**** BEGIN LOGGING AT Mon Apr 02 02:59:56 2007 Apr 02 03:05:11 Henryk: ping Apr 02 03:05:24 ah, hi rwhitby Apr 02 03:05:47 I'm answering your mail at this moment :) Apr 02 03:05:55 cool. :-) Apr 02 03:06:45 Henryk: when do you think we'll be able to move mokomakefile users to the 1.8 branch? Apr 02 03:07:49 I'm not sure. Does it build at all with 1.8? I think I remember having heard something about 1.8 needing a relatively current openembedded with task dependency information? Apr 02 03:08:18 I haven't tried 1.8 yet - I'm still building all my OE distros with the 1.6 branch. Apr 02 03:08:30 Henryk: IIRC, the dependency's the other way around.. the proposed dependency changes would require 1.8.2 minimum Apr 02 03:08:55 I can test 1.8 on nslu2-linux first (I should do that ASAP). Apr 02 03:08:57 NAbyss: yes, that's the other thing from two days ago or so Apr 02 03:09:25 I've been waiting for the svnnow patches to be forward ported to 1.8 before trying it out on openmoko. Apr 02 03:09:38 rwhitby: we'll (hmm, I think that is: mickey :) have to look into 1.8 anyway because my new remote revision code will be 1.8 based Apr 02 03:10:59 Henryk: ok. Let me know when it's time to change mokomakefile. (if anon svn access was working on projects.openmoko.org, I'd offer to give you write access) Apr 02 03:11:09 hmm, I have the functionality working for 1.8 (even in a somewhat clean manner), but performance is gonna suck, currently. 1.8 does fork() for different tasks, so I can't easily have cached results Apr 02 03:12:37 yeah, read that in the backlog. Apr 02 04:38:45 rwhitby: how about adding a make flash target to mokomakefile (with flash-uboot, flash-kernel and flash-rootfs subtargets) Apr 02 04:39:13 using dfu-util? Apr 02 04:39:20 let's presume that sha256sum isn't in my distro's coreutils.... anyone know where I can grab sources for just sha*sum? Apr 02 04:39:41 rwhitby: yes Apr 02 04:40:09 Henryk: sounds great. want to send me a patch? ;-) Apr 02 04:40:41 (i.e. do you have an implementation already, or is it just an idea at the moment) Apr 02 04:41:06 just an idea, because I'm kind of lazy and don't always want to look up the exact commands in the wiki Apr 02 04:42:07 laziness is what MokoMakefile is designed for :-) Apr 02 04:42:19 (and to stop typo errors when flashing :-) Apr 02 04:43:19 whoah. "More than one DFU capable USB device found, you might try `--list' and then disconnect all but one device" Apr 02 04:43:39 interesting. my bluetooth dongle wants to do DFU ... Apr 02 04:47:35 yeah, so does mine. Apr 02 04:47:52 another reason to put it in the makefile :-) Apr 02 04:48:34 indeed. forgetting to connect the neo while having bluetooth on might do interesting things ... Apr 02 05:54:13 good morning all Apr 02 05:55:53 koen: how do we modify matchbox_0.9.1.bb somehow so that it rdepends on either matchbox-panel or matchbox-panel-2? Apr 02 05:56:35 at the moment, I get a conflict in the OM image (built from OM SVN and OE head) Apr 02 05:56:59 I think I saw you and RP discussing it in the backlog. Was a decision reached? Apr 02 05:58:25 rwhitby: After I went to bed I came up with an easy solution Apr 02 05:58:42 do_install_append() { mv matchbox-panel matchbox-panel-2} Apr 02 05:58:43 cool Apr 02 05:59:03 and edit the openmoko session file Apr 02 05:59:37 how is that a solution? Apr 02 06:00:07 mb-panel and mb-panel-2 aren't comparable Apr 02 06:00:14 so don't make them conflict Apr 02 06:00:46 and the 'matchbox' meta package is bogus AFAIK Apr 02 06:01:21 i think the solution would be to edit matchbox-common and let it depend on matchbox-panel-2 instead of matchbox-panel Apr 02 06:01:39 s/solution/intention/ Apr 02 06:01:39 no Apr 02 06:01:53 matchbox--2 is unfinished crap right now Apr 02 06:02:09 I am NOT going to put all OE users through that Apr 02 06:02:22 hmm, ok. then remove/fix whatever depended on matchbox-panel-2 Apr 02 06:02:23 since they already have matchbox-common installed Apr 02 06:23:53 03koen 07org.oe.dev * r27aca49b... 10/ (1 packages/matchbox2/matchbox-panel-2_svn.bb): matchbox-panel-2: solve all the madness by making it parallel installable with matchbox-panel-1 Apr 02 06:23:58 03koen 07org.oe.dev * rdb900dbf... 10/ (1 packages/matchbox-panel/matchbox-panel_0.9.3.bb): matchbox-panel: bump PR Apr 02 06:24:04 03koen 07org.oe.dev * r19524f3e... 10/ (3 files in 3 dirs): openmoko-session: use matchbox-panel-2 instead of matchbox-panel Apr 02 06:27:46 03koen 07org.oe.dev * raa5818ff... 10/ (1 packages/matchbox-common/matchbox_0.9.1.bb): matchbox: bump PR Apr 02 06:27:54 03koen 07org.oe.dev * rb4a8b6e0... 10/ (1 packages/gpe-session-scripts/gpe-session-scripts_0.67.bb): gpe-sesssion-scripts: bump PR Apr 02 06:28:51 there, that should have a clean upgrade path Apr 02 06:29:18 (on the assumption that ipkg behaves like the docs say) Apr 02 06:38:43 good morning Apr 02 06:39:21 I'm having problems with perl-native (in dev), since yesterday evening Apr 02 06:39:36 how do I tell monotone to get me the previous rev for a file? Apr 02 06:45:27 oops (rtfm) Apr 02 06:52:49 * Bernardo quits fighting with mtn co for today and leaves for work Apr 02 07:04:00 mickeyl: guten morgen Apr 02 07:04:06 guten Morgen :)) Apr 02 07:04:11 indeed it's too early for me :) Apr 02 07:04:14 * mickeyl brews coffee Apr 02 07:04:16 how are you? Apr 02 07:04:18 * koen|lab gets on his bike and heads to th elab Apr 02 07:04:24 mickeyl: way too early :) Apr 02 07:15:52 re Apr 02 07:16:43 mickeyl: did you have a productive weekend? Apr 02 07:17:05 Bernardo : What kind of problems ? Apr 02 07:17:53 koen: yeah, it was great. although we didn't manage more than 60% of my (way too overloaded) agenda, we did a whole lot of things wrt. OE. Apr 02 07:17:56 OM, even Apr 02 07:18:39 mickeyl: I think I solved the matchbox-panel-2 vs matchbox-panel-1 problems Apr 02 07:19:00 mickeyl: could you sync openmok-session_svn.bb and matchbox-panel-2_svn.bb? Apr 