**** BEGIN LOGGING AT Wed Apr 12 09:59:57 2006 Apr 12 10:00:59 no card reader here Apr 12 10:01:03 :( Apr 12 10:01:12 I'd lend you mine, but we're too far Apr 12 10:01:28 Bernardo: I have one at home - 10-15 minutes from here Apr 12 10:01:34 bricked? that seems a bit unlikely. what did gpe do? Apr 12 10:01:37 hrw|work: use altboot to boot into init=/bin/sh and edit fstab Apr 12 10:02:05 CoreDump|home: altboot here has 'any key' problem Apr 12 10:02:14 damn Apr 12 10:02:23 pb__: keylaunch/matchbox fight for keyboard as usual Apr 12 10:02:38 oh, right. that's not a brick. Apr 12 10:02:45 bricks are much less responsive than that Apr 12 10:02:47 ctrl-alt-del Apr 12 10:02:56 CoreDump|home: zap-root-password... Apr 12 10:03:04 right... Apr 12 10:03:10 pb__: I have to reflash device to get it working again Apr 12 10:03:28 indeed Apr 12 10:03:48 so until I reflash I have nice brick Apr 12 10:03:51 but you still have a working bootloader with reflashing ability, right? Apr 12 10:03:54 CoreDump|home: /bin/sh password shows in plain text? Apr 12 10:03:58 hrw|work: isn't there a rescue rootfs on c7x0 like on cxxxx? Apr 12 10:04:15 Bernardo: yep, a bug Apr 12 10:04:16 CoreDump|home: it is but is unable to mount my flash Apr 12 10:04:24 a brick would be a device with no working firmware at all, only fixable with jtag (or not fixable full stop) Apr 12 10:04:36 hrw|work: it should be able to mount the rootfs Apr 12 10:04:42 pb__: I know I know.. Apr 12 10:04:48 03hrw 07org.oe.oz354x * r2b2868b4... 10/packages/enca/enca_1.9.bb: enca: added quotes in make call - close #834 Apr 12 10:04:53 03hrw 07org.oe.dev * r48fc87c0... 10/packages/enca/enca_1.9.bb: enca: added quotes in make call - close #834 Apr 12 10:05:01 is there a easy way to enter the rescue rootfs on the akita? Apr 12 10:05:08 CoreDump|home: will try Apr 12 10:05:23 Bernardo: yes there is a specal key combination during power-on Apr 12 10:05:30 forgot how to do it tho =) Apr 12 10:05:33 Bernardo: unplug power, take off battery. hold D+B, power on Apr 12 10:05:53 hrw|work: thanks... BTW, no .keylaunchrc in /home/root/ Apr 12 10:06:11 2.4.18-crapix oopses with 2.6.16 touched jffs2 Apr 12 10:06:20 i do not think the image breakage is caused by keylaunch Apr 12 10:06:42 with the keylaunch bug you get at least gpe-login to display Apr 12 10:07:00 right now I have a empty /home/root in cf Apr 12 10:07:06 and still it hangs Apr 12 10:07:18 CoreDump|home: ok - I will check it tomorrow then. Apr 12 10:10:11 anything else I can try to get gpe to boot? Apr 12 10:11:10 yeah, edit /etc/passwd and enable root-login Apr 12 10:11:36 launch xserver manually, then try launching gpe-login and see what happens Apr 12 10:11:50 wow.. usb card reader/writer 52-in-1 Apr 12 10:11:58 marketing.... Apr 12 10:11:58 heh Apr 12 10:12:34 CoreDump|home: for that I have to boot into CF using altboot, but block xserver from starting, right? Apr 12 10:13:08 once you've enabled root-login you can kill X with ctrl-alt-del Apr 12 10:13:29 alt is the first japanese key, right? Apr 12 10:13:35 yep Apr 12 10:16:35 no luck Apr 12 10:16:50 xserver starts again immediately Apr 12 10:17:02 try 4 times Apr 12 10:17:09 it'll die eventually Apr 12 10:20:15 Hi all!!! Is cs2005q3.2 uses kernel EABI interface or old ABI? Apr 12 10:20:15 it did... Apr 12 10:20:37 now how do I startx in gpe? Apr 12 10:21:11 slapin: It can use either depending upon the configuration Apr 12 10:21:51 Bernardo: gpe-dm (/etc/init.d/gpe-dm start) Apr 12 10:22:16 RP: and if I have kernel headers from 2.6.16 with EABI enabled in autoconf.h, will it use EABI interface? Apr 12 10:23:29 slapin: Use DISTRO = "angstrom-2006.9" Apr 12 10:24:05 slapin: You'll also need to set: PREFERRED_VERSION_gcc-cross="3.4.4+csl-arm-2005q3" Apr 12 10:24:05 PREFERRED_VERSION_gcc-cross-initial="3.4.4+csl-arm-2005q3" Apr 12 10:24:27 and disable those lines in the angstrom conf file in conf/distro Apr 12 10:25:10 RP: thanks, that gets me back to the same situation, can't launch gpe-login in the same line as it complains it can't open display Apr 12 10:25:27 RP: 2005q3.2 is gcc-3.44+glibc-2.3.5, will it be better to use gss-4.1 instead? I heard at debian wiki that it doesn't use kernel EABI interface but translates it to old ABI, is it true? Apr 12 10:26:13 anyway, it also complains of a missing gpe-setbg, /etc/gpe/locale.default, and a pango error complaining of missing modules Apr 12 10:26:52 slapin: We don't have the full 2005q3.2 toolchain in OE - we have gcc which will work with glibc 2.4. Angstrom is set to use gcc 4.1 by default Apr 12 10:26:58 the pango error suggested running "pango-querymodules" but that didn't help Apr 12 10:27:44 slapin: The kernel can optionally use the new EABI interface and has optional functionality to work with the old ABI too. Apr 12 10:28:01 another missing utility, /usr/bin/bl Apr 12 10:28:07 Bernardo: It sounds like some of the postinstalls didn't run Apr 12 10:28:38 RP: possibly. I'm trying hrw's gpe-image-alpha2 on a cf booted with altboot Apr 12 10:31:31 shit.. sharp maintaince lack wget ;( Apr 12 10:33:41 can someonehelp me out here. Apr 12 10:34:16 I am trying to build a bootstrap-image with bitkake, but i t allways failes when it trying to build the root_fs Apr 12 10:34:42 hrw|work: I'm giving up on your gpe-image for now Apr 12 10:35:09 * Bernardo also had no luck building one at home - bitbake complains of a missing diet-x Apr 12 10:35:30 Bernardo: ok - I will release RP images probably Apr 12 10:36:30 hrw|work: might also be some problem with altboot Apr 12 10:36:59 I just expanded gpe-image.tar.gz onto my cf root (formatted ext2, of course) Apr 12 10:37:43 Bernardo: you were seeing the same problems as I saw with a flashed install Apr 12 10:37:54 Bernardo: have same problem on my c7x0 Apr 12 10:38:20 ok Apr 12 10:38:38 so it wasn't me, this time... :D Apr 12 10:39:12 * Bernardo boots back into opie Apr 12 10:39:44 * hrw|work boot into sd Apr 12 10:40:26 * CoreDump|home uses his last CF card to try fixing CF booting Apr 12 10:40:41 * Bernardo was looking to the side, and is now in broken gpe again Apr 12 10:41:06 Bernardo: sound doesn't work at this stage otherwise I'd make it beep Apr 12 10:43:06 :) Apr 12 10:43:25 :) Apr 12 10:45:06 * Bernardo just decided to install initscripts, and a load of postinst scripts ran Apr 12 10:46:00 and now gpe login runs... Apr 12 10:46:16 hmm... Apr 12 10:47:00 will look Apr 12 10:47:31 The clock setting application loads all messed up, seems like it is trying to show up in portrait with the screen in landscape Apr 12 10:47:41 and the pointer is way off Apr 12 10:50:12 Bernardo: hmm.. have to check does gpe-dm starts before ipkg configure. Apr 12 10:50:43 configure runs at rc2.d, gpe-dm at rc5.d Apr 12 10:51:28 CoreDump|home: does not we boot into init 5 not into 2? Apr 12 10:51:47 ahem Apr 12 10:52:00 ignore my comment Apr 12 10:52:21 How to i override this ( iam trying to buil a bootsrap image ) task-bootstrap: unsatisfied recommendation for kernel-module-ipv6-2.6 Apr 12 10:52:39 that is not an error and can be ignored Apr 12 10:52:40 how is a kernel update supposed to work? i see that "kernel" depends on "kernel-image-${KV}", so "ipkg install kernel" installs a new kernel, if available - but it doesn't uninstall the older one, thus causing a conflict. what's wrong? can i tell ipkg to remove orphan packages? Apr 12 10:53:43 hrw|work: I'm now updating from my yesterday's build feed Apr 12 10:54:14 Got openzaurus-version, tslib-conf and zaurusd reinstalled, besides initscripts Apr 12 10:55:05 CoreDump: But it failes to build the bootstrap-image ( | ERROR: Cannot satisfy the following dependencies for task-bootstrap: Apr 12 10:55:05 | kernel-module-evdev-2.6 kernel-module-tsdv-2.6 kernel-module-asic3-mmc-2.6 kernel-module-mmc-core-2.6 kernel-module--g-ether-2.6 Apr 12 10:55:51 NOTE: package bootstrap-image-1.0-r0: task do_rootfs: failed Apr 12 10:55:58 hrw|work: seems like /usr/bin/and and /usr/bin/portmap aren't installed too Apr 12 10:56:07 you were quoting "unsatisfied recommendation" which is not a reason for the build to fail. The error is a missing dependency Apr 12 10:57:04 CreDump: How to i fix that missing dependency? Apr 12 10:57:31 well, dunno why these modules are not beeing built for you sorry Apr 12 10:58:00 ok, thx anyway Apr 12 10:58:06 ~lart gpe Apr 12 10:58:06 * ibot takes gpe to the vet for a "special" visit Apr 12 10:58:38 now I get almost to the desktop. It lets me login, shows the menu bar on the bottom, and then goes back to gpe-login Apr 12 10:59:05 Bernardo: hmm.. progress ;) Apr 12 10:59:32 Bernardo: http://hentges.net/tmp/do_not_use/Akita/ext2/gpe-image-3.5.4.1-alpha2-akita.rootfs.ext2.gz Apr 12 10:59:44 this one works Apr 12 11:00:13 CoreDump|home: can't download it from here - or my internet access might be terminated. Apr 12 11:00:22 heh Apr 12 11:00:24 I'll have to try downloading it tonight at home Apr 12 11:00:32 or in other words: any alpha2 which is not mine Apr 12 11:00:44 hrw|work: :( Apr 12 11:00:47 aw, what the hell... 20mb? Apr 12 11:01:08 CoreDump|home: I have oneliner for altboot Apr 12 11:01:11 compressed 150mb loop-image Apr 12 11:01:25 hrw|work: nice, shoot Apr 12 11:01:35 CoreDump|home: on device... have to boot first Apr 12 11:01:53 CoreDump|home: I mean, how big is that tarball? Apr 12 11:02:03 CoreDump|home: basically: if /etc/init.d/keymap exist then start it Apr 12 11:02:16 Bernardo: 17mb Apr 12 11:02:43 hrw|work: that is already bgeeing done Apr 12 11:02:55 CoreDump|home: in 1.0.6-rc2? Apr 12 11:03:13 Bernardo: and its not a tarball but a ready loop-image, tarball is in ../tarballs Apr 12 11:03:36 hrw|work: yep, check /etc/altboot.rc Apr 12 11:03:56 CoreDump|home: so that one is ready to expand into my cf card, right? Apr 12 11:04:06 CoreDump|home: ah.. I added it into /sbin/init.altboot... Apr 12 11:04:20 CoreDump|home: anyway - loading keymap does not help here ;( Apr 12 11:04:53 Bernardo: no, use the tarball from ../tarballs for that. The file i pointed you at is a loop-images to be placed into /media/cf/boot-images Apr 12 11:05:31 hrw|work: altboot reads kbd interrupts from /proc for that. the keymap doesn't matter Apr 12 11:05:43 wait a second Apr 12 11:06:01 ok Apr 12 11:06:23 found problem with my images. lack of /etc/rcS.d/S98configure Apr 12 11:06:23 cat /proc/interrupts | grep Spitzkbd on Z please Apr 12 11:06:46 CoreDump|home: c7x0 has corgikdb Apr 12 11:06:50 CoreDump|home: c7x0 has corgikbd Apr 12 11:06:57 * Bernardo is about to get fired...downloading hentges gpe image... Apr 12 11:06:59 thats the problem then ;) Apr 12 11:07:13 CoreDump|home: glad to help ;) Apr 12 11:07:26 hrw|work: what ipk should I reinstall? Apr 12 11:07:29 is the casing of corgikbd correct? no capital letters? Apr 12 11:07:48 CoreDump|home: all small Apr 12 11:07:59 CoreDump|home: "grep -i" anyway Apr 12 11:08:23 Bernardo: on my c760 when I did 'ipkg configure' then I have working image Apr 12 11:08:50 hrw|work: hre it shows "nothing to be done" Apr 12 11:09:00 I'll clean up the cf card and start again Apr 12 11:10:15 /me fears he has just killed /home Apr 12 11:14:48 hrw|work: ipkg configure seems to work, but the timezone app still shows up messed up Apr 12 11:15:42 I got the "welcome to gpe" dialog, then back to gpe-login, and now it is once again looping back to gpe-login after showing the menu Apr 12 11:16:34 1688 /home/hrw/devel/build/3541/tmp/rootfs/usr/lib/ipkg/info/ipkg.postinst: line 5: /local/pkg/oe/multi/tmp/rootfs//e tc/rcS.d/S98configure: No such file or directory Apr 12 11:16:42 argh Apr 12 11:16:51 as root I get to see the desktop, then I'm kicked to gpe-login Apr 12 11:17:04 * Bernardo has to go now, lunch time Apr 12 11:17:09 see you in a hour Apr 12 11:18:37 wow, finally a working java compiler and vm that supports floats on arm Apr 12 11:18:52 hrw|work: could you try http://hentges.net/tmp/init.altboot? Apr 12 11:19:05 CoreDump|home: could Apr 12 11:19:12 * hrw|work found problem ;)) Apr 12 11:19:17 ipkg in feed is fscked Apr 12 11:19:23 =D Apr 12 11:24:43 hi lardman|work Apr 12 11:25:42 CoreDump|home: Works! Apr 12 11:26:01 excellent Apr 12 11:26:13 hey hrw|work Apr 12 11:26:14 CoreDump|home: so 1.0.6-rc3? Apr 12 11:26:17 hi everyone Apr 12 11:26:23 yep Apr 12 11:26:35 later today Apr 12 11:29:35 I like fsck when boot from sd Apr 12 11:30:20 CoreDump|home: 'booting the selected roofs' frame could be resized to fit? Apr 12 11:31:04 I'll put it on my ToDo Apr 12 11:31:09 thx Apr 12 11:37:59 fetching 20M image takes time... Apr 12 11:41:49 flashing my gpe-image Apr 12 11:47:25 Anyone here know much about renting cars in Europe? Apr 12 11:47:49 Specifically would it be cheaper to rent one in Italy or Germany? Apr 12 11:52:49 What is this ? 'tosa kernel defconfig' Where do i find this file? Apr 12 12:13:11 CoreDump|home: I was thinking that the /home when I booted from SD wasn't my /home partition and it seems I lost all my data... :( Apr 12 12:13:29 ahem Apr 12 12:13:39 ouch :( Apr 12 12:13:54 the default is to use the real /home Apr 12 12:14:17 yep, but I had chosen to use the other Apr 12 12:15:30 hmm, worked fine last time I checked Apr 12 12:17:44 well, I'll look at "mount" output next time before trying such a stupid stunt as "rm -rf /home/*" Apr 12 12:18:15 :( Apr 12 12:19:25 hopefully, the nand flash I have here isn't very old... :) Apr 12 12:19:56 all I might have lost is my desktop background and a couple of appointments Apr 12 12:26:33 Is there a document available where OE's cross compiling process would be explained? Apr 12 12:26:50 all those gcc-cross gcc-cross-initial make me dizzy Apr 12 12:30:33 Cockroach-: gcc-cross-initial build libc which is needed to build gcc-cross Apr 12 12:33:13 hrw|work: And the idea is the same as in 'linux from scratch' process, removing dependency between the "actual" gcc-cross (built later) and the build hosts's libc? Apr 12 12:34:23 iirc yes Apr 12 12:34:50 but then I don't get it. I would understand it, if the 1st pass would produce non-cross compiling gcc Apr 12 12:35:28 FYI sox receive.wav -t raw -r 8012 -u -b -c 1 - | cat > /dev/dsp works perfectly Apr 12 12:35:29 if host gcc build gcc-cross-initial, then the resulting gcc would produce arm code? Apr 12 12:36:04 Cockroach-: yes Apr 12 12:36:50 hrw|work: how is it possible to use x86->arm gcc (gcc-cross-initial) to produce another x86->arm gcc (gcc-cross) ? Apr 12 12:37:54 Cockroach-: g-c-i/x86 build libc/arm. g-c/x86 then use libc/arm to produce binaries/arm Apr 12 12:38:28 Oh that's the way it goes Apr 12 12:38:32 hrw|work: thanks Apr 12 12:38:53 thats how I understand it Apr 12 12:41:19 I'm trying to figure that process out so I might be able to resolve my original problem some day Apr 12 12:43:53 the gpe-dm script hangs while dd'ing into /dev/fb0 on "stop" Apr 12 12:45:43 CoreDump|home: gpe-bootsplash is trying to write to / when / is ro Apr 12 12:46:11 eww, not again... Apr 12 12:46:49 CoreDump|home: maybe I have older version.. Apr 12 12:46:50 03mickeyl 07org.oe.dev * r65cea53b... 10/packages/tetex/tetex_3.0.bb: tetex: more libtool oddities Apr 12 12:46:55 re Apr 12 12:46:55 03mickeyl 07org.oe.dev * r4054569b... 10/classes/ (qmake-base.bbclass qmake.bbclass qt4x11.bbclass): qmake/qt build classes: make qmake and qmake2 coexist Apr 12 12:47:04 hrw|work: i'll have a look later Apr 12 12:47:25 CoreDump|home: 1.15-r4 is correct? Apr 12 12:48:04 yep Apr 12 12:48:11 CoreDump|home: anyway I rebuilt all my images and booted gpe-image/c7x0 into gpe Apr 12 12:48:24 thats great news Apr 12 12:55:17 * Bernardo has downloaded 40Mb worth of hentges gpe image... Apr 12 12:55:59 =) Apr 12 12:56:22 If I get fired, it's your fault... :) Apr 12 12:57:25 lets hope ext2fs doesn't kill my windows machine again Apr 12 13:00:57 * Bernardo does another flash backup before further tests Apr 12 13:10:45 03hrw 07org.oe.oz354x * rd05cf179... 