**** BEGIN LOGGING AT Wed Nov 22 02:59:58 2006 Nov 22 03:55:50 03lenehan 07org.oe.dev * raf60b15c... 10/ (1 packages/lilo-sh/lilo-sh_21.bb): Nov 22 03:55:50 lilo-sh: Lilo sh is an sh3/sh4 version of the lilo bootloader, so add a Nov 22 03:55:50 COMPATIBLE_HOST to indicate that this only works for sh3/sh4. Nov 22 05:10:44 03lenehan 07org.oe.dev * rfe03d95c... 10/ (7 files in 5 dirs): Nov 22 05:10:44 arpwatch 2.1a15: Add the latest version, 2.1a15. No source for older Nov 22 05:10:44 versions is carried on the distribution site. Nov 22 05:10:48 03lenehan 07org.oe.dev * r5344c304... 10/ (4 files in 3 dirs): (log message trimmed) Nov 22 05:10:48 arpwatch 2.1a15: Cleanups for arpwatch: Nov 22 05:10:48 - mark configuration so they don't get wiped on upgrade Nov 22 05:10:48 - don't pass -p, for nopromisc, as an option since that isn't Nov 22 05:10:51 supported (must be a debian patch which is not included in OE) Nov 22 05:10:53 - remove RUNAS which was used to specify a non-root user to run as Nov 22 05:10:55 since it isn't support (must be a debian patch which is not Nov 22 07:53:16 hi :) Nov 22 08:00:19 good morning all Nov 22 08:00:43 yo koen Nov 22 08:00:50 morning koen :) Nov 22 08:01:00 hi Nov 22 08:01:15 time to go to bed :) Nov 22 08:04:00 morning Nov 22 08:04:09 somebody from japan in the channel? Nov 22 08:06:08 morning all Nov 22 08:07:19 RP: patch for bitbake: http://rafb.net/paste/results/eddkgG59.html Nov 22 08:33:05 03mickeyl 07org.oe.dev * rd4083ecb... 10/ (5 files in 3 dirs): gnet: add cvs version that builds w/ modern compilers. remove outdated 2.0.7 version Nov 22 08:33:23 03mickeyl 07org.oe.dev * rb6527d93... 10/ (1 packages/refdbg packages/refdbg/refdbg_1.2.bb): add refdbg, a refcount debugger for GObject based programs Nov 22 08:37:46 mickey|grocery: good morning! Nov 22 08:38:20 morning. i'm darn sleepy. guess i will return for another nap Nov 22 08:38:27 return to bed, that is Nov 22 08:39:36 either that or brew coffee Nov 22 08:39:39 * mickey|grocery ponders Nov 22 08:39:42 coffee Nov 22 08:39:47 03koen 07org.oe.dev * r3adbc0be... 10/ (1 packages/git/git_snapshot.bb): Nov 22 08:39:47 git: RDEPEND on perl: Nov 22 08:39:47 root@h2200:/data# git clone git://git.cairographics.org/git/cairo Nov 22 08:39:47 /usr/bin/git-clone: line 325: perl: command not found Nov 22 08:42:36 hey tomas_f Nov 22 08:43:25 morning tomas_f Nov 22 08:43:32 * mickeyl brews coffee instead of taking another nap Nov 22 08:46:03 good morning everyone Nov 22 08:58:07 seems some package inside gpe-image pulls in some opie qt stuff .. Nov 22 08:59:23 rob_w: you probably have wrong preferred providers for sdl and stuff Nov 22 08:59:39 oh .. hmm lemme check Nov 22 09:00:09 03koen 07org.oe.dev * rfae04308... 10/ (1 packages/git/git_snapshot.bb): git: also rdepend on perl-module-file-path Nov 22 09:00:11 why does someone need to provide those .. Nov 22 09:00:35 because opie stinks and has hacked versions for common libraries Nov 22 09:00:46 s/opie/qte/ Nov 22 09:01:03 a sane distro should set those for you :) Nov 22 09:01:04 * rob_w moves his nose Nov 22 09:01:18 too bad that gpe is influenced by this Nov 22 09:01:53 well .. my 2core opteron gets over this quite quick Nov 22 09:02:44 can you tell which packages? Nov 22 09:04:08 it got over it ., but it felt that i have libsdl-qpe was build Nov 22 09:04:16 lemme grep for that Nov 22 09:04:23 *nod* then it's what koen suspected Nov 22 09:04:58 yeah, kdrive depends on sdl now to build Xsdl Nov 22 09:05:34 ya know, going from QtE to X will make at least one thing easy Nov 22 09:05:36 SDL! Nov 22 09:05:39 no more hacking Nov 22 09:05:46 a relief. Nov 22 09:05:47 so probably some kdrive library sucks sdl and i miss the prefferred thingi for that Nov 22 09:06:16 mickeyl, oh .. that sound nice .. will qte be dead soon ? Nov 22 09:06:32 no no Nov 22 09:06:46 i think there will be a usage for Qtopia Core. Nov 22 09:07:13 I'm just babbeling personally here Nov 22 09:07:30 since I'm going to use X for a while Nov 22 09:07:30 and QtE gave me headaches w/ SDL Nov 22 09:07:58 mickeyl: did you see hrw|gone's comment about his webpad? Nov 22 09:07:58 koen: no, on his blog? Nov 22 09:07:58 PREFERRED_PROVIDER_libsdl = "libsdl-x11" Nov 22 09:07:58 should do the trick Nov 22 09:08:13 mickeyl: no, on IRC: "XGA X display, but no XGA framebuffer, bye bye opie" Nov 22 09:08:21 heh Nov 22 09:08:36 how's that? X display, but no framebuffer? Nov 22 09:08:36 it has a gfx card? Nov 22 09:09:00 no, just a chip without a decent kernelfb driver Nov 22 09:10:17 or to put it in a different way "The kernel doesn't support opie" Nov 22 09:11:08 heh Nov 22 09:11:17 not much of a loss though Nov 22 09:11:30 direct framebuffer access is way too slow for large fb's Nov 22 09:11:31 it highlights the fallacy of insisting on fb Nov 22 09:12:00 yo mithro Nov 22 09:14:34 Hi koen, mickeyl, all. Nov 22 09:14:40 hey den-ros Nov 22 09:14:40 i'd like to see the numbers... Nov 22 09:15:19 "fully supported by Linux 2.6 (except the sound part) " Nov 22 09:15:43 how to contradict yourself in the same sentence part 1, presented by mr Aznar Nov 22 09:16:12 eugenebiz: did you catch my patch? Nov 22 09:16:33 no, where did post it? Nov 22 09:17:09 on a less off-topic channel :) Nov 22 09:20:16 ljp: o-hand have some screen benchmarks in their svn. It'd be really interesting to see how a qte backend for those testcases. they're at http://svn.o-hand.com/view/misc/trunk/fstests/ Nov 22 09:20:22 yo mickeyl Nov 22 09:20:26 s/how a qte backend/how a qte backend performs/ Nov 22 09:20:47 hi all, is there any know issues with name resolution in libc2.5 > Nov 22 09:20:50 how about qtopia core instead Nov 22 09:22:21 fair enough Nov 22 09:23:13 haha in devshell script : alias ./configure=oe_runconf Nov 22 09:23:14 <= :D can't work on my shell Nov 22 09:23:35 sound like you need a better shell Nov 22 09:23:50 Genesis: bitbake -c devshell usually works a lot better Nov 22 09:23:57 otherwise i'm happy it's fixed Nov 22 09:24:50 i don't know if it has been fixed the devshell class Nov 22 09:25:04 mickeyl, : hu ? Nov 22 09:30:08 any X11 guru's in da house? Nov 22 09:30:33 how can I get a Window* from a GdkWindow? Nov 22 09:30:36 morning Nov 22 09:30:37 yo hrw Nov 22 09:31:05 mickeyl: this webpad has 4M gfx card with own X11 driver Nov 22 09:31:14 cool Nov 22 09:31:38 and TS use another X11 driver Nov 22 09:31:44 mickeyl, display = gdk_x11_get_default_xdisplay(); Nov 22 09:32:37 mickeyl: there is one kernel driver for microtouch TS but I did not yet checked does it works at all on this device Nov 22 09:33:41 rob_w: that gets me a Display*, but