**** BEGIN LOGGING AT Mon Oct 23 02:59:57 2006 Oct 23 03:13:50 NOTE: package glibc-2.3.5+cvs20050627-r12: task do_package_write: started Oct 23 03:14:06 ERROR: Exception:exceptions.OSError Message:[Errno 2] No such file or directory: '/home/slug/debianslug/tmp/work/glibc-2.3.5+cvs20050627-r12/install/glibc-locale' Oct 23 03:14:09 anyone else seen this? Oct 23 03:20:18 a good time to switch over 2.4/2.5! ;-) Oct 23 03:20:22 night. Oct 23 05:10:18 how to list files of a ipk in build environment? Oct 23 05:57:45 sharprom ? I build failed due to this compiling . Oct 23 06:57:15 morning Oct 23 07:04:59 reading backlog... Oct 23 07:05:02 yo Oct 23 07:25:47 good morning Oct 23 07:26:01 yo likewise Oct 23 07:26:08 i was waiting after you Oct 23 07:26:22 what for? Oct 23 07:26:35 in you icecc.bbclass Oct 23 07:26:38 at the end Oct 23 07:26:41 print ICECC_VERSION="${@icc_version(bb,d)}" Oct 23 07:26:55 print is not a shell command Oct 23 07:26:59 isn't it ? Oct 23 07:27:08 ah yes, good point. Oct 23 07:27:18 :)) Oct 23 07:29:20 I am still having problems making it work OK. I do not get why the path of CC1 and CC1PLUS is made into /usr/bin into the tar file. Oct 23 07:30:02 During some tests, the tar file is transferred to remote clients, but they fail to execute the chroot compiled, probably a library or library loader issue. Oct 23 07:30:13 chroot compiled *gcc* Oct 23 07:31:11 likewise, : when you finished it , i'll be pleased to try it but i need a little doc ;) Oct 23 07:32:17 Genesis: well I am just modifying zecke's icecc.bbclass, with what docs I can find available on the net. never did any dist cc before. Oct 23 07:33:48 doc about how to deal with the bbclass , not with icecc Oct 23 07:34:03 morning Oct 23 07:37:23 Genesis: ok, yes Oct 23 07:37:53 Genesis: When I got it to work, I'll document it. But for the moment, any help *getting it to work* would be great! Oct 23 07:43:25 morning all Oct 23 07:56:05 koen|away: http://bugs.debian.org/cgi-bin/pkgreport.cgi?tag=ftbfs-gcc-4.2;users=tbm@cyrius.com can interest you maybe Oct 23 08:02:27 hrw|work: are there any known problems building glibc-2.3.5-cvs in OE at the moment? Oct 23 08:02:51 I get: Oct 23 08:02:54 ERROR: Exception:exceptions.OSError Message:[Errno 2] No such file or directory: '/home/slug/debianslug/tmp/work/glibc-2.3.5+cvs20050627-r12/install/glibc-locale' Oct 23 08:03:00 during glibc packaging Oct 23 08:03:51 rwhitby: I built it in last week Oct 23 08:04:36 any other major changes that I might have missed (have been away on business for a number of days) which might require me to rebuild, or blow away cache or tmp ? Oct 23 08:05:56 qemu bump but it should not be related Oct 23 08:06:31 thanks - I'll start looking deeper then. Oct 23 08:10:35 Hmm - directories like glibc-locale-be and glibc-locale-ca, etc exist, but it seems to be looking for just glibc-locale with no suffix. Oct 23 08:10:59 Is there some distro variable that we've forgotten to set for locales? Oct 23 08:12:38 hey koen Oct 23 08:13:00 hi koen Oct 23 08:13:17 hey rwhitby Oct 23 08:13:19 * koen is at an army base now Oct 23 08:13:25 rwhitby: ah.. do_split_locales etc are reworked Oct 23 08:14:06 hrw|work: ok, that sounds relevant Oct 23 08:14:15 what needs to change in the distro as a result? Oct 23 08:15:05 koen should know more then I Oct 23 08:15:26 koen: ERROR: Exception:exceptions.OSError Message:[Errno 2] No such file or directory: '/home/slug/debianslug/tmp/work/glibc-2.3.5+cvs20050627-r12/install/glibc-locale' Oct 23 08:15:33 when building slugos glibc Oct 23 08:15:48 directories like glibc-locale-be and glibc-locale-ca, etc exist, but it seems to be looking for just glibc-locale with no suffix. Oct 23 08:15:51 check the OE mailinglist Oct 23 08:16:23 ok, been busy with business lately, so haven't kept up with the list. where are the list archives now? Oct 23 08:16:36 specifically http://lists.linuxtogo.org/pipermail/openembedded-devel/2006-October/000655.html Oct 23 08:16:52 koen: thx muchly Oct 23 08:17:11 shouldn't someone have bumped the PR to do that automatically? Oct 23 08:17:46 (i.e. for autobuilders) Oct 23 08:17:59 * rwhitby considers doing that ... Oct 23 08:20:41 * rwhitby just blows away tmp/*/glibc-* on the autobuilders instead. Oct 23 08:22:36 03lenehan 07org.oe.dev * rb20a6d62... 10/ (3 files in 3 dirs): Oct 23 08:22:36 coreutils 5.3: Fix for glibc 2.5 by renaming the tee function to tee_files Oct 23 08:22:36 to prevent a conflict with the tee function definied by glibc 2.5 The name Oct 23 08:22:36 tee_files is what is used in newer coreutils distriubtions. See bug #1483. Oct 23 08:30:06 good morning Oct 23 08:31:41 hi florian_kc Oct 23 08:36:58 03lenehan 07org.oe.dev * rb20a6d62... 10/ (3 files in 3 dirs): Oct 23 08:36:58 coreutils 5.3: Fix for glibc 2.5 by renaming the tee function to tee_files Oct 23 08:37:00 to prevent a conflict with the tee function definied by glibc 2.5 The name Oct 23 08:37:04 tee_files is what is used in newer coreutils distriubtions. See bug #1483. Oct 23 08:37:58 hey florian_kc Oct 23 08:41:13 * koen stabs EDGE Oct 23 08:43:45 03lenehan 07org.oe.dev * rb20a6d62... 10/ (3 files in 3 dirs): Oct 23 08:43:46 coreutils 5.3: Fix for glibc 2.5 by renaming the tee function to tee_files Oct 23 08:43:46 meeting time ;( Oct 23 08:43:48 to prevent a conflict with the tee function definied by glibc 2.5 The name Oct 23 08:43:51 tee_files is what is used in newer coreutils distriubtions. See bug #1483. Oct 23 08:43:53 CIA-4: kick Oct 23 09:01:28 03rwhitby 07org.oe.dev * re327ba72... 10/ (1 packages/linux/ixp4xx-kernel_2.6.19-rc2.bb): ixp4xx-kernel: Updated 2.6.19 svn rev Oct 23 09:08:00 koen: small question: do you a LOFT board? Oct 23 09:09:54 03lenehan 07org.oe.dev * r25768be5... 10/ (1 packages/clearsilver/clearsilver_0.10.3.bb): Oct 23 09:09:54 clearsilver 0.10.3: Stage the headers and libraries so that C based Oct 23 09:09:54 applications that use clearsilver can be built under OE. Oct 23 09:10:05 * rwhitby realises that *every* OE program with locale needs to now be rebuilt :-( Oct 23 09:11:06 Surely someone could think of a way to trigger that automatically for the incremental autobuilders ... Oct 23 09:11:18 likewise: I have a loft board, yes Oct 23 09:11:21 brb Oct 23 09:11:28 koen: cf works ok? Oct 23 09:15:31 likewise: last time I tried, yes Oct 23 09:16:11 koen: is the IDE driver open-sourced? I see bootlogs with "Gateworks Avila IDE/CF driver" which is what I presume what the LOFT uses as well? Oct 23 09:16:12 rwhitby: wait some days with rebuilding from scratch, we're about to clean up a lot of packages containing spurious .debug directories Oct 23 09:16:33 "gateworks avila" if the official name for the loft Oct 23 09:17:18 koen: any idea where the IDE driver is at in the kernel source? Oct 23 09:17:26 * koen dunnos Oct 23 09:17:46 I think it needs rewriting to libata anyway Oct 23 09:18:15 koen: LOFT != http://www.gateworks.com/, is it? Oct 23 09:18:36 likewise: you'd have to ask [g2] when he wakes up Oct 23 09:20:29 afair this is loft: http://www.giantshoulderinc.com/hardware.html Oct 23 09:20:46 koen: I'll see if [g2] wakes up today, tnx Oct 23 09:22:33 argh Oct 23 09:22:37 * koen stabs read-elf Oct 23 09:22:44 * koen stabs pax-utils and binutils as well Oct 23 09:28:37 koen: interesting ARM glibc2.4 gcc 4.1.1 wierdness email on lakml Oct 23 09:32:19 koen: thx - we'll do a complete