**** BEGIN LOGGING AT Mon Aug 11 02:59:57 2008 Aug 11 07:18:21 good morning Aug 11 08:32:14 bonjour Aug 11 08:34:29 hi Aug 11 08:37:46 yo woglinde :) Aug 11 08:38:02 hi Aug 11 08:38:14 he genesis are the robots still fine? Aug 11 08:38:18 hi diego Aug 11 08:40:04 yeap Aug 11 08:40:46 better than amethyst :x Aug 11 08:40:50 hehe Aug 11 08:41:13 hm this is a nice board http://www.linuxdevices.com/news/NS4190784713.html Aug 11 08:41:30 * woglinde waits for qt checkout Aug 11 08:41:44 the doc dir takes ages Aug 11 09:16:20 good morning Aug 11 09:26:48 hi all Aug 11 09:27:00 I'd like to enable shadow password on oe Aug 11 09:27:00 hi florian Aug 11 09:27:06 hi unclejack Aug 11 09:27:13 why? Aug 11 09:28:05 trying to play with the security model Aug 11 09:28:42 I need to know if this has been already implemented on oe Aug 11 09:28:48 hm then you have to configure glibc diffrent if I remeber correctly Aug 11 09:28:53 no Aug 11 09:29:07 security isnt a high goal for now Aug 11 09:29:32 yeah, someone told me a million times, I think Aug 11 09:31:11 What I don't get is why we are using such old technologies ... I refer to simply passwd files with DES encryption Aug 11 09:31:34 I understand that developers are still programming assuming full privileges Aug 11 10:28:39 morning all Aug 11 10:51:10 mickeyl: good morning Aug 11 10:51:49 morning pb__ Aug 11 10:54:03 florian: good morning Aug 11 10:54:41 Unclejack: you shouldn't need to do anything special to enable shadow passwords, glibc should support them right out of the box. Aug 11 10:54:43 hi pb__ Aug 11 10:55:01 pb__: I sent an answer to Benno... Aug 11 10:55:06 florian: ah, very good Aug 11 10:56:12 florian: he seems to be scrutinising all the gpe i18n stuff. I had another pile of emails from him yesterday. Aug 11 10:56:57 pb__: oops.. so I should take a look at the remaining components and make sure they get updated. Aug 11 10:58:06 florian: I just saw your mail. did you really mean to cc ldecicco? Aug 11 10:58:38 pb__: argh Aug 11 10:58:45 no Aug 11 10:59:29 * florian should get some more coffee and a little bit of food Aug 11 10:59:34 :-} Aug 11 11:15:54 thank you /pb__ Aug 11 11:16:38 so I guess I need to reconf busybox and do some magic on /etc/shadow Aug 11 11:17:44 yes, probably. Aug 11 11:17:58 I think "shadowconfig on" should be about the extent of the magic you need. Aug 11 11:18:25 if you're using busybox then you will (obviously) indeed need to make sure that it's configured to support shadow. Aug 11 11:22:45 have to watch out for the root password squash stuff in OE as well Aug 11 11:22:53 make sure it squashes the right file Aug 11 11:25:45 yep Aug 11 11:26:06 I think you're talkin' about base-passwd package Aug 11 11:27:02 XorA: probably best to run shadowconfig after zap_root_password so that you only have to do the zapping in one place Aug 11 11:28:09 yeah, I couldnt remeber the proper name for that bit Aug 11 11:30:53 re Aug 11 14:35:03 re Aug 11 14:36:20 thesing: ping Aug 11 14:38:11 ant_office: pong Aug 11 14:38:21 hi Thomas Aug 11 14:38:33 I was testing the kexec bootmenu Aug 11 14:39:07 but I have one problem: I boot from /mtdblock2... Aug 11 14:39:23 the culprit is probably my commandline: cat /proc/cmdline -> console=ttyS0,115200 console=tty1 root=/dev/mtdblock2 rootfstype=jffs2 mtdparts=sharpsl-nand:7168k(smf),98304k(root),-(home) Aug 11 14:40:25 lets start from the beginning. How far did it work and what didn't ? Aug 11 14:40:40 it boots (I see openhanded splash) Aug 11 14:40:53 but then it finds root on mtd2 Aug 11 14:41:20 I mean, the cmdline (from u-boot) should be 'filtered' to intercept root= situations Aug 11 14:41:28 or not? Aug 11 14:41:57 so you see the bootmenu and can choose one entry but the selected kernel uses the wrong root device? Aug 11 14:42:08 no, it just skips your init Aug 11 14:42:15 and boots from flash Aug 11 14:42:20 the rootfs Aug 11 14:43:59 (I deployed an image with 2.6.26 on mtd2) Aug 