**** BEGIN LOGGING AT Mon Aug 27 02:59:57 2007 Aug 27 03:03:36 interfaith: I still don't understand why you need to patch the redboot partition detection code.\ Aug 27 03:04:17 if the partition table is located at the start of the last erase block in the flash chip, then the mainline code should find it. Why exactly do you need to patch it, and what exactly does your patch do? Aug 27 03:06:53 From what i can see, your patch removes the usage of the "directory" kernel parameter (which means that right there we can not apply it, cause that would break any other machine using the same ixp4xx kernel), and then sets offset to 0, and then looks at the second to last block. Aug 27 03:07:24 interfaith: It seems to me you could achieve the same outcome with no patch by simply supplying 'directory=-2' on the kernel cmdline. Aug 27 03:09:16 are you sure the flash configuration is set correctly for your board? perhaps your platform initialisation code is not reading the erase block size correctly? Having the partition in the second-to-last block instead of the last block is very much non-standard - every redboot implementation I've seen puts it in the last block. Aug 27 03:10:28 interfaith: BTW, 32 lines is not "just a few lines" - you should use a pastebin as I suggested next time. Aug 27 03:31:13 interfaith: I will update the ixp4xx-kernel defconfig to enable the IXDPG425 board code, but I won't apply that redboot patch of yours, cause it is superfluous and unacceptable as a kernel patch anyway. Aug 27 04:00:41 03rwhitby 07org.oe.dev * rdb972685... 10/ (3 files in 3 dirs): ixp4xx-kernel: Enable support for IXDPG425 boards in the defconfig Aug 27 04:23:03 i was just showing the code i had, not a planned patch, sorry Aug 27 04:29:31 fine, so simply running task-base will get the new defconfig ? perhaps the new redboot and new kernel will do the p30 as is. Aug 27 05:52:04 why would modpost crash in a segmentation fault ? Aug 27 05:58:36 could be hundred reasons Aug 27 06:01:03 ynezz, yeah i dont know how to get rid of it Aug 27 06:11:16 good morning Aug 27 06:27:07 can someone help to get madwifi cross compiled ? Aug 27 06:39:44 http://rafb.net/p/VN2Ki375.html is my modpost segfault ??? any hints ? Aug 27 06:52:53 good morning all Aug 27 07:09:38 steliosk, did you have had modpost segfaul ? Aug 27 07:10:54 rob_w : never seen it before in OE Aug 27 07:12:27 rob_w : what is it trying to execute in line 1 ? /bin/sh: line 1: 7183 Segmentation fault scripts/mod/modpost Aug 27 07:12:39 looks like Aug 27 07:12:45 which line 1 is that ? Aug 27 07:13:12 rob_w : from modpost i think Aug 27 07:13:26 but modpost is a binary ? Aug 27 07:19:12 rob_w : yes it is Aug 27 07:19:37 so which line 1 is that , which script is called by then ? Aug 27 07:36:49 rob_w: if I were you, I would use strace as next step Aug 27 07:37:10 strace inside oe ? hmm Aug 27 07:38:36 why not, modpost is host binary right? Aug 27 07:38:49 or what does 'file modpost' returns? Aug 27 07:39:32 ile modpost Aug 27 07:39:32 modpost: cannot open `modpost' (No such file or directory) Aug 27 07:40:30 well, would you do 'rm -fr /' if i told you? :p Aug 27 07:41:29 rob_w : no idea :( Aug 27 07:42:01 ynezz, today i wil do anything to get this straight Aug 27 07:42:11 file /data1/home/bob/Handhelds/oe/stuff/au1100-build/tmp/work/mipsel-angstrom-linux/linux-au1100-2.6.19-r1/linux-2.6.19.7/scripts/mod/modpost Aug 27 07:42:33 I might be wrong Aug 27 07:49:03 ynezz, ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), for GNU/Linux 2.6.4, dynamically linked (uses shared libs), for GNU/Linux 2.6.4, not stripped Aug 27 07:49:26 does that help somehow Aug 27 07:49:40 yes, it's host binary and you can strace it Aug 27 07:52:27 try this, cd /data1/home/bob/Handhelds/oe/stuff/au1100-build/tmp/work/mipsel-angstrom-linux/linux-au1100-2.6.19-r1/linux-2.6.19.7 Aug 27 07:52:58 strace scripts/mod/modpost -i /data1/home/bob/Han... Aug 27 07:53:42 yeah have that done already .. let my pastebin it to you .. hang on please Aug 27 07:54:17 http://rafb.net/p/HRgwvr95.html Aug 27 07:55:00 do you think the paths are probably too long ? some long time ago bitbake had some issues i think with such long stuff ? Aug 27 07:55:46 morning Aug 27 07:57:22 Hi! Aug 27 07:57:50 rob_w: no idea, but the path is about 130 chars long so it shouldn't be a problem Aug 27 07:58:23 i can do modpost on some objects inside the madwifi source by hand .. Aug 27 07:58:37 it seems its inside the ath_pci and so Aug 27 07:58:48 let me see which modules work and which not Aug 27 08:00:25 try to remove this from that modpost line -I /data1/home/bob/Handhelds/oe/stuff/au1100-build/tmp/work/au1100-angstrom-linux/madwifi-modules-1.0.0+svn-r0/madwifi/Module.symvers Aug 27 08:01:04 rerun with strace and pastebin again Aug 27 08:04:11 http://rafb.net/p/75rxyT61.html Aug 27 08:05:02 same thing .. some modules work when i call modpost on them along .. som ieee802.11 ones for example but ath_pci breaks Aug 27 08:05:32 so it's just this module causing problems? Aug 27 08:06:45 no there are some .. ath_rate_amrr , ath_rate_minstrel .. but some work Aug 27 08:11:22 damn i dont get it Aug 27 08:12:54 it's some bug Aug 27 08:16:00 can you upload somewhere you vmlinux and ath_pci.o ? Aug 27 08:16:06 s/you/your/ Aug 27 08:17:49 moin Aug 27 08:20:09 morning Aug 27 08:23:32 ynezz, yes hang on Aug 27 08:25:13 rob_w : disable the atheros stuff and use madwifi Aug 27 08:25:28 ? Aug 27 08:25:53 steliosk, what you mean ? i am using madwifi ? Aug 27 08:26:35 Does the error occur when building madwifi or your kernel ? Aug 27 08:26:53 building madwifi .. Aug 27 08:27:34 its possible that madwifi is not build correctly then Aug 27 08:27:52 i thought it was a kernel issue Aug 27 08:28:07 but modpost shouldn't segfault anyway Aug 27 08:28:40 well if wrong binaries from madwifi are used it would Aug 27 08:29:17 oh i thought, that some error message would be better :) Aug 27 08:35:17 morning all Aug 27 08:36:41 ynezz, http://www.linuxdevelopment.de/projects/Sanddune/mipsel/ Aug 27 08:50:12 FIIIIXXXEDD ! Aug 27 08:50:47 thanks to steliosk , i checked the compile options and rebuild the bb file .. now it does the job .. Aug 27 08:52:11 rob_w : nice :) Aug 27 08:52:44 madwifi bb files are a mess .. i tricked them already quite a bit .. Aug 27 08:53:16 now lets hope they load and work Aug 27 08:54:47 rob_w : btw which toolchain/distro you used ? I will be doing some builds for an infenion MIPS32 4Kc Aug 27 08:55:20 i am always using angstrom 2007.1 Aug 27 08:55:27 rob_w : i have tested the madwifi with x85 and ppc. I know some people use them on the slug also Aug 27 08:55:37 s/x86/x86 Aug 27 08:55:49 with the current bb files ?? Aug 27 08:56:13 yes Aug 27 08:58:59 yeah ,, they load and scan !! Aug 27 09:07:40 whoa!!!! 32gig SD cards Aug 27 09:10:27 morning Aug 27 09:10:34 XorA: sdhc by toshiba? Aug 27 09:11:37 hrw: yeah Aug 27 09:12:02 hrw: thats almost as much as harddrivews I just retired from my PC Aug 27 09:12:07 ;) Aug 27 09:12:22 could do an OE build on a zaurus Aug 27 09:12:38 without nfs Aug 27 09:12:40 I gave 17GB hdd for free some days ago - it was not used for more then two years Aug 27 09:12:54 XorA: spitz ones are able to do it now Aug 27 09:13:02 I think I have 15, 40, 60G sitting on shelf Aug 27 09:13:56 good day all Aug 27 09:15:02 hi likewise Aug 27 09:15:08 XorA: 3.5"? Aug 27 09:17:08 hrw: yes Aug 27 09:17:24 hrw: just replaced with 500G SATA Aug 27 09:20:27 I would replace 40G 2.5" in alix Aug 27 09:21:13 2.5" still stupid expensive :-( Aug 27 09:21:23 Id love to replace ps3 2.5 with 500G Aug 27 09:21:52 alix got 2.5" from laptop which got 80g Aug 27 09:22:39 good morning all Aug 27 09:23:01 hi koen Aug 27 09:29:14 that Tongfeng.Yang guy has the right kind of feature requests Aug 27 09:29:21 ones we already implemented :) Aug 27 09:30:14 rob_w: FATAL: section header offset=815726848 in file 'ath_pci.o' is bigger then filesize=153968 Aug 27 09:30:25 i think it's better then segfault Aug 27 09:30:37 although it says nothing :p Aug 27 09:30:49 ynezz, i ve it working .. thanks for your help Aug 27 09:31:42 I wonder where to submit that modpost patch Aug 27 09:33:26 kernel.org probably Aug 27 09:40:17 good morning Aug 27 09:41:00 morning all Aug 27 09:41:01 hi florian Aug 27 09:46:15 hello, all Aug 27 09:47:58 koen: do we have ekiga in OE? Aug 27 09:48:15 we have gnome-meeting Aug 27 09:48:19 I have a question. How to build kernel manually in do_compile_prepend()? Aug 27 09:48:27 ekiga depends on the WORST libraries ever Aug 27 09:48:38 ~curse pwlib and opal Aug 27 09:48:39 May the fleas of a thousand camels infest your most sensitive regions, pwlib and opal ! Aug 27 09:48:47 koen: gnome-meeting? which name of recipe it is? Aug 27 09:49:02 I need copy module in initramfs. Aug 27 09:49:12 hrw: hmm Aug 27 09:49:23 hrw: I fear it's somewhere in bugzilla Aug 27 09:49:35 AMD contributed it, but it didn't crosscompile Aug 27 09:49:54 ok Aug 27 09:50:24 hrw : voip stuff ? Aug 27 09:50:34 steliosk: friend asked me about it Aug 27 09:51:04 hrw : ekiga has issues, but its i think the best option at the moment Aug 27 09:55:42 I'd rather go for something like tapioca or empathy Aug 27 09:56:30 Linux alix 2.6.23-rc3-hrt2 #40 PREEMPT Mon Aug 27 11:42:39 CEST 2007 i586 unknown unknown GNU/Linux Aug 27 10:06:11 koen: still not fixed drupal to output news the right way up? Aug 27 10:12:27 XorA: nope Aug 27 10:13:22 koen: I had this happen on my drupal, but I cant remeber the fix, but it was quite simple Aug 27 10:13:46 koen: I think you can search the answer on drupal.org Aug 27 10:14:26 searching for that on drupal.org is like finding the right page on google by typing 'sex' Aug 27 10:14:45 koen: heh Aug 27 10:15:18 koen: I guess any of the first 10 results on google probably fit for average teenage male :-) Aug 27 10:15:27 rwhitby: ping Aug 27 10:21:13 zecke: gu Aug 27 10:24:50 rwhitby: Do you want to have a tags/bitbake-1.8 in BitBake's svn? We would make this have a copy of the latest stable source... Aug 27 10:24:57 need to get food... sorry bbl Aug 27 10:25:28 koen: where is 'how to add own packages into angstrom feed' doc? Aug 27 10:25:34 zecke: I'm not fussed. The releases are not that frequent, and it's no trouble to update mokomakefile when they happen Aug 27 10:27:16 hrw: somewhere in irc logs Aug 27 10:27:47 hrw: find . -name morgue | xargs rm -rf ; rsync . -av angstrom@linuxtogo.org:/home/angstrom/website/unstable/feed/unsorted/ Aug 27 10:28:26 rob_w: patch has been accepted by kbuild maintainer, but not considered critical since it was first report so it won't be applied in current -rc and in -stable Aug 27 10:28:56 but it will be fixed, thanks for help Aug 27 10:30:08 koen: angstrom@? Aug 27 10:30:44 yes Aug 27 10:31:00 koen: so I need to send you keys Aug 27 10:31:18 that's what I said multiple times on the -devel ml, yes Aug 27 10:32:00 sent Aug 27 10:34:57 added Aug 27 10:34:57 Crofton: ping Aug 27 10:36:04 Jin^eLD: zecke plans to attend EELC in linz Aug 27 10:37:48 cool! Aug 27 10:38:26 koen: btw, any news on bfin? Aug 27 10:42:07 Jin^eLD: sort of Aug 27 10:42:22 the board I have seems only to have linux support with secret ADI patches Aug 27 10:42:28 which I don't have access to Aug 27 10:44:01 03florian 07org.oe.dev * re09f0ed4... 