**** BEGIN LOGGING AT Sat Sep 29 02:59:56 2007 Sep 29 03:41:11 cbrake: I'm currently pondering over how to best include /usr/lib/mono/gac into the mono packages. The code doesn't get built at all during the cross-compile step (you had --disable-mcs-build, but I found that the mcs build is hard-coded disabled when cross-compiling) Sep 29 03:42:08 cbrake: instead http://www.mono-project.com/Mono:ARM suggests to do a buildhost-native build for the managed code and a cross-compile build for the targethost-native code Sep 29 03:45:44 cbrake: so one could just go ahead and re-package the files that get installed by mono-native. It's huge, though: 50MB for the gac directory (about half of which are .mdb files with debugging information) Sep 29 03:57:11 03mwester 07org.oe.dev * r56e95791... 10/ (3 files in 3 dirs): libgsmd: Fix return value from lgsm_send(), this fixes the dialer failure on boot. Sep 29 04:25:13 Crofton, ping Sep 29 07:06:40 now we know what distro harald uses: http://hughsient.livejournal.com/39074.html Sep 29 07:20:51 help Sep 29 07:21:04 Good Morning Sep 29 07:24:45 Just installed a fresh bitbake version from trunk. Unfortunately "setup.py" complains "cooker.py line 832 - finally:" syntax error. Anyone here to help ? Sep 29 07:57:50 03koen 07org.oe.dev * r4d114b16... 10/ (4 files in 3 dirs): dbus: sync with poky and use .inc file Sep 29 07:57:56 03koen 07org.oe.dev * r065d0169... 10/ (5 files in 3 dirs): gnome-keyring: sync with poky Sep 29 07:58:01 03koen 07org.oe.dev * r9672e7dc... 10/ (3 files in 2 dirs): iproute2: sync with poky Sep 29 07:58:08 03koen 07org.oe.dev * rcd35c63e... 10/ (6 files in 3 dirs): network-manager svn: sync with poky Sep 29 07:58:13 03koen 07org.oe.dev * rcffa4cb4... 10/ (17 files in 5 dirs): packages/gnome: sync with poky, while maintaining OE tweaks like avahi and gnutls Sep 29 08:21:12 03koen 07org.oe.dev * re26357a9... 10/ (21 files in 7 dirs): udev: sync with poky, def_pref 115 negative Sep 29 08:21:17 03koen 07org.oe.dev * r2dc57008... 10/ (1 classes/gnome.bbclass): gnome.bbclass: sync with poky Sep 29 08:26:28 * koen waits for the autobuilder to catch up Sep 29 08:40:02 koen: dbus staging fails for me since latest checkin Sep 29 08:40:53 | /home/moko/build/tmp/work/armv4t-angstrom-linux-gnueabi/dbus-1.0.2-r5/temp/run.do_stage.2317: line 528: cd: glib: No such file or directory Sep 29 08:41:22 (tried a rebuild of the dbus package, same result) Sep 29 08:57:32 03koen 07org.oe.dev * r6ed6e2ae... 10/ (1 packages/dbus/dbus.inc): dbus.inc: remove libinstall for dbus-glib Sep 29 09:04:14 koen: that fixed it, thx. Sep 29 09:43:45 03tmbinc 07org.oe.dreambox * r10c5f2fa... 10/ (3 files in 2 dirs): enigma2,enigma2-plugins: Add descriptions for plugins, add proper RDEPENDs for plugins Sep 29 09:45:09 hey zecke & pH5 Sep 29 09:45:37 03tmbinc 07org.oe.dreambox * r10c5f2fa... 10/ (3 files in 2 dirs): enigma2,enigma2-plugins: Add descriptions for plugins, add proper RDEPENDs for plugins Sep 29 09:46:21 hi koen, hi zecke Sep 29 09:50:32 moin Sep 29 09:52:36 03koen 07org.oe.dev * r69e0da18... 10/ (3 files in 2 dirs): iproute2: fix include ordering Sep 29 09:53:41 03tmbinc 07org.oe.dreambox * r10c5f2fa... 10/ (3 files in 2 dirs): enigma2,enigma2-plugins: Add descriptions for plugins, add proper RDEPENDs for plugins Sep 29 09:53:46 03koen 07org.oe.dev * r69e0da18... 10/ (3 files in 2 dirs): iproute2: fix include ordering Sep 29 10:03:08 I have a problem with binutils 2.18 during kernel compilation in certain situations. Sep 29 10:03:18 The standard OE kernel for my devices work fine, but on my custom kernel build (upstream + minimal device patches) Sep 29 10:03:29 the arch/arm/boot/Image ends up as a 3.1GB (sparse) file (it is created by objcopy -S -O binary from vmlinux), and the resulting zImage is 4.3 MB Sep 29 10:03:32 Any idea how to debug this? Sep 29 10:05:12 pH5: hehe Sep 29 10:05:25 pH5: well, there is a fix to the linker script Sep 29 10:05:48 pH5: it is used in poky and was posted to the arm lkml Sep 29 10:07:52 Zecke: Can you come to NL for t-dose ? Sep 29 10:09:23 zecke: heh, thanks. that is exactly the issue I'm seeing. Sep 29 10:09:25 pH5: http://svn.o-hand.com/view/poky/trunk/meta/packages/linux/linux-rp-2.6.22%2B2.6.23-rc4/binutils-buildid-arm.patch?rev=2726&view=markup Sep 29 10:09:49 koen: ta Sep 29 10:10:02 and http://lists.linuxtogo.org/pipermail/openembedded-commits/2007-September/009053.html Sep 29 10:10:11 Koen ping Sep 29 10:10:13 packages/linux/linux-davinci/binutils-buildid-arm.patch Sep 29 10:10:19 gerwinin: pong Sep 29 10:10:35 Koen kun jij een lezing geven over oe op t-dose ? Sep 29 10:10:42 