**** BEGIN LOGGING AT Wed Aug 23 02:59:57 2006 Aug 23 05:12:45 Is #openzaurus usually very quiet or have I just caught them at a bad time? (meaning no disrespect, I just want to know when is a good time to ask support in there) Aug 23 06:53:29 lo Aug 23 07:42:13 morning Aug 23 07:48:00 morning Aug 23 07:48:11 hey Aug 23 07:50:38 kergoth: switched to europian day? Aug 23 07:50:45 hehe, pretty close to it Aug 23 07:54:36 good morning all Aug 23 08:11:25 morning all Aug 23 08:11:30 hey RP Aug 23 08:11:40 hey RP & XorA Aug 23 08:12:48 RP: can you look at http://lkml.org/lkml/2005/9/9/158 and tell do we have it or not already? Aug 23 08:14:32 hrw|work: Its not in our patch series but its from a long time ago - was it applied to mainline? Aug 23 08:17:05 good monring Aug 23 08:17:10 hi florian_kc Aug 23 08:20:42 RP: will look Aug 23 08:35:13 hi zecke Aug 23 08:35:34 ho Aug 23 08:38:43 hi zecke Aug 23 08:40:05 hey zecke & hrw|work Aug 23 08:59:06 hi, folks. Aug 23 08:59:14 hey leoncamel Aug 23 08:59:44 which monotone server should I use now ? it seems that monotone.vanille.de is down.. :( Aug 23 08:59:51 hey, koen Aug 23 09:00:04 leoncamel: check GettingStarted, that's uptodate Aug 23 09:00:06 koen, how are you doing these days ? Aug 23 09:00:17 pretty good Aug 23 09:00:35 apart from all the exams coming this way Aug 23 09:00:42 how are you? Aug 23 09:00:44 koen, hmm, the question is where is the GettingStarted. ? ;) Aug 23 09:01:05 http://www.openembedded.org/wiki/GettingStarted :) Aug 23 09:01:18 koen, OH. which university ? Aug 23 09:02:11 leoncamel: www.utwente.nl/en Aug 23 09:02:34 OH. koen.. great. Aug 23 09:03:29 koen, OK. you folks switched to SVN now .. :) Aug 23 09:04:03 what? Aug 23 09:04:41 koen, I mean , I should checkout by Subversion intead of monotone ? Aug 23 09:05:15 ah Aug 23 09:05:40 leoncamel: that's read only Aug 23 09:07:45 kergoth, :). OK. It's my mistake. I misunderstanding the "bitbake subversion" :). sorry for that . :) Aug 23 09:07:54 hi, florian_kc .. :) Aug 23 09:08:00 bitbake != oe :) Aug 23 09:08:19 :-) Aug 23 09:08:44 hey kergoth Aug 23 09:09:02 hey Aug 23 09:09:31 RP: same result wth pre12 on corgi Aug 23 09:09:35 kergoth, :). yes. but I really *hate* monotone. I *love* git and svn .. :) Aug 23 09:09:49 XorA: ok, Liam broke it ;-) Aug 23 09:10:04 RP: just double checking, then ll email Aug 23 09:11:19 kergoth: does your jhbuild2oe script work well? Aug 23 09:12:58 hey, koen . I can't connect to monotone.openembedded.org Aug 23 09:13:32 koen@bitbake:~/OE/monotone/org.openembedded.dev/packages/driftnet$ mtn sync Aug 23 09:13:32 mtn: connecting to monotone.openembedded.org Aug 23 09:13:33 mtn: finding items to synchronize: Aug 23 09:13:33 mtn: certificates | keys | revisions Aug 23 09:13:33 mtn: 27943 | 28 | 9224 Aug 23 09:13:34 mtn: bytes in | bytes out | revs in | revs out Aug 23 09:13:35 mtn: 1.3 k | 1.7 k | 0/0 | 0/0 Aug 23 09:13:36 koen: needs some work.. currently it emits rudamentary packages and supports supplying templates for each that it can drop data into, ala autoconf type subst's Aug 23 09:13:37 mtn: successful exchange with monotone.openembedded.org Aug 23 09:13:39 that's what I get Aug 23 09:13:58 kergoth: ok, I'm reading up on bb collections and willgive it a spin Aug 23 09:13:59 koen: it needs some sort of templating that can encompass more than one package, though, some sort of rudamentary classification system Aug 23 09:14:13 * kergoth was hoping to make the next xorg package upgrade less of a pain in the ass Aug 23 09:15:09 kergoth: there's a script in the bugtracker to autogen xorg stuff Aug 23 09:15:37 is there? didnt notice Aug 23 09:15:42 getting the .bbs into OE isn't the problem Aug 23 09:15:49 having enough resources to test it is Aug 23 09:16:01 aye Aug 23 09:16:57 we need to get rid of the nonworking directory Aug 23 09:16:57 thanks to XorA and mallum we fixed some stupid bugs last month Aug 23 09:17:18 * XorA doesnt remeber doing anything Aug 23 09:17:19 koen: poky is using xorg now Aug 23 09:17:26 RP: cool Aug 23 09:17:27 i'm just happy to finally have a capable build machine so i can actually test oe stuff again Aug 23 09:17:39 RP: 7.1? Aug 23 09:17:40 think i may dive into the toolchain bits, they seem to be something of a mess at the moment Aug 23 09:17:43 heh Aug 23 09:18:05 kergoth: please fix our glibc2.4/gcc4.1 eabi toolchain bits Aug 23 09:18:29 kergoth: add --with-cpu to gcc Aug 23 09:18:33 koen: yes Aug 23 09:18:43 kergoth: fix enabling and disabling of languages, fix the fortran mess Aug 23 09:18:49 --with-cpu? Aug 23 09:18:55 only RP knows fortran anyway... Aug 23 09:18:59 hehe Aug 23 09:19:16 * RP hates fortran more than you can imagine... Aug 23 09:19:19 kergoth: ob gcc's./configure add a default CPU to optimize for Aug 23 09:19:23 * kergoth is glad he never had to learn that stuff.. the benefit of never having gotten an actual formal CS education Aug 23 09:19:32 Its also the only langague I've been trained in :) Aug 23 09:19:52 kergoth: we don't do fortran in CS education over here Aug 23 09:19:57 is fortran roughly what matlab uses? Aug 23 09:20:47 kergoth: I'm a physicist by training ;-) Aug 23 09:21:22 XorA: I don't think so Aug 23 09:21:34 then I have managed to avboid it :-) Aug 23 09:22:32 XorA: be grateful ;-) Aug 23 09:22:53 bitbake's generating an error and telling me to look at the config.log file -- does anyone know the default store location for this file? i just ls'd around and can't seem to find it. Aug 23 09:23:03 I rember at GEC the scary haired system modellers used it, but they were safely on the other side of the partition wall :-) Aug 23 09:23:40 bigd0g: I'm sure bitbake is not 'generating' the error Aug 23 09:23:55 bigd0g: it is a configure script and the configure script asks you to look into config.log Aug 23 09:24:17 i apologize for my semantical error Aug 23 09:24:26 anyway, do you know where that file is generally stored? Aug 23 09:24:41 bigd0g: it is within the '${S}' dir which is normally tmp/work/PackageName-PackageVersion-PackageRelease/* Aug 23 09:24:43 bigd0g: in the source/build tree of the package. Aug 23 09:24:57 bigd0g: about which package do we talk? Aug 23 09:25:40 well, i'm trying to build 'linux-omap1' ... it fails when trying to build glibc-2.3.5 Aug 23 09:25:52 and then points me to config.log Aug 23 09:26:53 I wonder what could break in glibc-2.3.5 Aug 23 09:27:51 hey mikearthur Aug 23 09:27:52 i don't know. would you like to see the failed notes? Aug 23 09:28:25 XorA: yo Aug 23 09:28:51 ~pastebin Aug 23 09:28:55 i heard pastebin is a place to paste your stuff without flooding the channel - try http://pastebin.com/ (BROKEN AND SUCKING NUTS), or http://pastebin.ca, or http://channels.debian.net/paste, or for #oe use http://oe.pastebin.com, or http://bzflag.pastebin.ca/, or http://paste.lisp.org/ for the lisp/scheme nuts, or http://paste-it.net Aug 23 09:28:59 bigd0g: out of interest yes Aug 23 09:29:12 zecke: prefer PM or