**** BEGIN LOGGING AT Sat Nov 11 02:59:57 2006 Nov 11 04:02:03 anyone here dev for openzaurus? Nov 11 04:24:44 zaurus ieee80211_crypt registered algorithm 'NULL' anyone know what thats about? Nov 11 06:20:52 anyone here know what to do about: Nov 11 06:20:52 NOTE: preparing tree for binary locale generation Nov 11 06:20:52 NOTE: generating locale es_NI (UTF-8) Nov 11 06:20:52 qemu-arm: symbol lookup error: qemu-arm: undefined symbol: _llseek Nov 11 07:00:00 03mwester 07org.oe.dev * rb1116e7c... 10/ (14 files in 6 dirs): Nov 11 07:00:00 Unslung: major change - Linksys busybox now subsumed by slingbox; cpio and find Nov 11 07:00:00 are no longer required; rc.bootdisk functionality introduced in unslung-rootfs Nov 11 07:08:52 good morning Nov 11 09:53:00 03koen 07org.oe.dev * re828cc82... 10/ (3 files in 3 dirs): sysvinit: fix divide by zero at reboot for angstrom Nov 11 09:53:50 RP: if you ever want to properly shutdown/reboot poky you need re828cc82.. :) Nov 11 10:21:11 03koen 07org.oe.dev * r8f6e401a... 10/ (6 files in 4 dirs): Nov 11 10:21:11 initscripts: fix RO root problem for angstrom Nov 11 10:21:11 * I suspect it applies to all 2.6 based devices Nov 11 10:38:22 koen: Nice, that double reboot enter corrects. Nov 11 10:38:27 Hi. Nov 11 10:49:10 03koen 07org.oe.dev * r68c7a099... 10/ (1 packages/xserver-common/xserver-common_1.11.bb): Nov 11 10:49:10 xserver-common: add 1.11 Nov 11 10:49:10 * should fix #1567 Nov 11 11:20:28 ~botmail for zecke: http://www.joelonsoftware.com/articles/fog0000000043.html Nov 11 11:33:12 03koen 07org.oe.dev * r9ee83d09... 10/ (1 packages/xserver-common/xserver-common_1.12.bb): xserver-common: add 1.12 which incorperates all patches from OE minus the setDPI.sh script Nov 11 11:49:38 morning guys Nov 11 11:54:07 morning mickeyl Nov 11 11:55:18 mickeyl: are there some better infos about releasing the openmoko and the code? Nov 11 11:57:00 openembedded.org frontpage :) Nov 11 11:57:00 and planet.linuxtogo.org Nov 11 11:57:19 greentux: the best info I can give you is January 2007, and even this is _very_ optimistic :D Nov 11 11:57:25 (between you and me) Nov 11 11:57:49 seriously, we try to release it as soon as the core stuff works Nov 11 11:58:17 mickeyl: and are there some distributors for europe? Nov 11 11:58:34 mickeyl: in the moment it is impossible to buy the real geekstuff in europe... Nov 11 11:58:44 greentux: no need for distributors. FIC will sell it worldwide direct Nov 11 11:59:05 mickeyl: mmh but in europe you need something like CE tests etc... Nov 11 11:59:28 greentux: good point. i have no idea about these things frankly. Nov 11 11:59:30 mickeyl: the same problem with the zaurus... Nov 11 11:59:42 yeah, you're right. i remember TriSofts problems Nov 11 11:59:49 i know TriSoft very well Nov 11 11:59:56 mickeyl: perhaps trisoft is interested Nov 11 11:59:59 i think i could talk them into doing that Nov 11 12:00:07 thanks for your suggestion. /me makes a note Nov 11 12:00:31 i talked to them because the nice sony ux-90, but they said: to much problems... Nov 11 12:00:43 which kind of problems? Nov 11 12:00:54 but there is a german ebay dealer now.. with warranty etc etc. Nov 11 12:01:49 I think es price for the sony is to high and they have no direct contact to sony. Nov 11 12:01:49 hey, please /join #openmoko, i don't want to pollute #oe Nov 11 12:01:56 booth should be better with the openmoko Nov 11 12:11:42 03koen 07org.oe.dev * re70382d5... 10/ (6 files in 3 dirs): Dbus: add 1.0.0 release, needs testing before removing DEFAULT_PREFERENCE = -1 Nov 11 12:34:55 good morning all Nov 11 12:36:34 hey cyrilRomain Nov 11 12:55:17 and the next commit is for mickeyl and his secret project :) Nov 11 12:56:01 03koen 07org.oe.dev * r06d5f378... 10/ (6 files in 3 dirs): Nov 11 12:56:01 mb-applet-inputmanager: apply patch from Phil Blundell to pop up input method on focus Nov 11 12:56:01 * requires libgtkinput to be loaded Nov 11 12:56:01 * see http://www.nabble.com/libgtkinput-t971843.html Nov 11 12:56:01 * this also brings mbinputmgr-tray.c in sync with o-hand svn rev 1163 Nov 11 12:58:59 Hi openembedded irc Nov 11 12:59:14 this cgi:irc is too slow Nov 11 12:59:46 hey acassis Nov 11 13:01:26 Hi koen, how are you? Nov 11 13:01:28 guys I created a bitbake package to ipatience (the patience game developed by Keith Packard), how can I submit it to GPE and OE ? Nov 11 13:01:42 attach it to the bugtracker Nov 11 13:02:34 sure, the .bb and the tarball package? Nov 11 13:02:59 acassis: just the .bb file Nov 11 13:03:40 'mtn diff packages/ipatience' would be fine as well Nov 11 13:03:46 sure, them I need to create a patch with my modifications Nov 11 13:04:01 them = then Nov 11 13:05:24 if you want it in gpe, send a mail to gpe-list@linuxtogo.org Nov 11 13:05:49 03koen 07org.oe.dev * re2fde1c5... 10/ (2 files in 2 dirs): matchbox-applet-inputmanager: clean up Nov 11 13:06:31 just wait a moment, I will try to found other cgi:irc to connect to freenode, I am at my university it block 6667 port Nov 11 13:10:31 testing Nov 11 13:10:34 good Nov 11 13:11:28 koen: I can convert most of OE's into an AST in about two seconds on my macbook Nov 11 13:12:50 nthere is one issue with the lexer and adding handlers (sanity.bbclass and tinderclient.bbclass) can't be parsed Nov 11 13:13:08 koen: the ipatience is on Packard's CVS, will it be moved to OE repository after? Nov 11 13:15:29 other question, you suggest ask at GPE list to include it on GPE, but it use FLTK and C , I think it is a problem. Maybe it need to be converted to GTK Nov 11 13:16:04 C plus plus, the plus signal don't appear Nov 11 13:20:26 acassis: OE does not mirror project repositories. Your .bb file should reference the real repositository (the Packard's one). Nov 11 13:21:23 acassis: it would be too space and time consuming mirroring and syncing more than 4000 packages Nov 11 13:22:31 03zecke123 * r637 10bitbake-parser/bitbakescanner.l: (log message trimmed) Nov 11 13:22:31 Fix parsing of tinderclient.bbclass and sanity.bbclass Nov 11 13:22:31 def foo(e): Nov 11 13:22:31 text text Nov 11 13:22:31 addhandler moo Nov 11 13:22:32 python booo Nov 11 13:22:34 The lexer entered state S_DEF as the def method began. When