**** BEGIN LOGGING AT Mon Jan 28 02:59:57 2008 Jan 28 03:03:00 03Laibsch 07org.oe.dev * r7c1e0973... 10/ (5 files in 3 dirs): Jan 28 03:03:00 libvncserver: add recipe for version 0.9.1 as proposed in 2411 Jan 28 03:03:00 * this has disabled x11vnc and the SDL viewer. DEFAULT_PREF is -1 Jan 28 03:03:00 * add Patrick to MAINTAINERS file as requested in 2125 Jan 28 03:03:05 03Laibsch 07org.oe.dev * rcec9ba77... 10/ (11 files in 5 dirs): fbvncserver + x11vnc: consolidate all VNC related stuff in packages/vnc/ Jan 28 04:19:44 XorA|gone: d01c50d1b11c46a3d67726552fb7fe4ff8ae6da5 has a side effect that: dbus-lib does not get depended by dbus, and hence libdbus was not installed when do_rootfs Jan 28 07:28:15 morning Jan 28 07:30:11 hrw: morning :) Jan 28 07:32:12 hvontres|home: on #angstrom people are asking sometime about poodle nand images Jan 28 07:32:21 hrw: It looks like the patches for #3715 didn't make it through :( But, it looks like helge has a better set of patches Jan 28 07:32:40 hrw: I saw that. I think CoreDump used to have some Jan 28 07:32:54 hrw: I only have a 250 available here Jan 28 07:33:05 I starts to hate OE monotone servers Jan 28 07:33:22 hrw: hehe.. Jan 28 07:34:12 hrw: Don't worry about it yet. I just told Helge to add his patches to #3715. RP should have a copy of them as well Jan 28 07:35:46 repushing now again Jan 28 07:37:00 hrw: How's the baby doing? Jan 28 07:38:12 not yet born Jan 28 07:38:13 hi steliosk Jan 28 07:38:29 morning all ! Jan 28 07:38:30 hvontres|home: Ania feels good Jan 28 07:38:38 hrw : hi ! Jan 28 07:38:43 hvontres|home: now it should be pushed for real Jan 28 07:38:49 ~curse monotone.openembedded.net Jan 28 07:38:50 May you be reincarnated as a Windows XP administrator, monotone.openembedded.net ! Jan 28 07:39:16 hrw: try monontone.openembedded.org? Jan 28 07:39:38 switched to it Jan 28 07:44:55 * XorA wishes people wouldnt leave completely untrue bugreports as onelines on IRC when you are away Jan 28 07:46:03 * hrw -> shop Jan 28 07:46:58 FYI : My OE presentation at last years Power Dev conference -> http://www.power.org/devcon/07/Session_Downloads/PADC07_Koroneos_OE-presentation-final.pdf Jan 28 07:47:09 http://drlauer-research.com/ ;) Jan 28 07:50:36 mornin Jan 28 08:03:19 03hvontres 07org.oe.dev * r36133f7d... 10/ (5 files in 3 dirs): linux-rp: misc fixes to make 2.6.24 compile for poodle - close #3715 Jan 28 08:27:09 RP: thanks, I'll try that out later today Jan 28 08:33:37 <_law_> hi i think there is some git problem in angstrom.stable Jan 28 08:34:04 <_law_> git.inc requires git-gui-install-mode-arg-spaces.patch Jan 28 09:07:12 re Jan 28 09:07:27 hrw: er Jan 28 10:02:25 hrw: BTW successfully compile/ran epdfview in OE on n810 Jan 28 10:03:02 hrw: unfortuneately I think we need to put in the maemo version of gtk+ because it doesnt quite work like it should Jan 28 10:08:03 XorA: now they use upstream + one big patch way so shorter recipes Jan 28 10:08:27 hrw: cool I shall look at this as some widgets seem to be broken in epdfview Jan 28 10:08:48 hrw: but thanks to rschuster the basics work, command line stuff should just work (tm) Jan 28 10:09:05 XorA: poky has os2008 kernel Jan 28 10:09:10 hrw: I know Jan 28 10:09:16 did not yet tested does it build Jan 28 10:09:24 I need to call expansys to push my order Jan 28 10:09:37 hrw: kernel seems to work fine so far, I just want to add to maemo for now Jan 28 10:10:44 I will probably add CFS backport to it. It apply fine Jan 28 10:14:37 hrw: we need to get their browser in OE Jan 28 10:14:46 hrw: then we can all just use poky on n800 Jan 28 10:15:16 * zecke continues to search for a new job Jan 28 10:15:18 hrw: does poky have cool network config tools yet? Jan 28 10:15:32 zecke: can you say "Want Fries" :-) Jan 28 10:16:33 XorA: well, I consider being a bump again and selling Street newspapers Jan 28 10:16:56 zecke: :-D Jan 28 10:17:17 zecke: well you can apply for nokia now they gonna own qt :-D Jan 28 10:17:18 XorA: networkmanager is in. not in default images Jan 28 10:17:31 hrw: including bluetooth like maemo? Jan 28 10:17:45 no idea - did not tries it Jan 28 10:17:57 XorA: don't get me started... Jan 28 10:18:15 hrw: BTW dont try navigating with n810 GPS, unless you want to end up over clifs Jan 28 10:18:54 XorA: OH guys already told me that internal gps suxx totally Jan 28 10:19:18 hrw: 50m consitently wrong here Jan 28 11:14:48 nokia owns qt? Jan 28 11:15:05 Crofton: not only that Jan 28 11:15:05 will Jan 28 11:15:22 is this news? Jan 28 11:15:28 hrw: they already have 66(,6666)% of the stocks :} Jan 28 11:15:33 or are you guys spreading rumours ? Jan 28 11:15:41 Crofton: it is on the trolltech.com website Jan 28 11:15:44 heh Jan 28 11:15:58 wow not even on /. yet :) Jan 28 11:15:58 Crofton: it might be hacked though Jan 28 11:16:11 http://trolltech.com/company/newsroom/announcements/press.2008-01-28.4605718236 Jan 28 11:19:36 nokia is buying everything Jan 28 11:23:04 * * OE Bug 1732 has been RESOLVED (FIXED) by Jan 28 11:23:06 * * fbvncserver and libvncserver-0.7.1 Jan 28 11:23:08 * * http://bugs.openembedded.org/show_bug.cgi?id=1732 Jan 28 11:28:41 Should a CVS bb file lock down a specific check out or date? Jan 28 11:29:30 Why is versioning policy " cvs packages should use +cvs, e.g. foo_1.3+cvs20051106" otherwise? Jan 28 11:37:04 Hi all, I have a consumer device (Zipit Z2) that comes with a 2.6.21 kernel installed into a PXA-270. On this device, I cannot see to kernel boot messages so I am blindly working at getting angstrom to run. Jan 28 11:37:20 I can build a filesystem that boots, but not angstrom. Jan 28 11:37:31 ~seen pgf Jan 28 11:37:35 Laibsch: i haven't seen 'pgf' Jan 28 11:37:37 what does boot? Jan 28 11:37:53 The angstrom filesystem plays dead, I suspect that it is the oabi vs eabi stuff Jan 28 11:38:44 If I set ARM_ABI = "oabi" in my machine spec, is that all that is needed to build an oabi system against angstrom-2008.1 ? Jan 28 11:39:55 Crofton: as far as I can tell, nothing works. I've tried echo-ing to /dev/tty0 & /dev/tty1. That will work on the old filesystem, I can spit text onto the screen via /etc/init.d/rcS Jan 28 11:39:57 why not just build eabi? Jan 28 11:40:17 well, I think it has to do with the kernel that is pre-installed Jan 28 11:40:29 ah, you are using pre-installed kernel? Jan 28 11:40:35 yes Jan 28 11:40:42 we can't reaplce it easily? Jan 28 11:41:27 I'm not sure the best way to switch to oabi in that case Jan 28 11:41:28 not sure, I've considered that. the cmdline of the existing kernel says "console=/dev/ttyS0 115200" so the kernel boot messages are going to a non-existant device. Jan 28 11:42:09 the var you mentioned may do the trick Jan 28 11:42:17 a cat of /proc/mtd does show three Flash partitions, so, I may be able to overwrite the existing Flash? Jan 28 11:43:28 I was thinking of buildiing the mtd modules, then trying to replace the existing kernel a try. but, I'd not done this before and not sure how to go about it. Jan 28 11:43:30 we should be able to build an oabi image Jan 28 11:43:55 Crofton: how can I tell which is which? Jan 28 11:44:09 not sure Jan 28 11:44:15 file maybe Jan 28 11:44:20 I've looked in /sys and /proc but no clues as to the kernel. Jan 28 11:44:37 as in eabi vs oabi? Jan 28 11:44:39 AFAICT, I've been building oabi in the old distro + machine files Jan 28 11:44:44 yes Jan 28 11:45:03 can you see kernel config? Jan 28 11:45:46 