in chan? Aug 23 09:29:22 bigd0g: I want the the log.do_configure* and the config.log to be pasted into two different logs Aug 23 09:29:26 bigd0g: we prefer pastebin Aug 23 09:30:00 sorry for the following n00b question: what's pastebin ? Aug 23 09:30:10 ~pastebin Aug 23 09:30:12 pastebin is probably a place to paste your stuff without flooding the channel - try http://pastebin.com/ (BROKEN AND SUCKING NUTS), or http://pastebin.ca, or http://channels.debian.net/paste, or for #oe use http://oe.pastebin.com, or http://bzflag.pastebin.ca/, or http://paste.lisp.org/ for the lisp/scheme nuts, or http://paste-it.net Aug 23 09:30:28 oh. ok. i've used similar sites before. no worries. Aug 23 09:30:38 ~snack Aug 23 09:30:40 * ibot snack? yes yeaaas yeaaaaaaaaaaas pleaaase :P Aug 23 09:30:46 ~snack ibot Aug 23 09:30:48 * ibot throws ibot a doggy treat Aug 23 09:33:47 kergoth: it would be nice if jhbuild2oe would append _svn, _git and _cvs to the recipe name :) Aug 23 09:36:52 zecke: pastebin for log.do_configure @ http://pastebin.com/773859 Aug 23 09:37:08 trying to find the config.log still ... one moment Aug 23 09:42:37 hey, folks. I can't connect to monotone.openembedded.org. hmm, is the monotone serve shutdown ? Aug 23 09:43:16 leoncamel: Which version of monotone are you using? Aug 23 09:43:40 0.25.. Aug 23 09:44:17 RP, OK. I got it. I should upgrade the monotone first. Aug 23 09:45:02 leoncamel: i think 0.28 is the latest Aug 23 09:45:09 0.29 Aug 23 09:45:17 see http://venge.net/monotone Aug 23 09:45:31 hmm, whichi one do you preferred ? Aug 23 09:45:33 oh. then i guess i'm a point release behind, too. and i just updated recently. Aug 23 09:45:44 leoncamel: latest Aug 23 09:46:35 hrw|work, ok. thanks. Aug 23 09:54:00 zecke: i'm not seeing config.log in my sources dir or the packages dir Aug 23 09:54:06 am i looking in the wrong place(s) ? Aug 23 09:55:44 bigd0g: tmp/work/glibc*/ Aug 23 09:55:50 bigd0g: no, but glibc uses BUILD DIR != Source DIR Aug 23 09:57:58 many thanks, hrw|work Aug 23 10:04:09 zecke: http://pastebin.com/773877 for config.log - http://pastebin.com/773859 for log.do_configure Aug 23 10:04:55 any help/direction is appreciated Aug 23 10:08:35 line 393 Aug 23 10:08:41 conftest.c:1: error: bad value (arm926ejs) for -mtune= switch Aug 23 10:09:07 arm926ejs has different name for gcc3 and gcc4 iirc Aug 23 10:09:24 yes Aug 23 10:09:27 Read the tune-*.inc file... Aug 23 10:09:33 ~lart gcc 'developers' Aug 23 10:09:34 bigd0g: check conf/machine/include/tune-arm926ejs.conf Aug 23 10:09:37 * ibot wallops gcc 'developers' with a main rotation server that needs rehubbing. It won't take long Aug 23 10:09:43 #if gcc breaks change arm926ejs to arm926ej-s Aug 23 10:09:44 TARGET_CC_ARCH = "-march=armv5te -mtune=arm926ejs" Aug 23 10:10:51 ok, checking that file and make suggested change(s). Aug 23 10:13:25 seems simple enough. trying build again. Aug 23 10:13:29 * bigd0g crosses fingers Aug 23 10:18:07 RP: spectrum_cs fix roxx Aug 23 10:18:30 /data/build/koen/OE/build/tmp/angstrom/staging/arm-angstrom-linux-gnueabi/lib/libopieui2.so: undefined reference to `QRect::intersects(QRect const&) const' Aug 23 10:18:38 so no opie for angstrom Aug 23 10:19:23 koen: haeh? how did you do that? Aug 23 10:19:34 koen: I got opie built Aug 23 10:19:44 Angstrom should contain the award winning Qtopia4x Aug 23 10:19:48 zecke: 'bitbake opie-taskbar' Aug 23 10:20:03 koen: and some kind of CFLAGS? Aug 23 10:20:04 koen: -rw-r--r-- 1 hrw pavilon 27951104 Aug 17 23:36 Angstrom-opie-image-test-20060817-tosa.rootfs.jffs2 Aug 23 10:20:22 zecke: the visibility stuff you pasted yesterday Aug 23 10:20:30 anyway Trolltech said they will GPL QtopiaPhone in September Aug 23 10:20:41 koen: only inlines hidden? Aug 23 10:21:28 glibc is trying to compile now. thanks, zeke et al. Aug 23 10:21:58 * zecke crosses the fingers Aug 23 10:22:10 zecke: http://rafb.net/paste/results/GUnMl094.html Aug 23 10:22:55 koen: a) your apple term sucks b) mine sucks too c) that was a link line Aug 23 10:23:02 koen: visibility happens on compiling Aug 23 10:24:28 FULL_OPTIMIZATION = "-fexpensive-optimizations -fomit-frame-pointer -frename-registers -Os -fvisibility-inlines-hidden" Aug 23 10:24:28 BUILD_OPTIMIZATION = "-Os -fvisibility-inlines-hidden" Aug 23 10:24:45 hmm Aug 23 10:29:12 NOTE: package glibc-2.3.5+cvs20050627-r12: task do_compile: failed Aug 23 10:29:15 hrm... Aug 23 10:29:15 ah Aug 23 10:29:25 monotone is workign again on osx Aug 23 10:33:12 03koen 07org.oe.documentation * r2056cc91... 10/ (1 usermanual/usermanual.xml): usermanual.xml: paste collections example into the appropriate section Aug 23 10:33:16 03koen 07org.oe.packaged-staging * rf1b6b0ac... 10/ (1 docs/packaged-staging.xml): docs/packaged-staging.xml: be more verbose and talk about chroot a bit Aug 23 10:33:20 03koen 07org.oe.packaged-staging * r350fd527... 10/ (1 docs/packaged-staging.xml): docs/packaged-staging.xml: more bits Aug 23 10:34:12 amazng, realtime checkin emails :-=) Aug 23 10:36:13 did pastebin just die? Aug 23 10:36:24 ...that, or i must've found a character limit Aug 23 10:37:00 cc1: warning: command line option "-fvisibility-inlines-hidden" is valid for C++/ObjC++ but not for C Aug 23 10:37:02 heh Aug 23 10:43:36 koen: CXXFLAGS += instead :-) Aug 23 10:44:36 why there is noone who use wlan-ng when I hack it.. Aug 23 10:46:54 zecke et al: can you take another look at the new config.log file? Aug 23 10:46:55 http://www.rafb.net/paste/results/MHkOL024.html Aug 23 10:46:57 hmmm Aug 23 10:47:23 I can fix mesa for x86->arm, but I don't know how to fix it properly Aug 23 10:47:36 NOTE: package wlan-ng-modules-0.2.4+svn20060823-r0: task do_build: completed Aug 23 10:47:52 zecke: i can't paste the new log.do_compile* because it's too large for any pastebin i can find Aug 23 10:48:08 bigd0g: post end part then Aug 23 10:48:18 kergoth: http://bugs.openembedded.org/attachment.cgi?id=837 and http://bugs.openembedded.org/show_bug.cgi?id=870 Aug 23 10:48:25 hrw|work: ok Aug 23 10:50:40 zecke et al: new log.do_compile* end @ http://www.rafb.net/paste/results/vH6gNa79.html Aug 23 10:55:28 hrm. it looks like glib compile correctly this time Aug 23 10:55:45 but it doesn't give much other direction as to what went wrong with the rest of the build Aug 23 10:57:33 n/m. glibc did fail again. >> NOTE: package glibc-2.3.5+cvs20050627: failed Aug 23 11:01:17 do13: ping Aug 23 11:02:33 mtn: 2 heads on branch 'org.openembedded.dev' Aug 23 11:02:36 heh. Aug 23 11:02:44 pushing now Aug 23 11:03:30 pushed Aug 23 11:04:47 03hrw 07org.oe.dev * r99d4c960... 10/ (11 files in 4 dirs): Aug 23 11:04:47 wlan-ng: fixing to get it build and updated - close #1260 Aug 23 11:04:47 - added 0.2.4+svn20060823 Aug 23 11:04:47 - bumped SVN to 0.2.4 Aug 23 11:04:47 - SVN are DEFAULT_PREFERENCE = -1 now Aug 23 11:04:47 - dropped pcmcia-driver.patch from wlan-ng-modules.inc Aug 23 11:04:49 - for PCMCIA cards we use HostAP not wlan-ng Aug 23 11:06:32 03koen 07org.oe.dev * r2098d6e0... 