**** BEGIN LOGGING AT Sat Feb 19 02:59:57 2011 Feb 19 03:15:49 03Rick Foos  07master * rf5a8b9cf63 10openembedded.git/contrib/testing/testscript.sh: Feb 19 03:15:49 contrib/testing/testscript.sh: Detect MIN_BB_VERSION Feb 19 03:15:49 * Fixes Bugzilla/5517 Feb 19 03:15:49 * extract the quoted string from BB_MIN_VERSION in config/sanity.conf. Feb 19 03:15:49 Should work for all versions of bitbake Feb 19 03:15:49 Signed-off-by: Khem Raj Feb 19 03:33:20 hooray fixing the libtool output seems to have fixed my gtk+ issue from earlier. stupid *.la files had a bogus libdir path Feb 19 05:35:59 My openembedded build-console is running for more than 12 hours on a 8core system with 1.7GHz speed connected to network of 10MBytes/sec. Is such a long time common? Feb 19 05:36:40 that seems like a pretty long time. the 8 cores only matters if you have setup things to go in parallel Feb 19 05:37:05 demigod2k, you can easily swamp the HD in most things with 8 cores setup in parallel Feb 19 05:37:41 sure I just mean check uptime, its possible you think the 8 cores are working but will still see a 1.000 load average like I did today Feb 19 05:37:48 could be network limited, could be a lot of things Feb 19 05:38:06 but no when I've been running things lately it has been 1-2 hours with most of the binaries already downloaded Feb 19 05:38:18 12+ seems awfully long Feb 19 05:38:24 demigod2k: you do a --fetchall (some switch)? Feb 19 05:38:54 I havent been using any switches. actually I'm just getting started with things, trying to bringup a board using OE for the first time Feb 19 05:39:08 I see. Feb 19 05:39:39 then, I don't know what really the issue issue. Does bitbake skip past failures on (target servers) or just keeps retrying? Feb 19 05:40:05 it prints what it's doing you should be able to see the status on stdout Feb 19 05:40:31 yeah, I do see it. but did not see any skip msg so far. Feb 19 05:40:59 and it shows what task its on. usually for me it's about 7000 total Feb 19 05:57:53 I don't see which task it is on, but I see the following msg. NOTE: package opkg-native-0.1.8+svnr599-r4: task do_rm_work: Succeeded Feb 19 05:58:03 It is at this point for a long time now. Feb 19 05:58:55 if long is >5 mins I'd guess something has gone wrong Feb 19 05:59:26 whenever I see a build it usually has moved onto a new task at least every minute Feb 19 06:02:05 I see. Feb 19 06:02:21 what can be done, if you think that something has gone wrong? Feb 19 06:03:21 ctrl-c and try it again? I dont know. If it has sat there for way too long then I suspect something is wrong Feb 19 06:07:05 I see. thanks, I shall browse to look for possible errors and go for ctrl-c as a last resort. Feb 19 06:07:45 well if the output isnt changing it sure would seem as if it has locked up Feb 19 06:51:04 03Mario Schuknecht  07org.openembedded.dev * r09da7564ee 10openembedded.git/recipes/u-boot/ (3 files in 3 dirs): (log message trimmed) Feb 19 06:51:04 u-boot-2009.03: use AMD fixups for AMIC too Feb 19 06:51:04 * AMIC flashes are AMD clones and need the same fixups otherwise e.g. the Feb 19 06:51:04 u-boot command saveenv may not work (depending on the sector layout) Feb 19 06:51:04 * patch has been send upstream too: http://patchwork.ozlabs.org/patch/82763/ Feb 19 06:51:04 Signed-off-by: Mario Schuknecht Feb 19 06:51:05 Signed-off-by: Steffen Sledz Feb 19 07:49:52 Tartarus: with actual head, when compiling console-base-image & qt4e-demo-image using angstrom 2008.1 and targetting qemuarm,, I still have QA issues in gnutls, shroedinger, libcdio and mysql5 Feb 19 07:58:40 morning ericben, we worked today on solutions. Feb 19 07:58:51 but now I must sleep...not much last night. Feb 19 07:59:00 ka6sox: hi ka6sox Feb 19 07:59:07 cool Feb 19 07:59:11 have a good night Feb 19 07:59:17 thanks... Feb 19 07:59:37 solutions fgor qa issues or for oestat/jenkins ? Feb 19 07:59:59 oestats/jenkins Feb 19 08:00:21 I am terrible with Java but I think it can be done. Feb 19 08:01:48 jenkins expects full time connection but I think we can make that work by telling it the next time we will connect Feb 19 08:01:55 that's a good new. now have a good night you may need it before going to jenkins :) Feb 19 08:02:04 true Feb 19 08:02:50 nn Feb 19 08:06:37 "bitbake nano", taking too long, anything smaller? Feb 19 08:07:25 usually breaks at the end Feb 19 08:08:48 which is better on at91sam9263ek with Qt embedded apps, libc or uclibc? Feb 19 08:19:59 nataraj: I using eglibc with angstrom-2010 on armv5 without much problem for qt4e Feb 19 08:29:11 ericben, is qt4e the latest stable, say 4.6.3 ? Feb 19 08:29:35 ERROR: 'virtual:native:/stuff/openembedded/recipes/automake/automake_1.11.1.bb' failed Feb 19 08:30:04 whats the error Feb 19 08:30:53 ericben: I dont understand why patchwork is freakin out on utf Feb 19 08:31:09 on contrary I applied a patch to fix utf in pw Feb 19 08:31:53 http://pastebin.com/7aMd9LQK, am i missing enything? Feb 19 08:32:12 getting this for 'bitbake nano' Feb 19 08:34:32 hmmm I wonder why it cant find aclocal Feb 19 08:34:42 what does your local.conf look like Feb 19 08:36:20 khem, aclocal of host? Feb 19 08:40:46 http://pastebin.com/5vzidWEN Feb 19 08:49:58 hi khem Feb 19 08:51:33 khem, do you remember of youc ommit c6a86c6075bfba1c51fc9b85f941d1d530cb3db0 ? log sayys libtool > 2.4, comment says < 2.4 and in the end a test is ("${LIBTOOL_HAS_SYSROOT}" == "yes") : is everything right in this patch ? Feb 19 08:53:53 khem: there is one issue in .la files in sysroot with angstrom 2008 and actual head. libstdc++.la, supc++, ssp.la and gfortan.la have a wrong libdir Feb 19 08:54:03 with a full path instead of /usr/lib Feb 19 08:59:12 gm Feb 19 09:12:50 hi eFfeM Feb 19 09:34:05 how are you doing eric ? Feb 19 09:34:12 (was afk for coffee) Feb 19 09:44:58 eFfeM: fine, lot of work actually but far better than no work at all ;-) Feb 19 09:46:23 same here Feb 19 10:38:41 03Koen Kooi  07org.openembedded.dev * ra73eb65b95 10openembedded.git/recipes/gnome/gvfs_1.7.2.bb: gvfs: add 1.7.2 Feb 19 10:38:50 