**** BEGIN LOGGING AT Fri Jul 28 02:59:56 2006 Jul 28 06:25:36 good morning all Jul 28 06:25:58 hi, I have set PREFERED_VERSION_glibc to "2.3.5", but bitbake always choose the glibc_2.3.5+cvs20050607 Jul 28 06:27:46 I have to delete the 2.3.5+cvs20050607 Jul 28 06:28:01 any idea? Jul 28 06:31:03 kewkew Jul 28 06:43:54 O_Neil: does 2.3.5 exist in OE? Jul 28 06:44:02 yes Jul 28 06:44:39 and if there is 2.3.2+cvs, it will use 2.3.2+cvs Jul 28 06:46:58 hi koen Jul 28 06:47:31 koen: will your new kernel work with regular builds or just the angstrom test images? Jul 28 06:47:43 koen: there is a 2.3.5 but it always chooses the cvs versions. I had the same problem Jul 28 06:48:06 xserver-kdrive doesn't seem to want to finish building Jul 28 06:48:20 it's dieing on 'Making all in input' Jul 28 06:48:24 hvontres|home: it has 'OABI compat' enabled, so it shoud work with regular images Jul 28 06:48:37 hvontres|home: provided that to put the right kernel modules in Jul 28 06:48:45 with undeclared so and so's in dmxevents.c: In function `dmxEnqueueExtEvent' Jul 28 06:49:10 (that's xserver-kdrive-X11R7.1-1.1.0-r0) Jul 28 06:49:28 O_Neil: there is no glibc 2.3.5, only a 2.3.5+cvs, so it naturally uses the +cvs version, since the other doesn't exist Jul 28 06:49:30 koen: ok. I might try it out later then...:) Jul 28 06:50:00 time to turn in :) Jul 28 06:50:18 koen: I added the glibc_2.3.5.bb to glibc Jul 28 06:50:48 anyone seen or solved this one yet? Jul 28 06:51:40 koen: O_Neil: sorry, i meant glibc-2.3.3, -2.3.2 and -initial-2.3.2 with and without +cvs... Jul 28 07:10:30 * koen finishes filing bugs Jul 28 07:13:08 morning all Jul 28 07:15:24 hey RP Jul 28 07:15:36 I found some spurious debug info in packages Jul 28 07:16:00 RP: http://bugs.openembedded.org/show_bug.cgi?id=1210 and http://bugs.openembedded.org/show_bug.cgi?id=1211 Jul 28 07:17:41 good morning Jul 28 07:17:47 koen: I think this bug is a sorting issue Jul 28 07:18:15 pH5: yeah, but I haven't come up with a good way to fix it Jul 28 07:18:22 pH5: xtscal works, btw Jul 28 07:18:24 koen: the ${PN} package is populated before ${PN}-dbg and thus takes the *.dbg files. Is this correct? Jul 28 07:18:26 yay! Jul 28 07:18:49 pH5: dbus, however doesn't Jul 28 07:18:59 it quits after an alignment trap Jul 28 07:19:16 strace and gdb should be in the angstrom feeds :) Jul 28 07:19:38 pH5: and udev firmware upload is broken as well, so hx4700_acx doesn't work Jul 28 07:19:56 oh shoot, I forgot to upload gdb 6.4. gdb 6.3 doesn't work with eabi. Jul 28 07:20:20 does Pawel Osiczko still maintain gdb? Jul 28 07:21:31 http://svn.w4py.org/Paste/trunk/paste/util/subprocess24.py looks promising Jul 28 07:21:49 we should think about switching to it instead of our current crappy os.system's and stuff Jul 28 07:21:57 can choose not to use a shell, just directly execv'ing Jul 28 07:22:27 cute Jul 28 07:26:00 03pH5 07org.oe.dev * rf43951bf... 10/ (1 packages/gdb/gdb_6.4.bb): Jul 28 07:26:00 gdb: add 6.4 Jul 28 07:26:00 - supports EABI stack Jul 28 07:26:00 - does Pawel Osiczko (still) maintain this package? Jul 28 07:27:42 03koen 07org.oe.dev * rcf1a0809... 10/ (3 files in 2 dirs): Jul 28 07:27:42 angstrom: fixes Jul 28 07:27:42 * fix typos Jul 28 07:27:42 * ship fonts Jul 28 07:27:42 * RRECOMMEND kernel-modules needed for basic operation Jul 28 07:28:20 pH5: do you know who's bright idea it was to make af-packet modular? Jul 28 07:28:35 dhcp kind of needs it :) Jul 28 07:31:44 koen: could be from the early sdgsystem defconfigs where everything was modular just because. Jul 28 07:31:54 koen: but I don't know for sure. Jul 28 07:32:32 pH5: any idea why firmware upload wouldn't work? Jul 28 07:32:56 pH5: http://www.openembedded.org/viewmtn/revision.psp?id=e7a2f39096ee46a8369ff9e117efe57cfff4f50e Jul 28 07:33:36 koen: I've seen it already. Good idea. Jul 28 07:34:33 Don't know about the firmware upload. I'll have a look. I'm curious about the dbus alignment error anyway :) Jul 28 07:35:57 new gdb is in the feeds Jul 28 07:36:08 hey XorA Jul 28 07:36:15 morning Jul 28 07:36:18 hey koen Jul 28 07:36:18 hi XorA Jul 28 07:36:32 hey pH5 Jul 28 07:37:13 someone who has managed to build xserver-kdrive... did you use preferred-xorg-versions-X11R7.1.inc? Jul 28 07:38:36 tomkp: I use the latest available packages. X11R7.1 might hit the new fontconfig/old libxfont incompatibility. Jul 28 07:39:02 hmm, I did try with the latest ones also Jul 28 07:39:48 hey zecke Jul 28 07:39:54 hey Jul 28 07:41:57 03clarson 07org.oe.dev * rc983506d... 10/ (1 classes/base.bbclass): Jul 28 07:41:57 base.bbclass: bugfix in oe_unpack_file Jul 28 07:41:57 * Fix an issue with file:// urls in SRC_URI which reside in WORKDIR. With Jul 28 07:41:57 this issue fixed, patches within tarballs can be successfully applied to Jul 28 07:41:57 ${S}. Jul 28 07:43:32 WOW Jul 28 07:43:44 zecke: did you full up /home on ewi? Jul 28 07:43:54 koen: I wasn't on ewi for ages Jul 28 07:43:59 kergoth is back :) Jul 28 07:44:04 indeed Jul 28 07:44:06 koen: to trigger the dbg bug, did you build with DEBUG_BUILD="1"? Jul 28 07:44:20 pH5: no Jul 28 07:44:24 pH5: which DBG bug? Jul 28 07:44:30 pH5: .dbg files in the wrong packages? Jul 28 07:44:31 heh, sadly enough, that was my first ever commit/push with monotone :P Jul 28 07:44:32 zecke: #1211 Jul 28 07:44:33 thats how long its been Jul 28 07:44:57 zecke: right. in ${PN} instead of {$PN}-dbg, it seems. Jul 28 07:45:19 pH5: these are due wrong FILES statements Jul 28 07:45:24 pH5: e.g. see openssl.inc Jul 28 07:45:42 pH5: libcrypto is doing FILES_libcrypto = ...so*" Jul 28 07:45:53 pH5: which includes .so.dbg as well -> package bigger Jul 28 07:46:53 zecke: check out http://svn.w4py.org/Paste/trunk/paste/util/subprocess24.py . worth using in bitbake to replace our crappy os.system's and such? Jul 28 07:47:17 zecke: and do you have any opinions on http://svn.o-hand.com/view/poky/trunk/bitbake/lib/bb/fetch/ssh.py?rev=524&view=markup ? Jul 28 07:47:43 (implementation wise... i'm somewhat rusty with bitbake :) Jul 28 07:49:10 * kergoth goes back to the code he was working on Jul 28 07:49:31 kergoth: the avoidance of os.system is welcomed :) Jul 28 07:49:40 kergoth: the subprocess24.py looks huge though Jul 28 07:50:18 zecke: so should we change all packages that touch PACKAGES or FILES_... and add -dbg support or can the code in packages.bbclass be made more clever? Jul 28 07:50:19 kergoth: oh, I had a scp fetcher as well, but yours is looking more polished Jul 28 07:50:37 pH5: the FILES variables of the packages are wrong anyway Jul 28 07:50:45 pH5: using openssl.inc as well Jul 28 07:50:58 pH5: libcrypto contains "libcrypto.so" and "libcrypto.so.dbg" Jul 28 07:51:08 pH5: the one should be in -dev the other in -dbg Jul 28 07:51:24 pH5: with fixing FILES of every bb files we fix -dev and -dbg package sin one go :} Jul 28 07:52:57 mickey|ibiza: stop pretendng your on holiday and checkin the rest of your libetpan stuff :-) Jul 28 07:53:01 kergoth: I didn't know commands, starts pydoc now Jul 28 07:54:20 zecke: so what should ${libdir}/libpng12.so.* be changed to, for example? ${libdir}/libpng12.so.[0-9]* ? Jul 28 07:54:58 pH5: that would still include .dbgs Jul 28 07:55:04 zecke: me neither, just happend to run across it. just a simple wrapper, but its simple and clean and easy Jul 28 07:55:24 pH5: but it would solve the -dev problems Jul 28 07:56:25 pH5: good question (it should work), also check the FILES inside bitbake.conf Jul 28 07:58:38 otoh, for libpng this should work too: http://en.pastebin.ca/103327 Jul 28 07:59:04 pH5: right Jul 28 08:01:00 koen: BTW: if you feel like taking the role of the release dude, leader, direction giver or whatever for bitbake, just inform me :) Jul 28 08:01:43 zecke: I know virtually nothing about bitbake, why do you ask? Jul 28 08:01:44 pH5: I would prefer in limiting the FILES globs Jul 28 08:01:49 oops :} Jul 28 08:01:54 kergoth: BTW: if you feel like taking the role of the release dude, leader, direction giver or whatever for bitbake, just inform me :) Jul 28 08:01:59 :) Jul 28 08:02:01 koen: I'm worse than Nixon Jul 28 08:02:21 "Bomb these suckers, throw nuclear bombs. I don't care I'm drunk" :} Jul 28 08:02:26 zecke: i may.. we'll see. have a lot of ideas for changes :) Jul 28 08:03:29 kergoth: Just inform me, and I won't respond with "screw you, go away you don't exist" Jul 28 08:04:23 why the hell is this saying my preferred version of glibc doesnt exist, despite it existing. /me shakes fist at it, then enables debugging Jul 28 08:04:46 kergoth: hehe Jul 28 08:05:24 kergoth: a) I have a unfinished bitquery tool b) bitbake needs better reporting skills Jul 28 08:05:39 * kergoth is amused by the DEBUG: bbmake: messages Jul 28 08:05:42 kergoth: glibc 2.4? Jul 28 08:05:42 bbmake :) Jul 28 08:05:48 koen: indeed Jul 28 08:06:03 zecke: bitbake needs a lot of things :P Jul 28 08:06:05 kergoth: bitbake will complain, but happily start bulding it Jul 28 08:06:23 mickeyl and I were amazed about it Jul 28 08:06:40 it wasnt so happy here :( Jul 28 08:06:42 * kergoth pokes at it Jul 28 08:08:10 we need a proper