**** BEGIN LOGGING AT Tue Jul 03 02:59:57 2007 Jul 03 03:43:15 * summatusmentis is lost... considers letting someone else debug mb-applet-wireless Jul 03 03:49:00 * * OE Bug 2592 has been created by shane.oe(AT)hathawaymix.org Jul 03 03:49:02 * * linux-rp_2.6.21.bb has the wrong Collie patches Jul 03 03:49:04 * * http://bugs.openembedded.org/show_bug.cgi?id=2592 Jul 03 04:55:13 03pfalcon 07org.oe.dev * r80fcddab... 10/ (4 files in 3 dirs): opie-qcop: Do not mess with debughandlers, they are for user control. Jul 03 04:55:18 03pfalcon 07org.oe.dev * rfe825c18... 10/ (3 files in 2 dirs): libqpe-opie: Fix packaging. Jul 03 04:55:23 03pfalcon 07org.oe.dev * ra62d80fb... 10/ (1 classes/opie.bbclass): opie.bbclass: Bother to expand palmtopdir. Jul 03 04:55:29 03pfalcon 07org.oe.dev * r3caac664... 10/ (3 files in 2 dirs): Jul 03 04:55:29 opie-console: Install one of missing images. Jul 03 04:55:29 * Another one (linewrap) is missing and doesn't appear to be present in CVS. Jul 03 04:55:34 03pfalcon 07org.oe.dev * rde76eef5... 10/ (5 files in 3 dirs): opie-init: Do not hardcode OPIEDIR and friends, substitute ${palmtopdir} build value. Jul 03 04:55:38 03pfalcon 07org.oe.dev * r60811f48... 10/ (1 packages/libopie/libopie2_cvs.bb): libopie2 cvs: Drop patch merged upstream. Jul 03 04:55:42 03pfalcon 07org.oe.dev * r0bb3c74d... 10/ (1 packages/opie-irdaapplet/opie-irdaapplet_cvs.bb): opie-irdaapplet cvs: Drop patch merged upstream. Jul 03 04:55:48 03pfalcon 07org.oe.dev * rdd0c3f07... 10/ (1 packages/opie-ttf-support/opie-ttf-support_1.1.bb): opie-ttf-support 1.1: Do not use hardcoded OPIEDIR and friends. Jul 03 04:55:53 03pfalcon 07org.oe.dev * rae88d476... 10/ (1 packages/qte/qte-mt_2.3.10.bb): qte-mt 2.3.10: Fix packaging. Jul 03 04:55:58 03pfalcon 07org.oe.dev * re131c955... 10/ (1 packages/opie-oxygen/opie-oxygen.inc): opie-oxygen: Elaborate DESCRIPTION. Jul 03 04:56:04 03pfalcon 07org.oe.dev * r08779813... 10/ (1 packages/opie-networksettings/opie-networksettings_cvs.bb): opie-networksettings cvs: Drop patch merged upstream. Jul 03 04:56:11 03pfalcon 07org.oe.dev * rc5035132... 10/ (1 conf/bitbake.conf): Jul 03 04:56:11 bitbake.conf: Switch OPIE prefix to /usr. Jul 03 04:56:11 * With the other latest changes, this produces bootable system, still needs Jul 03 04:56:11 more detailed testing. Jul 03 04:56:11 * The next stage is fix data/plugin file location. Jul 03 04:56:13 * Partially closes #2254. Jul 03 04:56:15 03pfalcon 07org.oe.dev * rb4ae3e32... 10/ (1 conf/machine/wl500g.conf): wl500g.conf: Add MACHINE_FEATURES. Jul 03 04:56:20 03pfalcon 07org.oe.dev * ra45766d5... 10/ (1 conf/distro/include/preferred-opie-versions-1.2.3-pre.inc): Jul 03 04:56:24 preferred-opie-versions-1.2.3-pre.inc: Bump SRCDATE to 20070702. Jul 03 04:56:26 * Few patches merged. Jul 03 04:56:27 * Improved Palm support. Jul 03 06:17:18 olá birunko Jul 03 06:38:57 good morning Jul 03 07:00:16 good morning Jul 03 07:15:21 good morning all Jul 03 07:16:21 hi koen Jul 03 07:16:30 hey djay-il Jul 03 07:19:10 hi koen, djay-il Jul 03 07:20:21 koen: I have a problem which requires your intimate knowledge of OE to solve I think. Jul 03 07:21:00 I need to use gcc-cross_csl-arm-2005q3.bb to compile a kernel, but I don't want to use that version to compile the rest of the rootfs. Jul 03 07:21:38 (the reason is that I have binary modules for the kernel from the vendor, and I need to kernel compiler to match those, but I want the rest of the rootfs to be Angstrom) Jul 03 07:21:56 rwhitby: GCC_CROSS_KERNEL should be able to do that Jul 03 07:22:15 is there an example of that anywhere? Jul 03 07:22:33 grepping right now Jul 03 07:22:38 me too :-) Jul 03 07:23:47 it used to be in the handhelds 2.4 kernels Jul 03 07:24:24 ah Jul 03 07:24:25 KERNEL_CCSUFFIX = "-3.3.4" Jul 03 07:25:39 So do I need to create a gcc-cross-kernel.bb of some sort? Jul 03 07:25:48 rwhitby: see kernel.bbclass how OE uses KERNEL_CCSUFFIX Jul 03 07:26:44 