**** BEGIN LOGGING AT Sat Jul 19 02:59:57 2008 Jul 19 07:21:52 03mickeyl 07org.oe.dev * rb6a2cead... 10/ (7 files in 6 dirs): frameworkd git add configuration for a780 (needs adjusting the keys) Jul 19 08:13:22 u-boot guru needed :) Jul 19 08:14:01 zaurus u-boot guru is better :) Jul 19 09:16:06 good morning guys Jul 19 09:17:19 Laibsch: :) Jul 19 09:19:16 morning Jul 19 10:32:53 03Laibsch 07org.oe.dev * rca62c82e... 10/ (4 files in 3 dirs): Jul 19 10:32:53 tree: update to 1.5.2. Closes 3208. Jul 19 10:32:53 * upstream had silently replaced the 1.5.1 tarball with a link to the 1.5.1.1 version Jul 19 10:32:53 leading to md5sum check failure and thus triggered bug 3208 Jul 19 10:32:53 * 1.5.2 compiles fine without any further changes Jul 19 10:38:06 * * OE Bug 3208 has been RESOLVED (FIXED) by Jul 19 10:38:08 * * Tree 1.5.1 doesn't build Jul 19 10:38:10 * * http://bugs.openembedded.net/show_bug.cgi?id=3208 Jul 19 10:46:05 morning all Jul 19 10:46:10 hi rp :) Jul 19 10:46:24 * RP is soaked Jul 19 10:46:35 hui Jul 19 10:46:46 motorcycle ride? Jul 19 10:46:51 serves me right for trying to ride a motorbike in leathers in summer. All I wanted to do was fill it with fuel :/ Jul 19 11:07:24 hi ph5 Jul 19 11:07:38 hi woglinde Jul 19 11:11:56 * RP ponders another abi break... Jul 19 11:11:59 hi pH5 Jul 19 11:12:03 rp why? Jul 19 11:12:20 multimachine breaks with armv7 and armv5 builds in the same tree Jul 19 11:12:32 hm and the solution? Jul 19 11:12:55 split staging and cross into armv5 and armv7 versions Jul 19 11:13:06 ah Jul 19 11:13:13 instead of arm-bla only Jul 19 11:13:14 okay Jul 19 11:13:29 I split staging locally and that was fine but now I realise cross has to change too :/ Jul 19 11:13:57 but the cross tripple will be the same? Jul 19 11:14:03 for compiliers Jul 19 11:14:12 or armv5-libc-moo-gcc? Jul 19 11:14:16 yes, hence we need different cross directories Jul 19 11:15:01 hi RP Jul 19 11:15:27 split the builds on different machines :) Jul 19 11:54:42 Crofton: That defeats the point of multimachine ;-) Jul 19 11:56:06 * * OE Bug 2691 has been RESOLVED (FIXED) by Jul 19 11:56:08 * * Sysklogd fails to build Jul 19 11:56:10 * * http://bugs.openembedded.net/show_bug.cgi?id=2691 Jul 19 12:05:41 03Laibsch 07org.oe.dev * rf9f4aeef... 10/ (6 files in 3 dirs): sysklogd: update to version 1.5 and unify. Hopefully closes 2691. Jul 19 12:11:16 03mickeyl 07org.oe.dev * rbb80418f... 10/ (5 files in 2 dirs): iptables 1.3.8 factor out .inc, remove some versions, add recommendation to nat iptable module Jul 19 12:43:54 mickey|zzZZzz: wake up! Jul 19 12:44:11 "mtn add packages/iptables/iptables.inc" Jul 19 12:44:16 ;-) Jul 19 12:44:34 hehe Jul 19 12:45:50 the giant awakes Jul 19 12:47:26 03mickeyl 07org.oe.dev * r0cfa04cd... 10/ (1 packages/iptables/iptables.inc): iptables: add .inc file Jul 19 13:06:08 Well, I just broke ABI in poky, I needed to fix the arm contamination problems. Lets see how it works out... Jul 19 13:30:00 who is ccsmart? Jul 19 13:30:06 !seen ccsmart Jul 19 13:30:09 Laibsch, I don't remember seeing ccsmart. Jul 19 13:45:02 03rpurdie 07org.oe.dev * r8db7f823... 