10/packages/ipkg/ipkg_0.99.159.bb: ipkg: bumped PR to fix feed Apr 12 13:10:49 03hrw 07org.oe.oz354x * r3c60ed32... 10/packages/openzaurus/openzaurus-version.bb: openzaurus-version: set PACKAGE_ARCH to 'all' Apr 12 13:10:53 03hrw 07org.oe.oz354x * rcc03be81... 10/packages/meta/ (meta-e-x11-core.bb meta-e-x11.bb meta-e.bb): meta-e-*: fixed feeds, set PACKAGE_ARCH to 'all' Apr 12 13:10:58 03hrw 07org.oe.oz354x * r0f4bab6c... 10/packages/meta/ (meta-gpe-extras.bb meta-gpe.bb): meta-gpe*: set PACKAGE_ARCH to 'all' Apr 12 13:13:10 ~seen pigi Apr 12 13:13:26 pigi was last seen on IRC in channel #oe, 12d 21h 3m 37s ago, saying: 'later all'. Apr 12 13:31:24 * Bernardo wonders for how long should he wait for pivoting root... Apr 12 13:31:45 seconds. Apr 12 13:31:49 Bernardo: Not more than a minute. Apr 12 13:32:48 ok, so it has hang for some reason Apr 12 13:32:57 I was trying hentges gpe now Apr 12 13:33:39 bbl Apr 12 13:37:09 5 seconds max I'd say Apr 12 13:52:41 ARM system image emulation with QEMU for OE images via nfs: http://www.rpsys.net/openzaurus/qemu/ Apr 12 14:00:42 03coredump 07org.oe.dev * r18e53df4... 10/packages/altboot/ (5 files in 3 dirs): altboot: Fix tar.gz installer to work w/o inserted cf or sd cards, fix mount_cf to work with udev Apr 12 14:01:05 03coredump 07org.oe.dev * rc600c8a6... 10/packages/altboot/ (5 files in 4 dirs): Apr 12 14:01:05 altboot: Apr 12 14:01:05 - Emmit a low-volume beep on alsa devices before running the timeout Apr 12 14:01:05 - Removed some redirects from init_rootfs Apr 12 14:01:05 - Fixed CF booting w/ kernel 2.6.16 Apr 12 14:01:07 - Fixed altboot menu on c7x0 running kernel 2.6 Apr 12 14:02:55 CoreDump|home: care to merge those to oz354x? Apr 12 14:03:33 CoreDump|home: Fixed CF booting w/ kernel 2.6.16 == Fixed CF booting w/ pcmciautils isntalled instead of pcmcia-cs? Apr 12 14:03:36 yep, but I have some more things to fix first Apr 12 14:03:45 hrw|work: right Apr 12 14:04:02 CoreDump|home: it's strange when booting hentges gpe image from a cf to have a altboot inside a altboot... Apr 12 14:04:10 Bernardo: =D Apr 12 14:04:15 I know the feeling Apr 12 14:04:20 I just let the one inside cf to timeout... Apr 12 14:04:31 indeed Apr 12 14:05:20 deleting /home/.gpe now... Apr 12 14:05:55 Bernardo: .gpe .matchbox .Xdefaults .keylaunchrc Apr 12 14:06:20 timezone applet shows weird too Apr 12 14:07:07 is BLOB able to work with initramfs ? Apr 12 14:07:16 Found the hostap-modules problem :) Apr 12 14:07:28 the "welcome to gpe" dialog stays up... How do I kill it? Apr 12 14:07:29 Bernardo: those EET/UST/GHT etc names? Apr 12 14:07:36 hrw|work: yes Apr 12 14:07:59 the applet only occupies half the screen, and the pointer seems way off Apr 12 14:08:28 but I got past that, and now I have a "Welcome to gpe!" dialog that won't close (but will move if I drag it). Apr 12 14:09:06 finally closed it, got the "ok" button highlighted and pressed the space bar Apr 12 14:09:07 bbl Apr 12 14:11:03 it seems that the pointer really was badly calibrated... strange, I pressed on all right spots Apr 12 14:11:42 how do I recalibrate it? Apr 12 14:11:53 settings / screensetup Apr 12 14:11:55 Bernardo|meeting: Run xtscal Apr 12 14:12:28 ok, seems fixed now Apr 12 14:12:34 really have to run, bbl Apr 12 14:18:15 03rpurdie 07org.oe.dev * re4253b12... 10/packages/linux/linux-openzaurus.inc: linux-openzaurus: Fix hostap-modules problem people were seeing (kernel.bbclass trampled a variable we were setting) Apr 12 14:18:19 03rpurdie 07org.oe.dev * rdbb7c9f1... 10/packages/fontconfig/fontconfig_2.2.95.bb: fontconfig: Use RREPLACES instead of RPROVIDES to prevent upgrades breakage Apr 12 14:19:14 That sounds a bit odd. What files is fontconfig replacing? Apr 12 14:22:09 pb__: libfontconfig-utils became fontconfig-utils Apr 12 14:22:47 In that case, it ought (probably) to provide/replace/conflict with the old package. Apr 12 14:23:04 Replaces on its own means that you want to overwrite some subset of the files in the old package. Apr 12 14:23:50 So you need to do all three? Apr 12 14:24:02 Right. Apr 12 14:28:56 * chouimat is back. Apr 12 14:29:31 morning Apr 12 14:30:08 those provide/replace/conflict makes me mad sometimes.. Apr 12 14:30:40 recently .oz354x got change in irk - they rconflict but do not rreplace/rprovide Apr 12 14:30:47 03rpurdie 07org.oe.dev * rde96ee4f... 10/packages/fontconfig/fontconfig_2.2.95.bb: fontconfig: Also RREPLACES and RCONFLICTS libfontconfig-utils (thanks pb for spotting) Apr 12 14:31:17 I suspect we have a developer misunderstanding or two on how this works... Apr 12 14:32:34 http://www.us.debian.org/doc/debian-policy/ch-relationships.html#s-replaces has some words about that stuff. Apr 12 14:35:17 03coredump 07org.oe.dev * r5b47c321... 10/packages/altboot/files/altboot.func: altboot: Do not echo the password when entering init=/bin/sh Apr 12 14:35:25 I had the impression Replaces had only the second function. I now know better :) Apr 12 14:38:51 http://openzaurus.org/wordpress/2006/04/10/second-alpha-release-of-openzaurus-3541/ Apr 12 14:39:49 mickey|zzZZzz: qt4 build seems to want to create a 64bit arm :-) http://pastebin.com/655697 Apr 12 14:40:17 hrw|work: Looks good to me Apr 12 15:16:45 spitz/borzoi/terrier are pxa270 or 272? Apr 12 15:20:45 hrw|work: 270 Apr 12 15:22:11 ok Apr 12 15:22:35 thx Apr 12 15:24:41 cu Apr 12 15:24:43 time to go Apr 12 15:24:49 03coredump 07org.oe.dev * r08f8592d... 10/packages/altboot/files/beep.raw: altboot: Add missing file Apr 12 15:32:32 03coredump 07org.oe.dev * r25365dea... 10/packages/altboot/ (altboot_0.0.0.bb files/init.altboot): altboot: Fix STDIN error messages on collie Apr 12 16:04:43 CoreDump|home: ping Apr 12 16:05:06 pong Apr 12 16:05:12 :) Apr 12 16:05:57 I've had a small problem with your gpe image Apr 12 16:06:12 it thinks it is running in portrait mode Apr 12 16:06:25 when I have my akita open Apr 12 16:06:32 bad install or damaged FS on card Apr 12 16:06:48 I use the GPE images day in day out, trust me =) Apr 12 16:07:09 :) Apr 12 16:07:16 I usually use opie... Apr 12 16:07:26 time to reformat the card and try again? Apr 12 16:08:22 yep Apr 12 16:09:18 ok, here it goes... :( Apr 12 16:09:52 formatted... time to extract the tarball Apr 12 16:17:04 bye Apr 12 16:17:29 bye Genesis, good evening everybody Apr 12 16:18:03 hi pH5 Apr 12 16:18:18 hi Bernardo Apr 12 16:21:05 hi pH5 Apr 12 16:21:20 hi CoreDump|home Apr 12 16:22:05 * Bernardo waits for the tarball... Apr 12 16:23:21 03coredump 07org.oe.oz354x * r979e958d... 10/packages/altboot/ (altboot_1.0.6-rc2.bb altboot_1.0.6-rc3.bb): altboot: Add 1.0.6-rc3, including all changes from .dev Apr 12 16:23:25 03coredump 07org.oe.dev * r24be272e... 10/packages/altboot/ (altboot_1.0.6-rc2.bb altboot_1.0.6-rc3.bb): altboot: Add 1..6-rc3 Apr 12 16:38:35 booting now... Apr 12 16:38:58 I wonder why it finds a .backup directory, when I told it to move /home to inside the image Apr 12 16:40:37 and the same problem once again. As soon as it passes gpe-login, gpe gets convinced it is running in portrait Apr 12 16:40:56 mmm... Apr 12 16:41:25 launching "screen setup" sends me back to gpe-login Apr 12 16:41:40 try again Apr 12 16:42:07 now it launched, and it thinks it is running in portrait mode Apr 12 16:42:31 I booted that image countless times off an NFS drive Apr 12 16:42:33 if I choose "landscape (left)" it shows in portrait mode Apr 12 16:42:41 I _know_ rotation is right ;) Apr 12 16:43:05 Bernardo: yeah that app is bugged in this regard Apr 12 16:43:23 brightness doesn't work Apr 12 16:43:45 it should Apr 12 16:44:07 the slider doesn't move Apr 12 16:44:13 do I need to login as root? Apr 12 16:44:41 nope Apr 12 16:45:15 also, I've just launched the audio mixer and can't close it Apr 12 16:45:49 that's a known problem with this app. On the left side there is a bar you can click IIRC Apr 12 16:46:10 yet, thnaks Apr 12 16:46:14 thanks Apr 12 16:46:36 I can't type ":" or ";" in firefox Apr 12 16:46:44 haven't checked on other apps Apr 12 16:47:09 will try in a second Apr 12 16:47:11 works in abiword Apr 12 16:50:16 also, /home still gets mounted. Could it be because I'm not using a loopback image, but a extracted tarball in /media/cf? Apr 12 16:50:31 i will investigate this bug Apr 12 16:51:37 thanks, sorry for all the trouble Apr 12 16:51:56 * Bernardo is going back to opie... :) Apr 12 16:55:45 bye, see you guys later Apr 12 17:16:16 hi mickeyl Apr 12 17:16:27 hey pb_ Apr 12 17:16:46 hi mickeyl, pb_ Apr 12 17:16:50 hi ph5 Apr 12 17:16:55 cheers pH5 Apr 12 17:21:30 Hrm: Apr 12 17:21:32 FATAL: This kernel is too big for the c7x0 and will destroy your machine if you flash it!!! Apr 12 17:30:54 03mickeyl 07org.oe.dev * r4b742ad8... 