indentation plugin for vim for bitbake metadata Jul 28 08:08:15 its -really- starting to annoy me Jul 28 08:16:23 03pH5 07org.oe.dev * r6dd6715d... 10/ (1 packages/libpng/libpng_1.2.8.bb): libpng-1.2.8: split out debug info, fix for #1235 Jul 28 08:16:32 woglinde: ho Jul 28 08:17:32 he zecke Jul 28 08:17:36 you got mail Jul 28 08:17:38 I guess Jul 28 08:17:53 woglinde: "Bewertung von E-Chalk" Jul 28 08:18:04 *puh* Jul 28 08:18:05 nope Jul 28 08:18:09 gamecube session Jul 28 08:18:10 pH5: mount -at fails -> no tmp -> broken firmware helper Jul 28 08:18:15 hi koen Jul 28 08:18:19 hey woglinde Jul 28 08:18:47 how come I can put virtual/imlib2 in DEPENDS, but not in RDEPENDS? Jul 28 08:18:48 koen: okay, the new mtn annotate is faster Jul 28 08:18:53 koen: but it is still slow Jul 28 08:18:57 yeah Jul 28 08:19:01 did you try the patch? Jul 28 08:19:11 currently trying Jul 28 08:19:19 koen: it is at rev 83 of 7.293 Jul 28 08:20:06 hey mallum Jul 28 08:20:16 koen: do you have any clue why *âš¡#☠ mount -at doesn't mount the tmpfs? Jul 28 08:20:25 hey mallum Jul 28 08:20:28 zecke: protoype.annotate.patch? Jul 28 08:20:34 pH5: busybox? Jul 28 08:20:38 koen: yes Jul 28 08:20:39 that's reason enough for me Jul 28 08:21:05 hey koen, zecke Jul 28 08:21:32 mallum: thanks for bringing kergoth back to OE Jul 28 08:21:39 wheeeee Jul 28 08:21:45 internal wlan of the hx4700 works Jul 28 08:21:50 zecke: :) Jul 28 08:22:09 hehe Jul 28 08:22:15 * kergoth has missed bitbake & oe Jul 28 08:23:13 koen: great. now watch that temperature :) Jul 28 08:23:15 hi kergoth Jul 28 08:23:20 hey Jul 28 08:23:52 hm Jul 28 08:24:46 koen: you are fine with using binutils version 2.15.94.0.1? Jul 28 08:25:20 no, angstrom need 2.17 Jul 28 08:25:21 for msp430 this patch http://bugs.openembedded.org/attachment.cgi?id=765&action=view or my patch where I check if the files are avalailabe is needed Jul 28 08:25:36 which version is preffered? Jul 28 08:25:39 koen: hehe, for openzaurus, to fix the bug report Jul 28 08:25:48 zecke: yes, that's ok Jul 28 08:26:34 woglinde: what dou you want to check? Jul 28 08:26:45 woglinde: if [ -f file ]; then mv -f ? Jul 28 08:27:17 RP: your touchscreen udev rule doesn't work Jul 28 08:27:43 zecke right Jul 28 08:29:22 koen: try removing the "14A," part in the modalias expression Jul 28 08:29:43 zecke http://page.mi.fu-berlin.de/~heinold/msp430.diff Jul 28 08:31:09 hm Jul 28 08:31:38 woglinde: attach the patch as well? Jul 28 08:33:13 zecke I will look what has changed meanwhile Jul 28 08:36:51 koen: 526 revs scanned Jul 28 08:38:29 pH5: my next commit is for you :) Jul 28 08:41:07 03koen 07org.oe.dev * rd85a2425... 10/ (1 packages/angstrom/task-angstrom.bb): task-angstrom: ship strace and gdb is DISTRO_TYPE = debug Jul 28 08:45:31 /opt/mvlcee *puke* Jul 28 08:48:40 ugh, yes Jul 28 08:48:44 its terrible Jul 28 08:50:04 is this an acceptable fix for fontconfig-dbg: http://en.pastebin.ca/103452 ? Jul 28 08:51:09 03koen 07org.oe.dev * r151f9633... 10/ (1 packages/angstrom/task-angstrom.bb): task-angstrom: add modutils-initscripts Jul 28 08:51:25 pH5: hmm, yes Jul 28 08:51:49 okay once we have fixed this .dbg issue Jul 28 08:51:49 koen: Which kernel are you using? There were kernel bugs that broke it... Jul 28 08:52:00 I will add -B to our CC line Jul 28 08:52:03 RP: 2.6.16-hh5 on a hx4700 Jul 28 08:57:11 zecke: btw, debian puts all the debug info into a separate folder (/usr/lib/debug) Jul 28 08:57:40 pH5: hmm Jul 28 08:57:56 koen: Can you find out the modalias of the touchscreen device (its in /sys) and check the regexp used in the udev rule matches it? Jul 28 08:58:02 pH5: we would not run into two packages writing the same file? Jul 28 08:58:15 anyone interested in good quality video recordings of a number of OLS talks? Jul 28 08:58:26 pH5, koen: Thanks for creating the -dbg packages bug - I guess we have some cleanup to do :) Jul 28 08:59:08 * zecke continues on his insane.bbclass Jul 28 08:59:24 zecke: they mirror the directory structure under /usr/lib/debug Jul 28 08:59:30 03freyther 07org.oe.dev * r89e38da8... 10/ (1 conf/distro/openzaurus-unstable.conf): Jul 28 08:59:30 conf/distro/openzaurus-unstable.conf: prefer 2.15.x of binutils to fix #1183 Jul 28 08:59:30 To fix bug #1183 we will use binutils 2.15.94.0.1 again. We do this Jul 28 08:59:30 by setting the PREFERRED_VERSION to this version. Jul 28 08:59:37 RP: input:b0000v0000p0000e0000-e0,1,3,kra0,1,18,mlsfw Jul 28 08:59:47 pH5, koen: In some cases where packages add to the standard FILES_${PN}, we'll need to add similar expressions to FILES_${PN}-dbg Jul 28 08:59:51 also, I wonder if we don't want just one -dbg package per .bb recipe instead of one for every generated .ipk with stripped files Jul 28 09:00:09 pH5: We only get one per .bb ;-) Jul 28 09:00:26 (like -dev) Jul 28 09:00:32 okay Jul 28 09:00:35 * zecke tries to remember Jul 28 09:00:47 RP: that's not true, actually Jul 28 09:01:09 RP: if that is policy, fine. then what shall I do about .bbs that do PACKAGES =+ "something-containing-stripped-files " ? Jul 28 09:01:10 ehm Jul 28 09:01:12 ignore me Jul 28 09:01:21 the .dbg ended up in subpackages Jul 28 09:01:47 (then my libpng fix is wrong, as is the fontconfig one) Jul 28 09:01:55 pH5: How does that cuase a problem? Jul 28 09:02:11 RP: they grab away the .dbg files in their folders before ${PN}-dbg can. Jul 28 09:02:50 RP: libpng12 for example takes ${libdir}/libpng12.so.*, including libpng12.so.*.dbg Jul 28 09:03:07 koen: In local.rules, it looks like the 14A isn't needed. I'll need to look up what that meant before I'll happily remove it Jul 28 09:04:36 hi polyonymous Jul 28 09:04:50 pH5: tricky :-/ Jul 28 09:04:56 woglinde: which games do we want to play? Jul 28 09:05:02 zecke zelda Jul 28 09:05:16 if it is playable by two person Jul 28 09:05:18 mom Jul 28 09:05:27 RP: we'd need some kind of "insert libpng12 before libpng but after libpng-dbg" Jul 28 09:05:36 or we just overwrite PACKAGES instead of adding to it Jul 28 09:06:21 zecke Jul 28 09:06:29 http://www.amazon.de/gp/product/B00009P51G/302-8251796-0644805?v=glance&n=301052 Jul 28 09:07:10 pH5: I'd prefer not to do that. Can we make the expression exclude the dbg file? Jul 28 09:07:31 good morning Jul 28 09:07:35 hi florian Jul 28 09:07:38 hi florian_kc Jul 28 09:08:21 I'm trying to build glibc-2.4, but it says glibc-2.4 is unavailble and it reverts to 2.3.5+cvs, how can I find out why it's doing this? Jul 28 09:08:58 myopiate: hacking bin/bitbake Jul 28 09:09:31 that one wa bugging me earlier Jul 28 09:10:00 RP: I'm not sure how. An easy way out would be to follow debian's example and put all debug info under /usr/lib/debug. Jul 28 09:10:28 pH5: That isn't a standard location looked in by gdb though Jul 28 09:11:06 kergoth, how did you fix? My python skills are broken... maybe I should fix that first. Jul 28 09:12:01 RP: the path is added to the binaries via --add-gnu-debuglink anyway Jul 28 09:12:30 pH5: No, the path is not, just the name Jul 28 09:12:33 kergoth, zecke: In python globing, can you exclude an expression? Jul 28 09:13:06 good morning Jul 28 09:13:11 RP: ah, I understand. But the path could be added there? Jul 28 09:13:46 RP: not that i'm aware of Jul 28 09:14:21 pH5: No, it can't - it doesn't store the path at all Jul 28 09:15:22 pH5: I already checked into this. There was something about creating a .debug directory though which is in the standard search path Jul 28 09:15:48 If we did that, we'd remove the current problem but still have gdb etc find the files Jul 28 09:16:56 hi bernado Jul 28 09:17:32 RP: Strange. I was sure debian's dh_strip stores the whole path there. Jul 28 09:18:04 pH5: If you look up the specification on the gnu-debuglink section, it says no path Jul 28 09:19:16 morning morning Jul 28 09:19:18 RP: they should... (guessed) Jul 28 09:19:28 damn i like mtn's speed now :) Jul 28 09:19:53 zecke: There is no need if you store the debug symbols in the same directory... Jul 28 09:20:05 hmm it looks like it can't Jul 28 09:20:36 back to a bit of work Jul 28 09:23:15 hey Bernardo & ade|desk Jul 28 09:23:41 RP: the objcopy manpage says that the parameter to --add-gnu-debuglink is path-to-file Jul 28 09:23:51 where did you read that no path is allowed? Jul 28 09:24:16 or is this a debian specific patch? Jul 28 09:26:00 pH5: Its in the specification for the gnu-debuglink section - its quite clear about what it contains. debian probably hack gdb to look in /usr/lib/debug by default Jul 28 09:26:21 pH5: That won't help oprofile or other things that want the symbol. Jul 28 09:26:44 pH5: objcopy does take a path, but only so it can find the file, it will drop the path component Jul 28 09:28:12 well, so that's not an easy way out either :-/ Jul 28 09:28:38 We need to know what BFD's default search paths are Jul 28 09:29:00 (libBFD's) Jul 28 09:29:55 pH5: http://developer.apple.com/documentation/DeveloperTools/gdb/gdb/gdb_16.html Jul 28 09:30:29 The question is how much of that is gdb behaviour and how much things like oprofile will do too... Jul 28 09:31:42 koen: Your touchscreen device doesn't support BTN_TOUCH which is strange. I guess tslib musn't require it... Jul 28 09:32:40 RP: where is this globaldebugdir set? Jul 28 09:32:52 pH5: Its a gdb variable Jul 28 09:33:29 koen: Can you try s/*k*14A,// in your local.rules file? Jul 28 09:37:13 03arjan 07org.oe.dev * r59ba4be7... 