I don't know if they put it in there, let me see if I can reach into the filesystem can copy out /boot/config Jan 28 11:46:34 set TARGET_OS = linux-gnueabi for eabi Jan 28 11:46:42 I think the alternative is just linux Jan 28 11:46:59 Crofton: I've pulled some things out of their filesystem, it is very tiny though. Jan 28 11:47:12 look in packages/linux/linux.inc for the config options that get played with Jan 28 11:47:48 I don't use the packages/linux stuff, I have my own bb file dir Jan 28 11:48:15 BBFILE_COLLECTIONS = "upstream local" Jan 28 11:48:23 that file shows the config options Jan 28 11:48:26 yeah Jan 28 11:48:59 * T0mW goes to dump filesystem. Jan 28 11:52:52 Crofton: you ever work with mtd ? Jan 28 11:53:32 Crofton: I still think that I should take the good filesystem kernel, set console=/dev/tty and replace it in Flash via mtd Jan 28 11:53:35 no Jan 28 11:53:47 Crofton: maybe prpplague has Jan 28 11:55:06 yeah Jan 28 11:57:55 OK, I n honor of completing office constructions I have added an OE todo note on my copmuter Jan 28 11:59:36 Crofton: btw how do you want this poodle packed? Jan 28 11:59:53 hmmm Jan 28 12:00:05 hrw: I've sent an email to the poodle guy Jan 28 12:00:12 Crofton: its in a box, but I could just tape it in some bubble wrap Jan 28 12:00:14 in such a way as I can claim I use it if they rumage through my bags :) Jan 28 12:00:24 how big is the box? Jan 28 12:01:06 6 inches in all directions Jan 28 12:01:21 Crofton: nots it original box Jan 28 12:01:41 hrw: I'm not very comfortable with hosting NAND backups. Someone might decide he has a copyright on some obscure part of the image and sue... Jan 28 12:01:45 lets use bubble wrap Jan 28 12:01:54 Crofton: ok Jan 28 12:01:58 CoreDump: drop them then Jan 28 12:02:10 * hrw drops zaurus support soon Jan 28 12:02:12 I'll just pretend I am paraoid Jan 28 12:02:22 not that they will bother me coming back in to us Jan 28 12:02:23 re Jan 28 12:02:25 I will be back in 20-30 minutes Jan 28 12:02:43 anyone used a crosstool tool chain with OE? Jan 28 12:02:57 this is not something I would normally do .... Jan 28 12:03:03 Crofton: you mean "just some" external one? Jan 28 12:03:08 oh, lol Jan 28 12:03:10 one that was not produced with OE? Jan 28 12:03:16 Crofton: so you dont want the Terrorists Handbook on poodle then Jan 28 12:03:17 yeah Jan 28 12:03:23 Crofton: I did a couple of times Jan 28 12:03:26 no, that what not be good Jan 28 12:03:28 ok Jan 28 12:03:41 anyway, I already have a copy of the anarchist cookbook Jan 28 12:03:47 of course it sucks to do so but sometimes there's no choice, needs some environment settings but works Jan 28 12:03:47 this retarded **** still has me under moderation on angstrom-devel. Teaches me right to try help someone I guess. Jan 28 12:03:59 Crofton: no filesystem as we know it. They have one 3Meg app, Zipit2, while appears to be their /sbin/init replacement Jan 28 12:04:14 heh Jan 28 12:04:16 s/while/which Jan 28 12:04:38 yeah, so, it looks like the console= thing so I can get kernel messages Jan 28 12:05:59 ping XorA Jan 28 12:06:21 xjqian: I dont beleive you Jan 28 12:06:50 XorA: the libdbus? Jan 28 12:07:02 xjqian: yes Jan 28 12:07:09 Package: dbus Jan 28 12:07:09 Version: 1.0.2-r9 Jan 28 12:07:09 Architecture: armel Jan 28 12:07:11 Maintainer: OpenEmbedded Team Jan 28 12:07:13 Depends: update-rc.d, libdbus-1-3 (>= 1.0.2), libc6 (>= 2.5), libexpat1 (>= 2.0.0) Jan 28 12:07:50 XorA: let me double-check my build env Jan 28 12:10:54 for some reason libdbus was not packed into the rootfs for me, weird Jan 28 12:14:08 must be time for lunch Jan 28 12:15:21 philippe: ping? Jan 28 12:24:44 if run.do_compile defines CC as: "CC="arm-angstrom-linux-gnueabi-gcc -march=armv5te -mtune=xscale", then it is compiling my app as eabi and not oabi? Jan 28 12:26:55 mickeyl: hi Jan 28 12:27:03 morning hrw Jan 28 12:27:03 re Jan 28 12:27:17 mickeyl: nice site you have as drlauer-research :) Jan 28 12:27:24 morning Jan 28 12:27:44 hrw: where? Jan 28 12:27:51 dr. mickey sounds way cooler =) Jan 28 12:28:29 hrw: heh, how did you stumble over it? Jan 28 12:28:46 just planning for the future Jan 28 12:28:49 :) Jan 28 12:32:17 zecke: drlauer-research.com Jan 28 12:32:51 mickeyl: whois on one of your domains Jan 28 12:33:30 aah Jan 28 12:33:32 the MX record Jan 28 12:34:11 what happened to fosdem2008 in the wiki? Jan 28 12:34:17 under events? Jan 28 12:36:09 hmm Jan 28 12:36:12 was it there once? Jan 28 12:36:17 i didn't add it yet Jan 28 12:36:40 * mickeyl adds it Jan 28 12:37:49 done Jan 28 12:38:12 mickeyl: you are forced to share a room with me (again) :) Jan 28 12:38:35 zecke: looks like it, yeah Jan 28 12:42:21 Crofton: http://www.openembedded.org/fosdem2008 Jan 28 12:43:47 how many of you will be attending? Jan 28 12:43:55 I'm going to pay a visit :) Jan 28 12:44:23 6 or 7 at least Jan 28 12:44:47 good, good :) Jan 28 12:47:19 it was there once upon a time Jan 28 12:51:46 pb_: hey Jan 28 12:56:58 <_law_> hi is bugzilla down? Jan 28 12:57:21 <_law_> ok wrong url Jan 28 12:58:34 works from here Jan 28 12:58:39 bugs.openembedded.net Jan 28 13:02:05 * * OE Bug 3716 has been created by law(AT)a1.net Jan 28 13:02:06 * * git-native 1.5.3-r1 doesnt build Jan 28 13:02:08 * * http://bugs.openembedded.net/show_bug.cgi?id=3716 Jan 28 13:04:19 hi zecke Jan 28 13:17:11 I do learn a lot from oe but currently I am really sucked on the glibc (I guess this is a very had part) Jan 28 13:18:04 http://paste-it.net/6049 this is where I am at now ( and I tried hacking arond for the last few days / changing makefiles etc Jan 28 13:20:54 keesj: did you google for that? Jan 28 13:21:48 keesj: is that build working for others? Jan 28 13:23:42 it contains arm related patches, it works for the mamona arm target Jan 28 13:24:37 somehow I have the feeling I am the first really trying to compile things for this arch Jan 28 13:25:23 for exampe I had to change classes/kernel-arch.bbclass so that x86_64 is seend a valid arch Jan 28 13:26:23 keesj: do you run on x86_64 or do you compile for it? Jan 28 13:26:43 x86_64,i386 x86 => x86_64 i386 x86 Jan 28 13:26:55 zecke: booth Jan 28 13:27:08 keesj: and yes, binutils, gcc, glibc have a love and hate relationship with each other and sometimes there are cross compile bugs, this means your host architecture is taken into account and not your target Jan 28 13:27:41 keesj: you are probably really the first, googling and changing binutils version until it is working is one option Jan 28 13:28:44 perhaps I can try a vannila glibc or a more recent one? Jan 28 13:31:05 lego fans, check out google Jan 28 13:31:41 the thing is that I only want to be able to play with it , perhaps I must "just" try to compile to and other arch like 586 Jan 28 13:32:16 Crofton: nice , usually I can click on the image to know why it is like that Jan 28 13:32:26 keesj: 50 years of lego bricks Jan 28 13:33:06 how cool is that :P Jan 28 13:33:27 maybe those data center crates of googles are really filled with lego :-) Jan 28 13:34:02 I did a preseantion last week about testing and I used lego to attach my devices Jan 28 13:38:02 XorA: you will attend fosdem? Jan 28 13:38:11 hrw: yes Jan 28 13:38:45 XorA: then respond to tosa thread on OE (reply to my just-sent mail) Jan 28 13:38:46 ;D Jan 28 13:39:14 hrw: I shall wait 4 hours for oe.org to send messages Jan 28 13:39:52 yep - thats ugly part of ltg Jan 28 13:41:04 yo lumag Jan 28 