10/packages/lyx/lyx_1.4.1.bb: lyx: add version 1.4.1 Apr 12 17:37:08 hey mickeyl Apr 12 17:37:44 hi xora Apr 12 17:38:24 mickeyl: did you see my earlier message? Apr 12 17:38:51 * mickeyl scrolls up Apr 12 17:39:23 mickeyl: http://pastebin.com/655697 Apr 12 17:39:28 oh dear Apr 12 17:39:46 as much as Id love a 64bit arm :-) Apr 12 17:40:51 hehe Apr 12 17:41:25 please try again with an updated HEAD Apr 12 17:41:42 in case you didn't have the classes update which i commited some hours ago Apr 12 17:42:01 pulling now Apr 12 17:42:07 then start with sending me log.do_* to mickeyl@handhelds.org Apr 12 17:42:15 and we'll start to investigate Apr 12 17:42:30 mickeyl: yeah, I just didnt have time to investigate in work today Apr 12 17:42:41 what's your DISTRO and MACHINE conf btw.? Apr 12 17:42:59 DISTRO="openzaurus-unstable" Apr 12 17:43:04 MACHINE="c7x0" Apr 12 17:43:21 right. same here. that makes diffing easier Apr 12 17:44:02 just looked at configure script, I think its actually managed to be worse than an m4 generated one :-) Apr 12 17:47:31 heh. right. those who don't understand autotools are forced to reinvent it in a worse way Apr 12 17:47:52 as much as i hate autotools, handcoded ones I even hate more Apr 12 17:47:55 *sigh* Apr 12 17:48:47 mail is on its way Apr 12 17:49:02 Necronom: Is that being built on a 64 bit machine? Apr 12 17:49:02 excellent. I'll look into that tomorrow Apr 12 17:49:11 hi mickeyl Apr 12 17:49:14 hi RP Apr 12 17:49:28 Necronom: Nope. 32 bit machine. Apr 12 17:49:52 Necronom: .dev? which DISTRO? Apr 12 17:52:39 XorA: aaaah Apr 12 17:52:41 XorA: you Apr 12 17:52:45 are on a 64 build system Apr 12 17:52:50 64 bit, even Apr 12 17:53:18 mickeyl: yes Apr 12 17:53:25 looks like i didn't swith all tests to using the target compiler yet Apr 12 17:53:38 that gives me enough info to look into it Apr 12 17:53:39 thanx Apr 12 17:53:41 (kind of ) Apr 12 17:53:42 :) Apr 12 17:53:56 mickeyl: no probs, and I ready to test any fixes as soon as I see them Apr 12 17:54:11 cool. i don't have access to a 64bit machine, so I need to rely on you Apr 12 17:54:47 mickeyl: I have found gentoo makes a good testing platform as /usr/include/* is effectively empty Apr 12 17:55:39 RP: cat org.openembedded.oz354fam083/MT/revision Apr 12 17:55:39 f14a8701d982daae63a37e5c8c5a0cbb6ee802eb Apr 12 17:55:52 This would be a LFS 6.1 box.. Apr 12 17:56:18 Necronom: You're using familar? Apr 12 17:56:50 grep DISTRO build/conf/local.conf Apr 12 17:56:50 DISTRO = "openzaurus-3.5.4" Apr 12 17:57:10 Necronom: For the best c7x0 support I'd use kernels from .dev or oz354x Apr 12 17:57:57 RP: prism3 cards w/ uploadable firmware don't work on resume Apr 12 17:58:04 RP: spitz/3.5.4.1-alpha2 Apr 12 17:58:17 XorA: right. i should use a gentoo just for building Apr 12 17:58:36 mickeyl: only on amd64 machines with multilib turned on though Apr 12 17:58:45 mickeyl: :-(. Was there code in pcmcia-cs to reupload the firmware on them? Apr 12 17:59:02 heya, if my glibc on my host was only compiled with support for my locale would it causes problems trying to compile different locales? Apr 12 17:59:14 mickeyl: that pushes the includes into ARCH dirs and /usr/include/* becomes #if ARCH #include ARCH/include.h Apr 12 17:59:42 mickeyl: There are perhaps two options 1) see if the kernel driver can handle the firmware through udev 2) if not, add/fix the scripts so they're run on resume Apr 12 17:59:51 mickeyl: re suspend thing, according to my Sandisk manual, PocketPC cant handle resuming a prism3 card either Apr 12 18:00:28 XorA, mickeyl: I'm guessing this did work though Apr 12 18:00:28 RP: i handled uploading the firmware by a resume script calling hostap-fw-load. Apr 12 18:00:35 RP: see contents of prism3-support package Apr 12 18:00:53 RP: its never worked for me, but I rarely need it to so never complained Apr 12 18:01:01 i had it working on 3.5.3 Apr 12 18:01:06 that was w/ 2.4 though Apr 12 18:01:22 mickeyl: Can you check dmesg and see if the kernel looked for firmware? Apr 12 18:01:28 ah, I missed 3.5.3 going straight to 2.6 kernel stuffs Apr 12 18:01:49 hmm. that doesn't look good Apr 12 18:03:04 RP; thansk. Apr 12 18:04:50 mmh Apr 12 18:04:54 can't log in Apr 12 18:05:36 account revoked? oh dear Apr 12 18:05:51 looks like. Apr 12 18:06:38 * RP -> back later Apr 12 18:08:48 * mickeyl same Apr 12 18:08:51 RP: here it is Apr 12 18:10:26 ~lart pastebin Apr 12 18:10:26 * ibot gives pastebin a "free" copy of Windows and then charges double for "Upgrades" Apr 12 18:10:47 http://pastebin.com/656145 Apr 12 18:11:03 that's the log right after pressing on/off to resume Apr 12 18:14:59 hi all Apr 12 18:16:03 hey koen|train Apr 12 18:16:16 hey pH5 Apr 12 19:00:07 * Philippe is back (gone 21:34:53) Apr 12 20:14:09 * RP returns Apr 12 20:25:31 hi koen Apr 12 20:26:30 koen: JustinP changed the e .inc file so it doesn't trample on distro settings. Can let Angstrom do things properly again now? Apr 12 20:26:52 RP: I'll change it now Apr 12 20:26:55 mickey|bbl: I've just tried hostap here and it works with my card. That dmesg you posted is worrying :-( Apr 12 20:27:20 * koen just got back from 3 days without net access Apr 12 20:27:51 koen: thanks. I don't mind doing it as I have that change locally - just wanted to check with you first :) Apr 12 20:28:09 RP: please go ahead Apr 12 20:28:40 koen: ok. Can I also add a comment about how to enable the gcc3 csl toolchain? Apr 12 20:31:14 RP: if you can please put the gcc3 bits in the .conf, but commented out Apr 12 20:31:51 koen: That's what I was asking :) Apr 12 20:31:59 :) Apr 12 20:43:00 03rpurdie 07org.oe.dev * rc6c7cc9d... 10/conf/distro/angstrom-2006.9.conf: angstrom-2006.9: Allow the user to override the toolchain versions again. Also comment on how to use an appropriate gcc version 3 Apr 12 20:43:45 RP: thanks Apr 12 20:44:06 koen: I'm pleased its in as I can't accidentally commit it now :) Apr 12 20:44:50 kergoth, you still in Texas? Apr 12 20:44:58 RP: :) Apr 12 20:45:00 Bernardo|away: I've identified the "/home" issue. Apr 12 20:45:02 hey treke Apr 12 20:45:17 Bernardo|away: udev mounts /home even if it isn't configured in fstab... Apr 12 20:45:23 treke: thanks again for bugs.oe.org Apr 12 20:45:38 koen: Glad it's still working well Apr 12 20:46:06 one of these weekends I still want to make source mirroring work again Apr 12 20:46:13 It is useful - stops things getting lost... Apr 12 20:46:31 Just hard to get motivated to prune old stuff Apr 12 20:46:43 treke: It would be nice to have the working again :) Apr 12 20:46:58 It never actually got turned off. Just sort of ran out of space Apr 12 20:47:31 I remember you saying - and identifying the files to keep is tricky :-/ Apr 12 20:47:44 yeah Apr 12 20:47:50 * CoreDump|home looks at /etc/udev and curses Apr 12 20:47:56 this is going to be a bitch Apr 12 20:48:18 I'd say this weekend, but I told myself the same thing last friday