10/ (1 packages/mesa/mesa_6.4.1+cvs20060101.bb): mesa 6.4.1: add a hack to use 'gcc' for building gen_matypes on x86 Aug 23 11:07:19 ~seen treke Aug 23 11:07:26 treke was last seen on IRC in channel #oe, 131d 10h 33m 47s ago, saying: 'at least as of a year and a half ago :)'. Aug 23 11:07:41 someone (other then treke) has control over OE bugzilla theme? Aug 23 11:07:54 not afaik Aug 23 11:08:11 ~seen ggilbert Aug 23 11:08:15 ggilbert is currently on #oe (1d 18h 10m 56s), last said: 'must resist urge to build an image for my treo'. Aug 23 11:09:08 it would be nice to have one *working* css for oe.org, viewmtn, tinderbox and bugs Aug 23 11:11:16 can i bother anyone to take a look at my log files and give me some direction as to what i should be doing next? Aug 23 11:11:31 i think it's failing because 'arm-linux-g++' isn't being found Aug 23 11:12:30 as per line 808 on http://www.rafb.net/paste/results/MHkOL024.html Aug 23 11:13:06 koen: ha! Aug 23 11:13:53 hrw|work: do you have a mail in your inbox about a test? Aug 23 11:14:31 checking... Aug 23 11:14:38 koen: when sent? Aug 23 11:14:46 just now Aug 23 11:15:15 'no new mail' Aug 23 11:15:44 openembedded@hrw.one.pl is the right address? Aug 23 11:16:39 yes Aug 23 11:17:01 will check spambox Aug 23 11:17:26 koen: X-Spam-Status: Yes, score=5.1 required=5.0 Aug 23 11:17:37 X-Spam-Status: Yes, score=5.1 required=5.0 tests=ADDRESS_IN_SUBJECT,BAYES_50, DNS_FROM_RFC_ABUSE,FORGED_RCVD_HELO,SUBJ_ILLEGAL_CHARS Aug 23 11:18:11 in other words - I got it Aug 23 11:18:19 cool Aug 23 11:18:33 so our key ID work as email adresses now Aug 23 11:18:42 (for people who registered) Aug 23 11:18:46 so if 'confdefs.h' is the program that is continually failing ... what can i do? Aug 23 11:19:00 ah we have an MX entry now? Aug 23 11:19:10 zecke: I think so Aug 23 11:20:29 bigd0g: uh, that's not a program Aug 23 11:21:14 i try to build oz with dbus 0.62 Aug 23 11:21:14 florian_kc: i know. it's a library. but the log says "program failed: confdefs.h" Aug 23 11:21:43 koen: good to know about new alias Aug 23 11:21:55 bigd0g: its a header file, but iirc that some generated file by autotools Aug 23 11:22:02 hrw|work: I'll leave it to mailguru mickeyl to announce it Aug 23 11:22:37 florian_kc: wouldn't it be wiser to try dbus 0.92? Aug 23 11:22:46 koen: you are going alone to oedem or with gf? Aug 23 11:22:46 i think it's being generated by the lack of arm-linux-g++ Aug 23 11:22:57 but bitbake built arm-linux-gcc ... why wouldn't it build arm-linux-g++ ? Aug 23 11:25:13 morning all Aug 23 11:25:18 hey liam Aug 23 11:25:27 koen: that was the latest complete set available... updating dbus is the next step then. first i need to find out if everything works with the new api Aug 23 11:25:31 hi lrg Aug 23 11:25:37 XorA: fixed the bug - my stupid Aug 23 11:25:58 hi liam Aug 23 11:26:12 hi Liam Aug 23 11:26:37 lrg: applying now Aug 23 11:35:48 Ah, mailing list moved. Will gmane usenet bridge be updated too? Aug 23 11:36:40 wrobbie: dunno, that depends on gmane Aug 23 11:36:49 so, what do we recommend for me? am i in the wrong channel requesting this type of help? Aug 23 11:37:15 bigd0g: sorry, missed your troubles and I will get some food now Aug 23 11:37:49 bigd0g: but generally you are right, now that you have changed the CFLAGS, this is one point where I would start over from scratch (rebuild everything) Aug 23 11:38:21 is there an easy way to do that? Aug 23 11:38:41 i've sort of been thrown into this bitbake project, so i'm just learning it as i go along Aug 23 11:40:23 florian_kc: in .dev it does Aug 23 11:40:33 florian_kc: it compiles :) Aug 23 11:41:13 koen: iirc the matching native or dbus-glib package was missing Aug 23 11:42:42 florian_kc: dbus-glib is present Aug 23 11:43:24 florian_kc: I'm trying to save you from doing two upgrades :) Aug 23 11:45:01 koen: ah okay yes, thanks - so far that's just a test Aug 23 11:46:55 03lenehan 07org.oe.documentation * rc23c92a4... 10/ (9 files in 2 dirs): usermanual: Grammer and consistency edits. Aug 23 11:47:14 hmm Aug 23 11:47:15 ERROR: Task 133 has circular dependency on 27 Aug 23 11:47:41 koen: clear message isn't it? Aug 23 11:49:20 hmm. I really need to improve that one :) Aug 23 11:49:52 someone have a link or something to somewhere that can help clean he necessary files to start a bitbake build from scratch? Aug 23 11:50:07 bigd0g: rm -rf tmp;bitbake glibc Aug 23 11:50:40 RP: can bitbake show from time to time info about tasks progress? I have it stuck on "NOTE: package ncurses-5.4-r8: task do_package: started" for some time... Aug 23 11:51:31 hrw|work: Is should dump the feedback from the task so its no better or worse than it used to be in that regard Aug 23 11:51:41 RP: kind of "NOTE: Running tasks: ncurses/do_package zaurus-updater/do_deploy" kind Aug 23 11:52:25 hrw|work: We need something like that but I think we're probably in need of a UI for it... Aug 23 11:52:31 ko Aug 23 11:53:22 koen: bump angstrom from wpa-supplicant-nossl to wpa-supplicant-0.5.x Aug 23 11:53:29 hrw|work: oh, just delete and re-run. got it. thanks. Aug 23 11:54:01 task-angstrom.bb: wpa-supplicant \ Aug 23 11:55:14 ERROR: Task 133 (do_package) required by /home/koen/OE/monotone/org.openembedded.dev/packages/flex/flex-native_2.5.31.bb has circular dependency on 27 Aug 23 11:55:50 morning Aug 23 11:56:38 koen: I'll add something to make the error message readable Aug 23 11:57:40 RP: http://rafb.net/paste/results/W9wmuS60.html Aug 23 11:59:33 koen: NOTE: package wpa-supplicant-nossl-0.2.6: started Aug 23 12:00:05 koen: angstrom-2006.9/tosa + bootstrap-image Aug 23 12:00:20 hrw|work: angstrom doesn't use bootstrap-image Aug 23 12:00:30 'bitbake angstrom-bootstrap-image' Aug 23 12:00:44 ok Aug 23 12:00:50 did not know ;D Aug 23 12:00:55 that doesn't use the junk in BOOTSTRAP_FOO Aug 23 12:01:05 and it includes gdb + strace Aug 23 12:01:18 I must get task-machine included... Aug 23 12:01:28 RP: empty oenote() are in output Aug 23 12:01:37 NOTE: Removing failed build target linux-omap1 Aug 23 12:01:37 NOTE: Aug 23 12:01:39 hrw|work: poky has a fix for that Aug 23 12:01:53 hrw|work: Its a base.bbclass bug Aug 23 12:01:56 ah Aug 23 12:02:34 koen: angstrom-bootstrap-image is very X11 related when it comes to building :D Aug 23 12:02:54 hrw|work: it will build everything in task-angstrom with the new bitbake :( Aug 23 12:02:56 i removed the glibc dir from the tmp dir and del'd the cache, but when i re-build... it fails earlier now Aug 23 12:03:02 still on glib, tho Aug 23 12:03:31 03rpurdie * r581 10bitbake/lib/bb/runqueue.py: runqueue.py: Improve circular depends error message Aug 23 12:03:33 bigd0g: after change in machine/include should do clean rebuild - rm whole tmp Aug 23 12:03:57 the whole dir? ok. Aug 23 12:04:00 thanks. Aug 23 12:04:27 yay! Aug 23 12:04:34 svn corrupted my data Aug 23 12:04:58 unless '<<<<<<< .mine' is valid python Aug 23 12:05:14 I