how to get a Window* ? Nov 22 09:34:06 mickeyl: have you looked at adding this benchmark to OE: http://folks.o-hand.com/jorn/pango-benchmarks/ ? Nov 22 09:34:35 koen: no -- first time i see this page Nov 22 09:35:16 good morning Nov 22 09:36:01 mickeyl, oh right .. Nov 22 09:37:03 koen: say author about binutils-symlinks ;D Nov 22 09:37:09 "and create a symlink from /usr/bin/arm-linux-objdump to /usr/bin/objdump Nov 22 09:37:11 " Nov 22 09:37:15 rob_w: end result is i want to use gdk_pixbuf_xlib_get_from_drawable to get a Pixbuf out of a complete rendered top level window, but that needs a Window* or a Drawable* as param. Nov 22 09:40:30 hrw|work: ok, I sent a patch to jorn :) Nov 22 09:46:02 yeeha Nov 22 09:46:17 gdk_x11_window_get_xid(GTK_WIDGET(widget)->window) Nov 22 09:46:25 this gets me my Window* Nov 22 09:46:47 right Nov 22 09:47:34 sometimes just reading header files helps Nov 22 09:48:05 hey lrg Nov 22 09:48:05 hey all Nov 22 09:48:26 * lrg is at Frankfurt airport on his way to Bangalore Nov 22 09:48:39 lrg: so you are close to mickeyl Nov 22 09:48:47 koen: yep. Nov 22 09:48:50 lrg: hi Nov 22 09:48:52 ping mickeyl Nov 22 09:48:54 hi Liam Nov 22 09:49:01 morning all Nov 22 09:49:03 lrg: pöng Nov 22 09:49:11 RP: hi töö Nov 22 09:49:18 yo RP Nov 22 09:49:19 hey RP, mickeyl Marcin, Florian and Koen Nov 22 09:49:30 lrg: when is your plane taking off? Nov 22 09:49:58 mickeyl: about 1 hour. Nov 22 09:50:04 ah that's too close Nov 22 09:50:10 :-) Nov 22 09:50:17 i can be at the airport in 30 mins Nov 22 09:50:24 but that's about time you start boarding Nov 22 09:50:35 03rpurdie * r699 10bitbake/lib/bb/fetch/git.py: git.py: Fix typo (from koen) Nov 22 09:50:59 i'm packaging some plugin for mozilla family Nov 22 09:51:02 hey thos Nov 22 09:51:10 hullo Nov 22 09:51:15 since they could compile against xulrunner , firefox , seabird , mozilla ... Nov 22 09:51:34 hmm... our kernel for the 770 doesn't want to compile, is this a known issue? Nov 22 09:51:50 should i introduice a PREFERRED_PROVIDER_virtual/mozilla ? Nov 22 09:51:52 mickeyl: I'm coming back on tghe 30th Nov 22 09:53:54 and in the plugin recipes , DEPENDS = "virtual/mozilla" and switch case on the right to get the good header ? Nov 22 09:55:45 * lrg -> take off Nov 22 09:57:42 morning all Nov 22 09:57:48 hi ade Nov 22 09:58:09 hey ade|desk Nov 22 10:09:05 florian_kc: drivers/i2c/chips/tlv320aic23.c ? Nov 22 10:09:47 koen: this one (fixed it) and usb/core/hub.c iirc Nov 22 10:10:25 florian_kc: let me commit a newer kernel (with the same bug) so you can add a patch :) Nov 22 10:11:21 koen: not that i have spare time but i'd like to replace my familiar based image for the 770 with a more recent Angstrom one. Nov 22 10:11:49 florian_kc: right, but it would be nice to have your fixes in OE Nov 22 10:12:10 03koen 07org.oe.dev * r384d70a3... 10/ (5 files in 2 dirs): linux-nokia770: update to osso26 (from poky) Nov 22 10:13:27 ok :-) Nov 22 10:14:07 Genesis: I have been busy with the recipe for mplayer plugin Nov 22 10:14:28 Genesis I run into some problems with faad and faac which are missing Nov 22 10:16:42 gerwinin: both faac and faad should be in OE Nov 22 10:19:24 gerwinin, : i told you i had quiet finish it Nov 22 10:19:29 and you remake it .. Nov 22 10:19:42 gerwinin, : have you corrected the FF 1.5 bb ? Nov 22 10:20:01 Genesis sorry didn't know you almost finished it Nov 22 10:20:07 adding a do_stage for cp the include directory ? Nov 22 10:20:13 Genesis : no Nov 22 10:20:27 how do you build mplayerplugin against it so ? Nov 22 10:20:38 Genesis can you post your files so I can try as well with your files ? Nov 22 10:21:04 it's not finish but if you want i pastit somewhere Nov 22 10:21:12 okay Nov 22 10:22:19 i've introduice virtual/mozilla Nov 22 10:23:34 * gerwinin will vote for a moment brb Nov 22 10:24:41 http://code.bulix.org/54h1qb-24201?raw Nov 22 10:24:52 => firefox Nov 22 10:25:45 http://code.bulix.org/drd9h7-24202?raw Nov 22 10:25:50 => mplayerplugin Nov 22 10:26:36 gerwinin, : for this last one , i've pb with the install command from OE Nov 22 10:26:37 Genesis: s:/etc:${sysconfdir}:g Nov 22 10:27:05 getting firefox 2.0 working would be nice Nov 22 10:27:17 no time to do that for the moment Nov 22 10:27:49 I know Nov 22 10:27:58 ff is a nightmare to package :) Nov 22 10:29:45 since i update this morning , bitbake refuse to see my glibc ipk , i rebuild them and same issue Nov 22 10:31:57 otherwise , happy to see mx21 support in kernel comes in same time than MS zune ^^ Nov 22 10:33:49 * koen hugs insane.bbclass Nov 22 10:34:19 * gerwinin voted Nov 22 10:34:29 Koen: did you vote ? Nov 22 10:34:42 gerwinin: not yet, will do later today Nov 22 10:34:57 Koen: who will get your vote ? Nov 22 10:35:29 the SP Nov 22 10:36:00 Koen: I voted for them years ago this I voted cda Nov 22 10:36:40 Koen: I was doubting this time between sp and cda Nov 22 10:41:21 03koen 07org.oe.dev * r6d82f3a7... 10/ (1 packages/librsvg/librsvg_2.16.0.bb): librsvg: add 2.16.0 Nov 22 10:42:38 I'm trying to do a task-bootstrap and it's failing on poodle-kernel-2.4-embedix-2.4.18-rmk7-pxa3-embedix-r46 Nov 22 10:43:29 apparently arm-linux-gcc-2.95 can't be found... but isn't this needed for other packages? Nov 22 10:43:41 ~zauruskernels Nov 22 10:43:45 methinks zauruskernels is by default Zaurus used 2.4.18 or 2.4.20 - now we have 2.6 working on all clamshells and work is progressing for tosa, poodle, collie. More info in OpenZaurus wiki: http://wiki.openzaurus.org/Kernel_Information Nov 22 10:45:47 zezom: http://www.openembedded.org/wiki/ZaurusKernels Nov 22 10:46:25 koen, thanks... I still find it weird that it didn't need it until now... Nov 22 10:46:43 only crappy software needs gcc 2.95.3 Nov 22 10:48:24 gerwinin, : so my bb ? :) Nov 22 10:49:26 Genesis I am trying it out now Nov 22 10:53:55 gerwinin, : ho i forget to say , i've modify the mplayerplugin files Nov 22 10:54:16 Genesis what did you change ? Nov 22 10:54:29 the header have some #include "foo.h" i modify to #include Nov 22 10:54:53 easy , if a file failed on no such directory , replace it Nov 22 10:55:24 Genesis: sounds like a missing -I Nov 22 10:55:49 03koen 07org.oe.dev * r311a862d... 10/ (1 packages/librsvg/librsvg_2.16.0.bb): librsvg: add cairo dep Nov 22 10:55:57 03koen 07org.oe.dev * r83d98662... 10/ (1 packages/cairo/cairo_git.bb): cairo: add floating git version Nov 22 10:56:26 koen : noticed :) Nov 22 11:11:23 hrw|work: The spca5xx bb I saw listed on your email, what was the problem with it? Nov 22 11:13:53 koen: its in... but deploy still fails Nov 22 11:15:03 koen: looks like the deploy/glibc/images/nokia770 subdir does not get created Nov 22 11:15:08 03florian 07org.oe.dev * r959a6162... 