Apr 12 20:48:21 :) Apr 12 20:49:31 oh, kergoth isn't even here :p Apr 12 20:50:04 treke: iirc he's in AZ now doing work for a defense subcontracter Apr 12 20:50:13 ah Apr 12 20:50:37 this is certainly the time to work for one :p Apr 12 20:53:21 How are things going on the pda front? Anything interesting going on with pda/phones Apr 12 20:53:55 palm still rules, everyone else still drools Apr 12 20:54:03 hehe Apr 12 20:54:04 OE built rootfs booted on a treo 650 Apr 12 20:54:08 hey CP|CELF Apr 12 20:54:09 oh god Apr 12 20:54:12 don't tempt me Apr 12 20:54:14 treke: oz support is looking nice with 2.6 kernels now :) Apr 12 20:54:19 RP: nice Apr 12 20:54:33 people are getting linux on the HTC universal phone Apr 12 20:54:38 koen: hey Apr 12 20:55:08 treke: http://gehennom.net/~tool/tiki-index.php?page=TreoLinuxPics Apr 12 20:55:15 CP|CELF: how did your presentation go? Apr 12 20:55:20 koen: good Apr 12 20:55:26 koen: couldnt possibly be any less stable :p Apr 12 20:55:29 that's good to hear Apr 12 20:55:37 koen: have you already heard about the CELF Test Lab? Apr 12 20:55:54 CP|CELF: no, I've been incommunicado for the past few days Apr 12 20:56:17 koen: running gpe? slick Apr 12 20:56:37 koen: They're setting up a lab to automatically build and run images on a number of different embedded boards Apr 12 20:56:46 the kicker is - they're using OE to do the builds Apr 12 20:56:47 CP|CELF: cool Apr 12 20:57:06 * koen is working on a similar setup at home Apr 12 20:57:19 They even have a web interface that you can select your machine/distro/meta image to build Apr 12 20:57:32 cute Apr 12 20:57:36 CP|CELF: This something at AMD? Apr 12 20:57:53 treke: no - its totally funded by CELF Apr 12 20:58:14 CP|CELF: but as zecke would say "why aren't we filthy rich yet?" ;) Apr 12 20:58:17 googling. Never heard of CELF :) Apr 12 20:58:36 koen: heh Apr 12 20:58:47 Consumer Electronics Linux Forum? Apr 12 20:58:53 treke: nod Apr 12 20:59:10 koen: apparantly they aren't making enough to actually hire a web designer Apr 12 20:59:22 CP|CELF: could you poke one of the CELF people to update the 'success strories' wiki? Apr 12 20:59:46 koen: good idea - the guy in charge is in the room right now, I'll talk to him at a break Apr 12 20:59:52 CP|CELF: thanks Apr 12 21:00:02 CP|CELF: and say hi to lrg if you see him Apr 12 21:00:15 also, tell him that koen and zecke want to be rich? Apr 12 21:00:38 koen: I will Apr 12 21:01:10 CP|CELF: do you know of any PDAs sporting an alchemy SoC? Apr 12 21:01:29 koen: the DigitalCube is being sold in Korea, iirc Apr 12 21:01:30 CP|CELF: See if there's anything they would fund to improve OE/bitbake ;-) Apr 12 21:01:34 who's the maintainer of udev? Apr 12 21:02:21 CoreDump|home: pH5 is as close as you'll get. I also tweak it now and again... Apr 12 21:02:30 ooooh Apr 12 21:02:35 i see Apr 12 21:02:35 http://www.dapreview.net/comment.php?comment.news.1972 Apr 12 21:03:39 udev's mount.sh script is used to auto-mount CF / SD / whatever devices on insertion. But it really shouldn't auto-mount flash partitions in my book ;) Apr 12 21:04:05 beeing able to not mount /home on boot is a very nice thing Apr 12 21:04:22 dibs on anhy istation machines shipped to OE devs ;) Apr 12 21:04:29 s/anhy/any/ Apr 12 21:04:56 koen: :-D Apr 12 21:05:09 koen: You might have a fight on your hands ;-) Apr 12 21:05:25 :) Apr 12 21:06:03 CoreDump|home: That is indeed not desireable - the udev scripts could use some work including the network one (which breaks under nfs) Apr 12 21:06:22 CoreDump|home: Any ideas on how to stop it mounting /home? Apr 12 21:06:27 RP: I'll filter mtdblock* devices then Apr 12 21:06:41 CoreDump|home: Won't work on c3x00 Apr 12 21:06:41 easier than I thought ;) Apr 12 21:06:54 argh Apr 12 21:07:03 /home on MD Apr 12 21:07:08 yes Apr 12 21:07:20 ~lart http://v43.i-station.co.kr/ Apr 12 21:07:21 * ibot shoves a crumpet down http://v43.i-station.co.kr/'s throat, happy now?! Huh? Want some JAM with that? Apr 12 21:07:42 maybe we could create a blacklist in /etc/udev which devices/partitions are to be ignored? Apr 12 21:08:24 CoreDump|home: and make it device dependent? Apr 12 21:08:31 yeah Apr 12 21:08:46 assuming that th /home partition on MD is always the same of course Apr 12 21:08:54 I did once wonder whether zaurusd/devmand could take on this kind of thing Apr 12 21:09:23 Ignoring all partitions on the microdrive would be the desired behaviour Apr 12 21:09:44 RP: MD is _always_ /dev/hda right? Apr 12 21:09:52 CoreDump|home: For 2.6, yes Apr 12 21:10:12 I'll do some hacking then ;) Apr 12 21:10:56 CoreDump|home: If you're in the area, the network.sh script should check if a network interface is already up before trying to do its thing... Apr 12 21:11:25 RP: It needs somelove indeed. Though I never had problems with it with "fake" root on NFS Apr 12 21:11:50 but it almost never launches an already inserted NIC on boot Apr 12 21:12:28 CoreDump|home: It really upset my qemu over root nfs setup... Apr 12 21:12:35 ah heh Apr 12 21:12:46 will have a look Apr 12 21:12:53 The quick fix was rm network.sh :-/ Apr 12 21:13:55 ahem, ouch Apr 12 21:14:27 I had other things to worry about at the time ;-) Apr 12 21:14:51 I gave altboot an option to not mount /home when booting off SD or CF. So, thinking that he was safe, a user rm -rf'ed /home Apr 12 21:15:11 well, altboot disabled /home in fstab alright... Apr 12 21:15:15 ;) Apr 12 21:15:41 :-/ Apr 12 21:16:18 mickey|bbl: Looking at this dmesg log, CS_EVENT_PM_RESUME is conspicuous by its absence. Looks like hostap_cs is crashing upon suspend for that card... Apr 12 21:17:03 CoreDump|home: I see that udev is to blame Apr 12 21:17:56 Bernardo: it's the first release w/ udev replacing hotplug, there were bound to be problems Apr 12 21:18:09 yep Apr 12 21:18:39 in sourcemage there were a lot of problems with firmware uploading Apr 12 21:45:55 n8 all Apr 12 21:58:07 Why is it that 'do_rootfs' failes. Is it because of the missing modules dependecies? Apr 12 21:58:43 goxboxlive: It usually tells you in the log Apr 12 22:04:12 Anyone know how i clear my failed bootstarp-image to start all over again ( just the compilation, not downloading the whole CVS kernel again)? Apr 12 22:08:24 goxboxlive: Just rerun bitbake bootstrap-image? Apr 12 22:13:09 RP: Even if i have made changes to the machine configuration? Apr 12 22:13:29 goxboxlive: Ah, no. You need to rebuild task-bootstrap Apr 12 22:14:01 goxboxlive: bitbake task-boostrap -c compile -f, then bitbake bootstrap-image should do it Apr 12 22:14:25 thx Apr 12 22:17:54 * Philippe is away: visual contact - melancholic dreams Apr 12 22:28:36 RP: WHen i tried to to rebuild the task-bootstrap it gave me an 'packages quilt-native-042 failed' Apr 12 22:30:17 goxboxlive: That sounds strange as it should have built that a long time ago. Check you were in the right directory with the right environment Apr 12 22:32:57 RP: I am in the 'build' directory. I also tried to put 'INHERIT += patcher' but that gave me another Error message Apr 12 22:34:59 goxboxlive: It sounds like you're no longer in your build directory or your enviornment isnt correct for bitbake. I'm afraid I can't really tell much more Apr 12 22:35:20 RP: Thx anway Apr 12 23:10:26 mreimer: We added both the comments for 1525 in exactly same time. For me it was not ok... Apr 13 00:26:06 hey people Apr 13 00:26:08 sup? Apr 13 03:52:08 where can i find a list of all the key symbols that loadkeys recognizes? Apr 13 03:53:35 ah, asked too soon Apr 13 03:53:56 dumpkeys --long-info Apr 13 03:56:07 hrm... no symbols for Japanese keyboards... Apr 13 03:59:23 you probably need a .jp font first Apr 13 03:59:34 * CoreDump|afk is off to work... Apr 13 04:00:28 the font isn't a problem Apr 13 04:00:38 I want to reenable the Japanese keys Apr 13 04:00:41 at least for Opie Apr 13 04:29:24 * emte thought some deveices had kanji support Apr 13 04:29:29 devices* Apr 13 04:29:53 i know raster was compiling it for his Apr 13 04:30:25 compiling image / using Apr 13 04:44:08 not sure, but it appears that the keys are unrelated to the keymap(loadkeys) Apr 13 04:44:24 if I can get KeyHelper working properly I'll just use that to remap them Apr 13 05:43:38 mickey|bbl: don't suppose you're around.... Apr 13 05:44:20 I need to install those evas modules to an arch specific directory....such as linux-gnu-i686 Apr 13 06:56:03 goodmorning Apr 13 06:56:14 is there any problem with the server ? Apr 13 06:56:18 stelios@Athena:~/oe$ monotone --db=./oe.db pull ewi546.ewi.utwente.nl org.openembedded.