when's t-dose? Sep 29 10:11:02 13 and 14 october Sep 29 10:11:10 We will be having a booth there and a slot Sep 29 10:13:24 The booth I can man but I do not want to tell something about oe because I am not to good at that (more user than developer) Sep 29 10:14:53 hmm, that's right after OEDEM Sep 29 10:15:24 I was thinking of asking leon as well Sep 29 10:21:20 morning all Sep 29 10:21:41 koen: Thanks for the poky merges earlier on :) Sep 29 10:21:54 koen: I'm getting a little worried about the classes getting out of sync again :/ Sep 29 10:33:34 RP: things need to get merged back as well :( Sep 29 10:34:05 e.g. our wpasupplicant fixes Sep 29 10:34:40 OE's wpa-supplicant is controlled via /etc/network/interfaces and works on ifup, pokies uses /etc/wpa-supplicant.conf and works on sysvinit Sep 29 10:34:49 (OE follows debian) Sep 29 10:35:19 koen: Ours is goverened by /etc/networik/interfaces too Sep 29 10:35:38 koen: The problem there is I have no idea what I'm doing and therefore tend to leave it alone :/ Sep 29 10:36:35 http://ewi546.ewi.utwente.nl/tmp/no-idea-what-I-am-doing-with-wpa.diff Sep 29 10:36:40 that's my latest try Sep 29 10:36:47 but since I don't use WPA.... Sep 29 10:37:04 ../../../WebCore/page/Frame.h:147: error: 'DragImageRef' does not name a type Sep 29 10:38:24 koen: I think I have even less idea :/ Sep 29 10:44:32 koen: Whats your opinion on the cross change I proposed? Sep 29 10:46:50 RP: I'd go for it Sep 29 10:47:01 I haven't looked at your .packaged changes yet Sep 29 10:47:21 koen: I just pushed the packaged ones ;-) Sep 29 10:50:15 Morning Sep 29 10:50:23 hi sirfred Sep 29 10:50:35 RP: Hi Sep 29 10:51:28 Just upgraded to gutsy, and monotone was bumped up to 0.35, is that supposed to work against OE? I've just seen the manual and talks about 0.32 or newer, so I suppose I'm safe, true? Sep 29 10:52:57 It asks me to migrate the mtn database, though. Any experience about that? Is it a safe process? Lengthly? Sep 29 10:54:43 * sirfred thinks monotone should change their pet picture, it's a little repulsive Sep 29 10:54:59 sirfred: I can't remember it taking any significant amount of time. and it should be safe, but as always, backup your database :) Sep 29 10:55:18 sirfred: I also suspect its ok to do but make a backup first ;-) Sep 29 10:55:44 The backup, I always forget that part. Sep 29 10:55:51 Thanks, guys ;-) Sep 29 10:59:46 It worked fine, pulling now Sep 29 11:04:36 RP: at OEDEM we need to talk about a formal review process as well for changes to e.g. udev, gcc and glibc Sep 29 11:05:41 with 'formal' I mean 'core developers are forced to look at it' Sep 29 11:06:23 koen: Agreed Sep 29 11:07:15 03rpurdie 07org.oe.dev * r64ce6ddf... 10/ (1 classes/insane.bbclass): insane.bbclass: Raise exceptions, not strings (from poky) Sep 29 11:13:28 03rpurdie 07org.oe.dev * r64ce6ddf... 10/ (1 classes/insane.bbclass): insane.bbclass: Raise exceptions, not strings (from poky) Sep 29 11:14:42 why does bot report same thing twice? Sep 29 11:15:18 Uff, the pidgin irc plugin is having a bad time trying to parse the colour control codes Sep 29 11:15:34 * sirfred considers reverting to xchat Sep 29 11:17:02 03rpurdie 07org.oe.dev * r64ce6ddf... 10/ (1 classes/insane.bbclass): insane.bbclass: Raise exceptions, not strings (from poky) Sep 29 11:17:10 03rpurdie 07org.oe.dev * r5e202399... 10/ (1 classes/kernel.bbclass): kernel.bbclass: Package vmlinux file (used by oprofile) (from poky) Sep 29 11:17:15 03rpurdie 07org.oe.dev * r63b088d8... 10/ (1 classes/kernel.bbclass): kernel.bbclass: Add menuconfig task (from poky) Sep 29 11:17:21 03rpurdie 07org.oe.dev * r75c6a806... 10/ (1 classes/package.bbclass): package.bbclass: Fix .packaged pkgmaps data creation to fix depchains. Clean up/optimise PACKAGES checks a bit Sep 29 11:17:27 03rpurdie 07org.oe.dev * rd7765554... 10/ (1 classes/package_deb.bbclass): package_deb.bbclass: Map TARGET_ARCH to DPKG_ARCH to keep dpkg/apt happy with 'valid' archs (from poky) Sep 29 11:17:32 03rpurdie 07org.oe.dev * r56fdbd9f... 10/ (1 classes/package_deb.bbclass classes/package_ipk.bbclass): package_(deb|ipk).bbclass: Rename package index stamp to make it more obvious what its for (from poky) Sep 29 11:17:38 03rpurdie 07org.oe.dev * r25a3f79f... 10/ (1 packages/meta/cross-linkage_1.0.bb): Add cross-linkage from poky Sep 29 11:17:43 03rpurdie 07org.oe.dev * r0ed57b3c... 10/ (10 files in 2 dirs): linux-libc-headers: Add dependency on cross-linkage, no longer need to install headers to cross/ (from poky). Sep 29 11:17:49 03rpurdie 07org.oe.dev * re2522944... 