11 14:44:39 the problem seems obvious: root=/dev/mtdblock2 Aug 11 14:45:05 and missing noinitrd ? Aug 11 14:51:35 ant_office: is initramfs-support enabled? the cmdline shouldn't matter if initramfs is used Aug 11 14:52:12 hm...I'll double check defconfig then Aug 11 14:52:36 thesing: Can I kexec-boot all kernels? Even 2.4? Aug 11 14:52:48 boot 2.4 from 2.6? Aug 11 14:53:26 Laibsch: yes. I never tested it but it should work. Aug 11 14:53:35 cool Aug 11 14:54:01 last time we talked, IIRC my collie hung Aug 11 14:54:13 I sent you a serial console log Aug 11 14:54:19 Any progress since then? Aug 11 14:54:43 ant_office: you need to have CONFIG_BLK_DEV_INITRD=y and CONFIG_INITRAMFS_SOURCE="initramfs.cpio.gz" Aug 11 14:55:05 yes the images start completely Aug 11 14:55:27 but touchscreen doesn't work for some reason. Aug 11 14:55:29 hey Aug 11 14:55:29 thesing: the size was bigger than the normal 2.6.26 so I'd suppose cpio.gz was included Aug 11 14:56:49 i need some help building libobjc.so, anyone help? Aug 11 14:58:58 gcc doesn't build it by default :/ Aug 11 15:01:13 mattface: actually, looking at my build tree, I think it is getting built but just not packaged correctly. Aug 11 15:01:47 might be as simple as adding an appropriate packages and FILES_.. to gcc-package.inc. Aug 11 15:03:24 worth a try Aug 11 15:03:30 gcc-package-cross.inc ? Aug 11 15:04:17 bbl Aug 11 15:04:18 er, yeah, that one Aug 11 15:04:31 you should be able to follow the example of libstdc++ Aug 11 15:06:56 oh, you'll also need to add some strategic 'mv' commands to do_install in the same file, again following the same procedure as for libstdc++ Aug 11 15:09:14 thesing when did the fix hit .dev? Aug 11 15:10:39 is the website down for everyone? Aug 11 15:10:55 bin1010: yes, see the topic Aug 11 15:29:51 I'm looking for the best way to put together a recipe that takes a kernel and initrd which are built from separate recipes and uses the kernel's Makefiles to create a powerpc cuImage.initrd (which is a single file containing kernel, initrd, and a bootwrapper compatible with older bootloaders) - the easiest way would be to have my 'bootstrap' recipe work in the same WORKDIR as the kernel recipe - is it possible to share WORDIR's Aug 11 15:29:51 across recipes? Aug 11 15:30:46 tharvey: that sounds similar to thesing's work for the collie Aug 11 15:31:00 it's certainly possible to share WORKDIRs, but I don't think it is a good idea Aug 11 15:31:09 it would probably be better for you to install the common bits into staging somewhere Aug 11 15:31:13 XorA, collie? Aug 11 15:31:31 tharvey: one of the zaurus machines, he has a kernel+initrd combo that works for that machine in OE Aug 11 15:32:06 pb__, thats what I was thinking as well... in this case 'vmlinux' may be all thats needed but I would really also want to make sure that my bootstrap recipe uses the same kernel version as the virtual/kernel - guess I need to do that manually? Aug 11 15:33:35 tharvey: yeah. you can find the staged kernel version in STAGING_DIR/kernel/kernel-abiversion or some such. Aug 11 15:35:17 pb__, ah... didn't even think of looking at what was in staging... vmlinux is already there as well as info about kernel version - looks like that may be all I need. I guess I'll have to have my recipe extract yet another kernel tree to do its work with Aug 11 15:36:32 XorA, thanks - I'll look at collie Aug 11 15:36:52 linux-kexecboot is I think the recipe in packages/linux Aug 11 15:37:10 have to track what uses that to make the image Aug 11 15:37:28 XorA, thanks - that saved me a lot of time Aug 11 15:37:58 there are so many recipes... hard to figure out what is used for what, whats obsolete etc Aug 11 15:41:18 pb__, if I 'did' want to share a workdir, would I make one recipe export its workdir so that the other knows where it lives? Aug 11 15:42:10 Laibsch: some days ago. Aug 11 15:42:14 tharvey: yeah Aug 11 15:42:32 tharvey: but, obviously, you would