can't believe people are still using svn Aug 23 12:05:20 koen: It does leave backup files ;-) Aug 23 12:05:48 i hate that i can't pip svn thru the corporate proxy natively. it doesn't pull the env vars Aug 23 12:05:52 s/pip/pipe Aug 23 12:06:25 koen: can you split task-angstrom into few tasks? Aug 23 12:06:54 bigd0g: even when use svn http:// you cannot? Aug 23 12:06:55 hrw|work: patches welcome :) Aug 23 12:07:19 hrm. well, bitbake calls svn on ftp servers and i have ftp_proxy set as an env var Aug 23 12:07:30 it doesn't seem to go thru the proxy correctly. just hangs Aug 23 12:07:51 koen: task-angstrom-x11.bb? Aug 23 12:09:24 koen: acx-firmware is hx4700 only? Aug 23 12:09:38 hrw|work: hx4700 + h4000 + htcuni Aug 23 12:09:59 RDEPENDS_angstrom-base-depends_append_ipaq-pxa270 = " tiinit acx-firmware " Aug 23 12:10:14 koen: it should be copied for h4000/htcuni then? Aug 23 12:10:15 it's the only device I can test on Aug 23 12:10:31 hrw|work: probably Aug 23 12:10:38 hx2xxx will also need it... Aug 23 12:10:51 RP: that's ipaq-pxa270 ;) Aug 23 12:11:19 koen: It needs splitting. hx2xxx needs lots of TLC when I get a free couple of weeks :-/ Aug 23 12:12:23 koen: http://ewi546.ewi.utwente.nl/tmp/hrw/t/angstrom.diff for start Aug 23 12:13:12 RDEPENDS_angstrom-x11-base-depends is useless Aug 23 12:13:26 virtual/libx11 is DEPENDS not RDEPENDS rather... Aug 23 12:16:43 RP: if 'bitbake -n' then subtasks should not be called rather... Aug 23 12:17:09 NOTE: package umbaumba-17.3-r9 started/completed x3 Aug 23 12:17:49 hrw|work: Did it do anything or was it a logging issue? Aug 23 12:18:16 RP: NOTE: package angstrom-bootstrap-image-1.0: started Aug 23 12:18:17 NOTE: package angstrom-bootstrap-image-1.0: completed Aug 23 12:18:22 03rpurdie 07org.oe.dev * r34f1a5cd... 10/ (1 classes/base.bbclass): base.bbclass: Remove empty messages Aug 23 12:18:25 03Marcin 07org.oe.dev * rd3b7dd0f... 10/ (1 packages/angstrom/task-angstrom.bb): task-angstrom: split it up a bit Aug 23 12:18:27 and it goes for each subtask Aug 23 12:18:42 hrw|work: It always did that Aug 23 12:19:43 koen: nasty commit it was Aug 23 12:19:58 koen: now gpe image for angstrom is fscked Aug 23 12:20:13 I hate the idea of having to carve up the task files due to debain.bbclass :-( Aug 23 12:20:52 kyk/cr Aug 23 12:21:29 NOTE: package gnupg-1.4.2.2: started Aug 23 12:21:29 NOTE: package gnupg-1.4.2.2: completed Aug 23 12:21:37 bitbake 1.6 show it just once Aug 23 12:21:58 hrw|work: Yes, but bitbake 1.6 doesn Aug 23 12:22:00 1.7 show it for each task Aug 23 12:22:13 ''t run each task separately Aug 23 12:22:21 yep Aug 23 12:22:39 I agree it needs addressing though Aug 23 12:22:43 RP: simulation mode does not need to run each task rather Aug 23 12:22:52 (but the bug is not just in -n) Aug 23 12:23:01 hrw|work: why not? Aug 23 12:24:19 RP: for me simulation mode == 'show me which packages you would run at all' (but I'm not yet used to 1.7 behaviuor) Aug 23 12:24:58 i just got bitbake recently and ended up with 1.4.2 ... is it worth upgrading to the latest? 1.6 is it? Aug 23 12:25:06 03Marcin 07org.oe.dev * r8e70c657... 10/ (1 packages/angstrom/task-angstrom-x11.bb): task-angstrom-x11.bb: add missing file Aug 23 12:25:14 bigd0g: I recommend 1.6 Aug 23 12:25:18 hrw|work: simulation mode means do what you'd normally do but don't actually run any task scripts or touch the stamps directory Aug 23 12:25:24 ok Aug 23 12:25:35 RP: like I said - not yet used to 1.7 behaviuor Aug 23 12:25:52 hrw|work: what sets 1.6 apart from 1.4.4 ? Aug 23 12:26:12 bigd0g: speed?! Aug 23 12:26:16 * bigd0g likes how there's no 1.5.x in-between :) Aug 23 12:26:27 bigd0g: old kernel style numbering Aug 23 12:26:43 bigd0g: 1.5 is (was) in svn Aug 23 12:27:03 maintaining odd == unstable and even == stable ? Aug 23 12:27:38 * zecke waits for the delivery service Aug 23 12:27:47 koen: next time when you will use --author to mark changeset as mine use hrw@oe Aug 23 12:27:47 ok. i'll try to upgrade when glibc fails in about 40-mins (when the build gets there) :) Aug 23 12:28:28 koen: you wrote my surname with error next time Aug 23 12:28:47 hmmm Aug 23 12:28:57 I could have sworn I copied the one from your ident Aug 23 12:29:20 I bet I forgot the s again Aug 23 12:29:29 yep Aug 23 12:36:15 lrg: Still get the static Aug 23 12:37:30 XorA: ok, best set the pop_time to 0 :( Aug 23 12:37:46 XorA: i though it might be worth a try Aug 23 12:41:16 lrg: yeah :-( Aug 23 12:50:20 is 'laptop mode' available on ARM? Aug 23 12:50:36 what is laptop mode? Aug 23 12:50:46 do13: ping Aug 23 12:50:56 a feature of the linux kernel to get more battery life Aug 23 12:51:07 less flushes to disk and stuff like that Aug 23 12:51:19 mikearthur: http://blog.windfluechter.net/?q=node/133 Aug 23 12:52:38 hrw|work: ping Aug 23 12:52:43 :D Aug 23 12:52:47 hrw|work: point taken Aug 23 12:59:33 good Aug 23 13:00:55 hrw|work: okay no more PING? So I will fall back to ARP Aug 23 13:01:01 ARP whohas hrw|work for zecke? Aug 23 13:01:15 ;D Aug 23 13:02:08 morning all Aug 23 13:02:14 hi dirk Aug 23 13:02:28 hey Marcin Aug 23 13:02:53 hey Dirk Aug 23 13:03:18 hey Liam Aug 23 13:03:39 lrg: Sorry no time yet to check you latest asoc Aug 23 13:04:28 damn, now I get "mtn: warning: missing packages/bitbake/bitbake_1.0.bb", etc. on an oz354x update Aug 23 13:04:28 what'd I miss when I was gone? :( Aug 23 13:04:53 do13: np Aug 23 13:05:33 Kerwood: its updated to 1.6.0 Aug 23 13:05:38 NOTE: package angstrom-bootstrap-image-1.0-r2: task do_rootfs: started Aug 23 13:05:41 finally Aug 23 13:07:16 hrw|work: I was using 1.4.4 at the time, but I'll assume you mean I should ignore the channel topic and upgrade to 1.6.0. np. Aug 23 13:08:18 Kerwood: .oz354x req 1.4.4 and works with it. 1.6.0 is faster Aug 23 13:08:36 koen: angstrom suxx when it comes to wifi support in bootstrap ;( Aug 23 13:09:08 hrw|work: what's missing? Aug 23 13:09:14 koen: wep, wpa support Aug 23 13:09:38 hrw|work: wep worked for me, so I suspect the zauruskernel have broken deps (again) Aug 23 13:10:06 koen: show me list of kernel-modules from your bootstrap then Aug 23 13:12:09 hrw|work: http://rafb.net/paste/results/0IJH9Z81.html Aug 23 13:13:36 koen: kernel-module-arc4 kernel-module-michael-mic kernel-module-aes kernel-module-ieee80211-crypt-tkip kernel-module-ieee80211-crypt-wep kernel-module-ieee80211-crypt kernel-module-ieee80211-crypt-ccmp are needed for wep/wpa Aug 23 13:14:14 We should add task-machine and teach it that... :) Aug 23 13:14:46 koen: and task-angstrom lack consistency - there is task-wifi but base also contain some of wifi stuff... Aug 23 13:14:52 RP: yes yes YES! Aug 23 13:16:29 do13_: have you run the sharp original firmware on your tosa? Aug 23 13:16:31 patched accepted (direct commit as well) Aug 23 13:17:54 If I'm getting complaints from mtn v0.27 about "missing packages", do I need to upgrade to v0.28? Aug 23 13:17:58 mikearthur: a few mins Aug 23 13:18:26 koen: task-angstrom should