rebuild on the weekend Oct 23 09:32:55 likewise: we have a libata patch in the nslu2-linux svn kernel repo Oct 23 09:33:25 http://trac.nslu2-linux.org/kernel/browser/trunk/patches/2.6.19 Oct 23 09:47:49 who is the glibc-2.5 guru? Oct 23 09:48:02 koen? Oct 23 09:48:18 dropbear won't start on the OSK anymore Oct 23 09:48:28 it fails at an assert in glibc Oct 23 09:50:58 assert is in ./sysdeps/unix/sysv/linux/check_pf.c Oct 23 09:52:28 Fails at line 222 in http://pastebin.ca/216867 Oct 23 09:56:37 morning all Oct 23 09:59:06 hi lrg Oct 23 10:02:28 Crofton: I would add: assert(offsetof(struct req, g) == sizeof(nlh)); before it. Oct 23 10:02:52 Crofton: if that fails, there are bytes between nlg and g. Oct 23 10:02:58 hi Liam, mickeyl Oct 23 10:02:59 s/nlg/nlh Oct 23 10:03:41 someone know how to get BUILD_ALL_DEPS behaviour again? Oct 23 10:03:52 Crofton: And add assert(sizeof(g) == 1); Oct 23 10:04:28 morning folks Oct 23 10:04:41 likewise, I ma not much of a glibc hacker ... Oct 23 10:04:50 I am hoping someone recognizes the problem :) Oct 23 10:05:20 thinking about it last night, the compiler should know the answer at compile time Oct 23 10:05:21 Crofton: is the pastebin code snippet from glibc or from dropbear? glibc? Oct 23 10:05:30 glibc Oct 23 10:06:08 in this file Oct 23 10:06:30 I think the assert is there to let you know the compiler botched some alignment thing Oct 23 10:07:39 mickeyl: do you know how to get built task-opie with ALL deps in other way then 'bitbake opie-image'? Oct 23 10:07:39 03mickeyl 07org.oe.dev * r539f35ff... 10/ (1 packages/libxml/libxml2_2.6.26.bb): libxml: revert SRC_URI change to tar.bz2 -- I don't see any bz2 on ftp.xmlsoft.org Oct 23 10:07:39 koen, did you see my glibc-2.5 problem? Oct 23 10:08:01 hrw|afk: bitbake task-opie doesn't do it? Oct 23 10:08:01 does not Oct 23 10:08:02 Crofton: just reading about it Oct 23 10:08:22 Crofton: I didn't see a problem with dropbear yet, but I'll to test in the evening Oct 23 10:08:26 mickeyl: so now I have to wait ~4-8h to get base system built Oct 23 10:08:36 mickeyl: zecke should join the dutch army as an MP, lots of chicks there :) Oct 23 10:08:38 looks like an alignment check is failing Oct 23 10:09:00 koen: we don't want zecke to be even more distracted by life :) he should rather get more into coding ;) Oct 23 10:09:23 * koen stabs readelf again Oct 23 10:09:35 koen: you have to be healthy to join army Oct 23 10:09:43 what are you doig koen ? Oct 23 10:10:00 Crofton: measuring people at an army base Oct 23 10:10:07 including said chicks Oct 23 10:10:08 hmmm Oct 23 10:10:44 koen: <20 years chicks too? Oct 23 10:10:54 * hrw|work watched clerks2 yesterday Oct 23 10:10:58 hrw|work: yes, it's recrute day today Oct 23 10:11:03 so lots of 17 year olds Oct 23 10:11:18 lol dirty old men Oct 23 10:11:31 * Crofton resembles that remark Oct 23 10:11:41 koen: what does Ninky say about it? :D Oct 23 10:12:00 17 year olds ... mhh... Oct 23 10:12:29 hrw|work: nynke gets to measure the 17 year old boys when she works :) Oct 23 10:12:47 heh. there is a huge difference between boys and girls with 17. Oct 23 10:12:48 mickeyl: rumours say that it easier to get 17 years old one do some nasty things then >30y one ;P Oct 23 10:12:57 hrw|work: it's not just rumours :D Oct 23 10:13:02 hrw|work: don't tempt me Oct 23 10:13:08 do I understand correctly - koen measures 17y chicks for work and hacks embedded linux on freetime? :] Oct 23 10:13:10 * mickeyl watches the ring glowing Oct 23 10:13:18 one ring to rule them Oct 23 10:13:25 suihkulokki: yes Oct 23 10:13:50 so, that means... next time koen is responsible for bringing the chicks to the aftershow parties Oct 23 10:13:57 i.e. FOSDEM07, OEDEM07, ... Oct 23 10:14:57 hmmm Oct 23 10:15:00 I wonder what measurements he is taking? Oct 23 10:16:39 Crofton: the ones so they can fit into these: http://bsst.de/ Oct 23 10:17:28 how the hell did koen wrangle a job measuring breasts? Oct 23 10:18:31 yeah. how? Oct 23 10:18:38 do they take freelancers? Oct 23 10:18:49 (the job, not the chicks... well) Oct 23 10:18:59 I bet he wears an FBI t-shirt as well Oct 23 10:19:04 * koen hides laptop Oct 23 10:19:13 hehe Oct 23 10:22:14 does nynke know about this? Oct 23 10:22:17 :-) Oct 23 10:22:40 likewise: it sounds like she measures the armoured cod peices Oct 23 10:22:47 * mickeyl whistels 'ignorance is bliss' Oct 23 10:22:51 before the days of body armor http://www.wind-drifter.com/VietNam/ReconReunion.htm Oct 23 10:35:28 hmm.. looks like jffs2 changed badly past 2.6.18 Oct 23 10:37:04 how bad? Oct 23 10:43:38 on lakml people report problems Oct 23 10:46:03 hrw|work, I turned on alsa support in the 2.6.18+got omap1 kernel Oct 23 10:46:13 I noticed Oct 23 10:46:27 hopefully it works ;) Oct 23 10:50:27 OT: I love didactic. To teach students addition we say that x+y=2 because it is more easy to understand Oct 23 10:50:54 mickeyl: something for you: Oct 23 10:50:55 NOTE: multiple providers are available (glibc, sharprom-toolchain-native); Oct 23 10:50:56 NOTE: consider defining PREFERRED_PROVIDER_virtual/libc Oct 23 10:51:16 should I use bb 1.7? Oct 23 10:51:48 Crofton: why? Oct 23 10:52:02 why not? Oct 23 10:52:14 Crofton: BitBake 1.7 rocks but it is still in development and considered as unstable Oct 23 10:52:19 I got burned forgetting to upgrade Oct 23 10:52:20 ok Oct 23 10:52:25 Crofton: so if you love unstable :) Oct 23 10:52:25 that was the wuestion :) Oct 23 10:52:32 not at the moment Oct 23 11:01:08 zecke: even without insanity my build is failing Oct 23 11:01:20 zecke: I'll try to track it down tonight/tomorrow Oct 23 11:01:30 koen: send me the config.log Oct 23 11:01:49 ljp: thanks for your response Oct 23 11:01:58 ljp: I still wonder about some stuff Oct 23 11:02:24 zecke: this should be it: ewi546.ewi.utwente.nl/tmp/config.log Oct 23 11:02:57 arm-angstrom-linux-gnueabi-g++ Oct 23 11:03:01 not found Oct 23 11:04:16 * koen bitbake gcc-cross-initial Oct 23 11:04:54 koen: could you paste the error as well? Oct 23 11:05:48 actually not having g++ at this point is okay Oct 23 11:06:21 /data/build/koen/OE/build/tmp/angstrom/cross/lib/gcc/arm-angstrom-linux-gnueabi/4.1.1/../../../../arm-angstrom-linux-gnueabi/bin/ld: cannot represent machine `arm:hx4700' Oct 23 11:06:25 is weird Oct 23 11:06:57 zecke: ewi546.ewi.utwente.nl/tmp/log.do_configure.11663 Oct 23 11:07:35 koen: your binutils or such try to create something weird :) Oct 23 11:08:18 let me unset MACHINE in shell env Oct 23 11:12:20 hi lala Oct 23 11:12:56 "Is is up to date with the latest cvs OZ" Oct 23 11:12:57 some people.... Oct 23 11:14:38 <[lala]> hi guys Oct 23 11:14:44 koen: oesf again? Oct 23 11:14:52 duh Oct 23 11:15:07 where else can something like that come from? Oct 23 11:17:22 koen: so either gcc emits something stupid, or binutils Oct 23 11:17:38 koen: but grep MACHINE | grep getnv didn't show anything Oct 23 11:18:07 zecke: I'll have a closer look when I'm not behind a sucky link Oct 23 11:19:00 cbrake: could you check if http://www.openembedded.org/viewmtn/revision.psp?id=46b07c78864648b21e7b885096464b61a30fa717 solves the problem with checkroot you were seeing? Oct 23 11:19:17 koen: sure Oct 23 11:19:24 thanks Oct 23 11:20:55 koen: yes, that should do it. I'll test and let you know if I run into problems. Oct 23 11:22:43 koen: you should like my reply - will send it in a minute Oct 23 11:23:12 