13:41:13 hi! Jan 28 13:41:15 hi XorA Jan 28 13:41:40 lumag: you are Dmitry Baryshkov ?? Jan 28 13:41:44 hi Дми́трий Jan 28 13:42:21 XorA, yes Jan 28 13:42:28 hrw: hi :))) Jan 28 13:42:40 lumag: we were just discusssing Tosa machines and FOSDEM :-) Jan 28 13:42:50 oh. Jan 28 13:42:55 XorA: and Dmitry Jan 28 13:42:55 !logs Jan 28 13:42:57 Channel logs for #oe are archived at: Jan 28 13:42:58 http://hentges.net/tmp/logs/irc/%23oe Jan 28 13:42:59 Live-logs are available at Jan 28 13:43:01 http://hentges.net/tmp/logs/irc/livelogs/%23oe.livelog Jan 28 13:43:02 See ?? help-logs for usage instructions Jan 28 13:43:13 lumag: basically - XorA will be at fosdem Jan 28 13:43:27 I will not. He also has OE tosa Jan 28 13:45:21 ok. Jan 28 13:45:29 lumag: will you be at FOSDEM Jan 28 13:45:32 ? Jan 28 13:46:19 Unfortunately, no. A week ago my wife gave bith to our child, so I'm currently a bit busy here :) Jan 28 13:47:05 lumag: ah, forking like hrw Jan 28 13:47:06 Hopefully one of my colleagues will visit FOSDEM. Jan 28 13:47:17 :) Jan 28 13:47:26 lumag: well if he does, give me a should before hand and let me know his name to find him Jan 28 13:47:34 s/should/shout/ Jan 28 13:47:49 of course :) Jan 28 13:48:05 and we can make sure an extra tosa gets to you Jan 28 13:48:21 got to get that child started early :-) Jan 28 13:53:48 good morning Jan 28 13:54:09 gm Jan 28 14:00:18 lumag: did the jacket arrive yet? Jan 28 14:00:35 mickeyl: not yet. Jan 28 14:00:46 hmm bummer. i don't know how long a parcel to .ru takes usually Jan 28 14:01:00 mickeyl, do you have any tracking number? Jan 28 14:01:15 no, it has been sent by airmail, not by courier Jan 28 14:01:15 it can be 2 or 3 weeks. Jan 28 14:14:11 03koen 07org.oe.dev * r26852238... 10/ (1 packages/xorg-xserver/xorg-xserver-common.inc): xorg-xserver-common.inc: fix typo Jan 28 14:14:19 03Laibsch 07org.oe.dev * rcc839991... 10/ (23 files in 21 dirs): Jan 28 14:14:19 divers: delete 22 empty directories Jan 28 14:14:19 * skipping packages/netbase/ since those seem to be used albeit incorrectly Jan 28 14:14:27 03Laibsch 07org.oe.dev * rf26d6980... 10/ (1 packages/vnc/fbvncserver_0.9.4.bb): Jan 28 14:14:27 fbvncserver: remove dependency on version 0.6 of libvncserver. 0.8.x and 0.9 are fine. Jan 28 14:14:27 * unbreaks build when RDEPENDS line is commented out (the kmodule still fails to build) Jan 28 14:14:27 * reorder fields according to StyleGuide Jan 28 14:14:28 * closes 1732 Jan 28 14:14:35 03koen 07org.oe.dev * r0ff7b192... 10/ (87 files in 9 dirs): linux-ezx: add 2.6.24 Jan 28 14:14:48 03Laibsch 07org.oe.dev * r847b7160... 10/ (1 packages/mysql/mysql_4.1.18.bb): mysql: make sure all files are packaged. Add -dbg package while we are at it. Jan 28 14:14:57 03Laibsch 07org.oe.dev * re5a4ea76... 10/ (10 files in 3 dirs): Jan 28 14:14:57 conf/machine: Go back to treating all Strongarm devices the same. Jan 28 14:14:57 * http://lists.linuxtogo.org/pipermail/openembedded-devel/2008-January/004031.html Jan 28 14:14:57 gcc does not differentiate between the different Strongarm flavors, why should OE? Jan 28 14:15:06 03mickeyl 07org.oe.dev * r2576615c... 10/ (5 files in 2 dirs): dbus 1.1.1 update to 1.1.4 (aka 1.2.0 RC2) Jan 28 14:15:14 03koen 07org.oe.angstrom-2007.12-stable * r7f38dc36... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom 2007: bump to r3 Jan 28 14:15:21 03Bernhard.Guillon 07org.oe.dev * rf8452fb3... 10/ (4 files in 3 dirs): (log message trimmed) Jan 28 14:15:21 Revision: 40ab8178005b44226c94200bdcb19bca0262e49c Jan 28 14:15:21 Ancestor: 0f7dd726f77b3066a94b7670832cc04148fcd721 Jan 28 14:15:21 Author: Bernhard.Guillon@opensimpad.org Jan 28 14:15:21 Date: 2008-01-13T12:16:57 Jan 28 14:15:23 Branch: org.openembedded.dev Jan 28 14:15:25 Modified files: Jan 28 14:15:55 /, finally found out about the TT deal Jan 28 14:16:05 Crofton: hehe Jan 28 14:16:32 philippe: are other osso guys online? Jan 28 14:31:51 morning Jan 28 14:41:20 mickeyl: you can ditch that DEBIAN_NAMES python in dbus_1.1.4.bb Jan 28 14:45:32 yep Jan 28 14:45:36 some further updates coming along Jan 28 14:45:43 *cleaning up that directory* Jan 28 14:45:50 mickeyl: just checking you realised that Jan 28 14:45:55 sure Jan 28 14:46:01 thanks Jan 28 14:46:43 * XorA had to fixup rschusters patch to get stuff compiling for n810 :-) Jan 28 14:47:16 can we compile compatible packages with OE? Jan 28 14:47:35 or did you yank the complete distro? Jan 28 14:48:09 mickeyl: we can compile compatible packages Jan 28 14:48:20 excellent Jan 28 14:48:22 mickeyl: rschuster did a distro.conf for it Jan 28 14:48:25 aah Jan 28 14:48:31 * mickeyl checks Jan 28 14:48:36 XorA: Which toolchain is it using? Jan 28 14:48:41 mickeyl: its in his svn Jan 28 14:48:43 XorA: The csl one or not? Jan 28 14:48:48 RP: csl Jan 28 14:48:59 XorA: oh. why is it not in OE? Jan 28 14:49:00 rschuster: ? Jan 28 14:49:14 mickeyl: the guy that bought dinner for the whole of OEDEM Jan 28 14:49:15 I have some fixes for the csl toolchain and sysroot Jan 28 14:49:31 XorA: i know. note the ':' :) Jan 28 14:49:57 RP: coolies. in with it :) Jan 28 14:50:05 http://evolvis.org/plugins/scmsvn/viewcvs.php/*checkout*/trunk/oe-overlay/conf/distro/chinook-compat.conf?rev=340&root=jalimo Jan 28 14:50:21 I made one change locally to take into account that I repackaged dbus Jan 28 14:50:21 pretty cool Jan 28 14:50:35 it should go in Jan 28 14:50:47 mickeyl: We're still waiting on me to merge the sysroot support. I keep trying and then the size size of the job overwhelms me ;-) Jan 28 14:50:56 * XorA didnt want to push it without rschuster go ahead Jan 28 14:51:01 mickeyl: Hopefully there will be some movement on that soon... Jan 28 14:51:15 ~lart MSI... Jan 28 14:51:16 * ibot executes killall -HUP MSI... Jan 28 14:51:25 RP: yeah, that would be much appreciated Jan 28 14:52:11 mickeyl: I think we need to get the maeom patched gtk+ into OE though, stuff doesnt quite work right Jan 28 14:52:14 Is OE.dev interested in the ton of -sdk packages I added to poky btw? Jan 28 14:52:37 RP: which ones, e.g.? Jan 28 14:52:56 mickeyl: qemu-sdk, libx11-sdk and the X stack Jan 28 14:53:02 sure thing Jan 28 14:53:04 XorA: if you think chinook-compat.conf is ok for OE put it in Jan 28 14:53:29 rschuster: its probably not perfect, but there is such an interest in it, people will quickly improve it Jan 28 14:53:53 I would love to see a maemo gtk in OE... Jan 28 14:54:10 ya, why not Jan 28 14:54:48 RP mickeyl: but for a long time this distro conf will only allow building compatible packages. not a whole distro Jan 28 14:54:49 chinook-compat.conf is on its way to OE Jan 28 14:54:56 XorA: thanks! Jan 28 14:55:13 rschuster: no thank you for doing 99.9999999% of the work Jan 28 14:55:16 hi mickeyl Jan 28 14:55:27 rschuster: that's fine. _some_ people are not interested in replacing everything on a device ;) Jan 28 14:55:31 hail pb_ Jan 28 14:55:33 rschuster: poky can generate sato images for the nokia so we have lots of the pieces in place though Jan 28 14:55:49 with working wifi now... Jan 28 14:55:56 RP: awesome Jan 28 14:56:10 RP: kernel included? Jan 28 14:56:43 rschuster: It can rebuild the kernel in a way the wifi module can be reused and since its on the initrd, you can still reflash the rootfs Jan 28 14:56:47 grr, gcc-cross-kernel 3.3.4 doesn't seem to stage anymore. I wonder how that got broken. Jan 28 14:56:50 rschuster: Poky has 2.6.21 kernel from os2008 Jan 28 14:57:03 ah ok. Jan 28 14:57:17 I should have n810 iin this week Jan 28 14:58:00 and then: good bye zaurus machines (except c760 used to test poky/armv4-5) Jan 28 14:58:15 * XorA watches the zaurus die Jan 28 14:58:44 * RP is still quite fond of them Jan 28 14:59:19 RP: spitz is nice as developer tool indeed Jan 28 14:59:25 native builds etc Jan 28 15:01:21 at least spitz has more accurate GPS than n810 ;-) Jan 28 15:01:23 ~spell consitently Jan 28 15:01:25 possible spellings for consitently: consistently constantly contently confidently Jan 28 15:01:49 consistently Jan 28 15:03:30 * ant wonders why linux-rp-2.6.24 is not yet the preferred kernel for Zaurus... Jan 28 15:03:42 03mickeyl 07org.oe.dev * r4f688461... 