e? No, we're generally moving *away* from that. Jan 30 21:18:41 JustinP: I do. We are?. Why's that? Jan 30 21:18:47 I'm having some trouble getting e2fsprogs to compile Jan 30 21:19:10 I think the problem is due to the fact the I am using an external toolchain Jan 30 21:19:11 nzg: because it's yet another thing to support. It's easier to use X and build on top of what is already out there Jan 30 21:19:15 not to mention more flexible Jan 30 21:19:24 ah Jan 30 21:19:46 getting warnings about ld-linux.so.2 not found ((try using -rpath or -rpath-link)) Jan 30 21:20:19 Qt seems easier from a device driver writers point of view, just write the framebuffer driver Jan 30 21:20:32 but then I've never tried to write an X hardware driver..... Jan 30 21:20:34 | have had this before and have fixed it by setting EXTRA_OEMAKE += " LDFLAGS ='${LDFLAGS} " Jan 30 21:20:51 and that does indeed fix the problem in this case too Jan 30 21:21:16 i agree with nzg Jan 30 21:21:34 however I hit another problem... which is the one I'd appreciate some help understanding Jan 30 21:22:12 we do still support opie but it could use some more love Jan 30 21:22:30 it's now complaining that strcasecmp is defined twice Jan 30 21:22:51 I can fix by adding -DHAVE_STRCASECMP in the cflags Jan 30 21:23:12 blaarg, cirrus patches are incompatible with oe patches... nothings easy I tell ya Jan 30 21:23:37 before I do that it thinks that HAVE_STRCASECMP is not defined and so attemps to define it Jan 30 21:23:52 hence the redefinition error Jan 30 21:24:02 * nzg closes his eyes and prays it's whitespace Jan 30 21:24:11 but I don't get why HAVE_STRCASECMP is undefined Jan 30 21:24:23 it is defined in the main util.h Jan 30 21:24:38 (the one from the toolchain) Jan 30 21:25:01 yet when autoconf runs, it doesn't detect it Jan 30 21:25:34 -I doesn;t include the toolchain perhaps? Jan 30 21:25:52 how do I get it to include that at the configure stage tho? Jan 30 21:26:07 it is set in CFLAGS Jan 30 21:26:33 (-I${PRE_BUILT}/include) Jan 30 21:26:35 then it should pull it in Jan 30 21:26:50 check the config.log's gcc command Jan 30 21:31:57 JustinP: You mean where it says 'checking for i586-stv-linux-gcc'? Jan 30 21:41:42 the toolchains include path is included at configure time as far as I can see. Jan 30 21:41:47 tkp: I mean the part where it checks for stcasecmp Jan 30 21:41:59 it looks like the test fails Jan 30 21:42:16 and it looks like it's because my LDFLAGS are not there Jan 30 21:42:35 warning: ld-linux.so.2, needed by /usr/local/i586/oe/lib/gcc/i586-stv-linux/3.4.4/../../../../i586-stv-linux/lib/libc.so.6, not found (try using -rpath or -rpath-link) Jan 30 21:43:21 that normally goes away when I have my LDFLAGS in there, which include: Jan 30 21:43:23 -L${PRE_BUILT}/lib -Wl,-rpath-link,${PRE_BUILT}/lib Jan 30 21:44:32 that results in lots of errors similar too: Jan 30 21:44:47 /usr/local/i586/oe/lib/gcc/i586-stv-linux/3.4.4/../../../../i586-stv-linux/lib/libc.so.6: undefined reference to `_rtld_global@GLIBC_PRIVATE' Jan 30 21:48:28 koen: I added a comment and closed the bug. Jan 30 21:48:51 tkp: well...the you should look into those I guess Jan 30 21:49:55 I remember trying to look into that one a while ago... the internet seems to not have much of a clue about it Jan 30 21:53:13 evening Jan 30 21:54:30 ok. I think I understand a little better... Jan 30 