10/ (4 files in 3 dirs): linux-nokia770: Add patches to fix compile. Nov 22 11:16:35 florian_kc: mtn pull :) Nov 22 11:18:07 03koen 07org.oe.dev * r0e31d33e... 10/ (1 packages/linux/linux-nokia770_2.6.16-osso26.bb): linux-nokia 770: fix deploy Nov 22 11:18:25 heh Nov 22 11:18:32 koen: thanks Nov 22 11:27:35 morning Nov 22 11:28:31 hi XorA Nov 22 11:32:03 NOTE: Task failed: Fetch failed: http://mxhaard.free.fr/spca50x/Download/spca5xx-20060202.tar.gz Nov 22 11:32:06 lardman: see? Nov 22 11:32:09 ERROR: Build of /home/hrw/devel/oe/org.openembedded.dev/packages/spca5xx/spca5xx_20060202.bb do_fetch failed Nov 22 11:42:23 re Nov 22 11:49:22 hrw|work: Ah, fair enough. Sorry for the silly question but I don't have a build machine attached to the network atm Nov 22 11:50:37 hrw|work: I can give you an alternative url/version for that file: http://mxhaard.free.fr/spca50x/Download/spca5xx-20060501.tar.gz Nov 22 11:51:06 hrw|work: Though that's for <=2.6.11 Nov 22 11:51:19 i am still getting aligment traps , although i use the same conf switches then hte h2200 ? any hints ? Nov 22 11:51:23 hrw|work: there;'s a new version (different name) for >2.6.11 Nov 22 11:51:38 s/<=/< Nov 22 11:55:44 can someone get me ahead on this alignment things and how to work them out ?? Nov 22 11:56:15 lardman: ok. I will check Nov 22 11:56:47 lardman: w8.. 20060501 != 20060202 Nov 22 11:57:15 hrw|work: I know, needs a version upgrade Nov 22 11:57:44 lardman: simple mtn rename is enough? Nov 22 11:57:55 hrw|work: Yes, for the time being Nov 22 11:58:00 lardman: and what with 20050419 version? drop? Nov 22 11:58:07 hrw|work: Will eventually require a kernel version check too Nov 22 11:58:17 hrw|work: Yep, drop that one as there's no source anymore Nov 22 11:58:20 ok Nov 22 11:58:24 thanks Nov 22 12:02:50 hi likewise Nov 22 12:02:59 hi all Nov 22 12:04:56 hey likewise Nov 22 12:06:13 koen: hoi Nov 22 12:09:06 so it seems hh.org kernels have a patch for alignment traps Nov 22 12:09:16 rob_w: no they haven't Nov 22 12:09:16 can someone point me to the that patch ?? Nov 22 12:09:42 rob_w: alignment trap catching is in every arm kernel since ages Nov 22 12:10:21 but the defconfig of a hh.org kernel shows a CONFIG_ALIGNMENT_TRAP=y Nov 22 12:10:36 so? Nov 22 12:10:38 hmm ok i do use the compulab kernel .. which then misses this .. crap Nov 22 12:11:02 koen: good thing in sending fetcherrorlist is that some people wake and promise to send updates Nov 22 12:11:17 hrw|work: that was indeed a pleasant surprise Nov 22 12:11:26 rob_w: in userspace or in kernel? Nov 22 12:11:31 hrw|work: sending it that list and saying "it's going to be in removal.txt" Nov 22 12:11:48 03hrw 07org.oe.dev * rb51d22a8... 10/ (6 files in 2 dirs): spca5xx: updated to 20060501, dropped 20050419 - both on lardman request Nov 22 12:13:24 even though I'm one of ones waking up to send an update, I'm very glad to see OE being kept clean :-p Nov 22 12:13:25 koen: and responce: please, dont! I promise that I will workout it. Nov 22 12:13:55 suihkulokki, userspace as waht i see Nov 22 12:15:11 someone remember how to force mtn to use prefered merge tool? there was something like MERGETOOL Nov 22 12:19:06 hi schurig Nov 22 12:19:15 hrw|work: hi! Nov 22 12:19:51 suihkulokki, any comments ? Nov 22 12:20:35 rob_w: for userspace, you can set the aligment trap behaviour in /proc/cpu/alignement Nov 22 12:20:48 i do have 3 already .. Nov 22 12:21:07 but yet i get filled with messages Nov 22 12:22:00 rob_w: 3 is "fix _and_ warn" Nov 22 12:22:55 rob_w: echo 5> to get get cores of the files (to FIX the problem) Nov 22 12:23:30 rob_w: echo 2> if you just want to get rid of the messages for now Nov 22 12:25:06 well i do notice a speed lag also .. will this be fixed by "2" also ? Nov 22 12:25:25 no Nov 22 12:27:25 suihkulokki, hmm Nov 22 12:29:49 so this is not a soultion .. what is my main problem .. do i compile the apps still with wrong compiler switches ?? Nov 22 12:30:09 someone said a pxa270 doenst have the alignment thingi Nov 22 12:30:30 only arm11 can do unpunished unaligned access Nov 22 12:30:40 and xscale is arm9(ish) Nov 22 12:31:07 food Nov 22 12:32:08 so what ? whats wrong with me ? Nov 22 12:37:39 brb Nov 22 12:44:26 rob_w, ; i've 3 alignement errors too Nov 22 12:44:40 and all works Nov 22 12:45:00 the default TARGET_OS for arm eabi is linux-gnueabi right ? Nov 22 12:45:17 i get tons of them on all apps Nov 22 12:45:25 almost Nov 22 12:45:37 vlo: or linux-uclibcgnueabi Nov 22 12:47:32 koen: ok ... and on what packets does this affect ? because I've just seen that I've been working for ages with only TARGET_OS=linux (shame on me!). Should I rebuild everything from scratch ? Nov 22 12:47:33 rob_w: I missed beginning of this, are you getting alignment errors for all gtk+ apps? Nov 22 12:47:55 vlo: OE should rebuild automagically Nov 22 12:48:58 Crofton: 're u here? did you finish your tests yesterday? Nov 22 12:48:59 XorA, seems so Nov 22 12:49:28 they built, haven't flashed them though Nov 22 12:49:38 rob_w: I get that with some configurations, when the moon has an apple on its head, a complete rebuild always seems to cure it Nov 22 12:50:48 oh Nov 22 12:50:59 ahm Nov 22 12:52:37 hmm i did rebuild already as i though i was missing IPKG_EXTRA_ARCHS Nov 22 12:53:21 rob_w: I see the problem, occasionally, but never managed to track down why it hits some images and not others Nov 22 12:54:22 XorA: I am also seeing alignment traps with GTK apps. Still running glibc2.3 ATM. Nov 22 12:55:01 cbrake: I see it on openzaurus-unstable builds Nov 22 12:55:16 cbrake: sirfred noticed fontconfig does unaligned access Nov 22 12:55:26 cbrake: don't with fontconfig version, though Nov 22 12:57:57 looking at the kernel code, it looks like a message is always printed if an unaligned access occurs. Nov 22 12:59:06 03koen 07org.oe.dev * r6bf51b9d... 10/ (1 packages/cairo/cairo_git.bb): cairo git: also compile performance suite Nov 22 13:04:46 hrm... which package provides xextensions.pc` Nov 22 13:05:21 libxext? Nov 22 13:05:39 xextproto? Nov 22 13:07:54 xextproto contain xextproto.pc Nov 22 13:08:06 ok Nov 22 13:08:06 bitbake shell is broken in trunk ;( Nov 22 13:09:19 there is xext.pc, but not xextensions.pc Nov 22 13:09:41 debian does not contain xextensions.pc now Nov 22 13:09:51 florian_kc: what needs xextensions.pc? Nov 22 13:10:01 florian_kc: I compiled a gpe image without that Nov 22 13:10:13 koen: Xomap Nov 22 13:10:14 and iirc xextensions is depecrated with the switch to Xorg Nov 22 13:10:25 florian_kc: ah, right, Xomap needs to get updated Nov 22 13:12:24 koen: why cairo_git is preferred? Nov 22 13:12:37 hrw|work: preferred? Nov 22 13:12:50 what is Xomap? Nov 22 13:13:00 koen: it is highest and does not have DEF_PREF -1 Nov 22 13:13:14 hrw|work: oops, copied over a local diff from 1.3.2 Nov 22 13:13:25 koen: ok, that will have to wait... Nov 22 13:13:38 unless someone has the update ready Nov 22 13:13:54 Crofton: kdrive for omap cpus Nov 22 13:13:59 brb, food Nov 22 13:14:18 or rather for the omapfb + epson chip nokia uses Nov 22 13:14:32 ah Nov 22 13:14:36 can anyone help me be less ignorant... the "getting started" on the wiki makes reference to a "build" folder. well... it dosn't exist. do I need to create it? & just put the file in there? or is there something I'm missing? Nov 22 13:15:19 03koen 07org.oe.dev * r2735a177... 