10/ (33 files in 6 dirs): clean up dbus directory, remove unused patches, make better use of .inc file Jan 28 15:03:48 03mickeyl 07org.oe.dev * r88b6d0d5... 10/ (1 conf/distro/maemo-1.0.conf): remove maemo-1.0.conf which is clearly outdated ("include familiar"...) Jan 28 15:03:53 03xora 07org.oe.dev * r416609d9... 10/ (1 conf/distro/chinook-compat.conf): Jan 28 15:03:53 chinook-compat.conf : a distro to make packages for n8x0 machines without Jan 28 15:03:53 reflashing them. From Robert Schuster. Jan 28 15:04:05 btw - maemo is still compiled for armv5te? Jan 28 15:05:55 three commits at the same time, the automerge will have fun :) Jan 28 15:06:07 (I pushed qemu changes that second too) Jan 28 15:06:10 heh Jan 28 15:06:24 monotone stability seems rock solid these days Jan 28 15:06:29 never had any problem with 0.38 Jan 28 15:06:32 you? Jan 28 15:06:38 03rpurdie 07org.oe.dev * r426e966e... 10/ (4 files in 2 dirs): Add qemu 0.9.1 from poky and make it the deafult since it fixed a ton of bugs Jan 28 15:06:40 I had Jan 28 15:06:46 03rpurdie 07org.oe.dev * rdc89094b... 10/ (68 files in 5 dirs): qemu: Remove versions with negative default preferences and cleanup Jan 28 15:06:51 but rather due to your server mickeyl... Jan 28 15:07:09 brb Jan 28 15:07:16 mickeyl: oh your server is already set up Jan 28 15:09:17 mickeyl: Just that its slow ;-) Jan 28 15:09:40 zecke: no, not really Jan 28 15:09:57 the server is ordered and will be shipped mid-februaray Jan 28 15:10:06 monotone.oe.net is pointing to my private machine atm. Jan 28 15:10:18 which runs an mtn server but is not autosynced Jan 28 15:10:23 so, just a fallback Jan 28 15:10:57 RP: ya, speed could be better... Jan 28 15:11:57 * mickeyl hunts food Jan 28 15:17:48 mickey|lunch: and do 'is not autosynced' I had to repush few times.... Jan 28 15:20:17 hello, after building opie-image i have things like this: ls: ./Settings: Input/output error Jan 28 15:22:17 GNUtoo: is your flash full? Jan 28 15:22:27 zecke, qemu... Jan 28 15:22:52 is your disk broken? Jan 28 15:23:05 GNUtoo: e.g is this on your machine or the machine you installed Opie on? Jan 28 15:23:06 pb_, i see no bad dmesg messages Jan 28 15:23:17 GNUtoo: where is the corelation between opie-image and ls: ./Settings? Jan 28 15:23:29 zecke, MACHINE = "qemuarm" Jan 28 15:23:37 that appears on qemu Jan 28 15:23:47 GNUtoo: what did you do after bitbake opie-image? flashed qemu? Jan 28 15:24:23 zecke, nothing just that: qemu-system-arm -hda Angstrom-opie-image-glibc-ipk-2008.1-test-20080128-qemuarm.rootfs.ext3 -kernel zImage-qemuarm.bin -append root=/dev/sda -M versatilepb -serial /dev/stdout -m 256 && reset Jan 28 15:25:20 GNUtoo: where is 'ls' then coming from? Jan 28 15:25:33 zecke, from inside the serial console of qemu Jan 28 15:26:42 GNUtoo: now try to put everything in a sentence that makes sense :) Jan 28 15:26:54 LOL Jan 28 15:27:34 GNUtoo: So you build an image, execute it using qemu, login, and then type ls and you get an error? Jan 28 15:28:06 yes Jan 28 15:28:35 GNUtoo: how big is the image, how big is the ext3 file, and what does the kernel think about the size? Jan 28 15:28:42 GNUtoo: did you run ext2fsck on it? Jan 28 15:29:13 zecke, ext3fsck yes and it was bad...but now i've rebuilt the opie image and it works fine...strange Jan 28 15:29:54 but qemu doesn't have a touchscreen and i'm stuck at "tap anywhere to on the screen to continue" in opie Jan 28 15:30:14 GNUtoo: what do you try to say? now things work? Jan 28 15:30:20 just hit the screen harder Jan 28 15:30:27 GNUtoo: sorry, I sadly don't understand Jan 28 15:30:30 zecke, yes after a rebuild Jan 28 15:30:44 GNUtoo: you can try the neo1973 version of qemu, it has emulation for more hardware Jan 28 15:31:03 BTW: anyone got the mail address from quim gil? Jan 28 15:31:06 zecke, ok thanks good idea, but there is no way to bypass this screen? Jan 28 15:31:21 (i already have it...must change gcc for building it) Jan 28 15:32:01 (because gcc-3.4.6-r2 is broken on gentoo) Jan 28 15:32:13 GNUtoo: if you don't have a touchscreen, how do you want to start and use any of the applications? Jan 28 15:32:58 zecke, isn't it like qtopia...with a variable we can switch between mouse and touchscreen Jan 28 15:33:02 GNUtoo: IIRC qemu can emulate a touchscreen using a wacom tablet Jan 28 15:33:11 hvontres|work, ok thanks Jan 28 15:34:05 GNUtoo: well, then your question is not how you can bypass the touchscreen calibration screen but how to use another mouse Jan 28 15:34:07 * hvontres|work wishes he had one Jan 28 15:34:13 GNUtoo: and how to configure QtE to show a cursor... Jan 28 15:34:36 GNUtoo: and seriously I have no interest in QtE 2 and Opie Jan 28 15:34:48 zecke, my question are 1) how to bypass the screen 2)how to use a mouse Jan 28 15:35:13 zecke, so what environnement do you use? none? another? Jan 28 15:35:33 GNUtoo: if you do 2nd first 1st will go away otherwise touch /etc/pointercal but these questions are off-topic here Jan 28 15:35:43 GNUtoo: I work on my own Epicenter Jan 28 15:36:13 that's a nice one... debian etch can't build 2.6.24 kernels... Jan 28 15:36:53 needs to get a newer binutils..... Jan 28 15:38:06 ok thanks Jan 28 15:38:28 GNUtoo: You should use qemu from poky Jan 28 15:38:34 Anything else never worked for me Jan 28 15:38:38 Laibsch: OE has Poky qemu now Jan 28 15:38:41 They also have packages for download Jan 28 15:38:57 hrw: Yes, if you want to build your own, fix PATH, etc. Jan 28 15:39:08 I did that when the packages had not been updated Jan 28 15:39:11 ? Jan 28 15:39:17 ok thanks Jan 28 15:39:24 but just installing the packages is easiest, ATM, I think Jan 28 15:39:30 hrw: what is unclear Jan 28 15:39:32 ? Jan 28 15:39:43 I believe GNUtoo wants to run an image. Jan 28 15:39:59 yes Jan 28 15:40:04 So, the best for him IMHO is to install the packages, including poky-scripts Jan 28 15:40:33 He could of course "bitbake qemu-native" and use the result, but that is more tedious Jan 28 15:42:52 Laibsch: For reference Poky's toolchain tarballs have a standalone qemu binary now Jan 28 15:43:14 * Laibsch prefers packaged stuff Jan 28 15:43:15 ;-) Jan 28 15:43:53 Laibsch: Thats fine if there is a package available ;-) Jan 28 15:44:13 The tarball only touches /usr/local/poky so its not so bad... Jan 28 15:44:16 Shouldn't we all be running ubuntu or debian ;-) Jan 28 15:44:59 * RP looks at his slackware install Jan 28 15:45:29 Laibsch: Seriously, that isn't an option for everyone and this gives us a way to solve a multitude of problems... Jan 28 15:45:37 sure Jan 28 15:45:40 I understand Jan 28 15:46:34 Laibsch: Even just maintaining packages to the various debians is in reality a nightmare for something