rwhitby: no, PREFERRED_PROVIDER_virtual/${TARGET_PREFIX}gcc${KERNEL_CCSUFFIX} = "gcc-cross" + PREFERRED_VERSION__virtual/${TARGET_PREFIX}gcc${KERNEL_CCSUFFIX} = "csl-arm-2005q3" should do it Jul 03 07:27:13 that should at least trick bitbake into some more helpfull error messages Jul 03 07:29:11 I presume the double underscore in PREFERRED_VERSION_virtual was just a typo? Jul 03 07:29:20 rwhitby: http://www.openembedded.org/snapshots/ update complete Jul 03 07:29:25 yes, that's a typo Jul 03 07:30:00 rwhitby: http://www.openembedded.org/snapshots/OE.mtn.bz2 is now a symlink pointing to a snapshot, no garantees on the version Jul 03 07:30:56 every snapshot is updated each morning (UTC) Jul 03 07:31:07 cool. Jul 03 07:31:39 so pointing the makefiles to OE-this-is-for-mtn-.mtn.bz2 should always work Jul 03 07:35:26 so that should be no change for any of the makefiles I know about then. Jul 03 07:38:47 good, that's what I hoped Jul 03 07:40:10 koen: how does the preferred provider stuff above stop the csl compiler being installed in tmp/cross/bin with the unsuffixed name? Jul 03 07:40:50 ah crap Jul 03 07:40:58 you do need a gcc-cross-kernel recipe Jul 03 07:41:37 but making on is not hard Jul 03 07:41:47 judging from gcc-cross-kernel-3.4.4_3.4.4.bb Jul 03 07:42:05 will it be called gcc-cross-kernel-csl-arm-2005q3_csl-arm-2005q3.bb ? Jul 03 07:42:43 I think so Jul 03 07:43:09 it's been >2 years since I use a kernel that needed CCSUFFIX, so I'm not clear on the details Jul 03 07:43:24 and it looks like it will be identical to gcc-cross-kernel-3.3.4_3.3.4.bb Jul 03 07:43:57 echo 'require gcc-cross-kernel-3.3.4_3.3.4.bb' > gcc-cross-kernel-csl-arm-2005q3_csl-arm-2005q3.bb Jul 03 07:46:28 ok, starting a build from scratch with that. Jul 03 07:47:06 koen: Freecom are using BE EABI 2.6.18 kernel on the new alpha firmware for the FSG-3 device. Jul 03 07:48:16 I'm trying to work out how I can install Angstrom, but still be able to use the TwonkyVision media application that comes with the vendor firmware. Jul 03 07:54:58 if all else fails, use a minimal chroot Jul 03 07:55:18 or poke Jin|away to make mediatomb kick twonkies ass :) Jul 03 07:57:19 heh Jul 03 07:57:27 koen: ever seen this before: "udevd-event[397]: run_program: open /dev/null failed: Directory not empty Jul 03 07:57:28 " Jul 03 07:58:13 actaully, never mind - that's SlugOS, which I'm going to immediately replace with Angstrom on the fsg3, so I don't care if SlugOS doesn't work. Jul 03 08:00:07 I have never seen that before Jul 03 08:00:23 but IIRC slugos uses a different udev version as the rest of OE Jul 03 08:00:28 100 vs 092 Jul 03 08:00:47 hmm - I wonder why we do that. Jul 03 08:01:21 I think it's a leftover of the aborted mdev attempt Jul 03 08:04:39 good, the udev error doesn't occur with Angstrom-minimalist-image-glibc-test-20070605-ixp4xxbe Jul 03 08:08:59 koen: no unstable angtsrom python feed for ixp4xxbe? Jul 03 08:09:22 morning Jul 03 08:09:33 g'day hrw Jul 03 08:11:48 rwhitby: it seems to be missing, let me see if it's a build problem or not Jul 03 08:11:53 hey hrw Jul 03 08:12:22 koen: gstreamer is not there either - but I don't even know what that is so I'm not concerned yet :-) Jul 03 08:13:02 gstreamer should be there Jul 03 08:13:25 unless you built the rootfs before I fixed the gsteamer typo :) Jul 03 08:14:06 ah, that'd be it then. Jul 03 08:14:27 Also: "ERROR: Cannot satisfy the following dependencies for task-proper-tools: Jul 03 08:14:28 coreutils findutils gawk sed patch tar diffutils less vim" for ixp4xxbe Jul 03 08:15:36 (I would have expected those to be in armv5teb) Jul 03 08:16:26 bitbake task-proper-tools start to build here Jul 03 08:16:33 for MACHINE=ixp4xxbe Jul 03 08:16:51 rwhitby: CSL? what you builds? Jul 03 08:17:02 hrw: Freecom FSG-3 vendor kernel Jul 03 08:17:17 (so I can reuse their binary modules for the wifi card) Jul 03 08:17:20 ah Jul 03 08:21:06 would someone tell me why fbset setting do not preserved and get reset to 8bpp all the time? Jul 03 08:27:18 or, how do I force specific mode in kernel? Jul 03 08:28:17 03koen 07org.oe.dev * r6c8b1580... 