{dev,dreambox} Apr 13 06:56:18 monotone: doing anonymous pull; use -kKEYNAME if you need authentication Apr 13 06:56:18 monotone: connecting to ewi546.ewi.utwente.nl Apr 13 06:56:18 monotone: network error: name resolution failure for ewi546.ewi.utwente.nl Apr 13 06:56:18 stelios@Athena:~/oe$ Apr 13 06:57:17 Looks like DNS is down.. try monotone.vanille.de Apr 13 06:59:46 NAbyss: it looks like it. :( Apr 13 07:24:03 hi, host ewi546.ewi.utwente.nl seems to be down, where may i download OE.db.bz2 ? Apr 13 07:36:14 is it save to build a bootstrap-image with the current state in oe.db, i will it break? Apr 13 07:39:11 It's certainly safe in the sense that nothing is likely to be damaged by trying. I think there's a fairly good chance of success, but it depends to a large extent on your target platform. Apr 13 07:40:38 * mithro pokes zekoZeko Apr 13 07:40:45 hrm that wasn't who i meant to poke Apr 13 07:46:24 is this normal ? Apr 13 07:46:29 monotone: finding items to synchronize: Apr 13 07:46:29 monotone: certs | keys | revisions Apr 13 07:46:29 monotone: 15,404 | 45 | 5,037 Apr 13 07:46:29 monotone: bytes in | bytes out | certs in | revs in | revs written Apr 13 07:46:29 monotone: 2.2 M | 279.2 k | 1,926 | 451 | 1 Apr 13 07:46:30 monotone: verifying new revisions (this may take a while) Apr 13 07:46:31 morning Apr 13 07:46:32 monotone: bytes in | bytes out | certs in | revs in | revs written Apr 13 07:46:34 monotone: 2.2 M | 279.2 k | 1,926 | 451 | 305 Apr 13 07:46:36 monotone: read from fd 6 (peer monotone.vanille.de) failed, disconnecting Apr 13 07:46:38 monotone: warning: discarding revision cert packet 07df82a1d560525f8842ebcb6004cad6a910d2c0 with unmet dependencies Apr 13 07:46:44 joy... I managed to get qpobox working with Opie Apr 13 07:46:49 i get 100+ discards Apr 13 07:46:52 finally have a half-decent Japanese input method Apr 13 07:49:04 good morning all Apr 13 07:49:23 morning Koen Apr 13 07:49:35 morning Apr 13 07:49:55 Ifaistos: http://handhelds.org/hypermail/oe/76/7675.html Apr 13 07:50:00 hey Ifaistos Apr 13 07:50:38 i have a (stupid) question about feeds Apr 13 07:50:40 how are they used ? can we build image without compiling anything - just use ipkgs in feeds, or are they only put in ipkg.conf to allow further package install, from the embedded device ? Apr 13 07:57:35 both Apr 13 07:57:49 you can build images purely from the feeds, without compiling anything locally, if you want. Apr 13 07:57:50 victorz: http://bugs.openembedded.org/show_bug.cgi?id=818 Apr 13 07:59:22 koen: but in this case, we have to remove all DEPENDS from bootstrap-image.bb ? Apr 13 08:00:26 (or any *image) Apr 13 08:04:11 morning Apr 13 08:05:25 morning all Apr 13 08:06:11 hi XorA RP Apr 13 08:06:31 ~seen koen Apr 13 08:06:37 koen is currently on #gpe (27m 21s) #oe (27m 21s) #handhelds.org (27m 21s) #openslug (27m 21s). Has said a total of 4 messages. Is idling for 8m 46s, last said: 'victorz: http://bugs.openembedded.org/show_bug.cgi?id=818'. Apr 13 08:07:17 do I need gcc-cross-2.95 to compile a Kernel 2.6 bootstrap-image for SL6000 ? Apr 13 08:07:46 no Apr 13 08:07:51 ~zauruskernels Apr 13 08:07:53 [zauruskernels] By default Zaurus used 2.4.18 or 2.4.20 which need GCC 2.95.3 to compile - all is described in OE wiki: http://openembedded.org/cgi-bin/moin.cgi/ZaurusKernels Now we work on moving all models to 2.6 kernels: http://openembedded.org/cgi-bin/moin.cgi/ZaurusKernels26 Apr 13 08:09:51 thanks, for the hint Apr 13 08:13:09 hey hrw|work Apr 13 08:15:14 koen: ewi died? Apr 13 08:15:38 uf.. working again Apr 13 08:15:46 hrw|work: no, the router at the university had an indigestion Apr 13 08:15:51 ah Apr 13 08:16:06 ~lart IOS Apr 13 08:16:07 * ibot squeezes IOS till IOS turns blue like papa smurf Apr 13 08:16:30 hrw|work: I suspect it started working again when I rejoined this channel Apr 13 08:16:40 ;) Apr 13 08:18:39 victorz: or build bootstrap-image.bb directly with "bitbake -b", yeah Apr 13 08:18:59 or, I guess, yet another option would be to add all the relevant packages to ASSUME_PROVIDED in local.conf Apr 13 08:19:36 pb_: ok, thx Apr 13 08:24:34 good morning Apr 13 08:24:53 hi Bernardo Apr 13 08:24:54 hey Bernardo Apr 13 08:24:56 hey RP Apr 13 08:25:24 morning all Apr 13 08:25:53 Bernardo did you unbrick your akita already? Apr 13 08:26:16 hey katossi Apr 13 08:26:45 RP: did you try to boot an EABI image already? Apr 13 08:27:20 koen: I have booted both gcc4 and gcc3 EABI images Apr 13 08:27:29 nice Apr 13 08:27:36 any difference between the two? Apr 13 08:27:57 Not that I noticed... Apr 13 08:28:46 they were only bootstrap images Apr 13 08:30:18 03coredump 07org.oe.oz354x * r2260646c... 10/packages/udev/ (7 files in 6 dirs): udev: Include a blacklist of devices which are not to be mounted automatically Apr 13 08:30:23 03hrw 07org.oe.oz354x * r2ca42de0... 10/conf/distro/openzaurus-3.5.4.conf: openzaurus-3.5.4: set binutils version to same as binutils-cross Apr 13 08:30:28 03hrw 07org.oe.oz354x * rf53da565... 10/conf/distro/openzaurus-3.5.4.1.conf: openzaurus-3.5.4.1: bump to alpha3 since alpha2 was released to users Apr 13 08:31:44 Hi Apr 13 08:31:49 hey Cobelius Apr 13 08:31:53 hi koen Apr 13 08:32:17 are there environment variables wich contains the path and file name to the toolchain? Apr 13 08:33:44 Cobelius: ${CROSS_DIR} and ${BUILD_PREFIX} Apr 13 08:34:01 Cobelius: most of those variables should be in conf/bitbake.conf Apr 13 08:34:21 RP: any big bugs in the eabi bootstrap images? Apr 13 08:34:45 katossi: I didn't brick it... If you're refering to yesterday, was booting gpe-image from altboot, so I had always opie as a fallback Apr 13 08:34:56 okay, to configure something, can I directly cd to the directory and run configure or make or is there somthing like a wrapper function provided by bitbake Apr 13 08:35:06 i read somthing like oe_runmake Apr 13 08:35:13 Bernardo sorry I misunderstood the problem but... a lot better :) Apr 13 08:35:19 koen: Not that I noticed. I did start to hit compile errors when I tried to build an x based image though (and ran out of disk space so haven't got back to them) Apr 13 08:35:24 Cobelius: 'bitbake devshell' Apr 13 08:35:25 :) thanks Apr 13 08:35:49 I was just testing hrw images for him Apr 13 08:36:44 http://www.theregister.com/2006/04/13/cisco_commute/ Apr 13 08:38:00 when i call oe_runmake in my .bb file, are is the toolchain in the path? Apr 13 08:40:18 oh man bitbake is tricky. Someone told me that it was quite easy, but man i have truoble Apr 13 08:40:20 Here is the error message Apr 13 08:40:25 http://pastebin.de/5874 Apr 13 08:40:42 he can't find the programs for compilation Apr 13 08:40:48 is it a path problem? Apr 13 08:40:52 a problem of the makefile? Apr 13 08:41:20 hrw|work: I was reading that article too Apr 13 08:41:37 http://pastebin.de/5875 this is the .bb file Apr 13 08:44:14 Bernardo: its sick Apr 13 08:45:25 CoreDump|afk: altboot lack LICENSE Apr 13 08:45:35 I have tried to build 'nano' and it build correctly, bit when i try to build a bootstrap-image for the blueangel it complains about : ERROR: Cannot satisfy the following dependencies for task-bootstrap: and it lists up a bounch of modules. But i know that they are in the blueangel_defconfig. Do i have to manually build them? Apr 13 08:45:42 03hrw 07org.oe.oz354x * rfbe7b112... 