as fast moving as qemu :( Jan 28 15:46:58 Is it? Jan 28 15:47:18 I am only just starting, but there are very helpful tools out there Jan 28 15:47:39 I assume you are aware of them? Jan 28 15:47:54 Laibsch: 4-6 versions of debian/ubuntu to support. each with other libs... Jan 28 15:48:09 Laibsch: I don't make the debs but the people who do are extrememly capable ;-) Jan 28 15:48:12 You guys know about PPA for ubuntu? Jan 28 15:48:26 It is an autobuilder Jan 28 15:48:54 I just upload the definition and compilation and libs for different distro releases are automatically being taken account of Jan 28 15:49:06 The definition hardly ever changes Jan 28 15:49:15 It is sort of like a bb file that you maintain Jan 28 15:49:29 Laibsch: The problem is this is a hacked qemu so upstream don't want to touch it Jan 28 15:49:32 I think there is no such thing for debian, yet Jan 28 15:49:49 Laibsch: also the definition does change a lot, there is a horrible patch set Jan 28 15:49:57 rp: that should still not interfer with creating the package at all Jan 28 15:50:05 Laibsch: Well, there's backports.org Jan 28 15:50:32 broonie: Is that automatically compiled for anybody uploading there *.dsc? Jan 28 15:50:37 broonie: I can hardly images Jan 28 15:50:41 imagine Jan 28 15:51:03 rp: You should really check out PPA for ubuntu. It is relatively new and makes things much easier Jan 28 15:51:13 Anyone with a key in the keyring, which I think is restricted to developers. Jan 28 15:51:17 I can upload your qemu stuff ever once in a while if you want me to Jan 28 15:51:29 given that that there's only one release at once it's noddy enough to do builds for stable anyway. Jan 28 15:51:32 broonie: good for those with a key ;-) Jan 28 15:51:56 Well, I can't see why anyone would bother using it except to publish stuff. Jan 28 15:52:24 we just need OE to become fully self hosting :-) Jan 28 15:52:38 broonie: We are talking about publishing stuff Jan 28 15:52:47 broonie: and taking the hassle out of maintaining packages Jan 28 15:53:04 publishing != autobuilding Jan 28 15:53:17 yes? Jan 28 15:54:38 rp: http://www.ubuntu.com/news/launchpad-ppa Jan 28 15:54:49 RP: https://help.launchpad.net/PPAQuickStart Jan 28 15:56:17 Laibsch: It only covers gutsy/hardy, not any of the useful ones... Jan 28 15:57:10 XorA: OE was selfhosting year ago Jan 28 15:57:31 XorA: I built angstrom-console-image under angstrom-console-image/i586 Jan 28 15:58:06 rp: not true Jan 28 15:58:12 you can go back further Jan 28 15:58:50 RP: it goes back to dapper Jan 28 15:58:54 6.06 Jan 28 15:58:56 https://launchpad.net/~r0lf/+archive Jan 28 15:59:12 -> "click on PPA supported series" Jan 28 16:03:19 Laibsch: I think last time our debian guys looked it didn't. I've mentioned it to them though Jan 28 16:03:25 have a nice evening all Jan 28 16:04:06 rp: Check it out. I think it will save you some time and help with fresher packages Jan 28 16:04:48 http://blog.haerwu.biz/2008/01/28/waiting-for-n810-to-arrive/ Jan 28 16:07:41 Laibsch: It doesn't help, you need a different source package for each distro you want to target Jan 28 16:07:58 no, again not true Jan 28 16:08:49 Laibsch: Its what I've just been told Jan 28 16:09:11 well, I have done it and did not need to change anything but ChangeLog Jan 28 16:09:27 First line of that file is the important one Jan 28 16:10:00 "debchange;debuild -s;dput -f ../*.changes" is usually all that is necessary Jan 28 16:10:12 "debchange;debuild -S;dput -f ../*.changes" is usually all that is necessary Jan 28 16:10:43 Laibsch: It is not as simple as that, there are more subtle problems. I trust the person telling me that too. Jan 28 16:11:21 Well, if the package *does* need to be compiled differently, than of course, there is no way around that Jan 28 16:11:46 The usual issues are created by the packaging tools and standards for them changing between releases. Jan 28 16:12:07 But taking OE as an example, you don't usually need to change a bb file because you want to compile for a different distro Jan 28 16:12:33 What you upload to the PPA is basically nothing but a bb file, upstream source and your patches Jan 28 16:12:41 Very similar to OE, indeed Jan 28 16:12:46 You'll also run into trouble if the build dependcies change incompatibilty. Jan 28 16:12:52 the PPA then compiles the stuff for you Jan 28 16:12:52 no, but on debian based distros the dependency names tend to change all over the joint Jan 28 16:12:59 broonie: I think thats the case with the qemu packages Jan 28 16:13:15 I'd expect so, it's got graphical stuff in it. Jan 28 16:14:48 reminds me, I need to learn to do debian packages properly at some point Jan 28 16:15:11 XorA: I'd probably find it easier to teach OE to make them :) Jan 28 16:15:20 RP: it already can :-) Jan 28 16:16:26 XorA: I mean with proper dependencies for the distros in question Jan 28 16:16:29 XorA: It's really not very hard unless you get into interesting install time configuration. Jan 28 16:16:51 broonie: yeah I made one a few weeks ago Jan 28 16:16:59 broonie: but its a bit of a mess Jan 28 16:17:11 RP: Where is the source for your qemu packages? Jan 28 16:17:25 lintian will spot most of the obvious problems for you. Jan 28 16:17:27 RP: well I had hoped the embedian guys would give in and teach OE how to do that :-) Jan 28 16:17:47 broonie: yeah, I just need to sit with the manual and learn Jan 28 16:18:01 Laibsch: http://svn.o-hand.com/view/misc/trunk/qemu-packaging/ Jan 28 16:18:04 Is there anyone actually working on embedian except Neil? Jan 28 16:18:06 XorA: That would be good Jan 28 16:18:52 broonie: I dont think so Jan 28 16:42:36 guys, a OT question - I'll have to travel to Brussels in a week, what (affordable) hotel you recommend? Which will you be staying at FOSDEM? Jan 28 16:42:43 Crofton: I see NYPD want to outlaw noses Jan 28 16:42:53 Bernardo: OE is at Astrid Jan 28 16:43:00 Bernardo: www.astridhotel.be Jan 28 16:43:25 XorA: thanks Jan 28 16:43:41 Bernardo: its ok as hotels go, nothing fancy, but comfy Jan 28 16:43:54 I was thinking of staying at Bedford, but I've seen some awful reviews Jan 28 16:44:23 Bernardo: novatel about 100 years down the road is fancier, but a bit more expensive Jan 28 16:44:28 100 yards Jan 28 16:44:36 or meters for wierd .eu types :-) Jan 28 16:44:45 any ideas on the average rates at the astrid? I don't need a fancy place, but wifi would be a bonus Jan 28 16:44:58 Bernardo: 70 euro a night for FOSDEM Jan 28 16:45:07 oe/oz sites back fine after the domain issues? there any dns changes that need to be made? we'll definately have to get the domains transferred to someone actually involved in the project nowadays sometime soon Jan 28 16:45:10 heh Jan 28 16:45:30 whoa, kergoth at a vaguely .eu timezone :-) Jan 28 16:45:48 kergoth: not got MV converted to OE yet then? Jan 28 16:46:01 :P it's not /that/ early, just 9:45am here, usually i'm at work around 9, on the days i dont work from home Jan 28 16:50:09 XorA: thanks, booked at the astrid Jan 28 17:04:49 ~lart kboot for recompiling all each time Jan 28 17:04:50 * ibot decapitates kboot conan the destroyer style for recompiling all each time Jan 28 17:08:55 * broonie wishes OE tasks took a consistent amount of time to run Jan 28 17:11:15 OE has $sbindir. one of my packages installs a file into $sbindir and I want to call it in postinst. How do I deal with the situation where $sbindir is not /usr/sbin ? Jan 