10/ (1 packages/linux/linux-h1940_2.6.17-h1940.bb): linux-h1940: add forgotten patch to make usb gadget work Jul 28 09:40:10 RP: oprofile contains references to /usr/lib/debug (DEBUGDIR in libop/op_config.h) Jul 28 09:44:05 is there not a xserver-kdrive- implementation for the SiS drivers? Jul 28 09:44:46 pH5: That suggests there is no standard location then :-/ Jul 28 09:47:51 morning Jul 28 09:47:58 hey lardman Jul 28 09:48:00 RP: but it means that debug info in this dir is supported by oprofile at least. What else wants to look at these symbols? Jul 28 09:48:01 hi lamikr Jul 28 09:48:04 eh lardman Jul 28 09:48:34 hi chaps Jul 28 09:49:28 RP: pH5 can you not just "fix" debuggers produced by OE to look in the "right" place Jul 28 09:49:34 RP: removing that worked Jul 28 09:49:59 root@ipaq-pxa270:/etc/udev/rules.d$ ls /dev/input/ Jul 28 09:49:59 event0 event1 event2 touchscreen0 uinput Jul 28 09:50:34 koen: Great :). We should update udev Jul 28 09:50:55 XorA: Then we need to fix every debug program forever. There has to be a better way... Jul 28 09:51:15 RP: apart from wget, the new prelink is working pretty good Jul 28 09:51:25 I'd like to have all debugging information in a separate directory, question is, is this /usr/lib/debug some kind of unwritten standard or not? Jul 28 09:51:36 pH5: Have a look at binutils/bfd/opncls.c - find_separate_debug_file() Jul 28 09:51:52 RP: thanks, will do Jul 28 09:52:16 pH5: Its not a standard and has to be set in the environment. We can use a .debug directory underneath the binary's location though Jul 28 09:52:24 pH5: btw, I misconfigured busybox for angstrom, so 'ps' and 'wget' are broken Jul 28 09:52:44 koen: prelink broke wget? Jul 28 09:52:50 RP: yes Jul 28 09:53:02 prelink --u wget solved it Jul 28 09:54:09 koen: Interesting. Someone should really find out why it broke at some point... :) Jul 28 09:54:22 prelink does make a nice speed difference Jul 28 09:56:01 wb mickeyl! Jul 28 09:56:12 hey mickeyl Jul 28 09:56:17 hi guys! Jul 28 09:56:27 mickeyl: hey Jul 28 09:56:44 congrats for completing the migration! Jul 28 09:57:35 mickeyl: thanks for migrating ciabot before leaving Jul 28 09:57:40 mickeyl: and most of the tests were skipped Jul 28 09:57:50 hi mickeyl Jul 28 09:58:23 zecke: good to see you have found it :) i wasn't sure because i did it just when the taxi for the airport was waiting, so i couldn't tell you :) Jul 28 09:58:44 mickeyl: I was about to copy from merlin Jul 28 09:58:57 mickeyl: and then found the existing dir on openembedded.org Jul 28 10:06:48 hey mikearthur Jul 28 10:08:10 mickeyl: it worked flawlessly Jul 28 10:08:20 mickeyl: we should publish our cioabot though Jul 28 10:08:30 mickeyl: and I will steal code from it to get the list of updated revs Jul 28 10:08:41 this will be a bonsai monitor Jul 28 10:08:55 hey mikearthur Jul 28 10:08:56 er mickeyl Jul 28 10:08:59 hey guys Jul 28 10:09:14 is org.openembedded.oz354x the stable branch? Jul 28 10:09:24 mikearthur: for OE, no, for OZ, yes Jul 28 10:09:25 it is a stable branch Jul 28 10:09:37 RP: do you think this is worth a try: http://en.pastebin.ca/103517 Jul 28 10:09:50 morning all Jul 28 10:10:06 zecke: because it won't build for the tosa Jul 28 10:10:08 hi mikearthur, do13_ Jul 28 10:10:09 crapping out on glibc Jul 28 10:10:17 cheers kergoth Jul 28 10:10:24 hey Dirk Jul 28 10:10:29 oh well, all 'branches' should be buildable all the time Jul 28 10:10:34 do13_: wlanng-svn is broken again :( Jul 28 10:10:39 hey pH5, koen, mickeyl Jul 28 10:10:40 mikearthur: on glibc? that is strange Jul 28 10:10:50 koen: which kernel version? Jul 28 10:10:57 mikearthur: what is it crapping out on? no more space available? Jul 28 10:11:05 do13_: a patch doesnn't apply (2.6.17) Jul 28 10:11:25 ../sysdeps/unix/sysv/linux/bind.S:5: Error: symbol `__bind' is already defined Jul 28 10:11:25 | make[2]: *** [/home/mike/oe/tosa/tmp/work/glibc-2.3.5+cvs20050627-r7/build-arm-linux/socket/bind.o] Error 1 Jul 28 10:11:38 koen: ahh. yes they have updated their svn :) Jul 28 10:11:40 mikearthur to new binutils Jul 28 10:11:43 mikearthur: ah-ha, what is that branch again? Jul 28 10:11:57 org.openembedded.oz354x Jul 28 10:12:03 woglinde: eh? Jul 28 10:12:05 * koen guesses a wrong DISTRO Jul 28 10:12:26 koen: Mark Shuttleworth is coming round to kick your ass Jul 28 10:12:35 he is? Jul 28 10:12:54 mikearthur: http://www.openembedded.org/viewmtn/revision.psp?id=89e38da8206b15620ba7c2bef11c31e2dda99dd8 Jul 28 10:12:54 koen: can you do a test without the patch? Jul 28 10:12:57 yep Jul 28 10:13:08 mikearthur: actually you should not have 2.16 of binutils in your branch Jul 28 10:13:11 pH5: Looks good to me Jul 28 10:13:14 do13: I just selected the non svn version Jul 28 10:13:19 hi Dirk Jul 28 10:13:22 mikearthur: oh well, binutils 2.17 Jul 28 10:13:27 do13: and I don't have a tosa to test on :) Jul 28 10:13:28 hey Richard Jul 28 10:13:29 mikearthur: what version of binutils were built? Jul 28 10:13:58 mikearthur: did you change branch without wiping TMPDIR? Jul 28 10:14:14 koen: they will release a new version soon. Jul 28 10:16:39 I'll try and build an image with debug info in .debug subdirs. Jul 28 10:16:56 koen: no, zecke: how do I check Jul 28 10:17:16 bbl Jul 28 10:17:18 pH5: While I think on, could you have it just run basename/dirname once and store the name in a variable? Jul 28 10:17:20 mikearthur: looking at the console? Jul 28 10:17:24 :p Jul 28 10:17:27 built yesterday Jul 28 10:17:29 RP: sure Jul 28 10:17:46 pH5: Should save a few fractions of a second ;-) Jul 28 10:17:56 binutils-cross-2.17-r0/ Jul 28 10:18:05 mikearthur: http://www.openembedded.org/filebrowser/org.openembedded.oz354x/packages/binutils Jul 28 10:18:18 mikearthur: the error you describe was introduced with binutils 2.17 Jul 28 10:18:23 ah ok Jul 28 10:18:33 mikearthur: so something is wrong on your site Jul 28 10:18:45 mikearthur: either you are not using oz354x Jul 28 10:18:53 mikearthur: and your distro is wrong or... Jul 28 10:19:52 RP: Could you clarify to the reply on the version number change to a opie file? Jul 28 10:20:06 RP: what was wrong with it? violation of the versinoning policy? Jul 28 10:20:32 zecke: When it was upgraded, I was hinting we could change it to the standardised version form Jul 28 10:20:59 I think we could do it Jul 28 10:21:23 Nobody is using opie CVS in a release so people probably wouldn't object Jul 28 10:21:40 Its not something I can change but if an opie developer did it... :) Jul 28 10:22:04 mickeyl: http://www.oesf.org/forums/index.php?showtopic=16669&view=findpost&p=136709 Jul 28 10:22:09 RP: do you played with qemu 0.8.2? Jul 28 10:22:33 do13_: I use an older cvs version - I'm not sure how it equates Jul 28 10:23:04 RP: changelog: - Support for AMD Flash memories (Jocelyn Mayer) Jul 28 10:23:20 koen: yeah! Jul 28 10:23:29 lets see what's in there... Jul 28 10:23:45 #ifdef mickeyl Jul 28 10:23:48 die(); Jul 28 10:23:53 possibly Jul 28 10:25:19 mickeyl: wow, these two 'mails' sound harsh :} Jul 28 10:25:45 which ones? Jul 28 10:26:05 i've been asking him to release the source since february. i guess that shows my patience Jul 28 10:27:46 that always such a lame excuse Jul 28 10:27:50 "oh I'm so busy poor me" Jul 28 10:28:18 sorry, I'm too busy to feed my child, its not my fault it died Jul 28 10:28:58 ~hail OE Jul 28 10:29:08 * ibot bows down to OE and chants, "I'M NOT WORTHY!!" Jul 28 10:29:13 hehe Jul 28 10:29:31 wtf Jul 28 10:29:35 I hate dropbear Jul 28 10:29:36 do13_: Interesting. Probably for the OLPC project... Jul 28 10:29:43 piece of crap with hardcoded paths inside the binary Jul 28 10:32:08 NOTE: package zudoku-1.0-r0: task do_build: completed Jul 28 10:32:09 :D Jul 28 10:32:54 does anyone have a working X installation for x86 - one that does not segfault? If so, I'd really appreciate it if I could take a look at your config... Jul 28 10:33:30 I have tried all sorts here... most of the time things just fail to build, and when I finally got kdrive to build, it just segfaults on me all the time Jul 28 10:33:49 I'm using all the latest version Jul 28 10:38:40 oh cool - all binutils in dev and oz354x are broken for the sdk Jul 28 10:41:31 tomkp: install gdb and debug? Jul 28 10:41:50 at the end of the day this is still a tech channel Jul 28 10:42:03 zeckie: how can I get symbols into the target libs? Jul 28 10:42:09 they are all stripped Jul 28 10:42:42 a) inhibit the stripping b) as you use the latest versions install the dbg packages from the feed Jul 28 10:43:05 zecke: There is no need for a if you use .dev Jul 28 10:44:03 zecke: from the feed? Jul 28 10:45:08 tomkp: well Jul 28 10:45:15 tomkp: get a ipkg manual Jul 28 10:45:33 did anyone her build an sdk from unstable lately? Jul 28 10:45:36 tomkp: feed ====== bunch of packages, downloadable Jul 28 10:45:54 florian_kc: I built a native