**** BEGIN LOGGING AT Sun Jul 10 23:59:57 2005 Jul 11 00:47:28 Hi Jul 11 00:55:24 morning Jul 11 01:01:15 morning all Jul 11 01:04:41 hi all Jul 11 01:05:00 anyone tried building an image with bluez included, lately? Jul 11 01:06:52 which RDEPEND will store the kernel image into filesystem. cant find zImage in my filesystem. Jul 11 01:12:53 morning Jul 11 01:14:05 oh balls! just apt-get'd monotone 0.20, whoops Jul 11 01:15:45 ade|desk: time for a dpkg hold :-) Jul 11 01:26:45 <__law__> is the OE repo incompatible with montone 0.20? Jul 11 01:33:44 How is a dependency to a package resolved? I still cant include bluez-utils.. Jul 11 01:34:17 because it depends on bluez-libs which builds a libbluetooth1.{PV}.ipkg Jul 11 01:34:17 __law__: monotone server and client must match. monotone server from vanille is 0.19 so client must be 0.19 too. Jul 11 01:46:11 i want to statically build all modules to be with in the kernel. the kernel has been compiled succesfully and itask do_package, function do_install fails saying that : /bin/sh: line 1: scripts/mod/modpost: No such file or directory Jul 11 01:46:24 re Jul 11 01:46:39 hi zecke Jul 11 01:47:09 i solved my missing dependencies problems. Jul 11 01:48:04 eJumbo: scripts/mod/modpost Jul 11 01:48:12 should be in your kernel directory Jul 11 01:50:37 cant find that in my kernel direcroty i have modpost.c and related files. Jul 11 01:52:12 i have a makefile in it which make modpost; i think it has not been executed. Jul 11 01:59:02 damn.. i cant get bitbake to work on my debian Jul 11 02:02:07 morning Jul 11 02:04:01 dolpp: I dount it has anything to do with debian Jul 11 02:04:51 i dont know.. but i just thought i'd mention my distro there Jul 11 02:06:05 eJumbo: does do_install outputs "no modules to install" in the beginning? Jul 11 02:06:25 installed it under /usr and data /usr/lib Jul 11 02:06:49 03mickeyl 07org.openembedded.dev * r5c435b81... 10/packages/busybox/ (busybox-1.00/defconfig busybox_1.00.bb): Busybox: enable the -l option for umount Jul 11 02:06:53 03mickeyl 07org.openembedded.dev * r9b55f620... 10/packages/initscripts/ (14 files in 14 dirs): Initscripts: add udev guard to all devices scripts Jul 11 02:06:57 03mickeyl 07org.openembedded.dev * r9158de30... 10/packages/linux-hotplug/ (files/mmc.agent linux-hotplug_20040920.bb): Hotplug mmc-agent: Increase delay to give udev based systems more time to create the device node Jul 11 02:06:58 guess 4 lines aint too much of flooding.. Jul 11 02:07:01 03mickeyl 07org.openembedded.dev * r317bfd00... 10/packages/python/ (3 files): Upgrade python-pylinda to 0.5 Jul 11 02:07:08 Traceback (most recent call last): Jul 11 02:07:08 File "/usr/bin/bitbake", line 27, in ? Jul 11 02:07:08 from bb import utils, data, parse, debug, event, fatal Jul 11 02:07:08 ImportError: cannot import name parse Jul 11 02:07:17 getting that after typing bitbake Jul 11 02:08:22 giel: yes. please see http://pastebin.com/311063 the problem gets resolved if i enable MODULE LOADING AND UNLOADING SUPPORT in the kernel defconfig Jul 11 02:08:40 dolpp: which version of bitbake are you using ? Jul 11 02:09:22 i would like to write a little serial driver. its only functions should be receiving an RTS interrupt and sending a CTS signal! can anybody help me? Jul 11 02:10:11 hmm.. think its 1.3.2 .. took it from berlios.de via svn Jul 11 02:10:21 ade|desk -^ Jul 11 02:11:20 bluez include update... managed to include bluez-utils Jul 11 02:12:06 had to add an RDEPEND on libbluetooth1 instead of bluez-libs Jul 11 02:12:29 this is meant to be that way? Jul 11 02:17:04 "instead" ? Jul 11 02:17:15 i don't see bluez-libs in RDEPENDS of bluez-utils Jul 11 02:23:36 anyone got suggestions bout my bitbake problem? Jul 11 02:28:23 what problem u r facing ? Jul 11 02:28:46 only after you already checked whether a) the module is actually there, and b) PYTHONPATH is correct [in case you used setup.py] Jul 11 02:28:53 dolpp: try bitbake without installing it. just c Jul 11 02:30:39 set PATH, PYTHONPATH to point to PATH => bitbake/bin PYTHONPATH => bitbake/lib/bb it may work Jul 11 02:31:08 it did not work when i installed; i am using bitbake with installing it. Jul 11 02:32:18 03mickeyl 07org.openembedded.dev * rc2c56935... 10/packages/ (10 files in 2 dirs): Upgrade BlueZ to 2.18 Jul 11 02:32:22 03mickeyl 07org.openembedded.dev * rb0e0a2f9... 10/packages/bluez-libs/bluez-libs_2.18.bb: Bluez-libs: adjust for removed headers Jul 11 02:32:50 mickeyl: sorry, was afk - I had problems including bluez-utils in the first place because bluez-libs was missing Jul 11 02:33:12 mickeyl: I got 'unresolved dependency' errors when building my rootfs image Jul 11 02:33:57 i don't see how that can happen Jul 11 02:34:03 bluez-utils DEPENDs on bluez-utils Jul 11 02:34:18 the shlibdeps code automatically makes bluez-utils.ipk RDEPENDS on the proper bluez-libs Jul 11 02:34:30 at least that's how it works here Jul 11 02:34:53 the strange thing is - the libs got installed into staging correctly, but not in the image Jul 11 02:35:05 bluez-utils did build fine, but I could not install it. Jul 11 02:35:38 see here: Jul 11 02:35:40 mickey@r2d2:/local/pkg/oe/c7x0$ ipkgcontrol tmp/deploy/ipk/bluez-utils-nodbus_2.18-r3_armv5te.ipk Jul 11 02:35:41 Package: bluez-utils-nodbus Jul 11 02:35:48 Depends: update-rc.d, libbluetooth1 (>= 2.18), libgcc1 (>= 3.4.4+csl-arm-20050603), libc6 (>= 2.3.5+cvs20050627), libusb (>= 0.1.10a) Jul 11 02:36:00 check if the dependencies are correct for your version of bluez-utils Jul 11 02:36:39 ok, just a moment.. Jul 11 02:37:39 what provides ipkgcontrol? Jul 11 02:38:05 mickeyl: One thing I'm wondering... Is there a correct way to add users to packages that require specific users? (openntpd, mysql, postfix, cyrus-imapd) Jul 11 02:41:03 NAiL: i'm not sure, i never did this. I would think that we could call 'adduser' etc. in the postinstall scripts. This is what the "big" distributions to as well, isn't it? Jul 11 02:41:13 koobla: ipkgcontrol is provided by my .bashrc Jul 11 02:41:26 ipkgcontrol () Jul 11 02:41:27 { Jul 11 02:41:27 ar p $1 control.tar.gz | tar -xzO ./control; Jul 11 02:41:27 return $? Jul 11 02:41:27 } Jul 11 02:41:31 mickeyl: thx Jul 11 02:42:56 mickeyl: Yes, that's what I'm doing. I was asked to check by rwhitby though, in case there were any oe things that would certainly break by it. I guess I can add a check to see if adduser is there at all, and just echo that a user needs to be added manually if not Jul 11 02:43:31 NAiL: yes, that would be best. even better would be if you can factor that code out into a bitbake class since it would be reusable. Jul 11 02:43:53 mickeyl: Ok, any documentation on how to do bb classes? :) Jul 11 02:44:07 NAiL: only 50 examples :) Jul 11 02:44:16 mickeyl: Package: bluez-utils-nodbus Jul 11 02:44:18 Version: 2.16-r2 Jul 11 02:44:19 Description: Linux Bluetooth Stack Userland Utilities. Jul 11 02:44:21 Section: console Jul 11 02:44:22 Priority: optional Jul 11 02:44:24 Maintainer: OpenEmbedded Team Jul 11 02:44:25 Architecture: armv5te Jul 11 02:44:27 OE: bluez-utils-nodbus-2.16 Jul 11 02:44:27 mickeyl: Oh, I need ~70 ;-) Jul 11 02:44:28 Depends: update-rc.d, libbluetooth1 (>= 2.16), libgcc1 (>= 3.4.4+csl-arm-20050603), libc6 (>= 2.3.5+cvs20050627) Jul 11 02:44:30 Provides: bluez-pan, bluez-sdp, bluez-utils Jul 11 02:44:31 Source: http://bluez.sourceforge.net/download/bluez-utils-2.16.tar.gz file://base.patch;patch=1 file://blueboxes.patch;patch=1 file://hcid.conf file://bluetooth.default file://bluetooth.conf file://default-manfid-0x0-to-bcps.patch Jul 11 02:44:58 thats the one that gets included into the rootfs correctly. Jul 11 02:45:33 NAiL: basically a bbclass is just a .bb file - they use the same syntax and semantics. Have a look at update-rc.d.class and see in dropbear how that's used. Jul 11 02:45:44 koobla: looks good to me Jul 11 02:45:47 mickeyl: Thanks a bunch :) Jul 11 02:45:49 must have been some odd distribution side-effect... will keep an eye on it Jul 11 02:46:01 np Jul 11 02:46:43 _keturn: ping? Jul 11 02:54:05 i hate that bitbake.. im just too dumb it seems Jul 11 02:54:13 dolpp: I see the same problem. Jul 11 02:54:47 mickeyl: Problem with bitbake, remove all the .pyc and rerun bitbake. You will end up with dollpp problem. Jul 11 02:55:11 hmm.. any solution? Jul 11 02:56:17 dolpp: Just an import problem. I'm looking it. Jul 11 02:56:45 parse seems to be missing an __init__.py Jul 11 02:57:11 mickeyl: zecke moved the parse into parse/parse_py. Jul 11 02:59:03 yeah Jul 11 02:59:05 mickeyl: Can I import parse.parse_py ? Jul 11 02:59:10 and that seems to miss an __init__.py Jul 11 02:59:19 not until you have an __init__.py Jul 11 02:59:20 :D Jul 11 02:59:25 * mickeyl does the parrot thing Jul 11 02:59:53 Ok, zecke likes putting indirections everywhere. Jul 11 03:00:47 i think current bitbake is broken, you should revert some revisions Jul 11 03:01:00 ouch Jul 11 03:01:41 openslug is using 269 with good success ;) Jul 11 03:02:51 try downgrading the parser directory to r278 Jul 11 03:03:44 rwhitby: Morning :) Jul 11 03:05:52 g'day Jul 11 03:07:13 morning Jul 11 03:07:17 hey rwhitby Jul 11 03:08:00 hey reenoo Jul 11 03:08:13 mickeyl: that downgrade thing was for me or proti or? hmm.. hows that done Jul 11 03:09:21 dolpp: cd lib/bb; svn update -r278 Jul 11 03:10:00 dolpp: -r 280 is good too. Jul 11 03:10:15 mickeyl: It includes my fix for pb's problem. Jul 11 03:12:02 mickeyl: 281-282 fuck up completely the bitbake. Can we repair it ? Jul 11 03:12:22 cya later Jul 11 03:12:32 03mickeyl 07org.openembedded.dev * r782ff10e... 10/packages/linux-hotplug/linux-hotplug_20040920.bb: linux-hotplug: make it use the first 'install' binary in the PATH. FIXES #129 Jul 11 03:13:04 guess i wasnt that dumb after all.. :) Jul 11 03:13:28 proti: let zecke handle that. Jul 11 03:13:43 proti: he is doing a parser redirection which automatically chooses the best available parser Jul 11 03:14:14 mickeyl: Seems to be missing from svn. It looks like a partial commit. Jul 11 03:14:19 proti: how is a removal of a contrib fucking stuff up? Jul 11 03:14:50 and what is broken anyway? Jul 11 03:14:53 zecke: parse/__init.py has been deleted and there is still import parse. Jul 11 03:15:09 zecke: remove all the *.pyc in bitbake and you'll see. Jul 11 03:15:10 oh __init__.py is missing Jul 11 03:16:18 yeah sorry, I will add the file Jul 11 03:16:29 proti: and to answer your question, it is not repairable I guess Jul 11 03:17:04 gimme a shout when the latest rev works Jul 11 03:17:38 03mickeyl 07org.openembedded.dev * r5a9a8832... 10/packages/libpcap/ (4 files in 2 dirs): libpcap: remove version 0.8.1 and fix do_stage for 0.8.3 Jul 11 03:18:13 zecke: Make bitbake work is enough. Jul 11 03:24:56 Bah. I think the Debian gcc 4 transition broke my build. uicmoc-native-2.3.10-r0 doesn't want to build any more. Jul 11 03:29:30 still getting that parse stuff after did that svn update -r278.. hmm Jul 11 03:29:34 Noodles: sounds like qte-2.3.10 needs some gcc4 patches. shouldn't be many though Jul 11 03:33:48 eJumbo: can you pastebin your .config? Jul 11 03:34:51 giel: i got it working by enabling MODULE LOADING AND UNLOADING in .config Jul 11 03:34:53 are u facing any problem ? Jul 11 03:35:14 no Jul 11 03:35:30 but your solution doesn't satisfy me :) Jul 11 03:35:46 imo you should be able to build a kernel w/o module support :) Jul 11 03:36:29 but if you're okay with this solution, it's okay with me as weel Jul 11 03:36:31 well* Jul 11 03:39:05 i can believe this :/ Jul 11 03:39:10 cant*¨ Jul 11 03:42:43 03mickeyl 07org.openembedded.dev * re5dfa0aa... 10/packages/bluez-utils/ (bluez-utils-common_2.18.inc bluez-utils-dbus/dbus.patch): bluez-utils: update dbus patch courtesy Patrick Steiner Jul 11 03:43:04 giel: i really dont know why i got that error. all i can say is that makefile creates mod which may be did not execute. Jul 11 03:44:04 mickeyl: Yeah, I figured as much; I'll have a poke later - currently listening to a talk on emdebian (I'm at Debconf). Jul 11 03:44:09 if you want a kernel without module support, it should not execute that specific makefile. question remains, what and why was looking for modpost, against your specific instructions Jul 11 03:45:00 after hitting "cd lib/bb; svn update -r278" bitbake says: Jul 11 03:45:01 Traceback (most recent call last): Jul 11 03:45:01 File "/usr/bin/bitbake", line 27, in ? Jul 11 03:45:01 from bb import utils, data, parse, debug, event, fatal Jul 11 03:45:01 ImportError: cannot import name parse Jul 11 03:45:39 do you have parse/__init__.py ? Jul 11 03:45:42 how big is it? Jul 11 03:46:15 lib/bb/parse/ Jul 11 03:46:15 -rw-r--r-- 1 root root 2399 Jul 11 06:43 __init__.py Jul 11 03:47:14 giel: http://pastebin.ca/17439 Jul 11 03:47:27 looks good Jul 11 03:47:36 no idea why it doesn't work now Jul 11 03:47:38 it does here Jul 11 03:49:19 maybe you should globally revert bitbake to r280 Jul 11 03:50:43 damn how do i do that.. dunno shit about svn Jul 11 03:51:51 -> rtfm :) Jul 11 03:52:08 svn -r 280 Jul 11 03:52:31 ie. cd bitbake; svn -r 280 up Jul 11 03:52:47 03mickeyl 07org.openembedded.dev * ref5052f6... 10/packages/zroadmap/zroadmap_1.0.8.bb: zroadmap: supply -rpath-link manually. FIXES #140 Jul 11 03:53:25 should this work just fine? Jul 11 03:53:25 svn co -r280 svn://svn.berlios.de/bitbake/trunk/bitbake Jul 11 03:53:31 i did use that one.. Jul 11 03:53:34 yup Jul 11 03:54:02 great Jul 11 03:54:52 i must be cursed Jul 11 03:55:02 eJumbo: i think it's the genksyms script which is forced to build in kernel.bbclass, even if module support is off Jul 11 03:56:08 dolpp: you're running /usr/bin/bitbake, did you install the version you just checked out there or is that an older one? Jul 11 03:56:57 no idean giel: i have not an expert in oe; just starting off. learning slowly with the help of all the guys over here esp. zecke. Jul 11 03:58:27 giel: deleted all the old files: /usr/bin/bitbake and bbimage, /usr/lib/bitbake and fetched clean revision. and then changed setup.py python -> python2.4 Jul 11 03:58:40 and then: ./setup.py install --prefix=/usr --install-data=/usr/lib Jul 11 03:59:06 and now when i type "bitbake" i get: Jul 11 03:59:07 Traceback (most recent call last): Jul 11 03:59:08 File "/usr/bin/bitbake", line 27, in ? Jul 11 03:59:08 from bb import utils, data, parse, debug, event, fatal Jul 11 03:59:08 ImportError: cannot import name parse Jul 11 03:59:09 dolpp: try just not running setup, but run bitbake from its own dir Jul 11 03:59:17 no reason to put it somewhere in your fs Jul 11 03:59:42 so not to run setup.py but instead go to bitbake/bin/ and run ./bitbake? Jul 11 04:00:04 or put bitbake/bin in your path for a moment Jul 11 04:00:21 and i think your original problem is the --install-data Jul 11 04:00:33 shouldn't the lib thingies go in /usr/lib/python or the like? Jul 11 04:02:19 oe wiki says: The package must be installed in /usr and the data in /usr/lib .. tho thats the case with fc3 Jul 11 04:02:46 gonna try without setup.py now Jul 11 04:04:21 mm'kay export PATH="$PATH:/home/dolphin/bitbake/bin" Jul 11 04:04:30 u can work on bitbake without installing it also. Jul 11 04:04:55 ./usr/bin/env: python: No such file or directory Jul 11 04:05:20 changed bitbake's first line to: #!