02 07:19:08 let me check Apr 02 07:19:12 (and files/session) Apr 02 07:20:02 mickeyl: hi :) Apr 02 07:20:32 cheerio! Apr 02 07:20:37 Ifaistos: can't build perl-native (claims make depend wasn't called) Apr 02 07:20:50 * Bernardo|away is still at home - running late Apr 02 07:21:01 Bernardo|away: do you have /bin/sh == dash? Apr 02 07:21:28 zecke: probably - I had reverted in edgy, but upgrading to feisty probably reinstated it Apr 02 07:21:49 Bernardo|away: it did for me and perl-native broke Apr 02 07:22:50 to anyone who's interested Apr 02 07:23:06 http://www2.informatik.hu-berlin.de/~tkunze/zaurus/patches/ contains a supposedly working SD driver for Zaurus/Collie (2.6) Apr 02 07:23:09 ok, it had dash again. I'll leave it building now, really have to run Apr 02 07:23:54 mickeyl: http://lists.linuxtogo.org/pipermail/angstrom-distro-devel/2007-April/000152.html :) Apr 02 07:24:25 ah Apr 02 07:24:27 i'm too late :) Apr 02 07:24:43 marcins reply says it all: http://lists.linuxtogo.org/pipermail/angstrom-distro-devel/2007-April/000153.html Apr 02 07:25:05 hehe Apr 02 07:26:28 hehe Apr 02 07:26:38 koen|lab: can you compile the linux kernel with LLVM? Apr 02 07:26:55 zecke: you should be able to Apr 02 07:27:01 dynamically self optimizing paging routines :} Apr 02 07:27:26 mickeyl: if you can take a step back, try to view PyPy and figure out how far they are Apr 02 07:27:34 03koen 07org.oe.dev * rf5edc19d... 10/ (1 packages/esound/esound_0.2.36.bb): esound: bump PR Apr 02 07:27:35 zecke, mickeyl: http://lists.cs.uiuc.edu/pipermail/llvmdev/2007-April/008420.html Apr 02 07:28:16 zecke: i'd like to , but not in the very short run... probably in two months Apr 02 07:28:34 OMG EABI :) Apr 02 07:28:34 * mickeyl waiting for 300revs to go in... Apr 02 07:28:36 zecke: leonardo rocks Apr 02 07:28:50 (santagada, not the turtle) Apr 02 07:30:02 ~time Apr 02 07:30:15 You are educated stupid and therefore too dumb to understand nature's perfect time cube! (2007.04.02 7:30:15 GMT) Apr 02 07:30:21 heh Apr 02 07:32:14 http://rafb.net/p/wpm2hR13.html Apr 02 07:35:26 odd question guys.. is there a way I can add a directory like /home/root to the install for a package? Apr 02 07:35:38 jselect: yes, although it's not recommended Apr 02 07:35:48 FILES_${whatever} += "" Apr 02 07:35:50 zecke: I had to make a correction to sha256sum in base.bbclass to eliminate it in favor of shasum. Patches are in mailing list. Apr 02 07:35:57 mickeyl: why isn't it recommended? Apr 02 07:36:23 including user directories in packages is prone to fail if someone installs without permissions Apr 02 07:36:59 and there's always the question of overriding stuff you might want to keep Apr 02 07:37:03 rpm handles that Apr 02 07:37:05 ipkg not Apr 02 07:37:10 i.e. there's no foobar.ipkgsave Apr 02 07:38:02 mickeyl: ok well part of it is a set of files to go in /opt which don't seem to be included by default either so it's not all bad Apr 02 07:38:20 mickeyl: thanks for the tip though.. I might see about adding it to the readme Apr 02 07:38:31 does anyone know the name of the shell script to make a Packages file for a local feed? Apr 02 07:38:42 ipkg-make-index ? Apr 02 07:39:04 mickeyl: Interesting.. I defined FILES_extra with the full path to the files but they weren't packaged. Apr 02 07:39:13 mickeyl: perhaps I need to include the directories themselves? Apr 02 07:39:36 darmou: bitbake package-index Apr 02 07:39:52 obviously the files in /etc and /usr/bin are packaged automatically if they're in the install dir Apr 02 07:39:53 Henryk, thanks:) Apr 02 07:39:55 jselect: if you add new packages you need to adjust PACKAGES as well Apr 02 07:40:03 PACKAGES += "${PN}-extra" Apr 02 07:40:08 FILES_${PN}-extra = "foobar" Apr 02 07:41:21 morniing Apr 02 07:41:27 cheers hrw Apr 02 07:41:36 hrw: tried SD/2.6 yet? ;) Apr 02 07:41:47 mickeyl: so if I added the files to FILES_${PN} then I wouldn't need to install ${PN}-extra.. seems smoother as they're not actually extras but required parts of the package Apr 02 07:42:04 ya Apr 02 07:42:20 mickeyl: ok thanks, trying that now Apr 02 07:42:33 np, please also review the default value of all these magic variables in bitbake.conf Apr 02 07:43:24 Ah.. It doesn't work if I add them with FILES_${PN} += then it dies trying to copy the /etc/X11 dir from the image dir to the install dir however if I define FILES_${PN} in total via just an = then it works. cheers Apr 02 07:43:40 mickeyl: I'll have a look in there as part of this package dev :) Apr 02 07:43:41 mickeyl: not yet - just ate breakfast Apr 02 07:43:56 hrw: good. a good breakfast is an important start :) Apr 02 07:44:29 hrw: received the tosa Apr 02 07:44:34 had to pay EUR 80 for customs :/ Apr 02 07:56:30 morning Apr 02 08:02:19 Henryk, for some reason I can update my package list ok but can't install anything from it, keeps compaining about missing packages Apr 02 08:02:35 Henryk, I have done an ipkg update of course Apr 02 08:03:16 hmm, don't look at me :) I've only started using ipkg one week ago Apr 02 08:08:53 this is really bizzar [root@spitz:/var/lib/ipkg] # ipkg list | grep mono Apr 02 08:08:53 mono - 1.2.2.1-r1 - Mono Programming Language Apr 02 08:09:15 but when I try to install mono it says that it is not avaliable, what gives? Apr 02 08:09:42 Cannot find package mono Apr 02 08:12:48 darmou: perhaps is has something to to do with achritechure. I also only started using ipkg quite recently but when I perform an ipkg update is showd some information about masked packages etc perhaps it's something like that? Apr 02 08:13:22 I installed mono using the ar packageName , tar zcvf method and it kinda worked Apr 02 08:13:39 Can anyone suggest a program to load an X background that doesn't require a window manager? I like xloadimage but it doesn't have a recipe in oe yet. I'd like to avoid creating one if possible :) Apr 02 08:13:51 keesj, :) done that before many a time too Apr 02 