one from .dev Jul 28 10:46:15 koen: did it work? Jul 28 10:46:18 zeckie: right, but there are no xserver-kdrive dbg packages, as far as I can see Jul 28 10:46:26 florian_kc: yes Jul 28 10:46:32 tomkp: then you don't use the latest version Jul 28 10:46:41 florian_kc: hello.c worked, as well as mathomatic Jul 28 10:46:44 tomkp: rebuild with and inhibit the stripping Jul 28 10:47:01 tomkp: or update before rebuilding and get -dbg packages Jul 28 10:47:03 well I haven't specified any versions.. so it should be using the latest Jul 28 10:47:08 tomkp: I would stay with the current version... Jul 28 10:47:09 it did xserver-kdrive-0.0+cvs20060312-r11 Jul 28 10:47:21 ah I thought latest version of OE Jul 28 10:47:30 zecke... well, yes Jul 28 10:47:34 I updated yesterday Jul 28 10:47:55 koen: i found out that binutils are broken for the cross sdks - they don't ship ldscripts Jul 28 10:48:02 tomkp: you have your options, take one, try it Jul 28 10:48:26 RP: http://rafb.net/paste/results/4f2Fgc11.html :) Jul 28 10:48:31 koen: and they create links like "gcc -> arm-linux-gcc" Jul 28 10:48:40 that blows up everything Jul 28 10:48:51 florian_kc: the native ones have 'gcc-symlinks' for that Jul 28 10:48:59 and binutils-symlinks Jul 28 10:50:00 zecke: well I'd like to install the dgb packages, but I just don't see where they come from Jul 28 10:50:19 koen: then i winder why do_install tries to do the same for all binutils Jul 28 10:50:26 s/winder/wonder Jul 28 10:50:44 we have thet crap in *every* binutils package and in binutils.inc Jul 28 10:50:49 tomkp: as I said. If you have latest version of OE, they are generated Jul 28 10:51:03 tomkp: as you don't have them, you don't have the latest version Jul 28 10:51:07 koen: Seems to work :) Jul 28 10:51:12 tomkp: I won't try to update OE now Jul 28 10:51:24 tomkp: so rebuild and inhibit stripping of packages Jul 28 10:51:36 tomkp: don't ask how, you should remember a line from local.conf.sample Jul 28 10:52:03 zecke: guess my update didn't complete properly or something Jul 28 10:52:09 I'll pull & update again Jul 28 10:52:46 RP: and I put them in a seperate feed to avoid scaring users: Downloading http://www.angstrom-distribution.org/unstable/feed/debug/dbus-dbg_0.61-r0_armv5te.ipk Jul 28 10:53:03 pretty cool stuff Jul 28 10:53:23 koen: ah wait, looks like the newer binutils for the sdk override this behaviour Jul 28 10:53:25 koen: Excellent. They should prove pretty useful :) Jul 28 10:53:55 it still doesn't solve the dbus crash, though Jul 28 10:54:02 * koen mails backtrace to pH5 Jul 28 10:54:18 koen: but it does let you debug it ;-) Jul 28 10:54:32 that it does Jul 28 10:55:13 mickeyl: http://freshmeat.net/projects/ttlinux/ <- I challenge you :) Jul 28 10:55:55 heh Jul 28 10:56:10 that's not what you would expect it to be Jul 28 10:56:20 it's the OS, not the app Jul 28 10:56:23 03mickeyl 07org.oe.dev * rd685f587... 10/ (6 files in 4 dirs): add zudoku, a sudoku game for Opie Jul 28 10:57:35 mickeyl: I got misled by the screenie Jul 28 11:00:12 koen: mickeyl RP pb: Do you consider coming to the Chaos Congress this december? Jul 28 11:00:26 i can consider Jul 28 11:00:32 iirc harald will be there as well Jul 28 11:00:33 I will turn in a talk proposal "Rants of a cross compileist" Jul 28 11:00:53 once I find a proper title, I will send it in Jul 28 11:00:56 sounds appealing :) Jul 28 11:01:12 i'll keep it in mind. atm i can't plan much into the future Jul 28 11:01:45 right :} Jul 28 11:02:14 * zecke imagines ranting in front of a couple of people about stupid Makefiles, Makefile.am's and m4's Jul 28 11:02:28 all ending with a faked heart attach Jul 28 11:02:53 mickeyl: How's your freelance work progressing? Jul 28 11:03:42 do13: first assignment starting in august while I'm still trying to get "Überbrückungsgeld für Existenzgründer" from Arbeitsagentur. Jul 28 11:04:16 still the not-yet-submitted-thesis is my 1st priority Jul 28 11:04:34 I can imagine this Jul 28 11:07:05 zecke: I'm not sure I'll be there but I'll keep it in mind Jul 28 11:08:45 * zecke should start learning and try to get an appointment to make his next exam Jul 28 11:10:12 03mickeyl 07org.oe.dev * ra5ef4c4c... 10/ (1 packages/tasks/task-qpe.bb): task-qpe: add zudoku to qpe-games Jul 28 11:13:55 hrm... damn, i pushed crap... Jul 28 11:14:52 03florian 07org.oe.dev * r9349d420... 10/ (1 packages/meta/meta-sdk.bb packages/tasks/task-sdk.bb): task-sdk, meta-sdk: Update libmimedir, add additional libraries. Jul 28 11:15:00 03florian 07org.oe.dev * r23f392c4... 10/ (1 packages/binutils/binutils_2.15.94.0.1.bb): Jul 28 11:15:00 binutils 2.15.94.0.1: Do not remove ldscripts, it breaks the SDK. Jul 28 11:15:00 Do not create links without cross prefix, it break the SDK. Jul 28 11:16:49 mickeyl: how about checking in your libetpan patch? Jul 28 11:21:39 03florian 07org.oe.dev * re273cc9b... 10/ (1 packages/binutils/binutils_2.15.94.0.1.bb): disapproval of revision '23f392c4a32f829ba2c33998e47b14cc08c51a26' Jul 28 11:23:06 XorA: good idea ;) Jul 28 11:25:30 03mickeyl 07org.oe.dev * r4c2bff0a... 10/ (1 packages/libetpan/files/honor-oe-lflags.patch): libetpan: add new lflags patch Jul 28 11:25:35 mickeyl: cheers Jul 28 11:25:48 ~lart so called web designers Jul 28 11:25:48 * ibot executes killall -HUP so called web designers Jul 28 11:26:29 ~lart marketing depts Jul 28 11:26:29 * ibot stuffs marketing depts into a shiny new tin can and vacuum seals it Jul 28 11:27:35 Do we stage before or after we package? Jul 28 11:27:49 zecke: depends on the branch Jul 28 11:27:57 zecke: in my branch after Jul 28 11:28:14 eek Jul 28 11:31:00 zecke: we also have two ways to stage in .dev: do_stage() and do_populate_staging() Jul 28 11:31:34 ah right Jul 28 12:19:27 zecke: how do I remove the previous binutils-cross, just install a different version? Jul 28 12:20:07 bitbake -b ../../packages/binutils/prev-file.bb -cclean Jul 28 12:20:19 mikearthur: I wonder which version it built! Jul 28 12:20:24 mikearthur: answer this please Jul 28 12:20:31 2.17 I think Jul 28 12:20:34 how can I check? Jul 28 12:20:40 2.17 is not in the branch ;) Jul 28 12:20:42 it was yesterday Jul 28 12:20:44 arm-linux-ar --version Jul 28 12:21:16 GNU ar 2.16 Jul 28 12:21:48 sure you didnt hit your host-ar? Jul 28 12:21:49 hmm Jul 28 12:21:54 and you build branch right? Jul 28 12:21:57 tosa/tmp/cross/arm-linux/bin/ar --version Jul 28 12:22:02 okay Jul 28 12:22:05 built .oz354 Jul 28 12:22:36 bleh Jul 28 12:22:44 adding debug spew to dbus fixed teh segfault Jul 28 12:22:58 koen aeh? Jul 28 12:23:25 woglinde: I sprinkled printf("%s(%d)""got here", __FILE__, __LINE__); everywhere Jul 28 12:26:22 koen but how fixed that the segfault Jul 28 12:26:33 I have no idea :( Jul 28 12:26:50 port valgrind, run it, find it out Jul 28 12:27:14 woglinde: you write to another area of memory, which leads to corruption but not to a segfault Jul 28 12:27:49 hm valgrind is x86 only? Jul 28 12:27:59 woglinde: no Jul 28 12:28:03 woglinde: it runs on PPC as well Jul 28 12:28:08 it got more portable Jul 28 12:28:16 hm Jul 28 12:29:45 still Fabrice didn't create an ARM port yet ;) Jul 28 12:30:01 zecke: any glibc ideas? Jul 28 12:30:45 mikearthur: do not use binutils 2.16 ;) Jul 28 12:30:54 :) Jul 28 12:30:58 2.15 then? Jul 28 12:31:03 mikearthur: It smells like you used binutils 2.17 Jul 28 12:31:14 mikearthur: and that you are not on the branch (even if you think you are) Jul 28 12:31:31 mike@steyraug ~/oe $ echo $BBPATH Jul 28 12:31:32 /home/mike/oe/build:/home/mike/oe/org.openembedded.oz354x Jul 28 12:33:06 and BBFILES? Jul 28 12:34:15 bollocks Jul 28 12:34:39 aye? Jul 28 12:34:55 forgot about BBFILES :( Jul 28 12:35:00 hehe Jul 28 12:35:11 in Germany we would say "Lutscher" Jul 28 12:35:15 what now? Jul 28 12:35:36 set BBFILES correctly? Jul 28 12:35:45 I habe a problem. I try to integreate a source code wich have a very complex system of makefiles. They are overwriting most the envorinment variables wich came from bitbake and openembedded. I changed them that they are using the compilers of the cross compiler environment. This is working. The the binaries still get linked against the libraries of the building System, not against the one in the staging directory. Jul 28 12:35:50 do I need to remove my tmp? Jul 28 12:35:59 mikearthur: it would be better Jul 28 12:36:06 :'( Jul 28 12:36:15 I already tried it to appent a -L to the STAGING_LIBDIR but it didn't help Jul 28 12:36:30 does exist any other flag that I could use? Jul 28 12:38:34 cobelius show us how the compiler is invoked at the end Jul 28 12:38:43 maybe you missed some -L Jul 28 12:38:53 and gcc takes the first lib he finds Jul 28 12:39:03 not that one which is suitable Jul 28 12:39:18 hm thats an idea Jul 28 12:39:23 to make gcc more robust Jul 28 12:40:07 "make sure that the lib is for the arch" Jul 28 12:40:14 http://pastebin.de/9809 Jul 28 12:40:17 shouldnt be so hard Jul 28 12:40:52 http://pastebin.de/9810 Jul 28 12:41:34 -L../../../lib/linux-x86/ is pointing to which real directory? Jul 28 12:41:49 Well, I hoped that the "-L" flag will do it, but it don'T Jul 28 12:42:19 to /var/tmp/bitbake/work/epics-3.14.7/lib/linux-x86 Jul 28 12:42:23 kill this second -L Jul 28 12:42:32 and try again Jul 28 12:42:47 These are some libs wich are compiled by epics and gets linked Jul 28 12:42:57 the libs wich are wrong are libc and libpthread Jul 28 12:43:00 okay I try it Jul 28 12:43:01 Cobelius: do yourself a favor. Shoot the guy who wrote that crappy Makefile Jul 28 12:43:19 The epics make file? Jul 28 12:43:48 aehm Jul 28 12:43:57 hm your compiling for x86 with oe to Jul 28 12:44:11 so guess whats ldd shows Jul 28 12:44:35 whats the stable distro/branch for an angstrom? Jul 28 12:45:05 03koen 07org.oe.dev * r0e8bd604... 