rdepend on task-angstrom-wifi? Aug 23 13:18:30 mikearthur: after 10 minutes I flashed openzaurus. Aug 23 13:18:46 do13_: did you ever try drawing lines in a pain program? I'm just wondering if sharp nailed the touch glitches 100% in kernel, or partially in userspace Aug 23 13:18:47 mikearthur: the next day I started 2.6 development Aug 23 13:18:55 do13_: :D Aug 23 13:19:00 mikearthur: I used sharprom for few days on tosa Aug 23 13:19:15 hrw|work: I have no strong opinion about that Aug 23 13:19:25 mikearthur: then few days of oz 3.5.4 with 2.4 kernel Aug 23 13:19:45 koen: we need task-machine/task-distro from poky.. Aug 23 13:20:06 hrw|work: I know Aug 23 13:20:22 I was hoping someone would merge that while I was away Aug 23 13:20:51 mikearthur: Afaik sharp did all this in their kernel Aug 23 13:21:37 do13_: yeh, I've been looking at their code, but can't produce perfect results, i.e. no glitches ever Aug 23 13:23:15 koen: http://pastebin.ca/146254 - much better now Aug 23 13:23:16 mikearthur: pain program? Aug 23 13:24:15 What version of bugzilla is running at bugs.openembedded.org? Aug 23 13:24:24 koen: using 2.4 on tosa is pain Aug 23 13:24:47 koen: It will get merged, I'm just trying to do too many things atm... Aug 23 13:24:48 so 'pain program' == anything under 2.4 kernel on tosa probably ;D Aug 23 13:26:47 03koen 07org.oe.dev * r60e64ed9... 10/ (1 packages/angstrom/task-angstrom.bb): task-angstrom: add wifi modules for wep/wpa Aug 23 13:27:19 * hrw|work decide to not touch packages/angstrom/* Aug 23 13:29:14 mikearthur: their code should work. Do you have the correct values? Aug 23 13:31:07 ibot: botmail to koen: I decided to not touch packages/angstrom/ as it does not have a sense when I show patch to review and it is applied before verification. Aug 23 13:31:14 ibot: thx Aug 23 13:31:16 np Aug 23 13:31:23 do13_:it works, just not perfectly Aug 23 13:31:32 03hrw 07org.oe.dev * r06b0b904... 10/ (1 packages/angstrom/task-angstrom.bb): Aug 23 13:31:32 task-angstrom: cleanup WiFi a bit Aug 23 13:31:32 - added all kernel modules needed for WEP/WPA Aug 23 13:31:32 - task-base RDEPEND on task-base-wifi Aug 23 13:32:14 re koen Aug 23 13:33:09 hey woglinde Aug 23 13:33:30 apparently the keyspan driver doesn't like to be unplugged with an active connection Aug 23 13:34:06 I'm not buying into the 'microkernels can't crash' nonsense Aug 23 13:34:07 mikearthur: do you have a picture? Aug 23 13:34:37 i've got some good news... and some bad news Aug 23 13:34:41 ok, i lied. just bad news Aug 23 13:34:49 the build failed on glibc, again Aug 23 13:36:15 do13_: I can't work out how to output sketchpad to an actual picture Aug 23 13:36:35 do13_: and I'm currently playing with the mask bit Aug 23 13:36:55 zecke: are you available to help me again? Aug 23 13:37:13 hmmm Aug 23 13:37:21 ~seen zecke Aug 23 13:37:30 I don't get this multikernel magic mickeyl is trying to do with the a780 machine Aug 23 13:37:37 zecke is currently on #gpe (5h 6m 58s) #oe (5h 6m 58s) #opie (5h 6m 58s) #handhelds.org (5h 6m 58s). Has said a total of 39 messages. Is idling for 36m 36s, last said: 'ARP whohas hrw|work for zecke?'. Aug 23 13:38:27 bigd0g: maybe Aug 23 13:39:09 zecke: well, i don't want to be a bother. if i should be asking somewhere/someone else, please just tell me. Aug 23 13:39:15 mikearthur: can you try fbgrab? Aug 23 13:41:04 mikearthur: gpe-sketchpad lack export but you can take screenshot with gpe-screenshot ('take screenshot' in 'start' menu) and upload it to handhelds.org/scap Aug 23 13:42:15 sketchpad saves .png pictures in .gpe/*/*.png Aug 23 13:42:24 XorA: skillz Aug 23 13:42:45 .gpe/sketchbook/2006-02-07_21-08-31.png Aug 23 13:42:48 for example Aug 23 13:42:59 XorA: good man Aug 23 13:43:06 which is a DnD game map if I remeber rght Aug 23 13:43:14 mikearthur: taking screenshot to handhelds.org/scap/ is faster if you have net on pda Aug 23 13:43:15 I hear you won't be starting till after I leave :( Aug 23 13:43:20 i just tracked down a breakage of the current OpenSlug HEAD due to a (recent ?) openembedded busybox update no longer seeming to support the cpio -p switch ... would anybody perhaps be able to shed some more light on why this support seems to have been removed from busybox ? Aug 23 13:43:29 mikearthur: earliest is now Sep 25th Aug 23 13:44:33 that you'd be starting, if you accept? Aug 23 13:44:44 mikearthur: I accepted today Aug 23 13:45:29 :D Aug 23 13:45:31 n1 Aug 23 13:45:43 good place to work Aug 23 13:45:46 except for lrg Aug 23 13:46:08 you'll have to pretend to be rubbish at pool and lose to him like I do Aug 23 13:46:33 zecke: new config.log at http://www.rafb.net/paste/results/RJEIrI83.html Aug 23 13:46:41 zecke: new log.do_compile at http://www.rafb.net/paste/results/RPMv6o96.html Aug 23 13:46:45 XorA: congrats Aug 23 13:46:52 zecke: thanks, in advance for any help/advice/direction Aug 23 13:51:14 hrw|work: NOTE: package angstrom-gpe-image-1.0: completed Aug 23 13:51:48 mikearthur: hehe Aug 23 13:52:13 bigd0g: arm-linux-as --version says what? Aug 23 13:52:39 arm-linux-as: command not found Aug 23 13:52:39 bigd0g: the binary is in tmp/cross/bin or "${CROSS_DIR}/bin" Aug 23 13:52:57 XorA: congratulation Aug 23 13:53:11 bigd0g: in OpenEmbedded almost every path is configurable Aug 23 13:53:30 zecke: i see it in tmp/cross/bin ... 1 sec Aug 23 13:54:36 zecke: hrm. i can't seem to execute. do i have to ln it or something? Aug 23 13:55:11 bigd0g: I'm sure you can execute it Aug 23 13:55:16 zecke: if i do '.{full path to arm-linux-as} --version" ... that should do it, right ? Aug 23 13:55:27 bigd0g: it might be -v Aug 23 13:55:37 oh Aug 23 13:55:45 no... i just blanked. Aug 23 13:55:56 bigd0g: I sit in front of a machine without binutils installed Aug 23 13:55:56 zecke: GNU assembler 2.17 Aug 23 13:56:24 bigd0g: before I continue make sure to send a patch for this one Aug 23 13:56:36 bigd0g: look at conf/distro/openzaurus-unstable.conf Aug 23 13:57:21 bigd0g: there are two lines about binutils (setting a preferred version) Aug 23 13:57:33 bigd0g: place these in your local.conf and happy rebuilding ;) Aug 23 13:57:56 zecke: found the file and the lines you were talking about Aug 23 13:58:11 zecke: do you think i'll need to do that for any other of the other apps ? Aug 23 13:58:36 koen: do13_ cheers Aug 23 13:59:04 XorA: so now we have a second person with a company creditcard at fosdem ;0 Aug 23 13:59:07 ;) Aug 23 13:59:15 Yeah Aug 23 13:59:27 Now we can have two tables Aug 23 14:00:21 ? Aug 23 14:00:52 zecke: do i have to do another clean build (i.e. removing the tmp directory again) ? Aug 23 14:00:58 or two days one table Aug 23 14:04:46 16:17 hrw@work:linux$ diffstat patches/orinoco-remove-all-which-are-in-hostap-HACK.patch orinoco_cs.c | 20 -------------------- Aug 23 14:04:53 XorA: you're going to become a l33t kernel sound driver haxzor? congrats :) Aug 23 14:05:26 RP: thats the plan, now lets hope I can rember that C stuff Aug 