10/ (1 conf/bitbake.conf): bitbake.conf: Set PACKAGES_DYNAMIC so locale packages can be used in runtime dependency variables (from poky) Jul 19 13:45:07 03rpurdie 07org.oe.dev * r6861a45c... 10/ (1 conf/bitbake.conf): bitbake.conf: For EXTRA_OEMAKE so it works with bitbake 1.8 svn correctly Jul 19 13:53:40 03rpurdie 07org.oe.dev * r6dbd04de... 10/ (31 files in 4 dirs): conf/machine: Set BASE_PACKAGE_ARCH which is then used to construct PACKAGE_ARCH instead of setting directly. This allows us to experiment with various changes which wouldn't otherwise be possible (from poky) Jul 19 13:59:07 * * OE Bug 4364 has been RESOLVED (FIXED) by heinold(AT)inf.fu-berlin.de Jul 19 13:59:09 * * alsa-lib doesn't work properly with uClibc Jul 19 13:59:11 * * http://bugs.openembedded.net/show_bug.cgi?id=4364 Jul 19 14:01:56 03woglinde2 07org.oe.dev * rbe3713b5... 10/ (4 files in 3 dirs): Jul 19 14:01:56 alsa-lib: fix avr32 builds and libmath detection Jul 19 14:01:56 * fix bug 4364 provided Geoffrey Wossum Jul 19 14:01:56 * add fix_libmath.patch to correct detection either -lm is needed or not Jul 19 14:01:56 * bump PR Jul 19 14:09:13 03rpurdie 07org.oe.dev * rdbfd8134... 10/ (15 files in 5 dirs): conf/bitbake.conf: Merge in multimachine making it the standard layout as discussed on the mailing list. Anyone wanting the previous layout can INHERIT the new singlemachine class Jul 19 14:14:42 Long live multimachine :) Jul 19 14:25:14 03rpurdie 07org.oe.dev * r5986a81c... 10/ (4 files in 3 dirs): opkg-utils: Add from poky Jul 19 14:25:19 03rpurdie 07org.oe.dev * r6b61e8f4... 10/ (1 conf/distro/include/sane-srcrevs.inc): sane-srcrevs.inc: Add opkg-utils Jul 19 16:36:34 good morning ! Jul 19 16:38:33 hi khem Jul 19 16:38:53 hey RP Jul 19 16:39:06 I was wondering what toolchain combos should I test Jul 19 16:39:18 with new toolchain sequence Jul 19 16:39:36 I have gone as far as gcc 3.4.4 and glibc 2.3.6 Jul 19 16:39:50 khem: Did you test meta-toolchain? Jul 19 16:39:52 Should I do for older than that too ? Jul 19 16:39:57 RP: no not yet Jul 19 16:40:05 khem: I'd actually be interested in your thoughts on the sdk business Jul 19 16:40:22 So far I have been testing various combinations Jul 19 16:40:34 khem: You could ask mwester about what nslu2 are using Jul 19 16:40:42 I think they're our oldest glibc users Jul 19 16:41:06 RP: good idea Jul 19 16:41:12 * RP is trying to write a definitive guide to OE policies :/ Jul 19 16:41:47 I had to do surgeries for glibc 2.3.6 and gcc 3.4.4 it seems no one really used them Jul 19 16:42:51 khem: We did use 3.4.4 and glibc 2.3.5+cvs in poky Jul 19 16:43:09 I'm trying to drop them though Jul 19 16:43:15 well, glibc 2.3.5 anyway Jul 19 16:43:16 RP: btw. I have done more on the patch so far I will send an update later today Jul 19 16:43:20 gcc 3.4.4 can be useful Jul 19 16:43:31 yeah I got gcc 3.4.4 working Jul 19 16:43:43 I dropped gcc 3.3.x from poky Jul 19 16:44:06 with one small patch that was needed to make libgcc stop pully unwind stuff when building bootstrap compiler patch Jul 19 16:44:15 RP: cool Jul 19 16:45:31 RP: I can work on getting all combos working but it will be good to depricate some of them if they are unused Jul 19 16:46:51 RP I will also try the combo you used on poky Jul 19 16:47:30 glibc 2.3.5+cvs + gcc 3.4.4 what binutils and kernel-header versions did u use Jul 19 17:11:11 khem: binutils 2.15.94.0.1 and 2.6.11.1 headers Jul 19 17:11:27 cvs 20050627 for glibc Jul 19 17:12:09 hrw|gone, mickey|away, XorA|gone, Crofton, Crofton|work: Please read the email I've just sent. I want to get this sorted once and for all... Jul 19 17:22:11 RP: thanks for pushing this along Jul 19 17:22:42 anyone else having trouble with bitbake parsing the current OE metadata: http://pastebin.ca/1076787 Jul 19 17:22:46 I'm using latest