/usr/bin/env python2.4 Jul 11 04:05:26 i got a similar error to yours when i installed it. so i used bitbake without installing. dolpp: try it Jul 11 04:05:53 and now its giving me: Jul 11 04:05:54 from bb import utils, data, parse, debug, event, fatal Jul 11 04:05:54 ImportError: cannot import name parse Jul 11 04:06:00 argh! Jul 11 04:07:19 there Jul 11 04:07:32 what? where? :D Jul 11 04:07:36 maybe it's because in parse/__init__.py it says python instead of python2.4 as well. try a ln -s /usr/bin/python2.4 /usr/bin/python Jul 11 04:07:54 (which is btw generally a good idea) Jul 11 04:08:00 dolpp: here, actually ;-) Jul 11 04:08:15 03ccsmart 07org.openembedded.nslu2-linux * r39ec5c10... 10/packages/ntp/files/ntp.conf: Adding local HW clock and default security. Jul 11 04:08:19 03nail 07org.openembedded.nslu2-linux * r25f9f814... 10/conf/distro/openslug.conf: Set PREFERRED_VERSION for kernel to 2.6.11.2 until mtd issue is resolved Jul 11 04:08:21 okey changed bitbake python line back and made symlink Jul 11 04:08:23 03jbowler 07org.openembedded.nslu2-linux * r1031d19d... 10/packages/nfs-utils/files/nfsserver: Jul 11 04:08:23 This is a sufficient startup script to get a disk-full nfs server working. Jul 11 04:08:23 On a flash based server or any (user) configuration where /var is in tmpfs Jul 11 04:08:23 or ramfs (so is not populated on boot) the script will not work because Jul 11 04:08:23 it must create the /var/lib/nfs directory and contents. Consequently at Jul 11 04:08:24 present the script is not in the build. Jul 11 04:08:26 03nail 07org.openembedded.nslu2-linux * rdbf7cecf... 10/packages/sane-backends/sane-backends_1.0.15.bb: Updated bb. (DaKa2) Jul 11 04:08:27 wohoo Jul 11 04:08:30 03nail 07org.openembedded.nslu2-linux * rc2343927... 10/packages/libgphoto2/libgphoto2_2.1.6.bb: Updated install part (DaKa2) Jul 11 04:08:34 03nail 07org.openembedded.nslu2-linux * rb4c635ab... 10/packages/ (3 files in 3 dirs): Bumped revisions Jul 11 04:08:36 no help.. life sucks Jul 11 04:08:38 03dyoung 07org.openembedded.nslu2-linux * r90985ad6... 10/packages/ctrlproxy/ctrlproxy_2.6.2.bb: ctrlproxy Jul 11 04:10:00 mickey|lunch: I think CIA stuck ;-) Jul 11 04:10:30 im gonna kill myself soon.. Jul 11 04:11:02 hey, cheer up man, there's more in life than a working bitbake ;) Jul 11 04:11:45 * NAiL closes two bugs Jul 11 04:11:55 very good Jul 11 04:11:56 j/k.. but this is really annoyinh Jul 11 04:11:57 g Jul 11 04:12:00 This is fun ;) Jul 11 04:12:07 heh, that i can understand dolpp Jul 11 04:12:09 I see that Uberhacker Lauer has been knocking them down at quite a rate as well. Jul 11 04:13:14 i guess i do need that bitbake installed if i wanna work with org.opemembedded.dreambox Jul 11 04:13:24 getting an usb-error: when i connect the gadget to host i get the error:usb 2-2: new full speed USB device using address 3 Jul 11 04:13:24 usb 2-2: device not accepting address 3, error -71 Jul 11 04:13:41 sorry for asking this offtopic question though Jul 11 04:13:52 your gadget driver is either missing or malfunctioning. Jul 11 04:16:56 03ccsmart 07org.openembedded.nslu2-linux * rd1dce877... 10/packages/cyrus-imapd/cyrus-imapd_2.2.12.bb: Removed snmp support. Jul 11 04:17:00 03ccsmart 07org.openembedded.nslu2-linux * r8b65cbd4... 10/packages/openslug-init/ (openslug-init-0.10/turnup openslug-init_0.10.bb): Make rmrecovery permanent active per default. Jul 11 04:17:04 03ccsmart 07org.openembedded.nslu2-linux * rb6beeff0... 10/packages/cyrus-sasl/cyrus-sasl_2.1.19.bb: Adding missing plugin libraries to package. Jul 11 04:17:08 03ccsmart 07org.openembedded.nslu2-linux * r6db0efe5... 10/packages/cyrus-sasl/cyrus-sasl_2.1.19.bb: Update staging dir for plugin libs. Jul 11 04:17:12 03jbowler 07org.openembedded.nslu2-linux * r0586ecf0... 10/packages/nfs-utils/ (files/nfsserver nfs-utils_1.0.6.bb): Jul 11 04:17:12 Working, tested, nfs-utils. This now pulls in all the dependencies on Jul 11 04:17:12 ipkg install and has a working startup script. Jul 11 04:17:16 03ccsmart 07org.openembedded.nslu2-linux * rb39c1942... 10/packages/ntp/ntp_4.2.0.bb: Upped rev to r3. Jul 11 04:17:19 03ccsmart 07org.openembedded.nslu2-linux * rcf55a1f4... 10/packages/openslug-init/openslug-init_0.10.bb: Upped rev. to r34. Jul 11 04:17:23 03ccsmart 07org.openembedded.nslu2-linux * r2900e539... 10/packages/cyrus-sasl/cyrus-sasl_2.1.19.bb: Upped rev. to r1. Jul 11 04:17:27 03nail 07org.openembedded.nslu2-linux * rc58b0633... 10/packages/meta/openslug-packages.bb: Cosmetic change to OPENSLUG_PACKAGES. Makes it easier to read Jul 11 04:17:30 03jbowler 07org.openembedded.nslu2-linux * re8103bb5... 10/packages/cyrus-imapd/ (cyrus-imapd-2.2.12/install-sh cyrus-imapd_2.2.12.bb): Jul 11 04:17:31 Build reliably against db 4.3 Jul 11 04:17:31 Fix the install so that the strip works Jul 11 04:17:34 03nail 07org.openembedded.nslu2-linux * rc51e54eb... 10/packages/ (5 files in 3 dirs): Added ISC BIND (DaKa2) Jul 11 04:17:38 03nail 07org.openembedded.nslu2-linux * r9beddf92... 10/packages/bind/bind_9.3.1.bb: Set bind MAINTAINER to nslu2-linux Jul 11 04:17:42 03nail 07org.openembedded.nslu2-linux * r32a88fee... 10/packages/base-passwd/base-passwd-3.5.9/ (openslug/root-home.patch root-home.patch): Back out ntpd/mysql users. Add them in the .bb instead Jul 11 04:17:44 pb_: gadget side i can assign address using ifconfig. ifconfig usb0 works for me. Jul 11 04:17:45 03nail 07org.openembedded.nslu2-linux * r823b0e21... 10/packages/ (mysql/mysql_4.1.12.bb openntpd/openntpd_3.7p1.bb): Make MySQL & openntpd create users on install Jul 11 04:17:49 03nail 07org.openembedded.nslu2-linux * r88420603... 10/packages/meta/openslug-packages.bb: added glib-2.0, gtk-doc, libxml, pcre Jul 11 04:17:53 03nail 07org.openembedded.nslu2-linux * r03d194f3... 10/packages/vlan/vlan_1.8.bb: Added vlan-package (DaKa2) Jul 11 04:17:57 03nail 07org.openembedded.nslu2-linux * r41be7c70... 10/packages/meta/openslug-packages.bb: Added vlan-package to openslug-packages Jul 11 04:18:00 03nail 07org.openembedded.nslu2-linux * r1388a66a... 10/packages/postfix/postfix_2.0.20.bb: Postfix Now adds user when installed Jul 11 04:18:04 03nail 07org.openembedded.nslu2-linux * re04c6d4a... 10/packages/linux/nslu2-kernel/2.6.12/defconfig: Add CONFIG_VLAN Jul 11 04:18:08 03jbowler 07org.openembedded.nslu2-linux * ree2ea524... 