koen: may I update viewmtn? Oct 23 11:25:27 time to create conf/distro/umbaumba-17.3.conf Oct 23 11:27:37 hrw|work: what is umbaumba? Oct 23 11:28:12 cbrake: my word which has any meaning Oct 23 11:28:19 kind of placeholder Oct 23 11:28:36 hrw|work: ahh Oct 23 11:28:39 hrw|work: Do you remember yesterday when we talked about moving Opie over to task-base? How do i start, wich files should i edit? Is it only task-base or is it others as well? Oct 23 11:29:23 goxboxlive: task-opie and opie-image Oct 23 11:29:34 thx Oct 23 11:30:01 goxboxlive: one question? does Opie link? Oct 23 11:30:18 cbrake: its like 'package foo depends on umbaumba' or 'so you are running oz 3.5.4.3 on sl-a300 and have problems with umbaumba?' Oct 23 11:30:29 SO if i do the same as koen has done with task-angsrom-gpe and angstrom-gpe-image it will install all kernel modules as in gpe right? Oct 23 11:30:47 goxboxlive: kernel modules are task-base thing Oct 23 11:30:59 zecke: I asked you yesterday where i should add the link stuff, but you never answerd me. Oct 23 11:31:12 hrw|work. thx Oct 23 11:31:42 goxboxlive: grep conf/ for -fvisibility-inlines-hidden Oct 23 11:31:57 goxboxlive: remove that Oct 23 11:32:04 goxboxlive: rebuild QtE at least and the app you tried Oct 23 11:32:56 ok, do upi mean in build/conf or in org.openembedded.org/conf? Oct 23 11:33:32 "The rumours exists that UmbaUmba 17.3 is available for sl-c3000 but no public release was done because it's maintainer does not like oesf users." Oct 23 11:34:03 koen: do you like it? Oct 23 11:34:43 * XorA is tempted to release UmbaUmba 18.1 :-) Oct 23 11:34:56 hrw|work: I'm tempted to respond with "it runs great on my 2ghz 800x600 arm gadget" Oct 23 11:35:14 zecke: i got 'no such file or directory' Oct 23 11:35:20 hrw|work: or "kinda slow on the sl-4000" Oct 23 11:36:02 * koen yawns Oct 23 11:38:55 goxboxlive: then search harder Oct 23 11:39:03 goxboxlive: use a substring, etc... Oct 23 11:43:12 ok Oct 23 11:46:22 NOTE: package opie-console-1.2.2-r0: task do_compile: started Oct 23 11:48:53 conf/distro/include/angstrom-glibc.inc:CXXFLAGS += "-fvisibility-inlines-hidden" Oct 23 12:09:02 zecke: bool Q_EXPORT QRect::intersects( const QRect &r ) const Oct 23 12:09:22 zecke: should not this unhide this method? Oct 23 12:09:22 hrw|work: but that is inline? Oct 23 12:09:38 hrw|work: -fvisibility is only about inlined methods Oct 23 12:10:08 zecke: it is not Oct 23 12:13:06 atleast it is nor marked as inline Oct 23 12:22:42 later all Oct 23 12:23:22 hrw|work: so the compiler shouldn't touch it at all :) Oct 23 12:23:54 hrw|work: it looks like it (the compiler) inlined it within QtE but not within userspace apps :) Oct 23 12:24:02 but that is only guessing Oct 23 12:24:56 maybe Oct 23 12:25:18 but as the compiler is older than me, I won't talk/speak up for the compiler Oct 23 12:26:50 ;))) Oct 23 12:26:53 foodtime Oct 23 12:28:24 hrw|work, ;). so, what kind of food are you eating ? Oct 23 12:41:51 leoncamel: not cjkv one ;) Oct 23 12:54:01 Hi! Oct 23 12:54:21 mickeyl, yep, http://ftp.gnome.org/pub/GNOME/sources/libxml2/2.6/ is where I saw tar.bz2 %) ... Oct 23 12:54:29 :) Oct 23 12:54:58 mickeyl: http://www.openembedded.org/viewmtn/diff.psp?id1=7f473638ef944831e2ffb5346ae969d18d2472b9&id2=0caa2fad92941b6d48a8b8a25ef1fc6f99482005&fname=packages/patcher/patcher_20040913.bb Oct 23 12:55:04 mickeyl: do you want such a feature? Oct 23 12:55:16 YEAH!!!!! Oct 23 12:55:22 mickeyl: it was there :) Oct 23 12:55:34 cool Oct 23 12:55:42 now could i see who has made the changes on the same page? Oct 23 12:55:44 and Oct 23 12:55:52 an accumulated view of all those diffs for one file?` Oct 23 12:55:55 (+ authors) Oct 23 12:55:56 http://www.openembedded.org/viewmtn/revision.psp?id=0caa2fad92941b6d48a8b8a25ef1fc6f99482005 Oct 23 12:55:59 that would really hel Oct 23 12:56:27 morning Oct 23 12:56:42 hrw|work / Zecke : I commented out this line "conf/distro/include/angstrom-glibc.inc:CXXFLAGS += "-fvisibility-inlines-hidden"" but it didnt help. Opie-console is still failing with the same complaines Oct 23 12:57:01 goxboxlive: then read the other 3/4 of what I said Oct 23 12:57:07 goxboxlive: did you rebuild QtE? Oct 23 12:57:13 being able to see all diffs on a file with their corresponding authors would be pretty cool Oct 23 12:57:25 zecke: i will Oct 23 12:59:49 mickeyl: I removed that feature as it was destructive and unused :) Oct 23 13:00:37 destructive? not at all Oct 23 13:03:58 mickeyl: mtn annotate? Oct 23 13:05:18 dunno. something that i can do on the web Oct 23 13:05:31 a function of the filebrowser or so Oct 23 13:10:52 okay annotate is still slow :} Oct 23 13:11:09 zecke: it build :-) Oct 23 13:11:12 +t Oct 23 13:11:41 also opie-taskbar Oct 23 13:12:52 okay, then we will need to enable visibility in QtE again, or dig into gcc Oct 23 13:13:18 ok, i'll report back if opie-image succed or not Oct 23 13:14:06 * zecke bites in his tounge Oct 23 13:14:26 on Oct 23 13:17:04 I'm using the generic distro which has now been updated to compile glibc 2.5. However when I build my image bitbake reports that version 2.5 of glibc does not exist, however when I check in the packages folder it does exist. Oct 23 13:18:04 Gerrath: misleading error message, some of the deps can not be fullfilled Oct 23 13:18:17 Gerrath: check what other distributions using glibc2.5 set as PREFERRED_PROVIDERS Oct 23 13:18:32 me too, but things still seem to work Oct 23 13:19:18 zecke, thanks. The generic.conf distro file I'm using is the latest from the OE repository. I'll try to find what is missing. Oct 23 13:19:50 Crofton, check your files, I think you will find that it built glibc 2.3 and not 2.5 Oct 23 13:20:31 nope, 2.5-r2 Oct 23 13:23:03 ~lart HP for fucking up printing on my OSX Oct 23 13:23:04 * ibot burns HP to a crisp with a laser for fucking up printing on my OSX Oct 23 13:23:11 Crofton, interesting.. when I run it it starts building 2.3 Oct 23 13:23:47 I am trying to use the toolchain stuff from angstrom Oct 23 13:23:47 guys - I'm thinking about RFC - granulate libraries packaging Oct 23 13:24:05 NOTE: Multiple libraries (libbfb.so.0, libmulticobex.so.1, libobexftp.so.0) found and LEAD_SONAME not defined Oct 23 13:24:08 etc. Oct 23 13:27:17 sounds like a good task for a bbclass Oct 23 13:28:44 mickeyl: which will split such big packages into separate ones? Oct 23 13:28:49 NOTE: multiple providers are available (glibc, sharprom-toolchain-native); Oct 23 13:28:52 ;D Oct 23 13:29:00 hrw|work: yeah Oct 23 13:29:14 recently i had a problem with sharprom-toolchain-native Oct 23 13:29:23 DEFAULT_PREFERENCE = "-1" would have no effect Oct 23 13:29:26 echo PREFERRED_PROVIDER_virtual/libc := "sharprom-toolchain-native">>conf/distro/include/angstrom.inc Oct 23 13:29:37 it would always be chosen for some packages Oct 23 13:29:43 if that happens for you as well, we need to fix it Oct 23 13:29:47 mickeyl: BROKEN = "1" Oct 23 13:29:57 will that be sufficient? Oct 23 13:30:23 dont know Oct 23 13:37:26 Crofton: "OK, I've pushed that part too and tagged 2.6.18-omap1." Oct 23 13:37:49 Crofton: time for 2.6.18-omap1.bb? Oct 23 13:39:54 hrw|work: rm sharprom-toolchain-native_1.0.bb works well Oct 23 13:40:22 ade|desk: I know, but prefer to get it fixed correctly Oct 23 13:41:15 Crofton: NOTE: package linux-omap1-2.6.18-omap1: started Oct 23 14:01:21 Crofton: NOTE: package linux-omap1-2.6.18-omap1-r0: task do_build: completed Oct 23 14:04:56 zecke / hrw|work : Opie image where build successfully, it's based on Angstrom distro. yeeahaaa :-) Oct 23 14:06:04 goxboxlive: not default angstrom config anyway ;( Oct 23 14:13:23 no i now, but it is distro angstrom. Oct 23 14:14:32 goxboxlive, so, it works ok? Oct 23 14:15:41 i havent tried it yet. I have to wait to later this evening. I just picked up my child atkondergarden so i have to play with him a little :-) But meanwhile i building kdepim image and other stuff. Oct 23 14:16:34 goxboxlive, Sure, have good time! ;-). But even "build completes ok" is great news. keep us posted. Oct 23 14:17:42 yes that is good news. Zecke and hrw gave some good tips. Oct 23 14:18:15 even opie-kdepim-image where build :-) Oct 23 14:19:00 goxboxlive: now do me a favor Oct 23 14:19:45 i will Oct 23 14:20:00 goxboxlive: debug gcc and see if this symbol got inlined :) Oct 23 14:20:19 how do i do that? Oct 23 14:20:39 do you mean rebuild gcc with -DDDD option? Oct 23 14:20:40 goxboxlive: dunno, find the place in gcc that decides to inline methods Oct 23 14:20:58 hack it :) Oct 23 14:21:10 zecke: i would if i could, but i even dont know what inline means. Oct 23 14:25:31 koen|away: http://www.cyrius.com/journal/debian/gcc-4.2-arm Oct 23 14:27:46 mtn: warning: restriction excludes addition of 'packages/linux/linux-omap1-2.6.18-omap1' but includes addition of 'packages/linux/linux-omap1-2.6.18-omap1/another-ide-cs-ids.patch' Oct 23 14:33:25 hrw|work: Hello. I was noticing that openzaurus-3.5.4.2.conf uses the CVS date from openzaurus-3.5.4.conf. Is that on purpose? Oct 23 14:33:55 03hrw 07org.oe.dev * r1e539074... 10/ (5 files in 3 dirs): linux-omap1: added 2.6.18-omap1 fetched via HTTP Oct 23 14:34:04 03hrw 07org.oe.dev * re9cc6897... 10/ (1 packages/linux/linux-omap1.inc): linux-omap1: use OBJCOPY instead of hardcoded binary (so it will work with EABI), use for temporary files Oct 23 14:34:12 hvontres|poodle: 3.5.4.x are compatible as much as possible with 3.5.4 Oct 23 14:34:43 hrw|work: OK. I'll just per-package overides for now then. :) Oct 23 14:43:48 NOTE: multiple providers are available (binutils-cross, sharprom-toolchain-native); Oct 23 14:43:48 NOTE: consider defining PREFERRED_PROVIDER_virtual/arm-linux-binutils Oct 23 14:43:52 mickeyl: another one Oct 23 14:45:22 yeah. i have no immediate solution. perhaps we should just define these providers? or outcomment the PROVIDES in the sharprom-toolchain.bb Oct 23 14:45:33 people who want to build for that would then need to edit it Oct 23 14:45:36 not a big loss if you ask me Oct 23 14:47:00 hi Oct 23 14:49:26 mickeyl: add IAM_USING_NASTY_TOOLCHAIN = "1" into sharprom-compatible.conf and check for it in sharprom-toolchain-native - and only if defined then PROVIDE? Oct 23 14:49:55 worth a try Oct 23 14:50:20 better SHAPROM_COMPATIBLE_TOOLCHAIN = "1" Oct 23 14:50:26 to match our new naming policy Oct 23 14:50:33 SHAPROM? :D Oct 23 14:50:37 ups Oct 23 14:51:50 mickeyl: can you take libqte-dev/libqpe-dev stuff from .oz354x into .dev? Oct 23 14:52:23 i'll take a look, yeah Oct 23 14:56:45 03mickeyl 07org.oe.dev * rbd0a87c6... 10/ (1 packages/libsdl/libsdl-native_1.2.11.bb): libsdl-native 1.2.11 inherit binconfig Oct 23 14:56:47 03mickeyl 07org.oe.dev * r23bb0a8c... 10/ (1 packages/qemu/qemu-native.inc): qemu-native * make it pick up sdl-config-native Oct 23 15:02:19 As far as i can see, the opie-image works as a familar based one. And i noticing that it is way faster than angstrom-gpe-image. Oct 23 15:03:27 But i hade lpotter to change something for me in the OPie source, but it seems like it doesent work anymore. Liek Bluetooth applett, IRDA applet, Backlight slider applet and automatic screen rotation. Oct 23 15:04:32 mickeyl: Even konqueror-embedded 2006 version where build :-) Oct 23 15:04:39 wow Oct 23 15:04:41 goxboxlive: keep in mind lorn is a fulltime paid Opie Core Developer :) Oct 23 15:05:16 mickeyl: I should apply at Oslo, just to sit next to Paul :) Oct 23 15:05:33 what is happening in Oslo zecke? Oct 23 15:06:03 goxboxlive: Trolltech's Office is there Oct 23 15:06:06 i can confirm the weather is grey,and a bit cold.Not long before we have snow here Oct 23 15:06:12 i know Oct 23 15:10:58 zecke: *nod Oct 23 15:11:07 mickey|grocery: hehe Oct 23 15:16:44 OT: anyone knows a good python irc lib? Oct 23 15:17:57 a+ Oct 23 15:27:07 Kristoffer: do you still have issues with keyboard+opie? Oct 23 15:27:23 goxboxlive, Very nice! Yes, OPIE is faster than GPE, that's why it is important to not let it go ;-) Oct 23 15:27:54 yes i agree Oct 23 15:28:21 psokolovsky_: hehe, I have seen this argument (speed) in all directions :) Oct 23 15:29:12 ;-) Oct 23 15:31:45 console is faster than Opie Oct 23 15:32:01 let gui die- won't be missed Oct 23 15:32:43 actually, don't need console, as long as we can login remotely Oct 23 15:33:31 remove the keyboard and screen, and you'd save a *ton* of battery too Oct 23 15:34:53 raduga: remove networking, really save on battery life Oct 23 15:34:55 raduga, doesn't work, everyone already has linux router ;-P Oct 23 15:35:17 raduga: do you own a filewalker device? Oct 23 15:35:21 ade: I/O is overrated anyway Oct 23 15:35:27 zecke: no Oct 23 15:35:30 better do vice-versa, take a router, put X on it, ... Oct 23 15:36:02 raduga: it was a cell phone without keyboard and display Oct 23 15:36:24 i have a few broken cell phones, but none running linux Oct 23 15:36:26 better to work why you need the device, not to work out if the device should run linux Oct 23 15:36:40 raduga: it was a linux phone without keyboard and display Oct 23 15:36:49 ade|desk: the device should run linux because linux is cool Oct 23 15:36:56 ade|desk: that's easy to work out Oct 23 15:37:20 koen: ugh. I'm trolling again :( Oct 23 15:37:25 does my watch need to be linux Oct 23 15:37:28 (sorry) Oct 23 15:37:37 zecke, sorry for the late reply. Yes the patch didnt do any changes. Keymap looks good inside textmode console but doesnt work properly inside Opie. I was thinking that I may need to rebuild more stuff in order for the change to take effect? Oct 23 15:38:14 ade|desk, you have watch? 8-) Oct 23 15:38:19 Kristoffer: QtE is fine, install opie-keyview and see which keys get reported to apps Oct 23 15:38:25 ;-) Oct 23 15:38:48 zecke, specificly we are having issues with pgup/pgdown and arrow keys Oct 23 15:38:57 psokolovsky_: yup , it nearly tells the time most days, sadly not even twice a day is it correct Oct 23 15:39:26 zecke, btw, who (on the OE team) is currently maintaing Opie? Feels like its slowly dying Oct 23 15:39:39 psokolovsky_: i go by approximation mostly Oct 23 15:40:13 Ken|JLime: mickeyl used to maintain, he has send a call for maintainer to the Opie lists Oct 23 15:40:42 Ken|JLime: besides one "it would be sad if it dies" no one volunteered so far Oct 23 15:41:23 I guess i know why the appletts doesnt work. It seems like the changes we aplied to the opie-cvs where after the opie-1.2.2 release. How do i let opie buidl form cvs instead of 1.2.2? Oct 23 15:41:48 goxboxlive: your distro sets preferred versions Oct 23 15:42:01 goxboxlive: modify this file, or better copy it and edit it :) Oct 23 15:42:34 ok, thx again zecke. Oct 23 15:42:50 goxboxlive: there is even one file for Opie which sets the versions :) Oct 23 15:44:09 zecke, pity Oct 23 15:44:30 Ken|JLime: No