10/ (1 packages/xserver-common/xserver-common_1.17.bb): xserver-common: Add 1.17 Aug 27 10:44:06 03florian 07org.oe.dev * r894bdfc4... 10/ (1 packages/gpephone/vochistory_0.1.bb): vochistory: Add 0.1 Aug 27 10:44:10 03florian 07org.oe.dev * rc65843aa... 10/ (3 files in 3 dirs): gpe-session-scripts-phone: Drop some environment crap. Aug 27 10:44:17 03florian 07org.oe.dev * rff54d40f... 10/ (1 packages/gpephone/libabenabler2_1.0.bb): libabenabler2: Add version 1.0 Aug 27 10:44:23 03florian 07org.oe.dev * r348a3cdc... 10/ (1 packages/gpephone/ptim-engine_svn.bb): ptim-engine: Package additional files Aug 27 10:44:29 03florian 07org.oe.dev * r5281adae... 10/ (1 packages/tasks/task-gpephone.bb): task-gpephone: Add vochistory Aug 27 10:48:44 koen: got mail? Aug 27 10:49:21 koen: I think I have an easy kit and a stamp 533 Aug 27 10:49:27 which one have you got? Aug 27 10:50:20 pong Aug 27 10:50:33 ping Aug 27 10:50:41 Crofton: the TI stuff... Aug 27 10:50:54 I am back from holidays, so I guess I could take a look now Aug 27 10:51:00 anything special I have to take care of? Aug 27 10:56:27 steliosk: did you used award flasher to update alix bios? Aug 27 10:56:58 Jin^eLD, the only real problem is if the davinci git kernel builds :) Aug 27 10:57:26 set machine to davinci-dvevm and see if it build Aug 27 11:05:34 Jin^eLD: I have the bf538f ez kit: http://www.flickr.com/photos/koenkooi/1192511235/ Aug 27 11:10:12 Crofton: ok, will try :) Aug 27 11:10:54 koen: aha.. well does not tell me much since I did not try anything with those boards yet; I'd have to see if I need any secret crap for the versions that I have Aug 27 11:11:18 the stamp boards have linux support Aug 27 11:11:38 koen: without the secret crap? Aug 27 11:11:44 yes Aug 27 11:11:50 so I could test what you did? or is it not yet in a testable state? Aug 27 11:12:03 not yet Aug 27 11:12:07 ok Aug 27 11:12:20 03koen 07org.oe.dev * r51d9e57d... 10/ (31 files in 3 dirs): Aug 27 11:12:20 gcc 4.2.1: add extra patches for uclibc build Aug 27 11:12:20 * uclibc still dislike 4.2.1, but it fails in a different place now Aug 27 11:12:24 well, I'll have my hands full with the davinci board for the start Aug 27 11:12:24 03koen 07org.oe.dev * rfe475621... 10/ (23 files in 2 dirs): gcc: drop lock files Aug 27 11:12:58 Crofton|home and khem have the davince stuff covered :) Aug 27 11:13:05 :) Aug 27 11:13:50 * Crofton|home is getting to hate working on stuff blind :) Aug 27 11:14:13 Jin^eLD, I have a set of images you can test, but the git kernel has been updated recently Aug 27 11:14:29 the problem with git is not many people contribute Aug 27 11:15:01 does anyone know if microperl is in OE? Aug 27 11:16:02 it isn't Aug 27 11:16:10 we do have micro-emacs Aug 27 11:16:51 sakoman's product uses microperl apparently Aug 27 11:17:21 apparently we are less bitrotted tan build root though :) Aug 27 11:17:30 heh Aug 27 11:17:43 I recommend cherokee over boa Aug 27 11:19:17 hwta re they ? Aug 27 11:22:39 webservers Aug 27 11:26:42 03koen 07org.oe.dev * r26b8940c... 10/ (5 files in 2 dirs): angstrom: start adding support for building .debs and .rpms instead of .ipk packages Aug 27 11:27:26 koen: 2007.1 to debs? Aug 27 11:28:01 urg u-boot git is failing for evm Aug 27 11:28:20 hrw: dunno, that's for the release and/or core team to decide Aug 27 11:28:36 and it needs bits from poky to work :) Aug 27 11:28:45 in fetch? (u-boot.git) Aug 27 11:37:08 time for "rsync . -av angstrom@linuxtogo.org:/home/angstrom/website/unstable/feed/unsorted/" Aug 27 11:38:12 so OE will use .deb ? Aug 27 11:38:42 in a future I mean Aug 27 11:38:50 ynezz: OE already can generate deb Aug 27 11:39:07 so it's possible to use apt? Aug 27 11:39:10 nice :) Aug 27 11:40:04 I do not know what people find so great in apt on small devices.. Aug 27 11:41:05 ynezz: so maybe you will tell me? Aug 27 11:44:33 any ideas on how to debug a failure by the git fetcher? Aug 27 11:53:11 03hrw 07org.oe.dev * r11e7fd7a... 10/ (3 files in 2 dirs): lmsensors-apps: upgrade to 2.10.4, build also sensors app Aug 27 11:53:15 03hrw 07org.oe.dev * re9851138... 10/ (3 files in 2 dirs): disapproval of revision '11e7fd7a1d2b01efd7be20943672ee11977561e2' Aug 27 11:53:20 03hrw 07org.oe.dev * r8cd13341... 10/ (1 packages/lm_sensors/lmsensors-apps_2.10.1.bb): Aug 27 11:53:20 lmsensors-apps: totally changed packaging + build of sensors application Aug 27 11:53:20 lmsensors-apps contain all utils which do not need libsensors Aug 27 11:53:20 lmsensors-sensors contain 'sensors' application + config (it depend on libsensors) Aug 27 11:55:36 hrw: since people starting putting multible gigabytes of storage in small devices it became less absurd Aug 27 11:56:29 hmm.. putting my feeds into ltg was not best idea Aug 27 11:56:34 its too much of it Aug 27 11:58:05 * koen remembers that steliosk's scripts need to get integrated with feedbrowser Aug 27 12:01:43 45606 100% 33.89kB/s 0:00:01 (xfer#901, to-check=3698/5235) Aug 27 12:01:59 hrw: i just like apt, that's it Aug 27 12:02:00 koen: looks like lot of pacages will be to sort out Aug 27 12:05:42 aargh Aug 27 12:05:50 the uni partitioned the internet *again* Aug 27 12:05:57 no oe builds for me today :( Aug 27 12:06:09 you need some competent admins Aug 27 12:06:15 or is this deliberate? Aug 27 12:06:19 dunno Aug 27 12:06:57 ot knocked out ebay, so I guess people high will complain :) Aug 27 12:07:06 heh Aug 27 12:08:31 Hum. Some (outdated) arm5vte packages are found via the feed browser, but not in http://www.angstrom-distribution.org/unstable/feed/unsorted/armv5te/ Aug 27 12:08:43 is that normal ? Aug 27 12:09:04 I'm thinking about e17 libs for instance Aug 27 12:09:27 the versions in the feed browser are 1 month old Aug 27 12:10:47 but the current versions build... (but there is a problem with the Package file as the versionning as changed, the old version gets in the package file and the new into morgue ;) ) Aug 27 12:11:51 is this because of the version sorting problem ? or do I simply not understand how packages are built / push into officials repo ? Aug 27 12:23:57 good morning! Aug 27 12:24:08 hey sakoman Aug 27 12:24:17 hey! Aug 27 12:24:30 Finally making some progress :-) Aug 27 12:24:36 * koen reads up on microperl Aug 27 12:25:13 http://search.cpan.org/src/NWCLARK/perl-5.8.8/README.micro Aug 27 12:25:22 Good for embedded devices that have web UI's Aug 27 12:25:40 I'd write webUIs in php instead of perl :) Aug 27 12:25:51 have you looked at recplacing boa with cherokee? Aug 27 12:26:08 regarding webUI in perl... mason is neat. Aug 27 12:26:28 Yes, it was one of the things that led me to OE Aug 27 12:26:46 No cherokee in buildroot Aug 27 12:26:55 i'm considering to use xavante(light httpd in lua) using luacgi Aug 27 12:27:03 for webui Aug 27 12:28:16 I'm biased because I spent some time with alvaro :) Aug 27 12:28:30 Tried lighttpd, it ended up being much bigger than boa Aug 27 12:29:46 tough keeping the image small enough! Aug 27 12:29:50 spot the cherokee devel: http://patux.net/blog/index.php?entry=entry051123-143507 Aug 27 12:31:16 nice pic, don't see any tacos though :-) Aug 27 12:31:46 we ate them all :) Aug 27 12:32:53 hew : can't remember what i used for the update :( Aug 27 12:33:00 * koen tries http://rafb.net/p/MvAY1Q26.html Aug 27 12:33:38 koen : what kind of integration ? Aug 27 12:34:37 ehm, what are you referring to? Aug 27 12:34:53 * koen is a teensy bit chaotic today Aug 27 12:35:33 koen remembers that steliosk's scripts need to get integrated with feedbrowser Aug 27 12:35:36 Tried angstrom-console-image (uclibc) --- build failed on openssl Aug 27 12:35:55 is it hopeless for a uclibc build? Aug 27 12:36:26 I keep forgetting you are trying to use uclibc .... Aug 27 12:36:31 steliosk: right, you sent me some updated sorting scripts that I still haven't bothered to fix and integrate Aug 27 12:36:44 Image size matters :-) Aug 27 12:36:51 I've set 'PREFERRED_PROVIDER_virtual/kernel="linux-rp"' in my local.conf. Why is the "linux_2.6.21.bb" unpacked and compiled? Why not the preferred one? Aug 27 12:36:55 ah Aug 27 12:37:05 ~lart python for building openssl Aug 27 12:37:06 * ibot beats python to within 2.54cm of his life for building openssl Aug 27 12:38:05 angstrom-minimak(glibc) bots fine on the OSK Aug 27 12:38:09 as of this mroning :) Aug 27 12:38:21 sysfa mount missing Aug 27 12:38:38 and /dev/initctl is not a fifo Aug 27 12:39:27 sounds like the same 2 I get with uclibc Aug 27 12:40:44 any theories yet on the mount issue? Aug 27 12:41:09 strange you don't get it -- happens with libc and uclibc on gumstix Aug 27 12:41:13 I added mount from util-linux to angstrom minimal image :) Aug 27 12:41:56 Same here, mount still fails on size Aug 27 12:42:07 I think my next OE goal is to understand how images are assmebeld from tasks Aug 27 12:43:04 At some point, I would nuke tmp and start a clean build Aug 27 12:43:09 just in case Aug 27 12:43:14 Crofton|home: it is easy too Aug 27 12:43:31 doing it as we speak, great minds Aug 27 12:43:33 hrw, I am sure it is, I just need to spend some time understanding :) Aug 27 12:43:36 heh Aug 27 12:43:44 I like to do it b4 bedtime Aug 27 12:43:50 or when I need to do something else :) Aug 27 12:43:56 I suspect the gumstux kernel configs needs some tweaking to work with an OE userspace Aug 27 12:44:03 yeah Aug 27 12:44:06 e.g tmpfs and friends Aug 27 12:44:14 this is why we need some gusmtix :) Aug 27 12:44:25 I'm trying! Aug 27 12:44:30 it seems to have tmpfs turned on Aug 27 12:44:48 tmp Aug 27 12:44:58 ignore that :-) Aug 27 12:46:35 http://ewi546.ewi.utwente.nl/tinderbox/showbuilds.pl?tree=Angstrom Aug 27 12:46:41 is very confusing Aug 27 12:50:30 koen : We are working on getting a menu based system to setup the feeds Aug 27 12:50:56 Crofton: btw, here are my gumstix diffs - http://www.sakoman.net/oe/gumstix-diffs.txt Aug 27 12:51:41 Just a few tweaks to move the verdex/connex specific stuff into the right files Aug 27 12:51:49 ok Aug 27 12:52:06 Crofton|home : Well you have to find out which one is your build in the tinderbox Aug 27 12:52:20 Hmm, also did a base-files/gumstix/fstab Aug 27 12:52:38 steliosk, exactly, the page is huge and it is hard to understand :) Aug 27 12:53:12 fosdem airfare is $612 atm Aug 27 12:53:22 well a crash course on tinderbox.... Aug 27 12:53:26 gumstix-verdex-glibc-build-image Linux 2.6.20-1.2962.fc6 Aug 27 12:53:38 is there something wrong with fetching from svn in 1.8.8? I get a very weird error Aug 27 12:54:09 means is a gunstix-verdex glibc build on a host machine with a 2.6.20-1 (probably fedora 6) kernel Aug 27 12:54:13 tinderbox? Aug 27 12:54:38 * steliosk wonders if zecke is abot waking up on certain keywords Aug 27 12:54:41 steliosk, I get the basics Aug 27 12:54:44 heh Aug 27 12:54:52 steliosk, you ok with the fire? Aug 27 12:54:57 steliosk: yes, I'm interrupt driven Aug 27 12:55:00 when fetching it added some /RCS/mypackage,v