10/packages/bind/bind_9.3.1.bb: Jul 11 04:18:08 Fix build break - bind apparently requires openssl for the configure Jul 11 04:18:08 step to succeed, this breaks the osuclibc build (at least) because openssl Jul 11 04:18:08 is not built (before bind). Jul 11 04:18:12 03jbowler 07org.openembedded.nslu2-linux * re94fdb8f... 10/ (5 files in 4 dirs): Jul 11 04:18:12 Official openslug-kernel_2.6.12 Jul 11 04:18:13 This is simply a copy of the nslu2-kernel_2.6.12 with one (minor) change Jul 11 04:18:15 to the defconfig to include the ACM kernel module (lost in the nslu2-kernel Jul 11 04:18:17 defconfig). Jul 11 04:18:19 03jbowler 07org.openembedded.nslu2-linux * rab5262ca... 10/packages/cyrus-sasl/ (cyrus-sasl-2.1.19/berkdb.m4.patch cyrus-sasl_2.1.19.bb): Jul 11 04:18:22 This is an attempt to regularise the dependencies of cyrus-sasl, the Jul 11 04:18:24 package will now reliably use BerkeleyDB 4.3 (like cyrus-imapd and Jul 11 04:18:26 unlike the previous incarnation which failed to even look for it). Jul 11 04:18:28 It will also *not* link against libpam - this would happen before if Jul 11 04:18:30 libpam was built, but it is extremely unlikely at present that libpam Jul 11 04:18:32 works correctly on any OE system. Jul 11 04:18:34 03jbowler 07org.openembedded.nslu2-linux * r992fa47b... 10/packages/meta/openslug-packages.bb: Fix the uclibc build - 'bind' does not build against a uclibc libc. Jul 11 04:18:39 03nail 07org.openembedded.nslu2-linux * r49e8d85b... 10/ (5 files in 3 dirs): Fix samba LINUX_LFS problem in a proper way. FIXES #58 (OE) Jul 11 04:19:17 ok, now CIA should be done ;) Jul 11 04:19:41 heh Jul 11 04:25:11 dolpp: when you do type bitbake, which bitbake is returned ? Jul 11 04:26:45 check this Jul 11 04:26:45 Traceback (most recent call last): Jul 11 04:26:46 File "/home/dolphin/bitbake/bin/bitbake", line 27, in ? Jul 11 04:26:46 from bb import utils, data, parse, debug, event, fatal Jul 11 04:26:46 ImportError: cannot import name parse Jul 11 04:26:56 so its the right one Jul 11 04:28:41 dolpp: Ok, I'm just checking. In /home/home/dolphin/, what does svn diff -r 280 yield ? Jul 11 04:29:18 dolpp: /home/dolphin/bitbake rather Jul 11 04:30:02 it says nothing Jul 11 04:30:06 "svn diff -r 280" Jul 11 04:31:00 dolpp: So it is ok. Your install is good. Does it says the same if you invoke bin/bitbake from that dir ? Jul 11 04:35:34 zecke: have you started a new branch ? Jul 11 04:39:02 dolpp: If you have the same error, have a look at the owner and the rights to lib lib/bb lib/bb/parser.py. Jul 11 04:46:53 ade|desk: it was a mistake on his side. we will remove that branch when we upgrade to monotone 0.20 Jul 11 04:51:40 mickeyl: Read my message on the monotone 0.19+ ? Jul 11 04:52:04 don't think so Jul 11 04:52:09 please repeat Jul 11 04:52:51 mickeyl: I was saying that the monotone-0.19+ was rather a 0.20-. Interface changed. I don't now for the netsync but Interface cmd line is like 0.20. Jul 11 04:54:29 are there daily snapshots of the OE monotone repo somewhere? Jul 11 04:56:00 proti: ah bummer, that means we can't use it. Jul 11 04:56:08 i am havin an usb ethernet gadget error: i can do ifconfig -a and see usb0 interface in the device side. when i connect it to the host side. i get the error saying device is not accepting the address. i am not passing any params to the modules (g_ether). dont know if i have to ? Jul 11 04:56:15 or at least not until we upgraded Jul 11 04:56:23 giel: not yet Jul 11 04:56:40 mickeyl: No point to using it if we can use the 0.20. Jul 11 04:56:59 proti: why not? the speedup patches didn't make it into 0.20 Jul 11 04:57:09 at least not all Jul 11 04:57:25 mickeyl: Ah. I didn't know that. I thought that the patches were part of the 0.20. Jul 11 04:59:18 proti: yes says the same.. Jul 11 04:59:21 unfortunately not Jul 11 04:59:51 proti: and all files owned by root:root Jul 11 05:00:06 dolpp: are you root when executing ? Jul 11 05:00:20 yep Jul 11 05:00:22 btw Jul 11 05:00:40 theres no lib/bb/parser.py Jul 11 05:00:57 dolpp: No, there should be a lib/bb/parser/__init__.py Jul 11 05:01:02 proti: -r Jul 11 05:01:14 giel: ? Jul 11 05:01:20 proti: lib/bb/parse Jul 11 05:01:22 -rw-r--r-- 1 root root 2399 Jul 11 07:03 __init__.py Jul 11 05:02:35 dolpp: Can you check the accessed files/dirs : 'strace -e open ./bin/bitabke' Jul 11 05:03:27 s/bitabke/bitbake/ Jul 11 05:07:45 giel: Still confused. Jul 11 05:46:59 mickeyl: ping? Jul 11 05:47:26 mickeyl: re: http://bugs.openembedded.org/show_bug.cgi?id=116 Jul 11 05:47:33 I am able to reproduce that... Jul 11 05:49:49 good. then you might be able to debug what's going on there Jul 11 05:50:12 btw., --strip shouldn't be allowed as an argument to install Jul 11 05:50:30 either overwrite that or use our patches coreutils Jul 11 05:50:47 install --strip doesn't know about different architectures hence it destroys a binary Jul 11 05:51:01 jbowler came up with a workaround Jul 11 05:52:10 This hits a few packages I try to build.. Jul 11 05:53:54 Is there any way to portable strip a binary before packing it up? Jul 11 05:56:22 jbowler's solution includes having an install.sh in files/ and using that to make sure it is stripped by the right strip. Jul 11 05:58:53 our ipkg packaging class takes care about stripping it, so we don't need to strip earlier. Jul 11 05:59:16 this is also important to have one setting, i.e. INHIBIT_PACKAGE_STRIP (or so) Jul 11 05:59:22 ah, yes Jul 11 05:59:40 ok. So it should be patched in the procps makefile then Jul 11 06:09:47 hah Jul 11 06:09:48 users Jul 11 06:09:51 you got to love 'em Jul 11 06:10:03 they always have something to complain... Jul 11 06:10:05 it's ridicolous Jul 11 06:10:18 see http://www.oesf.org/forums/index.php?showtopic=13646&hl= Jul 11 06:10:31 we don't even have released anything and already they're complaining about the name(!) Jul 11 06:10:41 * mickeyl shakes head Jul 11 06:21:09 mickeyl: they are probably stupid enough that if you said you'd put code in to disable the distro on any machine owned by complainers Jul 11 06:21:21 mickeyl: they'd beleive it Jul 11 06:23:08 hehehe Jul 11 06:23:10 good idea :) Jul 11 06:24:00 no seriously, I just don't get it. I would _never_ complain about someone else chosing a name even _if_ I found it stupid. Hell, it's their decision, what makes me feel I have a voice here? Jul 11 06:24:02 tsss Jul 11 06:24:39 let alone publically stating it Jul 11 06:25:12 perhaps I'm too weak here but those things make me a bit angry Jul 11 06:25:13 OSS is just an easy target for the professional complainers Jul 11 06:25:37 *nod* Jul 11 06:26:31 *whine* Jul 11 06:27:08 * NAiL complains about ... bugs! Yes. Bugs! Jul 11 06:27:18 in general Jul 11 06:27:30 * XorA sometimes gets fedup of being treated like he gets paid to fiddle with software Jul 11 06:42:11 * giel agrees with you guys Jul 11 07:07:35 when i install inetutils package. it does install inetd.conf file into /etc; should we modify the inetutils.bb to make it to include inetutils. some files missing are /etc/services /etc/protocols; which is the best place to add. Jul 11 07:11:32 I think those two files should be in netbase. Jul 11 07:15:05 i am trying to minimise the boot time. if i include netbase,initscripts the boot time is more. Jul 11 07:15:33 why? Jul 11 07:15:39 I don't think netbase has any initscripts. Jul 11 07:16:11 oh, I guess it does the "ifup" stuff. Jul 11 07:16:28 initscripts itself is taking too much time. right now my boot time is 30 Seconds Jul 11 07:16:34 well, feel free to move the data files into a new netbase-data package, and make the regular netbase depend on it if you