10/ (1 packages/cairo/cairo_git.bb): cairo git: remove local diff Nov 22 13:16:08 TheCompWiz: build folder is the one which you use for build Nov 22 13:16:24 TheCompWiz: its up to you how you will use it. Nov 22 13:16:28 ok.... meaning kernel build? Nov 22 13:16:36 OE build Nov 22 13:16:48 and that reminds me about my post which need update Nov 22 13:17:00 ok.... so... um... what is the org.openembedded.dev folder for? Nov 22 13:17:13 http://www.hrw.one.pl/2005/10/31/my-openembedded-enviroment/ Nov 22 13:17:49 rob_w: following suihkulokki suggestion of setting /proc/cpu/alignment to 5 and then running my app through gdb, I get the following: alignment trap: 0x40668604 in FcPatternNeededBytes () from /usr/lib/libfontconfig.so.1 Nov 22 13:18:14 rob_w: I'm only see the alignment issues on app startup. Nov 22 13:19:16 * TheCompWiz is still lost. where do I get the "build" folder from? Nov 22 13:19:27 make it Nov 22 13:19:31 how? Nov 22 13:19:35 mkdir? Nov 22 13:19:46 just make an empty directory... with nothing in it? Nov 22 13:20:00 with conf/local.conf Nov 22 13:20:01 is the bulid folder the folder for "changes"? Nov 22 13:21:31 Hi! Nov 22 13:21:36 hey psokolovsky Nov 22 13:21:38 hi, all! Nov 22 13:21:49 hey slapin_nb Nov 22 13:22:20 TheCompWiz, why don't you just do as written in GettingStarted, and get build going? maybe you get answers yourself then... ;-) Nov 22 13:22:26 Hi koen! Nov 22 13:25:25 psokolovsky... I have done... and no mention of creating the build directory Nov 22 13:28:49 cbrake, good point Nov 22 13:29:21 i will investigate those apps also with gdb .. in a while .. need to fix something else now .. Nov 22 13:35:34 morning Nov 22 13:35:54 hey chouimat Nov 22 13:36:30 hey zecke Nov 22 13:36:35 hi zecke Nov 22 13:36:47 wow, i'm happy not to use subversion for OE Nov 22 13:36:57 I just did a merge and it messed up the .svn dir so badly Nov 22 13:37:11 svn is evil Nov 22 13:37:24 in a merge the same file was deleted and added Nov 22 13:37:35 and this makes subversion fuck up your working copy Nov 22 13:37:35 svn can't even *check in* an OE tree when you use bdb as backend Nov 22 13:38:13 zecke: conflicts stay in files instead of being resolved.. Nov 22 13:39:02 hrw|work: no, again it fucked up .svn directory Nov 22 13:39:15 I can't commit, can't update, can't revert, just fucked up Nov 22 13:39:19 ah... Nov 22 13:39:25 ivil^2 Nov 22 13:45:54 ibot: dict compagnon Nov 22 13:45:57 could not find definition for compagnon Nov 22 13:46:06 ibot: dict companion Nov 22 13:46:17 pfft.. dosn't listen to me ;) Nov 22 13:46:27 Dictionary 'companion' (2 of 7): a traveler who accompanies you ;; one paid to accompany or assist or live with another. Nov 22 13:52:27 hrw|work: Is the angstrom feed browser source available anywhere online? Nov 22 13:52:48 RP: it's in OE .dev :) Nov 22 13:52:50 RP: contrib/feed-browser/ Nov 22 13:52:54 RP: under contrib Nov 22 13:54:11 hi Nov 22 13:54:27 hey CoreDump|home Nov 22 13:54:39 zecke, when I have had svn issues with files that svn can't fix, I delete the files in confict and then update my svn directory. This downloads the last checked in copy of the conflicted files. Nov 22 13:55:15 zecke, its not how you should have to do things but it works. Nov 22 13:55:42 Gerrath: read above the .svn/entries is corrupted Nov 22 13:55:50 Gerrath: you can not svn ci, revert, up Nov 22 13:56:00 zecke, sorry, I missed that. Nov 22 13:56:18 zecke: What I'd do then is a checkout into a new directory; find olddir/ -type d -name .svn -exec rm -rf {} \;, then rsync -r olddir newdir Nov 22 13:56:43 NAbyss: this is on merging trunk into a branch Nov 22 13:56:56 NAbyss: I NEED TO MERGE, no way of rsyncing Nov 22 13:57:01 zecke: Ah, smeg.. Nov 22 13:57:04 koen, hrw|work thanks :) Nov 22 13:57:08 sure I can apply the patch by hand Nov 22 13:57:14 or unfuck .svn dirs by hand :) Nov 22 13:57:23 RP: poky will get it too? Nov 22 13:59:26 hrw|work: quite possibly :) Nov 22 14:00:52 re Nov 22 14:01:41 does anyone know if there is a machine conf for the toshiba e740... or what spyro used in the past? Nov 22 14:02:43 TheCompWiz: spyro does not use OE Nov 22 14:03:07 o.O ... the old version if his work I had... (precompiled of course) had the OE logo during boot... Nov 22 14:05:00 hrw|work... just for curiosity... how do you know he dosn't use OE? Nov 22 14:21:52 bugzilla is slow today ? Nov 22 14:27:25 Crofton: did you check the glibc patch yesterday? Nov 22 14:35:42 http://www.hrw.one.pl/2006/11/22/my-openembedded-environment-ii/ Nov 22 14:36:30 some people will find that usefull ;D Nov 22 14:37:07 Jin^eLD, I'll go do that soon Nov 22 14:37:14 need to walk in to work first :) Nov 22 14:41:32 hi [cc]smart Nov 22 14:44:21 cbrake, XorA , i also had libfontconfig doing the alignment traps ,, so i upgraded from 2.3.91-r3 to 2.3.95-r2 and now it works !! Nov 22 14:45:46 rob_w: ah, I thought 2.3.95 had becvome the default Nov 22 14:45:53 Crofton: I see :) have you made a new setup for your environment? Nov 22 14:46:01 I still had no time to finish mine Nov 22 14:46:01 rob_w: Ill bet I have a PREFERRED_VERSION set locally then Nov 22 14:46:25 hrw|work: that's the link I was following yesterday ;) Nov 22 14:46:40 hrw|work: you have a local.conf per build environment, right? Nov 22 14:46:52 not in my local.conf XorA Nov 22 14:47:11 not yet, I will need to once I get my EFIKA board :) Nov 22 14:47:21 :> Nov 22 14:48:12 Jin^eLD: this is NEW post Nov 22 14:48:18 oh! Nov 22 14:48:51 cool Nov 22 14:49:07 what is what is rootfs_ipk_log_check? Nov 22 14:49:31 rob_w: excellent! Nov 22 14:50:00 i am still not happy with the performance Nov 22 14:51:03 brb Nov 22 14:55:43 rob_w: specifically, what performance problems are you having? Nov 22 15:02:53 