Nov 11 13:22:48 cyrilRomain: but if the CVS go out, it will fail, sure? Nov 11 13:23:00 acassis: yes Nov 11 13:23:04 CIA-4: respect my authority and do not trruncate log messages! Nov 11 13:23:21 acassis: it is up to the package provider to ensure his repository is up Nov 11 13:25:06 acassis: and it is up to the .bb maintainer to check the SRC_DIR is correct Nov 11 13:27:21 zecke: Sounds interesting :) Nov 11 13:27:44 zecke: You're going to hit the same problem as the C parser though? No unit tests? :-/ Nov 11 13:27:59 RP: it is the C parser :) Nov 11 13:28:05 RP: but another strategy to make use of it Nov 11 13:28:36 RP: where I hope to get a real speed improvement is stuff like this Nov 11 13:28:40 A = "100" Nov 11 13:28:51 A += "500" Nov 11 13:29:09 we will save all prepends, appends for each variable in a dict Nov 11 13:29:21 dict['A'] = [] Nov 11 13:29:31 for prepend we will prepend to the list, for append ... you know Nov 11 13:29:48 now we get to the end of evaluate, or include/require/inherit or an immediate assignment Nov 11 13:30:07 zecke: That sounds interesting and potentially quite useful :) Nov 11 13:30:09 and then and only then we will need to rollback these arrays into a string Nov 11 13:30:35 e.g. d.setVar('A', "".join(dict['A'])) Nov 11 13:30:52 Sounds good to me Nov 11 13:31:29 RP: i think there is a lot of potential. With the fast lexer/grammar the only performance key will be do less lookups Nov 11 13:31:32 do less expands :) Nov 11 13:32:00 Yes. expands are the cause of the slowdown... Nov 11 13:32:16 I have ideas about tracking variable dependencies Nov 11 13:32:47 now I need to do my physics assignment :( Nov 11 13:32:56 :( Nov 11 13:33:08 I need to find some food, then get back to building work ;-) Nov 11 13:33:26 building work? Nov 11 13:33:38 you found a house? or you are forced to repair your families house(s)? Nov 11 13:34:06 I volunteered to help with my brothers house... Nov 11 13:34:28 Its slowly gaining things like the correct number of walls and floors :) Nov 11 13:34:29 03zecke123 * r638 10bitbake-parser/bitbakescanner.l: Double new line was a bad idea this broke icecc.bbclass ;) Nov 11 13:34:53 BTW: holding the ast for all BBFiles, inc files, conf, classes Nov 11 13:34:59 is about ~50 MB of memory :} Nov 11 13:36:00 Not bad. We might be able to lose the cache eventually :} Nov 11 13:36:48 The biggest startup problem now with bitbake is we need to improve the file stat cache... Nov 11 13:36:58 It sucks badly if the kernels disk cache is cold Nov 11 13:37:17 fix the kernel :) Nov 11 13:37:20 but arguably that's the kernel sucking, not bitbake :) Nov 11 13:37:27 we need a warmup syscall Nov 11 13:37:32 hi zecke Nov 11 13:37:32 warmup(recursive, dir) Nov 11 13:37:54 it should be like READAHEAD with mmap but for whole directories Nov 11 13:38:07 hi pb_ Nov 11 13:38:12 hi rp Nov 11 13:38:14 pb_: hey master, how are you? Nov 11 13:38:41 * RP -> food Nov 11 13:39:07 * zecke -> GNUplot Nov 11 13:39:10 zecke: very well, thanks. how are you? Nov 11 13:52:56 * pb_ going to Enfield now Nov 11 13:52:57 later all Nov 11 14:03:06 mickeyl: Morning Nov 11 14:03:11 hey Nov 11 14:03:18 mickeyl: Do you mind taking a look at bug 1551? Nov 11 14:03:40 zecke, mickeyl: Is there somekind of debian package for bitbake around, or should the guy next to me start from scratch? Nov 11 14:03:44 BTW, I cannot wait to get my hand on that openmoko thing. I hope it also has WLAN. Nov 11 14:03:57 Laibsc1: it hasn't Nov 11 14:04:15 you know, wifi is a pain in the ass Nov 11 14:04:19 i want it, too Nov 11 14:04:20 but Nov 11 14:04:22 stefan_schmidt: There is one package openembedded-essential with the necessary as dependencies. Nov 11 14:04:23 stefan_schmidt: you can run bitbake straight from the svn checkout :) Nov 11 14:04:29 it's tough to find chipsets w/ open drivers Nov 11 14:04:31 power friendly Nov 11 14:04:38 and small Nov 11 14:04:55 I was not aware of that. Nov 11 14:05:04 CF card slot too big? Nov 11 14:05:12 much too huge, yeah Nov 11 14:05:24 we had to go micro-SD to make it small Nov 11 14:05:31 Laibsch: I know oe-essential, but like to have a bitbake package. Nov 11 14:06:10 Laibsch: re. 1551, you're supposed to have the 2.95.3 toolchain in your PATH before running bitbake. doesn't that fix it? Nov 11 14:06:10 koen: Of course, I'm doing this all the time. But have the released version officially in debian would be nice, right? Nov 11 14:06:18 stefan_schmidt: it would be nice, imo Nov 11 14:07:05 mickeyl: I think that is what I did. But let me recheck. Nov 11 14:07:47 stefan_schmidt: bitbake is a moving target so it is in CVS. At least that is my understanding. Nov 11 14:07:50 mickeyl: ok, he starts from scratch. I'll post an url for testing before we start to push this into the new queue. Nov 11 14:08:08 stefan_schmidt: cool, thanks. I'm not using debian so i can't test, but a lot of guys here do Nov 11 14:08:14 Laibsch: hmm, I see releases of it. Nov 11 14:08:55 stefan_schmidt: How about you preparing a package, then? ;-) Nov 11 14:09:04 I see that a lot of devs will stay with svn. But for people just want to have working bitbake releases are fine. Nov 11 14:09:45 i agree Nov 11 14:09:48 Laibsch: heh, bitbake 1.6.2 package is in progress right now. Nov 11 14:13:10 stefan_schmidt: nice to know. Nov 11 14:29:03 which python version are useable with bitbake? Only 2.4 or still 2.3? Nov 11 14:34:49 stefan_schmidt: the deb package says >= 2.3 Nov 11 14:34:57 So my guess is 2.3 is OK Nov 11 14:35:22 Laibsch: ok Nov 11 14:35:29 stefan_schmidt: Are you going to host that package anywhere? Nov 11 14:37:18 Laibsch: We will provide a temp repo for testing. Once all problems are fixed we will push it into debian. Nov 11 14:37:57 stefan_schmidt: You have commit access to debian? Even better. Nov 11 14:39:20 Laibsch: Shoragan, the guy next to me, is in the new maintainer process. He already is maintainer of some python packages. Nov 11 14:39:32 cool Nov 11 14:45:09 zecke: Is it right that lib/bb/methodpool.py is BSD licensed and not GPL? Nov 11 14:49:07 Wow, GPL, BSD and MIT lincense in bitbake Nov 11 14:53:52 mickey|sports: There is no mention of the need to fiddle with $PATH in