28 17:11:49 Laibsch: use ${sbindir} in postinst Jan 28 17:12:04 Laibsch: it will get expanded by OE before writing to package Jan 28 17:12:16 No, that is not what I want Jan 28 17:12:36 Laibsch: that is what you asked Jan 28 17:12:46 hm, not really Jan 28 17:12:51 I have a script Jan 28 17:13:06 Laibsch: if you change ${sbindir} value then everything will change accordingly Jan 28 17:13:07 Let me give you the diff Jan 28 17:13:59 it sounds like laibsch is using sbindir in his script that he installs, in which case he'd just need to sed it in the do_install, but he hasn't really been that clear about his question, heh Jan 28 17:14:14 XorA: http://oz.leggewie.org/wip/xora.diff Jan 28 17:14:34 kergoth: yes, sorry Jan 28 17:14:39 but you got my drift Jan 28 17:14:45 Was a bit difficult to explain Jan 28 17:16:10 RP: https://launchpad.net/~r0lf/+archive/+builds?build_text=&build_state=all took me the three commands I gave you earlier and I replaced one word with debchange Jan 28 17:16:21 The build took a bit longer to actually start Jan 28 17:17:17 Laibsch: what kergoth said, use sed Jan 28 17:17:32 Laibsch: or cheat and echo that line into the file from inside a .bb Jan 28 17:17:33 OK, thanks Jan 28 17:17:53 Maybe, I'll cheat and echo the whole file ;-) Jan 28 17:20:18 g'day kergoth Jan 28 17:20:55 Laibsch: Thats poky-scripts, not qemu Jan 28 17:21:14 I know Jan 28 17:21:21 qemu is homework for you Jan 28 17:21:22 ;-) Jan 28 17:21:33 It was just an example Jan 28 17:21:34 Laibsch: qemu is the hard one... Jan 28 17:21:43 Lucky you ;-) Jan 28 17:22:07 What price do I get if it builds the same way? Jan 28 17:23:00 Laibsch: eternal glory? Jan 28 17:23:05 mickeyl: wb Jan 28 17:23:09 RP: ah, font-util doesn't fetch now. It's looking in /releases/individual/data/ for the tarball, whereas it's actually in /.../font/. Jan 28 17:23:11 hey pb_, RP, florian Jan 28 17:23:19 zecke, mickeyl: You have mail... Jan 28 17:23:25 hi kergoth Jan 28 17:24:03 RP: OK, eternal glory it is Jan 28 17:24:04 ;-) Jan 28 17:24:12 pb_: ah, I'll fix that Jan 28 17:24:43 night! Jan 28 17:25:36 florian: I forgot to CC you on the mickeyl's response :} Jan 28 17:27:05 zecke: you always forget me ;) Jan 28 17:28:18 it might be on purpose? Jan 28 17:28:34 florian: replied you Jan 28 17:28:47 zecke: ! Jan 28 17:29:08 florian: the bits you didn't get where on: Der Vorstand kann bei der KDE e.V. die Rollen unter sich ausmachen Jan 28 17:29:10 pb_: fix pushed Jan 28 17:29:28 florian: and we need a "Notar" for the protocol and registration of the e.V. Jan 28 17:29:40 florian: for the founding assembly you are right we need a proper protocol.. Jan 28 17:30:36 back to work now :) Jan 28 17:30:36 RP: thanks Jan 28 17:30:39 will we have a proper founding ceremony? Jan 28 17:30:49 zecke: yes... i have some lines and a link to a comperhensive description in my second mail. Jan 28 17:31:05 also, in the prebuilt tool chain in the user manual Jan 28 17:31:13 Crofton: if we don't mess up yeah Jan 28 17:31:19 does the setting of machine and distro work Jan 28 17:31:26 zecke, work hard :) Jan 28 17:31:51 it seems like the example sets them to arbitrary values, do the files need to exist? Jan 28 17:43:31 anybody using apex with an ixp42x board in LE mode with angstrom? Jan 28 17:49:42 anyone familar with using an external tool chain know what the web page means by HOST_TOOLS path? Jan 28 17:49:59 Crofton: i can send you my sample script Jan 28 17:50:08 that would be great Jan 28 17:50:23 ok one sec Jan 28 17:50:27 thans Jan 28 17:50:31 np Jan 28 17:50:34 I am looking at the web page Jan 28 17:50:50 I suspect it is basically ok, we should clean it up some Jan 28 17:56:53 re, was on the phone, soz Jan 28 17:57:00 ok one sec :) Jan 28 17:57:41 np Jan 28 18:00:42 ok, on the way Jan 28 18:03:15 somthing strange is goin on... when I made build I got no space left on device... I free up space and after then it does only one package per bitbake.... each new package first fail to configure.. next this package goes ok Jan 28 18:03:51 Jin^eLD, did you have to create machine and distro files? Jan 28 18:09:02 Crofton: yes Jan 28 18:09:14 well, for machine I reused one of my configs Jan 28 18:09:18 it was ARM so easy stuff Jan 28 18:09:24 but for distro I had to match their crap Jan 28 18:09:35 so had to pick a matching glibc Jan 28 18:10:29 I was not building a full ditsro though, only separate packages, that worked out quite well Jan 28 18:10:48 I usually have to mess with external toolchains when I compile my UPnP server for various NAS boxes (i.e. as addition to vendor hardware) Jan 28 18:11:19 what system are you messing with? Jan 28 18:11:44 PPC based system Jan 28 18:13:36 a+ Jan 28 18:28:50 hi all Jan 28 18:30:19 RP: the latest qemu has circular dependencies on glibc_2.6.1 (do_package)... Jan 28 18:30:36 RP: I'm I supposed to rebuild it? Jan 28 18:31:01 s/Am/I'm/ Jan 28 18:31:30 well..is not my day.... Jan 28 18:31:44 s/I'm/Am/ Jan 28 18:32:52 Hello there. I'm a programmer from The Netherlands, and I'm trying to set up an OpenMoko development environment. I have an iMac G5 with Virtual PC installed, in which I installed Ubuntu (so as to have a standard environment). When using the MokoMakefile to make a development image, I get an error: "/usr/bin/env: bad interpreter". Does anyone have any idea what could be wrong? The error is... Jan 28 18:32:54 ...issued by bitbake. Jan 28 18:33:44 Witch: wrong channel Jan 28 18:34:13 ant_: I'll hang out in #openmoko as well, I'll stay there.. Jan 28 18:34:37 Witch: correct, some developers are common Jan 28 18:35:11 ant_: If you have any idea, you're welcome to hop over there, ofcourse ;-) Jan 28 18:35:45 Witch: sorry, cannot help with openmoko Jan 28 18:35:59 Witch: no idea Jan 28 18:37:06 /bin/sh is dash? Jan 28 18:37:22 that is a wild guess, based on what others have said about ubuntu Jan 28 18:37:23 ant_: Since this is BitBake _itself_, and nothing OpenMoke specific, I tried over here. Just typing 'bitbake/bin/bitbake' gives the error. Jan 28 18:38:02 Witch: is python installed? Jan 28 18:38:12 Crofton: lrwxrwxrwx 1 root root 4 2007-12-01 19:29 /bin/sh -> bash Jan 28 18:38:50 keesj: Python 2.5.1 (r251:54863, May 2 2007, 16:56:35) Jan 28 18:38:51 [GCC 4.1.2 (Ubuntu 4.1.2-0ubuntu4)] on linux2 Jan 28 18:38:52 Type "help", "copyright", "credits" or "license" for more information. Jan 28 18:39:28 Witch: did you *install* bitbake or did you copy it in a dir and updated by hand? Jan 28 18:39:53 ant_: The MokoMakefile retrieved and bilt bitbake for me. Jan 28 18:40:13 Witch: then I cannot help you... Jan 28 18:41:04 Witch: do a head -n 1 `which bitbake ` Jan 28 18:42:21 keesj: I don't think it's in my path. I know where it _is_, though... Jan 28 18:43:50 I don't remeber how it worked with openmoko. sorry and python is kown as python and not python2.5 or similar i hope Jan 28 18:48:02 keesj: No, python is just 'python' Jan 28 18:50:36 Witch: does you path contain "." ? Jan 28 18:51:52 keesj: no, it doesn't (and shouldn't, as far as I know...) Jan 28 18:57:21 hmm ok I run into the same problem with glic_2.6.1 as with 2.5. the first problem is that the kernel headers are missing mainly asm-x86_64 in my case (asm-x86_64/unistd.h that is pointed by from asm/unistd.h) Jan 28 19:09:40 hi all Jan 28 19:11:20 how can i force to build an image without EABI (so OABI) ??? using angstrom stable 2007.1 as distro ... some variable i