08:17:53 pH5: any plans on SRCDATE for libxcalibrate and calibrateproto? Apr 02 08:18:39 that's a good idea. just pick a date. Apr 02 08:20:56 hmm.. apparently I could set the background with fbsetbg or xv but I don't see packages for those either Apr 02 08:21:16 pH5: today it built with today date Apr 02 08:21:25 s/date/rev/ Apr 02 08:21:48 right, timeless git Apr 02 08:22:14 keesj, I think you are correct about the arch problem even trying to install it directly with the package on disk did not work, I'm not sure why package is a _armv5te type and I am using a 3100 zaurus very strange Apr 02 08:22:55 darmou: check /etc/ipkg/arches Apr 02 08:25:37 XorA, that file arch.conf was missing, fixed problem now, thanks for your help Apr 02 08:36:04 yay! Apr 02 08:36:09 64.3dB SINAD Apr 02 08:36:17 64.36 even Apr 02 08:41:55 * mickeyl has no idea what this means, but.... Congrats! Apr 02 08:44:04 pity the correct answer was 64.37 ... Apr 02 08:46:57 good morning Apr 02 08:48:37 morning Apr 02 08:49:23 is there a way to dictate the host-system, when running on a x86_64 box supporting 32bit userland? Apr 02 08:49:39 the 64bit gcc/binutils combo seem not to work :-/ Apr 02 08:54:29 should I set BUILD_ARCH in my local.conf? Apr 02 08:54:38 mickeyl: Signal to noise and distortion ratio Apr 02 08:55:05 the goal was 60db (e.g 6 million times less noise as signal) Apr 02 08:55:11 ah, cool Apr 02 08:58:37 re Apr 02 08:59:28 ~lart matlab for log being ln Apr 02 08:59:28 * ibot takes out a cattle prod and gives matlab a good jolt for log being ln Apr 02 09:08:21 ~lart OE for autoconf 2.59 Apr 02 09:08:21 * ibot executes killall -TERM OE for autoconf 2.59 Apr 02 09:13:48 NOTE: package coreutils-native-6.0-r1: task do_configure: completed Apr 02 09:13:50 uf.. Apr 02 09:28:54 03koen 07org.oe.dev * rb3b8e63b... 10/ (1 packages/openmoko-apps/openmoko-messages_svn.bb): openmoko-messages: remove patch that was applied upstream Apr 02 09:44:10 rwhitby, koen: I will push coreutils(-native) 6.0 into OE - what do you think about DEF_PREF=-1 for them? let do some builds and if they build then set it as default? Apr 02 09:46:49 mickeyl: I'd been hoping for something sooner than two months :/ Apr 02 09:47:12 mickeyl: I've replied to your lzo question btw, let me know if you need any more info Apr 02 09:49:19 re Apr 02 09:49:37 hrw: coreutils-native to get sha256sum? Apr 02 09:49:43 hrw: how do you want to trust that? Apr 02 09:50:35 zecke: I can compare it with debian coreutils Apr 02 09:51:01 hrw: how do you compare it? You can't Apr 02 09:51:23 zecke: results can be compared Apr 02 09:51:24 hrw: if you use a untrusted sha256sum implementation it is better to not use sha256 at all Apr 02 09:51:28 hrw: which results? Apr 02 09:51:48 zecke: so you suggest "stick with md5sum"? Apr 02 09:52:18 hrw: no, but building coreutils-native to get sha256sum is wrong, it is as stupid as accepting unknown certificates Apr 02 09:53:17 zecke: ok. then no sha256 from me. or you will give me trusted RHEL4 package Apr 02 09:54:21 * XorA guesses the only trusted sha256 in linux is in openssl Apr 02 09:54:45 as I beleive thats the only implementation that even stands a hope of passing FIPs or CC Apr 02 09:54:49 hrw: sure you can manually build coreutils-native and verify it Apr 02 09:55:03 RP: I imagined that, but please see it from a vendor perspective (non-OM here, I'm really wearing the OpenEmbedded hat here)... Apr 02 09:55:06 hrw: but putting a untrusted sha256sum into the core of your trust chain is bad Apr 02 10:04:58 03hrw 07org.oe.dev * r3dca1a73... 10/ (8 files in 3 dirs): gnupg: drop 1.4.0 directory, drop gnupg2 1.9.15 (too old) Apr 02 10:05:09 03hrw 07org.oe.dev * r696fafb8... 10/ (8 files in 3 dirs): coreutils: added 6.0 (DEF_PREF=-1) Apr 02 10:06:30 Regarding sha256sum, I think we should go back to shasum. As this can be easily installed with cpan Apr 02 10:06:52 mickeyl: I know. Perhaps I can can add half the patch to at least paper over the cracks for now whilst not breaking backwards compatibility yet... Apr 02 10:07:10 RP: that'd be helpful Apr 02 10:10:00 mickeyl: My other concern is we already have epoch code in OE.dev despite me trying to apply the brakes :/ Apr 02 10:10:18 ya, that's worrying me as well Apr 02 10:11:28 is this really dangerous? the xorg packages are selected by PREFERRED_VERSION anyway. Apr 02 10:11:33 hrw: sure I can do a coreutils build right now. Apr 02 10:12:00 zecke: sha1sum? Apr 02 10:12:42 hrw: what is that? Apr 02 10:13:04 hrw: sha1sum is SHA-1 and found to be insecure, we are on SHA-256 Apr 02 10:13:15 ok Apr 02 10:14:03 RP, mickeyl: yep - even my emails on the subject clearly pointing out the dependency on OM code didn't stop epoch dependency in .dev getting committed. Apr 02 10:15:15 but that's to be expected with .dev, so I can't complain. Apr 02 10:15:43 pH5: Its dangerous as people don't pay attention and you now need manual interaction to prevent feed contamination unless you use a new bitbake Apr 02 10:16:16 I guess we have to decide how much dev work is allowed in OE.dev... Apr 02 10:17:01 mickeyl: Going back to your point about vendors, a vendor shouldn't be using OE.dev directly and hence should just be able to filter out the changes in theory Apr 02 10:17:26 It comes back to the question of stability and a stable branch too :/ Apr 02 10:19:16 * hrw goes back to bitbake 1.8 Apr 02 10:19:39 or maybe not now.. Apr 02 10:21:03 hrw: ? Apr 02 10:21:09 as a user of oe I would love to have a official stable tree. Apr 02 10:21:49 mr_nice: Lots of people would love it. Can you write down the criteria for how it would be managed and who would manage it though? Apr 02 10:22:06 RP: 1.9 != devshell Apr 02 10:23:03 hrw: I should commit the interactive tasks proposal... Apr 02 10:23:31 hrw: I too can no longer build OE on my CentOS 4.4 box until I find a trusted sha256sum for it. I couldn't find sha256sum in the packages that