21:55:37 but it seems that the answer is infact to do what I was doing and passing in those LDFLAGS ( -rpath etc ) Jan 30 21:56:00 what is the best way to start truly grokking the OE bitbake system? Just start writing bb files? Jan 30 21:56:01 so... how can I get my LDFLAGS in at configure stage Jan 30 21:56:06 it's using autoconf Jan 30 21:56:26 and it managed to get my CFLAGS Jan 30 21:57:06 tkp: CFLAGS="..." LDFLAGS="..." ./configure Jan 30 21:57:19 tkp: that's how you normally do it. You should be able to just export them before the configure is run Jan 30 21:57:29 tkp: do_configure_prepend() or some-such Jan 30 21:57:37 I had them the other way round Jan 30 21:57:42 LDFLAGS first Jan 30 21:57:48 that matters? Jan 30 21:58:08 bradbev: sure. I suggest looking at some simple ones at first. Like packages/mpd/mpc_X.bb Jan 30 21:58:15 * JustinP shrugs Jan 30 21:58:19 hmm... not it says "checking for strcasecmp... (cached) no" Jan 30 21:58:21 it's how I've seem it done Jan 30 21:58:24 so that made some difference Jan 30 21:58:28 JustinP, thanks Jan 30 21:58:31 didn't say cached before Jan 30 21:58:40 well, it's probably because you just re-ran it Jan 30 21:58:53 you probably need to remove the automate cache or something Jan 30 21:59:17 remove the automatic cache? Jan 30 21:59:35 I bitbaked it clean Jan 30 21:59:54 or is that todo with ccache? Jan 30 22:02:13 might have to do with ccache...not sure Jan 30 22:02:22 I usually remove my ccache when I'm having such problems Jan 30 22:02:31 * JustinP knows little of those things really Jan 30 22:04:33 ok. so the LDFLADS were not in when the test run Jan 30 22:04:46 but further down the page it has ac_cv_env_LDFLAGS_value=".... Jan 30 22:04:49 and it's all there Jan 30 22:07:31 * JustinP doesn't know Jan 30 22:07:41 I don't think anyone understands auto* Jan 30 22:08:15 gotto be something in configure.in Jan 30 22:08:32 but I can't read than damn ugly file! Jan 30 22:43:42 I take it that bitbake -b frob.bb -c listtasks doesn't actually list the correct _order_ of the tasks that will run? Jan 30 22:43:54 is there a way to get that order? Jan 30 22:44:00 it is not? Jan 30 22:44:15 well, maybe, but the kernel order looks weird.. Jan 30 22:44:28 do_compile Jan 30 22:44:28 do_stage Jan 30 22:44:28 do_build Jan 30 22:44:28 do_mrproper Jan 30 22:44:28 do_fetch Jan 30 22:44:29 do_configure Jan 30 22:44:34 I see.. Jan 30 22:44:42 :-P Jan 30 22:44:54 not the order _I'd_ build a kernel :) Jan 30 22:44:58 bradbev: you are using trunk? 1.7.x Jan 30 22:45:13 bitbake? no, 1.6.?? Jan 30 22:45:28 1.6.2 Jan 30 22:45:35 time to upgrade? Jan 30 22:45:38 report a bug? Jan 30 22:45:57 :( Jan 30 22:46:22 Is "listtask" a bitbake thing, or is that an OE defined command? Jan 30 22:46:56 all tasks are defined by OE Jan 30 22:47:46 ah. So BB really is only the framework manager. I wasn't sure where the lines were drawn Jan 30 22:48:03 bitbake is task executor Jan 30 22:48:34 yeah, but I don't think I really understood that 30 minutes ago :) Jan 30 22:52:13 cu - sleeptime Jan 30 22:53:06 BB trunk is more broken. Where should I look to see where the code for the listtask command is? Jan 30 22:53:58 OE/classes/base.bbclass Jan 30 22:54:15 thanks. I should grep before I ask :) Jan 30 22:54:32 bye Jan 30 22:54:40 03philippe 07org.oe.dev * r348b7ebe... 10/ (1 packages/gpe-mini-browser/gpe-mini-browser_0.20.bb): Jan 