sharprom-compatible.conf. How about just setting $PATH in the bb file as suggested in the bug report? Any kind of drawbacks to that? I attached a patch to the bug 1551 (as an added bonus the patch will also correct a typo ;-)) Nov 11 14:58:28 stefan_schmidt: hehe, it is a licensing hell :) Nov 11 14:58:36 stefan_schmidt: old bits GPL, new bits BSD/MIT Nov 11 14:58:54 stefan_schmidt: I wonder what binary means with python files Nov 11 14:59:21 stefan_schmidt: if it eases some stuff I can changed from BSD to MIT Nov 11 15:03:18 zecke: hmm, what is the problem to have only one license? I'm not sure how compatible files with different licenses are in one distro. Nov 11 15:03:35 * stefan_schmidt says IANAL Nov 11 15:04:09 zecke: All people whcih contributed are still known? Nov 11 15:04:26 stefan_schmidt: well all these are GPL compatible. So we could relicense bitbake to GPL Nov 11 15:04:38 stefan_schmidt: but my license of the year is MIT/BSD :) Nov 11 15:04:52 stefan_schmidt: and yes the contributors are known and should be in this channel ATM Nov 11 15:05:51 stefan_schmidt: oh and the new parser is currently licensed as "Postcard license" which is likely to be GPL incompatible Nov 11 15:06:02 koen: I created the patch using it diff -urNd ../ipatience-0.1-r2/ipatience-0.1 ipatience > fix-patch.diff Nov 11 15:06:19 zecke: Personally I prefer gpl, but have no problem with BSD/MIT. It would good to have _one_ license. :) Nov 11 15:06:25 is it right? Nov 11 15:07:42 zecke: Is the new parser already in bitbake 1.6.2? Nov 11 15:07:43 you said me to use mtn, but I don't know how to create it with monotone Nov 11 15:07:43 stefan_schmidt: no, and it is a long way until it is done Nov 11 15:08:10 the man page don't help me Nov 11 15:08:29 stefan_schmidt: I think the list of contributors are: kergoth, pb_, mickeyl, schurig, hrw, rp, mithro, me Nov 11 15:08:55 some suggestion? Nov 11 15:09:10 stefan_schmidt: I have stopped GPLing stuff. People can run away with GPL software as well and then it is so ugly you don't want to even look at it :) Nov 11 15:10:00 zecke: As I said, I have no problem if you like to use BSD/MIT. :) Nov 11 15:10:25 stefan_schmidt: I just use this possibility to not finish my assignment ;) Nov 11 15:11:30 zecke: But we, Shoragan and me, have no idea if it is a legal problem to mix up all this licenses in the way bitbake is doing it. Nov 11 15:12:06 zecke: And it is no fun to get a package rejected after 4 weeks in the new queue because of legal problems. :) Nov 11 15:13:34 stefan_schmidt: IANAL, it is.. hmm the new GNU.org site is hard to navigate :( Nov 11 15:13:55 * stefan_schmidt hates to deal with legal prblems to much. :( Nov 11 15:15:12 stefan_schmidt: http://www.gnu.org/licenses/license-list.html#TOCGPLCompatibleLicenses Nov 11 15:16:34 stefan_schmidt: so X11==MIT License and BSD license without advertising clause is compatible with GPL Nov 11 15:17:00 stefan_schmidt: the reason is quite simple. You could just relicense the MIT/BSD code under GPL because the license is permissive enough Nov 11 15:17:57 Can someone help me with this problem: I want my bitbake package apply a patch on source download from cvs, I create a patch but bitbake is not found my patch Nov 11 15:18:50 zecke: ok, so there should be no problem. Nov 11 15:19:00 IANAL :) Nov 11 15:19:12 zecke: of course. :) Nov 11 15:19:25 i would be schocked if this would be illegal Nov 11 15:19:39 zecke: But it sounds good. ;) Nov 11 15:20:25 I am this error msg: Patch /"file://fix-make.patch does not exist Nov 11 15:20:36 iamtheacassis: where did you store the patch Nov 11 15:20:39 I getting this error msg: Patch /"file://fix-make.patch does not exist Nov 11 15:20:55 zecke: Anyway it could be helpfull for bitbake to have one license. But that is of course the decision of bitbake devs and not mine. Nov 11 15:21:29 zecke: I place it on org.openembedded.dev/packages/ipatience/ipatience-0.1 Nov 11 15:22:23 iamtheacassis: okay did you set PV to "0.1" as well? Nov 11 15:23:35 zecke: no, but before that I tried ipatience with no 0.1 attached and get the same problem Nov 11 15:23:57 I think I need add this patch to monotone Nov 11 15:24:12 iamtheacassis: no, you don't :) Nov 11 15:24:18 sure Nov 11 15:24:40 iamtheacassis: OE tries to find the patch in ${PN}-${PV}, then in files/ and then in '.' of the directory Nov 11 15:24:43 zecke: this is my SRC_URI = "cvs://anoncvs@keithp.com/local/src/CVS;method=pserver;module=ipatience \"file://fix-make.patch;patch=1" Nov 11 15:25:16 iamtheacassis: correct PN, correct PV or just store the patch in files Nov 11 15:25:16 zecke: sure, I will place PV = 0.1 and test Nov 11 15:30:41 zecke: still getting the same error Nov 11 15:31:31 iamtheacassis: then strace is your friend Nov 11 15:31:52 e.g. strace -e open bitbake ipatience.bb -cpatch Nov 11 15:32:02 and no you can't copy and paste it... Nov 11 15:35:54 zecke: I still getting error Nov 11 15:36:02 the same error Nov 11 15:36:11 16:51 < zecke> iamtheacassis: then strace is your friend Nov 11 15:36:11 16:51 < zecke> e.g. strace -e open bitbake ipatience.bb -cpatch Nov 11 15:36:11 16:52 < zecke> and no you can't copy and paste it... Nov 11 15:37:03 with strace you will see where BitBake tries to open the files... Nov 11 15:37:26 and see above for the lookup order of bitbake Nov 11 15:37:38 later Nov 11 15:38:17 zecke: sure, I will use it Nov 11 15:43:22 acassis: I think you should remove the double quote before your '"file://' Nov 11 15:43:46 if your paste here at 17:47 is correct Nov 11 15:58:09 cyrilRomain: but the other package use file:// Nov 11 15:58:18 others packages Nov 11 15:58:38 acassis: yes, but I think you have a double quote that should not be there Nov 11 15:59:07 acassis: the one after the '\' character Nov 11 15:59:27 hum, really I have just one " at begin and " at end SRC_URI Nov 11 15:59:48 I will append one Nov 11 15:59:52 acassis: it's ok then Nov 11 16:00:11 no, just two " is ok Nov 11 16:01:36 cyrilRomain: this is from gpe-go: SRC_URI = "file://fix-make.patch;patch=1" Nov 11 16:01:59 my SRC_URI is just like it Nov 11 16:03:23 acassis: the SRC_URI you paste at 17:47 have three " (not two). That's why I suggested you to check you should