http://www.openembedded.org/+http%3A//www.openembedded.org/dl/packages/rpm/openembedded-essential-1.1-1.noarch.rpm installs ... Apr 02 10:23:50 * rwhitby mutters something about carts and horses ... Apr 02 10:24:01 rwhitby: I only use that rhel4 box. Apr 02 10:24:17 rwhitby: do you have a shasum? Apr 02 10:24:40 hrw: luckily I build on both CentOS 4.4 and Debian Etch, so I can at least continue development on the OE core team's preferred platform ... Apr 02 10:25:13 zecke: not in the base install, I haven't looked closely on rpmforge and other 3rd party compatible sites yet. Apr 02 10:25:35 rwhitby: is using CPAN okay on CentOS? Apr 02 10:26:01 rwhitby: There is no preferred platform, our aim is to support multiple ones. We just need to find a way forward that lets us do so Apr 02 10:26:19 RP: yeah, I know. I was just having a poke :-) Apr 02 10:26:41 rwhitby: etch is not released, we are using oldstable Apr 02 10:27:32 RP: what would have been better perhaps was an RFC identifying that sha256sum would be a new required essential tool for OE development, and allow people to find and report solutions for installing sha256sum on the devel mailing list for about a week before that requirement was implemented ... Apr 02 10:27:35 zecke: maybe use sha256sum or shasum, whichever is available? Apr 02 10:28:14 I mean it's been insecure for how many years? One more week aint gonna make a difference ... Apr 02 10:28:16 rwhitby: yes, this was send half a year ago Apr 02 10:28:38 s/insecure/slightly less secure/ Apr 02 10:29:02 rwhitby: there was an RFC on which SHA to use, the only input was from Eric and we have settled on SHA-256 Apr 02 10:29:02 zecke: apologies if that's the case. I don't remember sha256sum being identified as the selected tool. I remember md5sum stuff ... Apr 02 10:29:13 rwhitby: zecke did mention it a while ago although I must admit I'd not realised the sha was required Apr 02 10:29:35 zecke: apologies then - I must not have read the RFC closely enough. Apr 02 10:29:36 rwhitby: but actually I didn't look into the required tools Apr 02 10:29:52 rwhitby: as shasum is preinstalled on OSX I assumed its pressence Apr 02 10:30:03 zecke: yep, these things happen. Apr 02 10:30:04 * RP ponders whether bitbake could build the shatools and then anyone wanting security could set ASSUME_PROVIDED Apr 02 10:30:21 more security that is Apr 02 10:30:30 BTW, coreutils-native-6.0 builds find on SlugOS/LE :-) Apr 02 10:30:35 s/find/fine/ Apr 02 10:30:47 RP: bitbake 1.9 should set PATCHRESOLVER = "noop" Apr 02 10:30:53 rwhitby: feel free to test "fis". I've committed it to head now. Apr 02 10:31:03 hrw: that is not bitbake jobs Apr 02 10:31:18 s/bitbake 1.9/OE with bitbake 1.9/ Apr 02 10:31:29 hrw: No, we should commit my interactive tasks patch Apr 02 10:31:41 hrw: noop should have been the default anyway... Apr 02 10:32:16 * NAiL pokes CIA-11 Apr 02 10:33:06 NAiL: fis_1.0-r1? Apr 02 10:33:14 no, -r2 Apr 02 10:33:15 03nail 07org.oe.dev * rcae3b433... 10/ (4 files in 3 dirs): fis: Update fis util to one that doesn't require boost/libstdc++. Looks like it's working, but needs testing Apr 02 10:33:25 I committed to mtn.oe.org Apr 02 10:33:54 I've semi-tested it myself, on a copy of a redboot fis partition Apr 02 10:34:21 I saw it's down from 800Kb to 8Kb now ... Apr 02 10:35:11 yeah, that's a stripped, shared binary on my laptop. Not sure what the ARM size would be, but I bet it won't be much larger. Apr 02 10:35:34 phil has been very helpful when he could spare the time :) Apr 02 10:36:22 Now someone needs to write a util to change the redboot config ;-) Apr 02 10:36:52 BTW: touch conf/checksums.ini Apr 02 10:37:10 and it will not check any sums Apr 02 10:38:01 NAiL: I thought someone wrote that ... Apr 02 10:38:37 really? Who, when what where? :) Apr 02 10:38:52 NAiL: fis - 8352 bytes. Apr 02 10:38:59 nice :-D Apr 02 10:42:16 So back to our SHA256 options Apr 02 10:42:41 a) require sha256sum from a recent coreutils which is not available on OSX, FreeBSD, RHEL, Mandriva Apr 02 10:43:04 b) require shasum which is installed by default on OSX, packaged in Ubuntu and Debian and installable using CPAN Apr 02 10:43:14 c) require python2.5 which has a hashlib Apr 02 10:43:38 zecke: I try not to use CPAN on my CentOS box, cause then I don't get automatic updates on perl version changes. Apr 02 10:43:57 yeah, actually all three options are no good for a production RHEL system. Apr 02 10:44:14 * RP notes his server is on python 2.4 :/ Apr 02 10:44:20 rwhitby: what would work for rhel? Apr 02 10:44:30 pH5: good question. Apr 02 10:44:33 pH5: probably not using sha256 Apr 02 10:44:40 :) Apr 02 10:44:41 md5sum is there ;-) Apr 02 10:45:00 rwhitby: http://andrzejekiert.ovh.org/software.html.en seems to be what I'm looking for. I'll try to get it into OE ;) Apr 02 10:45:09 zecke: shouldn't it be possible to use whichever of a),b),c) is available? Apr 02 10:45:34 pH5: d) write my own sha code Apr 02 10:45:38 zecke: What about d).. package sha256sum in OE and build as required, just as quilt etc. is? Apr 02 10:45:54 zecke: d) building a native package and playing with ASSUME_PROVIDED Apr 02 10:46:02 rwhitby: 20 -rwxr-xr-x 1 repvik repvik 17032 2007-04-02 12:46 fconfig Apr 02 10:46:06 NAbyss: yes, if we package sha256sum.c in OE Apr 02 10:46:14 rwhitby: That's not ARM, but it looks good ;-) Apr 02 10:46:14 zecke: I'd be happy with a flag which you have to set in local.conf to use an OE-built sha256sum Apr 02 10:46:55 rwhitby: currently you can disable checksum checking by creating an empty conf/checksums.ini in your BBPATH Apr 02 10:46:56 zecke: will you trust such sha? Apr 02 10:47:45 hrw: well base.bbclass and sha256sum.c are in the same manifest, and I hope mtn moves to sha256 soon Apr 02 10:48:12 zecke: here's the thing. I do agree with checksumming tarballs that come from various places. So I don't want to disable it. Apr 02 10:49:11 zecke: and I am happy to check the tarball that I download for coreutils 6.0, to make sure that the sha256sum that OE builds for me can be trusted. Apr 02 10:49:32 and then I set a flag in my local.conf and it's happy ever after ... Apr 02 10:49:54 okay so Apr 02 10:50:16 d) try shasum and sha256sum and if both are not present disable sha checking Apr 02 10:50:22 e) require openssl Apr 02 10:50:49 f) require all of them, runtime-check their speed and choose the fastest Apr 02 10:50:58 pH5: hehe Apr 02 10:51:16 g) same as f) but compare all results Apr 02 10:51:30 zecke: what's the openssl alternative? my openssl on RHEL4 has sha and sha1 commands Apr 02 10:51:49 rwhitby: http://nedron.net/cgi-bin/suid/dnedrow/fom?