cbrake, hard to say .. need more tests then i tell u Nov 22 15:05:58 RP: ping Nov 22 15:06:50 zecke: pong Nov 22 15:07:45 RP: 1) I will do regular poky builds again (at least until january) Nov 22 15:07:45 RP: 2) I miss flex 2.5.31, could you put it on your folks site? Nov 22 15:07:45 zecke: 1) cool, 2) ok :) Nov 22 15:08:17 nevermind, probably just another bad sf.net mirror Nov 22 15:08:28 zecke: ah :-/ Nov 22 15:08:58 ~lart bloxx Nov 22 15:08:58 * ibot whips out a hot clue gun and makes sure that bloxx is stuck to the floor Nov 22 15:12:23 RP: can it be that ctrl-c gets delivered differently due the new runqueue? Nov 22 15:12:52 zecke: http://www.openembedded.org/bonsai/query/description/ctrl/ Nov 22 15:12:56 zecke: Yes, I altered that in the last version Nov 22 15:13:22 koen: libsoup 2.2.98 released: no API change; XML-RPC minor syntax support addition, memory leak fix, cleanups Nov 22 15:16:16 * koen looks Nov 22 15:16:21 zecke: how's your FLUTE stuff doing btw? Nov 22 15:16:41 giel: I won't publish it due nokia :) Nov 22 15:16:46 bweuh :( Nov 22 15:16:57 too bad Nov 22 15:17:05 are you allowed to tell why? ;) Nov 22 15:17:22 giel: well, take a look at the mcl flute site Nov 22 15:17:31 k Nov 22 15:18:06 03koen 07org.oe.dev * red757515... 10/ (1 packages/gnome/libsoup_2.2.98.bb): libsoup: update to 2.2.98 Nov 22 15:18:07 giel: http://planete-bcast.inrialpes.fr/rubrique.php3?id_rubrique=2 Nov 22 15:18:57 thanks Nov 22 15:21:13 yes, the claims of that patent pretty much cover flute Nov 22 15:21:58 i'm not so sure there's no prior art... Nov 22 15:22:06 giel: I wonder how much alc is covered Nov 22 15:22:17 giel : I would be glad to skip that fdt stuff Nov 22 15:22:41 zecke: i might have a closer look at this Nov 22 15:23:19 zecke: you might be right indeed, btw, but i can't say anything yet ;) Nov 22 15:24:27 giel this is what keeps from releasing anything Nov 22 15:24:48 giel I have played a bit with mpeg4ip player and it is quite nice Nov 22 15:24:49 zecke: hm, i think they try to claim ALC as well, but i think there's prior art on that Nov 22 15:25:18 6. A device for transmitting datacast information comprising ... means for dividing the object into packets having headers; and means for transmitting the packets over a network. Nov 22 15:25:20 set fire to them damn lawyers Nov 22 15:25:25 good fucking luck with that, nokia Nov 22 15:27:19 zecke: from what i gather at a first glance, this is only FDT stuff, but this is by no means an official statement or anythng like that ;) Nov 22 15:27:42 just my own personal opinion that could be entirely wrong Nov 22 15:27:48 Nov 22 15:30:49 I have not read the claims Nov 22 15:35:54 hrw|work: could you by chance add your conf files to the post? especially site.conf with all common stuff would be interesting Nov 22 15:39:01 Jin^eLD: enjoy Nov 22 15:39:44 cool, thanks! Nov 22 15:42:16 is there an easy way to see if I am EABI or old abi at run time? Nov 22 15:43:18 I am 99% certain that glibc-2.5 will build ok for old abi now ..... Nov 22 15:43:36 with openomap distro and omap5912osk machine .... Nov 22 15:44:27 Crofton_: 'ipkg install binutils-symlinks ; objdump -p ' Nov 22 15:45:35 what is binutils-symlinks? Nov 22 15:46:05 didn't build it anywa Nov 22 15:46:17 my tool chain is arm-linux- .... Nov 22 15:48:11 in that case it's OABI Nov 22 15:48:24 Crofton_: binutils-symlinks is part of binutils Nov 22 15:50:02 if I run objdemp -p ... on the build machine, does that help? Nov 22 15:51:24 arm-linux-objdump Nov 22 15:51:43 ok Nov 22 15:51:55 and what is this tells me the abi version ? Nov 22 15:52:07 root@h2200:~# objdump -p /lib/libc.so.6 | tail -n3 Nov 22 15:52:07 0x0963cf85 0x00 05 GLIBC_PRIVATE Nov 22 15:52:07 private flags = 4000002: [Version4 EABI] [has entry point] Nov 22 15:52:41 private flags = 202: [APCS-32] [FPA float format] [software FP] [has entry point] Nov 22 15:52:47 (that's on the target) Nov 22 15:52:57 Crofton_: right, that's OABI Nov 22 15:53:16 ok, patch in glibc-2.5 fix OABI Nov 22 15:53:22 thanks Nov 22 15:54:53 Jin^eLD, the OABI stuff should work now (with glibc-2.5) Nov 22 15:55:51 I'll try and set up multimachine anyway Nov 22 15:55:54 Crofton: anything special that I would need to do? Nov 22 15:55:57 I need to be ready for EFIKA .... Nov 22 15:56:06 what target did you use now? Nov 22 15:56:26 I use openomap dsitro and omap5912osk machine Nov 22 15:57:11 for the last build I set TARGET_OS = "linux" in local.conf Nov 22 15:57:46 ok. once I finish setting up my environment I will try it again Nov 22 15:57:54 org.openembedded.dev/packages/glibc/glibc-2.5/glibc-check_pf.patch Nov 22 15:57:59 check for this file Nov 22 15:58:15 yep, it's there Nov 22 15:58:17 and that it is referenced in the glibc_2.5.bb file Nov 22 15:58:57 this is the patch that fixes the check_pf assert when you start dropbear Nov 22 15:59:11 ok Nov 22 15:59:13 thanks Nov 22 15:59:21 I will try it tonight Nov 22 15:59:25 good Nov 22 15:59:33 Crofton: org.openembedded.dev/packages/glibc/glibc-2.5/glibc-check_pf.patch won't work Nov 22 15:59:39 we try to keep improving quality Nov 22 15:59:41 um Nov 22 15:59:46 the glibc 2.5 .bb has FILESDIR = "${@os.path.dirname(bb.data.getVar('FILE',d,1))}/glibc-2.4" Nov 22 15:59:57 hmmm Nov 22 15:59:57 so it will never look in that dir Nov 22 16:00:22 it is also in the 2.4 dir ..... Nov 22 16:01:34 Jin^eLD, can you see if you have the file in both glibc-2.4 and 2.5 directories? Nov 22 16:02:01 let me check Nov 22 16:02:11 I think you should Nov 22 16:02:19 ./glibc-2.4/glibc-check_pf.patch Nov 22 16:02:19 ./glibc-2.5/glibc-check_pf.patch Nov 22 16:02:20 yep Nov 22 16:02:23 ok Nov 22 16:02:44 so you should be able to build either ABI and have dropbear work Nov 22 16:02:51 * koen starts cairo benchmark Nov 22 16:03:08 Crofton: have you commit the change to glibc_2.5.bb ? Nov 22 16:03:27 someone else commited it, you I think :) Nov 22 16:03:45 last week sometime Nov 22 16:03:46 RP: add sdl-config to sanity.bbclass Nov 22 16:03:52 Crofton: thx, will try it as soon as my setup is OK again, gotta leave now; l8r Nov 22 16:03:57 l8r Nov 22 16:04:22 zecke: That's a good idea :) Nov 22 16:04:34 I am going to get a book from the library Nov 22 16:15:39 RP: it is not good idea Nov 22 16:15:50 RP: what if I do not use qemu-native? Nov 22 16:19:23 mickeyl: morning Nov 22 16:19:56 wb mickeyl Nov 22 16:21:19 hi guys Nov 22 16:22:37 hrw|work: Not a good idea for OE. It would be appropriate for poky Nov 22 16:25:48 morning all Nov 22 16:27:06 hi Dirk Nov 22 16:27:49 hey Marcin Nov 22 16:35:27 hi mickeyl Nov 22 16:37:51 http://www.akihabaranews.com/en/en/news-12833-Sharp+Papyrus+PW-TC900,+the+first+dictionary+with+a+TV.html looks funny Nov 22 16:38:41 hi RP Nov 22 16:40:25 hey mickeyl Nov 22 16:40:33 cheers koen Nov 22 16:41:05 a+ Nov 22 16:45:18 * hrw|work goes home - cu Nov 22 17:06:34 re Nov 22 17:08:45 03koen 07org.oe.dev * r35ba4475... 