03Koen Kooi  07org.openembedded.dev * r246abfc865 10openembedded.git/recipes/gnome/ (gvfs/async.patch gvfs_1.6.6.bb): gvfs 1.6.6: add patch to fix async_connection_accept_new_fd( Feb 19 10:38:51 03Koen Kooi  07org.openembedded.dev * rd9f07823c0 10openembedded.git/conf/distro/include/angstrom-2010-preferred-versions.inc: Feb 19 10:38:51 angstrom next: prefer gvfs 1.7.2 (unstable series) Feb 19 10:38:51 Signed-off-by: Koen Kooi Feb 19 11:23:59 03Jan Lübbe  07master * r8694f8356e 10openembedded.git/recipes/ofono/ (4 files in 2 dirs): Feb 19 11:23:59 ofono: Ship test scripts as ofono-test-utils, ship udev rules Feb 19 11:23:59 Also add USB-ID for a Sierra modem and add INC_PR. Feb 19 11:23:59 Signed-off-by: Jan Lübbe Feb 19 11:23:59 Signed-off-by: Stefan Schmidt Feb 19 11:24:10 03Marcin Mielczarczyk  07master * re52a8bf769 10openembedded.git/ (5 files in 4 dirs): Feb 19 11:24:10 sciphone-g2: Initial support for Sciphone G2 (based on MT6235) Feb 19 11:24:10 This patch adds Sciphone G2 target to OpenEmbedded environment. Feb 19 11:24:10 It also defines osmocomBB repositories for U-Boot and Linux kernel. Feb 19 11:24:10 Signed-off-by: Marcin Mielczarczyk Feb 19 11:24:11 Signed-off-by: Stefan Schmidt Feb 19 11:52:17 good morning Feb 19 11:59:34 How can I add a task that needs to be done after image build? Feb 19 12:02:22 I think it can be done after do_deploy_to right? Feb 19 12:55:41 03Frans Meulenbroeks  07org.openembedded.dev * r984fd8b83a 10openembedded.git/recipes/file/ (file-5.04/reloc.patch file_5.04.bb file_5.05.bb): Feb 19 12:55:41 file: moved to 5.05 Feb 19 12:55:41 Signed-off-by: Frans Meulenbroeks Feb 19 13:01:10 03Otavio Salvador  07org.openembedded.dev * rd8a8a01d2b 10openembedded.git/conf/distro/micro.conf: Feb 19 13:01:10 distro/micro.conf: drop exec_prefix_native since it is unused Feb 19 13:01:10 Signed-off-by: Otavio Salvador Feb 19 13:01:10 Acked-by: Khem Raj Feb 19 13:01:10 Signed-off-by: Frans Meulenbroeks Feb 19 13:01:52 03Otavio Salvador  07org.openembedded.dev * r3e83cfe920 10openembedded.git/recipes/cups/ (9 files in 4 dirs): (log message trimmed) Feb 19 13:01:52 cups: drop unused versions Feb 19 13:01:52 Drop versions not being in use: Feb 19 13:01:52 - 1.1.23 and 1.2.7 - old and not preferred by any distro Feb 19 13:01:52 - 1.3.8 and 1.4.3 - D_P -1 and not preferred by any distro Feb 19 13:01:52 otavio: pushed most of your patches, but not the cmake ones Feb 19 13:01:52 Signed-off-by: Otavio Salvador Feb 19 13:01:53 Acked-by: Frans Meulenbroeks Feb 19 13:02:07 03Otavio Salvador  07org.openembedded.dev * r84b2080965 10openembedded.git/recipes/cups/ (cups.inc cups_1.2.12.bb): Feb 19 13:02:08 cups: merge cups.inc into cups 1.2.12 recipe Feb 19 13:02:08 Signed-off-by: Otavio Salvador Feb 19 13:02:08 Acked-by: Frans Meulenbroeks Feb 19 13:02:08 Signed-off-by: Frans Meulenbroeks Feb 19 13:02:17 03Otavio Salvador  07org.openembedded.dev * r3c1bb8fdd7 10openembedded.git/recipes/python/python-dbus_0.83.0.bb: Feb 19 13:02:17 python-dbus: update to 0.83.2 and fix rpath to use absolute path Feb 19 13:02:17 In case prefix and exec_prefix are empty, rpath can end up with no Feb 19 13:02:17 absolute paths thus failing to build due a linking issue. Feb 19 13:02:17 Signed-off-by: Otavio Salvador Feb 19 13:02:17 Signed-off-by: Frans Meulenbroeks Feb 19 13:02:18 eFfeM: due any specific reason? Feb 19 13:02:20 03Otavio Salvador  07org.openembedded.dev * rdb688aaa48 10openembedded.git/recipes/qt4/qt4-tools-sdk.inc: Feb 19 13:02:20 qt4-tools-sdk: use host include headers for dbus Feb 19 13:02:20 Signed-off-by: Otavio Salvador Feb 19 13:02:20 Acked-by: Khem Raj Feb 19 13:02:20 Signed-off-by: Frans Meulenbroeks Feb 19 13:02:39 03Otavio Salvador  07org.openembedded.dev * rc90af4040c 10openembedded.git/classes/sdk.bbclass: Feb 19 13:02:40 sdk.bbclass: cache ${libdir} as ${target_libdir} Feb 19 13:02:40 Signed-off-by: Otavio Salvador Feb 19 13:02:40 Signed-off-by: Frans Meulenbroeks Feb 19 13:02:47 03Otavio Salvador  07org.openembedded.dev * r917203ed7e 10openembedded.git/recipes/xkeyboard-config/ (4 files): Feb 19 13:02:47 xkeyboard-config: avoid recipe code duplication and depends on intltool-native Feb 19 13:02:47 Signed-off-by: Otavio Salvador Feb 19 13:02:47 Acked-by: Khem Raj Feb 19 13:02:47 Signed-off-by: Frans Meulenbroeks Feb 19 13:03:10 03Otavio Salvador  07org.openembedded.dev * r781b459ec2 10openembedded.git/recipes/cups/ (cups14.inc cups_1.4.4.bb): Feb 19 13:03:10 cups: merge cups14.inc into cups 1.4.4 recipe Feb 19 13:03:10 Signed-off-by: Otavio Salvador Feb 19 13:03:10 Acked-by: Frans Meulenbroeks Feb 19 13:03:10 Signed-off-by: Frans Meulenbroeks Feb 19 13:03:26 03Otavio Salvador  07org.openembedded.dev * r92f66ee0a4 10openembedded.git/recipes/libinih/libinih_git.bb: Feb 19 13:03:26 libinih: add Feb 19 13:03:26 This provides a C and C++ INI library that can be statically linked Feb 19 13:03:26 into projects. Feb 19 13:03:27 Signed-off-by: Otavio Salvador Feb 19 13:03:27 Acked-by: Khem Raj Feb 19 13:03:28 Signed-off-by: Frans Meulenbroeks Feb 19 13:03:38 03Otavio Salvador  07org.openembedded.dev * r9ffa0df733 10openembedded.git/recipes/cups/ (4 files in 2 dirs): Feb 19 13:03:38 cups: add textonly filter and split packages Feb 19 13:03:38 Signed-off-by: Otavio Salvador Feb 19 13:03:38 Acked-by: Frans Meulenbroeks Feb 19 13:03:38 Signed-off-by: Frans Meulenbroeks Feb 19 13:06:00 otavio: the cmake one did not have acks, the commit policy requires two acks for toolchain related stuff, and also I don't have too much knowledge on cmake, so can't really judge them Feb 19 13:08:26 * otavio rebases his tree to see what is pending Feb 19 13:08:58 eFfeM: good; Nice improvement since yestarday :-D Feb 19 13:11:43 :-) Feb 19 15:59:15 what bitbake version do I have to use and what OE branch to successfully build my project? Feb 19 16:01:20 I'm using bitbake 1.8.18. is it ok? Feb 19 16:05:11 Have question regarding post-installation scripts. Feb 19 16:05:37 I am building OS image and it looks like I have to clean those packages that has