can set in local.conf or my-machine.conf ... thanks! Jan 28 19:12:35 that should be a faq Jan 28 19:13:06 TARGET_OS = linux I think? Jan 28 19:15:34 Crofton the bad think is that i'm building for an arm9 ... so armv5te that support eabi and so all is eabi ... i have to disable (causae i'm doing some sort of legacy system) ... Jan 28 19:17:25 I understand Jan 28 19:17:45 I think TARGET_OS=linux wil buidl oabi binaries Jan 28 19:18:06 mickeyl, ping Jan 28 19:18:27 pöng Jan 28 19:18:56 did you see my post on generic distro? Jan 28 19:19:07 btw what TARGET_OS builds eabi stuff? Jan 28 19:19:18 linux-gnueabi I think Jan 28 19:19:43 I think it is set in distro, but you can override from local.conf Jan 28 19:19:49 * Crofton hopes he has this correct Jan 28 19:20:06 Jin^eLD, Crofton ... in my local.conf at now TARGET_OS is unset .... so probably angstrom set it ... let's try :) Jan 28 19:20:12 thx Jan 28 19:20:18 yeah I just found it, distro conf sets it Jan 28 19:21:38 Crofton: you're right. i just replied to that. Jan 28 19:21:41 thanks Jan 28 19:21:50 Jin^eLD that's sure ... but i don't know exactly how to override in local.conf :) Jan 28 19:21:51 I think the key thing is we clearly define what it is Jan 28 19:21:54 right Jan 28 19:22:09 to prevent it becomign the distro for people that do not like Angstrom :) Jan 28 19:22:11 i guess i just take a couple of hours tomorrow and strip angstrom 2007 down into minimal.conf Jan 28 19:22:32 If it builds without a line, leave it out :) Jan 28 19:22:34 (yes, this is the 3rd reason, but for diplomatic reasons i left it out) Jan 28 19:22:39 yeah Jan 28 19:22:49 me too Jan 28 19:23:15 we may need to look at how images build, they may have angstrom dependencies we should work around Jan 28 19:24:10 Wind River invited me to the 2008 Mobile WOrld Congress Jan 28 19:24:15 with a blank email Jan 28 19:24:17 oh Jan 28 19:24:17 morons Jan 28 19:24:18 that's nice Jan 28 19:24:20 heh Jan 28 19:24:22 that's dumb ;) Jan 28 19:24:24 yeah Jan 28 19:24:44 what mean 'blank email' ? Jan 28 19:24:54 no content Jan 28 19:24:59 only the unsubscribe line Jan 28 19:25:06 hahahah only subject/title ... HAHAHAH Jan 28 19:25:09 suspects all the content is html or something Jan 28 19:25:10 Crofton: maybe they have a secret message in the header? Jan 28 19:25:36 probably Outlook renders a webpage from X-Hidden-Header: Jan 28 19:25:49 ah view as html, I alwys set that to plain text Jan 28 19:26:05 what Limo Jan 28 19:27:44 Crofton: at least they did not send you a black mail with a .doc as attachment ;) Jan 28 19:27:49 yeah Jan 28 19:28:22 Witch: one more and last "idea" did you by any chance used Jan 28 19:28:35 wrong programs to transfer the files? Jan 28 19:28:56 what does file bitbake/*/bitbake say? Jan 28 19:37:15 anybody using apex did the switch from be to le? Jan 28 19:43:17 keesj, ant_: it turned out that I have my openmoko build dir on a different file system, and I needed the 'exec' permission in /etc/mtab! Jan 28 19:43:47 So, I'm now 'bitbaking' my openmoko files. Jan 28 19:43:56 Thanks for the pointers! Jan 28 20:06:05 Crofton, Jin|away .. th creation of the armv5te-angstrom-linux-gnueabi don't sound good :( Jan 28 20:15:23 how can i check if i'm building a oabi or eabi toolchain / libc ??? Jan 28 20:27:46 angstrom-glibc.inc contain TARGET_OS = ... that override all (both local.conf and my-machine-conf) grrr Jan 28 20:28:54 ERROR: Task /local/pkg/oe/org.openembedded.dev/packages/qemu/qemu-native_0.9.1.bb (do_package) has circular dependency on /local/pkg/oe/org.openembedded.dev/packages/glibc/glibc_2.5.bb (do_package) Jan 28 20:28:56 anyone seen that? Jan 28 20:28:58 news to me Jan 28 20:30:58 mickeyl: Yes, I think somebody already reported it about 30 mins ago Jan 28 20:31:03 ah Jan 28 20:31:09 RP ? Jan 28 20:32:25 mickeyl: Did you see Jan 28 20:32:29 !oebug 3711 Jan 28 20:32:31 * * Bug 3711, Status: NEW, Created: 2008-01-23 13:23 Jan 28 20:32:32 * * : offline wikipedia Jan 28 20:32:32 * * http://bugs.openembedded.net/show_bug.cgi?id=3711 Jan 28 20:32:44 I think we are quite close Jan 28 20:32:55 cool Jan 28 20:33:03 I mailed the guy and he told me that there are already people using his stuff on iphones Jan 28 20:33:16 I have not worked on it yet Jan 28 20:33:21 Laibsch: you nick highlighted me yesterday, what's up? Jan 28 20:33:26 But the steps seem quite clear Jan 28 20:33:44 summatusmentis: -> hentges.net Jan 28 20:34:24 aah Jan 28 20:34:26 do_package Jan 28 20:34:54 hmm, no Jan 28 20:34:55 damn Jan 28 20:35:27 ah, it's libsdl-native Jan 28 20:35:40 Laibsch: ah yes, my bad, wrong window Jan 28 20:46:55 mickeyl: hmm :/ Jan 28 20:47:22 kind of strange Jan 28 20:47:50 this is a fresh checkout Jan 28 20:47:59 on a new build machine Jan 28 20:48:00 The new circular dependency logic in bitbake should help with these kind of errors Jan 28 20:48:10 Mine is parsing now... Jan 28 20:48:28 Should be done in a couple of hours :/ Jan 28 20:48:40 * RP needs to bite the bullet and replace the server Jan 28 20:50:26 oooh Jan 28 20:50:29 heh Jan 28 20:51:46 hmm, no Jan 28 20:53:03 qemu-native.inc in poky is rather different to that in oe. Specifically the lack of an inherit native Jan 28 20:55:31 indeed Jan 28 20:55:36 hmm, who changed that last time? Jan 28 20:56:04 ah Jan 28 20:56:09 mickeyl: I guess each -native .bb did this... Jan 28 20:56:24 adding inherit native makes it work again Jan 28 20:56:48 Without that it doesn't inhibit the default dependencies and hence pulls in glibc Jan 28 20:56:52 yeah Jan 28 20:57:10 i just wonder which revision that came with Jan 28 20:57:13 must be very recent Jan 28 20:57:21 * mickeyl suspects a bogus merge Jan 28 20:57:22 I broke it earlier Jan 28 20:57:26 oh you? Jan 28 20:57:27 good. Jan 28 20:57:36 I assumed the -native had the inherit native Jan 28 20:57:49 right Jan 28 20:57:57 the proper fix is to do it in the -native.inc then Jan 28 20:58:01 will you or shall i? Jan 28 20:58:19 I'll fix, I just want to check something first Jan 28 20:58:23 k Jan 28 20:58:36 The qemu files need some cleanup :/ Jan 28 20:58:44 This was an attempt to start that... Jan 28 21:04:57 how do i download poky's qemu for gentoo...it's only for debian or apt-based distributions... Jan 28 21:06:00 GNUtoo: Use a toolchain tarball from the poky autobuilder then Jan 28 21:21:11 RP, i have made a find ./ | grep qemu and it didn't find qemu on both arm and i586 folders... Jan 28 21:21:34 GNUtoo: how about qemu-arm ? Jan 28 21:23:12 RP, the tarballs comes from there: http://pokylinux.org/releases/blinky-3.0/and are poky-3.0-blinky-arm-toolchain.tar.bz2 and poky-3.0-blinky-i586-toolchain.tar.bz2 Jan 28 21:23:43 hcid.conf for openmoko is wrong and contains class fro desktop as opposed to a phone Jan 28 21:23:43 GNUtoo: Try http://pokylinux.org/autobuild/toolchain Jan 28 21:25:04 03pfalcon 07org.oe.dev * re6d6ee58... 10/ (4 files in 4 dirs): Jan 28 21:25:04 packages/bluez/: Unkludge "files" dir. Jan 28 21:25:04 * Start with bluez-dtl1-workaround stuff. Jan 28 21:26:06 03pfalcon 07org.oe.dev * r4ba7dc2b... 10/ (4 files in 4 dirs): bluez-utils 3.23: Move hciattach-ti-bts.patch to version-specific dir. Jan 28 21:26:59 03pfalcon 07org.oe.dev * r1c44ae3b... 10/ (6 files in 4 dirs): Jan 28 21:26:59 bluez-utils: Move version-overridable stuff away from files/. Jan 28 21:26:59 * Or it will be pulled in randomly. Jan 28 21:27:30 03pb 07org.oe.dev * r5a2bd088... 