want. Jul 11 07:16:51 i am looking at something at around 7-9 seconds. Jul 11 07:17:30 I think that should be achievable with custom startup stuff. Jul 11 07:17:59 iirc, on our machines it takes about four seconds from power on to init running, and about another four to application startup. Jul 11 07:21:55 pb_: creation of device nodes too much. cant find how to limit the number of files in /dev Jul 11 07:22:55 eJumbo: if you don't have a requirement for hotplugging, I recommend you use a statically created /dev in your rootfs. Jul 11 07:23:15 if hotplugging is required, you will have to stick with udev and trim down its rules. Jul 11 07:24:19 i am not going to use hotplug facility. Jul 11 07:24:42 u mean to say create a package which statiscally creates nodes u want in the file system. Jul 11 07:27:07 yeah, or do it in the image construction step like the old zauruses do. Jul 11 07:27:24 see the makedevs invocation in image_ipk.bbclass Jul 11 07:29:28 is there no data files where i can edit saying these are the device files and thier major and minor numbers; file_table does not work for me. changed that file but it did not work for me Jul 11 07:30:13 yes; that's what device_table-initial.txt is for Jul 11 07:30:14 see the makedevs invocation in image_ipk.bbclass Jul 11 07:30:18 hi jamey Jul 11 07:32:48 hi pb_ Jul 11 07:39:47 i had a problem. i have enabled usb ethernet gadget in the kernel and got the kernel image as compiled by open embedded. when i connect it to the host i got an error saying the device has not accepted the address asssigned. but when i compile the same kernel source using the same .config file with cross tools. it works. Jul 11 07:40:21 i dont know if oe does something more when compared to cross tools which we cant see. Jul 11 07:40:29 what compiler versions are you using in each case? Jul 11 07:43:05 cross tools: arm-linux-gcc-3.4.1 and oe: 3.4.4 Jul 11 08:05:27 pb_:ping Jul 11 08:47:20 when i boot the image i get the error : can't access tty; job control turned off Jul 11 09:15:35 re Jul 11 09:21:51 zecke_: hail Jul 11 09:24:46 03zecke123 * r286 10bitbake/lib/bb/parse/__init__.py: Jul 11 09:24:47 Sorry for the inconvience and add the missing file Jul 11 09:24:47 I wonder what would happen if I only commit after getting Jul 11 09:24:47 enough donations for a file... Jul 11 09:24:47 Currently it defaults to the python implementation Jul 11 09:28:53 proti: could you confirm things work again? Jul 11 09:38:53 zecke_: It seems to be ok. bitbake runs and the expected packages. Jul 11 09:39:05 s/and/and build/ Jul 11 09:51:01 <_chronic> any nylon ppl here? Jul 11 09:53:31 didn't see any meshcube folks for quite a while Jul 11 09:53:37 ~seen br1 Jul 11 09:53:37 br1 <~br1@pD9E77FCC.dip.t-dialin.net> was last seen on IRC in channel #oe, 63d 1h 1m 52s ago, saying: 'when do i have to use "export VARIABLE = " in .bb files?'. Jul 11 09:54:43 <_chronic> mickeyl: ta for that. there is a workshop in london, uk this coming weekend, to structure a common dist around elements of nylon, familiar, and oleg's broadcom toolchain. (ie arm, mipsel & broadcom). Jul 11 09:55:13 <_chronic> i'm hoping to convince bruno to come .... Jul 11 09:56:07 that would be good. the last thing i heard from him was his monotone key for our new repository :) so he's still around somewhere Jul 11 09:56:32 <_chronic> i need to wrap my head around monotone - not done it yet.... :-/ Jul 11 09:56:48 if you've used bk in the past, it's pretty easy Jul 11 09:57:03 <_chronic> yeah, seen some of the wiki pages.... Jul 11 09:57:27 * _chronic is a bit of a version control weenie, though. Jul 11 09:58:02 <_chronic> rcs / cvs sure. bk just. monotone still to learn. Jul 11 10:04:11 <_chronic> are many #oe'ers going to 'what the hack?' i'd like to see a get-together. Jul 11 10:05:46 not me. have a pretty busy 2nd half of the year in front of me Jul 11 10:05:52 i hope to make next years FOSDEM Jul 11 10:06:01 but other than that, probably no conferences Jul 11 10:06:01 <_chronic> mickeyl: ok Jul 11 10:06:13 well, LinuxTag 2005 in Frankfurt Jul 11 10:06:20 <_chronic> mickeyl: heh Jul 11 10:06:24 but that's like 5 minutes travel from here Jul 11 10:06:24 heh Jul 11 10:07:12 03freyther 07org.openembedded.dev * rca92cc97... 10/packages/sqlite/ (4 files): Jul 11 10:07:12 sqlite3: Jul 11 10:07:12 -Update from version 3.2.1 to 3.2.2 Jul 11 10:07:16 03freyther 07org.openembedded.dev * r20220494... 10/packages/ (10 files in 8 dirs): Random syntax fixes. R_VALUE must be quoted Jul 11 10:07:28 _chronic: I'm considering going there Jul 11 10:07:46 mickeyl: -i s/LinuxTag/Linux World Expo/ ? Jul 11 10:07:52 zecke_: right Jul 11 10:08:33 mickeyl: the big Qtopia summit, hopefully they will discontinue Qtopia :} Jul 11 10:08:36 <_chronic> zecke_: cool. what dist takes your interest? Jul 11 10:08:55 ROAD dist Jul 11 10:09:29 zecke_: the amount of responses is a clear sign of how our hopes have been swamped away by how "cooperation" worked out in the last years Jul 11 10:09:32 mickeyl: do you remember if we ever considered a different init ystem? Jul 11 10:10:05 zecke_: yeah, we did. but none made it, because they were either too feature limited or unmaintained at that time. Jul 11 10:10:07 mickeyl: let us talk with the next Platform manager Jul 11 10:10:17 zecke_: yeah. who's in charge since Hogne left? Jul 11 10:10:37 mickeyl: We can replace init-scripts but what about all the others sysv5 init scripts? Jul 11 10:10:46 who should I report C3K problems to? Jul 11 10:10:49 mickeyl: ++Hogne Jul 11 10:11:12 Opie only starts in portrait mode with a newer OE image (but the widgets are larger, much better) Jul 11 10:11:20 also, GPE is starting up in the wrong orientation Jul 11 10:11:49 zecke_: no idea. i'd like to have an abstraction for init scripts. that would not only make switching init systems easier but also allow more polishing the startup messages Jul 11 10:12:18 JustinP: I didn't have time to test on c3k since long. atm., I'm concentrating on c7x0 again Jul 11 10:12:32 JustinP: sorry but patches welcome Jul 11 10:13:32 ok Jul 11 10:13:53 pointers on how to fix it? I'm willing to patch but I don't know my way around all of the source and libraries. Jul 11 10:14:09 check odevice Jul 11 10:14:12 check QWS_DISPLAY Jul 11 10:14:21 see messages on booting Jul 11 10:14:25 s/booting/launching/ Jul 11 10:14:25 I fixed the orientation thing for GPE by editing /etc/X11/Xserver, but now portrait is 180 degrees off Jul 11 10:14:28 qpe -nodaemon Jul 11 10:14:46 i think no one added CXK to gpe yet Jul 11 10:15:35 well, it seems to work fine except for the orientation is wrong. And as I said I got it to have the right orientation now but Portrait is 180 off. Jul 11 10:15:44 I believe there are patches hiding in some forum. but as usual people can't be bothered to post them to bugzilla Jul 11 10:17:05 "some forum" Jul 11 10:17:22 re Jul 11 10:17:26 I'd be happy to create patches from monotone (if that can be done, still learning) Jul 11 10:17:51 hi ! Jul 11 10:20:25 JustinP: yeah. "some forum". zaurus users in particular appear to believe that if they post bugs/fixes/patches in some random zaurus related forum, those will magically make it into OE Jul 11 10:20:51 mmm... Jul 11 10:21:11 heh. that goes along the same way as "i cooked an ipkg that