post-install scripts? Feb 19 16:06:10 GNUtoo|laptop: hya Feb 19 16:07:15 hi Feb 19 16:07:40 GNUtoo|laptop: WiFi AP works without any problems Feb 19 16:07:55 2.6.36 + latest stuff I had Feb 19 16:08:21 wow I tried AP too Feb 19 16:08:28 I finally get hostapd compile Feb 19 16:08:29 eFfeM: here? Feb 19 16:08:37 I've 2.6.30 Feb 19 16:08:40 which chip already? Feb 19 16:08:45 libertas_sdio_tf? Feb 19 16:08:49 yeah Feb 19 16:08:58 nice Feb 19 16:09:06 you ported it to 2.6.36? Feb 19 16:09:09 Tried like 10 times and it worked Feb 19 16:09:13 Yup Feb 19 16:09:17 buglabs would be *very* interested Feb 19 16:09:23 As latest version requires 2.6.36+ kernel Feb 19 16:09:24 they had a lot of issues with libertas_tf Feb 19 16:09:29 ok Feb 19 16:09:37 jconnolly|away, ^^^ Feb 19 16:09:51 grund, ^^^ Feb 19 16:10:02 so what's needed to make it work? Feb 19 16:10:08 2.6.36+ and? Feb 19 16:11:32 2.6.36 + latest libertas_tf drivers from Steve deRosier new driver repo + firmware he send me Feb 19 16:11:45 dv, at minimum you need 1.10.2 but probably better go for 1.12 Feb 19 16:11:52 that is if you are usign head Feb 19 16:12:32 1.12 released?! Feb 19 16:12:40 davidlt: i'm not sure about your Q, what did you want to clean and why Feb 19 16:12:44 yes Feb 19 16:13:22 if you want to build a new kernel eg because you locally changed your defconfig; do a bitbake -cclean virtual/kernel;bitbake virtual/kernel Feb 19 16:14:26 eFfeM: I made additional packages which holds bunch of configurations files for OS. I want them to overwrite old one. So I added that postinstall stuff and it only works then if this package was not build before. (clean OS image, build OS image -> no postinstall steps) (clean OS image, clean my package, build OS image -> postinstall executed) Feb 19 16:15:47 http://article.gmane.org/gmane.comp.handhelds.openembedded/42856/match=1.12.0 Feb 19 16:16:39 if you have a package that depends on kernel output you need to remove it Feb 19 16:16:54 if it is part of the kernel building cleaning and rebuilding the kenrel will fix it Feb 19 16:17:20 davidlt, ok Feb 19 16:18:24 eFfeM: my package contains such configs are syslog-ng.conf, named.conf, dhcpd.conf, etc. Feb 19 16:20:04 why would you wnat to clean them, these are independent of your kernel Feb 19 16:21:26 So this packages copies config files(.tmp) to required location and I have pkg_postinst for package that replaces current config files with new ones on build rootfs. Feb 19 16:21:56 But it looks like that I have to clean my package (contains configuration files) each time I build OS image. Feb 19 16:22:11 Otherwise pkd_postinst is not executed on build rootfs Feb 19 16:26:37 davidlt: better put these conf files in a separate package and add that to your image Feb 19 16:26:56 the os postinst should not install those conf files Feb 19 16:27:05 eFfeM: So I did that. Feb 19 16:27:07 that is, the linux kernel postinst Feb 19 16:27:12 postinst is on that package Feb 19 16:28:01 and I add that conf package to OS image Feb 19 16:28:19 eFfeM, how can I get bitbake if http://developer.berlios.de/projects/bitbake/ is down? Feb 19 16:28:27 davidlt: don't use postinst, rewrite your package that e.g. dns.conf is in the SRC_URI just like a patch is, then add it to FILES_${PN} Feb 19 16:28:47 But postinst from conf package is only executed if conf package is cleaned before Feb 19 16:29:09 You get conflicts Feb 19 16:29:10 dv: http://cgit.openembedded.org/cgit.cgi/bitbake/ Feb 19 16:29:27 As some of configs might already exist Feb 19 16:29:42 davidlt: there are ways to handle conflicts, don't know details though Feb 19 16:31:29 afk (food time) Feb 19 16:35:28 error: pathspec '1.12.0' did not match any file(s) known to git. Feb 19 16:35:30 hmm.. Feb 19 16:54:18 Hmm Feb 19 16:54:35 micro only seems buildable for meta-toolchain and minimal-image Feb 19 17:00:15 davidlt: http://cgit.openembedded.org/cgit.cgi/bitbake/snapshot/bitbake-1.12.0.tar.gz Feb 19 17:00:27 eFfeM: already running 1.12 Feb 19 17:28:43 hi, Feb 19 17:28:54 base-files_3.0.14.bb has: Feb 19 17:29:00 PR = "r99" Feb 19 17:29:07 how do I bump PR? Feb 19 17:29:13 is r100 > 99? Feb 19 17:29:27 r100 > r99 Feb 19 17:29:36 I bet so Feb 19 17:29:40 but it's ultra-core stuff Feb 19 17:43:08 davidlt, where did you clone the libertastf? Feb 19 17:43:22 http://dev.laptop.org/git/users/javier/libertastf.git/ dates from 2008 Feb 19 17:43:36 GNUtoo|laptop: http://dev.laptop.org/git/users/derosier/wireless-testing/ Feb 19 17:43:37 GNUtoo|laptop: it should be, you can ask dpkg to compare versions. same with opkg Feb 19 17:43:51 ok Feb 19 17:43:54 thanks Feb 19 17:44:11 GNUtoo|laptop: There should be some additional patches, but Steve didn't have time to include them Feb 19 17:44:23 davidlt, ok Feb 19 17:45:34 There is also new firmware as the old one has several serious bugs Feb 19 17:45:40 ahhh ok Feb 19 17:45:55 GNUtoo|laptop: I was having lots of problems with old version, was quite unstable Feb 19 17:46:04 maybe that's why..... Feb 19 17:46:21 buglabs people had issues with libertas_tf_sdio Feb 19 17:46:26 maybe because of the firmware Feb 19 17:46:49 " + firmware he send me" Feb 19 17:46:50 The old firmware might stop responding Feb 19 17:46:57 is that firmware public Feb 19 17:46:58 Especially if you bring up/down interfaces Feb 19 17:47:01 ? Feb 19 17:47:03 ok Feb 19 17:47:13 He said that it was unpublished Feb 19 17:47:20 ok Feb 19 17:47:24 how to get it? Feb 19 17:47:28 and how buglabs could get it? Feb 19 17:47:35 I think he send details to Gumstix mailing-list too Feb 19 17:47:44 Give me email and I send you Feb 19 17:48:01 GNUtoo A no-log.org Feb 19 17:48:12 replace the A wiht @ Feb 19 17:48:42 but could this firmware be used in commercial devices Feb 19 17:48:45 like the bug20? Feb 19 17:50:51 I'm not from buglabs btw, but I've good relation with them Feb 19 17:51:10 You should probably read license on web Feb 19 17:51:14 email -> done Feb 19 17:51:35 ok thanks a lot Feb 19 17:51:51 I'll try to import the sdio stuff to linux-next Feb 19 17:52:08 and then I'll