Oct 23 15:44:42 zecke, how you mean? Oct 23 15:45:10 Ken|JLime: I don't say it is a pity. We might not have managed to liberate/change the world but Oct 23 15:45:45 Ken|JLime: my skills has grown, OpenEmbedded has grown and allowed other Palmtop Environment to get portable to something != iPAQ Oct 23 15:46:58 Ken|JLime: and maybe some days Trolltech recognizes they need a community around Qtopia and will make it grow Oct 23 15:47:39 zecke: IN the angstrom.conf file this is what it saying: OPIE_VERSION = "1.2.2" Should i change it to this? OPIE_VERSION = "CVS" ? Oct 23 15:48:26 goxboxlive: setting it to something 'invalid' will work Oct 23 15:48:35 goxboxlive: as the CVS version is > 1.2.2 (hopefully) Oct 23 15:48:41 ok, thx Oct 23 15:48:53 goxboxlive, = "1.2.2+cvs", then set SRCDATE_*= Oct 23 15:49:22 Ken|JLime: e.g. I'm unscribed from the opie lists for +5 month because I couldn't stand the FUD from the only paid Core Developer Oct 23 15:49:34 libopie2_cvs.bb: PV = "${OPIE_CVS_PV}" Oct 23 15:49:35 goxboxlive: or listen to psokolovsky_ :) Oct 23 15:49:35 hm Oct 23 15:49:47 or not ;) Oct 23 15:50:54 well, I have PREFERRED_VERSION_libopie2 = "1.2.2+cvs-${CVSDATE_libopie2}" in Familiar, and that works as it should Oct 23 15:51:16 Can I use the OZ 3.5.4.1 feeds with this latest hentges release or does it have its own feed? Oct 23 15:51:20 didn't try OE.dev yet. but it's good idea to specify exact and correct version id ;-) Oct 23 15:51:22 zecke, eek what a mess Oct 23 15:51:23 duke: you can Oct 23 15:51:33 good Oct 23 15:51:48 Ken|JLime: yeah, I always felt like responding to his FUD until I recognized it is better to walk away Oct 23 15:52:20 zecke, but One developer cannot get much done? I cant see how it will survive with only one developer and no community support Oct 23 15:53:02 dos anyone here use xmms under 3.5.4.1? does it work (mp3 and ogg)? Oct 23 15:53:25 duke: this is still not a distribution support channel Oct 23 15:53:29 duke: go #openzaurus Oct 23 15:53:35 ok Oct 23 16:00:06 Ken|JLime, I guess, the current aim is to make sure that distros don't drop OPIE support, and at least libopie2 is maintained (i.e. new devices added). If OPIE is available for more and more devices, (new) people interested in it may appear. Oct 23 16:01:05 psokolovsky_, yeah but I think the Qt people should encourage more people when they submit stuff. Myself I've tried to add support for my platform but gotten no reply from #opie or their mailinglist Oct 23 16:01:50 Ken|JLime, QT or OPIE? ;-) If the latter, I can second this, zecke explained why is that ;-( Oct 23 16:01:58 Opie Oct 23 16:02:03 re Oct 23 16:05:57 Ken|JLime: I'm going home now, get some one to install opie-keyview and we can talk Oct 23 16:06:07 I will be awake until 22h (now it is 18:30h) Oct 23 16:08:46 Ken|JLime: who needs opie-keyview info? Oct 23 16:09:03 Jlime users Oct 23 16:09:15 ah Oct 23 16:18:31 yay!!! OE doesn't like my version of monotone Oct 23 16:19:38 chouimat: have you looked at http://www.openembedded.org/snapshots/ ? Oct 23 16:20:14 koen yes I have the 0.30 version of the snapshot I just try to update it and I get a monotone bug Oct 23 16:21:50 hi Oct 23 16:34:02 CIA-4: kick Oct 23 16:43:23 cu Oct 23 17:16:32 03rwhitby 07org.nslu2-linux.dev * rd685ebf6... 10/ (13 files in 4 dirs): Initial import from corrupted database - all history lost. Oct 23 17:16:37 03jbowler 07org.nslu2-linux.bitbake * r3f4bb86f... 10/ (53 files in 14 dirs): revision 287 from svn Oct 23 17:16:41 03rwhitby 07org.nslu2-linux.bitbake * r4fafd70a... 10/ (4 files in 4 dirs): Updated to SVN version 294 Oct 23 17:16:45 03rwhitby 07org.nslu2-linux.bitbake * r8b18357f... 10/ (7 files in 5 dirs): Updated to bitbake 1.3.2 Oct 23 17:16:53 03rwhitby 07org.nslu2-linux.bitbake * ra68af4ca... 10/ (5 files in 4 dirs): Updated to SVN revision 308 Oct 23 17:16:56 03jbowler 07org.nslu2-linux.bitbake * re1a19084... 10/ (4 files in 3 dirs): bitbake 313 Oct 23 17:17:01 03jbowler 07org.nslu2-linux.bitbake * r9f7b9c18... 10/ (16 files in 6 dirs): bitbake: release 330 Oct 23 17:17:07 03jbowler 07org.nslu2-linux.bitbake * rcd0a4d6c... 10/ (1 bin/bitbake lib/bb/fetch/wget.py): bitbake: release 332 Oct 23 17:17:10 03jbowler 07org.nslu2-linux.bitbake * r17970596... 10/ (1 bin/bitbake lib/bb/utils.py): Oct 23 17:17:10 bitbake: patch 332 to handle foo (ver) in RDEPENDS and make the Oct 23 17:17:10 monotone head build (temporary hack from RP) Oct 23 17:17:15 03jbowler 07org.nslu2-linux.bitbake * r334c0d2d... 10/ (1 bin/bitbake): temporary patch version 4 from RP Oct 23 17:17:18 03jbowler 07org.nslu2-linux.bitbake * rea870eec... 10/ (1 bin/bitbake lib/bb/shell.py lib/bb/utils.py): bitbake: bitbake revision 334 Oct 23 17:17:21 03jbowler 07org.nslu2-linux.bitbake * r802347a6... 10/ (1 bin/bitbake): disapproval of revision 'c50bd9ca9065d6f61495b07412f5c7a774eec3aa' Oct 23 17:17:33 03jbowler 07org.nslu2-linux.bitbake * r9aa6f5ec... 10/ (1 bin/bitbake): bitbake revision 336 Oct 23 17:19:08 03rwhitby 07unknown * rd685ebf6... 10/ (13 files in 4 dirs): Initial import from corrupted database - all history lost. Oct 23 17:19:16 03jbowler 07unknown * r3f4bb86f... 10/ (53 files in 14 dirs): revision 287 from svn Oct 23 17:19:21 03rwhitby 07unknown * r4fafd70a... 10/ (4 files in 4 dirs): Updated to SVN version 294 Oct 23 17:19:24 03rwhitby 07unknown * r8b18357f... 10/ (7 files in 5 dirs): Updated to bitbake 1.3.2 Oct 23 17:19:28 03rwhitby 07unknown * ra68af4ca... 10/ (5 files in 4 dirs): Updated to SVN revision 308 Oct 23 17:19:31 03jbowler 07unknown * re1a19084... 10/ (4 files in 3 dirs): bitbake 313 Oct 23 17:21:04 03rwhitby 07org.nslu2-linux.dev * rd685ebf6... 10/ (13 files in 4 dirs): Initial import from corrupted database - all history lost. Oct 23 17:21:16 03jbowler 07org.nslu2-linux.bitbake * r3f4bb86f... 10/ (53 files in 14 dirs): revision 287 from svn Oct 23 17:23:35 03koen 07org.oe.packaged-staging * ra3f4b3bb... 10/ (1 . README): README: add some blurb about the purpose of this branch Oct 23 17:25:59 koen, you working on oe yet? Oct 23 17:27:25 03rwhitby 07unknown * rd685ebf6... 10/ (13 files in 4 dirs): Initial import from corrupted database - all history lost. Oct 23 17:27:30 rwhitby??? Oct 23 17:28:01 mickeyl: no crappy ciabot script Oct 23 17:28:48 phew. ok Oct 23 17:28:53 i was getting worried Oct 23 17:29:27 ciabot picks up every rev in the db, not just the watchlist Oct 23 17:30:23 good evening Oct 23 17:31:39 hey cyrilRomain Oct 23 17:32:06 * koen still wonders where the 'arm:hx4700' is coming from Oct 23 17:32:35 psokolovsky_: I'm subscribed to the $I(@#*&%$(# OE list Oct 23 17:32:43 psokolovsky_: CC:ing is only going to piss me off Oct 23 17:32:58 psokolovsky_: your mail(s) have been deleted unread Oct 23 17:33:41 koen, cool ;-) I've added unconditional rule to not cc: Koen to my AI ;-) Oct 23 17:33:59 koen: compile but not assemble a simple test app Oct 23 17:35:13 koen, but I hope you understand why that was being done: many people receive tons of email via MLs. personal notices are usually of high priority, so I cc: them (and am being cc: by others) ;-) Oct 23 17:35:41 koen: y'know, if your mailer is so smart why don't you have it delete your duplicate messages....? Oct 23 17:36:30 JustinP: I know you are too stupid and/or to lazy to read whom you're sending mail to, so piss off Oct 23 17:38:34 koen: hey :) Oct 23 17:39:44 WOW Oct 23 