fixes it. why didn't it make it into oz " Jul 11 10:21:26 workflow is something that has to be teached Jul 11 10:21:46 taught ;-) Jul 11 10:21:52 *nod* Jul 11 10:21:57 ~lart irregular verbs Jul 11 10:21:57 * ibot grabs a large, mis-shapened log, with squirrels, and beats irregular verbs until only the nuts remain ... which the squirrels run off with Jul 11 10:22:49 i would like to set up a OE build machine but i need an advice. I have to choose between reconfiguring all my computer and partitionning, or free a 10gigas partition. Would 10 gigas be enough ? can i build a Openzaurus image and some progs in 10gigas or do i need absolutely 30 gigas ? Jul 11 10:23:40 let me check a standard opie-imag Jul 11 10:24:32 opie image shouldn't take too much Jul 11 10:24:37 it's world that takes a lot Jul 11 10:24:58 alan: should work Jul 11 10:25:34 hmm even du takes pretty long here :) Jul 11 10:25:50 is it different for GPE ? i mean just GPE ? i just want to learn whatever i'll be able to understand, so i won't need world... Jul 11 10:26:02 makes hardly any difference Jul 11 10:26:11 ok. good. Jul 11 10:26:29 10G should be definitly enough to build opie and gpe Jul 11 10:26:57 ok. that's cool. thank you very much for your answer and for your hard work in general. Jul 11 10:27:38 may i ask what requires much more than 10 if not OPIE & GPE ? what else is there in world ? Jul 11 10:27:48 hah Jul 11 10:28:01 opie and gpe are less than 1000 packages Jul 11 10:28:06 we have around 6000 packages in world Jul 11 10:28:26 ho... Jul 11 10:28:31 people are using OpenEmbedded to build software for fileservers, gadgets, pdas, routers, set top boxes, and much more Jul 11 10:29:10 ho. OE turns out to be really powerfull. ^^' Jul 11 10:29:18 *nod* Jul 11 10:30:31 thanks again. bye Jul 11 10:33:02 03mickeyl 07org.openembedded.dev * r9d0d3641... 10/packages/ (3 files in 2 dirs): Jul 11 10:33:02 Python: Jul 11 10:33:02 - add python-pybluez 0.2 (Linux Bluetooth Bindings) Jul 11 10:33:02 - add pybluez to task packages Jul 11 10:33:02 - add homepage information to task packages Jul 11 10:41:16 what is a good way to create new .bb files? Jul 11 10:41:29 i mean to prevent bitbake to read 2500 files Jul 11 10:42:04 reading the bitbake manual, the fosdem presentation, and a couple of examples Jul 11 10:48:02 with 2500+ packages the cycles of editing / build become long. What can be done to create do this in short cycles. I add BBFILES manually now and just use around 100 packages. When I have a new dependency I scan the whole tree. Just want to check if this is the way to go. Jul 11 10:50:06 first time scan everything and if your package needs recompiling use -b to specify your bb file Jul 11 10:50:17 then you only compile youre specific package Jul 11 10:52:02 I'll try the -b, thanks Jul 11 11:05:27 epronk_: use BBMASK to mask out files, use the bbshell for rapid 'prototyping', use -b file.bb to just build one file Jul 11 11:21:37 bbshell? Jul 11 11:22:54 Hi folks, starting to dig into oe and have coulple of newbie questions. I'm building a bot, based on epia board. I've just built a "generic" with "epia" machine, but I would need a custom kernel and couple of additional packages. What would be the best way to do it, add everyting to the local.conf or create new distro and new arch ? Jul 11 11:23:50 definately don't make a new machine conf Jul 11 11:24:05 CosmicPenguin: why not? Jul 11 11:24:12 Because one already exists Jul 11 11:24:18 ah okay Jul 11 11:24:41 lazy_marmot: update epia.conf and put it into your local conf/ directory Jul 11 11:24:51 lazy_marmot: set your preferred provider for virtual/kernel Jul 11 11:25:07 lazy_marmot: for packaging create your own bbfile to create an image Jul 11 11:25:15 lazy_marmot: you can control what packages to install Jul 11 11:26:39 zecke_: And this a second question, how is to add new packages to either local.conf, or new bbfile ? Reading manual on bitbake clear a lot, but still I don't understand how to request some package to be build. Jul 11 11:27:55 You want to create your own bb to build the pacakges you need - see package/meta/ for examples Jul 11 11:29:32 lazy_marmot: new packages == new bbfile Jul 11 11:29:43 when there is a BROKEN = "1" in a package .inc file, that's no good? Jul 11 11:30:27 Delta_: it is broken in some way and exlcluded from the world build Jul 11 11:30:28 CosmicPenguin: already looking. One more question. In the generic distro, kernel itself was built, but was not packaged in the rootfs. Is this normal ? Or how should I build a bootloader pointing to my custom kernel? That would be the right question. Jul 11 11:33:09 ah, ok, I see, it's the libxfce4util, it's broken when compiling for arm..it doesn't install it's files into the lib and include dirs.. Jul 11 11:33:10 I wonder if the guys doing the linux port to Axim's use BB/OE at all Jul 11 11:34:33 lazy_marmot: the whole x86 bootloader thing is sort of vague at this point, since most bootloaders need the actual media to install themselves on Jul 11 11:35:16 lazy_marmot: the kernel binary itself is available in a package (kernel-image, I think, but I might be wrong), and you can install that in your rootfs Jul 11 11:36:31 But you'll eventually need to write a bootloader config file on your rootfs, and then play the usual tricks to install grub/lilo on your CF or hard drive Jul 11 11:36:44 Of course, if you use grub, you'll only need to do that once Jul 11 11:38:25 CosmicPengiun: ok, basically I managed to install manually lilo on my image with small voodo on disk image, I just thought it kind of could be done automagically. Thanks for advice, I will try the grub. Jul 11 11:41:12 CosmicPenguin: with kernel, it's a bit strange, while kernel itself was built, and modules built and deployed on rootfs, but kernel itself not. Strange, ok, thanks for advice anyway I'll look more into existing examples Jul 11 11:41:47 lazy_marmot: yes - the kernel binary is packaged seperately from the modules Jul 11 11:43:21 lazy_marmot: as far as automagically running lilo, I'm sure you could do a script to handle the manual stuff, but its a bit outside of the scope of OE to run such scripts within the image making process Jul 11 11:43:25 At least I think so Jul 11 11:47:31 CosmicPenguin: np, I've just seen that it builds ext2 image itself and thought may be I don't understand something. I'll do my script. Thanks for the answers! Jul 11 13:50:28 uuh.. Jul 11 13:52:22 bitbake just doesnt like me :) Jul 11 13:52:38 ImportError: No module named bb Jul 11 13:53:16 Jul 11 13:54:59 found this but no help.. http://www.oesf.org/forums/index.php?showtopic=12006 Jul 11 13:56:36 dolpp: did you try to strace ? Jul 11 13:57:27 no.. changed box.. Jul 11 13:57:47 Could you do it please, and pastebin the result ? Jul 11 13:57:50 now im not getting that same error.. only this "No module named bb" Jul 11 13:58:44 hmm.. how do i forward to stdout to a file Jul 11 13:58:59 what was the strace cmd btw.. strace -e open... something Jul 11 13:59:30 strace -e open bitbake Jul 11 13:59:55 It will show you the path of the accessed files. Maybe it can give some hints. Jul 11 14:01:04 export PYTHONPATH= Jul 11 14:02:38 03mickeyl 07org.openembedded.dev * r0aa80132... 