try it Feb 19 17:52:22 but buglabs people have already libertas_tf_sdio working on bug20 Feb 19 17:52:30 only that it didn't work well Feb 19 17:52:34 I bet because of the firmware Feb 19 17:52:43 does sta work? Feb 19 17:53:28 STA stands for ? Feb 19 17:53:39 station Feb 19 17:53:46 normal mode Feb 19 17:53:51 that is you connect to an AP Feb 19 17:54:13 Haven't tried that, but it worked on all version without problems Feb 19 17:54:19 ok Feb 19 17:54:24 Gonna to full scale test in few days Feb 19 17:54:42 Just have to manage my configuration packages which don't want to work nicely :) Feb 19 17:55:10 Feb 16 16:46:11 normal 'managed' mode connections would freeze and drop Feb 19 17:55:44 didn't had this problem with old stuff, but as I said firmware could stop responding Feb 19 17:55:50 At leas the old one Feb 19 17:55:58 ok Feb 19 17:56:25 I tested in on master mode like 10 times and worked fine Feb 19 17:56:40 AP + DHCPD + BIND and it fastastic for now Feb 19 17:56:42 yes you told it Feb 19 17:56:42 ok Feb 19 17:56:45 *fantastic Feb 19 17:57:00 ah standalone setup? Feb 19 17:57:05 Yeah :) Feb 19 17:57:07 ok Feb 19 17:57:15 because dnsmasq can only froward Feb 19 17:57:21 the dns of your provider Feb 19 17:57:25 Now I can directly with iPhone connect to Smart Meter data via Gumstix Feb 19 17:57:30 + have some local stuff but limited Feb 19 18:03:02 I hate postinst Feb 19 18:04:58 Any other ways to overwrite files? Feb 19 18:05:13 It looks like the most reliable way is to overwrite on first boot Feb 19 18:11:02 davidlt, why do you want to override files? Feb 19 18:12:20 I don't need default configurations, I have custom interfaces, syslog, DNS, DHCP and etc configurations. Feb 19 18:12:41 why not using the overrdes? Feb 19 18:12:52 And there are several of them by changing one package you get different behavior of OS now. Feb 19 18:13:00 ah ok Feb 19 18:13:04 overrdes ? Feb 19 18:13:06 but you will have a big problem Feb 19 18:13:13 like for instance: Feb 19 18:13:47 recipes/xorg-xserver/xserver-xorg-conf Feb 19 18:13:56 theses are overriden files Feb 19 18:14:05 from file:// in SRC_URI Feb 19 18:14:21 because else....the file won't be handled by the package manager Feb 19 18:14:33 and an upgrade would overwrite the files Feb 19 18:14:38 or something like that Feb 19 18:16:27 right now I get conflicts because old configs exists and opkg does not nothing, but gives error Feb 19 18:16:57 Before that configs are included in OS with additional extension and on first boot files are changed and services restarted. Feb 19 18:17:24 So I am looking how to change those configs file before rootfs is cooked as image Feb 19 18:17:39 postinst works, but I can't get it. Feb 19 18:17:45 It doesn't work always Feb 19 18:18:08 So maybe sticking to 'on first boot' method is best choice Feb 19 18:18:51 I see that on one package it worked, on the other it didn't and I don't know why... Feb 19 18:21:10 I could do rootfs change manually everything and after that run build to make image Feb 19 18:21:19 But still doesn't sound nice Feb 19 18:43:20 hi Tartarus Feb 19 18:44:17 http://lists.linuxtogo.org/pipermail/openembedded-devel/2011-February/030140.html + http://patches.openembedded.org/patch/719/ fix all rpath problems for armv5+angstrom 2008.1 + custom & qt4e-demo-image Feb 19 18:45:17 ericben when you use the second patch Feb 19 18:45:20 -disable-rpath \ Feb 19 18:45:23 isnt needed Feb 19 18:45:26 I meant the sed Feb 19 19:38:10 davidlt, hi Feb 19 19:38:18 libertas_tf_sdio: failed to load helper firmware Feb 19 19:38:23 do I need a helper firmware Feb 19 19:38:30 or just a tf firmware? Feb 19 19:38:46 you need helper Feb 19 19:38:57 use the same from libertas-sd-firmware Feb 19 19:38:57 the standard one? Feb 19 19:39:04 ok thanks a lot Feb 19 19:39:27 I will probably later write instructions and add patches on Gumstix Wiki Feb 19 19:39:31 ok Feb 19 19:40:00 the patches apply nicely on linux-next => will someone send them to mainstream? Feb 19 19:40:21 I think, I found my problem with postint: empty else body in one place Feb 19 19:40:45 this is alpha driver (still) I doubt they would be included Feb 19 19:41:16 but sdio should be added into kernel 38 or later Feb 19 19:41:33 wow Feb 19 19:42:00 He told me that he is going to push that to kernel, to update current driver, but probably not soon Feb 19 19:42:03 sd8686tf_helper.bin is the right name for the helper? Feb 19 19:42:19 You can keep old name, name doesn't matter Feb 19 19:42:39 yes but I want to load it automatically with udev, not manually trough /sys/class/firmware Feb 19 19:42:51 and I've old name and it didn't load it Feb 19 19:43:01 I'll try with cp instead of ln -s Feb 19 19:43:08 I left it exactly as in litbertas-sd-firmware, adding 'tf' is not required Feb 19 19:44:02 cp litbertas-sd-firmware litbertastf-sd-firmware Feb 19 19:44:07 make small changes Feb 19 19:44:18 change firmware in image and that's all Feb 19 19:45:48 om the archive I have: sd8686tf.bin Feb 19 19:48:25 Yeah, that's the firmware and use helper from libertas (old one) Feb 19 19:49:22 ok Feb 19 19:49:29 I've that one: Feb 19 19:49:49 sd8686_helper.bin Feb 19 19:50:01 I'll look if buglabs has some infos as they use the same on bug20 Feb 19 19:50:15 I have never used plain OE, just Gumstix OE branch, but maybe I could try sending paches for kernel, firmware recipe, etc. Feb 19 19:52:08 yes Feb 19 19:52:11 that could be good Feb 19 19:52:50 I have other stuff that I could publish too probably Feb 19 19:52:53 Never tried Feb 19 19:53:16 There was wiki page I think how to publish patches to OE Feb 19 19:53:21 yes Feb 19 19:53:31 http://wiki.openembedded.org/index.php/How_to_submit_a_patch_to_OpenEmbedded Feb 19 19:53:51 and still this is alpha, so it might be rejected Feb 19 19:54:12 can you pastebin me the content of: Feb 19 19:54:15 find /lib/firmware ? Feb 19 19:55:44 Am... Probably not, I run tree build environment and I lack HDD space, so I always delete on of rootfs Feb 19 19:55:55 s/on/one/ Feb 19 19:56:05 yes but I can't