_recurse=1&file=1 Apr 02 10:51:50 h) d + WARNING Apr 02 10:52:18 s/warning/insult/g Apr 02 10:52:51 rwhitby: openssl dgst --help 2>&1 | grep sha256 Apr 02 10:52:51 yeah, we need more emotional software Apr 02 10:53:10 -bash-3.00$ openssl dgst -sha256 Apr 02 10:53:10 unknown option '-sha256' Apr 02 10:53:39 03nail 07org.oe.dev * rb97fbd84... 10/ (6 files in 4 dirs): redboot-utils: move package fis to redboot-utils, add package fconfig Apr 02 10:53:40 rwhitby: fconfig commit coming in. If you happen to have a config partition handy, feel free to test :-) Apr 02 10:54:06 my debian openssl has it, but not my mac openssl Apr 02 10:55:35 koen|lab: which openssl Apr 02 10:55:35 /space/dports/bin/openssl Apr 02 10:55:51 rwhitby: which sha's do you have? Apr 02 10:56:04 zecke: /usr/bin/openssl Apr 02 10:56:09 zecke: so how about the option of allowing a flag in local.conf to say that you trust the OE sha256sum ? Apr 02 10:56:20 -sha1 to use the sha1 message digest algorithm Apr 02 10:56:20 -sha to use the sha message digest algorithm Apr 02 10:56:48 openssl-0.9.7a-43.14 Apr 02 10:57:04 rwhitby: well, I can put sha256sum.c into OE, and this is equally trusted as our base.bbclass Apr 02 10:57:16 zecke: that works for me. Apr 02 10:57:42 rwhitby: so, now OE has userspace utils for both uboot and redboot manipulation :-) Apr 02 10:58:14 NAiL: which tool to tweak uboot? and how it works? Apr 02 10:58:30 NAiL: now we need a devirginator for slugs, which runs on vendor firmware, and turns it into SlugOS without even a reboot. Apr 02 11:00:01 * hrw will need to finally add serial port into PC Apr 02 11:00:28 my USB serial port works fine Apr 02 11:01:24 RP: you will need to help to add the proper quilt-native:do_fetch dependency on sha256sum.c Apr 02 11:03:08 Crofton: I have onboard serial - just need to connect cable Apr 02 11:03:46 ah Apr 02 11:09:49 ~lart collie patchset Apr 02 11:09:50 * ibot cuts collie patchset into thin stripes Apr 02 11:11:59 hrw: it's against 2.6.unknown? Apr 02 11:17:38 2.6.20 Apr 02 11:26:18 NAiL: we'll need to add checksum support for the NAS100d and DSMG600. Apr 02 11:26:26 (to the fis package) Apr 02 11:29:10 what do I have to do in Makefile.am to let a program be linked with the --export-dynamic ld option? Apr 02 11:29:10 NAiL: fconfig didn't emit any package ... Apr 02 11:29:20 03Laibsch 07org.oe.dev * r934ffc45... 10/ (1 packages/qpf-fonts/qpf-unismall_1.0.0.bb): qpf-unismall: hardwire SRC_URI for .jp sourceforge mirror. Seems to be necessary. Apr 02 11:29:21 zecke: Just let me know what kind of dependency we need Apr 02 11:34:34 * koen|lab witnesses a combinatorial explosion when trying to calculate the regression tests needed for a stable OE branch Apr 02 11:37:54 ~lart git for GNUism's Apr 02 11:37:54 * ibot puts git into a headlock and administers a mighty noogie, rubbing half of git's hair of for GNUism's Apr 02 11:38:10 uf. kernel built (too big but built) Apr 02 11:40:04 hi! Apr 02 11:40:18 what is shasum and where to get it? Apr 02 11:40:33 any -native package? Apr 02 11:42:29 coreutils-native 6.0 Apr 02 11:46:37 what package for ubuntu edgy? Apr 02 11:47:00 coreutils Apr 02 11:47:02 coreutils has md5sum but apparently not shasum Apr 02 11:47:12 Laibsch: coreutils 5.97 you need Apr 02 11:47:18 Laibsch: read the backlog Apr 02 11:47:44 Well, that version does not seem to be in edgy yet Apr 02 11:47:59 I'll go the "bitbake coreutils"-route Apr 02 11:48:14 coreutils-native Apr 02 11:49:30 Laibsch: bitbake coreutils-native-6.0 Apr 02 11:51:31 if anyone adds this line to the wiki, I will don't hesitate to shoot him/her Apr 02 11:53:28 20K to take.. Apr 02 11:57:56 hrw: chicken-and-egg-problem. Cannot build coreutils-native-6.0 without sha256sum. Moving checksums.ini out the way as suggested elsewhere did not fix this either. Apr 02 11:58:52 Wouldn't it be best to disapprove this change until some better solution has been found (such as including the code in base.bbclass)? Apr 02 12:21:42 Laibsch: keep your feet still :) Apr 02 12:22:28 Laibsch: but touching checksums.ini should work, if not then please take a look at base.bbclass and print the path Apr 02 12:58:47 is there some plans to make cxtest available in OE? Apr 02 13:00:42 that'd be nice tool for regression testing Apr 02 13:00:58 slapin_nb: hey :) Apr 02 13:01:09 robtaylor, hi! Apr 02 13:01:16 slapin_nb: how goes? Apr 02 13:01:33 robtaylor, fine, changing jobs :) Apr 02 13:02:05 slapin_nb: oh, cool. who's you're new employer? Apr 02 13:02:34 robtaylor, since 12 of April, Siemens-ru Apr 02 13:03:50 slapin_nb: very cool :) Apr 02 13:04:07 robtaylor, hope so :) Apr 02 13:04:11 slapin_nb: and doing open source stuff? Apr 02 13:12:44 robtaylor, yes :) Apr 02 13:12:51 :) Apr 02 13:13:11 robtaylor, in 5-letter company-style though, but anyway :) Apr 02 13:55:11 03xora 07org.oe.dev * rabc5baea... 10/ (4 files in 2 dirs): Apr 02 13:55:11 osb-jscore_svn.bb, osb-nrcit_svn.bb, osb-nrcore_svn.bb : update the PV to Apr 02 13:55:11 reflect svn updates, some small changes to fix building. Apr 02 14:01:29 is OE based on Debian? Apr 02 14:04:26 no Apr 02 14:04:53 OE use some policies from Debian, most of our distros use Debian naming of packages Apr 02 14:08:03 