30 22:54:40 packages/gpe-mini-browser/gpe-mini-browser_0.20.bb Jan 30 22:54:40 * add bb file for newly released version Jan 30 22:56:37 ok... so e2fsprogs doesn't use CFLAGS pr LDFLAGS for it's autoconf tests... Jan 30 22:56:49 instead it uses --with-ldopts and --with-ccopts Jan 30 22:57:54 I need the '-rpath-link,${PRE_BUILT}/lib' ldflags so I can use my external toolchain Jan 30 22:59:44 I cant set --with-ldopts="${CFLAGS} Jan 30 23:00:22 why not? Jan 30 23:01:40 because it complains that the -I opts are not valid options Jan 30 23:01:47 * tkp thinks quotes are needed... Jan 30 23:03:23 nowtf... quoting it resulted in: http://rafb.net/p/fKvJHc18.html Jan 30 23:03:47 that was from: EXTRA_OECONF += " --sbindir=${base_sbindir} --with-ldopts='${LDFLAGS}'" Jan 30 23:04:04 where did all the other qoutes come from! Jan 30 23:05:20 oh no... ok, that was bitbakes reproduction of the error Jan 30 23:05:38 but still configure: error: unrecognized option: -L/usr/local/i586/oe/i586-stv-linux/lib Jan 30 23:06:33 I need to somehow extract just the rpath LDFLAG if it exists, and pass that in to --ldopts Jan 30 23:14:37 tkp: you sure? You should be able to quote it correctly....just add ' or something Jan 30 23:18:46 JustinP: nope Jan 31 00:09:07 ~lart DOM Jan 31 00:09:08 * ibot installs a bad bootloader on DOM and turns DOM into a brick Jan 31 00:09:38 ~lart XPath Jan 31 00:09:38 * ibot acting on orders from an unspecified client drags XPath into court suing for $200 million Jan 31 01:32:23 hello...question: I'm trying to bitbake glibc; after compiling without any issues it tries to update locals, and it hangs for several hours. Output: http://pastebin.ca/333685 Jan 31 01:32:52 I tried adding ENABLE_BINARY_LOCALE_GENERATION="0" to local.conf but it gave same output Jan 31 01:33:10 you might need to rebuild then Jan 31 01:33:25 kk Jan 31 01:34:06 to do that I delete the glibc folder in work and all the files in stamps that have glibc in them? Jan 31 01:35:00 bitbake -v rebuild glibx Jan 31 01:35:05 glibc Jan 31 01:35:06 thx Jan 31 01:35:08 ack! Jan 31 01:35:12 bitbake -c rebuild glibc Jan 31 01:35:32 lol what did I just do to my database? Jan 31 02:01:32 hmmm same problem Jan 31 02:06:32 damn Jan 31 02:24:50 calmofthestorm7: perhaps you need to edit conf/distro/openzaurus* to make that change Jan 31 02:25:25 sorry, to make what change? Jan 31 02:25:33 I'm not fully sure what's wrong Jan 31 02:25:46 calmofthestorm7: ENABLE_BINARY_LOCALE_GENERATION = "0" Jan 31 02:27:01 ah Jan 31 02:27:07 sure lemmie try it Jan 31 02:27:29 sorry, is that the conf dir under build or under org.... Jan 31 02:27:40 org Jan 31 02:44:10 JustinP: So by disabling locale generation, will this break anything? Jan 31 02:44:32 possibly break i18n Jan 31 02:44:45 you're not too likely to encounter it if you don't use locales other than english Jan 31 02:44:52 so I'm only building glibc so I can build other programs Jan 31 02:45:00 eg, I'm not going to be using this build on the actual Zaurus Jan 31 02:45:06 I just want to be able to build gaim-encryption Jan 31 02:45:17 which links against gaim, which links against glibc Jan 31 02:45:26 then it should make no difference (I think) Jan 31 02:45:33 coolness Jan 31 02:45:33 thx Jan 31 02:45:46 yeah sorry for all the n00b questions;) Jan 31 02:45:53 np **** ENDING LOGGING AT Wed Jan 31 02:59:59 2007