10/ (4 files in 2 dirs): rsync: update to 2.6.9, remove old versions Nov 22 17:10:30 hi mallum Nov 22 17:15:08 florian_kc: hey Nov 22 17:21:00 Hello Nov 22 17:24:55 Program received signal SIGSEGV, Segmentation fault. Nov 22 17:24:56 0x00009d78 in _cairo_stats_compute () Nov 22 17:25:06 yay for the cairo performance suite! Nov 22 17:29:14 I suppose that cairo should be slow in a zaurus pda. Nov 22 17:29:36 not as slow as a few months ago Nov 22 17:29:40 still slow, though Nov 22 17:29:51 And I'm afraid there's no room for hardware accel. Nov 22 17:30:39 I've read that kaa does not implement acceleration for drawing lines, for example Nov 22 17:30:52 Or better said, does not expose Nov 22 17:34:42 Bitbake version 1.6.2 is required and version 1.6.0 was found Nov 22 17:34:47 This is new, I suppose. Nov 22 17:34:57 The sanity checker says Nov 22 17:46:47 When I run something like 'bitbake gpe-image' , how are the versions for every package selected? Nov 22 17:47:19 I mean, for example, why fontconfig-2.3.91 is chosen and not fontconfig-2.3.95 ? Where is that specified? Nov 22 17:47:30 if the distro doesn't define them it takes the highest version with the highest preference Nov 22 17:47:56 koen: And where the distro uses to define that preference? Nov 22 17:48:12 koen: I suppose that some PREFERRED_VERSION_package = somewhere ? Nov 22 17:48:15 see conf/distro/ Nov 22 17:48:19 es Nov 22 17:48:20 yes Nov 22 17:48:35 RP: i do not like this ctrl-c behaviour Nov 22 17:48:54 RP: sometimes stuff is stuck in the firewall and I do ctrl-c to kill the wget Nov 22 17:49:02 this nows downloads and stops my build completely Nov 22 17:49:24 koen: OK, I see. Nov 22 17:49:27 koen: Thanks. Nov 22 17:52:37 I see that running bitbake in dry-run mode seems to only show the packages that are still not built. Is there a way to see all the builts that a given package should trigger? Nov 22 17:52:44 Any documentation about setting up proper stuff for task-base? Nov 22 17:53:15 Kristoffer|Afk: the user-manual Nov 22 17:53:27 oki, thx Nov 22 17:53:30 I see, -v Nov 22 17:55:30 koen: It seems that the fontconfig-2.3.95-r2 package is not longer producing alignment errors. Nov 22 17:55:45 sirfred: yes, XorA found out today Nov 22 17:55:52 or cbrake Nov 22 17:56:27 koen: So, it was bumped up to 2.3.95 from 2.3.91 recently? Nov 22 17:56:47 no, I guess other distros use newer versions Nov 22 17:57:19 koen: So, that means that before today, there were some PREFERRED_VERSION_fontconfig somewhere in openzaurus-unstable distro conf file ? Nov 22 17:57:32 I wouldn't know Nov 22 17:57:40 I haven't built for OZ in months Nov 22 17:58:10 koen: I was just confused because my two days ago built was wearing 2.3.91 and not 2.3.95. That's why I was asking about how to force a given version. Nov 22 17:58:23 Anyway, it's fixed now. Great. Nov 22 18:00:00 koen: I originally added the 2.3.95 to OE because of the alignment errors problem, I thought it had become default, I have it by default locally which is why I dont see the problem Nov 22 18:01:24 XorA|gone: I think angstrom has it as default as well Nov 22 18:02:21 koen: the only reason it was not default was because libxfont needed an upgrade to fix problems as well, that was done ages ago, so its suitable to make 2.3.95 default all round I think Nov 22 18:02:38 RP: your populate_packages is wrong Nov 22 18:02:57 RP: http://rafb.net/paste/results/VjHJFc41.html Nov 22 18:02:58 koen: You asked me some time ago if the mplayer driver I made played nicely with Xw100. Nov 22 18:03:05 koen: The answer is NO Nov 22 18:03:10 sirfred: heh Nov 22 18:03:15 sirfred: how surprising Nov 22 18:03:17 koen: :) Nov 22 18:04:19 I have a dilemma. Reimplement the Xw100 driver to use my libw100 and so, be able to control access to the card at that low level. Nov 22 18:04:40 Or learn how an Xv extension is written Nov 22 18:05:35 is the bug tracker down? Nov 22 18:06:13 Gerrath: it seems that way Nov 22 18:06:34 koen: ok, thanks. Nov 22 18:12:08 cairo-xlib-surface.c: In function '_cairo_xlib_surface_show_glyphs': Nov 22 18:12:09 cairo-xlib-surface.c:2462: warning: cannot optimize loop, the loop counter may overflow Nov 22 18:20:06 03koen 07org.oe.dev * rc29c2d80... 10/ (1 packages/fontconfig/fontconfig_2.3.95.bb): fontconfig 2.3.95: up default pref so people will stop seeing alignment traps on arm Nov 22 18:21:03 koen: woowoo Nov 22 18:28:58 I suppose I should remove xerces-c 2.6.0 once I have 2.7.0 working, since the 2.6 source is no longer fetchabel? Nov 22 18:30:32 yes Nov 22 18:31:04 I guess the only people that use it have been living of the cached local copy :) Nov 22 18:31:26 the 2.7 source tarbal is about 1.5 M smaller than 2.6 .... Nov 22 18:31:35 * Crofton_ hates xml parsers Nov 22 18:31:51 * Crofton_ hates holidays Nov 22 18:31:56 * Crofton_ hates Nov 22 18:32:12 * koen loves holidays Nov 22 18:32:27 four day weekend Nov 22 18:32:35 thanksgiving? Nov 22 18:32:39 yeah Nov 22 18:32:56 my wife went to her sisters, they will fight over how to make dinner Nov 22 18:33:22 I have four days of peace and quiet Nov 22 18:33:22 * chouimat hates wearing a tie Nov 22 18:33:53 Crofton_ sound fun ;) Nov 22 18:34:29 * TheCompWiz will be playing WoW late... eating too much... and sleeping from time to time. Nov 22 18:34:46 http://www.openembedded.org/user-manual&dpage=ch07s10 says that there's no rev= for git... is it possible to fix git repository to some date, for patches to always apply correctly? Nov 22 18:35:04 last year we walked into the house about 11AM Thursday, they were fighting by 11:15 Nov 22 18:35:26 apparently we were supposed to get out of the car and start helping her cook etc Nov 22 18:35:32 not even have a beer Nov 22 18:35:52 NOTE: Multiple libraries (libxerces-depdom.so.27, libxerces-c.so.27) found and LEAD_SONAME not defined Nov 22 18:35:52 NOTE: Couldn't find shared library provider for libxerces-c.so.26 Nov 22 18:35:52 NOTE: Couldn't find shared library provider for libxerces-c.so.26 Nov 22 18:36:07 Crofton_ sound familiar to me ... it's why I'm currently thinking about staying here for xmas ... Nov 22 18:36:13 heh Nov 22 18:36:54 why is it going on about shared lib provider for .26? I am building .27 Nov 22 18:37:11 or does it know that other stuff was built against 26 already? Nov 22 18:38:01 hmm .26 is