only have two " in the SRC_URI Nov 11 16:03:33 cyrilRomain: you are right, I place 3 " Nov 11 16:05:30 hi :> when I am cross compiling, does the glibc version has to match the version of the target system? I created and compiled a package, but when I try to run it on the target plattform I get mediatomb: /lib/libpthread.so.0: version `GLIBC_2.3.2' not found (required by mediatomb) Nov 11 16:05:52 my problem is that I have to use a prebuilt toolchain, I guess that complicates things... Nov 11 16:06:03 any ideas? Nov 11 16:10:06 cyrilRomain: thanks bitbake found the patch Nov 11 16:10:23 acassis: you welcome Nov 11 16:11:10 acassis: there is a oe-stylize.py script in org.openembedded.dev/contrib that can maybe help you check your .bb follows the style guildelines Nov 11 16:11:11 but I get other problem the diff is not finding the files Nov 11 16:11:27 hum, good Nov 11 16:12:51 cyrilRomain: inside bitbake/tmp/work/arm-linux/ipatience-0.1-r2 I have 2 directories Nov 11 16:13:05 the dir ipatience is the cvs source Nov 11 16:13:45 eeew Nov 11 16:13:51 why do you store the source in OE? Nov 11 16:13:59 you should only have the .bb and the patch Nov 11 16:14:40 koen: no I don't do it Nov 11 16:14:48 it is on tmp Nov 11 16:15:21 I think I create the patch in the wrong way Nov 11 16:15:40 acassis: forget my last sentence I misread Nov 11 16:15:55 cyrilRomain: it is from download of CVS Nov 11 16:16:06 it is not on package dir Nov 11 16:21:26 First version of the debian package is ready. Please add the following to your /etc/apt/sources.list, test and report bugs. Nov 11 16:21:38 deb http://www.sicherheitsschwankung.de/~jluebbe/debian unstable/ Nov 11 16:56:30 mhm.. has anyone had this problem? "arm-softfloat-linux-gnu/bin/ld: ERROR: /opt/rootfs/lib//libpthread.so.0 uses hardware FP, whereas conftest uses software FP" I set TARGET_FPU to "hard", where is conftest taking the soft values from? Nov 11 16:58:21 cyrilRomain: whereis the right place to create the patch, the references is wrong Nov 11 16:58:24 Jin^eLD: probably the distribution you choosed Nov 11 16:58:42 I created one myself... Nov 11 16:59:54 I set TARGET_FPU to hard there Nov 11 17:00:40 Jin^eLD: hmm .. no idea then :-/ maybe do you use amsdelta or dht-walnut machine ? Nov 11 17:01:21 acassis: you can still tweak manually the patch you created to correct the path Nov 11 17:01:31 acassis: looking at the other patch available in OE Nov 11 17:01:53 acassis: but I recommend to learn how to use the patch command, or better how to use the quilt patch management tool Nov 11 17:02:30 cyrilRomain: what is amsdelta or "dht-walnut"? Nov 11 17:03:45 cyrilRomain: well, the story goes like that: I cross compiled my package that I wanted to port, turned out the dynamically linked version could not run because the glibc version was different, so I figured I would add the /lib path of the target root image (with prebuilt stuff) to my LDFALGS (-L/...), now when i run configure I can not pass the pthread check anymore Nov 11 17:03:50 because of this FP problem Nov 11 17:04:17 so I edited the configuration and set the TARGET_FPU to "hard", actually I was surprised this thing had a hard fpu Nov 11 17:04:31 but since the prebuilt libs were made that way I guess it's there... Nov 11 17:04:51 but nonetheless I get this error, eventhouth the target fpu is set to hard now. kind of lost at the moment... Nov 11 17:05:02 cyrilRomain: I do it, I think I can figure out it now Nov 11 17:08:18 Jin^eLD: well. I don't know what conftest is. I assume it is the nvidia related package present in OE, right ? Nov 11 17:08:39 Jin^eLD: I think you have to rebuild the package previously built (with "soft" presumly) Nov 11 17:09:21 nono, conftest is something that configure uses to figure out things, it is something used by autotools Nov 11 17:10:07 Jin^eLD: ok. But I'm afraid you have to rebuild all with TARGET_FPU="hard" Nov 11 17:10:32 hmm ok, I will try that Nov 11 17:10:37 Jin^eLD: I'm not sure changing target make bitbake rebuilt them automatically. Maybe do you have to clean first Nov 11 17:10:47 I think I need to, yes Nov 11 17:12:55 ha! I will probably have to rebuild the autools packages.. Nov 11 17:13:00 that was a good hint, thanks Nov 11 17:13:25 Jin^eLD: yes. and maybe all the others packages you built :-/ Nov 11 17:14:28 how do I actually figure out if the device has a hardware FPU at all? is there any hint in /proc/cpuinfo or something? Nov 11 17:14:45 does this say anything to you: Features: swp half thumb fastmult edsp java Nov 11 17:14:55 it's an ARM9 btw Nov 11 17:15:14 Jin^eLD: no this does not say anything to me Nov 11 17:16:09 I don't know how to determine if a device has a hardware FPU or not Nov 11 17:17:21 I just know that with "soft", applications run faster on Zaurus (armv5te) Nov 11 17:17:43 I see.. Nov 11 17:18:05 well, I have some prebuilt libs and stuff, so I guess I will have to go with what they have been compiled with Nov 11 17:19:34 Jin^eLD: I'm afraid yes Nov 11 17:19:36 Jin^eLD: Just as a comparison.. AMD x86 on desktop: Nov 11 17:19:40 flags : fpu ..... Nov 11 17:21:15 Jin^eLD: on Intel pentium-m x86 too: /proc/cpuinfo gives 'flags : fpu' Nov 11 17:21:18 NAbyss: well, I was not sure what the fastmult stuff is all about; and the second thing that confused me - if the prebuilt libpthread stuff is compiled with hardware FP... then it must have some sort of support for it? Nov 11 17:21:27 yeah, I know about the x86.. Nov 11 17:22:01 hi Nov 11 17:22:09 Jin^eLD: I'd be /guessing/ your CPU doesn't have native FPU.. most ARM cores don't, AFAIK Nov 11 17:23:12 NAbyss: any idea why this prebuild libpthread for it is using hard FP then? or how it is possible at all? Nov 11 17:23:30 Jin^eLD: hardware FP OPs are caught by a kernel trap Nov 11 17:23:38 Jin^eLD: Which is slow Nov 11 17:23:56 Jin^eLD: Which is why you saw soft FP is faster than native FP Nov 11 17:24:19 IIRC, each FP op results in two context switches and an interrupt handler.. and context switches are a killer on performance Nov 11 17:24:30 i see.. hmm. do I have to stay consistent? meaning: I can not link vs. this hardfp libpthread library but use soft with other libs that are not prebuilt ? Nov 11 17:25:04 Jin^eLD: I'm not entirely sure TBH.. all I know is there are some issues with mixing various soft/hard FP libraries under some ABIs Nov 11 17:25:13 Jin^eLD: You'd be best off asking one of the other devs here Nov 11 17:25:13 I see Nov 11 17:25:20 k Nov 11 17:25:22 hmm Nov 11 17:25:35 well, thanks, at least I now have more understanding about what is going on Nov 11 17:25:42 gotta leave now, Amon Amarth concert tonight ;) Nov 11 17:25:48 (anyone feel free to chime in if I'm wrong, but..) IIRC, the new EABI allows you to mix soft and hard floats, however old ones caused problems Nov 11 17:25:58 Seeyas Nov 11 17:26:00 I'll drop by tomorrow Nov 11 17:26:02 thanks again Nov 11 17:26:03 bye Nov 11 17:26:05 np Nov 11 17:26:13 NAbyss: that's it AFAIK Nov 11 17:30:20 cyrilRomain: my it is just like the gpe-go, nothing wrong, but it don't found the file Nov 11 17:30:56 the header is just like it: diff -urNd ../ipatience-0.1-r2/ipatience/table.cc ipatience-0.1/table.cc Nov 11 17:32:01 where ipatience-0.1-r2/ipatience is the module downloaded from CVS and ipatience-0.1 is my modifications Nov 11 17:35:15 I think the problem is because the cvs download file to ipatience-0.1-r2/ipatience and the bitbake create my package on ipatience-0.1-r2/ipatience-0.1 Nov 11 17:35:39 the 2 directories don't matchs Nov 11 17:37:16 I am goind out now Nov 11 17:38:01 I will find a way to solve it after Nov 11 17:38:06 acassis: indeed it could be that Nov 11 17:38:22 acassis: what contains your ipatience-0.1-r2/ipatience-0.1 ? Nov 11 17:38:22 thank you very much for your help Nov 11 17:38:38 just the directory patches Nov 11 17:38:46 so it is that Nov 11 17:38:59 the patches directory should be in ipatience-0.1-r2/ipatience Nov 11 17:39:19 at least you should have only one directory, whatever its name Nov 11 17:39:33 but the default to bitbake is create ipatience-0.1 Nov 11 17:40:54 well, I need go out now Nov 11 17:41:00 thank you very much Nov 11 17:41:05 acassis: you'll then problably have to define the S = "${WORKDIR}/path/to/your/src" variable Nov 11 17:41:15 acassis: cu Nov 11 17:41:23 cu2 Nov 11 17:56:04 ibot: wb Nov 11 17:56:13 It's great to be back! Nov 11 17:57:16 * chouimat send a nice heater to mickey|cold Nov 11 17:58:07 thanks. i catched a cold by a climate that was too enthusiastic Nov 11 17:58:13 or how's that cooling machine called? Nov 11 17:58:18 a/c Nov 11 17:58:20 right Nov 11 17:58:21 ;) Nov 11 17:58:41 mickey|cold: :) Nov 11 17:59:23 guess i better lay down a bit Nov 11 17:59:25 see you guys Nov 11 18:02:04 03mickeyl 07org.oe.dev * rd2f9a673... 10/ (1 contrib/python/generate-manifest.py): contrib/python/generate-manifest.py: python-pickle depends on python-io Nov 11 18:02:10 03mickeyl 07org.oe.dev * r114688a6... 10/ (3 files in 2 dirs): python: python-pickle needs python-io. bump due to updated manifest. Nov 11 18:02:16 03mickeyl 07org.oe.dev * ra83b40db... 10/ (154 files in 7 dirs): Nov 11 18:02:16 gtk+: remove a couple of obsolete versions, major stable versions remain Nov 11 18:02:16 gtk+-2.6.10: add menu styling patch based on gtk+ 2.7 Nov 11 18:11:53 * RP returns for a bit Nov 11 18:16:21 I am getting a lot of errors saying eol while scanning single-quoted string Nov 11 18:16:43 And next to that I have a problem with the depencies Nov 11 18:16:53 Any idea what I can do ? Nov 11 18:16:54 gerwinin: Sounds like a quoting problem in some common file Nov 11 18:17:07 RP local.conf ? Nov 11 18:17:22 gerwinin: Could be... Nov 11 18:18:45 or an ancient bitbake Nov 11 18:27:06 Koen: 1.6.3 Nov 11 18:29:00 that doesn't sound to ancient Nov 11 18:29:44 Koen I am compiling on freebsd Nov 11 18:30:00 never tried that Nov 11 18:30:06 only linux and darwin Nov 11 18:30:39 Well it seems bitbake is working but it gives me errors with depencies and ipkg native Nov 11 18:40:35 gerwinin: Can you pastebin the errors? Nov 11 18:41:36 RP ok Nov 11 18:43:45 http://pastebin.com/822012 Nov 11 18:46:22 gerwinin: Did you set BBMASK? Nov 11 18:46:34 gerwinin: check you don't have any bitbake.conf in your build/conf directory Nov 11 18:48:22 03rpurdie * r639 10bitbake/ (MANIFEST lib/bb/fetch/__init__.py lib/bb/fetch/ssh.py): Add ssh fetcher written by kergoth (from poky) Nov 11 18:51:22 this wifi sucks Nov 11 18:51:45 21:09 < RP> gerwinin: Did you set BBMASK? Nov 11 18:51:45 21:09 < cyrilRomain|bbl> gerwinin: check you don't have any bitbake.conf in your build/conf directory Nov 11 18:52:09 Hi all ... my OE can't fetch tslib via CVS (looks like pubcvs.arm.linux.org.uk is dead). Anyone knows if this is a temprary problem? Nov 11 18:52:46 Is there a way to specify a failback in .bb if the primary source can't be reached? Nov 11 18:52:55 pgfeller: tslib moved to berlios Nov 11 18:53:18 pgfeller: You can specify mirrors Nov 11 18:53:36 Cyril should I set it in local.conf ? Nov 11 18:53:52 RP: do I specify the mirrors in the .bb itself, - or in the .conf? Nov 11 18:54:06 pgfeller: read the botton of classes/base.bbclass Nov 11 18:54:37 RP: thanks - ´ll do. I´ll also check if .dev is updated already & submit a bug report for the .oz354x branch Nov 11 18:54:59 pgfeller: .dev is fixed, its definetely a oz problem Nov 11 18:55:01 gerwinin: ?? .. just check you don't have a bitbake.conf file in your oe/build/conf directory Nov 11 18:55:49 * cyrilRomain had strange errors one day with an old bitbake.conf file in his build directory.... Nov 11 18:56:33 I used to have it but I deleted it Nov 11 18:57:13 gerwinin2: or if you need for some reason a bitbake.conf in your build directory, update it from the one in org.openembedded.dev/ Nov 11 18:57:58 gerwinin2: Do you have an classes/opie.bbclass in there? Nov 11 18:58:26 gerwinin2: Somehow it looks like OPIE_CVS_PV isn't getting set Nov 11 19:01:53 03pfalcon 07org.oe.dev * r2b62cff8... 