then lose in a pretty bad way if someone tried to build with rm_work enabled Aug 11 15:44:51 pb__, understood thx Aug 11 15:45:21 tharvey: if you have questions about the stuff I did, feel free to ask me. Aug 11 15:49:18 thesing, thanks... I'm looking for the recipes now - could you point me to them? I assume you have a specific kernel recipe and a specific initrd/initramfs image recpie, then a third recipe that builds a kernel+initrd image? Aug 11 15:52:57 tharvey: look in packages/kexecboot Aug 11 15:53:12 thesing: hi, can you help me a bit with git on linuxtogo? Aug 11 15:53:34 tharvey: I have linux-kexecboot which is a kernel that has an initramfs build in. Aug 11 15:53:50 tharvey: the image that it uses is specified in the recipe Aug 11 15:54:02 mr_nice: sure. What do you want do know? Aug 11 15:55:13 thesing: I just created a project at linuxtogo for the simpad and I want to use git for it. I saw that you are also using linuxtogo for you git. Aug 11 15:55:41 thesing: I just can't figure out how to get git. I registerd with svn Aug 11 15:55:58 Laibsch: I gave you admin status for collie-kernel. What do you want to do with it? Aug 11 15:56:01 probably obvious, but i will ask anyways, so BBFILES is like a search path? Aug 11 15:56:10 thesing: nothing ;-) Aug 11 15:56:16 nichts Konkretes Aug 11 15:56:53 mr_nice: you need to login at projectname.linuxtogo.org (ssh) and create a repository. Aug 11 15:57:35 mr_nice: if you want it for your whole project create it in the project folder and give your group permissions to write. Aug 11 15:58:13 mr_nice: then you need to create a magic file in the repository to tell the git daemon to pick it up. Aug 11 15:59:03 mr_nice: the file is git-daemon-export-ok Aug 11 15:59:41 thesing, packages/kexecboot isn't in my tree - is it something new in org.openembedded.dev? Aug 11 16:00:08 tharvey: its only in .dev what tree are you using? Aug 11 16:01:22 I'm using dev... haven't pulled in a month or so though, pulling now Aug 11 16:01:50 thesing: thanks I will try that :) Aug 11 16:01:52 tharvey: pull from opensource.wolfsonmicro.com Aug 11 17:05:05 thesing, I'm not all that familiar with kexec - is the idea that your initial kernel+initrd uses kexec to not only bootstrap to a new rootfs but to a new kernel as well? Aug 11 17:23:32 tharvey: yes exactly. Aug 11 19:54:00 oe site dead again? Aug 11 19:54:08 I can't do an update Aug 11 20:01:35 mewyn: yes, see the topic Aug 11 20:02:13 he pb Aug 11 20:02:38 hi woglinde Aug 11 20:02:50 Durr :) Aug 11 21:59:56 Hi Jay7 Aug 11 22:48:01 thesing|away: here is the log of kexecboot on c7x0. It somehow works, but... Aug 11 22:48:07 thesing|away: http://www.pastebin.ca/1166928 Aug 11 22:48:40 thesing|away: as you see the mtdparts are lost after kexec... Aug 11 22:49:28 thesing|away: second small problem: the screen is rotated 90?CW on clamshell-Zaurus Aug 11 22:55:17 is there a recipe for getting the c++ libraries added? Aug 11 22:57:30 wb Aug 11 22:58:45 ant_: does uboot set the commandline? for collie the commandline it compiled into the kernel. Aug 11 22:59:09 yes. the cmdline is set in u-boot Aug 11 23:00:55 it seems there is an atag list somewhere in ram Aug 11 23:02:21 containing cmdline but dunno about sanity (seems 'garbled' and there are indeed unrecognized tags) Aug 11 23:03:17 I have cmdline in a file for collie. kexecboot uses this cmdline and only adds the root= stuff. Aug 11 23:03:33 we should grab the mtdparts string Aug 11 23:03:58 and add it to cmdline Aug 11 23:05:05 thesing: I used u-boot cause I'm lazy and kernel was way too big (1441720) Aug 11 23:05:25 but u-boot env can be read with u-boot-utils Aug 11 23:05:26 thesing, I'm curious as to why you needed to have your initramfs kexec to a new kernel - what made you require a 2stage boot and why couldn't you reuse the kernel? Aug 11 23:05:42 (for collie) Aug 11 23:05:51 I'm testing your work Aug 11 23:06:05 just to see how a menu can be added