get the name of the firmwares right Feb 19 19:56:25 Tartarus: I'm building iscsi-target to see if I get the QA issue for calamari. I haven't seen it for sheevaplug (but actually I have never build iscsi-target for either neek or calamari before) Feb 19 19:57:00 will take some time though as my rusty old core2duo needs to rebuild gcc, linux, libs etc Feb 19 19:57:38 GNUtoo|laptop: I can send you recipe + files Feb 19 19:57:49 ok thanks Feb 19 19:59:56 GNUtoo|laptop: done Feb 19 20:00:04 thanks a lot Feb 19 20:05:02 Proof that it works: http://img510.imageshack.us/img510/6561/iphonegumstixwap1.png Feb 19 20:06:46 by any chance are you guys using the PSP kernel for the gumstix? Feb 19 20:08:46 linux-omap3 is used Feb 19 20:09:00 But this naming I think is only on Gumstix OE branch Feb 19 20:09:57 and it is changed to linux-omap recently as omap3 and omap4 should be merged into one Feb 19 20:10:40 davidlt, btw about the recipe of the firmwares Feb 19 20:10:58 look in recipes/firmwares Feb 19 20:11:05 there is already one for some marvell firmwares Feb 19 20:11:10 some toughs: Feb 19 20:11:15 *there is a marvell license in oe Feb 19 20:11:20 use that if the license is the same Feb 19 20:11:30 and don't put LICENSE="unknown" Feb 19 20:11:46 then be carefull of the name clash Feb 19 20:11:50 that is to say Feb 19 20:11:56 ah bummer, i'm using linux-omap-psp Feb 19 20:12:01 make it possible to install 2 marvell firmwares in the rootfs Feb 19 20:12:27 for instance split the sd8686_helper.bin Feb 19 20:12:45 that is to say, make a recipe specially for that Feb 19 20:12:52 and depend on it Feb 19 20:13:25 Yeah, we could do it. There is license file for it on libertas-sd-firmware (I just deleted it for now :) ) Feb 19 20:13:48 And I don't know about libertastf firmware license Feb 19 20:13:56 ok Feb 19 20:14:04 we can't include stuff without license Feb 19 20:14:14 because else that means that it's not redistributable Feb 19 20:15:44 There is somewhere license, you just have to look for it Feb 19 20:15:58 ok Feb 19 20:16:14 libertas_tf_sdio: failed to load helper firmware Feb 19 20:16:14 libertas_tf_sdio: failed to load helper firmware Feb 19 20:16:15 hmmm Feb 19 20:16:53 I'll try to disable udev loading of firmwares Feb 19 20:16:56 and load it manually Feb 19 20:17:10 altough the files match in uevent Feb 19 20:17:13 *filenames Feb 19 20:21:23 how is it to have mac os x as host ? the wiki talks not too good about it ! Feb 19 20:21:46 rob_w: I would like that too! Feb 19 20:22:02 I think the problem is that not all software required for Bitbake is ported to Mac Feb 19 20:22:27 Maybe I will try one day to test it Feb 19 20:22:42 but as soon as all host and -native build it could work ? Feb 19 20:22:50 But I have Mac Pro, which would be great for build environment Feb 19 20:23:27 strange I load the firmware manually Feb 19 20:23:28 its not that i want that , but a collegue who is also a apple addict Feb 19 20:23:30 then nothing happens Feb 19 20:24:02 Probably, but some tools (I think) is hardware/kernel specific things, which are not easy to port Feb 19 20:24:10 GNUtoo|laptop: so it works? Feb 19 20:24:18 no Feb 19 20:24:21 no interface Feb 19 20:24:25 then nothing happens Feb 19 20:24:35 nothing in dmesg Feb 19 20:24:44 You did everything right? driver patched? kconfig patches? reconfigured kernel and etc? Feb 19 20:24:52 altough libertas_sdio worked Feb 19 20:24:57 yes Feb 19 20:25:02 I've ported all the patches Feb 19 20:25:08 git cherry-pick Feb 19 20:25:11 every single patch Feb 19 20:25:14 ahha Feb 19 20:25:16 wait a sec Feb 19 20:25:20 hang task Feb 19 20:25:40 http://pastebin.com/c1Hk94ae Feb 19 20:25:52 rob_w: come to Swiss, Apple is number one here Feb 19 20:26:06 invite me ! Feb 19 20:26:19 * GNUtoo|laptop prefers argentina where GNU/Linux is number one Feb 19 20:27:15 maybe the platform stuff? Feb 19 20:27:25 hi GNUtoo|laptop : is you rlibertasdriver compiled in kernel or as a module ? Feb 19 20:27:31 in the kenrel Feb 19 20:27:35 Hehe, I think Bahnofstrasse in Zurich just became the most expensive street in the world Feb 19 20:27:37 but it's libertastf_sdio Feb 19 20:27:40 not libertas_sdio Feb 19 20:27:45 hi btw Feb 19 20:28:08 try as a module. I had problems to load firmware as the timeout was reached and the filesystem was not mounted Feb 19 20:28:17 ok Feb 19 20:28:28 timeout as like in loading the firmware? Feb 19 20:28:35 I loaded the firmwares manually Feb 19 20:28:40 timeout was not reached Feb 19 20:28:55 because I echo 3600 to /sys/class/firmware/timeout as soon as I log in Feb 19 20:28:56 is your filesystem already present when it tries to load the firmware ? Feb 19 20:28:58 then I do that: Feb 19 20:29:01 yes Feb 19 20:29:05 as I do it manually Feb 19 20:29:06 ah ok Feb 19 20:29:08 like that: Feb 19 20:29:13 cd /sys/class/firmware/ Feb 19 20:29:16 you filrst load the loader ? Feb 19 20:29:17 cd mmc* Feb 19 20:29:22 echo 1 > loading Feb 19 20:29:36 cat /lib/firmware/foo_helper.bin Feb 19 20:29:40 oops Feb 19 20:29:48 I'll restart: Feb 19 20:29:58 cat uevent Feb 19 20:30:02 for getting the file name Feb 19 20:30:03 then Feb 19 20:30:07 echo 1 > loading Feb 19 20:30:13 cat /lib/firmware/foo_helper.bin > data Feb 19 20:30:16 echo 0 > loading Feb 19 20:30:17 cd .. Feb 19 20:30:23 cd mmc* Feb 19 20:30:45 echo 1 > loading; cat /lib/firmware/footf.bin> data;echo 0 > loading Feb 19 20:30:45 and th esame thing with the firmware Feb 19 20:30:46 cd .. Feb 19 20:30:54 yes Feb 19 20:30:57 that should work Feb 19 20:31:04 it works for the normal driver Feb 19 20:31:05 it's wired using spi ? Feb 19 20:31:11 but not for libertas thin firmware sdio Feb 19 20:31:17 ah ok Feb 19 20:31:23 never used this one Feb 19 20:31:24 that has AP Feb 19 20:31:37 basically with that one you can make an AP Feb 19 20:31:50 but it seems that you need a not-yet-released firmware to get it stable Feb 19 20:31:56 I understand Feb 19 20:31:57 which davidlt gave me Feb 19 20:32:16 this works with 8686 chipsets ? Feb 19 20:32:18 I *must* tell buglabs about it Feb 