stuff to the svn url Aug 27 12:55:06 and that failed, what is that? Aug 27 12:55:12 ,v is cvs Aug 27 12:55:16 or rather rcs Aug 27 12:55:19 heh Aug 27 12:55:23 I was going to point to rcs Aug 27 12:55:29 green == good, yellow == build in progress, red == bad Aug 27 12:55:31 why is it trying that then, my SRC_URI says svn:// Aug 27 12:55:38 whaty about black? Aug 27 12:55:44 Crofton: gumstix fstab is at http://www.sakoman.net/oe/fstab.txt Aug 27 12:55:45 and it did work with 1.6 Aug 27 12:55:48 but if you have green/red difficulties this is a bad indicator and I agree Aug 27 12:56:00 Crofton|home : Well not ok. We have a few hundrend square kilometers that were not burned, but our goverment makes sure that there are Aug 27 12:56:17 black in the header == last build vanished... (probably) Aug 27 12:56:18 they hate half finished jobs Aug 27 12:56:28 sometimes I think government is the root of all evil ... Aug 27 12:56:43 can anyone help me with the "kernel version" seletion of openembedded. Aug 27 12:56:48 steliosk: was the perl code of any use to you? Aug 27 12:57:05 it's like my preferred_provider is ignored. Aug 27 12:57:33 Crofton|home : i was down at the fires during the weekend, as my wife's family owns some land there (and i am a Red cross volunteer also) Aug 27 12:58:02 Crofton|home, steliosk: I just pushed the tinderclient and suppuku params into angstrom.inc Aug 27 12:58:19 ok Aug 27 12:58:20 INHERIT += "tinderclient" in local.conf is now all people need Aug 27 12:58:24 zecke : I was hoping to get some free time during the weekend but it did not happen Aug 27 12:58:27 ok Aug 27 12:58:30 for seppuky they need user and password as well Aug 27 12:58:57 so we can remove the tinderclient stuff, but leave seepp.... Aug 27 12:59:22 sakoman wanders off to get coffee while he does a clean build Aug 27 12:59:37 ughh . . . 6am is too early Aug 27 12:59:47 heh Aug 27 12:59:49 koen : hmm will try to test it, this week Aug 27 13:00:06 I love working with people to the west of me Aug 27 13:00:10 hrhr. 3pm over here ;) Much better time for a coffee, erm build! Aug 27 13:00:25 sakoman : For clean builds you have enought time to go and get the coffee from the trees Aug 27 13:01:26 koen, did you look at what sakoman has done to the gumstix configs? Aug 27 13:01:59 Shoudl I set the common machine features in .inc Aug 27 13:02:12 can anyone help me with the "kernel version" selection feature of oe? my preferred_provider is ignored / not used. Aug 27 13:02:13 cyberdeck : Working with xilinx ? Aug 27 13:02:15 and += the usb-host/gedget in teh machine spcific files? Aug 27 13:02:27 steliosk: yes. with the xupv2p board Aug 27 13:02:42 Crofton|home: it's pretty much cosmetics Aug 27 13:02:47 yeah Aug 27 13:03:06 * koen cycles to the photo dude to pick up the negs Aug 27 13:03:07 but if we are using a .inc file Aug 27 13:03:09 later all Aug 27 13:03:10 cyberdeck : Haven't tested on that yet. Aug 27 13:03:12 bye Aug 27 13:03:13 steliosk: i tried to simply duplicate (with some modifications) your ml403 config. Aug 27 13:03:45 and ? Aug 27 13:04:02 steliosk: building a helloworld-image succeded. But a "bitbake virtual/kernel" always try to build the "normal" linux_2.6.21 one. Aug 27 13:04:47 steliosk: but it should try to build the linux-xupv2p-slab-2.6.x_git one Aug 27 13:05:36 cyberdeck : R u using Angstrom ? Aug 27 13:05:42 yes. Aug 27 13:05:58 you need to override PREFERRED_KERNEL in your local.conf then Aug 27 13:06:30 or add the xilinx kernel to kernel.inc ;) Aug 27 13:06:51 i know a lot of people would prefer the second, as a lot of kittens will be saved Aug 27 13:07:08 :) Aug 27 13:07:09 03florian 07org.oe.dev * rcbccca81... 10/ (3 files in 3 dirs): fbreader: Get rid of iconv crap Aug 27 13:07:15 03florian 07org.oe.dev * rb23b25b1... 10/ (7 files in 2 dirs): xserver-common: Add 1.18, drop some older versions. Aug 27 13:07:17 duplication bad Aug 27 13:07:20 03koen 07org.oe.dev * rfd79f1d3... 10/ (1 conf/distro/include/angstrom.inc): angstrom: add sane defaults for tinderclient and seppuku Aug 27 13:07:23 yeah i would even prefer the second one. Aug 27 13:07:25 03hrw 07org.oe.dev * r98112d31... 10/ (7 files in 3 dirs): portmap: added 6.0 from Poky Aug 27 13:07:32 03hrw 07org.oe.dev * r2f5bd4fb... 10/ (5 files in 3 dirs): alsa: added 1.0.14 from Poky Aug 27 13:07:42 03koen 07org.oe.dev * rab56ad72... 10/ (1 conf/distro/angstrom-2007.1.conf): angstrom: add preferred provider for avahi Aug 27 13:07:48 03koen 07org.oe.dev * r1adc2610... 10/ (3 files in 2 dirs): avahi: attempt #1 to stop it building python by default Aug 27 13:07:52 03koen 07org.oe.dev * r1be7f2bf... 10/ (5 files in 2 dirs): machine tune files: add FEED_ARCH to the usual suspects to make out-of-tree machines work Aug 27 13:09:08 * koen really gone now Aug 27 13:10:45 03hrw 07org.oe.dev * r783abb45... 10/ (4 files in 3 dirs): Aug 27 13:10:45 linux: switched ALIX to 2.6.23-r3, added MFGPT patches, changed config Aug 27 13:10:45 - USB Host/HID compiled in (to get keyboard working after boot) Aug 27 13:10:45 - VIA Rhine compiled in (to get networking after boot) Aug 27 13:10:45 - netfilter enabled Aug 27 13:10:46 - more Geode stuff enabled (gpio, framebuffer) Aug 27 13:14:53 steliosk: PREFERRED_KERNEL="linux-xupv2p-slab-2.6.x" did not work. did i miss something? Aug 27 13:15:48 yes Aug 27 13:16:03 cyberdeck: PREFERRED_PROVIDER_virtual/kernel = "linux-xupv2p-slab-2.6.x" Aug 27 13:16:56 hi Aug 27 13:21:58 hrw: didn't work. a "bitbake virtual/kernel" results in unpacking linux_2.6.21.bb. Aug 27 13:23:09 cyberdeck: bitbake -g virtual/kernel and take a look at the graph Aug 27 13:23:26 cyberdeck: and if you spelled preferred correctly bitbake will tell you it can't satisfy your wish Aug 27 13:34:13 cbrake: I have an image booted now. I did build a release image and forgot about it. is there a default login that you know of ? Aug 27 13:34:24 anyone understand why I have a line in fstab for OSK trying to mount sysfs on /sysfs Aug 27 13:34:27 root root? Aug 27 13:34:40 but, the machine when running has already mounted sysfs on /sys? Aug 27 13:34:50 is this a typo I should fix? Aug 27 13:34:54 crofton not really Aug 27 13:34:59 fix it yes Aug 27 13:35:07 well, we get an error booting up :) Aug 27 13:35:09 ok Aug 27 13:35:24 the error is propagating :) Aug 27 13:37:26 zecke: hmm. even the graphviz stuff shows "linux_2.6.21.bb" for the kernel. Aug 27 13:37:53 cyberdeck: so you misspelled something, or it gets set a later point again? Aug 27 13:38:53 cyberdeck try bitbake bla -e Aug 27 13:39:08 and look out whats set for kernel Aug 27 13:40:13 woglinde: a "bitbake virtual/kernel -e | grep kernel" gives a friendly 'PREFERRED_PROVIDER_virtual/kernel="linux-xupv2p-slab-2.6.x"'. Aug 27 13:40:29 cyberdeck: and you have such recipe somewhere? Aug 27 13:40:42 args Aug 27 13:40:47 yes now I see it Aug 27 13:40:57 linux-xupv2p-slab-2.6.x is really bad choosen name Aug 27 13:41:53 ohh. steliosk name some of the files and i just copied them over. Aug 27 13:42:11 what would be a correct name? Aug 27 13:43:07 correct would be that you provide patches and integrate it in the 2.6.21.bb Aug 27 13:45:16 woglinde, sometimes we need time to figure out what the correct patches are :) Aug 27 13:45:53 crofton *g* Aug 27 13:46:27 devpts /dev/pts devpts gid=5,mode=620 0 0 Aug 27 13:46:39 is in the fstab for gumstix, should I also do this in the OSK? Aug 27 13:46:50 or does it depend on kernel config? Aug 27 13:47:35 woglinde: but collie and other devices uses a virtual/kernel = "linux-rp"? isn't this "allowed anymore? Aug 27 13:48:10 cyberdeck, intelectually speaking, we all agree we need fewer kernel configs Aug 27 13:48:19 to save kittens Aug 27 13:48:20 woglinde: and don't be afraid that i'll push something like "linux-xupv2p-slab-2.6.x" upstream. it's only a test image. Aug 27 13:48:21 cyberdeck rp provides a lot of patches Aug 27 13:48:34 and historical has his tree in Aug 27 13:48:35 Crofton|home: yes. Aug 27 13:49:00 sakoman back with mug of freshly roasted coffee Aug 27 13:49:31 didn't pick the berries, but did roast them Aug 27 13:49:37 cybderdeck try to rename it to linux-test_2.6.21.bb and try aganin Aug 27 13:49:53 hm, roast them in the popcorn popper? Aug 27 13:50:07 sakoman, I am going to the fstab now Aug 27 13:50:15 no, have a hottop roaster -- http://www.sweetmarias.com Aug 27 13:50:35 woglinde: i setup "virtual-kernel" to "linux-rp" to figure out if my build setup is correct. at least it should use "linux-rp-2.6.21.bb". Aug 27 13:50:46 the guys in the lab use the popcorn popper method I think Aug 27 13:51:27 I needed to do more than 3 oz at a time :-) Aug 27 13:51:38 my wife & I are coffee addicts Aug 27 13:51:38 but I think they get the beans from there Aug 27 13:51:50 Nothing like fresh roast! Aug 27 13:52:14 Any issues with the patches? Aug 27 13:52:15 why didntou put the tmpfs on /dev? Aug 27 13:52:28 woglinde: hmmm. even a 'PREFERRED_PROVIDER_virtual/kernel="linux-rp"' did _not_ work. any other ideas? Aug 27 13:53:06 cyberdeck hm try bitbake bla -D Aug 27 13:53:06 cyberdeck: you set it in local.conf or machine config? Aug 27 13:53:22 hrw: i set it at the local.conf Aug 27 13:53:33 cyberdeck: then check does machine set it too Aug 27 13:53:34 cyberdeck: and your machine is not changing it? Aug 27 13:53:46 cyberdeck: or just see above for the same question... Aug 27 13:54:36 zecke: my machine.conf (selfwritten) did not change it. Aug 27 13:54:38 hmm . . . good question Aug 27 13:54:44 heh Aug 27 13:55:03 I am going to make the OSK fstab look like the gumstix one, if possible Aug 27 13:55:49 was doing a merge of gumstix buildroot fstab, the generic OE, and the OMAP5912OSK Aug 27 13:55:54 late at night Aug 27 13:56:01 Jin^eLD, compiliing kernel for dvevm now Aug 27 13:56:07 ok Aug 27 13:56:18 I think it was because the orginal gumstix didn't have that -- checking Aug 27 13:56:21 yeah Aug 27 13:56:29 I suspect they do use udev Aug 27 13:56:34 er do not Aug 27 13:56:35 Crofton: I did not yet start, still at work Aug 27 13:56:56 They use udev-101 Aug 27 13:57:05 hmm Aug 27 13:57:37 well, I suspect it is personal taste whether to create dev nodes on ram, or flash :) Aug 27 13:57:45 Yep, no tmpfs mount for /dev in the gumstix buildroot fstab Aug 27 13:58:16 Shouldn't make a big differenc should it? Aug 27 13:58:27 not sure Aug 27 13:58:29 woglinde: i think i spot the problem. "bb -D" says "DEBUG: providers for virtual/kernel are: ['linux']". Is this the problem? Aug 27 13:58:40 I'll will commit /dev on tmpfs Aug 27 13:59:29 sounds fine to me, I'll give it a try here Aug 27 13:59:40 cbyerdeck yes Aug 27 13:59:41 as soon as the build completes Aug 27 14:00:10 cyberdeck but its hard to figure out where it cames from Aug 27 14:02:23 woglinde: I suspect the "angstrom-2007.1.conf" or one of the includes in there. Aug 27 14:02:54 cyberdeck: what about pastebin local.conf, machine config and kernel recipe? Aug 27 14:05:38 hrw: http://pastebin.ca/672327 for the confs. a specific kernel receipe didn't exist yet. i just copied over one from steliosk's ppc405 kernels. Aug 27 14:06:14 