probably already in staging .... Nov 22 18:47:44 * chouimat|away is away: interview bbl Nov 22 19:02:57 what does bitbake -b path do? Nov 22 19:04:25 build only that recipe Nov 22 19:04:57 I am trying to go through hrw's notes Nov 22 19:05:06 he defined a rebuild commnad Nov 22 19:05:23 bitbake -c rebuild Nov 22 19:05:29 bitbake -b -c rebuild Nov 22 19:05:38 yeah, he wrote this last year :) Nov 22 19:08:25 so inheriting from multimachine just sorts out the build paths based on the conf files used for the build? Nov 22 19:09:01 no, it sorts out the build paths from TARGET_OS and friends Nov 22 19:09:19 well, the conf files contain that info Nov 22 19:11:27 yes, but you can also do 'MACHINE=omap5192osk bitbake blah' Nov 22 19:11:43 details :) Nov 22 19:12:14 ok, so basically, you put config files in build/osk-linux build/osk-eabi etc Nov 22 19:12:25 ? Nov 22 19:12:27 and set BBPATH to appropriate directory Nov 22 19:12:33 local.conf Nov 22 19:12:37 site.conf etc Nov 22 19:13:48 Crofton_: http://rafb.net/paste/results/abVuXc47.html Nov 22 19:14:11 Who has access to update the bugs.openembedded.org dns info? Nov 22 19:14:25 ggilbert_: kergoth Nov 22 19:15:02 There was a small mixup in communication and no one told me the IP range for the server was being changed this morning :) Nov 22 19:16:27 so bitbake knows about auto.conf? Nov 22 19:16:34 Crofton_: yes Nov 22 19:16:39 Crofton_: see conf/bitbake.conf Nov 22 19:16:53 s/bitbake knows/oe knows/ Nov 22 19:16:55 :) Nov 22 19:16:59 right Nov 22 19:16:59 whats up with the dns now? Nov 22 19:17:05 do not confuse the two :) Nov 22 19:17:29 * koen oemake -b history_1.0.oe Nov 22 19:17:36 why the cleanings in betweem machines Nov 22 19:17:43 kergoth: just found out about 10 minutes ago that the ip for my machine is now 64.136.190.225 Nov 22 19:18:25 ah Nov 22 19:18:25 k Nov 22 19:18:28 will update it Nov 22 19:18:39 Sorry for the late notice, but I didn't find out it was changing until the machine disappeared Nov 22 19:21:08 rehi, all! Nov 22 19:21:28 how could I add -dev package to image I build? Nov 22 19:25:41 hello world Nov 22 19:26:25 hello eindhoven Nov 22 19:30:14 hey likewise Nov 22 19:30:47 hello europe Nov 22 19:34:06 the sanity checker is complaing about TARGET_OS not being set Nov 22 19:34:25 do I need to inherit after I set the TARGET_OS? Nov 22 19:34:28 so set it. Nov 22 19:34:36 it is set Nov 22 19:43:12 any ideas on how to add -dev package to built image? Nov 22 19:50:07 good evening all Nov 22 19:50:31 hey cyrilRomain Nov 22 20:03:53 why is the sanity checker compaing about TARGET_OS not being set, but I have it set in openomap.conf .... Nov 22 20:40:59 what does TARGET_OS ?= "linux-gnueabi" do? Nov 22 20:41:32 If the variable is not yet set, it is set to linux-gnueabi. Nov 22 20:42:02 GNU EABI is something related to ARM's embedded application binary interface, koen is the expert here. Nov 22 20:42:26 "expert" Nov 22 20:42:40 chief curse-slinger would be closer Nov 22 20:42:47 yeah, I know that, having a stupid problem Nov 22 20:45:47 the TARGET_OS ?= "linux-gnueabi" isn't being seen when it is in the openomap distro filel ... Nov 22 20:51:24 is something setting it before that? Nov 22 20:51:32 ?= is a weak assignment Nov 22 20:51:36 shoudln't Nov 22 20:51:48 I can set it in local.conf Nov 22 20:51:54 but it was commented out Nov 22 20:52:22 sanity checker complains it isn't set Nov 22 20:53:07 currently commented out in local.conf and set with ?= in distro Nov 22 20:53:27 and failing sanity check Nov 22 20:53:35 if I set in local.conf, it works Nov 22 20:53:36 Crofton_: set it hard with = Nov 22 20:54:01 Crofton_: for unknown reason it never worked to weak set TARGET_OS Nov 22 20:54:08 I want to override Nov 22 20:54:12 hmmm Nov 22 20:54:23 koen: rob_w discovered today that fontconfig 2.3.95 works better Nov 22 20:54:41 yes Nov 22 20:54:53 but xora added it to solve those problems :) Nov 22 20:55:03 yep Nov 22 20:55:11 so a lot of people knewm but didn't remember Nov 22 20:55:43 hrw, I will set hard in local.conf and leave weak in distro, in case the problem goes away one day .... Nov 22 21:25:00 what is the best way to autoload kernel modules in a OE distro? Nov 22 21:26:03 module_autoload_foo = "module-name" in the machine,conf or kernel.bb Nov 22 21:26:20 module_autoload_pxa27x_udc = "pxa27x_udc" Nov 22 21:33:03 RP, mickeyl: setup.py of bitbake install into /usr/bitbake/{classes,conf} instead of /usr/share/bitbake Nov 22 21:33:10 I'll push fix Nov 22 21:35:49 git is broken in OE Nov 22 21:36:05 it looks for /usr/ git for target? Nov 22 21:37:06 yes Nov 22 21:42:09 hmm. I'm using linux-openzaurus-2.6.18 and all the modules are correctly installed except the ohci-hcd which is installed in lib/modules/2.6.17 instead of lib/modules/2.6.18 .. any ideas ? Nov 22 21:42:38 I tried bitbake -c clean linux-openzaurus && bitbake linux-openzaurus but this is the same Nov 22 21:42:58 check KERNEL_RELEASE in recipe Nov 22 21:44:01 hrw: in linux-openzaurus_2.6.18.bb ? Nov 22 21:44:14 yes Nov 22 21:44:17 looks ok Nov 22 21:45:05 my build for spitz has it in 2.6.18 Nov 22 21:45:39 hmm I switched from 2.6.17 to 2.6.18 without rebuilding all from scratch, so I had to rebuild wlan-ng-modules and utils Nov 22 21:45:55 I also rebuilt task-base Nov 22 21:48:21 hrw: well maybe it is time to rebuild from scratch. (even if I'd like to avoid that for only one module) Nov 22 21:51:36 should not be needed Nov 22 21:52:58 hrw: yep, I'm still looking in log files to figure out what happend Nov 22 21:59:44 * koen stabs busybox patch Nov 22 22:03:12 * koen ipkg install patch Nov 22 22:04:06 * cyrilRomain is about to find his error and prepare for an auto ~lart Nov 22 22:29:02 cu Nov 22 22:29:18 cu hrw|gone Nov 22 22:42:30 my .config (and .config-2.6.18) from tmp/staging differ from the defconfig-tosa used in linx-openzaurus, ishould it be the case ? Nov 22 22:43:35 the .bb does some sed magic on the defconfig Nov 22 22:44:51 see linux-openzaurus.inc Nov 22 22:45:03 koen: ok thanks Nov 22 22:52:38 koen: but my tmp/work/.../linux-openzaurus-2.6.18-r5/.config should match thedefconfig-tosa from linux-openzaurus-2.6.18, right ? Nov 22 22:53:09 no, since the .bb does some sed magic on the defconfig Nov 22 22:53:17 koen: yes Nov 22 22:53:29 koen: I mean at least on non magic part ;) Nov 22 22:53:36 ah, yes Nov 22 22:53:55 so I found my problem Nov 22 22:54:57 linux-oz.inc is turning into deep voodoo :) Nov 22 23:02:15 wierd, I have no login prompt on console anymore, anyone think of a change that would cause that? Nov 22 23:02:36 XorA|gone: Is init still running? Nov 22 