from git Jul 19 17:23:54 cbrake: We will get there ;-) Jul 19 17:24:06 cbrake: Looks like a broke regexp somewhere Jul 19 17:24:38 * cbrake removes his new mono recipes ... Jul 19 17:27:11 We should fix bitbake so it shows the broken regexp Jul 19 17:37:28 RP: patch to print failing regex: http://bugs.openembedded.net/show_bug.cgi?id=4444 Jul 19 17:38:33 RP: libsigc++-1.2-locale-* is the failing regex -- wonder if that is something to do with the PACKAGES_DYNAMIC = "${PN}-locale-*" change? Jul 19 17:39:03 * cbrake heads off to a meeting -- hopefully recruit more OE users :-) Jul 19 17:44:06 cbrake_away: That would be my patch :/ Jul 19 17:45:06 cbrake_away: ah, is this the last bitbake release? Jul 19 17:46:32 hello, i just updated my copy of the org.openembedded.dev repository(and checked if there wasn't any update of bitbake but there wasn't) and now bitbake always crashes ...it's something about regular expressions...:http://pastebin.com/m20e4b721 Jul 19 17:48:10 hi Gnutoo Jul 19 17:48:19 Gnutoo: Its a known issue :/ Jul 19 17:48:31 RP, ok thanks so i'll wait...hi Jul 19 17:48:53 I will push a fix Jul 19 17:49:05 ok Jul 19 17:49:07 thanks a lot Jul 19 17:51:26 03rpurdie 07org.oe.dev * r32531173... 10/ (1 conf/bitbake.conf): bitbake.conf: Disable PACKAGES_DYANMIC until we upgrade bitbake Jul 19 17:51:33 there you go Jul 19 17:53:53 thanks Jul 19 17:58:28 03rpurdie * r1076 10/ (4 files in 4 dirs): providers.py: When a regexp fails to compile for PACKAGES_DYNAMIC, print a more useful error (#4444) Jul 19 18:39:15 RP: I am looking at slugos.inc Jul 19 18:39:26 and it uses PREFERRED_VERSION_binutils ?= "2.18" Jul 19 18:39:27 PREFERRED_VERSION_binutils-cross ?= "2.18" Jul 19 18:39:27 PREFERRED_VERSION_binutils-cross-sdk ?= "2.18" Jul 19 18:39:27 PREFERRED_VERSION_gcc ?= "4.1.2" Jul 19 18:39:27 PREFERRED_VERSION_gcc-cross ?= "4.1.2" Jul 19 18:39:29 PREFERRED_VERSION_gcc-cross-sdk ?= "4.1.2" Jul 19 18:39:31 PREFERRED_VERSION_gcc-cross-initial ?= "4.1.2" Jul 19 18:39:34 PREFERRED_VERSION_gcc-cross-intermediate ?= "4.1.2" Jul 19 18:39:36 PREFERRED_VERSION_glibc ?= "2.5" Jul 19 18:39:39 PREFERRED_VERSION_glibc-initial ?= "2.5" Jul 19 18:40:19 Unslung is much older. Jul 19 18:41:08 Also, I think we should check with rwhitby to see if he would like to be able to continue to build the current-1 release of SlugOS from OE. Jul 19 18:43:14 mwester: thank you Jul 19 18:44:41 Thanks for trying to clean some of this old stuff up. :) Just FYI, the plan was to freeze Unslung by taking a snapshot of OE and putting it in its own repo somewhere -- I had planned to do that when we migrated to git, but that keeps getting pushed out. Jul 19 18:45:10 If we can get Unslung out of the normal OE repo, that's one less 2.4 kernel and some very old bb files that we can clean up. Jul 19 19:00:36 mwester: cool sounds a nice plan Jul 19 19:01:59 RP: rev 1076 broke bitbake for me Jul 19 19:05:01 This patch here http://pastebin.com/m52f3b874 fixes it Jul 19 19:07:14 hi khem Jul 19 19:07:28 khem is your patch ready? Jul 19 19:08:43 woglinde: yes but I could only test backwards untils glibc 2.3.6 and gcc 3.4.4 and bintutils 2.16.1 Jul 19 19:09:08 I think thats enough Jul 19 19:09:21 put it in as fast as possible Jul 19 19:11:02 woglinde: I can barge it in if all think its ok Jul 19 19:11:52 woglinde: I have also fixed some toolchain combinations for arm and x86 during this testing process. Jul 19 19:12:04 woglinde: I think if Jul 19 19:12:13 