17:46:25 hi Oct 23 17:48:13 koen: has we already talked before, I worked a bit on a matchbox task (I needed one). Would you add this to angstrom? http://pastebin.ca/217552 Oct 23 17:49:07 nicolasfr: not in that form Oct 23 17:49:27 nicolasfr: and the packages need to be in OE as well Oct 23 17:49:37 nice. I see the initng initiative is picking up steam. They are up to 0.6.4 now. Oct 23 17:50:19 koen: how can I improve / modify / so that it fits ? Oct 23 17:50:43 nicolasfr: remove everything that isn't in OE Oct 23 17:50:51 nicolasfr: removing everything that is already in task-base Oct 23 17:51:16 koen: ok Oct 23 17:51:46 dope. Posted in the wrong channel. sorry. Oct 23 17:52:03 It seems I'm not going to have net access issues afterall :) Oct 23 17:53:32 RP|hel, are you on holiday ? Oct 23 17:53:51 RP|hel, or helsinki ? Oct 23 17:54:04 lrg, helsinki Oct 23 17:54:41 RP: hope the weather is ok and the women and beer cheap ;) Oct 23 17:55:02 lrg, beer is expensive and its raining ;-) Oct 23 17:55:42 lrg, and since I was up at 3:45am, I'll probably have an early night ;-) Oct 23 17:56:25 RP: XorA|gone and I are heading to Linuxworld tomorrow, but are not having to get up at 3.45 :) Oct 23 17:57:38 lrg: Its a shame I'll not be at Linuxworld as its where we first met. Was that really only 1 year ago?! Oct 23 17:57:50 lrg: meeting up with the o-hand dudes at LW? Oct 23 17:58:55 koen: meeting people from Intel Oct 23 17:59:26 koen: and anyone from o-hand if they are going Oct 23 17:59:38 lrg: prepare for more 'wireless ' and massive proof by handwaving ;) Oct 23 17:59:48 heh Oct 23 18:00:32 iwCPU, iwNOR iwRAM iwUSB iwDVI iwSolder Oct 23 18:00:34 etc Oct 23 18:02:01 koen, Not the mention the iwWires ;-) Oct 23 18:02:14 :) Oct 23 18:03:12 and not forgetting iwInside (tm) Oct 23 18:04:13 intel upstreamless pxa-linux.sf.net Oct 23 18:38:42 CCSUFFIX_sh3 = "-3.4.4" oki? Oct 23 18:55:50 mickeyl, I saw some IRC traffic earlier from you regarding the Sharp toolchain building by mistake.. I'm having the same issue were if I don't specifiy the toolchain bitbake builds the sharp version. Oct 23 19:04:40 hi Oct 23 19:09:45 hrw|tv, I suppose I should check the real .18 Oct 23 19:16:48 Crofton: not omap one? Oct 23 19:16:58 the one you committed Oct 23 19:17:01 koen: I reworked the task, using gpe-task as base, but removed GPE specific and install only matchbox core components. Only one thing would need to be imported from Poky (simple script in xserver-nodm-init) to run without dm. http://pastebin.ca/217586 Oct 23 19:17:35 nicolasfr: I don't think angstrom is going to support root logins Oct 23 19:17:38 ah Oct 23 19:18:01 koen, have you thought about my glibc/dropbear problem? Oct 23 19:18:24 Crofton: only thought about it, sorry Oct 23 19:18:44 well let me know I need to refresh your memory :) Oct 23 19:19:03 sounds like you had important business to take care off Oct 23 19:19:21 nicolasfr: there's still duplicate stuff in there Oct 23 19:19:23 so far it has only imapcted dropbear Oct 23 19:19:26 and nano? Oct 23 19:19:43 koen: duplicate? oh Oct 23 19:20:04 nano was my stuff, never mind, remove it Oct 23 19:20:15 NOTE: package python-2.4.3-ml7: task do_build: completed Oct 23 19:20:25 finally on X11-less machines Oct 23 19:20:40 nicolasfr: gtk+ in OE already recommends some pixbug loaders Oct 23 19:22:07 koen: the pixbuf loaders are in gpe-task, I thought they were need here Oct 23 19:22:30 NAiL: can you confirm #1447? Oct 23 19:23:21 koen: in fact one thing I am not sure is weither the dependencies of packages listed in task-* needs to be put here also, or are they added automatically? Oct 23 19:24:17 automatically :) Oct 23 19:25:36 mickeyl Once upon time you explained for me how to get konqueror-embedded to work in Opie, and not like this: http://gnulinux.biz/files/Universal/tmp/screenshots/KonqE.png Oct 23 19:26:07 goxboxlive: you clicked the QMessageBox away? Oct 23 19:26:18 goxboxlive: I think it clearly said you need to set KDEDIR Oct 23 19:26:49 yes i know that, but what about the fontdir as helvetica? I didnt understand how to do that Oct 23 19:27:11 hmh, qtopia + windows media player phone, or something stranger? http://aving.net/usa/news/default.asp?mode=read&c_num=27090&c_code=01&sp_code=35&btb_num=232 Oct 23 19:27:12 goxboxlive: DO NOT CLICK AWAY ERROR MESSAGES AND THINK THEY DON'T HAVE ANY MEANING ;) Oct 23 19:27:37 zecke: i dint, it miss fonts Oct 23 19:28:06 goxboxlive: how did you start it? Oct 23 19:28:20 from console, i dont have the icon Oct 23 19:28:23 "A radioactive cat has eighteen half-lives." Oct 23 19:28:33 suihkulokki: TT did announce qt for wince Oct 23 19:29:34 goxboxlive: and you sourced /etc/init.d/opie? Oct 23 19:29:56 oof Oct 23 19:30:00 what do u mean? Oct 23 19:30:11 suihkulokki: what's your general feeling about gcc 4.2 on arm? Oct 23 19:30:12 goxboxlive: go figure ;) Oct 23 19:30:32 goxboxlive: http://doc.trolltech.com/2.3/winsystem.html learn about QtE Oct 23 19:30:41 zecke: Mickel told me thius: 2.) Edit qt fontdir, copy vera entry and call it "Helvetica" Oct 23 19:30:42 goxboxlive: specially, http://doc.trolltech.com/2.3/running-qws.html Oct 23 19:30:46 koen: it compiles slower than before Oct 23 19:31:10 * suihkulokki looks at the contact list in the first photo Oct 23 19:31:43 goxboxlive: and then look at /etc/init.d/opie and check if any of these important variables... Oct 23 19:31:54 goxboxlive: then take a look at POSIX.3 and /bin/sh and sourcing Oct 23 19:32:19 hmm ok Oct 23 19:32:45 at the end of the day this is a developer channel :) Oct 23 19:32:56 And you should know the technology you use! Oct 23 19:33:14 suihkulokki: I hope gcc 4.2 doesn't inject 'clz' isns when I compile for armv4t Oct 23 19:34:07 goxboxlive: A little test. Which environment variable mst be set to find the FONTDIR? Oct 23 19:34:11 zecke: hmm guess i have to quit before that time then Oct 23 19:34:33 goxboxlive: which init script contains these variables? Oct 23 19:34:43 no idea Oct 23 19:34:53 goxboxlive: and this is one of the other reasons Opie died. From a developer I was turned into a support guy Oct 23 19:35:05 goxboxlive: I have used my limited free time to support... Oct 23 19:35:18 goxboxlive: instead of fixing bugs, creating new bugs :) Oct 23 19:35:28 :) Oct 23 19:37:04 "Up to GCC 4.1, a popular way to specify the target tools used during the build was to create directories named gas, binutils, etc. in the build tree, and create links to the tools from there. This does not work any more when the compiler is bootstrapped. The new configure option --with-target-tools provides a better way to achieve the same effect, and works for all native and cross settings." Oct 23 19:37:09 * koen mumbles Oct 23 19:37:15 zecke: I respect that, but if you look further up i didnt write your name in that qustion. But as mather of fact i am very greatfull for your help i have got here, You have realy helped me alot, but i can take a hint. Oct 23 19:38:19 goxboxlive: sure, but this is a developer channel. And you should have basic UNIX knowledge. e.g. be familiar what 'sourcing' a script means Oct 23 19:38:41 zecke: 'getting filthy rich' Oct 23 19:38:45 goxboxlive: and read the pointers you get, and read again :) Oct 23 19:39:05 goxboxlive: So I have pointed you to the QtE specific site at Trolltech Oct 23 19:39:24 goxboxlive: it clearly says what is needed to run a QtE