10/ (6 files in 3 dirs): motorola ezx phones: add an include that sets the basic stuff, remove killswitch from a1200, e680 because the machines need different defconfigs Jul 03 08:28:25 hmmm uboot for fic fails to git Jul 03 08:57:12 03koen 07org.oe.dev * r42d82be3... 10/ (4 files in 2 dirs): kazehakase: add cvs version Jul 03 08:57:16 03koen 07org.oe.dev * rfca5fc25... 10/ (4 files in 3 dirs): angstrom: add support for htcwallaby and htctornado Jul 03 09:02:52 hmm too old git Jul 03 09:12:22 hey lrg & pH5 Jul 03 09:12:33 hey koen Jul 03 09:13:07 hey koen & lrg Jul 03 09:18:33 so someone , whats the deal with AMD64 host`s .. what problems will have have with oe when i upgrade my host to his arch ? Jul 03 09:19:24 rob_w|mis, Marek submitted a patch that supposedly fixes qemu problem, haven't tried it yet, though... Jul 03 09:19:56 polyonymous, that was the only issue .. and only to create the glibc-locales or such ? Jul 03 09:20:08 so i am secure to update my host pc , cool Jul 03 09:20:27 rob_w|mis, Well, that's the only issue I have and I am on amd64 Jul 03 09:20:47 wonderfull .. so let the money rain in and i am off to the stores !! Jul 03 09:33:36 rob_w|mis: :) Jul 03 09:33:50 too bad that my amd64 is out of stock in shopes Jul 03 09:34:20 hmm seems AMD AM2 is the current stuff Jul 03 09:34:41 and I have 939 Jul 03 09:35:04 change from 939 -> am2 costs too much Jul 03 09:35:41 ive a "old" AMD XP 3200 .. so i am good to get my whole crap new Jul 03 09:37:24 yea!! I've got gpe-dm working, finally Jul 03 09:38:52 can anyone suggest how can I make a screenshot for proof of deed? Jul 03 09:39:38 install fbgrab Jul 03 09:51:55 first ever attempt at using bitbake to compile a kernel for an h4000 based device - process errored out whilst building gcc-cross-initial-4.1.2 with | CROSS COMPILE Badness: /usr/include in INCLUDEPATH: /usr/include/libffi Jul 03 09:51:55 | cc1: internal compiler error: in add_path, at c-incpath.c:362. Any pointers ? Jul 03 09:53:12 hrw: fbgrab? Jul 03 09:56:36 djay-il: that or gpe-scap or scap or just cat /dev/fb0 > rgb565-file Jul 03 09:56:42 3yes Jul 03 09:58:00 * * OE Bug 2491 has been RESOLVED (FIXED) by florian.boor(AT)kernelconcepts.de Jul 03 09:58:02 * * gpesyncd build failure (blocks angstrom-x11-image) Jul 03 09:58:04 * * http://bugs.openembedded.org/show_bug.cgi?id=2491 Jul 03 10:03:20 pH5: just cat would do it?? Jul 03 10:05:09 djay-il: yes, you'll end up with the raw frame buffer data, so you need to remember width/height/pixel format when converting that into some image file format. Jul 03 10:10:37 hi lrg Jul 03 10:15:50 djay-il: http://www.angstrom-distribution.org/repo/?action=details&pnm=fbgrab Jul 03 10:23:32 Hi all ! Jul 03 10:25:41 When bitbaking qt4-X11-free for x86, I get the following message from the qt build process: Jul 03 10:25:45 #error "Qt has not been ported to this architecture" Jul 03 10:26:25 Should I use a different setting for machine / architecture to avoid this ? Jul 03 10:31:30 Laibsch: the kazehase browser can compile against anthy Jul 03 10:32:01 koen? Jul 03 10:32:17 maybe that rings your CJKV bell Jul 03 10:32:20 anthy is an input framework. Jul 03 10:32:39 koen: It surely does. But I don't understand the link between anthy and kazehakase Jul 03 10:33:59 I only know anthy is for cjkv and kazehakase is written by someone that is from the cjkv part of the world Jul 03 10:34:16 did I already show you http://lists.maemo.org/pipermail/maemo-developers/2007-June/010632.html ? Jul 03 10:34:23 my memory is getting a bit fuzzy Jul 03 10:35:46 koen: thanks, got it! Jul 03 10:42:02 When bitbaking qt4-X11-free for x86, I get the following message from the qt build process: Jul 03 10:42:03 morning Jul 03 10:42:08 #error "Qt has not been ported to this