10/ (8 files in 2 dirs): glibc-initial: Can install to staging instead of cross due to cross-linkage Sep 29 11:17:55 03rpurdie 07org.oe.dev * ra0308061... 10/ (15 files in 2 dirs): glibc: No longer need to install to cross as well as staging (from poky) Sep 29 11:18:02 03rpurdie 07org.oe.dev * r8f071ffb... 10/ (12 files in 2 dirs): gcc-cross: Update paths to reference staging instead of cross (from poky), bump PR to make sure gcc-cross reinstalls after cross changes to ensure everything works correctly Sep 29 11:18:06 03rpurdie 07org.oe.dev * r90cc554a... 10/ (9 files in 2 dirs): glibc-initial: Clear PACKAGES_DYNAMIC to avoid bogus values Sep 29 11:20:20 * koen makes the autobuilder to a build from scratch Sep 29 11:45:08 scruggs, pong Sep 29 11:46:24 what benefit do I get from RP's latest changes? Sep 29 11:47:02 Crofton|home: a clean rebuild ;) Sep 29 11:48:40 working on srarting that already :) Sep 29 11:48:56 does this mean one less build of gccc/glibc? Sep 29 11:55:00 koen: You shouldn't have to do a rebuild Sep 29 11:55:15 koen: rebuilding gcc-cross should be sufficient (which is why I bumped the PRs) Sep 29 11:55:46 Crofton|home: it means two less "make install" cycles of glibc (and they're slow) Sep 29 11:56:23 Crofton|home: It also means we have one libs and one headers directory instead of two with random build failures due to files being in one and not the other etc Sep 29 11:57:05 RP, thanks, that helps with the questions "Why do I have to nuke tmp again" Sep 29 11:57:32 Crofton|home: As I said, you shouldn't have to Sep 29 11:57:49 Well, I don't trust you :) Sep 29 11:58:00 and I need to work on non OE stuff Sep 29 11:58:09 so I am runing my autobiuld Sep 29 11:58:35 best find any problems early Sep 29 11:58:46 I don't mind if people rebuild, just don't make it sound like it was forced ;-) Sep 29 12:00:47 koen: Angstrom is no longer a Linux-only? ;-) Sep 29 12:04:13 hm, anyone here have icecc+OE working? Sep 29 12:05:04 psokolovsky: it never was Sep 29 12:05:21 koen: indeed? ;D Sep 29 12:05:48 it only seemed so because idiots kept adding 'linux' into the wiki Sep 29 12:05:51 ~lart wikis Sep 29 12:05:51 * ibot blasts wikis with a huge firehose then strangles wikis with it Sep 29 12:06:03 03pH5 07org.oe.dev * r7c6cf112... 10/ (1 packages/python/python-2.5-manifest.inc): python-xml: RDEPEND on python-netclients (saxutils imports urrlib) Sep 29 12:06:09 03pH5 07org.oe.dev * rdf803cf5... 10/ (1 packages/python/python-2.5-manifest.inc): python-netclient: RDEPEND on python-stringold (urllib imports string) Sep 29 12:06:16 03pH5 07org.oe.dev * r6a695cd0... 10/ (3 files in 3 dirs): Sep 29 12:06:16 python-native: drop NOTEs from cross-distutils.patch Sep 29 12:06:16 * these unexpected prints to stdout can confuse the SCons build system Sep 29 12:07:24 koen: Oh, I bet, it always has been philosophical concept on monad-like nature. But stupid people love specifics, they have no idea what is "a distribution" (indeed, what it is)? they better understand "an OS distribution". And if you stuff specific OS name there, there will never mix it with Windows, you know. Sep 29 12:08:21 * RP -> back later Sep 29 12:09:13 l8r Sep 29 12:14:32 koen: also, poor you, you wasted lotsa time fighting stupid gmane for putting angstrom under "embedded", or what it is? Now, you need to raise crusade #2, to overcome it being under linux distros. Do it right this time - request top level group "angstrom", because, eh, poor people will never understand what it is anyway. Sep 29 12:16:40 koen: also, make sure you exterminate *all* occurances of "linux distribution" from wiki - those assholes which use google to search for stuff should die. On the contrary, my neighbour, who runs cat food distribution service, should have it as the first hit. Sep 29 12:16:56 psokolovsky: it's wasn't fighting gmane as much as fighting someone who took the liberty to subscribe angstrom to gmane without asking Sep 29 12:18:54 koen: you understand what I'm talking about, right? Stop confuse people with abstract logic, use pragmatic logic instead. Angstrom *does* run on Linux, so it *Linux* distribution. I fit runs also on BSD, call it "Linux and BSD distribution", not just "distribution". Disclaimer: YMMV Sep 29 12:21:56 koen: and btw thanks for fixing wiki Sep 29 12:33:39 ho Sep 29 12:33:44 and brb Sep 29 12:34:38 erm Sep 29 12:34:44 what happened with oe site? Sep 29 12:34:50 Unable to connect to database Sep 29 12:38:30 hm, someone DDoSing site? Sep 29 14:26:50 Does psyco work with x86_64? Sep 29 14:30:44 dcordes: it doesn't Sep 29 14:36:56 morning Sep 29 14:39:48 koen: is there a reason why the class setting in /etc/bluetooth/hcid.conf