to u-boot Aug 11 23:07:02 tharvey: you only have 1mb flash for the kernel for collie. It uses some non-free sharp bootloader so we can't change that and there is no other bootloader for collie. Aug 11 23:07:51 I suppose the kernel mantainer could easily create the needed machine.conf for u-boot Aug 11 23:07:54 thesing, ah... ok Aug 11 23:08:13 tharvey: now kernels can be as big as they want and as a bonus we can upgrade them with package mgmt and multibooting of distros is possible Aug 11 23:08:25 thesing, dang I'm impressed that you were able to get the kernel 'and' initramfs down to 1MB Aug 11 23:09:19 tharvey: klibc makes it possible. The whole thing is 700K now and could be made a bit smaller if necessary. Aug 11 23:10:33 ant_: it't not this easy. One would have to write drivers for too many stuff (SD, lcd, pcmcia) Aug 11 23:11:24 thesing, oic - I didn't realize that klibc had an sh - thats cool Aug 11 23:11:25 ant_: collie uses two custom chips so we couldn't reuse much. Aug 11 23:12:19 tharvey: you don't have regular expressions in klibc so you don't have expr, sed and friends. Aug 11 23:12:53 tharvey: that't why I wrote kexecboot in C and didn't take the existing shell implementation. Aug 11 23:16:34 thesing, ah... I see your init.sh kicks off kexecboot which must do the majority of the work Aug 11 23:22:21 Anybody else ever dropped their spitz from theird desk? Aug 11 23:22:29 Mine apparently did not like it that much Aug 11 23:22:46 I hardly get any reaction from it now Aug 11 23:23:08 The flash screen shortly comes alive for the split of a second, but then goes black again Aug 11 23:23:21 I don't think I'm seeing anything on the serial console Aug 11 23:23:36 XorA: Maybe you have an idea? Aug 11 23:23:41 HD dea? Aug 11 23:23:43 dead Aug 11 23:24:01 Shouldn't I be seeing at least something on the serial console? Aug 11 23:24:08 Laibsch: never managed to break mine Aug 11 23:24:20 * XorA should be asleep Aug 11 23:24:21 You didn't try hard enough ;-) Aug 11 23:24:32 good night Aug 11 23:24:59 Laibsch: did you try the diag menu? Aug 11 23:25:28 That is the one I am talking about, I think Aug 11 23:25:38 OK button and on/off Aug 11 23:25:45 Where you flash the device Aug 11 23:26:03 thesing: BTW, the collie still hangs at the same place, I think: http://rafb.net/p/lmMjjr69.html Aug 11 23:26:25 Did I need to replace the stuff in flash as well? Aug 11 23:26:34 I only updated what is on the SD card Aug 11 23:27:15 the problem was the flash part. Aug 11 23:27:23 is there any way to override the files that a recipe uses from a different path? For example, I have a 'local' tree and an 'oe' tree where local is my changes/additions to the untouched oe.dev tree. I wish to use a different file for the 'interfaces' that the netbase package installs without modifying oe/packages/netbase/files (ie put mine in local/packages/netbase/files) Aug 11 23:27:27 so you only need to replace it. Aug 11 23:27:54 Laibsch: I thing there is another diag menu on spitz. Aug 11 23:28:09 without making a local copy of netbase that is. I'm not sure I understand the rules that are used via SRC_URI to find files Aug 11 23:28:53 gn Aug 11 23:29:09 ant_: night Aug 11 23:29:48 thesing: Well, as you know that part is much more tricky for me Aug 11 23:29:53 I hate touching that CF card Aug 11 23:29:56 I'll replace it Aug 11 23:30:13 (both the card and the flash portion on the collie ;-)) Aug 11 23:30:58 tharvey: Not sure I understand your question Aug 11 23:31:10 But how about you put a copy of netbase there Aug 11 23:31:12 Laibsch, I don't think I phrased it well Aug 11 23:31:22 and make sure it is a "newer" version Aug 11 23:31:27 you mean make my own copy of the netbase recipe? Aug 11 23:31:33 Copy the file and increase rX Aug 11 23:31:43 That is what I would do Aug 11 23:31:56 ya, thats what I typically do, but I'm wondering if there is a way to replace just 1 file that the netbase recipe uses without having