23:02:50 RP: yes Nov 22 23:03:51 ls Nov 22 23:03:55 oops Nov 22 23:05:02 RP: 1:2345:respawn:/sbin/getty 38400 tty1 seems to have gone missing in inittab Nov 22 23:05:16 XorA|gone: That would do it... Nov 22 23:06:17 RP: Ill have to work out what revision caused that regression tomorrow Nov 22 23:08:17 anyway time for sleep Nov 22 23:08:17 'night XorA|gone Nov 22 23:13:40 it seems voodooo don't work here. '${WORKDIR}/defconfig-${MACHINE}' >>'${S}/.config' is not doing what intended. Nov 22 23:15:12 there is CONFIG_USB_OHCI_HCD=m in defconfig-tosa but not in the tmp/work/.../linux-2.6.18/.config one :( Nov 22 23:15:49 cyrilRomain: Did its name change in the kernel? Nov 22 23:15:54 check the configure logs Nov 22 23:16:35 * cyrilRomain checks logs Nov 22 23:18:33 no OHCI occurence in logs nor run files Nov 22 23:19:24 re Nov 22 23:19:26 cyrilRomain: Does the option still exist in 2.6.18 kernels? Nov 22 23:20:45 RP: ah not sure Nov 22 23:22:33 RP: I guess yes since my host 2.6.18 kernels config have them (unless they could be ignored ?...) Nov 22 23:24:41 03coredump 07org.oe.oz354x * r81b90234... 10/ (3 files in 2 dirs): altboot: Update to latest version Nov 22 23:24:46 03coredump 07org.oe.oz354x * r638e42fb... 10/ (1 conf/distro/openzaurus-3.5.4.2.conf): openzaurus-3.5.4.2.conf: Update PREFERRED_VERSION_altboot Nov 22 23:24:50 03coredump 07org.oe.oz354x * rd57cf54c... 10/ (1 packages/altboot/altboot_wip.bb): altboot: Do not overwrite the config file on upgrades Nov 22 23:24:55 03coredump 07org.oe.oz354x * r7c5ac330... 10/ (1 packages/altboot/altboot_wip.bb): altboot: Update to latest version Nov 22 23:25:00 03coredump 07org.oe.dev * r5baf3dd3... 10/ (3 files in 2 dirs): altboot: Update to latest version Nov 22 23:25:11 03coredump 07org.oe.dev * rd4a51260... 10/ (4 files in 4 dirs): altboot: Unb0rk Poodles 2.6 conf from last commit and store the master password in a new file (altboot.pwd) Nov 22 23:25:23 03coredump 07org.oe.dev * rd911fb03... 10/ (1 packages/altboot/altboot_wip.bb): altboot: Update to latest version Nov 22 23:40:17 RP: do you speak voodoo ? ;) I have some questions regarding linux-openzaurus.inc Nov 22 23:40:40 cyrilRomain: I wrote most of it ;-) Nov 22 23:41:21 RP: line 104 to 109 could be removed since line 149-150 CONFIG_AEABI and CONFIG_OABI_COMPAT are removed, right ? Nov 22 23:42:10 RP: and line 149 does not have an '=' before '/d' like the other, is it normal ? Nov 22 23:43:03 cyrilRomain: No, line 104-109 are not removed Nov 22 23:43:44 cyrilRomain: We add our own CONFIG*ABI* lines to .config in lines 104-109 and lines 149-150 filter out any line in the incoming defconfig Nov 22 23:44:06 RP: ah ok :) Nov 22 23:44:25 the missing = shouldn't be an issue either Nov 22 23:44:32 yep Nov 22 23:45:20 those sed commands delete matching lines Nov 22 23:46:19 so I really have no idea why or where OHCI are filtered out Nov 22 23:47:04 cyrilRomain: The kernel could be removing them. Keep in mind it runs make oldconfig during configure too Nov 22 23:49:02 RP: yes. I'll check for that by copying the .config after the sed command, just to be sure Nov 22 23:52:44 RP: to test images, do you usually flash your hardware, use altboot, or qemu ? Nov 22 23:53:58 cyrilRomain: Depends what I'm testing Nov 22 23:57:50 qemu crashes right after the start when I try angstrom images, and I'd like to avoid flashing too much my Z so I think I'll give altboot a try. Nov 22 23:58:10 cyrilRomain: qemu doesn't like EABI Nov 22 23:58:24 RP: shame :( Nov 22 23:58:39 cyrilRomain: I'd not worry about reflashing. Normal use is much worse for it ;-) Nov 22 23:58:57 ;) Nov 22 23:59:13 Even if you reflash every 10 minutes, it would take years to reach a number of reflashes which would trouble the flash Nov 23 00:00:24 but altboot allow you to keep a working stable system on your Z Nov 23 00:01:51 I'm not very used to flash it but I think I'm going to have the habbit to help debugging angstrom :) Nov 23 00:02:46 If you search the logs here, you'l find me worrying about the number of kernel reflashes I was making about 3-4 years ago ;-) Nov 23 00:03:00 * cyrilRomain hmm my english becomes quite bad with tire Nov 23 00:06:15 RP: you're right, this is the make oldconfig that remove the OHCI Nov 23 00:08:40 cyrilRomain: So now you need to work out why the kernel won't let you enable it :) Nov 23 00:08:45 koen: Still around? Nov 23 00:09:44 RP: yes, thanks for the voodoo clarification :) Nov 23 00:11:28 koen: http://www.rpsys.net/openzaurus/temp/maemo-mapper-nohildon_svn.bb - very rough but builds. Untested as yet ;-) Nov 23 00:11:35 'night all Nov 23 00:11:54 good night RP Nov 23 00:18:31 night all Nov 23 00:22:23 * rwhitby starts thinking about how to migrate SlugOS to use the Angstrom framework. Nov 23 00:22:46 Does Angstrom build with GCC 4? jacques says that qemu falls over when trying to build glibc locales ... Nov 23 00:23:00 (SlugOS currently builds nicely with GCC 4.x) Nov 23 00:24:07 qemu fails during configure with this error: Nov 23 00:24:22 ERROR: "gcc" looks like gcc 4.x Nov 23 00:24:22 QEMU is known to have problems when compiled with gcc 4.x Nov 23 00:24:22 It is recommended that you use gcc 3.x to build QEMU Nov 23 00:24:32 To use this compiler anyway, configure with --disable-gcc-check Nov 23 00:24:32 FATAL: oe_runconf failed Nov 23 00:24:56 so it depends on the version of the native gcc Nov 23 00:26:19 unfortunately multiple native compiler version functionality is distro-dependent Nov 23 00:28:00 I would say do the "--disable-gcc-check" thing but when I last read the qemu list, it really did have problems with gcc4 Nov 23 00:29:45 ~seen psokolovsky Nov 23 00:30:10 psokolovsky is currently on #htc-linux #oe #handhelds.org. Has said a total of 9 messages. Is idling for 59m 12s, last said: 'ok, will have a look after peering at mainline git first'. Nov 23 00:31:22 psokolovsky : which distro are you using for opie images in .dev right now? Nov 23 00:32:05 angstrom Nov 23 00:40:53 psokolovsky : could you give me your conf/bb files? I would like to try building a test image for poodle, but the .dev branch seems to be missing the angstom-opie stuff....suprise surprise Nov 23 00:41:34 hvontres|poodle, add following to local.conf: Nov 23 00:41:35 PALMTOP_USE_MULTITHREADED_QT = "yes" Nov 23 00:41:35 QTE_VERSION = "2.3.10" Nov 23 00:41:35 OPIE_VERSION = "1.2.2" Nov 23 00:41:35 require conf/distro/include/preferred-opie-versions.inc Nov 23 00:42:44 psokolovsky :ok. and then just build opie-image ? Nov 23 00:43:30 yep. should work Nov 23 00:44:15 psokolovsky :ok thanks **** ENDING LOGGING AT Thu Nov 23 02:59:57 2006