application (read the whole page before executing it with -qws) Oct 23 19:39:29 koen: yeah I heard about that problem. I don't we've checked yet if the current snapshot is producing clz Oct 23 19:39:31 goxboxlive, then you will also become frustrated with TT Oct 23 19:39:43 wat does clz do in the first place? Oct 23 19:39:53 Count Leading Zeros Oct 23 19:40:00 Crofton: did .18-omap1 boots? Oct 23 19:40:08 not sure, just pulled Oct 23 19:40:32 Crofton: ops. you were talking about it 0.5h ago Oct 23 19:40:43 I keep hoping koen will magically solve my glibc/dropbear problem so I can check both in one go :) Oct 23 19:40:55 I have a cold Oct 23 19:41:04 * koen also has a cold Oct 23 19:41:05 mind is working like molasses now Oct 23 19:41:31 I'm having a bit of trouble trying to build a toolchain right now Oct 23 19:41:52 hrw|tv, I have changed the prefered version so the new kernel goes in the next build Oct 23 19:42:11 zecke: btw: http://ewi546.ewi.utwente.nl/tmp/b0rkage Oct 23 19:42:14 Crofton: same like I did in celinux-test Oct 23 19:42:32 crypto/built-in.o: In function `init': Oct 23 19:42:32 lzf.c:(.init.text+0x6): undefined reference to `crypto_register_alg' Oct 23 19:42:34 ;( Oct 23 19:42:37 How can you be innocent and remorseful at the same time? Oct 23 19:43:20 hehe Oct 23 19:43:55 http://sourceware.org/ml/binutils/2002-01/msg00245.html Oct 23 19:45:25 zecke: right, expect the machine I'm building on is 'bitbake' Oct 23 19:45:35 http://sourceware.org/ml/binutils/2001-05/msg00443.html Oct 23 19:45:59 see OUTPUT_ARCH=${ARCH}:${MACHINE} Oct 23 19:46:47 koen: so indead if you have ${MACHINE} set it will be put into OUTPUT_ARCH Oct 23 19:47:01 heh Oct 23 19:47:02 koen: so unset $MACHINE in do_configue and do_compile Oct 23 19:47:51 koen: we need to dump the Postgres Database again Oct 23 19:48:00 koen: I have a new addition to tinderbox I'm going to finish Oct 23 19:51:50 goxboxlive: progress? Oct 23 19:51:58 zecke: Just wonder in something. If i have some qustions about OE, like the one today with failing building opie-image. Where should i ask or reposrt about bugs like that? I mean i have reported a lot to get the HTC Universal in to OE, and i mean, it couldnt have been there without all my qustions. Oct 23 19:52:23 I agree about the konqueror bug, it belongs in the Opie channel. Oct 23 19:53:19 goxboxlive: no, the place is probably right Oct 23 19:53:31 goxboxlive: just some basic knowledge is missing :) Oct 23 19:53:57 that's true. Oct 23 19:54:11 zecke, got a quick question about your kernel->opie keymap patch Oct 23 19:54:17 Kristoffer: sure Oct 23 19:54:26 1) should loadkeys change keymap in opie (after starting opie) Oct 23 19:54:36 Kristoffer: no! Oct 23 19:54:43 Kristoffer: load the keymap first Oct 23 19:54:46 Kristoffer: then start Opie Oct 23 19:55:01 zecke, oki. Currently that has no effect for us. I've recompiled qte and libopie Oct 23 19:55:24 Kristoffer: do me a favor, use opie-keyview before talking any further :) Oct 23 19:55:31 :D Roger that Oct 23 19:55:46 Kristoffer: it is no use talking if you don't have that running Oct 23 19:56:03 zecke, yeah understod as much. Just as always looking for the quick way out Oct 23 19:56:23 Kristoffer: Quick way == install opie-keyview Oct 23 19:56:31 re Oct 23 19:56:32 :D Oct 23 19:56:35 Kristoffer: press the keys and check if they report 0xffff (which would mean unmapped) Oct 23 19:56:44 Kristoffer: can't be any quicker Oct 23 19:57:10 zecke, aha. In which case it would mean that opie keymap hasnt mapped that key? Oct 23 19:58:48 Kristoffer: right Oct 23 19:59:42 0xffff -> unmapped the reason for that can have two reasons. 1.) this key is unknown to QtE 2.) it is not mapped in the kernel Oct 23 20:00:11 zecke, I know that the keymap when loaded in console (textmode) works fine Oct 23 20:00:15 so shouldnt be a kernel issue Oct 23 20:00:31 Kristoffer: no, but the kernel keymaps forgives more than I do :) Oct 23 20:00:40 :D Oct 23 20:00:57 Oki, will let you know once keyview is installed and ready Oct 23 20:00:59 I'm Mr. Pain/Punisher and punish bad keymaps Oct 23 20:01:12 Kristoffer: I hope it is in OE :} Oct 23 20:01:30 Kristoffer: Did you try to remove patches in qte-mt like i told you. That fixed the same problem on the HTC Universal. Oct 23 20:01:52 'night all Oct 23 20:01:54 goxboxlive, no I didnt. What patches was that? Oct 23 20:02:21 hold on and i'll pastebin my monotone diff. Oct 23 20:02:31 goxboxlive: send patches :) Oct 23 20:04:11 zecke: I did, but Mickeyl told me that qte-common_2.3.10.inc where to custom allready. Oct 23 20:07:24 Kristoffer: http://paste.uni.cc/11036 Oct 23 20:08:07 try it and rebuild qte (or qte-mt) Oct 23 20:08:40 goxboxlive, hmmm Oct 23 20:08:40 yeah removing/not applyign these zaurus patches :) Oct 23 20:08:43 Kristoffer: and it would be interesting to see which keyboard driver is loaded/used Oct 23 20:09:10 zecke, does keyview provide that? Oct 23 20:09:47 Kristoffer: no Oct 23 20:09:59 zecke: I tried to remove one by one, it took me ages to figure out wich ones i needed to remove. But it worked at the end. Oct 23 20:12:25 zecke, Up Arrow: : unicode 0x38 Oct 23 20:12:38 modifiers 0x4000 Oct 23 20:13:01 modifiers are set? Oct 23 20:13:10 apperently Oct 23 20:13:19 Kristoffer: did you press some? Oct 23 20:13:55 zecke, one of developers is trying it out Oct 23 20:13:58 I'll get him in here Oct 23 20:14:05 Good Morning Oct 23 20:14:23 B_Lizzard, what does arrow right give? Oct 23 20:14:34 Kristoffer: hehe I said I will in bed at 22h ;) Oct 23 20:14:48 0x34 for unicode and keycode Oct 23 20:14:51 zecke, :D Sorry man. Just that you are the guru Oct 23 20:14:58 goxboxlive: summary for QtE. Source ". /etc/init.d/opie" :) Oct 23 20:14:59 0x4000 for modifiers Oct 23 20:15:12 B_Lizzard: pressing 'a' gives modifiers as well? Oct 23 20:15:26 isPress and autoRepeat are at 0x0, for obvious reasons Oct 23 20:15:29 wait... Oct 23 20:15:39 a gives 0x0 at modifiers Oct 23 20:15:52 B_Lizzard: interesting Oct 23 20:16:38 zecke, We could collect a couple of data and post a bug on openembedded? Oct 23 20:17:01 many non-supported keys give off a 0x4000 for "modifiers" Oct 23 20:17:11 B_Lizzard: tomorrow evening is bad, can you be around on wednesday? Oct 23 20:17:25 B_Lizzard: I would have to look into the kernel keymap driver first :) Oct 23 20:17:29 I'm pretty sure I can Oct 23 20:17:35 Evening? Oct 23 20:17:44 B_Lizzard: I'm in Berlin/Germany and currently it is 22:36 Oct 23 20:17:55 zecke: I'm dumping the db right now, should I mail it to you? Oct 23 20:17:59 one hour behind me Oct 23 20:18:00 koen: no no Oct 23 20:18:06 koen: stop, wait :) Oct 23 20:18:27 B_Lizzard: I assume you can compile QtE? Oct 23 20:18:33 yes Oct 23 20:18:36 libqte? Oct 23 20:18:36 * koen stops & waits Oct 23 20:18:40 B_Lizzard: yes Oct 23 20:18:44 aye Oct 23 20:18:46 koen: just don't care fow now Oct 23 20:18:49 I can compile anything Oct 23 20:18:53 :D Oct 23 20:18:59 Unless it doesn't :0 Oct 23 20:18:59 koen: soon #123 will link to bug tracker Oct 23 20:19:08 zecke: cool Oct 23 20:19:24 zecke: I want bonsai to work so we can start getting rich with tarballs Oct 23 20:19:37 zecke, We will be here. Big thx for taking some ZzzZ time to help us out Oct 23 20:19:42 koen: rich tarballs? Oct 23 20:19:54 B_Lizzard: do a quick check Oct 23 20:20:00 yes? Oct 23 20:20:10 B_Lizzard: ls -la /proc/`pidof qpe`/ Oct 23 