I commit I will get wider testing :) Jul 19 19:12:22 forced on developers Jul 19 19:12:36 if ro says its okay Jul 19 19:12:38 than do Jul 19 19:12:40 ups rp Jul 19 19:12:42 *g* Jul 19 19:44:09 woglinde: RP will be testing it on poky Jul 19 19:44:15 so I will wait Jul 19 19:44:53 :) Jul 19 21:12:15 03rpurdie * r1077 10/ (2 files in 2 dirs): providers.py: Fix indentation errors Jul 19 21:13:26 khem: Your patch wasn't 100% right either but it should be fixed now :) Jul 19 21:13:39 hurray Jul 19 21:13:45 khem: Thanks for spotting :) Jul 19 21:19:46 RP: as long as it is fixed I am happy :) Jul 19 21:20:37 RP: I will send out my patch for toolchain build sequence change. I have done few other tweaks which I found while building various combinations. Jul 19 21:21:53 khem: Sounds good. I will try and run some tests on it as soon as I sort out the gremlins my recent poky changes have caused ;-) Jul 19 21:22:34 RP: heh ok Jul 19 21:22:44 RP: I am testing your combo right now Jul 19 21:22:46 on x86 Jul 19 21:22:55 he zecke Jul 19 21:23:25 gcc-3.4.5 glibc 2.3.5-cvs binutils-2.15.94.0.1 and kernel-headers 2.6.11.1 Jul 19 21:23:34 gcc 3.4.4 Jul 19 21:24:08 woglinde: hi and bye Jul 19 21:24:14 bye? Jul 19 21:25:25 woglinde: going to bed... it is bright already Jul 19 21:25:39 oh okay Jul 19 21:25:46 I thought you were awake again Jul 19 21:25:48 *g* Jul 19 21:26:46 9 failures from the poky autobuilder in less than that number of hours :( Jul 19 21:27:10 RP: gruaimín ( irish ) Jul 19 21:30:36 RP: I am seeing another error hopefully its bitbake Jul 19 21:30:38 /home/kraj/work/oe/build/glibc/tmp-qemux86/work/i586-angstrom-linux/glibc-2.3.5+cvs20050627-r18/temp/run.staging_helper.22652: line 753: syntax error near unexpected token `" | grep -v "' Jul 19 21:31:00 khem: What do you see in that script at that line/ Jul 19 21:31:02 ? Jul 19 21:33:24 RP: a moment pleas Jul 19 21:33:58 echo "$pkg OPP $depends;" | grep -v "(" | grep -v ")" | grep -v Depends | sed -e 's:,::g' -e 's:-:_:g' -e 's:\.:_:g' -e 's:+::g' |sed 's:OPP:->:g' >> ${TESTLAB_DIR}/depends.dot Jul 19 21:34:47 RP: | grep -v looks like it duplicated Jul 19 21:34:54 holy shit Jul 19 21:36:45 ouch Jul 19 21:36:57 I don't think I did anything to do that :/ Jul 19 21:38:48 let me revert to 1075 Jul 19 21:49:44 RP: same error with 1075 Jul 19 21:50:03 khem: and 1074? Jul 19 21:50:13 RP: let me try that Jul 19 21:50:17 hm but the grep seems false Jul 19 21:50:19 khem: you're suspecting bitbake? Jul 19 21:50:28 RP: no Jul 19 21:50:44 RP: its happening when building glibc-2.3.5+cvs20050627 Jul 19 21:50:55 and I think thats where problem is Jul 19 21:51:06 but bitbake changes happened in the middle Jul 19 21:51:09 I can't even see where this TESTLAB_DIR stuff makes it into the staging_helper function :/ Jul 19 21:51:18 Is that some local change? Jul 19 21:51:25 no Jul 19 21:51:34 its angstrom 2008 Jul 19 21:51:37 I am using Jul 19 21:52:26 Hmm, it looks like it comes from testlab.bbclass Jul 19 21:52:38 I have no idea about that stuff, never seen it before Jul 19 21:54:41 yeah I am seeing the same bbclass file Jul 19 21:55:42 and it only happens for above recipe Jul 19 21:57:51 khem: Sounds like you've hit a really odd interaction of some sort Jul 19 22:03:14 RP: yeah seems like that Jul 19 22:03:30 2.3.6 works ok I am retesting just to maintain my sanity Jul 19 22:03:59 is anyone using the old glibc? Jul 19 22:04:38 2.3.5+cvs is the old favourite from OZ days (and the