10/ (1 MAINTAINERS): Nov 11 19:01:53 MAINTAINERS: Add myself as opie packages mainatiner. Nov 11 19:01:53 * Somehow mail archive didn't capture Michael Lauer's approval of Nov 11 19:01:53 my maintainership proposal, but here's my response quoting it: Nov 11 19:01:53 http://handhelds.org/hypermail/opie-devel/30/3097.html Nov 11 19:02:11 stupid typos %) Nov 11 19:02:34 RP it says OPIE_CVS_PV = "1.2.2+cv-$(srcdate)" Nov 11 19:02:38 RP it says OPIE_CVS_PV = "1.2.2+cvs-$(srcdate)" Nov 11 19:04:44 I now remember the error I had: GPE_SVN was not set because my old build/conf/bitbake.conf was read prior to the org.openembedded.dev/conf/bitbake.conf. So similar error with a missing OPIE_CVS_PV would not strike me Nov 11 19:04:53 Greetings to all ! Nov 11 19:19:18 "So no one can ever again complain parsing is slow. If it is slow then lex and lemon must suck, but they don't suck." Nov 11 19:19:27 ~hail zecke Nov 11 19:19:29 * ibot bows down to zecke and chants, "I'M NOT WORTHY!!" Nov 11 19:24:07 I'd still guess it will remain slow, though not that much slow. The issue is not speed of parsing, but amount of stuff to parse. So, the solution would be smart, incremental reparsing, not the fast parser. Though, we'll see... Nov 11 19:24:36 howover, why I'm not subscribed to bb commits? %) Nov 11 19:24:41 psokolovsky_: http://planet.linuxtogo.org/ Nov 11 19:25:31 "can parse almost all bb, inc, conf, bbclass files in about two seconds" Nov 11 19:27:17 what?! but that's 100+ times speedup! Nov 11 19:27:36 it takes ~6min to parse OE.dev tree here Nov 11 19:28:12 the link I gave you has more info Nov 11 19:28:18 two seconds ... 8-O Nov 11 19:28:22 `sweet Nov 11 19:29:00 koen, yeah, thanks. will be waiting for ann when it is ready for test ;-) Nov 11 19:30:37 psokolovsky_: zecke's blog posts are announcements :) Nov 11 19:44:50 03mwester 07org.oe.dev * r08f49885... 10/ (3 files in 3 dirs): Unslung: Slingbox: df: fixed handling of root pseudo-devices for Unslung Nov 11 19:44:55 03mwester 07org.oe.dev * r2514e901... 10/ (1 packages/nslu2-binary-only/unslung-rootfs/unsling): Unslung: rootfs: unsling: skip the lost+found directory Nov 11 19:46:27 Greetings to all Nov 11 19:46:39 hey Ifaistos Nov 11 19:47:19 koen : Hi ! Back after 3 weeks in the US and 1 week traveling in Europe Nov 11 19:47:28 Ifaistos: cool Nov 11 19:47:55 koen : Did not make it to the OE meeting though :( Nov 11 19:48:18 koen : next time Nov 11 19:51:01 I had a discussion while in the US regarding the licensing status of the .bb files. What license are the bb under ? GPL ? nothing is mentioned in te file itself Nov 11 19:51:14 MIT Nov 11 19:51:38 Ifaistos: http://www.openembedded.org/repo/org.openembedded.dev/COPYING.MIT Nov 11 19:55:07 koen : Thanks for clearing this. I new about MIT but was not sure. Shouldn't that be mentioned in all .bb files ? Nov 11 19:56:55 Ifaistos: no, the toplevel COPYING.MIT is a blanket license Nov 11 20:02:35 Koen can you help me a bit out with a problem I have Nov 11 20:03:00 gerwinin2: I can try Nov 11 20:03:10 Koen: http://pastebin.com/822012 Nov 11 20:03:29 Koen: I am using oe on a amd64 on freebsd Nov 11 20:06:30 gerwinin2: you'd have to ask zecke, he's the bsd hippy over here Nov 11 20:09:03 ah but zecke is not here Nov 11 20:09:11 I will ask him tomorrow Nov 11 20:09:15 gerwinin2, koen: maybe because there is 'require ${PN}.inc' in libopie2_cvs.bb ? Nov 11 20:09:42 PN is determined from the .bb filename, right ? Nov 11 20:09:46 Cyril I have some other error as well it doesn't compile anything Nov 11 20:09:57 it looks like it cannot find the path or something Nov 11 20:10:33 gerwinin2: as RP said, it seems that OPIE_CVS_PV is not defined, which means that opie.bbclass is not read Nov 11 20:11:37 * cyrilRomain just trying to figure out.... Nov 11 20:13:02 cyril let me check this Nov 11 20:14:00 gerwinin2: well no ..libopie2_cvs.bb looks fine (at least through my noob eyes) Nov 11 20:16:43 gerwinin2: well ... I'm out of clue then :-/ Nov 11 20:18:45 gerwinin2: ah you didn't answer RP question: did you set a BBMASK in your local.conf ? Nov 11 20:22:16 no Nov 11 20:22:28 cyril what should I put there ? Nov 11 20:22:37 nothing special Nov 11 20:22:49 Okay with me it says nothing Nov 11 20:23:03 you could have opie.bbclass or something like that that prevent opie.bbclass to be parsed Nov 11 20:24:11 I'm out of ideas, sorry. Nov 11 20:25:01 me too :( Nov 11 20:25:57 gerwinin2: could you try bitbake trunk? Nov 11 20:26:02 (1.7.3) Nov 11 20:28:14 Koen you think that will be better ? Nov 11 20:28:23 it could be Nov 11 20:28:28 trunk seems to work on osx Nov 11 21:00:04 koen: I am testing the ixp4xx/angstrom build. Inheriting rootfs_ipk fails for me, did you have that enabled? Nov 11 21:08:23 The trunk versio also didn't work Nov 11 21:08:33 I will install it on my fedora box :) Nov 11 21:19:21 likewise: yes Nov 11 21:20:29 koen: I can build the image, working around the qemu-native x86_64 using ENABLE_BINARY_LOCALE_GENERATION = "0" Nov 11 21:22:11 koen: rootfs_ipk fails for me, could you spend a minute seeing why? http://pastebin.com/822115 Nov 11 21:23:21 hmmm Nov 11 21:23:28 could you update your bitbake and oe? Nov 11 21:24:21 bitbake 1.7.2, oe of yesterday Nov 11 21:24:59 koen: will try it with bitbake trunk, is that what you used? Nov 11 21:25:20 yes Nov 11 21:35:38 likewise: http://rafb.net/paste/results/rcVvzB56.html Nov 11 21:35:57 likewise: do you have any extra stuff in your local.conf? Nov 11 21:37:21 koen: I don't you yours, but I can show mine :-) Nov 11 21:38:54 koen: http://pastebin.com/822133 Nov 11 21:39:08 koen: I just moved INHERIT up, it was way down before. Nov 11 21:39:42 likewise: why do you inherit it in local.conf? Nov 11 21:40:01 it's already in the angstrom configs Nov 11 21:42:15 koen: probably a left-over from some older local.conf... So the error is I might be double-inheriting it, and/or at the wrong place. Nov 11 21:44:49 03zecke123 * r640 10bitbake-parser/standalone/ (11 files): Nov 11 21:44:49 Add a small, compilable utility to parse bitbake files. Nov 11 21:44:49 ./bitbake-helper "/space/the/new/stuff/packages/*/*.bb" "/space/packages/*/*.inc" Nov 11 21:44:49 This can be used by brave guys to fixup the dev branch to be parsable Nov 11 21:44:49 (mainly the missing quotes) Nov 11 21:45:56 hey zecke Nov 11 21:46:08 03zecke123 * r641 10bitbake-parser/standalone/lemon: No binaries in svn Nov 11 21:46:22 hmm 173 packages to fix :( Nov 11 21:46:26 zecke: hi there Nov 11 21:46:41 zecke: send a list to the oe-ml Nov 11 21:47:25 hey zecke Nov 11 21:47:47 gerwinin2: hey, how are you? Nov 