20:20:21 B_Lizzard: and check if /dev/tty0 is opened Oct 23 20:21:04 B_Lizzard: probably useless test but still see if qpe opened /dev/tty0 Oct 23 20:21:21 B_Lizzard: ls -la /proc/`pidof qpe`/fd/ Oct 23 20:21:22 I see no such thing Oct 23 20:21:28 ah... Oct 23 20:21:38 B_Lizzard: needed to ssh into a linux machine :) Oct 23 20:21:46 :) Oct 23 20:22:11 doing that last command gave me an output of about 10000 lines Oct 23 20:22:15 grep it? Oct 23 20:22:57 "ls -la /proc/`pidof qpe`/fd/ | grep tty" gives me a line Oct 23 20:23:21 B_Lizzard: okay, a keyboard is opened :) Oct 23 20:23:24 but we knew that Oct 23 20:23:31 it's a file named 6 which is a link to /dev/tty0 Oct 23 20:23:36 hehehe Oct 23 20:23:52 B_Lizzard: QWSTtyKeyboardHandler src/kernel/qkeyboard_qws.cpp and add a printf into the c'tor Oct 23 20:24:15 The whati whatten? Oct 23 20:24:36 What? Oct 23 20:24:37 B_Lizzard: yes, check if it gets loaded :) Oct 23 20:24:52 zecke, he's no programmer :D so you will have to explain abit easier Oct 23 20:25:04 I didn't understand a thing of what you're saying :D Oct 23 20:25:06 B_Lizzard: ah okay, then let us postpone it to wednesday :) Oct 23 20:25:10 hehehe Oct 23 20:25:22 Ok Oct 23 20:26:54 Gratias Ago, zecke Oct 23 20:28:36 koen: check out the admin interface of django Oct 23 20:28:39 koen: really awesome Oct 23 20:35:52 hmm that dropbear issue is really odd Oct 23 20:36:03 who was working on debugging that? Oct 23 20:36:04 yep Oct 23 20:36:11 I found it Oct 23 20:36:18 I am not sure how to proceed Oct 23 20:36:40 it would appear glibc has a check that a structure is laid out in memory properly Oct 23 20:37:13 I suspect the compiler knows the answer at compile time Oct 23 20:38:21 can you diff against what glibc-2.4 does in the same place? Oct 23 20:38:40 hmmm Oct 23 20:38:46 let me see Oct 23 20:39:19 /proc/cpu/alignment ? Oct 23 20:39:37 I am unzipped 2.5 no Oct 23 20:39:40 er now Oct 23 20:40:33 mickeyl: Is you guitar running an OE-built distro yet? Oct 23 20:40:54 koen: nah, i'm an old school guy when it comes to acoustic instruments. Oct 23 20:41:08 koen: actually i'd love to work on a project for a GUI on a synthesizer Oct 23 20:41:12 first glance, 2.4 does not have assert .... Oct 23 20:41:21 koen: zecke promised to get me in touch with some guys Oct 23 20:41:30 for diff, do you do old first? Oct 23 20:41:34 Crofton|home: yea Oct 23 20:41:35 mickeyl: have you seen jokosher? Oct 23 20:41:44 koen: no, who/what's that? Oct 23 20:42:03 mickeyl: http://www.jokosher.org/ Oct 23 20:42:04 -u or -c prefered? Oct 23 20:42:11 mickeyl: python + gstreamer Oct 23 20:42:40 mickeyl: not a synth, but cute nonetheless Oct 23 20:42:57 Croftono: unfied Oct 23 20:43:01 unified, even Oct 23 20:43:22 Crofton: diff -Nurd, easy to remember :) Oct 23 20:43:23 i recommend Oct 23 20:43:24 alias diff='diff -udbBPr' Oct 23 20:43:36 Nurd works also, yeah Oct 23 20:44:20 i find bBP pretty helpful as well Oct 23 20:44:43 http://pastebin.ca/217825 Oct 23 20:44:52 2.5 adds the check Oct 23 20:45:05 and the padding Oct 23 20:45:11 zecke: still there? Oct 23 20:45:17 make_request is the problem area Oct 23 20:45:31 the assert fails Oct 23 20:45:33 ok, can you please do two things Oct 23 20:45:42 1.) patch out the assert and check whether that helps Oct 23 20:45:53 2.) write a mail to dropbear author? Oct 23 20:46:23 I can patch the assert Oct 23 20:46:36 I am not sure what to say to the dropbear author Oct 23 20:46:42 the truth :) Oct 23 20:46:51 it appears like the problem is in the glibc2.5 alignment Oct 23 20:46:53 heh Oct 23 20:47:01 that I can do Oct 23 20:47:15 i'm not so sure about that. sending him that diff + a problem description might get him to revisit his code Oct 23 20:47:31 but it looks like all the things that effect the assert are known at glibc compile time Oct 23 20:47:46 remember, the code that fails is in glibc Oct 23 20:48:47 OE-developers: I understand that you get tired of n00b qustions from us that is not developers. Why cant we make a #oe-newbie channel or something, so that this channel would get flooded by rookie qustions? Oct 23 20:49:17 goxboxlive: because only newbies would be in that channel Oct 23 20:50:35 koen: so wouldnt be there from time to time to help out? Oct 23 20:50:50 so you Oct 23 20:51:54 the problem isn't really with newbie questions, the problem is that *users* (not developers) are picking up OE and get mightily confused when it errors out Oct 23 20:52:09 in a sense OE makes crosscompiling too easy Oct 23 20:52:53 Crofton: i understand that, but it could be that the glibc check is correct and that dropbear does something that works by luck, but isn't supposed to Oct 23 20:52:57 *shrug* Oct 23 20:53:03 koen: Yes that's true, it's realy easy to cross compile with this tool Oct 23 20:58:36 zecke: you're 78 minutes late Oct 23 20:58:43 zecke: let me lead by example Oct 23 20:58:47 'night all Oct 23 20:59:33 hehe Oct 23 21:07:34 mickeyl, here is the line from fropbear Oct 23 21:07:34 dropbear: ../sysdeps/unix/sysv/linux/check_pf.c:68: make_request: Assertion `sizeof (req) - ((size_t) &((struct req *)0)->pad) == 3' failed. Oct 23 21:08:45 I think all those arguments are known at compile time Oct 23 21:08:58 hmm Oct 23 21:08:58 I am working on patching out the assert Oct 23 21:09:16 and I will send him the email Oct 23 21:09:20 that's really a bit odd Oct 23 21:09:26 i wonder whether it's a name clash or something Oct 23 21:09:31 char pad[3] Oct 23 21:09:38 it's there Oct 23 21:09:38 how could that differ!? Oct 23 21:10:12 do we supply any odd alignment requests to gcc for glibc compilation? Oct 23 21:11:09 http://pastebin.ca/217876 Oct 23 21:11:16 there is the section of code Oct 23 21:12:29 it seems like it claculates the sizeof the overal structure, then subtracts the offset of the pad array and check that the results is three Oct 23 21:12:47 hmm Oct 23 21:12:53 sounds like a job for a couple of printf's Oct 23 21:13:02 can't see why this should be going wrong :/ Oct 23 21:13:42 let me patch in some printfs Oct 23 21:17:48 printf("sizeof req = %d\n", sizeof(req)); Oct 23 21:17:56 um, I have been in c++ too long ... Oct 23 21:18:04 is that the correct printf syntax? Oct 23 21:20:11 Crofton|oe: sizeof is unsigned at least Oct 23 21:20:22 Crofton|oe: unsigned long long or size_t :) Oct 23 21:21:10 %d will print any int? Oct 23 21:24:14 ya Oct 23 21:27:08 where do I put patches for glibc-2.5? Oct 23 21:28:58 mkdir glibc-2.5 Oct 23 21:29:15 I think the bb file gets the patched from glibc-2.4 Oct 23 21:29:26 FILESDIR = "${@os.path.dirname(bb.data.getVar('FILE',d,1))}/glibc-2.4" Oct 23 21:29:39 well, either put it into 2.4 dir or Oct 23 21:29:48 add glibc-2.5 as a valid FILESDIR entry Oct 23 21:29:50 I put it in 2.4 Oct 23 21:29:56 fair enough Oct 23 21:30:15 I am printing offset and sizes of about everything in the struct .... Oct 23 21:34:43 I am going to wipe tmp and rebuild Oct 23 21:34:48 I'll chek in my AM Oct 23 22:35:41 * v8jlene is back Oct 23 22:52:26 OK, glibc-2.5 compiling with debug patch for check_pf in place, thanks for the suggestion mickeyl Oct 23 22:52:39 cool Oct 23 22:53:34 I won't get around to booting until tomorrow afternoon Oct 23 22:53:44 I may lay on the couch and watch TV tomorrow morning Oct 23 22:53:53 unless I am feeling better then Oct 23 22:55:04 leaving Oct 23 22:55:10 oops **** ENDING LOGGING AT Tue Oct 24 02:59:57 2006