woglinde, zecke, hrw, steliosk: and thank you all for ur great support ! Aug 27 14:07:09 cyberdeck : Don't worry will send the bill later :) Aug 27 14:07:36 steliosk: ok. please send it to devnull@secretlab.de ;) Aug 27 14:07:41 cyberdeck: then look at COMPATIBLE_MACHINE in kernel recipe Aug 27 14:07:52 hrw hm Aug 27 14:07:58 right Aug 27 14:08:22 bye I am heading home Aug 27 14:10:37 hrw: omg. thank you! i didn't changed that line. ah crap! Aug 27 14:11:41 cyberdeck : devnull works there also ? this guy is sneaky :) Aug 27 14:19:57 hrw: you pinged earlier? Aug 27 14:20:07 Crofton: removed the sysfs line and added the tmpfs /dev mount on gumstix Aug 27 14:20:16 no sysfs error on boot Aug 27 14:20:26 no other apparent side effects Aug 27 14:20:33 CosmicPenguin: ah, yes Aug 27 14:20:41 ok, I will push something like that Aug 27 14:20:51 CosmicPenguin: alix lack free mfgpt ;( Aug 27 14:21:09 Clean build still going, perhaps I'll go plant some coffee trees :-) Aug 27 14:21:26 CosmicPenguin: and one more question - why under linux geode lx is seen as 'i586' when it is rather 'i686' cpu? Aug 27 14:21:27 hrw: ohhh - yeah Aug 27 14:21:44 hrw: there is a way around that, but you have to be careful Aug 27 14:22:07 The Geode is officially a 686 Aug 27 14:22:12 oops Aug 27 14:22:13 sorry Aug 27 14:22:14 586 Aug 27 14:22:45 CosmicPenguin: until geode I was thinking that i686 == "pentium with mmx" or better Aug 27 14:23:04 It turns out there are some native i686 opcodes Aug 27 14:23:23 CosmicPenguin: but better to build for i586? Aug 27 14:23:30 We ran into a problem with that last year with OLPC - a new version of gcc started using the native opcodes for memory management, and it blew up the Geode bad Aug 27 14:23:49 That depends on which version of gcc you have Aug 27 14:23:57 CosmicPenguin: amd does not plan to add -mcpu=geode into gcc4? Aug 27 14:23:57 mainline has a dedicated geode target Aug 27 14:24:01 CosmicPenguin: 4.1.x Aug 27 14:24:23 CosmicPenguin: gcc manual does not list geode... Aug 27 14:24:32 Trust me - its in there Aug 27 14:24:34 for mainline Aug 27 14:24:40 I'm pretty sure 4.1 doesn't have it Aug 27 14:25:13 neither 4.2.1 according to docs http://gcc.gnu.org/onlinedocs/gcc-4.2.1/gcc/i386-and-x86_002d64-Options.html#i386-and-x86_002d64-Options Aug 27 14:25:33 yeah - because GNU always keeps their documents up to date Aug 27 14:25:51 CosmicPenguin: gcc 4.3 got geode Aug 27 14:26:01 http://www.gnu.org/software/gcc/gcc-4.3/changes.html Aug 27 14:26:07 * CosmicPenguin shrugs Aug 27 14:26:15 My akita's training the battery while suspended Aug 27 14:26:18 whatever - I don't think very much of -mcpu=geode, so I haven't kept up to date on it Aug 27 14:27:08 CosmicPenguin: I plan to make usage of alix so prefer to know which way will be better for it ;) Aug 27 14:27:10 s/training/draining Aug 27 14:27:27 eekee: check bugtracker and add note/bug Aug 27 14:27:28 hrw: your best options for 4.1 will be k6 or k6-2 Aug 27 14:27:34 k Aug 27 14:27:40 CosmicPenguin: just like I was thinking Aug 27 14:28:00 PACKAGE_ARCH = "geode" sounds even good Aug 27 14:28:16 pentium-mmx isn't bad either, IIRC Aug 27 14:29:13 CosmicPenguin: pentium-mmx will ignore mmxext and 3dnow stuff Aug 27 14:29:19 oh, it will? Aug 27 14:29:24 thats bad Aug 27 14:29:29 I think Aug 27 14:29:58 atlist 3dnow will be ignored Aug 27 14:30:49 As for the MFGPTs here is the story Aug 27 14:31:11 the BIOS likes to use some of them for a timer - and many BIOS vendors just allocate all the timers for themselves, even if they don't use them Aug 27 14:31:36 Which is bad for us, because MFGPTs are program-once timers, so once they are configured, they are useless to us Aug 27 14:32:10 There is a scret way to unlock the MFGPT and erase the configuration, but before you use it, you have to be positive you pick a MFGPT that the BIOS vendor isn't actively using Aug 27 14:32:11 ;( Aug 27 14:33:28 I would also recommend contacting alix and telling them that you would like to use dynamic ticks on their board, but their BIOS is preventing it Aug 27 14:33:39 morning Aug 27 14:34:57 CosmicPenguin: I have draft of mail to pc engines (alix vendor) already Aug 27 14:35:08 CosmicPenguin: and also have to update bios Aug 27 14:35:19 Good Aug 27 14:36:10 CosmicPenguin: thx for technical infos Aug 27 14:36:16 not a problem Aug 27 14:40:17 steliosk: alive? Aug 27 14:42:14 sakoman, why did you remove the MACHINE_EXTRA_RRECOMMENDS? Aug 27 14:42:49 They were just a remnant of us trying to get /lib/modules populated Aug 27 14:42:54 ok Aug 27 14:43:11 Trying to reduce junk dna :-) Aug 27 14:44:44 yeah Aug 27 14:45:05 I have some juk that is still getting copied Aug 27 14:45:23 that should be discussed at OEDEM at least briefly Aug 27 14:45:39 sent Aug 27 14:46:41 koen: ping Aug 27 14:46:49 zecke: pong Aug 27 14:47:15 koen: do you have an opinion on my ELC 'proposal'? Aug 27 14:47:55 koen: it is not a pure OE talk, but I'm happy to mention OE on every second slide Aug 27 14:47:55 Jin^eLD, I will upload some current evm images in a bit Aug 27 14:48:05 koen: and when you are back to life - I finished uploading of packages. Aug 27 14:48:22 hrw : barely Aug 27 14:48:27 Crofton: you mean precompiled? Aug 27 14:48:29 cool Aug 27 14:48:36 steliosk: did you updated bios on alix? Aug 27 14:48:51 yeah Aug 27 14:49:04 i just ran a build to see if it built Aug 27 14:49:18 the kernel comes from git Aug 27 14:49:22 and so does u-boot Aug 27 14:49:37 is that good or bad? :) Aug 27 14:49:55 can't say, no evm :) Aug 27 14:50:00 hehe Aug 27 14:50:20 I'll see soon enough Aug 27 14:50:32 just have to finish some work first Aug 27 14:50:45 zecke: your proposal looks good to me Aug 27 14:51:15 http://rafb.net/p/9PpNIO18.html Aug 27 14:51:29 sakoman, I am going to run a build with this patch Aug 27 14:51:36 and see what happens Aug 27 14:51:39 then push Aug 27 14:51:49 Good luck :-) Aug 27 14:53:19 Crofton|home: add /sys/ to base-files/base-files/fstab so you will be able to drop gumstix*/fstab Aug 27 14:54:23 hrw, let us see if the gumstix fstab stays like this after do some more tests Aug 27 14:54:39 then when can try the default fstab Aug 27 14:54:53 ok Aug 27 14:55:07 in the back of my mind, I am thinking the same thing Aug 27 14:55:17 especially as teh OSK looks similar Aug 27 14:56:40 Crofton: I've started a list of outstanding gumstix issues/workarounds at http://www.sakoman.net/oe/gumstix-issues.txt Aug 27 14:58:00 steliosk: and if you did then which tool you used? standard award flasher? Aug 27 14:58:50 Crofton: Clean build looks much better :-) Aug 27 14:59:17 ok, running my build for the gumstix+osk Aug 27 14:59:25 /lib/modules is getting populated now! Aug 27 15:00:01 sakoman: for alignment you had a kernel patch to set it iirc.. and if you want to make it in clean OE way you need to create directory 'gumstix-connex' in packages/initscripts/initscripts/ with alignment.sh taken from arm/ dir Aug 27 15:00:22 * hrw -> afk Aug 27 15:00:31 created a gumstix-minimal-image since I couldn't get angstrom-console-image to build with uclibc Aug 27 15:00:32 hrw : i did. i think i got the utility from the pcengines site Aug 27 15:01:09 sakoman: console-image should now build with uclibc Aug 27 15:01:30 koen: Great! Thanks :-) Aug 27 15:02:01 steliosk: used it under msdos or freedos? their 'sb.com' reboots machine here under freedos Aug 27 15:02:26 hrw: Before I do that I want to figure out why the alignment errors with OE when I don't get them with buildroot Aug 27 15:03:36 The patch is just a workaround to quiet the noise while I get other stuff working :-) Aug 27 15:08:49 hrw : i had an msdos CF and used that, can look to see if i still have it and send you an image. its 16MB Aug 27 15:11:24 steliosk: would be great Aug 27 15:13:24 Crofton: I got the board ready... where is the image? Aug 27 15:13:53 Jin^eLD, uploading now Aug 27 15:14:02 the ext2 image is stupidly large Aug 27 15:14:13 no tar.gz for nfs? Aug 27 15:14:45 but you did not reproduce their 1GB rootfs with rpm and stuff like that? Aug 27 15:14:58 hang on Aug 27 15:15:05 that was really kind of huge :) Aug 27 15:15:36 http://ossie.wireless.vt.edu/~balister/davinci_dvevm/ Aug 27 15:15:48 the new builds are uploading now Aug 27 15:15:52 the old one is there Aug 27 15:16:12 I think the ext2 fs is created with a size of 64M Aug 27 15:16:24 I should disable its creation .... Aug 27 15:16:41 bzip2 it before uploading Aug 27 15:16:44 what about the .tar that you have there? Aug 27 15:16:49 koen, I am lazy Aug 27 15:17:05 that should be the rootfs suitable for nfs Aug 27 15:17:07 actually you have quite a number of images Aug 27 15:17:25 which one do I pick? Angstrom-minimalist-image-glibc-test-20070812-davinci_dvevm.rootfs.tar ? Aug 27 15:17:28 bitbake angstrom-minimal-image :) Aug 27 15:17:37 that would be a start Aug 27 15:17:45 Crofton: and 'tar.bz2' is a better choice that 'tar' :) Aug 27 15:17:45 the Aug 12 stuff did boot Aug 27 15:17:53 (OE will do it automagically) Aug 27 15:17:54 not sure about the uboot iamge Aug 27 15:17:59 heh Aug 27 15:18:04 I can't yet, have to finish work first before I can build an image myself Aug 27 15:18:11 well I will not reflash Aug 27 15:18:12 * Crofton knows he needs to go over local.conf Aug 27 15:18:43 the kernel will need to match the fs Aug 27 15:19:05 yes, but not the bootloader Aug 27 15:19:13 ok, I need to go look for the lunch crowd Aug 27 15:19:16 I get the kernel via tftp Aug 27 15:19:20 ok Aug 27 15:19:39 let me know how it goes, this afternoon I am doing non OE work Aug 27 15:19:53 ok I will Aug 27 15:19:59 other than keeping an eye on some builds Aug 27 15:20:01 l8r Aug 27 15:50:40 hmm, I have a package, a library, say, "foo", it is being build and libfoo is installed to my root in do_rootfs, but in the end I still get an error "Cannot satisfy the following dependencies for myimage-base-depends: foo" allthough foo is not even listed there (tried listing it, same result) Aug 27 15:50:47 would could possibly lead to something like that Aug 27 15:50:51 everything else seems normal Aug 27 16:06:33 RDEPENDING on a lib shouldn't be needed Aug 27 16:06:46 since you link against it, OE will include it automatically Aug 27 16:07:45 ok.. I removed it from the RDEPENDS list and did a -c rebuild my task Aug 27 16:07:49 but it still complains Aug 27 16:07:52 do I have to clean up more? Aug 27 16:10:18 it is there, its even not rebuilding it again because it finds it (did the -DDD run) Aug 27 16:10:47 and it is installing the ipk durint the do_rootfs process, so why does it crap out in the end? what else could cause it? Aug 27 16:14:27 koen: I should be able to send you a "working" board soon. Should have the ok to order tomorrow :) Aug 27 16:14:38 cool Aug 27 16:15:08 I was about to install visual dsp++ to see if it improved the past few years Aug 27 16:15:23 koen: heh, never used it Aug 27 16:15:45 I did for a mechatronics project in my second year Aug 27 16:16:10 * lrg wonders if it works under wine Aug 27 16:16:14 design a control loop, click "generate C" and visual DSP would flash it Aug 27 16:16:25 (of course it didn't work that way) Aug 27 16:20:44 * koen thinks happy thought and escape from entering rant mode Aug 27 16:20:51 doh... I am really stuck here; the only thing I see in the log is ERROR: Cannot satisfy the following dependencies for my-demo-base-depends: foo Aug 27 16:21:09 and no further indications... and its not even listed in RDEPENDS anymore Aug 27 16:21:16 did you bump PR on your task? Aug 27 16:21:16 any hints where else to look? Aug 27 16:35:05 Crofton: I guess the kernel was built without nfs support Aug 27 16:35:09 argh Aug 27 16:35:12 ok Aug 27 16:35:20 let me look at updating defconfig Aug 27 16:35:33 I'll build my own version once I finish work Aug 27 16:35:38 hmm spacemonkey ran root over nfs? Aug 27 16:35:41 ok Aug 27 16:35:52 I'd anyway want to reproduce your build Aug 27 16:35:59 have to fix my own first Aug 27 16:36:00 :) Aug 27 16:38:35 the files are still copying also ..... Aug 27 16:38:54 I need to fix up my local.conf to cut the sizes ... Aug 27 16:42:44 I think I also have a fix for openssl and arm/uclibc Aug 27 16:42:56 ~praise koen Aug 27 16:42:57 All hail koen! Aug 27 16:43:28 and I just proved that buildroot can't do eabi Aug 27 16:52:08 03koen 07org.oe.dev * rcc4ae698... 