11 21:47:50 Zecke I am trying to install oe in freebsd Nov 11 21:47:52 I am fine Nov 11 21:48:01 But no luck so far Nov 11 21:48:20 Zecke: Is it better to change to fedora ou think ? Nov 11 21:48:32 gerwinin2: depends on how much time you have Nov 11 21:49:04 Zecke: Which of the two is more work ? Nov 11 21:49:22 gerwinin2: switching to fedora because rpm based distris are disgusting Nov 11 21:49:34 gerwinin2: but there is a lot work to do on NON GNU systems Nov 11 21:50:20 gerwinin2: what's your aim? build on, or build towards BSD systems? Nov 11 21:50:29 Zecke: build on bsd system Nov 11 21:50:44 Sorry likewise : build on bsd systems Nov 11 21:51:03 gerwinin2: If you have time to spend a month on it, stay with BSD Nov 11 21:51:11 gerwinin2: this will help us to fix more crappy build systems Nov 11 21:51:43 Zecke: I will keep my bsd system on and in the meanwhile continue to build with fedora Nov 11 21:52:04 Zecke : I had a shit load of hardware that I want to port to oe Nov 11 21:52:49 Zecke: can I drop some of this hardware to you @t-dose so you can distribute it further Nov 11 21:53:11 hehe Nov 11 21:53:13 zecke: still coming to t-dose? Nov 11 21:53:45 Zecke: via send me some stuff (for which I had to pay the vat :( ) Nov 11 21:53:57 :( Nov 11 21:54:14 koen: i think i will, if no test conflict Nov 11 21:54:18 Zecke: I got some kind of weird arm board from a taiwanese guy Nov 11 21:54:45 Zecke: A thin client Nov 11 21:55:00 Zecke: some stuff from gigabyte Nov 11 21:55:07 Zecke: some stuff from jetway Nov 11 21:55:12 that is a shit load of stuff... Nov 11 21:55:21 * zecke searches some food Nov 11 21:55:29 * koen hears 'thin client' and 'arm' Nov 11 21:55:34 Yeah I went to Taiwan and just asked them Nov 11 21:55:54 I will concentrate on the epias Nov 11 21:56:25 Koen , I am not sure what the thin client is for architecture Nov 11 21:56:43 The arm thing is comming from cirrius logic and is running now netbsd Nov 11 21:57:14 I moved my cirrus arm board to my new desk today :) Nov 11 21:57:25 hopefully power and ethernet will follow Nov 11 21:57:30 :) Nov 11 21:57:46 likewise: new images at http://www.openembedded.org/~koen/ixp4xx/ Nov 11 21:57:50 I found out that we need to move to asia Nov 11 21:58:01 yeah, much better food over there Nov 11 21:58:21 Yeah and chicks that understand us :) Nov 11 21:59:04 "missing quotes on assignments PATCH = "f00" vs. PATCH = "f00"" Nov 11 21:59:14 * koen tries to find the distinction Nov 11 22:00:26 koen: I have an AFT error here, sorry. Nov 11 22:00:28 hehe Nov 11 22:02:29 gerwinin2: you walked into their offices and asked for h/w? Nov 11 22:04:58 zecke: ParseError: /space/hacking/embedded/oe/org.openembedded.dev/packages/gnome/gnome-doc-utils_0.7.1.bb:13 Nov 11 22:05:58 zecke: what's the error there? Nov 11 22:07:52 koen: '}' need to be at the start of line Nov 11 22:10:09 koen: oe-stylize.py should remove such extra space/tab :) Nov 11 22:10:22 03rwhitby 07org.nslu2-linux.dev * rd685ebf6... 10/ (13 files in 4 dirs): Initial import from corrupted database - all history lost. Nov 11 22:10:26 huh? Nov 11 22:10:43 NAiL: hehe Nov 11 22:10:43 again? Nov 11 22:10:44 * koen stabs ciabot_monotone.py Nov 11 22:10:48 hehe Nov 11 22:10:48 Isn't that like the third time? Nov 11 22:12:19 NAiL: yes, ciabot_monotone.py loves to replay history Nov 11 22:12:49 gah Nov 11 22:16:34 * koen gets some lieslaarzen and braces for a flood Nov 11 22:16:59 03freyther 07org.oe.dev * r2bc9f47d... 10/ (1 conf/bitbake.conf): conf/bitbake.conf: Fix quotes for the more strict parser Nov 11 22:17:04 03freyther 07org.oe.dev * r6f1ed32f... 10/ (21 files in 5 dirs): Add quotes for assignments to make the new parser happy Nov 11 22:17:12 03koen 07org.oe.dev * r0c748e1d... 10/ (3 files in 3 dirs): Nov 11 22:17:12 faad2: add a gcc4 fix Nov 11 22:17:12 * as a side note: thank $deity that we don't use that line-ending mangling bitkeeper anymore, yay! Nov 11 22:17:14 03koen 07org.oe.dev * r8a47d0d0... 10/ (5 files in 2 dirs): angstrom: fix some parse errors to make Holger happy Nov 11 22:17:20 03koen 07org.oe.dev * rff140283... 10/ (1 packages/eds/eds-dbus_svn.bb): eds-dbus: fix parse error Nov 11 22:17:51 03koen 07org.oe.dev * rc370f7ef... 10/ (1 packages/gnome/gnome-doc-utils_0.7.1.bb): gnome-doc-utils: fix another parser error to make Holger even happier Nov 11 22:18:12 koen: I would have been more happy if you ordered a pizza for me :) Nov 11 22:19:17 I could buy you a frikandel speciaal at t-dose Nov 11 22:19:30 but first I will go to sleep Nov 11 22:19:33 'night all Nov 11 22:19:43 koen: I will eat all frikandel at the train station+ Nov 11 22:23:31 03zecke123 * r642 10bitbake-parser/ast.py: AST: __str__ for one class :) Nov 11 22:24:08 03zecke123 * r643 10bitbake-parser/parse-all.py: Parse everything, add it to the list of ASTs and then sleep to analyze memory usage Nov 11 22:26:35 hi all ... have problem to build glibc-intermediate: configure: error: Need linker with .init_array/.fini_array support. Nov 11 22:27:05 pgfeller: and you have one Nov 11 22:27:18 pgfeller: but glibc uses readelf to look at your binary Nov 11 22:27:27 pgfeller: check a patch with 'zecke' for glibc Nov 11 22:28:31 zecke: do I understand right: it is broken in .dev & you have a patch? Nov 11 22:28:45 pgfeller: no, it is broken glibc, and I patched one version Nov 11 22:28:56 pgfeller: but as usual, take a look at the config.log Nov 11 22:29:04 pgfeller: draw your conclusion why the patch fails Nov 11 22:29:42 zecke: I have PREFERRED_VERSION_glibc = "2.3.2" in my config ... which one did you patch/should work? Nov 11 22:30:08 23:49 < zecke> pgfeller: but as usual, take a look at the config.log Nov 11 22:30:08 23:49 < zecke> pgfeller: draw your conclusion why the patch fails Nov 11 22:30:21 but again, take a look at the config.log Nov 11 22:30:26 see why the test failed Nov 11 22:30:45 ~pastebin Nov 11 22:30:50 somebody said pastebin was a place to paste your stuff without flooding the channel - try http://pastebin.ca, or http://channels.debian.net/paste Nov 11 22:31:49 @ http://pastebin.ca/246153 I posted the relevant section - any idea? Nov 11 22:32:20 pgfeller: 23:50 < zecke> 23:49 < zecke> pgfeller: but as usual, take