architecture" Jul 03 10:42:19 morning XorA ! Jul 03 10:42:30 What could be the reason for that ? Jul 03 11:00:30 Good morning ! Jul 03 11:02:14 koen: OK, the two are unrelated then Jul 03 11:02:32 Yes, you showed me that mail. Thank you for the continued input. Jul 03 11:07:12 re Jul 03 11:10:39 hey XorA Jul 03 11:11:38 koen: I missed checkin logs, did you add a kazehakase cvs? Jul 03 11:12:11 XorA: yes, I still need to figure out the mozilla trouble Jul 03 11:12:11 how kazehakase works? good? usable? ssl works? Jul 03 11:12:25 segfaults like crazy Jul 03 11:12:35 but after spawning a thread, so gdb is useless Jul 03 11:12:56 I'm committing almost finished bits from my tree before going on holiday Jul 03 11:28:11 koen: what ipkg update in angstrom does? Jul 03 11:28:21 if you please :- Jul 03 11:28:45 it downloads the package index files Jul 03 11:28:55 so you can use 'ipkg list' and 'ipkg search' Jul 03 11:29:59 cool, thanks. But if I want to build a package for it, I'd use bitbake, right? Jul 03 11:35:07 yes Jul 03 11:37:22 how then I solve the dependency problem? Jul 03 11:38:56 which dependency problem? Jul 03 11:43:35 I've built gvim and when I'm trying to install it, it says libxt6.so is missing Jul 03 11:48:17 tried 'ipkg update ; ipkg install gvim_.ipk' ? Jul 03 12:01:20 koen: http://pastebin.ca/601237 Jul 03 12:02:07 koen: did you have any more luck with blackfin recently? Jul 03 12:02:39 djay-il: time to configure networking on board? Jul 03 12:02:46 djay-il: look into /etc/resolv.conf too Jul 03 12:03:15 Jin^eLD: there is some slight progress, but nothing ready to commit Jul 03 12:03:36 roger that Jul 03 12:04:29 djay-il: ah, 'cd /etc/ipkg ; for i in *feed.conf ; do sed -i s:arm:armv5te:g $i ; done' Jul 03 12:07:10 hrw: Jul 03 12:07:30 koen: sec Jul 03 12:08:33 03rwhitby 07org.oe.dev * redcd7cc3... 10/ (3 files in 3 dirs): arm-kernel-shim: Updated the FSG-3 configuration to allow the bootloader cmdline to pass through. Jul 03 12:08:38 03florian 07org.oe.dev * r035b2607... 10/ (1 conf/distro/include/sane-srcdates.inc): sane-srcdates.inc: Update gpesyncd srcdate to become compatible with latest libraries. Fixes #2491 Jul 03 12:08:42 03koen 07org.oe.dev * r5b87de51... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom: add FEED_ARCH for compulab-pxa270 Jul 03 12:08:50 koen: thanks. Is this documented somewhere? Jul 03 12:10:41 hrw: yea, what I wanted to say is that resolv.conf is ok... Jul 03 12:11:00 djay-il: your next build should be OK Jul 03 12:11:14 koen: next build? Jul 03 12:11:48 djay-il: yes, the one after you update to r5b87de51 :) Jul 03 12:14:05 Marex: your qemu seems to work for me Jul 03 12:14:49 XorA, ok, thanks :) Jul 03 12:15:06 XorA, but hrw has probably better idea - merging poky qemu Jul 03 12:15:15 XorA, I havent tested it on amd64 though Jul 03 12:15:20 Marex: will that not have the same bug? Jul 03 12:15:37 XorA: it has other fix Jul 03 12:15:54 I dont know, it has many other patches and some seems to fix the same issue Jul 03 12:16:05 hrw, and probably better, yeah Jul 03 12:16:49 Hi! Jul 03 12:16:57 psokolovsky, hi ;-E Jul 03 12:17:52 koen: Re: "use $bindir directly" - and what about overriding it? opie is big system, and having a "master switch" to override it for all packages instead of each one individually would help Jul 03 12:17:54 psokolovsky: could you renable mtd for h2200? Jul 03 12:18:07 psokolovsky: $bindir is a masterswitch Jul 03 12:18:09 Marex: I' glad you work on qemu update issues Jul 03 12:18:21 don't add abstractions to abstractions Jul 03 12:18:28 koen: will check mtd Jul 03 12:19:25 psokolovsky, I' dont really work on it, I just fixed it, it worked so I put it in bugtracker, but anyway, poky version is surely better, they have more experienced developers than me Jul 03 12:19:33 s/me/I am/ Jul 03 12:19:34 koen: can you please provide scenario how to produce "install base to flash, opie on SD" image then (separating per dir is