10/ (3 files in 3 dirs): Jul 28 12:45:05 dbus 0.90: add a patch to fix the segfault when trying to access the system bus Jul 28 12:45:05 * NOTE: this is not a solution, but just moves the memory corruption in to a non-segfault area Jul 28 12:45:09 mikearthur: DISTRO=angstrom-2006.9 and .dev Jul 28 12:45:13 ldd shows that it takes the own libs from the work directory and the "standard" libs from /lib Jul 28 12:45:17 koen: cheers Jul 28 12:45:33 so, my software is running on the Build machine but not on the target image Jul 28 12:45:51 there I get a segmentation fault because the libc and libpthread are too different Jul 28 12:46:14 cobelius hm Jul 28 12:49:50 http://www.linuxdevices.com/news/NS2701809386.html Jul 28 12:49:52 hm Jul 28 12:50:10 @zecke: The problem with the makefiles is that it tries to provide a confotable envirnmoent for cross-compiling by it self. So it's a little bit difficult to intergrate it in openembedded (even RTAI was easier to integrate) Jul 28 12:50:41 Cobelius: well that is the point. The Makefile is authored by a moron not known what he was doing Jul 28 12:50:52 zecke jesus Jul 28 12:50:59 bye Jul 28 12:51:04 you can say it other words Jul 28 12:51:08 woglinde: he? Jul 28 12:51:20 you are real offending today Jul 28 12:51:25 something wrong? Jul 28 12:51:26 woglinde: lol Jul 28 12:51:33 woglinde: no, not at all :) Jul 28 12:51:50 woglinde: I have just read too many 'clever' makefiles Jul 28 12:52:01 yes Jul 28 12:52:16 we should make workshops on Makefiles and autotools Jul 28 12:52:29 woglinde: it is common practice to influence the build by exportting/setting CC and CXX and CPPFLAGS and CFLAGS Jul 28 12:52:56 woglinde: the next thing is putting CFLAGS += "-I/usr/include/ssl" into your makefile Jul 28 12:53:05 what are the criterias of a good makefile? Jul 28 12:53:29 Cobelius: beleive what the user passes you, dont make replace it with wrong stuff Jul 28 12:53:30 Cobelius: hehe, I don't know. I can't author Makefiles, I just need to fix them every second day Jul 28 12:53:38 there is no gnash bb , isnt it ? Jul 28 12:53:56 Cobelius: I think Keep It Stupid Simple (KISS) is the best design strategy for Makefiles Jul 28 12:54:04 hmm Jul 28 12:54:07 use $(CC) when compiling $(AR) for other stuff Jul 28 12:54:13 what ever happed to my rules of thumb of open source buildsystems document? Jul 28 12:54:15 Cobelius: and now the master will raise the word Jul 28 12:54:25 cobelius hm I think with a good Makefile you can compile, crosscompile, install uninstall Jul 28 12:54:26 kergoth: tell us :) Jul 28 12:54:32 it was on a wiki Jul 28 12:54:36 pretty sure it wa sthe oe one. Jul 28 12:54:39 i cant find it there now. Jul 28 12:54:55 kergoth: I have the wiki 'src' at home. I can check it Jul 28 12:55:06 kergoth: I never came across this document though Jul 28 12:55:58 kergoth: we were in the process of migration the old wiki to the new wiki when suddenly oe.hh.org was gone Jul 28 12:56:10 at which point we stopped migrating Jul 28 12:56:48 kergoth: hmm, I think that statement is wrong Jul 28 12:57:31 the old wiki src is at oe.org Jul 28 12:58:46 to correct it. We were asked if we need oe.hh.org Jul 28 12:58:51 and we said "No" Jul 28 12:59:10 zecke: it should be in ~/wiki Jul 28 12:59:29 zecke: so people should stop complaining about missing stuff Jul 28 13:02:17 koen: when do you want to hack on .pc and .la files? Jul 28 13:02:35 zecke: when the first error appears :) Jul 28 13:03:22 zecke: bitbkae has a CACHETIMESTAMP error :( Jul 28 13:05:49 any ideas? Jul 28 13:06:04 mikearthur: no, what should that be? Jul 28 13:06:11 zecke: http://openembedded.org/cgi-bin/moin.cgi/Buildsystems_2fRulesOfThumb was it, long long ago Jul 28 13:06:17 bitbake has a CACHETIMESTAMP error Jul 28 13:07:04 mikearthur: what should that be? Jul 28 13:07:13 I'm confused Jul 28 13:07:24 mikearthur: is it saying that it has one? Jul 28 13:07:25 theres an exception in python Jul 28 13:07:33 pastebin Jul 28 13:07:35 yeh, bitbake is crashing out Jul 28 13:07:35 sure Jul 28 13:07:43 and hope RP listens Jul 28 13:07:50 mikearthur: after changing BBFILES Jul 28 13:07:54 mikearthur: did you remove tmp? Jul 28 13:08:47 We should ask hh.org for a copy of that wiki data from the backups Jul 28 13:08:58 RP: read what I said Jul 28 13:09:37 RP: we have the stuff Jul 28 13:09:48 zecke: removed tmp, ran it again with broken envs, broke bitbake, read what you said, removed tmp again, fixed bitbake :) Jul 28 13:09:52 ah, ok, that's fine. I missed that in the general noise, sorry Jul 28 13:10:54 mikearthur: I would like to see the exception though Jul 28 13:11:00 sure Jul 28 13:13:06 okay Jul 28 13:13:11 I will head home Jul 28 13:13:19 bye till later Jul 28 13:14:44 zecke: http://pastebin.ca/103707 Jul 28 13:16:08 where is the best files to desactivate locales for gcc-cross ? Jul 28 13:16:17 mikearthur: thanks :) Jul 28 13:16:39 s/files/way Jul 28 13:17:44 bbl Jul 28 13:19:47 aargh Jul 28 13:19:55 root@ipaq-pxa270:/usr/lib/ipkg/info$ localedef --inputfile=/usr/share/i18n/locales/en_US --charmap=UTF-8 --prefix=/tmp/locale en_US Jul 28 13:19:55 cannot map archive header: Invalid argument Jul 28 13:21:29 jffs2? Jul 28 13:21:37 no, localegen Jul 28 13:21:40 kergoth: do you remember PACKAGEFUNCS? Jul 28 13:21:57 ENABLE_BINARY_LOCALE_GENERATION ? Jul 28 13:22:08 not really, no Jul 28 13:22:27 Genesis: this will only toggle using qemu for running localedef Jul 28 13:22:38 kergoth: good answer, it is flawed :} Jul 28 13:22:52 and qemu doesn't support EABI Jul 28 13:23:01 i want to deactivate local compile when it compile glibc , perharps i should use uclibc Jul 28 13:23:32 Genesis: turn it off in the libc bbfiles Jul 28 13:23:45 Genesis: or try using uclibc Jul 28 13:23:52 oki Jul 28 13:23:59 i thought there was a better way Jul 28 13:24:00 thx Jul 28 13:24:28 because if i modify a bb file , i should modify each time there is an update Jul 28 13:24:46 Genesis: only if the file was touched Jul 28 13:25:00 Genesis: and then monotone will launch kdiff3 or meld to help you merge the change :) Jul 28 13:25:33 03koen 07org.oe.dev * r929ac3e1... 10/ (1 packages/glibc/glibc-package.bbclass): glibc.bbclass: remove spurious '/' Jul 28 13:25:34 Genesis: and RP had a patch to only enable locale generation for some locales Jul 28 13:25:48 Genesis: but there were some issues I do not remember Jul 28 13:25:56 ouch oki :) Jul 28 13:26:04 i'll try uclibc i think Jul 28 13:26:38 Genesis: I hope you don't want to try Opie? as uclibc and Opie are not the best friends yet Jul 28 13:27:06 i think about matchbox Jul 28 13:27:19 that's oki with uclibc ? Jul 28 13:27:22 that might work Jul 28 13:27:37 but I have no clue Jul 28 13:27:43 i think Jul 28 13:28:51 * koen starts buying intel and amd stock Jul 28 13:30:20 03koen 07org.oe.dev * rd9d27407... 10/ (11 files in 2 dirs): glibc: and bump PR after last change Jul 28 13:34:20 03freyther 07org.oe.dev * rceeb4762... 10/ (1 classes/insane.bbclass): Jul 28 13:34:20 classes/insane.bbclass: Remove GPL text, add more planned checks Jul 28 13:34:20 The initial idea was to use the shell magic from ebuild, Jul 28 13:34:20 but I do not plan this anymore and will do it using python. Jul 28 13:34:20 This is why I'm removing the GPL license. Jul 28 13:34:20 Add some more planned checks and assume they are already Jul 28 13:34:22 implemented. Jul 28 13:34:26 03freyther 07org.oe.dev * r375c0460... 10/ (5 files in 2 dirs): Jul 28 13:34:28 packages/pax-utils: Update to 1.1.13, make the effect of make -e up Jul 28 13:34:31 Update to pax utils 1.1.13 as the source vanished and the Jul 28 13:34:33 new version works nicely. Jul 28 13:34:34 The Makefile of PAX would add -D_GNU_SOURCE but we are using -e Jul 28 13:34:36 as command line option and this option was not added to CPPFLAGS. Jul 28 13:34:39 03freyther 07org.oe.dev * r252cec78... 