10/ (7 files in 3 dirs): openssl: add arm/uclibc support Aug 27 16:55:49 koen: doing a build of angstrom-console-image after pull/update, so we'll see soon Aug 27 16:55:49 Jeez that target builds a lot of stuff! Aug 27 16:56:06 I'm doing the same here :) Aug 27 16:58:14 I have a build of angstrom with the x11 image, but when gpe-dm starts, there are boxes for the characters instead of the characters. Is this due to modules not being loaded or is there something else that I should check for why the text does not show up? Aug 27 16:58:41 missing font or pango modules would be my bet Aug 27 16:58:57 is there a way for pango to be in without modules? Aug 27 16:59:10 I do get a bunch of errors with modules.dep not found etc Aug 27 16:59:22 we need to create an image of the dependencies in angstrom-console-image and make a wall hanging ... Aug 27 16:59:30 I do see the font directory on there Aug 27 16:59:58 * * OE Bug 2877 has been created by  Aug 27 17:00:00 * * Build and package microperl during perl build Aug 27 17:00:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2877 Aug 27 17:01:01 v8jlene: could you have a look at #2877 when you have some time? Aug 27 17:09:38 somehow my ansgtrom build is hanging after "package task-boot-1.0: completed" Aug 27 17:12:01 koen : is the site up ? browser waits for ever Aug 27 17:33:37 koen: hello Aug 27 17:33:45 koen: this change Aug 27 17:33:46 +#PREFERRED_VERSION_glibc = "2.6.1" Aug 27 17:33:47 +#PREFERRED_VERSION_glibc-intermediate = "2.6.1" Aug 27 17:33:47 +#PREFERRED_VERSION_glibc-initial = "2.6.1" Aug 27 17:34:05 in angstrom-2008.1 is not going well with eglibc Aug 27 17:34:13 hmmm Aug 27 17:34:23 I wonder why glibc version will affect eglibc Aug 27 17:34:27 that should not have an effect Aug 27 17:34:36 but it does Aug 27 17:34:55 I get those unmet dependencies if these options are used Aug 27 17:35:20 hey Zta Aug 27 17:35:23 oops wronig window Aug 27 17:37:29 1 Aug 27 17:37:47 khem: you'd have to ask RP about that, he gets the whole dependency stuff Aug 27 17:38:30 RP ok Aug 27 17:38:53 but do you know how are these variables used ? Aug 27 17:39:20 should there be suck implicit variable for all recipes Aug 27 17:39:28 (s/suck/such/) correction: should there be such implicit variable for all recipes Aug 27 17:40:03 like PREFERRED_VERSION_eglibc etc for eglibc Aug 27 17:43:06 03koen 07org.oe.dev * rfbe42a1c... 10/ (1 packages/linux/linux-handhelds-2.6/h2200/defconfig): linux-handhelds: include wifi support in h2200 kernel and enable usb drivers Aug 27 17:54:43 RP: hello Aug 27 18:09:43 hello. can anyone tell me how to use the pango-module-basic-x that is built? the gpe-dm isn't showing any text on the login prompt and I am assuming something is setup incorrectly with pango. I am using angstrom distro. Aug 27 18:09:56 I see it in the ipk/armv4t directory Aug 27 18:10:24 can someone tell me what is needed to load it? is it something in the kernel perhaps? or does it need to get copied to the kernel? Aug 27 18:20:08 what does pango-query-modules return? Aug 27 18:28:39 hi all Aug 27 18:28:51 ciao Aug 27 18:28:57 hey vivijim Aug 27 18:29:12 hi koen Aug 27 18:29:26 just a question : eglibc can work also on oabi platform (armv4, strongarm 1100) ? Aug 27 18:29:49 gremlin[it]: it should work on all platforms glibc supports Aug 27 18:29:56 * koen looks at khem Aug 27 18:30:21 good ... :) what to enable it on local.conf ? Aug 27 18:30:54 ANGSTROM_MODE "eglibc" Aug 27 18:31:17 ok let's try, thanks koen Aug 27 18:32:29 ERROR: Unable to parse conf/bitbake.conf (conf/local.conf:95: unparsed line: 'ANGSTROM_MODE "eglibc"') Aug 27 18:32:33 hem :) Aug 27 18:33:10 ANGSTROM_MODE = "eglibc" Aug 27 18:34:03 ahahha true ... i'm dumb today ... first day of work after 3 weeks of vacations :) Aug 27 18:43:14 koen: do you have private contact with Tim Bird? Aug 27 18:48:29 koen, www.ti.com/donate Aug 27 18:48:38 only good for uni's Aug 27 19:09:58 hi people Aug 27 19:15:57 zecke: hrw has iirc Aug 27 19:16:05 hey Philippe_ Aug 27 19:16:37 howdy Aug 27 19:18:24 fhi koen Aug 27 19:19:12 Philippe: did you look at midori yet? Aug 27 19:19:19 that wraps webkit Aug 27 19:19:41 did someone mention mexican food? Aug 27 19:19:53 not really. I did not have much time yet. Aug 27 19:20:40 mickey said it would be interesting to have some help with openmoko-browser. Aug 27 19:21:00 guess I will have to convince zecke to co-operate :-) Aug 27 19:21:37 Philippe: as things look, I will get food+shelter in september to work on a browser Aug 27 19:22:04 Philippe: you are free to come to berlin :) Aug 27 19:22:19 zecke: sounds good. Make sure we can synchronize as coming to Berlin might be a bit hard. Aug 27 19:23:02 * Philippe figures out how to build a gumstix image Aug 27 19:24:02 Philippe, we have just gone through a round of debugging for gumstix Aug 27 19:24:37 Crofton: sounds good. Aug 27 19:24:51 I have a few updates that I need to push Aug 27 19:25:01 Crofton: any recommendations/ needed help Aug 27 19:25:02 just waiting for a build to run Aug 27 19:25:26 None really, I have been help a couple of guys with actual hardware Aug 27 19:26:21 Crofton: that's good Aug 27 19:26:29 yeah Aug 27 19:26:40 we have built stuff that works on connex and verdex Aug 27 19:26:48 not much testing with peripherals yet Aug 27 19:27:07 I have a soundcard and network card Aug 27 19:27:15 cool Aug 27 19:27:40 I am just wondering what the best image to build is Aug 27 19:27:56 what processor board do you ahve? Aug 27 19:28:16 basically, we have gumstix-connex and -verdex Aug 27 19:28:26 we could add basix as well Aug 27 19:28:27 I would use angstrom-minimal-image Aug 27 19:28:39 connex and basix use the same stuff Aug 27 19:28:42 I am told :) Aug 27 19:29:19 Crofton: +/- correct. Connex just has one more connector Aug 27 19:29:38 same kernel, which means we should be able to use smae machine Aug 27 19:30:01 Crofton: That should be the case indeed Aug 27 19:30:05 I suspect it gets a little crazy trying to make one image support all possible ways to assemble a system :) Aug 27 19:31:02 Crofton: my guess is that one basix/connex kernel should be enough (with a full set of modules) Aug 27 19:31:48 I'll wait for your push and will make a build myself to check for bugs Aug 27 19:31:55 ok Aug 27 19:32:09 mainly adjustments to the macine configs Aug 27 19:32:13 and the fstab Aug 27 19:32:28 I'll ping you when I push Aug 27 19:33:06 Crofton: great Aug 27 19:34:57 * * OE Bug 2878 has been created by autobuild(AT)openembedded.org Aug 27 19:34:59 * * linux-2.6.22+2.6.23-rc3-r0-do_patch Aug 27 19:35:01 * * http://bugs.openembedded.org/show_bug.cgi?id=2878 Aug 27 19:36:21 koen: BTW: I assume that the error logs are not attached because tinderclient is deleting them before they get uploaded Aug 27 19:43:53 RP: may I abuse your native speaker skills? Aug 27 19:49:59 zecke: ah, hmm Aug 27 19:50:44 shouldn't that be in site/common-glibc as well? ac_cv_func_realloc_0_nonnull={ac_cv_func_realloc_0_nonnull=yes} Aug 27 19:51:07 zecke: so unsetting TINDER_AUTOBUILD should do the trick, right? Aug 27 19:51:10 koen: you can ask tinderclient not to delete, and seppuku to delete Aug 27 19:51:18 koen: right, and seppuku has auto deletion as well Aug 27 19:51:31 koen: I don't have a build failure here to test this hypothesis Aug 27 19:55:07 any native english speaker around to read something and propose corrections? Aug 27 19:56:11 03koen 07org.oe.dev * r81566ccb... 10/ (1 conf/distro/include/angstrom.inc): angstrom.inc: let sepukku delete the workdir on failure, not tinderclient Aug 27 19:57:12 koen: did you test it? Aug 27 20:03:58 * * OE Bug 2878 has been RESOLVED (FIXED) by Aug 27 20:04:00 * *  linux-2.6.22+2.6.23-rc3-r0-do_patch Aug 27 20:04:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2878 Aug 27 20:16:10 Leen: did you put the ffmpeg wrapper in your path? Aug 27 20:16:13 oops wrong window Aug 27 20:17:58 * * OE Bug 2879 has been created by utx(AT)penguin.cz Aug 27 20:17:59 * * kernel misses USB host/client switch signalling Aug 27 20:18:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2879 Aug 27 20:20:56 koen: uclibc angstrom-console-image built without error Aug 27 20:21:07 burning flash now to see if it works Aug 27 20:22:11 wow, almost 12MB! Aug 27 20:22:26 gonna need to start with something a bit smaller :-) Aug 27 20:25:26 zecke: native english speaker here if you still need one Aug 27 20:27:45 03tmbinc 07org.oe.dreambox * r582aaec6... 10/ (4 files in 3 dirs): enigma: add blindscan plugin, add menu entry Aug 27 20:27:50 03tmbinc 07org.oe.dreambox * rf089487f... 10/ (1 packages/enigma/enigma/rotor_fix.diff): enigma: update rotor fix Aug 27 20:27:56 03tmbinc 07org.oe.dreambox * r1622178e... 10/ (1 packages/images/dreambox-image.bb): dreambox-image: add enigma blindscan plugin to image Aug 27 20:28:02 03tmbinc 07org.oe.dreambox * rc3d4fbf8... 10/ (1 packages/dreambox/dreambox-blindscan-utils.bb): dreambox-blindscan-utils: install all blindscan tools Aug 27 20:28:08 03tmbinc 07org.oe.dreambox * rc51d5ce6... 10/ (1 packages/dreambox/dreambox-dvb-modules.bb): dreambox-dvb-modules: update versions for dm500+, dm600pvr Aug 27 20:28:13 03tmbinc 07org.oe.dreambox * r25f18de0... 