19 20:32:19 yes Feb 19 20:32:25 buglabs on their bug20 had that Feb 19 20:32:30 but I bet they had the old firmware Feb 19 20:32:39 so it wasn't stable in station/managed mode Feb 19 20:32:54 so they now use libertas_sdio by default Feb 19 20:34:04 ericben, btw I've got 1 ack for the bugbase support in linux-next: Acked-by: Uwe Kleine-König Feb 19 20:34:11 I did stupid stuff Feb 19 20:34:24 because I always did that as the last stuff before going to bed.... Feb 19 20:34:43 so I was a bit tired and forgott to use whitespace.el during the v2 Feb 19 20:35:52 I also forgott a , at the end of an array Feb 19 20:36:12 that is not very professional..... Feb 19 20:36:22 but I've a lot of stuff to do.... Feb 19 20:37:13 GNUtoo|laptop: I(ve seens on arm-kernel ml Feb 19 20:37:16 ok Feb 19 20:37:24 so that will be in 2.6.39 Feb 19 20:37:37 I bet 1 ack is not enough Feb 19 20:37:59 GNUtoo|laptop: it will be taken by Sascha in its tree Feb 19 20:38:04 ah ok nice Feb 19 20:38:11 I think Feb 19 20:38:20 I think I should get a bug20 Feb 19 20:38:31 ping them in a few time if that's not the case Feb 19 20:38:36 ok Feb 19 20:38:36 the one with omap ? Feb 19 20:38:39 yes Feb 19 20:39:09 they say it has less hardware issues Feb 19 20:39:40 http://pastebin.com/qDZBkVLX Feb 19 20:39:49 bitbake 1.12.0, oe master Feb 19 20:39:59 lot of rpaths issues it seem Feb 19 20:40:09 how to fix it? Feb 19 20:40:16 I had some rpath issues with boost Feb 19 20:40:18 I abandoned Feb 19 20:40:33 as I didn't have it in my nokia900 tmpdir Feb 19 20:40:51 basically bosot uses jam Feb 19 20:40:51 dv : http://lists.linuxtogo.org/pipermail/openembedded-devel/2011-February/030140.html + http://patches.openembedded.org/patch/719/ Feb 19 20:40:54 which isn't great Feb 19 20:41:05 see the comments in the files..... Feb 19 20:41:21 ah nice Feb 19 20:41:39 ok back to sysadmin tasks, not fun but I have to do that this weekend :) Feb 19 20:41:41 ye Feb 19 20:41:43 bye Feb 19 20:41:51 lol ok bye Feb 19 20:42:01 I've to do some sysadmin stuff from time to time too Feb 19 20:44:46 off-topic: any one has netbook? Feb 19 20:45:45 davidlt, asus eee Feb 19 20:46:02 dv: was it worth the money as secondary computer? Feb 19 20:46:29 btw.. should do some sysadmin tasks too Feb 19 20:47:44 davidlt, It's my doughter pc.If I'd need portable pc, I'd bay it for myself too Feb 19 20:48:54 I always have computer with myself, so I am thinking of getting something lighter, more battery life and 3G. Feb 19 20:49:16 As you always can get magic call saying that there is urgent demo and we need something fixed asap. Feb 19 20:50:17 davidlt, eee701, supported by oe Feb 19 20:50:28 Lenovo ThinkPad X120e looks amazing Feb 19 20:50:37 davidlt, urgent demo? seems like design issue Feb 19 20:51:15 dv: there are tons of issues :D But I can rewrite the whole stuff it wasn't even my job Feb 19 20:51:42 davidlt, depends on task. for admin tasks eee is the best choice Feb 19 20:52:27 I remember those 7" eee pc, didn't look look or feel good Feb 19 20:52:56 We actually even tried running Netbeans once on them (had to make modifications), that was funny Feb 19 20:54:59 So far that ThinkPad looks amazing, holds ~5:30 as remember, 3 usb ports, 1 of them allows to charge devices without turned machine, AMD Zacate, 3G, etc. Feb 19 21:02:16 If I want to run 'update-rc.d' in my recipe, do I have to inherit update-rc.d Feb 19 21:02:19 ? Feb 19 21:03:20 http://pastebin.com/qDZBkVLX Feb 19 21:03:27 don't know how to fix this issue Feb 19 21:03:42 what do I have to update? Feb 19 21:04:06 dv: I answered you a few minutes ago Feb 19 21:04:07 dhcp-server is not started during boot Feb 19 21:05:11 dv: apply the patches, rebuild gcc-cross (at least) then rebuild gnutls Feb 19 21:05:44 ericben, that two patches??? Feb 19 21:06:21 Jay7, ping? Feb 19 21:06:29 ka6sox-away: pong Feb 19 21:06:36 davidlt: yes check recipes/watchdog/watchdog_5.6.bb for example Feb 19 21:06:44 Jay7, know any java programmers? Feb 19 21:06:53 dv: at least the one from patchwork Feb 19 21:07:10 ka6sox-away: I know one Feb 19 21:07:19 what do you needed? Feb 19 21:07:34 Jenkins plugin Feb 19 21:07:55 ericben, fantastic :) Feb 19 21:08:19 ka6sox-away: have you some description ready? Feb 19 21:08:38 * mwester mutters about using a tool that was written by developers for developers, instead. Feb 19 21:09:09 mwester, you talking to ME? Feb 19 21:09:19 he he Feb 19 21:09:21 Not specifically. Just generally. Feb 19 21:09:28 Java == COBOL. Feb 19 21:09:44 mwester, I had to LEARN COBOL...it was required Feb 19 21:09:44 It has its place, and that place is as far away from build tools as is possilble. :D Feb 19 21:09:57 Yeah, I learned COBOL. and RPG. Feb 19 21:10:05 PL/1 too Feb 19 21:10:11 PL/1 Subset G Feb 19 21:10:16 righ Feb 19 21:10:17 t Feb 19 21:10:45 Which is why I'm qualified to pass judgement on the worth of COBOL and JAVA as integration languages. Feb 19 21:10:50 They are not. Feb 19 21:11:29 mwester, I"d love to use something else but I'm sure they dont' have clean interfaces. Feb 19 21:11:33 I agree, java is overkill for build tools Feb 19 21:11:40 But, apparently, it's the ONLY language taught in schools anymore, so that's what the new kids at work think of when they think "programming"... which isn't so bad, except that they also think "JAVA" when someone says scripting. Feb 19 21:11:43 we have python already :) Feb 19 21:12:02 Bingo, Jay7 -- that's a great language. Feb 19 21:12:07 Jay7, not sure if the interfaces are clean enough Feb 19 21:12:42 I dislike both, but we have python required for bitbake Feb 19 21:12:43 so.. Feb 19 21:13:01 but we have no jenkins/hudson written in python Feb 19 21:13:14 Jay7, mwester we need to build statemachines I fear to deal with making jenkins happy. Feb 19 21:13:15 It would *really* annoy me if I had to have a full java on my buildhost, just to report stats. That would really SUCK. Feb 19 21:13:36 mwester: +1 Feb 19 21:13:49 hi kas6sox confrenece started? Feb 19 21:14:00 okay if we have a better tool lets find it. Feb 