10/ (1 classes/package.bbclass): Jul 28 13:34:40 classes/package.bbclass: Only set PACKAGEFUNCS once! Jul 28 13:34:45 Set PACKAGEFUNCS only once. Now if more than one bbclass inherits Jul 28 13:34:46 package both additional functions will be executed. The new packagefuncs Jul 28 13:34:48 will be appended to the list and the old ones will not be lost Jul 28 13:34:53 03freyther 07org.oe.dev * r526a073f... 10/ (1 classes/insane.bbclass): Jul 28 13:34:54 classes/insane.bbclass: Add skeleton for the tests Jul 28 13:34:56 Run each test from within the PACKAGEFUNC, add a special Jul 28 13:35:00 task for checking .la and .pc files from within staging. Jul 28 13:35:26 insane.bbclass, eh? Jul 28 13:35:45 kergoth: Want to talk about sanity with me? Jul 28 13:36:01 kergoth: everything what is insane will be detected within this class Jul 28 13:36:06 ah Jul 28 13:36:18 kergoth: dangerous rpath, bogus file permissions, stupid .pc and .la files Jul 28 13:36:31 kergoth: now even planned wrongly placed .dbg files Jul 28 13:37:22 sounds nice Jul 28 13:37:29 koen: lol Jul 28 13:37:34 koen: what a bogus patch :} Jul 28 13:37:44 zecke: for dbus? Jul 28 13:37:50 that reminds me, from the environmental issues department.. the glibc builds need to reset the TERMINFO variable Jul 28 13:37:57 yes, it's bogus, but it fixes the segfault Jul 28 13:38:00 'fixes' Jul 28 13:38:09 i ended up with a bunch of /home/kergoth/.terminfo files listed as unpackaged in teh glibc install/package :P Jul 28 13:38:11 koen: send upstream Jul 28 13:38:21 Hi Jul 28 13:39:14 * zecke stops working and buys burger+icecream Jul 28 13:40:19 ~botmail for woglinde I won't be online until tomorrow morning, mail the time we will have lunch Jul 28 13:42:07 later Jul 28 13:43:51 for uclibc , TARGET_OS="linux-uclibc" in the distro . conf ? Jul 28 13:47:00 oki it's that ): Jul 28 13:51:08 Hm I am still trying to crosscompile qtopia core 4 with open embedded. Before I used Embedix and it cross compiled without problem when I just set the PATH environment variable to the arm binarys. Is there any howto to cross compile an app or library with openEmbedded? Jul 28 13:52:24 ~botmail for zecke: https://bugs.freedesktop.org/show_bug.cgi?id=7675 Jul 28 13:52:57 mickeyl: didn't you add qtopia4 to OE already? Jul 28 13:54:28 koen qtopia4 doesn't exist just qtopia core 4 Jul 28 13:55:11 I know Jul 28 14:09:05 yes, i added qt/e 4 months ago. Jul 28 14:09:14 had to patch the guts out of it to make it work Jul 28 14:09:53 i removed it because i didn't want to play catch up with all those patches Jul 28 14:10:26 i don't see a use for it until you want to develop a complete new environment based on it Jul 28 14:11:17 although i'd be interested in seeing some recent benchmarks with that Jul 28 14:12:06 PolicyBased: take a look at our qte2.x and qt4.x bbfiles to get an idea of how to compile qt4. Jul 28 14:12:16 qt4/e, that is Jul 28 14:12:40 I'm not yet ready to accept that rename... Jul 28 14:13:20 mickeyl and where do I find the qt4.x bbfiles? Jul 28 14:13:29 packages/qt ? Jul 28 14:16:19 mickeyl: fwiw: mysql doesn't fetch so qt3x11 doesn't compile Jul 28 14:16:54 mickeyl, done with thesis? Jul 28 14:18:01 Crofton: no, unfortunately not yet. first release candidate is subject to discussion with my advisor on monday. i'm afraid he still has some points he want me to take into account before submitting it. *sigh* could take a month or two more :/ Jul 28 14:18:20 bummer Jul 28 14:18:32 I noticed you had dropped thesis from your nick :) Jul 28 14:18:53 I need to tell my advisor I need to schedule time into my life to write mine Jul 28 14:19:03 to the detriment of some projects we have Jul 28 14:19:53 of course, part of the plan is to do more OE work for the OSK :) Jul 28 14:20:01 that'd rock :) Jul 28 14:20:10 yeah Jul 28 14:20:27 another guy at school is doing some work on the TI davinci eval baord Jul 28 14:20:29 time management is critical. i just learned that in the past couple of years when I spent too much on silly things instead of bringing my research forward Jul 28 14:20:34 yeah Jul 28 14:20:51 mickeyl: So how I understand it with qt-x11-free_4.0.1.bb I can cross compile qt 4 for my device and so qt 4 app's would work on it? Jul 28 14:20:55 also dealing with stuff other students should be doing ...... Jul 28 14:21:25 PolicyBased: that's right. qt4/x11 works fine and I already added a couple of qt4 apps to OE. Jul 28 14:21:35 *nod* Jul 28 14:21:41 what will your thesis be about? Jul 28 14:21:59 something about software defined radio Jul 28 14:22:02 :) Jul 28 14:22:23 cool. pretty technical stuff isn't it? Jul 28 14:22:43 yeah Jul 28 14:23:00 I suspect I will write up some implementation and performance stuff Jul 28 14:23:12 bare minimum needed for signatures :) Jul 28 14:23:39 ah ok Jul 28 14:23:39 hehe. cool Jul 28 14:24:18 But mickeyl: qtopia core 4 would be better, because it is optimized for an embedded system? Jul 28 14:24:25 just need to get that behind me Jul 28 14:24:35 anyway off to the office Jul 28 14:24:43 Crofton: ya. l8er Jul 28 14:24:48 but for the beginning qt-x11 4 is perfect for me to. With this I could live ;-) Jul 28 14:25:16 PolicyBased: the problem has a broader scope Jul 28 14:25:28 it's questionable whether you want to lock out thousands of apps Jul 28 14:25:40 by not using X Jul 28 14:25:55 that's more or less the major point to decide on Jul 28 14:27:01 wonder if anyone thought of libqt11 along the lines of libw11 done for windows Jul 28 14:27:20 mickeyl: when I have openzaurus with opie, should qt-x11 work or do I need gpe, because of the x11? Jul 28 14:28:08 PolicyBased: both will do after installing the necessary packages after the install. it's easier to base on a gpe-image though, because that already contains an X server. Jul 28 14:28:35 I understand Jul 28 14:30:13 XorA: there's x/qt if you really need an x-server based on qte. Jul 28 14:30:22 XorA: but like all emulation layers, it's slow. Jul 28 14:30:40 * mickeyl going out a bit Jul 28 14:30:41 l8er Jul 28 14:30:51 mickeyl: I have used it, but lib11 stuff works by implentng libx11 in different toolkits Jul 28 14:37:27 mickeyl: so you tryied to compile qt/e 4, but it is to complicated? I compiled this week qtopia core 4 with Embedix crosscompiler, so I thought it should be also easy with OpenEmbedded Jul 28 14:52:02 cu thanks Jul 28 14:53:28 hmn Jul 28 14:53:41 the libstdc++ und libgcc is not in my staging dir Jul 28 14:53:45 and not in my image Jul 28 14:53:51 is this normal? Jul 28 14:54:45 in the work directory of gcc-cross are perapred directories for packages of this library Jul 28 14:54:49 But they are empty Jul 28 14:55:23 seem to be unable to build libxrender since it's looking for Xrender.h, which is not in my staginginc dir (although I do have render.h in there) Jul 28 14:55:37 have filed (another) bug report Jul 28 14:57:30 hi - need help on geting list of preferred versions from bitbake - 'bitbake -s' only seems to work if i actually build something - is there another way? Jul 28 14:57:36 (or rather Xregion.h/region.h Jul 28 14:57:37 ) Jul 28 15:00:30 anyone know what type of chip a "m8k" is? Jul 28 15:02:39 a m68k with a missing 6? Jul 28 15:02:48 hehe, apparently not Jul 28 15:02:58 we're trying to track down some CPU used for a project next year Jul 28 15:03:11 All the API includes refer to something called M8K Jul 28 15:03:15 They're chips called Cypress Micros. Jul 28 15:04:05 botmail read Jul 28 15:04:11 ~botmail read Jul 28 15:04:13 hms Jul 28 15:05:38 mikearthur: maybe one of those VHDL CPU designs for sticking in FPGAs ? Jul 28 15:05:48 maybe Jul 28 15:05:54 looks like a m88k derivative Jul 28 15:06:03 re Jul 28 15:06:11 koen are you around? Jul 28 15:06:15 yes Jul 28 15:06:19 hm Jul 28 15:06:34 is my oe key arrived? Jul 28 15:07:56 yes Jul 28 15:08:02 I'm waiting for mickey_away to ack Jul 28 15:08:07 oh Jul 28 15:08:08 okay Jul 28 15:08:21 thanks for info Jul 28 15:13:00 rp: hi Jul 28 15:13:17 hi raduga Jul 28 15:13:22 hi raduga Jul 28 15:13:38 rp: thanks for repatching the patch :) Jul 28 15:14:08 raduga: np, it was a silly typo and I should have spotted it :) Jul 28 15:14:21 rp: you commented "this was noticed and fixed in the last kernel update"; where is the updated/fixed kernel? Jul 28 15:14:36 raduga: checked into oe's .dev Jul 28 15:14:47 good. Jul 28 15:15:17 let me try pulling .dev again Jul 28 15:17:13 does tosa try to use 2.4 by default? Jul 28 15:17:18 for stable branch Jul 28 15:17:21 re Jul 28 15:17:25 mikearthur: yes Jul 28 15:17:38 can someone give me a pointer on how to use jffs2dump? Jul 28 15:17:44 how do I make it use 2.6? Jul 28 15:17:47 ierk: #mtd? Jul 28 15:18:05 and another question, as the oz channel is very low traffic right now Jul 28 15:18:12 RP: thnks Jul 28 15:18:20 to the other question Jul 28 15:18:40 X11 pkgconfig seems to be screwed up Jul 28 15:18:44 mikearthur: There's some variable to set, read tosa.conf Jul 28 15:18:53 * XorA feels underloved, he is only rated low traffic :-) Jul 28 15:19:03 all the libs are in include/X11, yyet the pc files just point to include/ Jul 28 15:19:05 I've just gotten myself an c3200. the OZ page tells me that when i install OZ the flash is untouched. can i still boot the existing sharp rom? Jul 28 15:19:37 rp: should it have been fixed in .dev as of GMT 20:00 yesterday? Jul 28 15:20:13 tomkp libs are in include? Jul 28 15:20:20 sorry... Jul 28 15:20:21 headers Jul 28 15:20:52 rp: your code is good. I'm trying to get a good process going, so that I can recognise when things are already fixed, so I don't submit redundant bugs Jul 28 15:20:58 raduga: It was committed Wed at 23:56 Jul 28 15:21:19 ok, thanks :) Jul 28 15:21:38 rp: so i have something old on