hrw: thanks Apr 02 14:08:19 I have to get back on my desktop, this console isnt cutting it! Apr 02 14:12:19 hrw: uboot-utils, to fiddle with uboot config from userspace Apr 02 14:12:47 requires working mtd though, someone tested on fic neo and couldn't get it to work. Apr 02 14:12:54 ok Apr 02 14:13:11 rwhitby: didn't it make a package? That's ... odd. Apr 02 14:13:24 It did when I built it here ;) Apr 02 14:25:22 rwhitby: found the problem, committing fix now. Apr 02 14:27:39 03nail 07org.oe.dev * r4d3d0b6a... 10/ (1 packages/redboot-utils/fconfig_20060419.bb): fconfig: Actually install binary, so it gets packaged Apr 02 14:35:29 zecke: ROFL http://gcc.gnu.org/ml/java-patches/2007-q2/msg00000.html Apr 02 14:46:33 chouimat: ;DDD Apr 02 14:51:43 :) Apr 02 15:00:10 added the simpad development update as bugrep 2073. hopefully this time I made no mistakes. Apr 02 15:05:58 mr_nice: well, even if you did mistakes. There is nothing that can't be recovered/fixed Apr 02 15:07:41 CIA-11: kick Apr 02 15:09:25 03mickeyl 07org.oe.dev * r5f616224... 10/ (16 files in 5 dirs): ipkg: remove a couple of outdated ones, update is_processing patch courtesy OpenMoko, Inc. Apr 02 15:09:30 03xora 07org.oe.dev * rd40606a5... 10/ (3 files in 3 dirs): Apr 02 15:09:30 kazehakase_0.4.5.bb : interesting gtk-webore based browser, currently Apr 02 15:09:30 requires svn versions of osb-nr* and osb-jscore stuff. Apr 02 15:09:34 03pH5 07org.oe.dev * r6bc2006e... 10/ (1 packages/openmoko-base/openmoko-theme-standard-qvga_svn.bb): Apr 02 15:09:34 openmoko-theme-qvga: don't package the vga theme and label as qvga, Apr 02 15:09:34 use the qvga theme instead Apr 02 15:10:23 zecke: its about getting integrated - less mistakes -> faster integration :). often make misstakes -> more effort for the integrator to read it -> long time to get integrated even if it is fixed in the second patch. Apr 02 15:12:03 s/integrated/included Apr 02 15:37:57 hi all Apr 02 15:40:19 ah my mini sha256 code seems to work Apr 02 15:41:07 mrdata: hi Apr 02 15:41:26 mr_nice: hi too Apr 02 15:43:47 koen: do you think naming my sha256sum code zhasum is appropriate? Apr 02 15:43:59 zecke: vlasum Apr 02 15:44:33 hehe, I want to avoid confusion with the real implementation Apr 02 15:47:02 ~lart me for forgetting to start the webcore build :} Apr 02 15:47:02 * ibot duct-tapes zecke to the floor and drools on him for forgetting to start the webcore build :} Apr 02 15:48:34 the new svn ones? Apr 02 15:48:52 pH5: oops sorry Apr 02 15:49:03 koen: yes Apr 02 16:14:05 !oebug 2074 Apr 02 16:14:07 * * Bug 2074, Status: NEW, Created: 2007-04-02 09:20 Apr 02 16:14:08 * * r.farina(AT)adelphia.net: opie-rdesktop build failure Apr 02 16:14:08 * * http://bugs.openembedded.org/show_bug.cgi?id=2074 Apr 02 16:14:15 if someone could take a peek at that for me please Apr 02 16:20:16 RP: we can now hook shasum-native as primary depedency into OE :} Apr 02 16:21:30 For bitbake 1.8 this would be do_fetch[depends] = "shasum-native" and do_fetch[depends] = "" in the .bb Apr 02 16:21:52 zecke: Older bitbakes are ugly and will need playing with DEPENDS :/ Apr 02 16:22:26 s/shasum-native/shasum-native:do_populate_staging/ Apr 02 16:23:05 RP: as we have epoch handling in OE, we can consider to require 1.8.2? :} Apr 02 16:23:34 zecke: Perhaps now is the time for an anon python function to automatically inject svn-native, cvs-native, git-native as needed too :} Apr 02 16:23:56 zecke: mickeyl said two months earlier for that. We need to discuss this... Apr 02 16:37:48 RP: I always try to avoid this situation. With Bitbake1.4 and 1.6 we did fairly well. I would like to keep such changes at a minimum :} Apr 02 16:38:50 RP: we could also up the dep without landing the changes Apr 02 16:39:22 once we require such things, we can add the 'injects' step by step Apr 02 16:40:03 does OE still have the old and insecure openssl? Apr 02 16:40:29 probably we do :( Apr 02 16:40:47 we should just remove OpenSSL from OE Apr 02 16:40:52 I vaguely remember something about patches in the bugtracker Apr 02 16:41:29 zecke: We probably could make older versions a special case and add code to inject things into them maybe. We can't sort the epoch problem easily though... Apr 02 16:43:02 how do I get out of the shell I get dropped into if a patch fails to apply, without interrupting OE? Apr 02 16:43:21 philippe: pester RP to commit his patch :) Apr 02 16:43:49 zecke: ok... Apr 02 16:43:51 RP: does your kernel have hostap firmware download enabled? Apr 02 16:44:05 ~pester RP Apr 02 16:44:10 RP: Are we there yet? .. Are we there yet? .. Are we there yet? see http://beverlys.net/LJ/BuggingYou.swf Apr 02 16:44:28 ~poke RP Apr 02 16:44:30 * ibot cuts down a small tree, sneaks up behind RP, pokes RP repeatedly, hilarity ensues. Apr 02 16:44:41 ~emulate stewie Apr 02 16:44:43 Damn you, damn the broccoli, and damn the Wright Brothers! Apr 02 16:44:46 actually it seems to work now.... CTRL+D seems to continue Apr 02 16:45:08 RP: for the shasum problem DEPENDS = "shasum-native" inside the base.bbclass Apr 02 16:45:12 ~emulate zecke Apr 02 16:46:59 philippe: hey? Apr 02 16:47:47 philippe: There is a patch on the mailing list which I need to commit since nobody has objected. I'm presuming this is bitbake 1.8+? Apr 02 16:47:58 Zero_Chaos: No, its not enabled and won't be as standard Apr 02 16:49:22 Actually 1.6, but it is the one I use for the openmoko builds... Apr 02 16:49:31 RP: since it is a willful act to update the firmware, and almost always needed on the crappy CF cards. I am certain your reason is because it is such a "danger", but honestly, it is more of a pain in the butt to not have upgrade ablity IMHO. Apr 02 16:49:39 philippe: Right, Ctrl+D is the right thing for that Apr 02 16:49:56 I installed this new shasum-native and I have tmp/staging/native-machine/bin/sha256sum, but the sanity checker keeps complaining that it is missing. What am I