10/packages/python/python_2.4.1.bb: Python-2.4.1: DEPENDS on readline to get lib-dynload/readline.so get built Jul 11 14:02:51 mickeyl: Should be take care of by the sys.path.append, shouldn't it ? Jul 11 14:03:36 tried: Jul 11 14:03:54 export PYTHONPATH="/home/dolphin/oe/bitbake/lib/bb" Jul 11 14:04:00 no, that's wrong Jul 11 14:04:06 i said PYTHONPATH= Jul 11 14:04:14 bb lives not in bb Jul 11 14:04:18 bb lives in lib Jul 11 14:04:22 whats the bb module Jul 11 14:04:26 oki lemme see Jul 11 14:05:13 ERROR: Unable to open conf/bitbake.conf Jul 11 14:05:20 seems better Jul 11 14:05:24 Good! Jul 11 14:05:37 halleluujaa Jul 11 14:05:38 here we are Jul 11 14:06:35 we have to look at that sys.append Jul 11 14:07:59 hmm... the way I read it, if bitbake is in your path, then the argv[0] won't have the right path info Jul 11 14:08:17 CosmicPenguin: ya, that argv[0] sounds bogus Jul 11 14:08:21 why cant it open bitbake.conf btw :D Jul 11 14:08:23 better use __file__ Jul 11 14:08:24 Of course, if one installs bitbake in a binary directory, it would follow that lib/bb would be put in the appropriate python directory Jul 11 14:08:47 dolpp: now that you got the binary running, time to read GettingStarted again Jul 11 14:08:56 dolpp: Set BBPATH correctly. Jul 11 14:09:49 yeah got it.. sweet Jul 11 14:10:29 now i need some bash script for those export lines Jul 11 14:18:06 works like a charm Jul 11 14:21:25 does anyone know why the firefox ipkg doesn't have a *binary*? Jul 11 14:21:31 it's impossible to use the package without the program Jul 11 14:30:42 CosmicPenguin: what happend to pub/bootmenu? Jul 11 14:31:35 well no one seems to care. I've asked in many forumes beefore now. It's been broken ever since I started playing with GPE more than a year ago (I think) Jul 11 14:32:54 zecke_: I killed it by apathy/accicent Jul 11 14:32:59 I'll put int on another domain Jul 11 14:33:18 CosmicPenguin: vs 0.6 and 0.7 are missing Jul 11 14:34:00 * CosmicPenguin boggles Jul 11 14:36:02 zecke_: fixed Jul 11 14:36:45 thanks Jul 11 14:40:35 is xfce4 the smalles windowmanager in oe? Jul 11 14:40:49 matchbox is probably a fair bit smaller Jul 11 14:41:10 ah, ok I take a look at it.. Jul 11 14:41:37 JustinP: http://hyboria.mine.nu/~dp/feed/ grab the firefox from there. The problem with firefox is its so large it seems to break ipkg on some machines. Jul 11 14:41:59 cute :) Jul 11 14:42:08 xfce is big because of the dependancies Jul 11 14:42:59 XorA|gonehome: thanks, I'll try that one. Any idea what "breaks" ipkg? I would think my machine could handle it. Jul 11 14:43:24 XorA|gonehome: also, who's feed si that? Will it work correctly with the newer OZ releases? Jul 11 14:43:45 JustinP: well on mine tar puts in flags that ipkg doesnt understand, I did report to ipkg guys and they couldnt figure it. Youll probably get a couple of warnings with my one, but it does work Jul 11 14:43:54 JustinP: its my feed Jul 11 14:44:12 JustinP: fresh off my build machine via ADSL Jul 11 14:44:13 OMG! i got it to work Jul 11 14:44:24 Handling BitBake files... Jul 11 14:44:28 I tried the xfce-image metapackage, I had to put in the fonts by hand, but now it just seg faults when I try to run xfwm4.. Jul 11 14:44:57 2569 bitbake files.. thats a lot.. wonder if all of those are used Jul 11 14:45:28 Delta_: what porcessor? Jul 11 14:45:33 processor even Jul 11 14:45:36 pxa270 Jul 11 14:46:18 I should install strace, maybe that tells me more.. Jul 11 14:46:39 XorA: I don't see firefox in that feed Jul 11 14:47:17 JustinP: crap ask me tomorrow Jul 11 14:47:47 JustinP: I need sleep Jul 11 14:48:12 ok Jul 11 14:48:13 JustinP: forgot I wiped it for a fresh install couple of days ago, havent built everything yet Jul 11 14:48:31 perhaps I'll try playing with it myself.... Jul 11 14:50:28 XorA: indeed, that data.tar.gz in the ipk has a firefox-bin....interesting... Jul 11 14:51:56 XorA: I'll try the real tar instead of busybox and see if that helps Jul 11 14:54:11 * mickeyl sleeping Jul 11 14:54:15 g'night Jul 11 14:55:38 'night mickey|zzZZzz Jul 11 15:00:56 does anyone know which variable is passed to --prefix in a do_configure? Jul 11 15:01:42 tried the STAGING_* but then the products are not packaged Jul 11 15:02:25 epronk_: you normally don't need to mess with any of that, it usually takes care of itself Jul 11 15:03:14 true, for autotools project with a ./configure Jul 11 15:03:25 it's about CMake Jul 11 15:04:04 they have a ./bootstrap script and i want to pass the parameters manually Jul 11 15:05:39 There are two usages of the prefix - 1) where previously build libraries live, and 2) where the built libraries should go Jul 11 15:05:54 in OE, these are not always the same thing Jul 11 15:08:04 epronk_: read autotools.bbclass Jul 11 15:09:32 'night all Jul 11 15:22:47 ok, i'll try this --prefix stuff tomorrow, cya Jul 11 15:47:40 ok, who do I talk to about ipkg problems. As in ipkg the command. Jul 11 15:48:42 There's a few ipkg guys that hang out in here, I guess Jul 11 15:49:09 well, my previous problem with FireFox isn't the ipk. It's ipkg. Jul 11 15:49:19 If I extract the files everything is there Jul 11 15:49:28 but some of the files are "lost" when ipkg installs Jul 11 15:49:46 they're in the /usr/lib/ipkg/lists/firefox.list file but not ont he filesystem Jul 11 15:50:47 I get some kind fo error from ipkg when I try to install as well Jul 11 15:50:49 one sec... Jul 11 15:51:31 I'll get them again Jul 11 15:54:45 * JustinP waits for ipkg.... Jul 11 15:57:48 I'm seeing 3 error messages Jul 11 15:57:57 well, 2, but one repeats Jul 11 15:57:58 ipkg: Unknown typeflag: 0x4c: Illegal seek Jul 11 15:57:58 ipkg: Unknown typeflag: 0x4c: Illegal seek Jul 11 15:57:58 ipkg: Don't know how to handle /./@LongLink: No such file or directory Jul 11 15:58:21 Before this is the Installing line and after is the COnfiguring line Jul 11 15:58:28 anyone have any clue? Jul 11 15:59:01 Any idea what platform your packages were created on? Jul 11 15:59:25 Sort of sounds like a 64bit->32bit issue to me - somebody's structs are messed up Jul 11 16:00:23 possible....but not likely IMHO Jul 11 16:00:27 I'm on my own system Jul 11 16:00:31 an older Ahtlon processor Jul 11 16:00:34 *not* 64 bit Jul 11 16:00:45 and the target is a Sharp C3000 Jul 11 16:00:59 the same thing has happened with all official FireFox builds as well Jul 11 16:01:06 on both my SL-5500 and C3000 Jul 11 16:01:17 these seem to be errors from libbb/unarchive.c Jul 11 16:01:29 oh, and as I said, I extracted the ipk and the files are in there Jul 11 16:01:38 ar x firefox-...ipk Jul 11 16:01:43 tar zxvf data.tar.gz Jul 11 16:01:47 and the files are all in there Jul 11 16:01:51 Right, but if the header was screwed up, all bets are off Jul 11 16:02:16 well...on *my Zaurus* I can extract them like that and copy them and FireFox now runs Jul 11 16:04:22 so either ipkg is handling the tarball incorrectly or "tar" is more fault tolerant than ipkg Jul 11 16:05:04 I see no errors when I untar the archive Jul 11 16:11:45 Its not an error in the archive Jul 11 16:11:49 thats what I keep telling you Jul 11 16:15:25 ok, so you agree it must be a problem in ipkg's archive handling Jul 11 16:15:52 more likely ipkg is getting confused with the headers Jul 11 16:16:20 There might be a odd patch its going down in the archive code, but for the most part, that code is pretty well excercised Jul 11 16:17:33 well....as far as I can see tar handles it fine and I'm inclined to trust tar Jul 11 16:19:53 ipk =! tar Jul 11 16:19:55 !