10/ (1 packages/dreambox/dreambox-secondstage.bb): dreambox-secondstage: update to build #53 for dm500+,dm600pvr Aug 27 20:28:18 03hrw 07org.oe.dev * rc386b75d... 10/ (3 files in 2 dirs): linux: added forgotten alix patches - close #2878 Aug 27 20:28:35 Packaged contents of cvs into /home/hrw/devel/build/angstrom/tmp/deploy/glibc/ipk/geode/cvs_1.12.9-r0_geode.ipk Aug 27 20:28:39 sakoman: consonle image includes gdb and stuff, which you probably don't need Aug 27 20:28:41 some devnotes for openmoko hackers: http://w1xer.de/OpenMoko_DevNotesDRAFT.txt Aug 27 20:28:41 morning Aug 27 20:28:47 hiya hrw|tv .. Aug 27 20:28:57 koen: IMAGE_FEATURES from Poky would be nice to have Aug 27 20:29:03 heh, funky domain name Aug 27 20:29:07 please provide feedback on this draft: http://w1xer.de/OpenMoko_DevNotesDRAFT.txt .. will work on it some more tomorrow and hopefully finish it too. Aug 27 20:29:37 koen: nur wenn mann deutsch sprechen kann! Aug 27 20:29:49 naturlich :) Aug 27 20:29:57 * koen lives about 3km from the border Aug 27 20:30:11 i'm australian, living in austria. Aug 27 20:30:26 koen: No, I hope not! Any image midway between minimal and console? Aug 27 20:30:28 did 5 years of time in deutschland, registering w1xer.de was the most productive thing i did there. . . ;) Aug 27 20:30:28 torpor: 12GB? with rm_work? Aug 27 20:30:33 yup. Aug 27 20:30:40 torpor: no fsck.. Aug 27 20:30:43 hrw|tv: actually maybe its less. Aug 27 20:30:55 i'll get the details sorted before i post it.. because that is important. Aug 27 20:30:58 sakoman: comment out the DEBUG_APPS in angstrom-console-image.bb Aug 27 20:31:10 OK, thanks Aug 27 20:32:21 torpor: those notes looks pretty ok Aug 27 20:32:26 thx Aug 27 20:32:28 torpor: command is bitbake, tool is BitBake Aug 27 20:32:36 right, i'll fix that up. Aug 27 20:33:06 going to add a bit more detail about how to use the libs too, from a developer perspective, and maybe a bit of a tour of the stuff thats included in openmoko so people can get an understanding of how to use them .. Aug 27 20:33:33 i also think the sample .bb needs some fixups to be proper. Aug 27 20:33:38 torpor: split getbigger SRC_URI into lines Aug 27 20:33:48 hrw|tv: right, done. Aug 27 20:33:56 torpor: libmokoui2 not mokoui Aug 27 20:34:02 do'h! Aug 27 20:34:11 remove #S line Aug 27 20:34:34 torpor: do not use install -d -m -- install -d is enough Aug 27 20:35:01 torpor: remove all comments from recipe Aug 27 20:35:20 torpor: remove PR = "r0" -- this is default Aug 27 20:35:38 remove PV = "0.1" -- bitbake will get it from filename Aug 27 20:36:17 I would also remove do_install() and let Makefile handle it ;) Aug 27 20:36:30 hmm. Aug 27 20:36:32 really? Aug 27 20:36:36 yes Aug 27 20:36:42 coz i thought about that and realized i didn't wnat to have an install stage in the Makefile. Aug 27 20:36:54 autotools do_install is "oe_runmake install DESTDIR=${D}" kind Aug 27 20:37:02 the recipe part is only for the OE-related stuff, and Makefile would be for whatever is non-OE specific. Aug 27 20:37:15 okay i'll play with that and adjust tomorrow accordingly Aug 27 20:37:21 torpor: install part in Makefile is very good to have in each project Aug 27 20:37:27 i'm also going to wokr a bit on the app itself to show off some of the stock features that coders can use. Aug 27 20:38:05 torpor: you do not DEPEND on cairo but use it in configure.ac - not nice Aug 27 20:38:15 ok Aug 27 20:38:16 fixed. Aug 27 20:38:25 torpor: add cairo to DEPENDS or switch configure.ac to gtk 2.10.0+ Aug 27 20:38:42 no i'll add it to DEPENDS .. sort of makes it a bit more coherent to newbies. Aug 27 20:38:45 change PKG_CHECK_MODULES(OPENMOKO, openmoko-libs >= 0.0.2) Aug 27 20:38:54 remember i'm not trying to build a totally slick example, just one that illustrates the issues. Aug 27 20:39:37 sure Aug 27 20:39:50 but good advice, all, so i've adjusted it. Aug 27 20:40:25 i'll be back tomorrow with a revised version and a working tarball. Aug 27 20:40:46 one question: is openmoko/hak really a terrible place to put code - should i move this outside the openmoko tree and explain why i've done that as well? Aug 27 20:41:06 I would avoid "bitbake -b" for newbies Aug 27 20:41:21 hmm .. Aug 27 20:41:22 torpor: put recipe in any place listed in BBFILES Aug 27 20:41:53 be Aug 27 20:41:55 * torpor makes notes to include BBFILES section. Aug 27 20:42:45 last part (about pkg-config dir) is something which I would remove or rephrase Aug 27 20:42:47 okay so if its in one of the paths in $BBFILES, they can just do "bitbake myrecipe.bb" then, and this will be simpler .. Aug 27 20:43:24 user should not check tmp/staging/ for anything. BitBake takes care about it - all it needs is proper content of DEPENDS var Aug 27 20:43:44 hrw|tv: i will extend this actually to be a bit more explicit about what this dir is, and why its fun to investigate it.. because the fact is that once you grok a .bb and autotools.ac configuration, the contents of this dir is quite intriguing - especially to new coders who have no clue whatsoever what capabilities they can use onboard... Aug 27 20:44:24 i think i'm looking at it from a new coder who wants to know what libs he/she can use right away and you'r looking at it more from the perspective of how to add whatever libs are needed/wanted to the mix .. Aug 27 20:44:57 torpor: I think that openmoko already defines list of libs available by default. Aug 27 20:45:13 new devs can just use it. Aug 27 20:45:13 yeah but this isn't readily obvious to a newcomer, thats the point. Aug 27 20:45:45 the reason i point out this dir is that once you've done your first build of openmoko from the makefile, this dir is going to tell you what you've built already and can start coding for right away without having to dig deeper into openembedded. Aug 27 20:46:12 if there is an easier/better way to understand what you can link with right away, i'd like t oknow. Aug 27 20:47:09 maybe i'll just do a section on "whats included in the moko package for you to use in coding" and include a hint that the staging/ dir can show you the results of the most recent build .. as well as a hint for how to access the onboard libs within configure.ac/.bb, etc. Aug 27 20:48:07 but .. okay .. good points, and thank you guys for taking the time to point these out. i'll get up tomorrow and have a fresh approach to it all and let you know when there's a new version for review too .. Aug 27 20:48:11 now its goodnight time. Aug 27 20:48:12 goodnight! ;) Aug 27 20:48:22 bye torpor Aug 27 20:58:01 re Aug 27 21:00:45 hi Aug 27 21:18:14 koen, what changes did you make in the sepp/tinder configs? Aug 27 21:18:19 I am updating now Aug 27 21:19:25 Philippe, pushing gumstix updates now, build tested only :( Aug 27 21:19:49 Crofton: good enough, will give me something to play with :-) Aug 27 21:19:54 thanks Aug 27 21:20:16 good time to do it, there are a couple others also caring about them atm Aug 27 21:20:24 Crofton: thank you for all the gumstix work. Worst case I just need to fine-tune it Aug 27 21:20:51 yeah, it is getting close, I think the next phase is getting network interfaces to come up at boot Aug 27 21:20:52 A bit more than fine tuning is still required :-) Aug 27 21:20:53 if present Aug 27 21:20:59 well Aug 27 21:21:11 gross work is "booting" Aug 27 21:21:14 Working on networking now Aug 27 21:21:20 fine work is booting sanely :) Aug 27 21:21:35 Indeed it is a very pleasant change!! Aug 27 21:22:10 http://en.wikipedia.org/wiki/What_to_Do_After_You_Hit_Return_or_P.C.C.'s_First_Book_of_Computer_Games Aug 27 21:22:15 found a reference at last Aug 27 21:22:27 this book taught me basic :) Aug 27 21:22:28 I'm doing a gumstix-minimal-image and adding packages one at a time as needed Aug 27 21:22:42 ah, but you are tuning to your application now Aug 27 21:22:50 We can figure out how to do it right after I figure out all that is needed! Aug 27 21:22:52 so we have special gumstix images Aug 27 21:22:54 ? Aug 27 21:22:56 no Aug 27 21:23:06 angstrom-minimal-image Aug 27 21:23:08 No still generic stuff to get basics working Aug 27 21:23:11 bye Aug 27 21:23:15 None of my special sauce Aug 27 21:23:24 and angstrom-console-image are "generic" angstrom images Aug 27 21:23:29 ah Aug 27 21:23:34 ah ok. sakoman got me confused for a sec with gumstix-minimal-image Aug 27 21:23:49 angstrom-minimal image isn't enough Aug 27 21:23:57 I think he needs the image between minimal and console :) Aug 27 21:24:06 and angstrom-console is *way* too much! Aug 27 21:24:11 Yup Aug 27 21:24:21 Turn around is much faster Aug 27 21:24:51 if you get the network going, you should be able to install from the feed Aug 27 21:25:04 That would be sweet! Aug 27 21:25:10 it does work Aug 27 21:25:31 Crofton: I had both sepukku and tinderclient clean out failed builds Aug 27 21:25:45 zecke theorized that was causing missing buildlogs Aug 27 21:25:49 you changed defaults Aug 27 21:25:56 so now only seppuku cleans out failed builds Aug 27 21:26:09 well, builds that fail in a way OE can catch it Aug 27 21:26:13 do I need to adjust local.conf? Aug 27 21:26:19 do_patch errors don't always get caught Aug 27 21:26:50 Crofton: you can remove almost everything from local.conf except the INHERIT lines and SEPPUKU_USER and SEPUKKU_PASS Aug 27 21:27:08 almost only counts in horse shoes .... Aug 27 21:29:12 can we override the "zecke-rocks" variable? Aug 27 21:29:28 you mean failfast? Aug 27 21:29:39 no Aug 27 21:29:47 "DISTRO" : data.getVar('DISTRO', d, True) or "Unknown", Aug 27 21:29:47 "zecke-rocks" : "sure", Aug 27 21:29:47 } Aug 27 21:29:55 :) Aug 27 21:29:58 :) Aug 27 21:30:28 I am trying to see how the vars are defaulted to sane values so I can remove the correct ones from local.conf :) Aug 27 21:31:03 reading the python is not showing me the answer :) I hesitate to call Python "source" Aug 27 21:32:35 Crofton: if you have any indication that this is a false statement we can remove it Aug 27 21:33:02 and rewrite it to use bugzilla xml-rpc is welcome as well :) Aug 27 21:33:16 ok, I found tinder.conf Aug 27 21:33:26 but the vars are commented out? Aug 27 21:34:13 you can remove the zecke-rocks line and commit if you feel like... Aug 27 21:34:15 Crofton: check conf/distro/include/angstrom.inc Aug 27 21:34:29 ah Aug 27 21:34:31 thanks Aug 27 21:34:37 koen: I turned in a proposal for ELCE Aug 27 21:34:38 That is what you said this morning Aug 27 21:34:46 :D Aug 27 21:35:08 Philippe: if you do a connex build you might want to check this page for a workaround to quiet all of the alignment trap messages you'll get Aug 27 21:35:10 * koen hugs adobe lightroom Aug 27 21:35:44 koen: how did the yashica-mat pics turn out? Aug 27 21:36:04 sakoman: pretty nice Aug 27 21:36:04 03crofton 07org.oe.dev * re0e9a6cd... 10/ (1 packages/base-files/base-files/omap5912osk/fstab): base-files : Update fstab for omap5912osk to clear up a boot time error. Aug 27 21:36:09 03crofton 07org.oe.dev * r4a5850b8... 10/ (1 conf/machine/omap5912osk.conf): omap5912osk.conf : Clean up machine file. Aug 27 21:36:13 03crofton 07org.oe.dev * r3dc200b4... 