defaults to PDA? Would there be any objection to changing it to more capable class value (a computer that can do networking (dund, pand), object transfer (obex), etc ? Sep 29 14:40:38 aren't class and services 2 completely different things? Sep 29 14:41:08 * koen is pretty sure sdp exists for a reason Sep 29 14:41:26 yes, I guess what I mean is to set the services to something more capable, and set the class to something more generic. Sep 29 14:41:56 koen: a lot of devices (like the Treo650) use the service to present a list of devices for pairing. Sep 29 14:41:58 the class should match the machine (we should add it to machine.conf) Sep 29 14:42:36 I agree we should start and advertize more services by default for most machines Sep 29 14:42:43 if the device does not have the right service (i.e. networking), then you can't select it for pairing for dund or pand (on the Treo650 this is true at least). Sep 29 14:42:55 ok, good. Sep 29 14:43:12 I'd like sdp, dund and pand servers started (and advertised) by default. Sep 29 14:43:42 just checking what the n800 needs right now to see the neo Sep 29 14:45:18 the n800 needs the class iirc Sep 29 14:45:26 (braindead, but that's nokia for you) Sep 29 14:45:58 there was a big thread on "n800 doesn't see my windows mobile smartphone" on the maemo lists Sep 29 14:46:15 I see the service, and sdp, as the coarse, and fine-grained selection mechanisms. Sep 29 14:46:34 (respectively) Sep 29 14:48:28 you could use bluez-gnome on the neo to configure all that Sep 29 14:48:52 'could' because mickeyl deliberately broke tap-n-hold, so you'll have to fix that locally first Sep 29 14:49:23 I think I'll do it in /etc/bluetooth/hcid.conf - the gui people can't break that. Sep 29 14:49:36 koen do you have a minute ? Sep 29 14:49:50 (and then it will also work on the nslu2 and fsg-3) Sep 29 14:54:56 Crofton, ping Sep 29 14:55:01 pong Sep 29 14:55:08 hey Sep 29 14:55:26 just registerning with the bugtracker now Sep 29 14:55:41 thanks Sep 29 14:55:47 I may look this afternoon Sep 29 14:56:04 no rush Sep 29 14:56:43 does it work? Sep 29 14:57:10 haven't flashed a rootfs with it yet, but will in the next hour Sep 29 14:57:41 the build output is exactly what I get in the gumstix buildroot, so I'm pretty optimistic Sep 29 14:59:39 great Sep 29 14:59:50 bother procps just failed to build for me Sep 29 15:07:16 Crofton: just submitted as bug 3080 Sep 29 15:08:09 thanks Sep 29 15:08:16 while building procps I get thhis: Sep 29 15:08:18 http://rafb.net/p/9RDPQx20.html Sep 29 15:11:32 koen: in what extent does psyco influence the speed of OE builds averagely? I'm trying to find out if is better using 64 or 32 bit enviornemnt on my machine. Sep 29 15:12:15 I ready on the psyco website that it makes pythong processes 2-10 times faster but what does that mean for OE Sep 29 15:12:35 s/ready/read Sep 29 15:12:44 scruggs, I can not read the tarball Sep 29 15:13:06 hmmm.. I'll try again Sep 29 15:14:36 unzipping it gets me a text file containing a directory listing Sep 29 15:14:56 bbiab Sep 29 15:15:14 must have fat-fingered the command Sep 29 15:15:19 :) Sep 29 15:24:37 Crofton, 2nd attempt at the upload seems corrupt too :/ Sep 29 15:25:18 koen: do you remember a merge where you manually had to resolve the conflict? Sep 29 15:28:45 can you untar b4 uploading? Sep 29 15:29:09 yep Sep 29 15:29:15 email it to me Sep 29 15:29:28 tar -czf wifistix.tar.gz wifistix/ Sep 29 15:29:30 will do Sep 29 15:33:42 mwester, I ran bitbake -g openmoko-devel-image and then grep'd both depends.dot and task-depends.dot but still no pango-directfb just pango so it must be randomness :-/ Sep 29 15:35:04 scruggs, got it Sep 29 15:35:18 great Sep 29 15:58:25 03pH5 07org.oe.dev * r6ee8d457... 10/ (1 conf/distro/include/sane-srcrevs.inc): moko-gtk-engine: fix and update SRCREV in sane-srcrevs.inc Sep 29 16:09:47 03pH5 07org.oe.dev * r6ee8d457... 10/ (1 conf/distro/include/sane-srcrevs.inc): moko-gtk-engine: fix and update SRCREV in sane-srcrevs.inc Sep 29 16:11:40 03pH5 07org.oe.dev * r6ee8d457... 10/ (1 conf/distro/include/sane-srcrevs.inc): moko-gtk-engine: fix and update SRCREV in sane-srcrevs.inc Sep 29 16:25:03 03pH5 07org.oe.dev * r6ee8d457... 10/ (1 conf/distro/include/sane-srcrevs.inc): moko-gtk-engine: fix and update SRCREV in sane-srcrevs.inc Sep 29 16:37:48 03pH5 07org.oe.dev * r6ee8d457... 10/ (1 conf/distro/include/sane-srcrevs.inc): moko-gtk-engine: fix and update SRCREV in sane-srcrevs.inc Sep 29 16:39:23 CIA-3: enough already Sep 29 16:41:04 03pH5 07org.oe.dev * r6ee8d457... 