=* Jul 11 16:20:03 tar handles tar files Jul 11 16:20:07 ipkg handles ipk Jul 11 16:21:55 I understand that Jul 11 16:22:09 however, ipk is an "ar" which includes in side of it a "tar/gzip" Jul 11 16:22:21 well, 2 tar/gzip files Jul 11 16:29:56 Well, you can spend your time looking through the archive code if you want, but its probably a dead end - you're looking at a classic case of a bad header Jul 11 16:30:40 if you lseek to the wrong place for the extraction, then for sure the ar is going to be corrupted Jul 11 16:30:46 ok....so whose fault is it then? What needs to fix this? Jul 11 16:31:03 if it's not the ipkg client code it must be the ipkg archiving code Jul 11 16:31:29 but I have a hard time believing it's not the ipkg installer code as "tar" can still unarchive the same files with ease Jul 11 16:31:48 unless, as I said, it's more fault tolerant somehow... Jul 11 16:36:36 Well, I'm not going to argue it with you - I told you want I thought and I stand by that Jul 11 16:36:59 A quick search of the bugzilla shows that nobody has reported a similar problem, so its possible you have found a new bug Jul 11 16:37:14 I'm not sure what you're standing by. You said it's not a problem with the archive and then you said it's not a problem with ipkg. So where is the fault? Jul 11 16:37:23 but more likely, you should prepare yourself for a storm of "Cannot reproduce" Jul 11 16:37:37 no, XorA in the GPE channel said that he saw this a *long* time ago. Jul 11 16:37:56 and, as I previously said, this has been an issue in OpenZaurus with the FireFox package for more than a year now Jul 11 16:38:56 But in the last year, dozens of people have built and installed firefox ipkgs on a wide variety of platforms Jul 11 16:39:05 and used them? Jul 11 16:39:14 so if there was something strange about firefox+ipkg, I feel fairly confident they would have been flushed out by now Jul 11 16:39:17 which platforms? any Zaurii? Arm? PXA? Jul 11 16:39:41 *but* if there was something wrong with the way that your ipkg was created, well then - that has nothing to do with the ipkg program, and everything to do with the environment in which it was created Jul 11 16:39:44 well I can assure you that this is 100% reproducable in both a Zaurus SL5500 and a Zaurus C3000 Jul 11 16:39:49 And I believe thats where your problem is Jul 11 16:40:10 ok.....so how do you explain "tar" and "ar" unarchiving the exact same file correctly? Jul 11 16:41:05 I can give you the ipk if you like. Jul 11 16:43:23 it could very well be the packaging, but AFAICS opembedded is using ipkg-build, as it should be. And it looks like ipkg-build is using "tar" and "ar". And, as "tar" and "ar" on both the target and host machines can extract correctly I must assume that it is *ipkg*'s archive code which is failing here. Jul 11 16:48:41 Well then, post a bug and hope for the best - if you're looking for sympathy you won't find it here - I regularly install many different ipkgs on a daily basis, including Firefox, and the only time I've ever seen your error is when I had a 64bit->32bit issue Jul 11 16:48:53 Apparently your millage does vary Jul 11 16:48:57 Good luck Jul 11 16:58:38 CP|AtHome: ok, bug filed: http://bugs.openembedded.org/show_bug.cgi?id=143 Jul 11 17:15:44 is there an easy way to update a feed? I have *lots* of packages in my directory and would like to "sync" the Packages file Jul 11 17:15:50 also, is there an easy way to remove duplicates from a feed? Jul 11 17:28:10 kergoth: thanks much for fixing mpd Jul 11 17:28:34 I have a new package bb file (mpc). It was very easy to add. How should I submit it? Jul 11 17:28:43 send it to the mailing list Jul 11 17:28:46 * kergoth hasnt done anything in oe in months Jul 11 17:30:16 then perhaps it was a general fix Jul 11 17:30:19 it works anyway Jul 11 17:30:47 kergoth, how are you doing? Jul 11 17:49:38 03nail 07org.openembedded.nslu2-linux * ree06322d... 10/packages/linux/openslug-kernel-2.6.12/defconfig: Added CONFIG_VLAN_8021Q, required for vlan package Jul 11 17:49:42 03rwhitby 07org.openembedded.nslu2-linux * rfa9875cd... 10/packages/linux/openslug-kernel_2.6.12.2.bb: Reinstated openslug-kernel 2.6.12.2 Jul 11 17:49:46 03rwhitby 07org.openembedded.nslu2-linux * rb44bd1f0... 10/packages/linux/openslug-kernel-2.6.12.2/ (README defconfig): Reinstated openslug-kernel 2.6.12.2 defconfig and README Jul 11 17:49:49 03rwhitby 07org.openembedded.nslu2-linux * r74309263... 10/packages/ (ixp425-eth/ixp425-eth_1.1.bb ixp4xx/ixp4xx-csr_1.4.bb): Bumped the PR on the ixp modules, becuase we changed the OpenSlug default kernel version. Jul 11 17:54:04 while trying to compile xserver-kdrive_20050624, I'm getting "syntax error before "ShmFuncsPtr"" included from fb/fboverlay.c. Any ideas? Jul 11 17:55:28 03nail 07org.openembedded.nslu2-linux * rdd3df045... 10/packages/xinetd/ (xinetd/xinetd.conf xinetd/xinetd.init xinetd_2.3.13.bb): Added xinetd package (DaKa2) Jul 11 19:14:16 <[g2]> anyone running 2.6.12.x on ARM/Xscale Jul 11 19:30:07 I built gpe-image for familiar-0.8.2, it boots, does the calibratiom, but then it gets stuck in a loop with just the X server's "X" on the screen, and some text flashing that it's loading something I believe. It sounds like GPE isn't starting. Is this a known problem? What's the fix? Jul 11 19:30:24 Btw, I ran into this: http://www.handhelds.org/hypermail/gpe/current/3474.html . Is that the fix people use? Jul 11 19:30:57 (I'd like to be able to build the same gpe-image for familiar-0.8.2 that they distribute) Jul 11 19:31:03 Theirs works. Jul 11 19:34:18 for nearly a week :( NOTE: package linux-openzaurus-2.6.12-mm1-r4: task do_patch: failed Jul 11 19:52:16 The text that flashes is "Please wait: booting...". Jul 11 20:11:31 so um, I'm sure you all get this a lot, but.... my bitkeeper stopped working and says "No free bk after July 1, 2005", so what do I do? Jul 11 20:11:49 pay USD$5 per year? Jul 11 20:11:58 USD$5K per year that is Jul 11 20:12:43 (that's if you want to keep using BitKeeper of course) Jul 11 20:12:57 I want to keep using OE, so what do I do? Jul 11 20:12:57 otherwise you can switch to Monotone like everyone else :-) Jul 11 20:13:04 thanks Jul 11 20:13:09 is there a FAQ somewhere? Jul 11 20:13:23 there was a posting to the oe mailing list I believe Jul 11 20:13:57 and it is only on the front page of www.openembedded.org .... Jul 11 20:14:21 where were you looking? Jul 11 20:15:23 I just dropped in and checked second :-) Jul 11 20:19:47 rwhitby: thanks for the help ;-) Jul 11 20:19:54 np Jul 11 22:08:17 Hi Jul 11 22:26:16 03rwhitby 07org.openembedded.nslu2-linux * r69eb896c... 10/packages/mgetty/mgetty_1.1.30.bb: Debian mgetty-1.1.30 source has disappeared, and upgrading is too hard for a 5 minute job, so nslu2-linux is now mirroring the source. Jul 11 22:26:20 03rwhitby 07org.openembedded.nslu2-linux * r28471e00... 10/packages/meta/openslug-packages.bb: Added mgetty back into openslug-packages Jul 11 23:22:40 having a problem creating a directory in /dev/pts ? Jul 11 23:23:09 i did create it. but did not get reflected in the file system. Jul 11 23:23:20 where should i do it ? **** ENDING LOGGING AT Mon Jul 11 23:59:57 2005