10/ (4 files in 3 dirs): Aug 27 21:36:13 gumstick-*.conf : Move machine specific parts from gumstix.inc to the gumstix-{connex,verdex} Aug 27 21:36:13 machine files. Aug 27 21:36:16 sakoman: the scanner arrives in ~48 hours :) Aug 27 21:36:18 03crofton 07org.oe.dev * r610425fd... 10/ (5 files in 4 dirs): Aug 27 21:36:18 base-files : Add machine specific fstabs for gumstix-{connex, verdex}. In the Aug 27 21:36:18 future these may be replaced with the generix fstab. Aug 27 21:36:20 sakoman: which page? Aug 27 21:36:44 oops: http://www.sakoman.net/oe/gumstix-issues.txt Aug 27 21:37:24 koen: looking forward to seeing the scans Aug 27 21:37:32 * koen too Aug 27 21:37:42 that camera takes really sharp images Aug 27 21:37:53 amazing detail Aug 27 21:37:55 sakoman: thanks Aug 27 21:38:05 * koen needs to buy a set of tweezers as well Aug 27 21:38:30 Philippe: just make the edits & reboot, all will be well Aug 27 21:38:35 or stop eating candy before touching the negatives Aug 27 21:38:42 We'll eventually get to root cause Aug 27 21:38:54 sakoman: is this on uclibc? Aug 27 21:39:19 I remember woglinde complaining about uclibc causing spurious alignemnt traps Aug 27 21:39:40 happens with both uclibc and libc in my experience Aug 27 21:39:51 but I only did one libc build Aug 27 21:40:02 and tried one from koen Aug 27 21:40:08 I checked on a different armv5 with glibc and only webkit caused alignment errors Aug 27 21:40:13 glibc 2.5, that is Aug 27 21:41:08 I'll try libc again after I get networking going Aug 27 21:41:23 once I fix gcc 4.2.1 with uclibc I can try that Aug 27 21:41:25 koen: webkit, alignment errors? Aug 27 21:41:25 ok tar.gz works as an image type? Aug 27 21:41:32 yes Aug 27 21:41:56 zecke: yes Aug 27 21:42:10 sakoman: I love B&W films that uses C41 Aug 27 21:42:37 I can get the negatives back in ~2 hours Aug 27 21:42:38 Yes, wasn't sure they still made it Aug 27 21:42:51 Used to develop my own back in the day Aug 27 21:43:02 even c41 Aug 27 21:43:03 no space for that Aug 27 21:43:34 Yeah, I moved & ended up selling all of the darkroom equipment Aug 27 21:44:55 koen: on glibc, or uclibc? reliable? Aug 27 21:46:01 ok rebuilding four machines again ... Aug 27 21:47:49 how is gpe-login executed after bootup? Aug 27 21:48:01 I mean what script or application initiates it Aug 27 21:48:10 zecke: glibc Aug 27 21:49:12 dcordes: I cannot think about the exact script but I think it is /etc/init.d/gpe-dm Aug 27 21:50:19 Philippe: any idea how I can replace that with the openmoko ui? Aug 27 21:50:33 dcordes: sorry no? Aug 27 21:51:10 koen: already on start? Aug 27 21:51:29 zecke: I don't have the logs with me, sorry Aug 27 21:51:52 dcordes: I haven't looked into the openmoko details Aug 27 21:52:18 replace the matchbox session file Aug 27 21:52:30 /etc/matchbox/session or ~/.matchbox/session Aug 27 21:52:44 koen: I installed openmoko-session2 which created /etc/matchbox/sesion Aug 27 21:53:07 if you don't have one in ~/.matchbox 'gpe' should pick that up Aug 27 21:53:16 I get into gpe-login screen after bootup and then login as root and get into the moko sesison Aug 27 21:53:32 AFAIK gpe-login just launches matchbox session Aug 27 21:53:42 okay that's the point Aug 27 21:53:47 ah right Aug 27 21:53:52 I would like to run the matchbox session directly after bootup Aug 27 21:54:03 without login Aug 27 21:54:05 remove gpe-dm, gpe-session-script, gpe-login and xserver-common Aug 27 21:54:24 will mokoui work without xserver-common? Aug 27 21:54:25 install xserver-kdrive-common , the openmoko session stuff and xserver-kdrive-nodm-something Aug 27 21:54:42 ok let me try Aug 27 21:54:51 * koen still wants a way for openmoko to work with gpe-dm+xserver-common Aug 27 21:55:02 it sucks that if X goes down it doesn't restart Aug 27 21:55:17 I have no idea what it does but it sounds critical :x Aug 27 21:55:27 that's why I asked Aug 27 21:55:58 xserver-kdrive-common is a fork of xserver-common Aug 27 21:56:03 I see Aug 27 21:56:08 they basically do the same Aug 27 21:56:14 but with some differences Aug 27 21:56:40 ok removed the bunch Aug 27 21:56:42 now installing Aug 27 21:57:01 what is that "kdrive" for? Aug 27 21:57:09 what does it mean Aug 27 21:57:10 Keith's driver Aug 27 21:57:22 kdrive was keith packards playground Aug 27 21:57:23 heh Aug 27 21:58:01 it removes all compat with computers and software from 1971 for example Aug 27 21:58:18 oh crap Aug 27 21:58:34 we must save the old machines! Aug 27 21:59:02 say that to the SGI and SUN machines in my bike shed :) Aug 27 21:59:09 heh Aug 27 21:59:11 and hp xterminals Aug 27 21:59:32 I think all the ones here have been surplused Aug 27 21:59:33 koen: there is no pack with xserver-kdrive-nodm* Aug 27 21:59:45 or taken home be people that needed terminals at home Aug 27 22:00:44 * koen looks Aug 27 22:00:55 thanks Aug 27 22:01:08 http://www.angstrom-distribution.org/repo/?action=details&pnm=xserver-nodm-init Aug 27 22:01:53 oh no :D Aug 27 22:01:56 I'm trying to make openmoko-today2 work outside of openmoko Aug 27 22:02:02 that startet a gpe-desktop Aug 27 22:02:40 koen: if openmoko-today2 is that thing where I have 3 tabs in the bottom and can make calls and stuff in the first screen that is working for me on angstrom Aug 27 22:03:02 it works Aug 27 22:03:14 but I want it to work without the openmoko theme and stuff Aug 27 22:03:20 ah ok Aug 27 22:03:21 and with matchbox-panel-1 Aug 27 22:03:35 and on qvga :) Aug 27 22:03:39 any idea why I got into a gpe-looking session now? Aug 27 22:03:48 no, sorry Aug 27 22:03:50 after installing xserver-nodm-init Aug 27 22:04:41 oh somehow openmoko-session2 was removed Aug 27 22:07:05 koen: that didn't help either Aug 27 22:07:32 now not the gpe looking thing is started but nothing at all Aug 27 22:09:30 a reboot made it work Aug 27 22:15:27 dcordes: could you send a mail to the angstorm mailinglist(s) with the steps needed for the conversion? Aug 27 22:15:54 Question on how drivers get installed into the image Aug 27 22:15:57 of course Aug 27 22:16:00 just a second Aug 27 22:16:11 gumstix build is building smc91x ethernet driver Aug 27 22:16:11 ask the question :) Aug 27 22:16:27 but it isn't getting installed Aug 27 22:16:35 thing is I don't know if a reboot at the end of the procedure is required or not Aug 27 22:16:43 or if a depmod -a would possibly do Aug 27 22:16:53 MACHINE_EXTRA_RRECOMENDS ? Aug 27 22:16:56 how do I get it installed :-) Aug 27 22:17:11 should install it, if avialable Aug 27 22:17:18 Ah, OK, or perhaps MACHINE_EXTRA_DEPENDS? Aug 27 22:17:21 is my understanding :) Aug 27 22:17:22 koen: I'm only registred on the oe bugtracker Aug 27 22:17:47 EXTRA_RDEPENDS would be my guess, since it is a run time dependency Aug 27 22:17:57 I'll try RRECOMMENDS first Aug 27 22:18:04 * Crofton is hoping someone is doube checking Aug 27 22:18:05 nvever EVER but a kernel module in RDEPENDS Aug 27 22:18:15 reason? Aug 27 22:18:19 that's what it is :-) Aug 27 22:18:25 hell will break loose when you make it in the defconfig Aug 27 22:18:43 for a gumstix rrecommends is better since someone without a network card my want to remove it Aug 27 22:18:47 ah Aug 27 22:19:03 if it is a depends, and it is not available Aug 27 22:19:06 the build fails Aug 27 22:19:10 gumstix buildroot version always includes it Aug 27 22:19:25 oh I tried adding module to extra install once and couldn't build for a week Aug 27 22:19:28 and guess why :D Aug 27 22:19:56 sakoman, gumstix builds ot into the kernel? Aug 27 22:21:57 somewhere I used to have something about autoloading a module Aug 27 22:23:37 module_autoload_ohci-hcd_omap5912osk = "ohci-hcd" Aug 27 22:23:45 I have this in linux-omap1.inc Aug 27 22:24:14 I am not entirely sure why, and what it really does Aug 27 22:24:51 Crofton: sorry stepped away Aug 27 22:25:05 Gumstix builds network driver as a kernel module Aug 27 22:25:14 ok Aug 27 22:25:16 koen: how can I install all packages beginning with task-openmoko? Aug 27 22:25:44 sakoman, I am pretty sure the module_autoload forces the driver to load at boot time Aug 27 22:25:59 but, it is voodoo I do not fully understand Aug 27 22:26:06 yes, but it doesn't get copied into rootfs :-) Aug 27 22:26:29 That was the question, I'm lousy at asking! Aug 27 22:27:02 So RRECOMMENDS should do it? Aug 27 22:27:08 I understand the original question also :) Aug 27 22:27:14 this is what I believe :) Aug 27 22:27:21 if it gets built and packaged Aug 27 22:27:42 the OSK kernel has networking built in Aug 27 22:27:54 since the NIC is on the PC board :) Aug 27 22:28:09 Trying to do it gumstix style so Craig has less work to do to adopt OE Aug 27 22:28:19 yeah Aug 27 22:28:47 once it gets installed, I suspect you will need to force loading at boot Aug 27 22:28:52 He says today is the day for giving OE a try Aug 27 22:29:15 tell him I will swap tech support for hw :) Aug 27 22:29:38 I've been shameless in pushing for free hw for you :-) Aug 27 22:29:43 thanks :) Aug 27 22:30:16 Next step is to tell him I'll buy it if he won't send it Aug 27 22:30:21 heh Aug 27 22:30:25 hopefully he sees the light Aug 27 22:30:27 Start using shame and guilt! Aug 27 22:30:56 how to ipkg install task-openmoko(anything) Aug 27 22:31:00 ? Aug 27 22:31:48 one of my other procrastination tasks has been looking at the Maniac Challenge software Aug 27 22:32:04 I am thinking a gumstix with wifi would be interesting Aug 27 22:32:19 I am slowly adding packages needed for it Aug 27 22:32:41 and have talked to the guys in charge about getting the "special sauce" piece Aug 27 22:32:50 when I kill/restart/crash xserver-kdrive, I can't get back into the session again. Any way to achieve that? Aug 27 22:33:11 startx didn't work Aug 27 22:33:53 Wifi is next on the list after ethernet, since that is what my product uses Aug 27 22:34:03 excellent Aug 27 22:34:17 sakoman: what kind of product are you makink? Aug 27 22:34:43 solar energy/battery monitor with web UI Aug 27 22:34:46 hahaha Aug 27 22:35:10 so I can war drive looking for people's solar controllers, and monitor their systems ? Aug 27 22:35:15 not lawn mower with wifi? Aug 27 22:35:17 you bet! Aug 27 22:36:05 you can monitor mine: http://www.sakoman.net:8080 Aug 27 22:36:11 uplink is slow Aug 27 22:36:20 but you'll get the idea Aug 27 22:36:48 very cool Aug 27 22:37:06 click on mini-graph icon for plots Aug 27 22:37:28 cool Aug 27 22:37:42 click on plot for bigger plot :-) Aug 27 22:38:03 Not bad for a 200Mhz arm Aug 27 22:38:05 hmm, what happened at noon? cook lunch? Aug 27 22:38:26 Need to keep base image size small for a year's worth of data Aug 27 22:38:57 Swimming pool pump kicks on :-) Aug 27 22:39:04 hah Aug 27 22:39:18 Hot today! See the temp? Aug 27 22:40:45 It can send alert emails for battery full/low etc Aug 27 22:42:23 sakoman: can I test your lawnmower ui? Aug 27 22:42:43 A more detailed question -- how do I specify with MACHINE_EXTRA_RRECOMMENDS that I want smc91x installed? Aug 27 22:42:57 dcordes: still under design :-) Aug 27 22:43:07 suggestions? Aug 27 22:43:38 what is the module called in the machine specific ipk directory in deply Aug 27 22:43:39 you should add gsm Aug 27 22:43:39 tried kernel-module-smc91x, didn't work Aug 27 22:43:46 for sms reports Aug 27 22:43:48 bother Aug 27 22:43:58 that was pretty much my suggestion Aug 27 22:44:04 why doesn't this work? Aug 27 22:44:57 kernel-module-smc91x_2.6.21-r7_gumstix-connex.ipk Aug 27 22:45:21 * dcordes notices he never used any of the gpe applications and is happy about mokoui Aug 27 22:46:04 hi :) Aug 27 22:46:11 << neo-1973-dev Aug 27 22:46:40 hello Aug 27 22:46:43 hmm Aug 27 22:46:55 dcordes: good idea, been using the carrier email gateway for sms Aug 27 22:47:32 sakoman, how did you rebuild the image? Aug 27 22:47:48 -c rebuild Aug 27 22:47:53 on the image Aug 27 22:48:11 try bitbake -c clean task-base, Aug 