10/ (1 conf/distro/include/sane-srcrevs.inc): moko-gtk-engine: fix and update SRCREV in sane-srcrevs.inc Sep 29 16:41:10 03pH5 07org.oe.dev * r151b70ba... 10/ (1 packages/angstrom/angstrom-x11-base-depends.bb): angstrom-x11-base-depends: inherit task.bbclass Sep 29 16:41:15 03pH5 07org.oe.dev * rcd21d75c... 10/ (4 files in 3 dirs): guichan: make libguichan-sdl actually link against sdl-image Sep 29 16:41:20 03pH5 07org.oe.dev * r8621de58... 10/ (5 files in 3 dirs): neod: make it work with magician and hx4700 Sep 29 16:44:39 ERROR: Unable to generate local paths for SRC_URI due to malformed uri: ${SRC_URI_EXTRA} while parsing /home/mteira/oeroot/org.openembedded.dev/packages/dbus/dbus_1.0.1.bb Sep 29 16:45:13 And the only usage of SRC_URI_EXTRA is in dbus.inc Sep 29 16:45:43 Is that deprecated or something so? Sep 29 16:53:12 sirfred: that might be a sync with poky gone wrong Sep 29 16:53:14 koen, RP? Sep 29 16:55:17 Sounds a bit like it although I didn't merge that bit... Sep 29 17:00:20 http://www.openembedded.org:1081/revision/info/4d114b16039cc9044c3762876ed8924145eba0f3 Sep 29 17:01:16 pH5: That's where the SRC_URI_EXTRA got added Sep 29 17:43:09 03pH5 07org.oe.dev * r1d32c1d9... 10/ (3 files in 3 dirs): xserver-kdrive-common: fix path for Xw100 and Ximageon Sep 29 17:47:37 pH5: what about making Xw100 and Ximageon machine specific in the kdrive scripts? Sep 29 17:48:20 the if -e /usr/bin/Xw100 check blows up if you reuse rootfs (e.g. on nfs) Sep 29 17:57:36 procps build is broken .... Sep 29 18:03:19 03pH5 07org.oe.dev * r1d32c1d9... 10/ (3 files in 3 dirs): xserver-kdrive-common: fix path for Xw100 and Ximageon Sep 29 18:03:42 03koen 07org.oe.dev * r2e2e83df... 10/ (1 packages/dbus/dbus.inc): dbus.inc: remove unused var Sep 29 18:09:50 koen: hm, right. and I wonder how to add the -mouse and -keybd parameters for 1.4.. Sep 29 18:09:54 does anything still use Xipaq? Sep 29 18:10:13 Xipaq is over 4 years old AFAIK Sep 29 18:13:39 Crofton: the wifistix drivers load but I'm not seeing the mwlan0 device get created Sep 29 18:13:58 any messages? Sep 29 18:14:07 I don't think it appears in /dev Sep 29 18:14:14 try iwconfig Sep 29 18:15:11 didn't add iwconfig to this image - thought I'd at least see it with ifconfig -a Sep 29 18:15:40 <4>cfio: module license 'unspecified' taints kernel.<4>Before register driver<4>After register driver Sep 29 18:16:07 I see those same messages with buildroot, but some added net messages too Sep 29 18:16:55 hmm Sep 29 18:16:58 no idea :( Sep 29 18:17:02 loading the pcmcia modules also doesn't seem to find my cf card, which usually shows up as /dev/hda1 Sep 29 18:17:26 the driver has a debug option, so I'll try that and see if that leads me anywhere Sep 29 18:17:31 koen: what do you think about something like this: http://en.pastebin.ca/719853 ? Sep 29 18:17:31 ok Sep 29 18:17:41 I can check pcmcia behavior on the OSK Sep 29 18:17:52 but not much real thinking work until Monday Sep 29 18:17:59 Today is cut and paste day :) Sep 29 18:18:04 that would be great Sep 29 18:18:13 I understand :) Sep 29 18:18:34 scruggs: could you send me a copy? I'll take a look too Sep 29 18:18:57 sakoman, sure thing Sep 29 18:19:04 let me try and push what he sent me also Sep 29 18:19:31 I need to wait until I get my tool chain rebuilt here though Sep 29 18:19:39 I did have to run update-modules before modprobe would find them - if that makes any difference Sep 29 18:19:55 I had rebuild today too Sep 29 18:21:20 sakoman: http://www.flickr.com/photos/koenkooi/1458801696/ Sep 29 18:23:07 koen: wow! that is an amazing sky shot. did you use a filter? Sep 29 18:23:17 red or polarizing? Sep 29 18:24:17 sakoman: orange Sep 29 18:24:36 pH5: yes, expect that you confused collie/poodle with shepher/corgi/boxer :) Sep 29 18:25:11 Has fic anounced when the next version of OpenMoko hardware is due out? I thought the anouncemant was due this month.. Sep 29 18:25:49 sakoman: just emailed you the files Sep 29 18:25:52 wifistix-modules_5.0.16.p0.bb Sep 29 18:25:53 I keep checking their website and the current one is sold out so I'm assuming you won't be able to get one until the next version is out.. Sep 29 18:25:59 Is this an ok name for a bb file? Sep 29 18:26:03 mainly the PV? Sep 29 18:26:24 there is still add added revision to that Sep 29 18:26:30 but I was wondering myself Sep 29 18:29:37 svolpe_gerrath: they said hardware will be available in december, no word on if that's the developer preview or mass market Sep 29 18:29:46 svolpe_gerrath: I guess it's the developer preview Sep 29 18:30:06 I'm going to set PR back to 1 Sep 29 18:30:31 you mean to r0 ? Sep 29 18:30:53 thats what I meant! Sep 29 18:30:58 really .... Sep 29 