doing wrong? Apr 02 16:50:10 Zero_Chaos: It was a distro decision and not one I'm going to argue with Apr 02 16:50:21 RP: it is okay as long as it works. I believe last time it stopped but that might be because I used exit Apr 02 16:50:43 RP: if the choice was not yours, can you tell me who I should beg? Apr 02 16:50:58 Anyway 1.8 is a great improvement for multi-cpu machines Apr 02 16:54:17 so this weekend I made a babycam using a spare epia board, an old quickcam webcam, an ide to cf adapter, and a 32mb cf card running OE :) Apr 02 16:54:26 cu Apr 02 16:55:27 it works pretty well, but the wireless connection is dropping after a while and I can't figure out why Apr 02 16:55:34 I'm using the webcam_server software Apr 02 16:55:50 and I had to make a .bb file for the quickcam driver Apr 02 16:57:54 03florian 07org.oe.dev * re93b10cf... 10/ (1 packages/gpephone/gpe-applauncher_0.7.bb): gpe-applauncher: Remove obsolete patch. Apr 02 16:57:58 03florian 07org.oe.dev * raff69a14... 10/ (1 packages/gpephone/gpe-applauncher-0.7/default-icon.patch): gpe-applauncher: remove patch file Apr 02 16:58:03 03mickeyl 07org.oe.dev * r68025812... 10/ (1 classes/openmoko-base.bbclass): classes/openmoko-base.bbclass: add support for inputmethods Apr 02 16:58:07 03mickeyl 07org.oe.dev * read450bb... 10/ (4 files in 4 dirs): openmoko-keyboard: catch up with changes in OM svn Apr 02 16:58:11 03freyther 07org.oe.dev * r4182d550... 10/ (8 files in 4 dirs): Apr 02 16:58:11 packages/shasum: Add a sha256sum utility with little dependencies Apr 02 16:58:11 To avoid weird requirements add a sha256sum tool to OpenEmbedded Apr 02 16:58:11 directly. This will be part of the OE manifest and be guarded Apr 02 16:58:11 by the monotone OE trust chain. Apr 02 16:58:57 is the css weird on openembedded.org ? (the paragraphs highlight when i move my mouse over them) Apr 02 17:02:32 03florian 07org.oe.dev * r77ed88b5... 10/ (1 packages/gpephone/gpe-applauncher_0.7.bb): disapproval of revision 'e93b10cfd979b1f989c686de6d2f75006362d34e' Apr 02 17:02:38 03florian 07org.oe.dev * re4274d99... 10/ (1 packages/gpephone/gpe-applauncher-0.7/default-icon.patch): disapproval of revision 'aff69a14340dd1655bfe774109fde80cc509eed6' Apr 02 17:10:52 what is a popular platform to run this on? Apr 02 17:14:01 hi Apr 02 17:16:04 I'm a moron.. Apr 02 17:23:24 RP: http://rafb.net/p/06RGAW92.html comments? Apr 02 17:27:33 zecke: It shouldn't depend on INHIBIT_DEFAULT_DEPS, just that PN != "shasum-native" Apr 02 17:28:51 03philippe 07org.oe.dev * r26684890... 10/ (1 packages/openmoko-inputmethods/openmoko-keyboard_svn.bb): Apr 02 17:28:51 packages/openmoko-inputmethods/openmoko-keyboard_svn.bb Apr 02 17:28:51 * correct the svn source url Apr 02 17:29:04 xuumbi: hehe... nice use of resources :) Apr 02 17:29:28 philippe: ehm, that should have been held back for 4.5 hours Apr 02 17:29:43 philippe: people that don't use SRCDATE=now will get an error now Apr 02 17:31:26 koen: oops sorry. Apr 02 17:32:52 but I don't get the exact issue. It probably does not build for them anyway as the original svn url is broken anyway. Apr 02 17:33:40 philippe: the svn fetcher fetches the state of last midnight Apr 02 17:33:55 and at that point it was still in /applications/ Apr 02 17:34:22 oh yeah... true.... Apr 02 17:34:41 (unless you use SRCDATE=now) Apr 02 17:34:57 well the problem will solve itself in 4.5 hours... Apr 02 17:35:00 indeed Apr 02 17:35:28 I think midnight GMT even Apr 02 17:35:32 most people will be building OpenMoko with SRCDATE=now anyway Apr 02 17:35:55 people that do that probably use the svn overlay as well Apr 02 17:36:18 SRCDATE=now without the openmoko bitbake patch == breakage Apr 02 17:37:22 I haven't being looking into detail in that stuff as I have been using the MokoMakefile. Apr 02 17:37:29 heh Apr 02 17:37:35 I did not want to screw up my other OE build env Apr 02 17:37:36 too stupid to setup OE? Apr 02 17:38:05 nope, wanted to keep OM apart from regular OE with 1.8 Apr 02 17:39:21 there is no need for the svn overlay IMO Apr 02 17:39:30 unless you want the lastest and untested kernel Apr 02 17:48:13 RP: inhibit default deps won't work?! Apr 02 17:56:02 woglinde: http://www.openembedded.org/~zecke/opengl/ Apr 02 17:56:34 hm ogg Apr 02 17:57:11 woglinde: vlc can play them Apr 02 17:57:31 woglinde: hi Apr 02 17:57:40 hi mr_nice Apr 02 17:57:44 zecke mplayer too Apr 02 17:58:17 zecke and now you writing beryl-ng? Apr 02 17:58:32 woglinde: no, that is just a compiz plugin Apr 02 17:58:42 woglinde: I hope to have some OpenGL ES hardware soon :) Apr 02 17:59:59 * koen should fix mesa in OE to build for powerpc Apr 02 18:00:20 zecke, so we're heading towards Xgl-kdrive ? ;) Apr 02 18:09:50 zecke: should I print a MVMD shirt and wear it at ELC? Apr 02 18:12:44 well, try to create a MVMD pin Apr 02 18:13:34 RP: http://rafb.net/p/xzjkH479.html Apr 02 18:13:40 RP: this works with BitBake1.6 Apr 02 18:16:23 I'm back in a hour or two (going to swim a bit= Apr 02 18:20:10 oh another error to fix Apr 02 18:22:03 * chouimat says some very nasty thing about "incompetence and workplace politcs" Apr 02 18:22:57 Good evening Apr 02 18:23:43 koen: and what would MVMD stand for? Apr 02 18:26:10 philippe: google has some interesting suggestions...but I don't think those are what koen has in mind...I am guessing the MV is for Monta Vista... Apr 02 18:26:53 hvontres|poodle: I got that far :-) Apr 02 18:27:56 hvontres|poodle: of course knowing how koen feels about MV maybe the Muzzel Velocity Measuring Device could be close...:) Apr 02 18:28:34 Master Dungeon? Apr 02 19:00:39 * chouimat is slowly sailing toward the mythical land of Insania Apr 02 19:35:14 llvm-native-1.9+cvs20070402 Apr 02 19:35:24 ehm Apr 02 19:35:25 NOTE: package llvm-native-1.9+cvs20070402: completed Apr 02 19:35:27 :D Apr 02 19:35:29 whee Apr 02 19:40:49 03koen 07org.oe.dev * ra36c2e4f... 10/ (1 packages/llvm/llvm-native_cvs.bb): llvm: add cvs version of llvm-native, this adds ARM support to the 1.9 version Apr 02 19:43:10 koen: :) Apr 02 19:50:35 drat, disk is full Apr 02 19:50:45 Hmm, must be mid-terms, someone made a bomb threat on one of the academic buidlings Apr 02 19:50:57 What school, Crofton? Apr 02 19:51:04 Virginia Tech Apr 02 19:51:41 Hokie High Apr 02 19:55:25 * koen taps foot, waiting for rm to complete Apr 02 19:56:36 koen: mv dir dir.