27 22:48:15 which image? Aug 27 22:48:19 | Collected errors: Aug 27 22:48:19 | Cannot find package kernel-module-net-smc91x. Aug 27 22:48:19 | Check the spelling or perhaps run 'ipkg update' Aug 27 22:48:32 ah, a clue Aug 27 22:48:36 arghh no -net!!! Aug 27 22:48:41 idiot! Aug 27 22:48:47 :) Aug 27 22:48:50 whew Aug 27 22:49:11 riot: how can one rebuild the list of apps in the moko today applaunching section? Aug 27 22:49:30 dcordes: you again ;) Aug 27 22:49:37 dcordes: hmm.. all i did, was reboot Aug 27 22:51:04 riot: when you restart x on your moko- what does happen? Aug 27 22:51:20 sakoman, must have turned of the do as I mean, not as I type feature Aug 27 22:51:24 hmm.. in pygtk (and e.g. mrxvt), all the fonts are HUGE. How can i make them weeetiny? Fontsize 8 is big, here. Dunno if this is offtopic here, so pardon me, if it is :/ Aug 27 22:51:36 dcordes: uh, it restarts the xserver? Aug 27 22:51:49 Crofton: I use old age as an excuse, how about you? Aug 27 22:52:00 dcordes: and starts outputting all the stdout/err on the console which i used to make it happen Aug 27 22:52:06 riot: for me it sows up screensfull of kdrive errors and then I get to command line Aug 27 22:52:14 You can only be young once, but you can be immature forever! Aug 27 22:52:21 :-) Aug 27 22:52:49 it is hard to get old in a college town Aug 27 22:53:14 True, I miss living in one Aug 27 22:53:15 however, my faculties are not what they used to be Aug 27 22:53:27 Crofton: right, its impossible for me, actually. Aug 27 22:54:14 koen: I'll take a look at 2877, could be a few days before I get around to it though. Aug 27 22:54:45 * riot is drinking a beer alone in the HQ of the students sailing club, while hacking some embedded python apps Aug 27 22:54:58 v8jlene, I think this is miniperl? Aug 27 22:55:05 Crofton: Yep. Aug 27 22:55:27 I believe the URL sakoman just posted shows an app running miniperl Aug 27 22:55:29 dcordes: hmm.. not ever seen here. I'm still running that image from 23.Aug. Aug 27 22:56:14 Crofton: yes it is microperl Aug 27 22:56:48 It sound easy enough to build... but nothing perl cross-compile related is ever easy! Aug 27 22:57:10 it builds on buildroot, so it must be easy! Aug 27 22:57:25 dcordes: what actually is kdrive? Aug 27 22:58:39 gnf.... the HUGE font in mrxvt really kills my nerves. Aug 27 22:58:50 "kdrive was keith packard's playground" Aug 27 22:59:14 riot: we should take this to #openmoko Aug 27 22:59:25 dcordes: possibly Aug 27 23:00:19 sakoman: It might be easy ;) Just looking now and since there's almost no configuration and no module stuff it looks like it probably will be. Aug 27 23:05:32 v8jlen: if it builds for you and you are interested, I have a couple of patches I needed to apply to the microperl build to fix broken syscalls and environment access Aug 27 23:05:53 both relatively important to actual use of microperl :-) Aug 27 23:11:43 sakoman: If you could attach the patches to bug 2877 then I could add them in when I'm looking at it. Aug 27 23:21:35 OK, haven't used the OE bug system yet, but I'll give it a try Aug 27 23:22:29 Now that I think about it, perhaps I should wait to see of the OE build of microperl has the issues Aug 27 23:22:38 Might have been fixed upstream Aug 27 23:23:19 My patches are for perl-5.8.8, not sure what you will be building Aug 27 23:24:03 sakoman: Yeah, no problem. Someone to test it when it's done would be handy. It'll probably be 5.8.8 since that's what's currently in OE. Aug 27 23:24:04 a quick glance suggests 5.8.8 .... Aug 27 23:24:33 OK, I'll attach the patches then Aug 27 23:32:17 sakoman, you have a solar controller biz right, what is the url? Aug 27 23:32:43 a friend of mine is interested in these things, he has a system in his camping trailer Aug 27 23:33:00 site is "under construction", but is at: www.sakoman.com Aug 27 23:33:25 In beta test right now Aug 27 23:34:52 your displays are much better then the ones in his trailer :) Aug 27 23:35:58 Yeah, the competition is with 7 segment led displays :-) Aug 27 23:37:08 Or 3 line lcd's at the high end Aug 27 23:38:19 Consulting is sucking a lot of time right now, so finishing final features is slow Aug 27 23:38:37 But it pays the bills :-) Aug 27 23:39:06 heh Aug 27 23:39:09 money is useful Aug 27 23:39:29 Ray mentioned mignight solar Aug 27 23:39:44 midnite solar Aug 27 23:39:50 well yeah Aug 27 23:39:51 Yup, I know those guys Aug 27 23:40:24 They are doing some interesting stuff on the mechanical side Aug 27 23:40:38 And also on an advanced charge controller Aug 27 23:40:49 I suspect that at some level, the mechanics are the hard part .. Aug 27 23:41:07 yeah Ray's response to the email I sent mentioned controller stuff Aug 27 23:41:09 Yeah, lots of wires Aug 27 23:41:41 do you just do the monitor? Aug 27 23:42:20 designer is a fellow by the name of boB Gudgel Aug 27 23:42:53 Yeah, just the monitor. Designed to work with any brands of panel, charge controller, inverter, batteries Aug 27 23:43:24 well, I'll suggest my friend put one in his trailer :) Aug 27 23:43:27 Fun project, I get to do hardware, software, mechanical Aug 27 23:43:55 The wifi interface is an issue for really small installations like a trailer Aug 27 23:44:04 Sucks power :-) Aug 27 23:44:27 Still working on getting power lower Aug 27 23:45:27 BTW, boB Gudgel is the fellow who is designing the high end charge controller for midnite Aug 27 23:45:37 not for my stuff :-) Aug 27 23:46:44 so the network module did make it onto the gumstix? Aug 27 23:46:59 I need to to wack out a few more lines and go home Aug 27 23:47:02 sigh, no Aug 27 23:47:25 hmm, I was hoping the spelling would fix it Aug 27 23:47:34 it does look like it works for me Aug 27 23:47:47 maybe, it is a minimal, versus console problem Aug 27 23:47:56 In gumstix.inc: Aug 27 23:48:00 MACHINE_EXTRA_RRECOMMENDS = "kernel-module-smc91x \ Aug 27 23:48:00 kernel-module-gumstix-smc91x \ Aug 27 23:48:08 Error still: Aug 27 23:48:21 Collected errors: Aug 27 23:48:21 | Cannot find package kernel-module-net-smc91x. Aug 27 23:48:21 | Check the spelling or perhaps run 'ipkg update' Aug 27 23:48:21 | Aug 27 23:48:30 May need a clean build Aug 27 23:48:46 you are doind minimal? Aug 27 23:49:00 my edited minimal Aug 27 23:49:23 which adds the loadable module stuff Aug 27 23:49:30 RDEPENDS on task-boot Aug 27 23:49:58 so try bitbake -c clean task-boot Aug 27 23:50:08 tried that, no joy Aug 27 23:50:49 Time for a break, so I may rm -rf tmp & rebuild Aug 27 23:51:15 Doing 3 things at once:-0 Aug 27 23:51:19 I'm just flipping through the bb files and task.bbclass Aug 27 23:51:20 yeah Aug 27 23:51:22 me too Aug 27 23:51:59 I think it's really close, my build env make just be screwed up after all this messing around Aug 27 23:52:02 yeah Aug 27 23:52:05 that happens Aug 27 23:52:17 even with buildroot :-) Aug 27 23:52:20 can you pastebin the RRECOMMENDS line Aug 27 23:52:24 you thinnk should work? Aug 27 23:52:34 rm -rf build-arm-nofpu in that case :-) Aug 27 23:52:42 I will add , I have a long build running Aug 27 23:53:53 MACHINE_EXTRA_RRECOMMENDS = "kernel-module-smc91x \ Aug 27 23:53:53 kernel-module-gumstix-smc91x \ Aug 27 23:54:01 these are correct names? Aug 27 23:54:08 yes Aug 27 23:54:28 here's what I am adding over angstrom-minimal Aug 27 23:54:54 sysfsutils \ Aug 27 23:54:54 module-init-tools \ Aug 27 23:54:54 update-modules \ Aug 27 23:54:54 kernel-module-rtc-core \ Aug 27 23:54:54 kernel-module-rtc-dev \ Aug 27 23:54:54 kernel-module-rtc-proc \ Aug 27 23:54:56 kernel-module-rtc-sysfs \ Aug 27 23:55:00 kernel-module-rtc-sa1100 \ Aug 27 23:55:02 kernel-module-smc91x \ Aug 27 23:55:04 kernel-module-gumstix-smc91x \ Aug 27 23:55:06 ipkg \ Aug 27 23:55:22 as RRECOMENDS? Aug 27 23:55:58 Well, up to now as part of the gumstix-minimal-image I created for testing Aug 27 23:56:36 Just moved to RRECOMMENDS after the smc91x stuff wasn't making it into the rootfs Aug 27 23:57:13 So that may be why my build env is hinky Aug 27 23:57:29 from task-boot http://rafb.net/p/CvDjun42.html Aug 27 23:57:39 Will move them all to RRECOMMENDS and do a clean build Aug 27 23:57:43 task boot is where the stuff for minimal image comes from Aug 27 23:58:31 the subject of the "best" way to create custom images for "vendors" should be addressesd better I think Aug 27 23:58:51 Hmm . . . so better make them MACHINE_ESSENTIAL_EXTRA_RDEPENDS Aug 27 23:59:04 er RECOMENDS Aug 27 23:59:10 you can set those things in local.conf Aug 27 23:59:27 cleaning task-boot should force them to be re-read Aug 27 23:59:36 is that the OE way? Aug 27 23:59:55 seems like it wouldn't help new gumstix folks if it were hidden there Aug 28 00:00:01 I am not certain at which point MACHINE_EXTRA_RRECOMMNEDS gets sucked in Aug 28 00:00:10 should be in machine.conf I would think Aug 28 00:00:16 I think the modules should come from macine file Aug 28 00:00:17 yes Aug 28 00:00:24 OK Aug 28 00:00:46 It is 8PM here, I need to finish up and go home Aug 28 00:00:55 Have a nice night! Aug 28 00:01:01 thanks you too Aug 28 00:01:11 I'll look in my images in the morning Aug 28 00:48:58 * * OE Bug 2880 has been created by autobuild(AT)openembedded.org Aug 28 00:49:00 * * linux-handhelds-2.6-2.6.21-hh12-r2-do_compile Aug 28 00:49:02 * * http://bugs.openembedded.org/show_bug.cgi?id=2880 Aug 28 00:53:16 03mickeyl 07org.oe.dev * r6b34ae43... 10/ (1 packages/midori packages/midori/midori_0.0.5.bb): add midori, a lightweight web browser based on webkit-gtk Aug 28 00:53:21 03mickeyl 07org.oe.dev * r0b24529a... 10/ (1 packages/qt/qt4-x11-free_4.3.0.bb): qt4-x11 4.3.0 remove bogus OE_QMAKE_LIBS_X11 from pkgconfig file Aug 28 00:53:25 03mickeyl 07org.oe.dev * rfe0b08c9... 10/ (3 files in 2 dirs): Aug 28 00:53:25 lyx: update to 1.5.1 Aug 28 00:53:25 NOTE: This version is based on Qt4/X11 Aug 28 00:53:30 03mickeyl 07org.oe.dev * r7c5d468e... 10/ (4 files in 3 dirs): linux-gta01: remove 2.6.21.5 Aug 28 01:17:50 sakoman, ping Aug 28 01:18:58 I'm thinking that for gumstix, we shoudl rrecommend all modules built from teh kernel config Aug 28 01:28:18 03mickeyl 07org.oe.dev * r3b73bd5a... 10/ (4 files in 2 dirs): qemu: factor out gcc3 check function and include it Aug 28 01:50:10 in the truly distrubed category, I am doing two seperate OR euns on the smae machine at the same time Aug 28 02:01:09 grmbl... emotion build segfault on my amd64 Aug 28 02:01:36 no e-x11-image tonight :( Aug 28 02:11:59 are there written procedures regarding bug management in OE ? Aug 28 02:13:48 for instance, what if I have a fix for a bug ? Should I change its status to "assigned" once I have provided a patch so that official maintainter can review them ? Aug 28 02:16:19 and what are the procedures to close a bug ? if I'm not the first reporter, but found that it's fixed, should I resolve it to FIXED or to WORKFORME until the original reporters close it ? Aug 28 02:17:11 I'm accustomed to bug management à la debian, but each project has its own rules and customs... Aug 28 02:38:35 * nodens just found bug 2194 Aug 28 02:39:04 progressing slowly but steadily ;) **** ENDING LOGGING AT Tue Aug 28 02:59:56 2007