10/packages/irk/ (irk-belkin_0.11.bb irk-targus_0.11.0.bb): (log message trimmed) Apr 13 08:45:42 irk-belkin/irk-targus: make them Conflict but not Replace/Provide Apr 13 08:45:42 - When they had Conflicts:/Replaces: each other then instaling them was Apr 13 08:45:42 nearly impossible: Apr 13 08:45:42 "ipkg install irk-targus" install irk-belkin (and vice versa) Apr 13 08:45:43 but on "ipkg upgrade" they were exchanged for proper one. Apr 13 08:45:47 - Now user can install proper one in normal way. If will try to install Apr 13 08:46:28 pb_: I know that this changeset is incorrect way of handling this but atleast it works Apr 13 08:49:48 RP: which way of giving list of needed modules for akita/spitz you prefer: include zaurus-cxx00.conf, set them in akita.conf and let spitz.conf include them? Apr 13 08:50:38 hrw|work: I have a 20-30 minutes comute, I just can't imagine what 3.5 hours each direction would be Apr 13 08:51:30 hrw|work: Can't we set them in zaurus-clamshell-2.6.conf? Apr 13 08:52:41 RP: thats good too - will look into Apr 13 08:52:56 Bernardo: I have ~10m, Ania has 30-50m Apr 13 08:53:24 RP: alpha2 lacked some modules Apr 13 08:53:47 hrw|work: Which ones? Apr 13 08:54:31 RP: wep ones, usb host Apr 13 08:55:17 hrw|work: ok. They probably need added to .dev too Apr 13 08:55:18 or users got wrong on them - checking rootfses now Apr 13 08:55:37 hrw|work: if I go on foot, I have the same 20-25 minutes as by car - and my wife also works near me... So we only really use the car when we have to. Apr 13 08:56:07 hrw|work: anything you need me to test today? Apr 13 08:59:21 Bernardo: usb host? Apr 13 09:00:22 RP: ohci-hcd? Apr 13 09:00:30 anything in particular about usb host? Apr 13 09:01:05 I can check with a usb memory dongle, or a bluetooth dongle Apr 13 09:01:17 Bernardo: what need to be added to get it working in alpha2 Apr 13 09:01:27 hrw|work: ohci-hcd, yes, thx ;) Apr 13 09:01:52 well, we need something like CoreDump|afk applet, to let the user switch between host, storage and network... Apr 13 09:02:21 obergix[work]: kernel-module-ohci-hcd kernel-module-usbcore kernel-module-usbhid kernel-module-usbkbd kernel-module-usbmouse kernel-module-usbnet kernel-module-usbserial kernel-module-usb-storage Apr 13 09:02:24 let me see how it reacts to the dongles Apr 13 09:02:26 obergix[work]: something more? Apr 13 09:03:46 hrw|work: hmmm...dunno, that seems to be fine... but I must say i haven't played alot with alpha2 since I was digging the session stuff for matchbox 1.0 crash... Apr 13 09:04:43 hrw|work: sd_mod isn't being loaded automatically, even though scsi_mod is Apr 13 09:05:13 hrw|work: could you add gpe-windowlist_0.1.bb to .oz354x and the upgrades feed? Apr 13 09:05:33 also, my usb dongle has no partition table, so no sda1, only sda, and it won't automount like that Apr 13 09:06:35 koen: can Apr 13 09:06:45 Bernardo: that sd-mod will get added Apr 13 09:06:47 hrw|work: thanks Apr 13 09:06:49 other problems: there is no /media/usbhdd, cp932 isn't installed Apr 13 09:07:00 why cp932 for vfat, btw? Apr 13 09:07:31 03koen 07org.oe.dev * r72f4172a... 10/packages/gpe-windowlist/ (gpe-windowlist_0.1.bb gpe-windowlist_cvs.bb): gpe-windowlist: add 0.1 and fix metadata for _cvs Apr 13 09:07:33 Bernardo: dont know Apr 13 09:07:45 koen: btw - what it is? Apr 13 09:08:10 hrw|work: a windowlist applet like mb-applet-windows Apr 13 09:08:33 but without the libwnck dependency Apr 13 09:08:33 hrw|work: It's in fstab, I think it should default to 437, and the user could change it to its local cp (mine is 860) Apr 13 09:08:58 Bernardo: about sd-mod, I files a bug suggesting appropriate rules for udev... Apr 13 09:09:38 hrw|work: sure the module is needed, but also some fixes to udev... /me is looking for the bug tracker Apr 13 09:10:16 obergix[work]: #811? Apr 13 09:10:19 ~time Apr 13 09:10:21 somebody said time was 1 dimensional, or everlasting, an illusion, or 2006.04.13 9:18:44 GMT Apr 13 09:10:50 http://bugs.treke.net/show_bug.cgi?id=811 indeed Bernardo Apr 13 09:11:25 RP: http://pastebin.com/657257 Apr 13 09:11:54 I'll edit my udev rules and see if that also works with my dongle. Apr 13 09:12:09 btw, what packages have nls codepages? locale-* ? Apr 13 09:12:38 Bernardo: kernel-module-nls-cp860 rather Apr 13 09:13:22 * Bernardo has no kernel-module-nls* Apr 13 09:14:04 I assume they aren't being generated for oz3.5.4.1? Apr 13 09:14:37 anyone know what the time is in Germany? Apr 13 09:14:43 mithro: 11:23 am Apr 13 09:14:53 mithro: GTM +2 (due to daylight saving stuff) Apr 13 09:15:36 koen: any idea where zecke is? :P Apr 13 09:15:52 gone home for easter? Apr 13 09:16:16 iirc this isn't a holiday in .de Apr 13 09:16:22 tomorrow should be Apr 13 09:16:35 mithro: funeral Apr 13 09:16:40 ah, right Apr 13 09:16:44 * koen forgot Apr 13 09:16:52 09 23:15 < zecke> hrw|tv: I will be at a funeral from Tuesday to Thursday, so I could have an r33 by next week Apr 13 09:17:36 hrw|work: poor zecke, any idea who? Apr 13 09:17:48 mithro: not asked Apr 13 09:18:25 obergix[work]: your rule fixes sd_mod loading for my dongle too Apr 13 09:18:51 i have almost finished a present for him :P Apr 13 09:19:16 hrw|work: that looks fine. It's perfectly legitimate for packages to Conflict but not Replace or Provide each other. Apr 13 09:20:07 hrw|work: in particular, Provides is only appropriate if two packages are genuinely interchangeable. Those two sound like they are for different hardware, which would make them (I assume) distinctly non-interchangeable. Apr 13 09:20:30 Bernardo: good news... feel free to followup in the bugtracker ? Apr 13 09:21:13 only i need to ask about some bitbake internals before I can finish it Apr 13 09:21:13 anyone else done bitbake stuff that might be able to help? Apr 13 09:21:14 pb__: I always had problems with understanding that part of debian policy. Apr 13 09:21:35 pb__: I'm about to add libtododb 0.10, do you want me to keep the MAINTAINER field from 0.09? Apr 13 09:21:36 pb__: do you know anything about ConfHandler.py? Apr 13 09:21:48 obergix[work]: already did... :) Apr 13 09:22:10 koen: yes, please Apr 13 09:22:29 pb__: ok, thanks Apr 13 09:22:29 Bernardo: cool, then ;) Apr 13 09:22:30 mithro: a bit Apr 13 09:22:35 what floating point "stuff" (vfp or fpa) does the pxa255 use Apr 13 09:22:42 TheMasterMind1: none Apr 13 09:23:06 floating point is so hype these days ;) Apr 13 09:23:06 pb__: whats the difference between all the equals? ie .=, =. and = Apr 13 09:23:23 mithro: .= and =. are catenations Apr 13 09:23:29 hrm i thought those were soft float implementations? Apr 13 09:23:37 catenations? Apr 13 09:23:39 = is a simple assignment Apr 13 09:23:50 TheMasterMind1: that would be softfpa and softvfp Apr 13 09:24:10 TheMasterMind1: which are software stuff, not hardware Apr 13 09:24:12 oh, ok. well softfpa or softvfp, or either works cuz its just software Apr 13 09:24:16 (does that make sense?) Apr 13 09:24:18 which is preferred Apr 13 09:24:35 so where is = found in ConfHandler? I can't find its "name" Apr 13 09:24:57 hrw|work: You need to add spaces to the append line and add in the same line for akita for usbhost. Otherwise, its fine Apr 13 09:25:06 RP: I did it already Apr 13 09:25:14 ahh .= doesn't do space Apr 13 09:25:41 hrw|work: fine then :) Apr 13 09:25:45 RP: pushed Apr 13 09:26:15 i.e. when i use TARGET_FPU = "soft", which implementation is used Apr 13 09:26:39 TheMasterMind1: softfpa with non-EABI toolchains, softvfp with EABI toolchains Apr 13 09:27:15 03hrw 07org.oe.oz354x * r377b1172... 10/conf/machine/ (zaurus-clamshell-2.4.conf zaurus-clamshell-2.6.conf): zaurus-clamshells: added kernel modules for WEP/WPA and USB Host ones for akita/spitz Apr 13 09:27:15 03koen 07org.oe.dev * r02b5b706... 