23 14:05:38 koen: remove all angstrom task packages, clean them and then try gpe-image Aug 23 14:05:42 XorA: First job - get the wm8731 working on corgi :) Aug 23 14:06:10 XorA: second job: get ASoC working on the netbook :) Aug 23 14:06:24 XorA: gratulations Aug 23 14:06:50 :) Aug 23 14:08:29 zecke: hey! it got by binutils! yay! tyvm! Aug 23 14:10:43 do13_: pm Aug 23 14:14:27 3rd job: get ASoC on poodle, 4th: ASoC on collie? madness Aug 23 14:17:19 NOTE: package linux-openzaurus-2.6.17-r17: task do_patch: completed Aug 23 14:17:21 uf. Aug 23 14:19:09 hrw|work: asoc is on poodle thanks to RP/lrg Aug 23 14:20:23 XorA|hiding: on corgi too but still got some problems... Aug 23 14:23:03 hrw|work: corgi is almost perfect, we are just trying to iron out the last annoyance Aug 23 14:24:36 XorA|hiding: then you can work on poodle - wm9731 is known to you already ;) Aug 23 14:25:21 hrw|work: NOTE: package angstrom-gpe-image-1.0: completed (after switching to a new machine) Aug 23 14:25:39 hrw|work: from what RP said, is 90% there, just a speaker bug Aug 23 14:26:02 koen: ok Aug 23 14:26:20 It is mostly there - just a couple of annoyances... Aug 23 14:26:59 RP: do you have something other then spectrum_cs fix to add into OZ kernels? Aug 23 14:27:01 that damn static :-) Aug 23 14:27:22 RP: I would release 3542-rc2 for poodle/tosa this week (no collie this time rather) Aug 23 14:28:11 hrw|work: No, just that and any asoc changes I can get working Aug 23 14:28:20 ok Aug 23 14:28:34 I'm testing ugly hack for orinoco<>hostap now Aug 23 14:28:52 2.6.17 builds with it, fetching 2.6.16 now Aug 23 14:29:22 hrw|work: mikearthur did some progress on the tosa touch. we should wait for this Aug 23 14:29:55 ok Aug 23 14:29:55 hrw|work and do13_: should be done and dusted late this week Aug 23 14:30:06 then I will release rc2 next week Aug 23 14:30:29 I've got it massively reducing the glitches already Aug 23 14:30:30 :D Aug 23 14:30:36 do13_: any news on usbhost suspend/resume? Aug 23 14:30:46 mikearthur: less heartbeats? Aug 23 14:31:17 hrw|work: sadly yes, you'll need to mark the heartbeat feature as removed. No more exciting hospital drawings from the tosa anymore Aug 23 14:32:08 hrw|work: no yet. I'm busy because of payed kernel work in my spare time. Aug 23 14:32:39 hrw|work: I'll ping chris humbert Aug 23 14:32:45 What could cause OE not to create a packages file in deploy/ipk? Aug 23 14:32:56 florian_kc: no INHERIT += "package_ipk" Aug 23 14:33:02 florian_kc: ipk or Packages? Aug 23 14:33:24 koen: Packages Aug 23 14:33:28 florian_kc: rootfs_ipk Aug 23 14:33:37 florian_kc: rootfs_ipk.bbclass Aug 23 14:33:49 florian_kc: you didn't make an image or didn't do 'bitbake package-index' Aug 23 14:34:22 well... i'm building the imho current gpe-image target from oz354x Aug 23 14:35:01 This gpe-image.bb has a "inherit image_ipk" Aug 23 14:36:02 florian_kc: which inherit rootfs_ipk Aug 23 14:36:19 do13_ hrw|work: who wrote the tmiofb? Aug 23 14:36:36 slash, who is Chris Humbert? Aug 23 14:36:40 building image == creating rootfs + converting rootfs into images Aug 23 14:36:54 mikearthur: tosa usbhost support maintainer Aug 23 14:36:56 mikearthur: spyro Aug 23 14:37:16 mikearthur: http://oz.drigon.com/ Aug 23 14:37:18 mikearthur: spyro wrote tmiofb for toshiba e800 iirc Aug 23 14:37:37 spyro wrote a previous version Aug 23 14:37:53 spyro and do13_: is there an easy way of getting the current blank status of the fb, as I can't see one Aug 23 14:38:04 spyro and do13_: however, I am a noob Aug 23 14:38:53 mikearthur: this should be handled via tosa-lcd.c Aug 23 14:39:34 do13_: ok cheers Aug 23 14:40:50 do13_: sorry, it is already in there, or should be added? Aug 23 14:41:13 mikearthur: It should be added Aug 23 14:41:39 mikearthur: have a look at corgi_ts.c. we should pass this via a machinfo struct Aug 23 14:41:42 do13_: sucks ;) Aug 23 14:41:49 ;) Aug 23 14:42:03 do13_: aye, thats what I was looking at Aug 23 14:46:14 do13_: presumably then, we want a tosats_machinfo in sharpsl.h? Aug 23 14:47:03 do13_: ignore me, should be in wm97xx.h Aug 23 14:47:18 mikearthur: done Aug 23 14:47:34 do13_: done as in ignored? Aug 23 14:48:08 ignore me -> done Aug 23 14:48:15 ;) Aug 23 14:48:40 :D Aug 23 14:49:52 Are there any known issues with Firefox 1.5 on XScale-based systems with OE? Aug 23 14:50:24 mattcamp: yes, building it is a horror story Aug 23 14:50:27 mattcamp: from what I remember it build but dont start Aug 23 14:50:38 mattcamp: 1.5.0.6 is in .dev Aug 23 14:51:15 koen, hrw|work: looks like it is a different issue: it tries to load the old packages file while generating a new one. This fails because it is a fresh build Aug 23 14:51:58 For me, Firefox 1.5.0.6 builds but segfaults when I run it. Even if I just run regxpcom it segfaults. Aug 23 14:52:37 florian_kc: ah. that is probably related to my recent checkins.. Aug 23 14:53:03 florian_kc: checking... Aug 23 14:54:31 later all Aug 23 14:55:08 can anyone see ftp://mirrors.kernel.org/gnu/gmp/gmp-4.1.2.tar.gz up? Aug 23 14:55:20 i'm hung at "Logging in as anonymous..." Aug 23 14:55:44 mattcamp: you could try stracing it or running it under catchsegv Aug 23 14:56:30 re Aug 23 14:57:01 woglinde: wb Aug 23 14:57:04 What does catchsegv do, compared to gdb? I'm guessing I'll need a debug build to get useful results with either. Aug 23 14:57:26 no need for debug builds if you are using .dev Aug 23 14:57:44 just install firefox-dbg Aug 23 14:59:31 When was firefox-dbg added? Aug 23 14:59:44 a few weeks ago Aug 23 15:00:00 03hrw 07org.oe.oz354x * r4ed3d94f... 10/ (4 files in 2 dirs): machine configs: moved wpa-supplicant to poodle/zaurus-clamshell so collie does not get it Aug 23 15:00:05 03hrw 07org.oe.oz354x * r36334dd0... 10/ (1 classes/rootfs_ipk.bbclass packages/meta/package-index.bb): rootfs_ipk/packages-index: fix problem when Packages file does not exist Aug 23 15:00:46 Hmm, I have firefox 1.5.0.6 but no -dbg package. Aug 23 15:00:55 I'll make sure I have the latest OE snapshot. Aug 23 15:01:12 I mean the latest code from .dev Aug 23 15:02:32 maybe my update removed -dbg.. Aug 23 15:03:50 Yeah, I don't have a firefox-dbg package. Aug 23 15:04:04 I have -dbg packages for plenty of other libraries, though. Aug 23 15:06:29 I modified my copy of firefox_1.5.0.6.bb to inherit from binconfig, and modified my firefox.inc to copy the header files, libraries, and IDL files to staging, so I could build PyXPCOM as a separate package. Would that affect whether I have a -dbg package? I'd think not. Aug 23 15:06:52 (Firefox 1.5 segfaults on my device even without PyXPCOM.) Aug 23 15:06:55 hello all Aug 23 15:06:59 hi sirfred Aug 23 15:07:23 one day I will learn how to use git and move bitbake git fetcher to use cogito instead of plain git Aug 23 15:09:06 hrw|work: why would one want that? Aug 23 15:09:15 he zecke Aug 23 15:10:10 zecke: seriously I first have to learn both those tools and then will look Aug 23 15:12:06 How do -dbg packages get built anyway? To build a -dbg package for a given package, does something have to be done in the .bb file? Aug 23 15:14:05 zecke: I feel like fetching from git get stuck on packing... rest of build take less time then packing Aug 23 15:15:11 will all the oe kernel patches eventually hit mainline? Aug 23 15:15:26 mikearthur: Z kernels? yes Aug 23 15:15:47 mikearthur: except those marked 'hack, unmergable, unmergable-hack' Aug 23 15:16:05 cool Aug 23 15:16:15 its just cool to know some of my work might hit mainline :D Aug 23 15:17:09 mikearthur: describe patches Aug 23 15:17:26 mikearthur: linux/Documentation/SubmittingPatches is nice read Aug 23 15:18:01 hrw|work: how do you mean "describe patches" Aug 23 15:18:38 mikearthur: linux/Documentation/SubmittingPatches has all info Aug 23 15:19:00 cool Aug 23 15:19:02 mikearthur: 'this patch do that on this machine.