18:31:09 koen, do you have an issue with the PV? Sep 29 18:31:23 koen, I wish I would not have missed out on the fist go around, oh well, December is not too long to wait :-) Sep 29 18:31:37 cf8385-5.0.16.p0-26306.tbz Sep 29 18:31:42 is the base file name Sep 29 18:31:50 my dear lord Sep 29 18:32:06 is that made by mplayer developers? Sep 29 18:32:15 and its being downloaded directly from gumstix - not from Marvel Sep 29 18:32:39 afaik that source isn't downloadable anywhere else, but it has a GPL license Sep 29 18:33:06 apart from http://projects.linuxtogo.org/projects/marvell8385/ and olpc git? Sep 29 18:33:19 ah Sep 29 18:33:39 is there already a bb file for this? Sep 29 18:33:45 could be - I know as a gumstix user we could only get binary drivers up till a few months ago Sep 29 18:34:36 built by Craig Sep 29 18:35:52 hmmm Sep 29 18:36:13 It looks like no one has create any bb file for this yet Sep 29 18:37:08 in the interests of getting the gumstix stuff working, I'm going to push the gumstix solution Sep 29 18:37:33 but we should look hard at getting the one from linuxtogo working for gumstix also Sep 29 18:38:16 I'll ask Craig if there is any reason why he wasn't using the ltg version Sep 29 18:38:49 good idea Sep 29 18:39:08 this will be a bit of an education for me, I haven't worked with any kernel module bb files Sep 29 18:39:36 me neither :) Sep 29 18:56:47 ~praise the autbuilder for saving me a bug report! Sep 29 18:56:48 All hail the autbuilder for saving me a bug report!! Sep 29 19:01:01 got mwlan0 to show up! Sep 29 19:01:18 what is the trick? Sep 29 19:01:18 had to make sure the mmc modules didn't get loaded Sep 29 19:01:33 weird Sep 29 19:01:35 and make sure to modprobe pxa2xx_cs Sep 29 19:02:20 guess I need to build wireless tools to make sure it really works Sep 29 19:04:30 this is probable an easy one, but how do I configure what modules are autoloaded? Sep 29 19:04:48 not as easy as you thinnk ... Sep 29 19:05:04 there is an example in the omap5912osk machine Sep 29 19:05:21 so it has to be set when you build the image? Sep 29 19:05:25 the problem is how do we decide what to autload for a gumstix Sep 29 19:05:37 since there are so many combinations :) Sep 29 19:05:56 hopefully we can specify modules to load in images Sep 29 19:06:14 and end up with special images for each combination on gumstix boards maybe .. Sep 29 19:06:16 with that in mind, I say a vanilla machine config that users can tweak to their liking Sep 29 19:06:27 not sure how it is handled in the buildroot images atm Sep 29 19:06:31 yeah Sep 29 19:06:38 with collections it should be easy? Sep 29 19:06:45 I hope so Sep 29 19:07:13 Crofton: Craig autoloads all of the driver modules Sep 29 19:07:32 and just deals with the failures? Sep 29 19:07:35 you just get an error message in the boot log if the hw isn't there Sep 29 19:07:38 ok Sep 29 19:07:38 yep Sep 29 19:07:47 so lets do that for now Sep 29 19:08:23 thats what is happening to me now with angstrom - it autoloads the smc driver, but I don't use it, so I just see the error at boot Sep 29 19:08:37 we're used to that with gumstix :) Sep 29 19:09:20 or you ask ?s about how to configure module autoloading to get rid of annoying error messages :) Sep 29 19:11:48 koen: whoops :) Sep 29 19:17:02 scruggs: got the files, doing a clean build so it will be a while :-) Sep 29 19:19:03 sakoman, probably plans to sit by the pool during the build .... Sep 29 19:24:38 Crofton: yup, roasting coffee outside by the pool Sep 29 19:25:08 well, I am in my office working on a report Sep 29 19:25:20 gm Sep 29 19:25:27 gm Sep 29 19:25:37 Philippe: http://linuxdevices.com/news/NS4642269161.html <- gpe minibrowser Sep 29 19:25:43 hey likewise Sep 29 19:26:25 hey koen Sep 29 19:26:28 hey all Sep 29 19:27:52 anyone heard from stelios? Sep 29 19:29:09 not me, but is he in the US? Sep 29 19:29:19 I think so Sep 29 19:29:19 no, only two days ago or so, here? Sep 29 19:29:37 He has been traveling, I'm curious how his talk went Sep 29 19:29:52 me too Sep 29 19:29:57 and the feedback he got Sep 29 19:30:03 yeah Sep 29 19:31:44 http://www.linuxdevices.com/articles/AT9272065302.html Sep 29 19:33:11 Crofton: that dude managers to misspell at least two names Sep 29 19:33:35 probably all Greek to him :) Sep 29 19:34:09 Crofton: I suspect the procps failure is due to the cross linkage stuff Sep 29 19:34:18 luckily I know PATH_MAX is 4096 :) Sep 29 19:34:19 I'm guessing that Sep 29 19:34:55 iirc there now is a patch_max() macro Sep 29 19:35:46 well, it built a few days ago Sep 29 19:37:31 03pH5 07org.oe.dev * rf6f643fc... 10/ (3 files in 3 dirs): xserver-kdrive-common: fix tabs/indentation of Xserver script Sep 29 19:37:45 03pH5 07org.oe.dev * r13230c32... 