to duplicate the entire recipe Aug 11 23:32:12 I think that would be the easiest way Aug 11 23:32:20 tharvey: do you have a new machine or distro? You can have a file specific to it. Aug 11 23:32:34 when you use SRC_URI I'm unclear how bitbate searches for the files referenced... typically they are in a subdir named the same as PN but sometimes they are in files subdir Aug 11 23:32:40 I think you don't even need to increase r if you use an overlay which I think you do Aug 11 23:32:56 tharvey: mom Aug 11 23:33:00 Laibsch, correct, if you have a local overlay and local is first in BBPATH you don't have to use a new version Aug 11 23:33:44 thesing, I do have my own machine/distro - how would I define a specific file for say 'interfaces' that would be picked up by the existing netbase recipe? (or do you mean modify the recipe to use a variable that I define in machine/distro?) Aug 11 23:34:52 tharvey: grep FILESPATH conf/bitbake.conf Aug 11 23:35:07 tharvey: I don't know how it works but the base-files recipe does this. Aug 11 23:35:24 The order is defined in bitbake.conf Aug 11 23:35:41 If you are really inclined to do so, you could hack FILESPATH Aug 11 23:35:51 I am not sure that is what I would do Aug 11 23:36:06 But OE gives you ample opportunities to shoot yourself in the foot ;-) Aug 11 23:36:28 the default filespath stuff is documented in the manual I think. Aug 11 23:36:38 ~oe-manual Aug 11 23:36:46 ~manual Aug 11 23:36:47 extra, extra, read all about it, manual is at http://www.ibot.org/guide-0.43.x.html, or there is no manual, ibot.org is gone, read the source. Aug 11 23:37:05 hm. what was the right word again? Aug 11 23:37:25 It is at bec systems right now Aug 11 23:37:36 http://bec-systems.com/oe/html/ Aug 11 23:37:37 oe site is down right now... can't get to manual Aug 11 23:37:45 ah... thx Aug 11 23:38:11 ibot: oe-manual is http://bec-systems.com/oe/html/ Aug 11 23:38:12 Laibsch: okay Aug 11 23:38:19 ~oe-manual Aug 11 23:38:19 oe-manual is probably http://bec-systems.com/oe/html/ Aug 11 23:38:24 cool Aug 11 23:38:38 cool... thats a good explanation of FILESPATH/FILESDIR Aug 11 23:38:44 tharvey: http://www.pokylinux.org/doc/poky-handbook.html has some more documentation Aug 11 23:38:55 yeah, the docs are pretty good right now, I think Aug 11 23:38:58 * tharvey loves the poky manual Aug 11 23:39:00 I use it all the time Aug 11 23:39:26 Sometimes I hope the poky people would try harder to push stuff back Aug 11 23:39:38 But then again, they have to make some money ;-) Aug 11 23:39:46 And if they are making money, that is good Aug 11 23:39:56 * Laibsch goes through the spitz manual Aug 11 23:40:10 not sure where that is discussed in the poky manual Aug 11 23:42:16 thesing: Your kexec stuff is superawesome, BTW! Aug 11 23:42:29 I hope it will become default boot method for most OE devices Aug 11 23:42:35 tharvey: Maybe they don't discuss it. But the manual is a great place to learn some basic oe. Aug 11 23:43:15 Laibsch: thanks. I would like that too. Aug 11 23:43:46 thesing: Is it the button in the middle of the cursor keys that I need to hit or the OK button on the collie? Aug 11 23:44:04 doesn't matter. Both work. Aug 11 23:44:12 OK Aug 11 23:44:18 And there is a timeout, too? Aug 11 23:44:21 Right? Aug 11 23:44:37 no. It just takes some time to load the new kernel. Aug 11 23:44:48 It takes quite a while after hitting the button for anything to happen which is why I am always unsure which one is correct Aug 11 23:45:01 Maybe you can print some short message to the screen? Aug 11 23:45:16 hmm... I see that I could make an overlay of the netbase recipe which modifies FILESPATH to include my local dir first then includes the netbase recipe Aug 11 23:45:54 tharvey: I would look at the base-files recipe first. Aug 11 23:46:23 thesing, ya - I'm looking at it now - not sure I see how distros/machines specify their own files such as 'interfaces' Aug 11 23:47:03 tharvey: machines