\n signed-off-by: I ' Aug 23 15:25:00 RP: 2.6.17+git fails for tosa - known issue? Aug 23 15:25:13 Regarding firefox-dbg, it turns out that the files which should be in firefox-dbg are in the main firefox package, for 1.5 at least. Aug 23 15:25:20 1.5.0.6 specifically Aug 23 15:25:42 mattcamp: in that case gdb will work without extra packages Aug 23 15:25:55 * koen waits for his firefox build to complete Aug 23 15:26:08 hrw|work: yes Aug 23 15:26:11 koen: get a coffee... and a pizza ;) Aug 23 15:26:23 hrw|work: Do not use +git atm until I check in some fixes Aug 23 15:26:38 ok Aug 23 15:26:38 what except of gpe stuff uses dbus in oz? Aug 23 15:26:48 florian_kc: nothing? Aug 23 15:26:49 florian_kc: my girlfriend is getting some burgers and fries from across the streeet right now :) Aug 23 15:27:04 florian_kc: gconf/gnomevfs/avahi Aug 23 15:27:24 Ha. /bin/tinylogin isn't owned by 0:0. Aug 23 15:27:39 Does the .debug directory have to be in LD_LIBRARY_PATH, or will gdb see the files with the debug symbols automatically? Aug 23 15:27:48 hrw|work: so if gpe 2.8 needs dbus > 0.50 it would be safe to switch :-) Aug 23 15:27:54 Which should I use, gdb or catchsegv? I'm not familiar with the latter. Aug 23 15:28:07 koen: well.. not used in the other images Aug 23 15:29:07 gkdial, gpsd, hal, opensync, python-dbus, bluez, avahi, galago Aug 23 15:29:19 the complete list: http://ewi546.ewi.utwente.nl/tmp/dbus.txt Aug 23 15:30:03 do13: is the tosabl device the same as the tosafb device? Aug 23 15:30:29 koen: thanks Aug 23 15:30:37 mikearthur: no Aug 23 15:30:45 RP: cheers Aug 23 15:31:44 mikearthur: bl is backlight, fb is framebuffer Aug 23 15:31:46 When I run regxpcom from Firefox 1.5.0.6 inside gdb, I get this from backtrace after the segfault: Aug 23 15:31:46 #0 0x40009e50 in _dl_cache_libcmp () from /lib/ld-linux.so.2 Aug 23 15:31:46 #1 0x00070498 in ?? () Aug 23 15:35:33 hrw|work: n1, thanks Aug 23 15:36:38 NOTE: Not creating empty archive for firefox-inspector-1.5.0.6-r0 Aug 23 15:36:38 NOTE: Not creating empty archive for firefox-dbg-1.5.0.6-r0 Aug 23 15:36:39 Packaged contents of firefox into /data/build/koen/OE/build/tmp/angstrom/deploy/ipk/firefox_1.5.0.6-r0_armv5te.ipk Aug 23 15:36:39 NOTE: Not creating empty archive for firefox-doc-1.5.0.6-r0 Aug 23 15:36:39 Packaged contents of firefox-dev into /data/build/koen/OE/build/tmp/angstrom/deploy/ipk/firefox-dev_1.5.0.6-r0_armv5te.ipk Aug 23 15:37:04 gah, all you people on x86 stop showing off :-) Aug 23 15:41:56 are you not on an x86 rig? :p Aug 23 15:43:23 bigd0g: no Aug 23 15:44:31 what are you on? Aug 23 15:45:23 XorA|hiding: good idea.. will run x86 build Aug 23 15:45:30 bigd0g: amd64, firefox is broken for crosscompile Aug 23 15:45:33 or other strange for me one Aug 23 15:45:40 koen: So are you going to install your newly built Firefox on a device and try to debug ig? Or were you mainly interested in whether it's packaged correctly? Aug 23 15:46:26 correct packagin Aug 23 15:47:39 tme to get tran Aug 23 15:50:16 How can I be sure that the symbols from the .debug directories under /usr/lib/firefox-1.5.0.6 are being loaded by gdb? Aug 23 15:50:41 mattcamp: Does it say no symbols found? Aug 23 15:51:36 No. Aug 23 15:53:05 mattcamp: Then you're probably ok :) Aug 23 15:53:49 But when I run "info sources", there are no files under the heading "source files for which symbols have been read in" Aug 23 15:54:07 And I know that the files under /usr/lib/firefox-1.5.0.6/.debug contain source file names. Aug 23 15:59:42 ugh. failed on the linux-omap1 files themselves. Aug 23 15:59:50 something to investigate later tonight/tomorrow Aug 23 16:00:00 thanks for the help, zecke et al. Aug 23 16:00:20 I go home Aug 23 16:00:39 * zecke too Aug 23 16:00:55 machine=epia, distro=generic started Aug 23 16:03:07 bye zecke Aug 23 16:03:10 bye hrw Aug 23 16:03:30 hm Aug 23 16:03:46 right I wanted a initrd for pcmcia support Aug 23 16:03:46 bye Marcin, zecke Aug 23 16:03:57 for simpad and nfsroot Aug 23 16:06:39 Oops, should have installed glibc-dbg an hour or so ago. Aug 23 16:09:12 Now when debugging regxpcom in Firefox, I have a much more complete backtrace, but still no arguments or source file names. Do the -dbg packages not contain that much debugging info? Aug 23 16:09:33 mattcamp hm it should Aug 23 16:09:35 strange Aug 23 16:10:37 woglinde: I want to go out this weekend... Aug 23 16:10:40 I'll install -dbg packages for more of the libraries on the device and see what happens. Aug 23 16:11:33 zecke we have our yearly basketball tournament, sorry Aug 23 16:12:21 I'm going home now... Aug 23 16:20:09 RP: ping. Aug 23 16:22:41 03koen 07org.oe.dev * r68fc7216... 10/ (1 packages/gpe-autostarter/gpe-autostarter_0.12.bb): gpe autostarter: add 0.12 Aug 23 16:23:56 03koen 07org.oe.dev * r784b6c66... 10/ (1 packages/gpe-beam/gpe-beam_0.2.9.bb): gpe-beam: add 0.2.9 Aug 23 16:29:33 i wonder if monotone will maanage this Aug 23 16:29:50 hi sirfred Aug 23 16:30:04 RP: Hi. A little question about the w100 framebuffer driver. Aug 23 16:30:18 RP: I've seen that when using 320x240 mode, the external memory is disabled. Is that true? Aug 23 16:41:07 koen: maybe you want to update gpe-edit too? :-) Aug 23 16:46:38 Do the files in glibc-dbg have full debugging information, including prototypes for each function so gdb can show arguments in the backtrace? Aug 23 16:47:12 I know now that regxpcom in Firefox 1.5.0.6 is segfaulting when it calls dlopen, but now I want to know what library it's trying to load. Aug 23 16:48:26 mattcamp: maybe strace would be more helpfull with that? Aug 23 16:48:41 I'll try that. Aug 23 16:50:26 03koen 07org.oe.dev * r587f9108... 