my end Jul 28 15:21:50 raduga: You could always watch the oe commits mailing list filtering on linux-openzaurus/linux-oz Jul 28 15:22:18 Providing I remember to spell it correctly. We've had beef flavoured linuz-ox before :) Jul 28 15:22:50 rp: i try to follow the CIA reports in here ;) Jul 28 15:23:07 raduga: That would also mention it :) Jul 28 15:23:35 ok. i've pulled a nice, new .dev Jul 28 15:23:39 let's see what it says.. Jul 28 15:27:51 hi Jul 28 15:32:33 afternoon Jul 28 15:33:02 hi reenoo, CoreDump|home Jul 28 15:33:20 hey RP Jul 28 15:33:26 re reeno Jul 28 15:34:01 hi woglinde Jul 28 15:37:07 RP: ZKERNEL_VERSION = "2.6" in my tosa.conf.... Jul 28 15:39:52 mikearthur: For the OZ branch? That should be fine then... Jul 28 15:40:03 I'm doing dev on 2.6 though :( Jul 28 15:40:24 sorry, misunderstood you Jul 28 15:40:41 yeh, I've got that for the oz354x branch, but yet, its trying to build the poodle-kernel :s Jul 28 15:41:16 The poodle kernel sounds like the wrong MACHINE... Jul 28 15:42:06 BBPATH="/home/mike/oe/build:/home/mike/oe/org.openembedded.oz354x" bitbake gpe-image Jul 28 15:42:14 and in my local conf MACHINE="tosa" Jul 28 15:42:47 Did bitbake consider a different kernel before trying to poodle one? Jul 28 15:43:06 looks like it Jul 28 15:43:14 define PREFERRED_PROVIDER_virtual/kernel? Jul 28 15:53:21 RP: Sorted Jul 28 15:56:50 How can I get commit access to the main OE db? Jul 28 15:58:00 Crofton: felate someone who currently has it Jul 28 15:58:12 ( I don't know ) Jul 28 15:58:18 I know about that approach :) Jul 28 15:59:03 Crofton: Ask koen/mickeyl Jul 28 15:59:45 and by "Ask" he means.... Jul 28 16:00:03 I'm hoping one/the other is awake now :) Jul 28 16:02:23 see the monotonephrasebook wiki Jul 28 16:05:45 as in go ahead and mail my public key? Jul 28 16:07:31 hey all - jkp here on tom's book Jul 28 16:08:45 so, we've been banging our heads against x11 for days. tom tells me that he has been told the latest stuff in the repo works. we've done clean checkouts and builds and cannot for life of us get it to build. atm its failing looking for Xregion.h which is not in this version at all Jul 28 16:09:03 we have it in a tmp dir from a later version we once built Jul 28 16:09:39 my question is this: the known working versions - are those simply the versions that get pulled with *no* preffered entries in any configs when you checkout from the current repo? Jul 28 16:09:52 because this is what we are doing and its taking us knowhere :/ Jul 28 16:11:43 hmm, why is bootstrap-image using task-bootstrap-unionroot? Jul 28 16:12:09 anything anyone can suggest or tell us at all will be much appreciated...i know we've been naging on about this for a while but its the last hurdle for our protoype! Jul 28 16:17:39 Hey guys is there any way I can get SDL support with a GPE image without using Qt libs? Jul 28 16:21:07 good day / bonne soirée Jul 28 16:26:21 dtx: you need to set a preferred provider Jul 28 16:27:18 add to your local.conf : Jul 28 16:27:20 PREFERRED_PROVIDERS += " virtual/libsdl:libsdl-x11" Jul 28 16:27:48 iirc the default is libsdl-qt Jul 28 16:28:28 * raduga wonders if there should be a third option, for console SDL without x or qt support.. Jul 28 16:44:16 Hey guys is there any way I can get SDL support with a GPE image without using Qt libs? Jul 28 16:48:00 dtx: yes, install sdl-x11 Jul 28 16:51:40 rp: i just pulled the latest (17:00, friday) .dev from monotone.openembedded.org Jul 28 16:51:57 and it still made reference to the old asoc patch. Jul 28 16:52:27 (and still grabbed it, patched with it, died with it) Jul 28 16:56:51 there haven't been any changes to linux-oz since spitz.c broke the build Jul 28 16:57:06 at least, none that fix the error I'm seeing Jul 28 16:57:32 koen: the fix is pretty simple. Jul 28 16:57:46 koen: RP says its merged, but I can't see that Jul 28 16:58:23 koen: line 323, add a ) at the end of the expression Jul 28 16:58:39 Yes. And I was silly enough to file a bug with the fix :) Jul 28 16:58:59 polyonymous: :) Jul 28 16:59:27 But it's always a pleasure to watch your bug being closed :) Jul 28 17:00:53 rp: Jul 28 17:00:57 *polyonymous* It is not fixed in mtn Jul 28 17:01:56 hmm... could it be that it's fixed without filename changes and the file is already downloaded? Jul 28 17:01:57 ok. Jul 28 17:02:24 polyonymous: i've wiped the sources and restarted Jul 28 17:02:33 I see. I haven't :) Jul 28 17:02:44 rp said: raduga: It was committed Wed at 23:56 Jul 28 17:03:01 uh... my last pull was an hour or two ago... Jul 28 17:03:10 after a change is committed, does it still have to be approved, and merged by someone else? Jul 28 17:03:26 I don't think so. Jul 28 17:03:31 okay. Jul 28 17:03:34 not that I know much about mtn. Jul 28 17:03:45 * raduga is learning too Jul 28 17:04:02 I don't even learn ;-) I just know pull, update and diff ;-) Jul 28 17:04:08 * koen recommends http://venge.net/monotone/monotone.html Jul 28 17:04:28 Well, I think mtn command line help is quite comprehensive. Jul 28 17:04:47 At least if you know similar tools. Jul 28 17:05:04 koen: that is helpful, but I'm also trying to learn/understand the structure of OE development, as a social network Jul 28 17:05:36 koen: bugs are created by people, bugs are fixed by people. To understand the bug, you need to understand the person :) Jul 28 17:06:40 quite a philosophy :) Jul 28 17:06:56 Do you need to understand the bug before you fix it? Jul 28 17:07:03 not necessarily Jul 28 17:07:10 Like do you need to understand a real bug before you kill it? :) Jul 28 17:07:19 people do stuff without understanding, all the time Jul 28 17:07:39 Yes and sometimes this is very destructive... Jul 28 17:07:57 too much analysis can be destructive too Jul 28 17:08:33 Nothing is perfect under the skies? :) Jul 28 17:08:35 * raduga advises polyonymous stick to the 'pull, update, diff' strategy Jul 28 17:08:56 polyonymous: are things perfect above the skies? :) Jul 28 17:09:11 apparently Jul 28 17:09:15 not sure. If there's no "above" quite likely. Jul 28 17:09:18 otherwise, why would the proverb exist? Jul 28 17:10:34 CosmicPenguin, well, there's a story about black ship and astronomer, who wrote down "all ships in scottland are black", physicist who wrote "some ships in scottland are black" and mathematician who wrote "there's at least one field in Scottland with at least one ship which looks black from at least one side" ;-) Jul 28 17:11:17 So, not knowing what's above it's only safe to talk about "under the skies" :) Jul 28 17:11:47 As for mtn, I'd love to learn more if I liked it, but I don't ;-) Jul 28 17:15:49 raduga: The fix is not in the asoc patch, its in asoc_fixes2-r1.patch Jul 28 17:15:56 raduga: Are you using .dev? Jul 28 17:22:22 re Jul 28 17:29:15 RP, two hours old pull from .dev still has the bug. Jul 28 17:29:42 And yes, I believe raduga did mention .dev Jul 28 17:30:54 uh.. arm-linux-gcc: /.2/home/hacker/OE/build/tmp/staging/arm-linux/lib/.libs/libz.a: No such file or directory, why .libs? Jul 28 17:32:00 libtool Jul 28 17:32:19 definitely Jul 28 17:32:34 what gcc are people using to build the kernel these days? Jul 28 17:32:39 * RP -> back later Jul 28 17:32:51 polyonymous: I'll look into it later on Jul 28 17:33:00 crofton which kernel for wich arch? Jul 28 17:33:02 Crofton: 4.1.1 Jul 28 17:33:16 ARM Jul 28 17:33:22 hm Jul 28 17:33:29 specifially TI OMAP 5912 Jul 28 17:33:33 RP, sure, no problem, I don't mind fixing it by hand as I don't compile kernel every ten minutes ;) Jul 28 17:34:33 koen, and what is there to do about this libtool intervention? (trying to build freetype) Jul 28 17:38:00 Owell, I'll symlink .libs ;-) Jul 28 17:38:16 hi tmbinc Jul 28 17:41:42 koen: can you add woglinde or should i do? Jul 28 17:42:55 mickeyl: you may have the honours Jul 28 17:43:26 Crofton: 4.1.1 or 3.4.4 look reasonable Jul 28 17:44:32 koen: do you know whether a restart is necessary after changing the .monotonerc? i used to do that but i wonder whether it is really necessary Jul 28 17:45:11 * koen has no idea Jul 28 17:45:18 it can't really hurt Jul 28 17:50:11 woglinde: key added. confirmation mail sent. welcome on board. Jul 28 17:51:09 ~lart mysql Jul 28 17:51:10 * ibot explains, ever so gently, that if mysql doesn't give the channel more information, they can't help Jul 28 17:51:21 thanks mickeyl thanks koen Jul 28 17:51:23 hm Jul 28 17:51:53 schould the variable be named GCC_THREADS or something like THREADS_YES_NO Jul 28 17:52:10 for what? Jul 28 17:52:29 msp-430 has no threads Jul 28 17:52:35 GCC_USE_THREADS would be my taste. Jul 28 17:52:38 we already have Jul 28 17:52:41 PALMTOP_USE_THREADS Jul 28 17:52:42 or so Jul 28 17:52:49 okay Jul 28 17:52:50 could help to try to keep it a bit consistent Jul 28 17:53:00 thats why I am asking Jul 28 17:53:03 or wait Jul 28 17:53:08 * mickeyl still in vacation mode Jul 28 17:53:11 it actually is called Jul 28 17:53:18 Is there any fixed-point OpenGL implementation in OpenEmbedded? Jul 28 17:53:23 PALMTOP_USE_MULTITHREADED_QT Jul 28 17:53:29 mickeyl Jul 28 17:53:31 yeah, GCC_USE_THREADS would be ok Jul 28 