provide there own fstab file there. Aug 11 23:47:03 the FILES_* vars? Aug 11 23:49:48 no I think its some default stuff. You could try to create a subdir for you machine and put your interfaces there. Aug 11 23:49:50 thesing, you mean the CONFFILES_* vars? I thought that was for telling the packager what config files belonged to it Aug 11 23:50:35 thesing, yes I see that in the files dir... not sure how it works yet, but I would think that it would have to be in FILESPATH meaning I would have to touch the oe tree and not my local tree Aug 11 23:51:22 what is so bad about touching the oe-tree? Aug 11 23:51:40 well, its not the model were using for our OE project... Aug 11 23:52:52 its interesting... clearly FILESPATH seems to accept ${PN}/${DISTRO} and ${PN}/${MACHINE} as valid location for local files but I'm not sure I see where/how thats specified Aug 11 23:58:41 thesing: http://rafb.net/p/LjxVIZ86.html Aug 11 23:58:46 The thing is still hanging Aug 11 23:58:53 with a blank screen now Aug 11 23:59:51 tharvey: its in base.bbclass I think. My python is not the best, but I bet on the base_set_filespath function. Aug 12 00:01:07 Laibsch: opie image again? console-image works fine for me. (a few times until it gets corrupted somehow.) Aug 12 00:02:01 Laibsch: maybe you could try a kernel with psplash=false or without console=tty1 in cmdline? Aug 12 00:03:42 thesing: yes, opie-image Aug 12 00:03:49 and it seems to be an opie problem Aug 12 00:04:11 I looked more closely at the output on the collie this time Aug 12 00:04:28 there is a touchscreen problem now. Aug 12 00:04:50 this is probably opies problem. Aug 12 00:04:53 the screen goes blank after Angstrom bootloader finishes and opie tries to start Aug 12 00:05:20 Oh, bluelightning actually told me something about that Aug 12 00:05:42 thesing, yes I think your right... looks like it adds OVERRIDES to the filespath allowing you to do things like 'OVERRIDES .= ":myconf"' to allow myconf subdir to be searched first - but it would still be confined to the recipes path Aug 12 00:05:56 well you both have given me some good info to go off of - I'll dig in deeper Aug 12 00:05:58 He said that udev 092 fixes a touchscreen problem on spitz Aug 12 00:06:08 I assume it will fix it on collie, too Aug 12 00:07:13 tharvey: be carefull with OVERRIDES it affects many stuff. Read about it in manual (I hope its in there) Aug 12 00:07:56 Laibsch: maybe I didn't try yet. Aug 12 00:08:47 thesing: does udev have any bearing on the stuff in flash? Aug 12 00:08:59 or is it present only in the tarball stuff on SD? Aug 12 00:09:34 only in the the sd. Aug 12 00:10:10 everything that affects the system is on the sd. Aug 12 00:10:20 OK Aug 12 00:10:27 I'll redo that part Aug 12 00:11:06 Ideally you'll never have to reflash your collie. (except for new features like sdhc support) Aug 12 00:12:39 sdhc sounds cool ;-) Aug 12 00:12:44 thesing, OVERRIDES is not in the manual unfortunately but I know what your saying - it is primarily used to deal with overriding vars - I don't think its completely what I want anyway Aug 12 00:12:54 time to head out... night Aug 12 00:12:56 But I'd be happy to just have a working device again Aug 12 00:14:43 especially since your spitz is broken now? Aug 12 00:15:06 thesing: yes, maybe that one is really broken now. I'll discuss that tomorrow. Aug 12 00:15:27 The other ones are just that software isn't doing what I want it to do. Aug 12 00:16:11 I always liked the collie so I would not mind be stuck with that Aug 12 00:16:19 being Aug 12 00:16:40 thesing: I assume you are quite proficient with git? Aug 12 00:16:53 you being a kernel hacker and all Aug 12 00:17:11 I am starting to think about how to move apt-getable OE over to git Aug 12 00:17:19 well I know it good enough for daily use. Aug 12 00:17:48 you want to have the packages in git? Aug 12 00:18:16 no Aug 12 00:56:56 night all **** ENDING LOGGING AT Tue Aug 12 02:59:56 2008