a look at the config.log Nov 11 22:32:23 23:50 < zecke> 23:49 < zecke> pgfeller: draw your conclusion why the patch fails Nov 11 22:32:26 23:50 < zecke> but again, take a look at the config.log Nov 11 22:32:28 I have time, I can copy and paste Nov 11 22:32:30 23:50 < zecke> see why the test failed Nov 11 22:32:34 take a look at the config.log Nov 11 22:34:25 zecke: sorry - missunderstood Nov 11 22:34:36 pgfeller: you are welcome, I have plenty of time now Nov 11 22:35:46 zecke: could ./tmp/org.openembedded.dev-collie/work/arm-linux/glibc-intermediate-2.5-r2/build-arm-linux/config.log be the file I should check? Nov 11 22:36:09 pgfeller: if that is the version that was built Nov 11 22:36:27 pgfeller: you might should set some more PREFERRED_VERSIONS as well... Nov 11 22:37:06 zecke: yep - it is; is there somewhere a list of PREFERRED_VERSIONS recommended for "beginners"? Nov 11 22:37:30 pgfeller: not that I know of, beginners normally don't maintain distributions Nov 11 22:37:42 pgfeller: okay but back to the config.log, try to find the test that was executed Nov 11 22:37:57 zecke: I only try to update some pacakges & to use the devshell ... Nov 11 22:38:32 zecke: is oe only recommended to maintain/build distros? Nov 11 22:38:47 pgfeller: you know that familiar is using git and forked from OE,BitBake Nov 11 22:38:58 pgfeller: no, but you build packages targeting a distribution Nov 11 22:39:35 zecke: yep ... currently I try to build something for OZ354x - & found that tslib is broken there (moved to berlios) Nov 11 22:39:53 zecke: so I thought lets switch to .dev & check it there .... Nov 11 22:39:59 pgfeller: it is probably wise to use the oz354x branch then Nov 11 22:40:15 pgfeller: but back to your error, check the config.log Nov 11 22:40:24 pgfeller: and you probably see that an readelf call failed Nov 11 22:41:20 zecke: ... see - this is to high for me; I'll copy the tslib.bb from .dev to .oz354x - maybe that's enought already to help with my problem. thanks anyhow :-) Nov 11 22:41:58 pgfeller: why is reading a text file too high? Nov 11 22:42:06 hehe, sorry then and good luck Nov 11 22:42:26 zecke: the reading is ok - the draw your conclusion is the problematic part ;-) Nov 11 22:42:42 pgfeller: do me a favor and paste the file Nov 11 22:42:51 pgfeller: which host system do you have? Nov 11 22:42:53 zecke: ok ... Nov 11 22:43:16 pgfeller: I'm fighting with Mr. glibc about this issue Nov 11 22:45:00 zecke: the file is too big for pastebin - is there an alternative? Nov 11 22:45:33 pgfeller: yes, grep for init_array Nov 11 22:45:39 pgfeller: should not match more than five times Nov 11 22:46:33 zecke: 2 matches: Nov 11 22:46:33 configure:5580: checking for .preinit_array/.init_array/.fini_array support Nov 11 22:46:33 configure:5612: error: Need linker with .init_array/.fini_array support. Nov 11 22:46:48 pgfeller: and now paste the context of the methods :) Nov 11 22:46:55 +- 20 lines Nov 11 22:49:47 zecke: http://pastebin.ca/246175 Nov 11 22:50:08 /stuff/tmp/org.openembedded.dev-collie/cross/lib/gcc/arm-linux/4.1.1/../../../../arm-linux/bin/ld: cannot represent machine `arm:collie' Nov 11 22:50:22 koen was the first to have this issue :) Nov 11 22:50:35 unset MACHINE when building binutils, or don't set it at all Nov 11 22:51:48 zecke: thanks - will try that Nov 11 22:52:21 zecke: will remove MACHINE from local.conf & use TARGET_... instead Nov 11 22:52:31 pgfeller: no that is okay Nov 11 22:52:36 but don't do export MACHINE=foo Nov 11 22:52:41 in your shell Nov 11 22:52:57 zecke: I see - ok. Nov 11 23:02:41 03zecke123 * r644 10bitbake-parser/ (5 files): Nov 11 23:02:41 Handle the ParseError. If we have an error propagate it through lex_t. Nov 11 23:02:41 Check the error when returning to python and raise an exception Nov 11 23:04:18 to which mailing list should I suscribe to get bugs reports ? Nov 11 23:04:32 cyrilRomain: openembedded-bugs? Nov 11 23:04:54 zecke: thanks :) Nov 11 23:05:01 oh no wait Nov 11 23:05:14 I think I had the brilliant idea to call it issues *confused* Nov 11 23:06:50 cyrilRomain: http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-issues Nov 11 23:06:58 hmm hard to navigate to... Nov 11 23:07:16 oh and a redirect to ltg Nov 11 23:07:46 zecke: thx Nov 11 23:10:14 zecke: is your 'comming soon' parser part of the bitbake-ng development ? Nov 11 23:10:39 no Nov 11 23:10:48 heh Nov 11 23:11:06 the new parser, GNU Hurd and GPL Qtopia compete Nov 11 23:11:11 I hope to release sooner(tm) Nov 11 23:11:33 the recent e-mail by RP made clear that bitbake-ng is nearly abandoned Nov 11 23:12:19 JustinP: ok thanks (I didn't read all the last mail :-/) Nov 11 23:13:07 actually we have a bitbake-ng with 1.7 :) Nov 11 23:13:35 zecke: I still have a patch to make bitbake accept regular expression in the -I argument. Nov 11 23:13:48 ah :) Nov 11 23:14:09 I have a new dot output ready Nov 11 23:14:19 this time for the parse tree Nov 11 23:14:24 zecke: :) Nov 11 23:14:25 This is OE == Picture of the Syntax Trees of BitBake Nov 11 23:14:32 of OE... Nov 11 23:14:34 *tired* Nov 11 23:15:14 mickey|bed: ping! Nov 11 23:15:19 mickey|bed: wake up! Nov 11 23:15:44 mickey|bed: I need python counseling, immediately Nov 11 23:17:22 zecke: 20:22 < mickey|cold> guess i better lay down a bit Nov 11 23:17:27 :-/ Nov 11 23:17:54 4 hours, that is plenty of time :) Nov 11 23:25:25 zecke: 4 hours, that's about the amount of time you would need to eat all frikandels at Eindhoven Station. Nov 11 23:25:47 hehe Nov 11 23:26:00 in contrast to sleeping I would not feel better afterwards Nov 11 23:26:22 zecke: I'll make sure all frikandel windows are stuffed when you arrive ;-) Nov 11 23:26:28 hehe Nov 11 23:26:43 Well, I'm off to bed now. Otherwise I will be leon|cold as well. Nov 11 23:26:57 * zecke as well Nov 11 23:27:02 sleep tight. Nov 11 23:27:02 cya Nov 11 23:27:04 cya Nov 11 23:28:15 nice meetin' ya likewise Nov 11 23:28:48 HopsNBarley: likewise. See you around here. Nov 11 23:50:09 03rpurdie * r645 10bitbake/lib/bb/fetch/ (8 files): fetchers: Refactor interating through urls into the fetcher core (preparing for common md5/tarball stash code) **** ENDING LOGGING AT Sun Nov 12 02:59:57 2006