17:53:31 aeh Jul 28 17:53:37 okay Jul 28 17:53:55 maybee I will change it if glibc needs something similar Jul 28 17:53:57 DataBeaver: not for X unfortunately Jul 28 17:54:04 Damn. Jul 28 17:54:07 DataBeaver: for qte2 we have something iirc Jul 28 17:54:36 mickeyl: doesn't Xorg include opengl-ES? Jul 28 17:54:37 I suppose it can't run on top of plain framebuffer? Jul 28 17:54:55 anyone feeling like updateing the autofoo.patch for mysql from .18 to .20 ? Jul 28 17:55:00 DataBeaver: no Jul 28 17:55:12 mickeyl I hate mysql Jul 28 17:55:20 but maybe I will do it later Jul 28 17:55:27 koen: does it? would be good - i have some apps that i can't add to OE because of a missing opengl Jul 28 17:55:34 woglinde: cool. i hate autofoo :) Jul 28 17:55:44 m4 makes me sick Jul 28 17:56:00 guess i need to come over this habit Jul 28 17:56:00 hi mickeyl Jul 28 17:56:00 mickeyl: opengl-ES is a fixed-point substandard of openGL Jul 28 17:56:02 hey chouimat Jul 28 17:56:05 but anyway, mesa is in OE Jul 28 17:56:21 Yeah, but it uses floats and is horribly slow Jul 28 17:56:39 Or at least I suppose it uses floats since it's so slow... Jul 28 17:56:54 03koen 07org.oe.dev * racf1f2ab... 10/ (1 conf/distro/openzaurus-unstable.conf): oz-unstable: use sane-srcdates Jul 28 18:08:21 ~praise RP for -dbg Jul 28 18:08:23 All hail RP for -dbg! Jul 28 18:44:01 damnit Jul 28 18:44:13 my computer died because of the heat Jul 28 18:44:18 bummer :/ Jul 28 18:44:34 hope it will rain here Jul 28 18:44:41 so I can test my patchtes Jul 28 18:44:47 22 C here Jul 28 18:44:48 bearable Jul 28 18:44:56 * mickey|dinner going to drink something Jul 28 18:44:57 l8er Jul 28 18:53:50 * CoreDump|home PC needs 4 case fans to stay somewhat cool Jul 28 18:54:28 what shows acpitool -j? Jul 28 18:54:37 what shows acpitool -e Jul 28 18:54:39 sorr Jul 28 18:54:40 y Jul 28 18:55:10 +mhentges@mhcln01:..p/deploy/images/ext2loop >acpitool Jul 28 18:55:10 bash: acpitool: command not found Jul 28 18:55:12 =) Jul 28 18:55:30 pff Jul 28 18:58:52 heh, extend acpi information? :) Jul 28 18:58:55 extended Jul 28 18:59:19 more crap then ever before! Jul 28 19:01:19 even more? Jul 28 19:01:28 even more - its extended! Jul 28 19:01:49 Ah, yeah. Jul 28 19:01:52 'evening Jul 28 19:01:59 'evening Jul 28 19:29:25 How do I specify a preferred version for u-boot? Jul 28 19:29:49 hi andersee Jul 28 19:30:02 crofton hm like gcc? Jul 28 19:30:40 well, I am not having much luck Jul 28 19:32:06 maybe because I can't spell PREFERED though .... Jul 28 19:32:44 :))) Jul 28 19:33:44 there are also files call u-boot in the uboot directory ... Jul 28 19:43:35 hi rwhitby Jul 28 19:44:30 http://www.dmochowski.pl/img/inne/bug3.jpg Jul 28 19:46:20 chouimat *g* Jul 28 19:46:37 bother u-boot-1.1.4 fails to build Jul 28 19:47:45 woglinde: http://www.openembedded.org/the-people-behind-openembedded Jul 28 19:49:53 koen msp430 Jul 28 19:50:02 *g* Jul 28 19:51:32 ah right Jul 28 19:51:38 my tv card has a 3400 Jul 28 19:58:17 woglinde: btw, I think I convinced OE to generate EABI stuff for armv4(t) Jul 28 20:00:24 koen wau Jul 28 20:00:30 hm Jul 28 20:00:37 how did you it? Jul 28 20:00:43 hm Jul 28 20:01:04 gcc-inital-cross does what fits msp430 Jul 28 20:01:50 woglinde: sed -i s:-Werror:: ${S}/configure Jul 28 20:02:08 what? Jul 28 20:02:18 how can that break eabi Jul 28 20:02:36 warning: armv4 doesn't support thumb interworking Jul 28 20:02:51 and warning are being treated as errors Jul 28 20:02:57 hm Jul 28 20:03:24 lol Jul 28 20:03:38 some is playing bagpipe here Jul 28 20:06:26 koen when can I test? Jul 28 20:07:36 I'm building for armv4t right now Jul 28 20:07:44 ah okay Jul 28 20:07:56 if that doesn't have armv5 insns I'll try armv4 Jul 28 20:08:21 simpad images Jul 28 21:00:31 coucou Jul 28 21:16:20 go ircbot Jul 28 21:16:37 aeh ciabot Jul 28 21:17:34 03woglinde 07org.oe.dev * r06a64510... 10/ (1 conf/bitbake.conf): Jul 28 21:17:34 conf/bitbake.conf: allow new field custom for TARGET_SYS Jul 28 21:17:34 some architecture has no sub-arch like msp430, so Jul 28 21:17:34 we should honor this because otherwise the configure-scripts Jul 28 21:17:34 for binutils and gcc won't work the right way Jul 28 21:18:16 yes Jul 28 21:56:31 hey kerwood, rp or anybody: im trying to use 'bitbake -s' to list the preferred versions of all packages but it doesnt work unless i build something - do i have something set wrong somewhere? or is there another way to do this? Jul 28 22:01:33 poushag: I don't know. did you read the bitbake manual? Jul 28 22:02:42 yes i read it - thats how i discovered the -s switch Jul 28 22:03:00 how do you list the preferred packages, kerwood? Jul 28 22:03:02 poushag: I've never used the -s switch :) Jul 28 22:03:31 well how do you know what packages versions are going to be in the stuff you build? Jul 28 22:03:48 I know the conf files too well ;-) Jul 28 22:04:33 poushag: yah, usually you build your desired package and find out after-the-fact Jul 28 22:04:39 well i wonder if the -s switch is broken Jul 28 22:05:08 jesus!!!! it works for me!! it ran off the screen! Jul 28 22:05:24 bitbake -s works for me too Jul 28 22:05:39 'night all Jul 28 22:05:41 yeah well for me it seems to just hang - let me try again Jul 28 22:05:53 poushag: make sure your env vars are correct Jul 28 22:05:54 what directory did you run it from? Jul 28 22:06:16 I suspect it will need a bit of thinking power to work that lot out with the current dependencies code Jul 28 22:06:21 my bbpath and bbfiles are correct - are there any other vars? Jul 28 22:07:20 I ran it from the samedirectory I build everything from -- the one which has the build, org.open..., and tmp directories Jul 28 22:07:44 goodnight RP Jul 28 22:08:06 is the ;handling bitbake files' step supposed to take a long time? - it seemed to me that since it's a cache it should be quick afterthe first time Jul 28 22:08:27 night rp Jul 28 22:08:43 now that you mention it, it seemed to take longer than usual to get going Jul 28 22:09:03 bitbake -s works here too Jul 28 22:09:38 bitbake -s is running right now - but reading the cache (if thats what 'handling files' means) is taking at least 10 minutes for 3912 files Jul 28 22:09:53 yes the first time Jul 28 22:09:53 (using .dev branch) Jul 28 22:10:16 the cache is the cache Jul 28 22:10:18 then Jul 28 22:10:19 the Jul 28 22:10:20 seems like every time to me Jul 28 22:10:33 which version of bitbake do you use? Jul 28 22:10:40 i mean when does the cache get cleared out? Jul 28 22:10:58 1.4.2 - just started using it this week Jul 28 22:10:59 when you work on your local.conf Jul 28 22:11:13 did your clear your tmp? Jul 28 22:11:24 bitbake 1.4.2 works here too Jul 28 22:11:54 if i update mtn will that clear the cache? Jul 28 22:12:27 yes because you have to download a nee oe.mtn Jul 28 22:13:08 ok that must be what happened this time - does the cache also get cleared when i log out my userid? Jul 28 22:13:41 no Jul 28 22:13:47 ok good Jul 28 22:13:52 remove the cache clears the cache Jul 28 22:14:15 maybe if they timestamp is too long the cache is renewed Jul 28 22:14:23 but I dont know excatly Jul 28 22:14:35 doing mtn update - does that clear all of tmp? or just cache? Jul 28 22:14:41 never looked in the source-files of this part Jul 28 22:14:54 it can make the cache invalid Jul 28 22:14:59 but not tmp Jul 28 22:16:03 alright - im still trying to understand all this oe stuff - i really want to disable builds of all the locale stuff i will never use Jul 28 22:16:14 can that be done in a fairly simple way? Jul 28 22:16:40 I am forget 5 seconds after someone post here Jul 28 22:16:44 all the time Jul 28 22:16:52 maybee its in the handbook Jul 28 22:16:54 let me see Jul 28 22:17:30 poushag: not feeling so cosmopolitan? Jul 28 22:17:36 :) Jul 28 22:18:34 heheh - well i know a little espanol and a few words of tagalog but i cant see why i should have to clog up my collie with it Jul 28 22:19:06 hm the locales only get build Jul 28 22:19:12 but not installed Jul 28 22:19:21 sorry cant find it Jul 28 22:19:23 actually, it's not clear to me that you have to install that baggage. I thought ipkg did that checking for you Jul 28 22:20:32 well my build machine is pretty slow too - so i dont want to build stuff i dont need - even if it wont get installed Jul 28 22:33:01 nite all Jul 28 22:33:28 goodnight woglinde Jul 29 00:24:43 * kergoth plays around with crazy bitbake/oe ideas Jul 29 00:32:31 kergoth: like a unified development environment for embedded apps, against all contrary forces? Jul 29 00:33:00 When do we start singing hosanna? Jul 29 00:33:21 :) Jul 29 01:00:01 do13: ping Jul 29 01:02:20 Kerwood_: all sorts of random things, things i wanted to see done differently back when we first implemented them :P Jul 29 01:14:26 kergoth: *sigh* the price of collaboration... Jul 29 01:14:26 kergoth: the trade is fair, yet there are still regrets, I suppose... Jul 29 01:36:49 03justinp 07org.oe.dev * r32f06dad... 10/ (339 files in 62 dirs): x11: change virtual/x11 to virtual/libx11 Jul 29 01:41:07 JustinP: arrrg.. and I just pulled :( **** ENDING LOGGING AT Sat Jul 29 02:59:56 2006