10/ (3 files in 3 dirs): xserver-kdrive-common: only consider Xw100/Ximageon if the device has such a chip Sep 29 19:37:52 koen: could you have a look if I did this right? Sep 29 19:37:56 I'm curious how procps got broken Sep 29 19:37:58 sure Sep 29 19:38:50 grrr, I have an outline of stuff to talk about, I I can't quite find anything to cut and paste in ..... Sep 29 19:40:00 pH5: a lot better, thanks Sep 29 19:40:55 hello Sep 29 19:41:48 need some advise: I'd like to develop a device using the LPC2103 processor Sep 29 19:42:14 The prototype board from Keil is USD150 Sep 29 19:42:31 However, their software is fantastically expensive Sep 29 19:42:52 Should I be looking at gcc and cross compiling? Sep 29 19:43:02 probably .. Sep 29 19:43:19 Are there flash tools available? Sep 29 19:43:27 dfu-util? Sep 29 19:43:53 I don't need Linux, just a dedicated aplication on the 2103 Sep 29 19:56:50 DukeOfURL: wrong channel. hold on. Sep 29 19:57:04 k Sep 29 19:57:31 DukeOfURL: http://www.gnuarm.org/ and I think there are enough LPC fora on the Internet. Sep 29 19:57:39 thx Sep 29 20:02:17 there, webkit builds again Sep 29 20:05:07 is webkit realativly new to OE? Sep 29 20:05:12 http://ftp.gnome.org/pub/GNOME/sources/gnome-keyring/${@gnome_verdir("2.20")}/gnome-keyring-2.20.tar.bz2 fails on me with http://www.pastebin.ca/719939 Sep 29 20:06:05 bitbake 1.8.8 should be ok with current metadata, right? Sep 29 20:11:32 boo tear him away :-) Sep 29 20:21:30 Crofton: after changing the autoloaded modules in the machine conf, what do I need to rebuild to get those to be present in the rootfs? Sep 29 20:28:35 [RFC]: org.openembedded.dev/packages/gnome$ mv gnome-keyring_2.20.bb gnome-keyring_2.20.0.bb Sep 29 20:29:08 hmm Sep 29 20:29:15 maybe kernel Sep 29 20:29:18 I forget Sep 29 20:29:30 clean kernel, rebuild image Sep 29 20:29:51 ok - I'll give that a go - thx Sep 29 20:31:24 http://www.fosdem.org/2008/ Sep 29 20:44:49 03likewise 07org.oe.dev * rc512da70... 10/ (3 files in 2 dirs): gnome-keyring: Change package version from 2.20 to 2.20.0. Fixes bb parsing. Sep 29 20:45:23 hi all Sep 29 20:47:06 03crofton 07org.oe.dev * r912a4116... 10/ (15 files in 4 dirs): wifistix : Add driver for Marvell cf8385 wireless interface driver for gumstix. Sep 29 20:48:20 sakoman, scruggs I have pushed the driver. If you make changes, use mtn diff to create a patch for me Sep 29 20:48:36 you should probably rename what you have before pulling and updating also Sep 29 20:48:52 I only made a minor change, the directory name and to PR Sep 29 20:49:55 great - thx Sep 29 20:50:24 thank you, you have saved me the trouble of working out how to get the driver from buildroot Sep 29 20:51:37 np - actually, there is a prob - kernel is giving me a dump when I run iwconfig Sep 29 20:51:45 http://pastebin.com/m5b868f50 Sep 29 20:52:59 oops Sep 29 20:53:41 hmm Sep 29 20:53:47 Monday I need to look over the kernel patches for changes and to get the wired drivers for the netwifimicroSD stix Sep 29 20:54:23 in case you didn't catch it on the gumstix list, Sep 29 20:54:57 they found a problem with the SMC driver for the newwifimicro when used with DMA, so make sure you don't have it enabled Sep 29 20:55:59 ok Sep 29 20:56:23 I'm on so many lists that sometimes this stuff becomes a blur Sep 29 20:56:36 good to have several sets of eyes Sep 29 21:07:05 http://www.paeps.cx/weblog/fosdem/fosdem_beer_2008.html Sep 29 21:12:50 Crofton: hmm, no other premises for the event. Sep 29 21:13:11 FOSDEM that is, the beer event seems to be relocated Sep 29 21:14:08 I think it is easier to relocate the beer event Sep 29 21:17:52 http://fifisramblings.blogspot.com/ Sep 29 21:32:20 hmm, anyone had such issue or should i report bug? http://pastebin.ca/720020 -- current OE/angstrom 2008.1 Sep 29 21:39:52 koen: hehe, u`ve just confirmed my thoughts about your iq Sep 29 21:53:05 Crofton, sakoman: the driver continues to work after the kernel oops. I've got it pinging out now, but I'll still have a look at where the error is coming from Sep 29 21:54:20 scruggs: wonderful! I'm occupied with other stuff for a while but will have a chance to look at it later this evening Sep 29 21:54:58 no rush, just wanted to report some success Sep 29 22:05:39 does it also occur in the build root? Sep 29 22:05:47 My kconfig is likely different Sep 29 22:06:26 not for myself in buildroot, but there have been other reports of it to the list recently Sep 29 22:06:32 ok Sep 29 22:06:41 success is better than failure :) Sep 29 22:06:51 for sure Sep 29 22:07:18 ok, back to lawn mowing Sep 29 22:07:27 woo hoo! Sep 29 22:36:20 03koen 07org.oe.dev * rab9178ff... 10/ (4 files in 3 dirs): webkit-gtk: catch up with gdk->gtk rename Sep 29 22:38:38 03koen 07org.oe.dev * rab9178ff... 10/ (4 files in 3 dirs): webkit-gtk: catch up with gdk->gtk rename Sep 29 22:41:29 03koen 07org.oe.dev * rab9178ff... 10/ (4 files in 3 dirs): webkit-gtk: catch up with gdk->gtk rename Sep 29 22:41:35 03koen 07org.oe.dev * ra46dc978... 