19 21:14:03 mwester: btw, there is more.. you will report stats from python bbclass anyway :) Feb 19 21:14:07 woglinde, I go thursday to setup. Feb 19 21:14:21 but starts friday-sunday Feb 19 21:14:34 so you have some time left Feb 19 21:14:38 which will be posted to your local jenkins via HTTP POST, which will POST it to master jenkins Feb 19 21:14:40 "Let's see -- 4GB of RAM, let's give a GB to the JRE and hope it doesn't run out of memory, not because the app does anything with the memory, but because today's programmers are taught that they don't have to take care with memory management, apparently the almighty and magical JAVA garbage collector will solve all! - Bah humbug! :D :P Feb 19 21:15:20 I need a local "jenkins" too?!!!! No chance of that. Feb 19 21:15:27 Ridiculous. Feb 19 21:15:57 Like making me buy a Range Rover -- so that I can get to the bus stop to catch a bus... makes no sense to me. Feb 19 21:16:30 * mwester resigns himself to not being able to report on builds from his local builder. :( Feb 19 21:16:52 okay if this isn't practical I need something that *is*. because the current setup is broken. Feb 19 21:17:14 If all we're lacking is someone to write some code, someone please post some specs somewhere. Feb 19 21:17:20 the is an API for hudson and example in python : http://wiki.hudson-ci.org/display/HUDSON/Remote+access+API Feb 19 21:17:41 hudson == Jenkins ATP Feb 19 21:17:46 so that would work. Feb 19 21:17:48 but... Feb 19 21:18:01 the JVM is a question mark. Feb 19 21:18:04 yes, if you prefere here is the link on jenkins :-D http://wiki.jenkins-ci.org/display/JENKINS/Remote+access+API Feb 19 21:19:00 The JVM is a big problem IMO. I want all my RAM to go to building, not set aside for possible use by a java program. Feb 19 21:19:45 mwester: easy way is rewrite current oestats-client to store stats into bb persistent storage Feb 19 21:19:54 and post it after build Feb 19 21:20:07 then adapt server side to deal with this Feb 19 21:20:25 My build machine is only a 32-bit machine, and the motherboard only supports 2GB RAM. I would like to upgrade that, but it would annoy me if I was *forced* to upgrade that just because I had to have a JVM running all the time during the build! I'm probalby the exception, though. Feb 19 21:20:53 Jay7, I'll take a look at it. Something for the airplane. Feb 19 21:21:17 I don't know Python very well, but it can't be that hard. Feb 19 21:21:22 mwester: most of my customeer run OE in a virtualbox under windows so you are not an exception ;-) Feb 19 21:21:50 03Denis 'GNUtoo' Carikli  07org.openembedded.dev * r92f586853e 10openembedded.git/recipes/base-files/ (base-files/om-gta02/fstab base-files_3.0.14.bb): (log message trimmed) Feb 19 21:21:50 base-files: update om-gta02's fstab and bump PR Feb 19 21:21:50 The update is based on the default fstab, Feb 19 21:21:50 it fixes a long standing bug where the rootfs was hardcoded Feb 19 21:21:50 to /dev/mtdblock6 with the jffs2 filesystem. Feb 19 21:21:50 The om-gta02 machine can boot from microsd or from mtd(jffs2 or ubifs), Feb 19 21:21:51 so not hardcoding the defult rootfs is good. Feb 19 21:22:38 * mwester goes back to trying to rescue his MythTV server so he can watch the news this evening... Feb 19 21:23:38 mwester: read first 7 sections of tutorial and you ready to go with Python Feb 19 21:23:55 there is bb's persistent storage Feb 19 21:24:11 not plain python but bb internals :) Feb 19 21:25:06 I think we should propose 2 RFC to ML Feb 19 21:25:30 That's probably a good idea Jay7 Feb 19 21:25:40 one about enchancing current oestats and one about migration to jenkins Feb 19 21:25:49 and let the battle begins Feb 19 21:25:53 Yep. Feb 19 21:26:13 Jay7: both can exist Feb 19 21:26:16 As I understand it, we will have *no* oestats until we fix it, right? Feb 19 21:26:48 I can live with jenkins when it will be running on server-side :) Feb 19 21:27:00 mwester: yep Feb 19 21:27:05 we already have no oestats Feb 19 21:27:28 mwester: I submitted a patch which could help to recover oestat : http://lists.linuxtogo.org/pipermail/openembedded-devel/2011-February/030136.html Feb 19 21:27:41 so we may adapt current oestats-client to post data to jenkins server Feb 19 21:27:59 but if jenkins can process it Feb 19 21:28:04 mwester, that is correct..till it gets resolved we can't run oestats Feb 19 21:28:22 Jay7: current oestat is interesting for existing installations (private tinderbox fore xample) Feb 19 21:28:30 mwester: I have conversation with kergoth about new scheme stored somewhere Feb 19 21:28:43 if you are interested I may pastebin it Feb 19 21:28:55 to I think it's better to start a jenkins-client.bbclass Feb 19 21:29:13 or something like this if jenkins is chosen Feb 19 21:29:45 well.. may be good idea too Feb 19 21:30:19 ericben, that is 1/2 the equation Feb 19 21:30:25 Jay7: actual oestat doesn't scale well for OE's community but is interesting on a company level Feb 19 21:30:50 the other half is that jenkins expects clients to remain connected during the entire build process or it says the build failed. Feb 19 21:31:31 so if it takes 15hrs for a build and in the 14th hour it loses connection it calls that build "failed" Feb 19 21:32:05 ka6sox-away: even is this case : Monitoring executions of externally-run jobs, such as cron jobs and procmail jobs, even those that are run on a remote machine. For example, with cron, all you receive is regular e-mails that capture the output, and it is up to you to look at them diligently and notice when it broke. Jenkins keeps those outputs and makes it easy for you to notice when something is wrong Feb 19 21:32:11 ? Feb 19 21:32:40 the number 2 in meet jenkins's section of their site Feb 19 21:33:27 external reporting is what jenkins-client.bbclass will be Feb 19 21:33:29 however. Feb 19 21:33:33 I've quick look at API Feb 19 21:33:47 the jenkins client still requires full time connection during build process. Feb 19 21:33:57 that is the problem. Feb 19 21:34:00 well.. seems we can use it to post tasks