$$ Apr 02 19:56:43 koen have you any comparsion for llvm with handheld sutff? Apr 02 19:56:45 koen: rm -rf dir.newname Apr 02 19:57:00 woglinde: [<----- that much better ----->] Apr 02 19:57:05 woglinde: I'll have to get it to compile something first :) Apr 02 19:57:20 koen hehe okay I see Apr 02 19:57:41 woglinde: the indt people seemed to be pretty excited about it Apr 02 20:00:52 zecke: hi, can i ask you one ore two questions, dumb i know Apr 02 20:01:30 mrdata: just ask, maybe others can answer as well Apr 02 20:01:36 okay Apr 02 20:02:02 dev->suspend is a pointer of a function Apr 02 20:03:08 mrdata: I can believe that Apr 02 20:03:23 how can i see the adress of the pointer inside a printk( 0x%08X, dev->suspend ) Apr 02 20:03:55 mrdata: good question, %p but I don't know if printk supports it Apr 02 20:04:09 printk( 0x%08X, dev) doesn't do it? Apr 02 20:04:47 the values are not ever the same Apr 02 20:05:50 equal functions, many runs, zero pointer, next than real one Apr 02 20:05:50 mrdata: so dev and dev->suspend might be uninitialized? (even if the kernel memsets everything...) Apr 02 20:06:47 but also mq200, that can not be true or? Apr 02 20:07:16 mrdata: printk from another place, e.g. in the mq200 init Apr 02 20:07:38 mrdata: the question is if someone makes the pointer invalid... Apr 02 20:08:04 RP: koen hrw|gone : http://rafb.net/p/zIyIXx36.html comments? Apr 02 20:08:28 the simpad try suspend to ram, so pm gives suspend and later suspen_late run Apr 02 20:09:17 mrdata: I don't know enough context to give a meaningul answer, the printk looks okay, sane Apr 02 20:09:29 okay, the short answer for printk and pointer should try %p Apr 02 20:09:31 mrdata: e.g. print dev as well to see if this is sane/insane Apr 02 20:09:56 mrdata: I don't think %p is known, so your code is good enough, specially on 32 bit archs Apr 02 20:10:36 the serial console have problems with dev_dbg Apr 02 20:10:38 zecke: looks pretty ok Apr 02 20:11:02 someone with bitbake 1.8 needs to test it :) Apr 02 20:12:23 printk(0x%08X, *dev->suspend) is also not the correct one, was my first try Apr 02 20:13:22 zecke: I don't have 1.8 Apr 02 20:13:27 zecke me? Apr 02 20:13:35 hm not today Apr 02 20:13:47 koen: what do you use? trunk? Apr 02 20:13:52 yes Apr 02 20:15:59 and thanks for ouer help, i forgot to say Apr 02 20:20:21 hmm Apr 02 20:20:23 configure.in:2226: error: possibly undefined macro: AS_FOR_TARGET Apr 02 20:27:11 hmm^2 Apr 02 20:27:13 GNUmakefile:16: /CoreOS/Standard/Standard.make: No such file or directory Apr 02 20:37:59 RP: this doesn't work for bitbake 1.8 Apr 02 20:44:03 03freyther 07org.oe.dev * r2a4ed2b2... 10/ (5 files in 3 dirs): Apr 02 20:44:03 classes/base.bbclass: place shasum-native as the first package to build Apr 02 20:44:03 classes/base.bbclass: Make sure shasum-native does not depend on quilt Apr 02 20:44:03 classes/sanity.bbclass: No need to require sha256sum as we use oe_sha256sum Apr 02 20:44:03 packages/shasum: Change the name, make the -native version not have any deps Apr 02 21:04:12 good night all Apr 02 21:21:45 03rwhitby 07org.oe.dev * rd0db7417... 10/ (1 packages/tasks/task-slugos.bb): task-slugos: Remove altboot Apr 02 21:52:52 Good night Apr 02 22:15:20 03coredump2 07org.oe.dev * rb3f0c4c4... 10/ (3 files in 2 dirs): Apr 02 22:15:20 altboot: Add SNVR28 Apr 02 22:15:20 NSLU2: Remove /sbin/altboot error message on virgin systems Apr 02 22:15:20 NSLU2: Failsafe fallthrough to the flash-based rootfs should work now on all die() and mdie() calls. Apr 02 22:30:41 CoreDump|home: hey :) Do you think those die() and mdie() mods would work on a Z as well? Apr 02 22:30:50 I get an error of "arm-linux-ar: not found" when trying to compile a package. What am I missing? Apr 02 22:31:24 hvontres|poodle: the current trunk is not useable on anything but NSLU2 at all ;) That will change eventually of course Apr 02 22:36:21 CoreDump|home: Ok. :) That does sound like a very usefull feature for automagically going into rescue mode when booting from SD fails. Apr 02 22:37:11 hvontres|poodle: That is indeed exactly what it does and will eventually be available on all Z as well Apr 02 22:37:50 CoreDump|home: Cool :) how's life in nightshift treating you? Apr 02 22:38:19 actually it's late-shift, not night-shift and I really like it =) Apr 02 22:39:13 CoreDump|home: Heh... I guess it makes shopping a bit easier :) Apr 02 22:39:50 * hvontres|poodle remembers the old Ladenschluss houres from 1988 Apr 02 22:39:56 it does Apr 02 22:40:18 hvontres|poodle: heh "Ladenschluss" is getting weaker and weaker every year. Apr 02 22:40:58 * hvontres|poodle likes 24 Hr Suppermarkets Apr 02 22:41:37 hvontres|poodle: we are getting there, slowly =) Apr 02 23:11:25 03justinp 07org.oe.dev * r6af41da9... 10/ (1 packages/gpsd/gpsd.inc): gpsd: export BUILD_SYS and HOST_SYS as setup.py expects them to be set Apr 02 23:11:29 03justinp 07org.oe.dev * r49597257... 10/ (1 packages/gpsd/gpsd.inc): gpsd: export BUILD_SYS and HOST_SYS as setup.py expects them to be set, add python to DEPENDS Apr 02 23:14:07 Laibsch: binutils Apr 03 00:18:23 NAiL: Thank you for the hint. I wonder why it was not build. I have restarted compilation. Apr 03 00:19:49 np Apr 03 00:49:57 could anyone suggest a program for displaying an image on the root of an x display? Apr 03 00:52:00 jselect, esetroot, feh Apr 03 00:52:08 wmsetbg Apr 03 00:52:17 thanks meister, I've looked at xloadimage, xwud, xsetroot, fbsetbg **** ENDING LOGGING AT Tue Apr 03 02:59:56 2007