one poky uses for OABI) Jul 19 22:04:49 woglinde: RP said poky is using Jul 19 22:04:54 rp hm uh okay Jul 19 22:04:58 For OABI only Jul 19 22:05:12 but its a good test of the toolchain Jul 19 22:05:16 to many problems with oabi and newr glibc? Jul 19 22:05:32 woglinde: No, just that if it ain't broke, don't fix it ;-) Jul 19 22:05:55 oabi is not maintained that well upstream glibc now a days Jul 19 22:12:53 morning Jul 19 22:12:58 hi ant Jul 19 22:13:23 woglinde: Du bist unermuedlich! Jul 19 22:13:32 noe Jul 19 22:13:42 I am watching tv Jul 19 22:13:46 since serval hours Jul 19 22:14:02 but you are right I fixed one bug earlier Jul 19 22:14:06 vor avr32 folks Jul 19 22:15:11 Which language is "noe"? Jul 19 22:15:16 german Jul 19 22:15:18 woglinde: could you give a deeper look at the libtool2 / autoconf issue? Jul 19 22:15:36 woglinde: Hmm, I never came accross that one :) Jul 19 22:15:39 ant apt? Jul 19 22:15:43 rp nö Jul 19 22:15:50 short for nein Jul 19 22:15:53 which is no Jul 19 22:16:08 I would have understood nein :) Jul 19 22:16:11 ant apr I meant Jul 19 22:16:26 no, speech-dispatcher Jul 19 22:16:38 and my "fix" Jul 19 22:16:39 ah right Jul 19 22:16:46 hm oh your fix Jul 19 22:16:54 is it in bug tracker? Jul 19 22:17:12 yo Jul 19 22:17:30 ah right Jul 19 22:17:38 okay now I fixed alsa-lib Jul 19 22:17:41 I could try Jul 19 22:17:54 but have to fix alsa-oss with no libio.h Jul 19 22:18:04 and remember pls 4427 has similar symptoms Jul 19 22:19:31 okay thanks Jul 19 22:20:01 as 4391 (error: m4_require: circular dependency of AC_LANG_COMPILER(C++)) Jul 19 22:20:22 'night all Jul 19 22:20:28 nite rp Jul 19 22:20:39 RP: buonanotte Jul 19 22:22:03 woglinde: 4427 is configure.ac, 4391 configure.in Jul 19 22:22:14 its only names Jul 19 22:22:18 .ac is newer Jul 19 22:24:51 hms Jul 19 22:25:02 seems libio.h is not available with uclibc Jul 19 22:27:55 woglinde: for the patch I got this bits: Jul 19 22:27:58 http://bugs.gentoo.org/show_bug.cgi?id=229901 Jul 19 22:27:59 okay I found a fix for that Jul 19 22:28:10 https://bugs.gentoo.org/show_bug.cgi?id=220543 Jul 19 22:30:29 'If they use C++, AC_PROG_CXX should have been there anyway, and not implicitly Jul 19 22:30:29 called by old libtool.' Jul 19 22:31:21 ? Jul 19 22:31:38 with the old libtool it was ok... Jul 19 22:31:47 ?? Jul 19 22:38:42 hms Jul 19 22:38:52 pulseaduio I have to fix too Jul 19 23:25:13 mwester, khem: We only need to be able to build current SlugOS from OE - for N-1 binary release we branch it off to SVN and maintain it there. Jul 19 23:26:28 mwester, khem, RP: as mwester said, we are planning to move Unslung out of OE altogether, since it's so old and will never be updated toolchain-wise or kernel-wise. Jul 19 23:28:31 rwhitby: thx Jul 19 23:29:13 rwhitby: I have aleady tested the combo gcc-4.1.2+glibc-2.5+binutils-2.18 that is currently in OE and it builds fine Jul 19 23:30:22 rwhitby: so current slugOS will build with the new tc sequence Jul 19 23:30:43 khem: excellent, thx. Jul 19 23:31:04 bitbake -c build -b glibc_2.3.5+cvs20050627.bb Jul 19 23:31:04 ERROR: Unable to match glibc_2.3.5+cvs20050627.bb (0 matches found): Jul 19 23:31:41 recipe it there Jul 19 23:32:17 s/it/is/ Jul 19 23:34:06 anything that has name_+.bb is not parsable by bitbake Jul 19 23:35:29 bitbake -c build -b ../org.openembedded.dev/packages/glibc/glibc_2.3.5+cvs20050627.bb Jul 19 23:35:34 works Jul 19 23:52:57 nite **** ENDING LOGGING AT Sun Jul 20 02:59:56 2008