Feb 19 21:34:18 so we need only client class on buildhost side Feb 19 21:34:18 ka6sox-away: so I don''t undderstand how they can monitor cron or procmail jobs ? Feb 19 21:34:45 its an "external" input. Feb 19 21:34:58 and becomes information passed back thru client Feb 19 21:35:45 but the fundamental issue I have is that during a build the master and slave MUST remain connected or the master says that the build failed. Feb 19 21:36:30 so if the build is in .ru and the master is in .us then if there is a dropout of connection the build fails. Feb 19 21:36:33 ka6sox-away: is that for the case of a build farm ? Feb 19 21:36:44 yes. Feb 19 21:36:49 this is far over the case of oestat Feb 19 21:37:06 its a hudson/jenkins thing that is wrong for distibuted builds. Feb 19 21:37:18 but do we care of distributed builds ? Feb 19 21:37:23 yes Feb 19 21:37:26 we neeed reporrting Feb 19 21:37:28 ah ok Feb 19 21:37:36 right..reporting Feb 19 21:37:38 so some people care of this :) Feb 19 21:37:43 well. Feb 19 21:37:52 I will have "local" builders as well. Feb 19 21:38:06 I think there are 2 differents usage : reporting & build farm Feb 19 21:38:27 right Feb 19 21:38:27 build farm is cool and will be a nice feature to have Feb 19 21:38:48 right I have promises of help to create a "local" buildfarm. Feb 19 21:39:06 but reporting is similar to oestat and is what is actually missing Feb 19 21:39:21 and for reporting we don't need java on the build host Feb 19 21:39:37 only a class to send log to a jenkins server Feb 19 21:39:41 am I wroong here ? Feb 19 21:39:46 but because of the size of the project and number of configurations in the matrix we will have "external" for a long time. Feb 19 21:41:08 maybe we could build a "connector" that is really local and bridge the jenkins-client to jenkins @ the server end. Feb 19 21:42:06 so clients would actually be "local" @ jenkins server with a plugin that converts. Feb 19 21:42:22 but not running java on each node. Feb 19 21:42:30 and we could do it in python. Feb 19 21:43:18 means the jenkins server needs GB of RAM but thats cheaper than everyone having to upgrade their servrers. Feb 19 21:45:31 seems sane Feb 19 21:48:03 * mwester after a bit of reflection, wonders if we should collect some requirements from the community. Feb 19 21:58:39 hi guys Feb 19 21:58:40 03Marcin Juszkiewicz  07org.openembedded.dev * rc58b36f60a 10openembedded.git/contrib/marketing/oe-flyer.pdf: Feb 19 21:58:40 OE official flyer - update to version which was printed in 2009 Feb 19 21:58:40 This is probably the only one existing copy of current version. Sources Feb 19 21:58:40 should be somewhere on Mickeyl's harddrive. Feb 19 21:58:40 We are using them on all events since LinuxTag 2009 where I took all Feb 19 21:58:41 2500 of them which were printed in Poland as a donation from me to Feb 19 21:58:42 OE e.V. Feb 19 21:59:24 mickey|zzZZzz: please dig your systems and harddrives for sources of OE flyer and push it to repository. it took me some time to find PDF with current version. Feb 19 22:00:02 mickey|zzZZzz: otherwise we will have to find someone who will recreate it from scratch for next updates Feb 19 22:00:15 have a nice rest of weekend Feb 19 22:02:44 jenkins-client.bbclass however needs to batch reports and send them when a particular package is done and when a particular "section" is done Feb 19 22:02:55 so when parsing is done send Feb 19 22:03:26 that way we can keep status of where we are. Feb 19 22:05:06 so, as Crofton said, we need some working group to collect req's and discuss it Feb 19 22:14:51 Jay7, agreed...if there is better place to discuss I'd like to. Feb 19 22:21:41 ka6sox-away: I think we should raise it in ML Feb 19 22:22:03 if no one will do it before I'll send mail a bit later Feb 19 22:25:32 Jay7, that would be good as I am doing 5things at one time here. Feb 19 22:30:54 ka6sox-away: other idea to have page on wiki Feb 19 22:31:25 Jay7, do it...perfect idea Feb 19 22:31:34 we put requirements document there. Feb 19 22:31:38 and all edit. Feb 19 22:31:52 then send to ML where it is. Feb 19 22:31:56 btw, do we have disabled registration still? Feb 19 22:33:22 I'll do in hour or two Feb 19 22:35:11 Jay7, if you need one I'll create one for you. Feb 19 22:35:34 I have login Feb 19 22:35:37 but yes, till we find a plugin that allows it to NOT be automatic then we have to disable. Feb 19 22:39:52 I think we may reuse Tinderbox page Feb 19 22:40:24 because current content it is deprecated anyway Feb 19 22:43:08 heho zecke Feb 19 22:43:30 03Khem Raj  07master * r034012e13e 10openembedded.git/recipes/xinetd/ (3 files in 2 dirs): Feb 19 22:43:30 xinetd: Update recipe from 2.3.13 -> 2.3.14 Feb 19 22:43:30 * Fix autoconf so that it can be regenerated with Feb 19 22:43:30 newer autconf versions Feb 19 22:43:30 * Should work on armeb automagikally now Feb 19 22:43:30 Signed-off-by: Khem Raj Feb 19 22:49:44 And khem even updated xinetd! Feb 19 22:49:58 * mwester crosses off two items on his to-do list. Thanks khem! Feb 19 22:50:49 Jay7, I'd rather not use that page as it will NOT be oestats/tinderbox anymore...to avoid confusion Feb 19 22:51:06 hm who uses xinetd these days? Feb 19 23:03:39 * mwester is guilty of using xinetd today... perhaps someone knows a better way? Feb 19 23:49:30 http://wiki.openembedded.org/index.php/BuildStats Feb 19 23:49:40 ka6sox-away: ^^ Feb 19 23:50:18 mwester, ka6sox-away, Tartarus: please add your ideas/wishes/comments there Feb 19 23:50:50 I've not mentioned any technical solution yet there Feb 19 23:58:41 now mentioned Feb 20 00:04:13 Jay7, okay I'll look. Feb 20 00:04:28 RFC sent to ML Feb 20 00:12:10 * Jay7 -> sleep Feb 20 00:21:04 kk Feb 20 02:46:24 Tartarus, db 5.1.19 isn't quite as upwards compatible as it claims, it would seem. squid fails with the new version. Thought I'd mention that, before re-instating the 5.0.x version and pinning SlugOS to use that older version. Feb 20 02:48:18 http://pastebin.com/jc57GxFt in case anyone cares. **** ENDING LOGGING AT Sun Feb 20 02:59:57 2011