10/ (1 packages/gpe-edit/gpe-edit_0.32.bb): gpe-edit: update to 0.32 Aug 23 16:50:34 florian_kc: there you go :) Aug 23 16:50:37 koen: thanks Aug 23 16:50:46 ~lart monotone Aug 23 16:50:47 * ibot changes monotone's permissions to 0777 and tells the world Aug 23 16:52:23 looks like regxpcom is choking on the files in /usr/lib/firefox-1.0.6/components/.debug Aug 23 16:52:56 koen: mtn: warning: rename target conflict: nodes 17017, 17007, both want parent 554, name gpe-autostarter_0.12.bb Aug 23 16:53:08 * koen points to njs Aug 23 16:53:17 koen: i removed my local versions already Aug 23 16:53:31 mtn drop -e? Aug 23 16:54:05 mtn: skipping packages/gpe-beam/gpe-beam_0.2.9.bb, not currently tracked Aug 23 16:54:15 tricked it :-/ Aug 23 16:55:27 found the problem with Firefox 1.5. Because regxpcom attempts to dlopen all files in /usr/lib/firefox-1.5.0.6/components *and its subdirectories*, it tries to load the .so files under .debug, which of course aren't really shared libraries. Aug 23 16:55:35 hmm no, it tricked me Aug 23 16:57:28 So we can move all the .debug directories under /usr/lib/firefox-1.5.0.6 into a firefox-dbg package, but then Firefox won't run when the -dbg package is installed. Aug 23 16:58:03 I guess I could patch XPCOM to explicitly skip the .debug directory. Aug 23 16:58:26 That's probably the best short-term solution. Aug 23 17:04:17 * koen hugs autopanosift Aug 23 17:06:45 So I'll patch Firefox to skip the components/.debug directory when searching for components, then start rebuilding it, eat lunch, pet the cats, and I'll probably still have to wait for it to finish building. Aug 23 17:07:16 mattcamp: I figured out how to exclude the debug stuff Aug 23 17:08:18 I know the .debug directories need to be moved to a firefox-dbg package, but firefox should still run when those directories are there. Aug 23 17:08:32 indeed Aug 23 17:09:38 And the reason Firefox is segfaulting is that it's trying to load the files in components/.debug using dlopen. Aug 23 17:09:43 Do you have a solution for that? Aug 23 17:09:48 nope Aug 23 17:10:17 I'm working on that now. Aug 23 17:12:45 mattcamp: is your patch to add staging bits available somewhere? Aug 23 17:13:00 Not yet. Aug 23 17:13:11 What's the best way to submit patches? I don't have write access to the repository. Aug 23 17:13:27 bugs.openembedded.org Aug 23 17:15:30 I'll push the fix to the -dbg package when the rebuild finishes Aug 23 17:16:50 For the curious, I'm doing all of this for a company that's developing a talking PDA for the blind. Our browser app is based on Firefox and uses PyXPCOM. Aug 23 17:17:00 cool Aug 23 17:17:02 i love monotone :-( Aug 23 17:19:36 hi all Aug 23 17:20:54 x29a: ping Aug 23 17:21:32 Morning Aug 23 17:21:43 Did someone have changes for the bugzilla theme? Aug 23 17:23:11 ggilbert: hrw|gone was talking about it Aug 23 17:23:16 ah Aug 23 17:23:38 sirfred: yes, that is true Aug 23 17:23:48 RP: Any reason for that? Aug 23 17:24:01 sirfred: power consumption really Aug 23 17:24:09 sirfred: and its what the 2.4 driver did Aug 23 17:24:33 RP: OK. Aug 23 17:24:37 sirfred: sorry for the delay, got distracted by a phone call Aug 23 17:24:47 koen: Which file(s) will you be updating for the firefox-dbg fix? Aug 23 17:25:05 RP: I was thinking that 320x240 modes could benefit from the external memory (backbuffer, surfaces,...) Aug 23 17:25:09 firefox.inc and a PR bump in firefox_1.5.06.3.2.1.34.bb Aug 23 17:25:48 mattcamp: http://ewi546.ewi.utwente.nl/tmp/ff.diff Aug 23 17:26:20 RP: I've started this: http://libw100.sourceforge.net Aug 23 17:26:45 sirfred: agreed, if we're going to make use of the memory Aug 23 17:26:48 I have read a lot old ibot threads about arnold block and simpad - it seems there was the possibility to build it with openembedded. is this also possible nowadays? Aug 23 17:27:18 RP: Anyway 320x240 modes are not a priority for me now. Aug 23 17:27:26 sirfred: I wonder if we could have it switch via some option to fbset? Aug 23 17:27:49 koen: Thank you. Aug 23 17:27:49 sirfred: good news on the libw100 front :) Aug 23 17:27:59 RP: Well, my english is pathetic, sure. Aug 23 17:28:21 RP: And spent a day trying to setup the wiki on sourceforge. Aug 23 17:28:38 RP: But well, we have something public, finally. Aug 23 17:29:00 sirfred: yes, its a start :) Aug 23 17:29:51 koen: Have you verified that the firefox package has no .debug dirs now? Aug 23 17:29:52 RP: I've published a first version of the library and also pushed it into the SVN repository provided by sourceforge. Aug 23 17:30:05 mattcamp: I'm about to do that Aug 23 17:30:15 Perhaps more people gets involved. Aug 23 17:30:21 sirfred: do you have a .bb already? Aug 23 17:30:33 koen: For the library? Aug 23 17:30:41 yes Aug 23 17:30:52 koen: It's oriented for experimenting at this stage. Aug 23 17:31:40 koen: But a .bb would be an easy task. Aug 23 17:35:04 sirfred: I know :) Aug 23 17:38:07 mattcamp: NOTE: package firefox-1.5.0.6-r1: task do_package: started Aug 23 17:51:13 koen: still packaging? Aug 23 17:51:33 mattcamp: tweaking it now Aug 23 17:52:15 one file left: ar x firefox_1.5.0.6-r1_armv5te.ipk ; tar ztvf data.tar.gz | grep debug Aug 23 17:52:55 I can do the next rebuild if you want. Aug 23 17:53:01 I'll need to anyway. Aug 23 17:53:42 Perhaps it's enough to remove the do_package stamp and just let it re-run do_package, instead of rebuilding Firefox in its entirety. Aug 23 17:54:06 I bumped pr, so it rebuilt is completely Aug 23 17:54:25 bitbake -b firefox_1.5.0.6.bb -c package -f :) Aug 23 17:55:02 mattcamp: mtn pull :) Aug 23 17:56:00 03koen 07org.oe.dev * r84eb5c05... 10/ (3 files in 2 dirs): Aug 23 17:56:00 firefox 1.5: put debug files in the -dbg package, the main package still has empty debug dirs: Aug 23 17:56:00 koen@bitbake:/data/build/koen/OE/build/tmp/angstrom/deploy/ipk$ ar x firefox_1.5.0.6-r1_armv5te.ipk ; tar ztvf data.tar.gz | grep debug Aug 23 17:56:00 drwxr-xr-x root/root 0 2006-08-23 20:07 ./usr/lib/firefox-1.5.0.6/components/.debug/ Aug 23 17:56:00 drwxr-xr-x root/root 0 2006-08-23 20:07 ./usr/lib/firefox-1.5.0.6/plugins/.debug/ Aug 23 17:56:03 drwxr-xr-x root/root 0 2006-08-23 20:07 ./usr/lib/firefox-1.5.0.6/extensions/inspector@mozilla.org/components/.debug/ Aug 23 17:56:46 2.0M firefox-dbg_1.5.0.6-r1_armv5te.ipk Aug 23 17:59:53 koen: Can you give me the final patch? Aug 23 18:00:12 mattcamp: it's in the repo now Aug 23 18:00:32 is it the same as the one you posted earlier? Aug 23 18:00:43 no, a bit more refined Aug 23 18:01:10 mattcamp: http://lists.linuxtogo.org/pipermail/openembedded-commits/2006-August/000025.html Aug 23 18:09:02 wow Aug 23 18:09:08 ggilbert_: after this Aug 23 18:09:42 mattcamp: thanks for bringing it to our attention Aug 23 18:10:12 koen: No problem, and thanks for fixing it. Rebuilding now. I'll submit my other patches soon. Aug 23 18:11:35 mr_nice_slacker: pong Aug 23 18:15:39 hrw: it also needs to lose the / at the end Aug 23 18:16:00 JustinP: are you sure? Aug 23 18:16:06 should be valid Aug 23 18:17:02 03koen 07org.oe.dev * rc02b3cdf... 10/ (1 packages/mozilla/firefox.inc): firefox.inc: put inspector bits in the inspector package Aug 23 18:17:05 is ok in html4 Aug 23 18:17:23 yesh Aug 23 18:17:27 but this is
Aug 23 18:17:31 it's valid XML Aug 23 18:17:36 but browsers may not like it Aug 23 18:17:41 JustinP: ehm Aug 23 18:17:49