10/ (1 conf/distro/include/sane-srcrevs.inc): sane-srcrevs: bump openmoko-today, openmoko-theme2 and webkit Sep 30 00:06:31 what architecture does The OpenEmbedded Autobuilder run on? Sep 30 00:11:04 Gothic? Sep 30 00:11:14 x86 I'm pretty sure. Sep 30 00:13:04 mwester: hm nevermind then. you probably came across oebug http://bugs.openembedded.org/show_bug.cgi?id=3081 ? Sep 30 00:24:02 Nope. I had a full sucessful set builds -- synced with OE about 20 hours ago. So it must be a recent problem then? Sep 30 00:24:13 s/set/set of/ Sep 30 00:25:55 mwester: yes indeed. Like 8 hours ago Sep 30 00:26:42 Ah, ok. Yep, pretty much a constant state of being with OE -- something always needs to be fixed, it seems. :( Sep 30 00:27:52 is that a serious problem with procps? Sep 30 00:32:48 dcordes: Dunno. Probably just a library incompatabilty resulting in a header file issue. The problem is finding the right way to solve it. Sep 30 01:00:11 hmm, I need to do a buildhost-native build and a cross-compile build within the same recipe (because I need to mix and match files from both builds to end up in the same package). Is there a recommended way to do that? Sep 30 01:05:12 * cbrake_away notes Henryk must be working on mono :-) Sep 30 01:05:24 hehe, yeah Sep 30 01:05:32 Henryk: I've not come up with anything I like yet ... Sep 30 01:06:19 cbrake_away: i've been trying different things all day. I've now come to the final conclusion that it's not possible to build the mcs directory within the cross-compile environment Sep 30 01:07:03 Henryk: nod, just not meant to be ... Sep 30 01:07:24 cbrake_away: on the other hand I don't want to try to use the mcs build from the mono-native build because it's been compiled with a whole different prefix Sep 30 01:08:50 Henryk: good point -- it might not matter, but its not right Sep 30 01:09:47 cbrake_away: but while you're here: I do want to make the individual ipk packages as small as possible, e.g. one assembly per package. Do you have any opinions on the name? Debian uses libmono-system1.0-cil, libmono-system2.0-cil, libmono-winforms1.0-cil, libmono-winforms2.0-cil, libmono-accessibility1.0-cil, etc. pp. Sep 30 01:09:49 Henryk: we could make a mono-mcs package built native, and only package the mcs stuff Sep 30 01:10:06 cbrake_away: right, that would be the third option Sep 30 01:10:58 Henryk: that seems a lot cleaner than pulling stuff out of the native build into the cross build Sep 30 01:11:21 Henryk: I like consistency, and since OE has such strong ties to debian, I think that is a good starting place Sep 30 01:11:39 Henryk: most embedded systems will not need the 1.0 and 2.0 profile, so it makes a lot of sense to separate them Sep 30 01:11:39 cbrake: the problem with the extra mono-mcs recipe is that we really want to package things from mono and mono-mcs into the same ipk. E.g. mono will install the gacutil shell script, but the necessary gacutil dll is in the mcs build Sep 30 01:12:39 Henryk: yeah, that gets messy, but is there any reason they can't be installed by separate ipkg's Sep 30 01:13:00 Henryk: the python packaging is very granular and might be a good reference if you have already looked at it Sep 30 01:13:47 so what I'm currently planning is to following along the lines of http://www.mono-project.com/Mono:ARM and build a native and a cross-compile version, then install the native version into a tmptree, then over-install the cross-compile version into the same tmptree (all platform-specific code should be overwritten by that) and then cherry-pick files from that tmptree into individual ipkgs Sep 30 01:15:55 Henryk: sounds good. So the trick is to figure out how to do that from one recipe. I guess separate ones would not be the works thing, but then you could not build with the rm_work (or whatever that is) flag set. There is probably a way to do it in one recipe, but will take some work. Sep 30 01:16:41 cbrake_away: well, for the two-recipe version I would just tar the tmptree up and store it somewhere (staging maybe?) Sep 30 01:16:52 Henryk: ahh, good idea Sep 30 01:17:38 Henryk: that seems clean **** ENDING LOGGING AT Sun Sep 30 02:59:56 2007