obvious easy solution) Jul 03 12:20:28 Marex: so, I'm glad you're working on reviewing it and making one good solution suitable for OE merge, please lead it to finish Jul 03 12:20:44 psokolovsky, can you check this http://planet.hackndev.com and this http://step2back.blogspot.com/ ? Jul 03 12:20:52 psokolovsky: if you install to sd you can chroot anyway Jul 03 12:20:58 psokolovsky, after all it seems what s2b pushed to OE is broken :-E Jul 03 12:21:50 Marex: please put priority into defconfigman to solve that in sustainable way Jul 03 12:22:18 koen: what's r5b87de51? Jul 03 12:22:37 psokolovsky, I should rather explain who is maintainer of palm lifedrive to some people ... Jul 03 12:22:44 [14:08] CIA-3: koen org.oe.dev * r5b87de51... / (1 conf/distro/angstrom-2007.1.conf): angstrom: add FEED_ARCH for compulab-pxa270 Jul 03 12:22:53 btw, I get "connection refused" now... Jul 03 12:23:04 psokolovsky, I have defconfigman in my todo, dont worry, but now Im fixing stuffed opie Jul 03 12:23:38 Marex: I hope you noticed it was updated to 20070702 to include your stuff Jul 03 12:23:42 djay-il: try again Jul 03 12:24:11 koen: ok, worked now Jul 03 12:24:24 psokolovsky, I havent noticed it ... if you included the stuff I put into bugtracker some time ago, then that is deprecated Jul 03 12:25:25 koen: so, what's r5b87de51? and why would my life become better with it ? :-) Jul 03 12:25:30 psokolovsky, I cant really do ten things at once Jul 03 12:25:42 djay-il: its a mtn revision Jul 03 12:25:50 djay-il: he means mane sure your mtn is upto date Jul 03 12:26:53 XorA: oh, I don't do that intentionally atm, because I've made changes in the tree. I will update mtn after I'm sure I can restore all the changes Jul 03 12:28:25 psokolovsky, I think if there is now some deprecated defconfig in OE, we should rather wait some time for defconfigman-crafted one ... or I can supply you with latest version directly so you can fast-fix it Jul 03 12:29:50 Marex: yes, if possible, just post last, known working version of defconfig for palmld, I'll commit it immediately, and close all those previous ones. Jul 03 12:30:20 psokolovsky, ok, I'll build kernel right away Jul 03 12:30:39 I have to strip my wifi-stuff from it since palmld wifi is still not finished Jul 03 12:38:38 djay-il: mtn diff, save the diff, then reaplly the bits you need Jul 03 12:38:55 djay-il: but as long as your changes dont conflict with server changes, mtn wont remove them Jul 03 12:41:57 XorA: thanks :-) Jul 03 12:42:13 btw, is there a way to make my device to work with external mouse? Jul 03 12:42:13 psokolovsky, what is the current version of hackndev kernel in OE? Jul 03 12:42:23 Marex: hnd2 Jul 03 12:42:31 psokolovsky, Ill prepare defconfig for r1101 Jul 03 12:42:43 psokolovsky, can you somehow set that revision in OE? Jul 03 12:44:27 djay-il: ask CoreDump|afk when he comes back online, afaik he had some magic for mixing mice and touchscreens Jul 03 12:47:12 Marex: you should tag your releases in svn, and submit .bb's with matching defconfigs for that. Jul 03 12:52:31 djay-il: plug it in? Jul 03 12:54:00 can someone either w/ a neo, or w/ the virtual env setup see if running mb-applet-wireless segfaults? Jul 03 12:55:48 summatusmentis: mb-applet-wireless is for matchbox-panel1 and openmoko use mb-panel2 Jul 03 12:56:21 hrw: oh.. ok Jul 03 12:56:33 hrw: sorry, I just woke up, for some reason I thought this was #openmoko Jul 03 12:57:53 summatusmentis: and mb-panel1 is considered obsolete in OH so no new code for it Jul 03 12:58:10 hrw: angstroms x11-image still uses it Jul 03 12:58:15 I know Jul 03 12:58:16 right? Jul 03 12:58:33 mb-panel1 is more usable then mb-panel2 for now Jul 03 12:58:41 ok, so then... will that change? or is the plan to just leave that as is? Jul 03 12:59:00 * * OE Bug 2593 has been created by jstoefer(AT)schuett-elektronik.de Jul 03 12:59:02 * * Qt4-x11 broken for generic x86 target Jul 03 12:59:04 * * http://bugs.openembedded.org/show_bug.cgi?id=2593 Jul 03 13:02:17 hrw: so no new code fixing something for mb-panel1 will be accepted, even for angstrom? Jul 03 13:02:40 it can be accepted into OE Jul 03 13:02:53 I do not know does it will be accepted in OH Jul 03 13:03:11 oh, just that OH will not be releasing anything new, etc. alright Jul 03 13:03:13 thanks Jul 03 13:04:21 summatusmentis: networkmanager applet would be better then that wireless applet Jul 03 13:04:57 networkmanager is fully working apart from a couple on install tweaks needed Jul 03 13:05:13 and someone to investigate using HAL instead of pseudo libhal that mmp wrote Jul 03 13:05:16 XorA: and it fetch libgnomeui and few other craps? Jul 03 13:05:36 or use gpe libs for nearly no gui Jul 03 13:05:50 hrw: well, partly I was working on fixing wireless applet, due to my minimal coding skills, and ignorance about the state of mb-panel1 Jul 03 13:05:59 djay-il: USB mouse should just work as long as you have the right options in the kernel enabled. Jul 03 13:06:15 djay-il: touchscreen and mouse can be used at the same time. Jul 03 13:09:04 cbrake: hi!! I'm so happy I finally did the thing... Jul 03 13:10:40 now, I just need to get it working with normal mouse... Jul 03 13:10:52 djay-il: excellent! So is compulab going to start using OE? Jul 03 13:10:57 cbrake: I have usb mouse detected, it just doesn't work Jul 03 13:11:01 djay-il: PS2 or USB? Jul 03 13:11:02 cbrake: definitely Jul 03 13:11:08 cbrake: usb Jul 03 13:11:31 and I can fly to Paris now in good mood Jul 03 13:12:13 djay-il: hmm, that should "just work". Have you verified you have all the HID stuff enabled in the kernel? Jul 03 13:12:55 psokolovsky, http://85.132.236.161/upload/defconfig please check this one Jul 03 13:13:35 Marex: please submit to tracker Jul 03 13:14:54 jeez Jul 03 13:17:58 cbrake: well, besides that kernel detects my mouse and its model, what should I check? Jul 03 13:18:40 good afternoon Jul 03 13:19:50 I still have aproblem bitbaking qt4-x11 for generic x86 target. have filed the bug some minutes ago Jul 03 13:20:22 seemms like there is a problem with the architecture the qt build process sees Jul 03 13:20:27 Any idea ? Jul 03 13:21:04 jan_js: first guess i486 doesn't have the atomic test and set operation? Jul 03 13:21:32 jan_js: open the file from the log (iirc qatomic_bla.h) and check what it checks Jul 03 13:21:55 has anyone had experience with maintaining multiple local.conf files with a single 'tmp' area? Jul 03 13:22:17 jsheldon: why multiple local.conf? Jul 03 13:22:35 o.k., I will have a look into that. Jul 03 13:23:06 I'll be building two different root file systems, one using ulibc, another using glibc, I assumed that couldn't be done with a single local.conf Jul 03 13:23:47 jsheldon: angstrom-glibc and angstrom-uclibc? Jul 03 13:25:30 well the glibc image uses a custom distro config file, but yes, the uclibc image uses angstrom Jul 03 13:26:04 if you mix distros I would not recommend sharing TMPDIR Jul 03 13:26:18 I'm just wondering if I can avoid having to use two different tmp trees Jul 03 13:26:31 ah, ok that's good to know Jul 03 13:26:33 just use angstrom :) Jul 03 13:26:59 psokolovsky, http://bugs.openembedded.org/show_bug.cgi?id=2594 Jul 03 13:27:12 koen: I might Jul 03 13:27:15 psokolovsky, at least try building it, my oe tree might be a bit obscure Jul 03 13:27:30 OE completely switched to bb 1.8 now? Jul 03 13:27:38 Jin^eLD: yes Jul 03 13:28:02 what is the main difference between 1.6 and 1.8? I will probably have to upgrade my local repos to 1.8 too then, but wonder how risky that may be Jul 03 13:28:39 less bugs and being able to do multithreaded builds Jul 03 13:28:52 Jin^eLD: main: multithreading and task dependencies instead of recipes dependencie Jul 03 13:28:54 s Jul 03 13:29:00 * * OE Bug 2594 has been created by marek.vasut(AT)gmail.com Jul 03 13:29:02 * * palmld defconfig Jul 03 13:29:04 * * http://bugs.openembedded.org/show_bug.cgi?id=2594 Jul 03 13:29:11 Jin^eLD: on fast machines build time is now shorter Jul 03 13:29:17 ok, that sounds cool Jul 03 13:29:31 so should not break a little older stuff? Jul 03 13:29:37 * hrw would like to do bitbake run on sun niagara or similiar 32-64 cores machine Jul 03 13:29:57 Jin^eLD: bitake 1.7/1.8 is in use for over half year Jul 03 13:29:59 I have an oe snapshot that I am using internally, not syncing it everyday with dev Jul 03 13:30:27 oh ok.. well it looked like 1.6 was quite popular, so I did not switch to higher versions yet Jul 03 13:30:48 zecke: in qatomic_arch.h there are checks like .. if defined(QT_ARCH_GENERIC) ... if defined(QT_ARCH_I386) and so on. Jul 03 13:31:01 djay-il: I'm not sure if tslib config would affect X and USB mouse -- might want to look into that. I suspect not. Jul 03 13:31:06 depending on result the aapropriate header is included. Jul 03 13:31:07 Jin^eLD: and you have now information how much tasks left to build Jul 03 13:31:18 Jin^eLD: NOTE: Running task 521 of 2569 (ID: 2372, /home/hrw/devel/OH/poky/meta/packages/lzo/lzo_1.08.bb, do_patch) Jul 03 13:31:39 Jin^eLD: output of bitbake is now harder to read Jul 03 13:31:46 jan_js: well, you have a line with #error, read the context. E.g. findout how the preprocessor gets there Jul 03 13:31:50 so how could I insert e.g. QT_ARCH_I386 for qt ? Jul 03 13:31:50 Jin^eLD: but you will get used to it Jul 03 13:32:02 ok, let me try it :) Jul 03 13:32:14 jan_js: read the qglobal.h and see how it is defined Jul 03 13:32:39 jan_js: check what TARGET_ARCH OE is passing to ./configure (see the configuration log bitbake wrote) Jul 03 13:33:06 o.k, tahts some hint, thanks ! Jul 03 13:33:49 jan_js: we used to have a mapping from TARGET_ARCH to what Qt's configure expects, this might be wrong for qt4 and the generic case Jul 03 13:34:04 has anyone ever gotten the following? I tried changing the configure args to include '--with-libiconv=gnu' that I found on google but to no avail: gconvert.c:48:2: error: #error GNU libiconv not in use but included iconv.h is from libiconv Jul 03 13:34:11 I am building glib-2.0 Jul 03 13:34:24 for arm9 (armv4t) Jul 03 13:34:36 Christopher: glibc, uclibc? Jul 03 13:34:37 Christopher: you built libiconv Jul 03 13:34:49 I hadn't built libiconv until I got that error. Jul 03 13:35:03 then, I tried that in addition to see if that would help and changed the bb file. Jul 03 13:35:06 koen: I can guarantee someone is deleting iconv.h from the staging area :} Jul 03 13:35:15 nothing has worked thus far. Jul 03 13:35:45 Christopher: I can only say that glib-2.0 built for armv4t here using uclibc. But the last time I tested was a week ago Jul 03 13:36:08 zecke: ok. thanks. Jul 03 13:36:48 where do I tell it to use uclibc? it looks like it is using glibc Jul 03 13:36:55 zecke: there were some cset doing "DEPENDS = libiconv" for packages, and the committer didn't respond to my questions why those where added Jul 03 13:37:00 btw, there is a iconv.h in staging area Jul 03 13:38:22 koen: ah okay Jul 03 13:38:26 ./packages/php/php_5.1.4.bb:4:DEPENDS = "zlib libxml2 mysql libiconv" Jul 03 13:38:27 ./packages/php/php_5.2.0.bb:4:DEPENDS = "zlib libxml2 mysql libiconv php-native" Jul 03 13:38:37 there were more, but I reverted those Jul 03 13:38:46 Christopher: listen to koen Jul 03 13:38:49 koen: be nice :) Jul 03 13:38:59 :) Jul 03 13:39:28 should I just get an update then? and how do I specify to use uclibc? Jul 03 13:39:46 Christopher: i thought you were using uclibc, ignore this :) Jul 03 13:39:50 remove tmp, rebuild is the safest bet Jul 03 13:41:30 I see it in the conf file now. I hadn't changed that Jul 03 13:41:37 so, delete tmp and start over with uclibc? Jul 03 13:42:39 ignore uclibc :) Jul 03 13:42:47 glibc and uclibc are both fine for glib Jul 03 13:43:08 I have seen "libiconv" and thought of uclibc Jul 03 13:44:28 zecke: ok. which should be used in general? if