10/ (1 packages/gcc/gcc-cross-kernel-3.3.4_3.3.4.bb): Jan 28 21:27:30 gcc-cross-kernel: permute targets in do_stage to avoid 'No such file or Jan 28 21:27:30 directory' Jan 28 21:28:11 03thesing 07org.oe.dev * rdbc92be1... 10/ (6 files in 5 dirs): commit latest collie work from thesing (zaurus-2.6.inc, linux-rp*, task-base) Jan 28 21:29:25 03rpurdie 07org.oe.dev * rd039068c... 10/ (1 packages/xorg-font/xorg-font-common.inc): xorg-font-common.inc: Fix SRC_URI Jan 28 21:30:20 03pfalcon 07org.oe.angstrom-2007.12-stable * r654323da... 10/ (6 files in 2 dirs): Jan 28 21:30:20 applied changes from 9814a7ac8b5ebeadd190ad216196e5efd3a70ff3 Jan 28 21:30:20 through 33c44c7c7ce5fce5da6c869be191126d5fca6f96 Jan 28 21:30:20 bluez-utils 3.23 and friends: Set FILESPATH explicitly and correctly. Jan 28 21:30:48 hi everybody Jan 28 21:31:14 03pfalcon 07org.oe.angstrom-2007.12-stable * r2037125f... 10/ (1 BACKPORTS.txt): BACKPORTS.txt: Note 33c44c7c7ce5fce5da6c869be191126d5fca6f96 Jan 28 21:31:54 mickeyl: xapian is as easy as pie to compile in OE (re offline wikipedia) Jan 28 21:32:05 mickeyl: How would one compile http://users.softlab.ece.ntua.gr/~ttsiod/quickstartindex.cc Jan 28 21:32:12 It uses xapian Jan 28 21:32:23 Would we need a xapian-native then Jan 28 21:32:28 Or libxapian? Jan 28 21:34:19 i need to look into that Jan 28 21:34:21 can't tell it offhand Jan 28 21:35:00 well, it's a header file Jan 28 21:35:06 so i guess you need to add libxapian Jan 28 21:40:30 RP, i have ./poky-qemu: line 52: .: filename argument required for # ./poky-qemu "/oe/tmp/deploy/glibc/images/qemuarm/zImage-2.6.23-qemuarm-20080127233927.bin" "/oe/tmp/deploy/glibc/images/qemuarm/Angstrom-opie-image-glibc-ipk-2008.1-test-20080128-qemuarm.rootfs.ext3" Jan 28 21:41:21 GNUtoo: I suspect it can't guess the machine or image type from those filenames Jan 28 21:41:48 qemuarm usually uses ext2 Jan 28 21:43:32 I just had bibtake nano fail because it could not find ipkg-build Jan 28 21:45:32 RP,it doesn't work even with MACHINE=spitz , i'll try to make an ext2 image Jan 28 21:46:29 GNUtoo: I'm just guessing at the problem, those scripts are simple so you could always debug them directly too Jan 28 21:47:24 XorA: hcid.conf for openmoko is wrong and contains class for a desktop as opposed to a phone, so it gets reported as a desktop when scanned Jan 28 21:49:54 hi ljp Jan 28 21:52:03 hey pb Jan 28 21:53:12 RP, thanks a lot it works Jan 28 22:00:21 03rpurdie 07org.oe.dev * r8aaa63b7... 10/ (7 files in 2 dirs): qemu: Fix missing inherit native and cleanup various files Jan 28 22:03:09 ljp: file a bug please Jan 28 22:05:00 ok Jan 28 22:11:09 ljp: I remember being told that was intentional so the device could advertise more features... Jan 28 22:19:28 re Jan 28 22:19:40 wb florian Jan 28 22:22:32 hello, what happened to oe bugzilla? Jan 28 22:23:26 ahh, .net :) Jan 28 22:31:54 qemu-native seems to have a staging error: http://rafb.net/p/GL0mvF43.html Jan 28 22:35:25 apaulsen: building right now..task 157 Jan 28 22:45:20 03thebohemian 07org.oe.dev * r69912ac0... 10/ (1 conf/distro/chinook-compat.conf): Jan 28 22:45:20 chinook-compat: Jan 28 22:45:20 - Added description and a note Jan 28 22:45:25 03thebohemian 07org.oe.dev * r8e7f10e1... 10/ (5 files in 4 dirs): Jan 28 22:45:25 dbus-java: Added recipe. Jan 28 22:45:25 libmatthew: Added recipe (dependency for dbus-java). Jan 28 22:46:24 apaulsen: My fault I think. not my day today... Jan 28 22:47:39 RP: no worries, I just commented out the '-s' and proceeded Jan 28 22:47:59 apaulsen: built fine on c7x0 angstrom DISTRO_VERSION = "2008.1-test-20080128" Jan 28 22:48:19 apaulsen: Actually, I misread that log, there is another problem too : Jan 28 22:48:21 :/ Jan 28 22:48:32 RP: need logs ? Jan 28 22:50:42 ant_: No, I know what the problem is thanks Jan 28 22:54:25 RP: don't forget the gcc-3.4.4 (native) elegant solution for qemu Jan 28 22:54:57 RP: with all that talk about external toolchain...this is an important bit IMHO! Jan 28 22:55:04 ant_: Don't you think I've broken enough today ;-) Jan 28 22:56:04 RP: that gcc works perfectly and avoid big pain Jan 28 23:00:16 ant_: ok, I'll commit something to help with that too Jan 28 23:00:53 RP: be confisdent Jan 28 23:01:18 RP: be confident if there are some obscure bug I'll stumble in it! Jan 28 23:01:41 RP: ever heard about Cassandra... Jan 28 23:05:40 RP: any particular reason for the delay adopting 2.6.24 for Zaurus? Jan 28 23:05:51 ant_: Lack of testing Jan 28 23:06:08 RP: you know, I'm patching sharpsl.c... Jan 28 23:06:21 ant_: What for? Jan 28 23:06:41 * ant feels dirty and confess: u-boot Jan 28 23:06:59 ant_: You know my view on uboot ;-) Jan 28 23:07:12 in this context anyway Jan 28 23:07:53 ok, native.bbclass fix, and gcc-native3 changes pushed.... Jan 28 23:08:26 RP: about that kernel: fixed cmdline is dictaded by the sharp bootloader passing bogus atags, isn't? Jan 28 23:09:05 RP: solved in 2.6.24, right (first atags, then cmdline)? Jan 28 23:09:06 ant_: It passes bogus data in the parameter area Jan 28 23:09:28 RP: ok, that's why I'm playing with mtdparts Jan 28 23:09:32 ant_: 2.6.24 accept a commandline through atags Jan 28 23:09:52 atags != parameter area Jan 28 23:09:52 RP: preferred over cmdline? Jan 28 23:10:16 ant_: atags win over everything else Jan 28 23:10:25 RP: ok, the most important param is the partitioning schema Jan 28 23:10:31 RP: atm Jan 28 23:10:33 ant_: The bootloader needs to pass them in r1 Jan 28 23:11:33 RP: that's done by the angel bootloader isn't? Jan 28 23:11:51 RP: the first hw init Jan 28 23:11:59 ant_: no, they're written into address 0xa0001000 iirc Jan 28 23:12:28 RP: I have to study deeper the sources.... Jan 28 23:12:57 RP: but can boot all possible images using your kernel... Jan 28 23:17:32 RP: anyway the orig params in mtd1 (http://externe.net/zaurus/flash/nandmap.jpg) get overwritten using u-boot Jan 28 23:23:08 ant_: Doesn't matter... Jan 28 23:24:12 RP: sorry to bother you again on the matter...hope this is the last time... Jan 28 23:24:21 RP: thx Jan 28 23:24:31 gn Jan 28 23:24:46 03rpurdie 07org.oe.dev * r674ed77b... 10/ (1 classes/native.bbclass): native.bbclass: Change prefix so its within the build-arch directory, hence moving on step closer to sysroot. Nothing much uses the prefix path from native.bbclass apart from qemu which this change fixes a problem with Jan 28 23:24:56 03rpurdie 07org.oe.dev * r2386d883... 10/ (1 packages/gcc/gcc-native_3.4.4.bb): gcc-native-3.4.4: Add PROVIDES gcc-native3 Jan 28 23:25:05 03rpurdie 07org.oe.dev * redeeced1... 10/ (1 packages/qemu/qemu-native.inc packages/qemu/qemu_0.9.1.bb): qemu: Bump PR to make sure nobody is bitten by the native.bbclass prefix issue. Add DEPENDS on gcc-native3 Jan 28 23:25:19 03rpurdie 07org.oe.dev * re5bcfd0c... 10/ (1 conf/local.conf.sample): local.conf.sample: Document how to avoid compiling gcc-native3, cleanup other ASSUME_PROVIDED documentation Jan 28 23:28:44 hmm fun Jan 28 23:28:52 i can no longer stop qemu-arm no longer Jan 28 23:28:57 kill -9 doesn't work Jan 28 23:28:58 wtf Jan 28 23:29:04 and it consumes 100% CPU Jan 28 23:29:04 *sigh* Jan 28 23:29:20 mickeyl: :( Jan 28 23:29:25 ps -efaux Jan 28 23:29:26 oops Jan 28 23:30:13 mmh Jan 28 23:30:15 reboot time Jan 28 23:46:18 * Crofton hearts OE Jan 29 00:25:37 g'night Jan 29 00:26:46 n8 mickey|zzZZzz Jan 29 01:45:59 anyone helpful around? **** ENDING LOGGING AT Tue Jan 29 02:59:57 2008