10/packages/gpe-taskmanager/gpe-taskmanager_0.20.bb: gpe-taskmanager: add 0.20 Apr 13 09:27:17 03koen 07org.oe.dev * rdb0641e2... 10/packages/libxsettings-client/libxsettings-client_0.17.bb: libxsettings-client: add 0.17 Apr 13 09:27:30 which one does OE build by default Apr 13 09:27:51 TheMasterMind1: depends on your $DISTRO Apr 13 09:28:10 hrw|work: can you also add modules for the nls codepages? I'd like at least 860... :) Apr 13 09:28:39 RP: thats question to RP - he maintain kernel defconfigs Apr 13 09:29:08 RP: can you add kernel modules to nls codepages, please? :) Apr 13 09:29:50 what could be the reasons when the toolchain isn't find? Apr 13 09:30:31 Bernardo: ok, I'll try and remember to enable them - If I forget, remind me :) Apr 13 09:30:47 03hrw 07org.oe.oz354x * ra1c631c9... 10/packages/meta/task-bootstrap.bb: task-bootstrap: bump PR due to changes in zaurus-clamshell config Apr 13 09:30:57 hrw|work: I've a few things to fix in the kernel, then I'll sync .dev to oz354x Apr 13 09:31:16 koen: thanks Apr 13 09:31:54 hmmm... viewmtn fails :( Apr 13 09:32:27 ok Apr 13 09:33:47 ?= means assign if empty? Apr 13 09:34:02 yes Apr 13 09:34:02 if not set iirc Apr 13 09:34:09 if not set - right Apr 13 09:34:22 if(!isset(VAR)) VAR = "value" Apr 13 09:34:28 hrw|work: the bluetooth dongle worked without any problem Apr 13 09:34:42 Bernardo: k Apr 13 09:34:55 yay working free java on arm! Apr 13 09:35:06 TheMasterMind1: where? Apr 13 09:35:37 as for the wifi dongle, I have the rt2570 driver building well, but the usb otg port doesn't provide enough power for wifi Apr 13 09:35:41 TheMasterMind1: gjc has been quite functional for while now :) Apr 13 09:36:01 none of the VMs have worked with floats Apr 13 09:36:19 03koen 07org.oe.dev * r6cc83c7e... 10/packages/libtododb/libtododb_0.10.bb: libtododb: add 0.10 Apr 13 09:36:23 03koen 07org.oe.dev * rd5f19a09... 10/packages/gpe-calendar/gpe-calendar_0.72.bb: gpe-calendar: update to 0.72 Apr 13 09:36:28 03koen 07org.oe.dev * rd5621697... 10/packages/libeventdb/libeventdb_0.21.bb: libeventdb: add 0.21 Apr 13 09:36:32 03koen 07org.oe.dev * r143b0545... 10/packages/gpe-clock/gpe-clock_0.24.bb: gpe-clock: add 0.24 Apr 13 09:36:59 TheMasterMind1: cvm did, but that isn't free Apr 13 09:37:25 koen: which of those gpe updates are good to provide for users? Apr 13 09:37:28 right Apr 13 09:37:46 koen: jamvm 1.4.2 + classpath-0.90 + patch works Apr 13 09:37:49 RP: is ntfs support enabled for the akita kernel? Apr 13 09:38:07 http://gcc.gnu.org/bugzilla/show_bug.cgi?id=27142 Apr 13 09:38:11 Bernardo: I'd doubt it Apr 13 09:38:51 hey RP Apr 13 09:39:22 hi mithro Apr 13 09:40:19 i'm doing some bitbake work atm :P Apr 13 09:40:46 mithro: What's your aim? Apr 13 09:41:03 I was thinking on later testing with my usb hd, but it is formatted ntfs (80Gb) Apr 13 09:41:08 to finish zecke's c parser Apr 13 09:41:33 mithro: That would be great :) Apr 13 09:41:42 Gah. Mainline have broken pcmcia on the zaurii :-( Apr 13 09:41:50 problem solved Apr 13 09:41:55 RP: i don't think the speed up is going to be as big as it first appears Apr 13 09:42:03 Bernardo: it worked for me... but more testing can't hurt ;) Apr 13 09:42:14 Bernardo: with a hub though ;) Apr 13 09:42:21 do we know where most of our cpu time is going? Apr 13 09:42:25 hrw|work: libeventdb, libtodotb and gpe-calendar Apr 13 09:42:34 mithro: The work zecke did lead him to the coclusion the problem was elsewhere Apr 13 09:42:51 mithro: We think its disk IO - look at the size of tmp/cache Apr 13 09:43:09 hrw|work: but I think you can wait a few days for even newer versions to appear in .dev Apr 13 09:43:11 what does tmp/cache actually do? Apr 13 09:43:12 mithro: We've discussed some ideas about a secondary cache to work around this Apr 13 09:43:31 mithro: Speeds up secondary runs of bitbake to save it reparsing all the files again Apr 13 09:43:37 RP: tried running is from a tmpfs? Apr 13 09:43:48 obergix[work]: this has a power adapter, so won't need a powered hub Apr 13 09:43:54 koen: no, but that would be interesting Apr 13 09:44:08 koen: ok - give me info when thery will be ready Apr 13 09:44:15 obergix[work]: the only thing I'd like to work but needs a powered hub is my wifi dongle Apr 13 09:44:16 koen: I still thing that processing 300MB of data takes too long, esp. when we can avoid doing it all together Apr 13 09:44:26 RP: indeed Apr 13 09:44:33 RP: i think the c parser might reduce the need for the cache Apr 13 09:44:35 RP: it's ~400MB over here.... Apr 13 09:44:48 hrw|work: will do that Apr 13 09:44:52 however i think you are loosing alot of speed to string concat Apr 13 09:44:57 koen: It grows everything we add classes :-/ Apr 13 09:45:17 (specially in the parser side) Apr 13 09:45:36 mithro: With the speedups zecke made to the python parser, I'll be interested to see what happens without the cache. I think the cache disable option isn't working entirely correctly atm though... Apr 13 09:45:40 TheMasterMind1: what is the problem with jamvm and floats exactly? Apr 13 09:46:02 the python parser is still horribally inefficent Apr 13 09:46:11 mithro: agreed Apr 13 09:46:29 this python code makes my brain hurt Apr 13 09:46:30 mithro: Hopefully we can benefit from all the different things to make bitbake fast :) Apr 13 09:46:50 * koen drops the word "multithreaded" Apr 13 09:47:19 koen: I know how to do this and will at some point :) Apr 13 09:47:23 katossi: the problem is in classpath, in converting floats to ascii Apr 13 09:47:33 koen: It will be a machine killer :) Apr 13 09:48:11 RP: ask sun for a T1 to T2 machine to test on :) Apr 13 09:48:40 or CP|CELF for a 8 way opteron Apr 13 09:48:42 TheMasterMind1: how big is your experience with gcj in arm? Have you tried with complex software?( I mean basically GUI) Apr 13 09:48:53 * RP ponders the effect on ewi's bandwidth :) Apr 13 09:49:19 katossi: i've never tried gcj. Apr 13 09:49:25 how will multithreading help? Apr 13 09:49:32 isn't parsing bb's basically linear? Apr 13 09:49:43 mithro: It will help the actual imae build speed Apr 13 09:49:45 the eclipse java compiler (which is written in java) though should work fine on arm Apr 13 09:49:51 mithro: Its a separate objective Apr 13 09:50:01 ahh okay Apr 13 09:50:41 RP: ~16 alpha2 was released Apr 13 09:50:42 does anyone know of a usb host in cf card format, I would like to extend my SL-5600 Apr 13 09:50:57 sws: ratoc card Apr 13 09:51:10 sws: they cost a small fortune Apr 13 09:51:20 are there any prebuilt arm-ucblic-linux toolchains availabe? Apr 13 09:51:21 TheMasterMind1: oh really? I asked some devs before and they told me ecj would not work with arm Apr 13 09:51:21 mithro: The basic idea is to have the build chain + depencency table dumped into an SQL database, then spawn multliple bitbake tasks in parallel Apr 13 09:51:40 katossi: i tested it last night. you have to use it with jamvm, not standalone. Apr 13 09:51:50 what is the difference between require and include? Apr 13 09:52:12 mithro: require raises an error if it can't fulfil the request Apr 13 09:52:24 mithro: I'd assume the difference is it falls over if require fails, whereas include would generate a warning? Apr 13 09:52:44 include is silent upon failure Apr 13 09:53:52 how well are the ratoc cards supported in zaurus kernel? Apr 13 09:54:04 TheMasterMind1: those are great news to me Apr 13 09:55:21 sws: the ratoc cards work pretty well Apr 13 09:55:23 sws: they are supported Apr 13 09:55:29 sws: at least the one on my desk does Apr 13 09:57:38 if i take my ${BUILD}/cross directory and put it on another machine in the same location it should work correct? Apr 13 09:58:32 yes Apr 13 09:58:51 TheMasterMind1: I moved whole builddir between machines (both was athlonxp boxes) Apr 13 09:59:15 thanks, only the price prevents me from buying one ;) **** ENDING LOGGING AT Thu Apr 13 09:59:57 2006