uclibc is best long term, I'll go with that Jul 03 13:44:54 Christopher: There is no rule of thumb Jul 03 13:45:02 Christopher: and glibc is more mature than uclibc Jul 03 13:45:09 zecke: ok. Jul 03 13:45:33 Christopher: but if you are interested in 700kb rootfilesystems with working ssh, uclibc will be of interest Jul 03 13:46:52 zecke: Previously I have used buildroot. However, there are so many changes and so few other packages implemented, I found open embedded hoping to find something suitable to build a RFS. keeping it that small is not as important as getting it to compile and work. :) Jul 03 13:47:20 Christopher: what is RFS? Jul 03 13:47:38 Christopher: I have seen your error for the first time Jul 03 13:47:39 wbx: Root file system Jul 03 13:48:09 zecke: does that mean that you have never heard of that or that you just recreated it? Jul 03 13:49:42 Christopher: ah, okay. there are other solutions to get a working rfs ;) Jul 03 13:50:03 Christopher: in my point of view, they might sometimes be simpler than oe. sometimes not. Jul 03 13:51:10 wbx: well, I have a working file system. I guess with other packages that I may need, it is nice to have something like oe to keep the build environment in check so that I don't have to worry about defining every last detail in a configure to get it to compile. I like the idea of picking a package and having it included in. buildroot did work and I do have a file system. Just need more things (graphics in particular) for a project Jul 03 13:52:27 03koen 07org.oe.dev * r41bd7025... 10/ (1 packages/php/php_5.1.4.bb packages/php/php_5.2.0.bb): php: depend on virtual/iconv Jul 03 13:53:16 Christopher: yes, x11 and graphics needs can be very good fulfilled with oe Jul 03 13:54:25 wbx: that's the way it seems. I am obviously new to it so I am trying to get the packages to compile and then tailor a distro or something to get the file system created properly. it isn't for a zaurus. just a at91rm9200 custom board. Jul 03 13:55:10 Christopher: I have to force my friend to push his at91rm9200 board configs and tweaks back into OE Jul 03 13:55:35 hrw: that would be great! what's he waiting for? :) Jul 03 13:55:41 Christopher: no idea Jul 03 13:55:55 hrw: :) Jul 03 13:56:29 Christopher: http://blackmesaeast.com.pl/projects/electronics/sarge-single-board-computer/ Jul 03 13:56:34 hrw: I know there is a kb9202 file out there and that board has a at91rm9200 so I have been using it. Jul 03 13:58:29 hrw: yep, that's in effect what I have. pretty funny project there :) Jul 03 14:00:38 hrw: he has some good detail there. thanks! Jul 03 14:01:04 zecke: did you say you saw the error when you tried it? Jul 03 14:01:38 Christopher: no problem Jul 03 14:01:51 Christopher: your error? no, I have never seen this error. But as koen pointed out it is due mixing glibc's iconv.h and libiconv Jul 03 14:02:47 how do I tell the configure to only use glibc's iconv? When I tried this the first time, I didn't have libiconv built at all through oe...then, I did a clean of libiconv and tried it all again... Jul 03 14:06:00 Christopher: well, for 99,9% glib-2.0 just build out of the box Jul 03 14:06:23 :) ok. that sounds like my luck. Jul 03 14:06:30 Christopher: so what is specific with your invocation for bitbake glib-2.0? what did you built before? Why did you manually built libiconv? Jul 03 14:06:52 cbrake: mouse works, I just don't see a cursor in matchbox..... Jul 03 14:07:21 well, i had built a lot of things before that. built nano as the example said...then, built matchbox desktop but when I got to matchbox-wm, it failed on glib. so, I built that by itself. Jul 03 14:07:32 it said it needed libiconv so I tried building that. Jul 03 14:07:43 I see the following when I do a find/grep: ./glib-2.0/glib-2.0_2.8.6.bb:DEPENDS += "virtual/libiconv virtual/libintl" Jul 03 14:07:43 ./glib-2.0/glib-2.0_2.4.6.bb:DEPENDS += "virtual/libiconv virtual/libintl" Jul 03 14:07:43