**** BEGIN LOGGING AT Wed Feb 07 02:59:57 2007 Feb 07 07:24:23 goooooooood morning all! Feb 07 07:25:29 hello Feb 07 07:25:54 Eblis: I'll test you mono recipes when I get back from the lectures Feb 07 07:27:14 ok, thanks Feb 07 07:27:22 there's also a portable .net recipe :) Feb 07 07:51:26 morning Feb 07 07:52:58 03eblis102 07org.oe.dev * r6e3df1ac... 10/ (6 files in 4 dirs): Feb 07 07:52:58 mono: add 1.2.2.1 Feb 07 07:52:58 * only works on arm Feb 07 07:52:58 * closes #1853 Feb 07 07:53:06 Eblis: thx Feb 07 07:53:25 Eblis: can you provide informations how to make it working on !arm targets? Feb 07 07:53:38 hrw: i didn't try Feb 07 07:53:43 you have to have the hardware first Feb 07 07:53:47 it's a bit tricky Feb 07 07:53:55 once you have the hardware you have to build mon on it Feb 07 07:53:57 Eblis: I have arm, mipsel, x86, x86_64here Feb 07 07:54:18 then you have to build the cpu info for that target, this is the part that actually needs to run on the target platform Feb 07 07:54:19 also there is qemu... Feb 07 07:54:27 then you can get the cpu-X.h file and cross-compile Feb 07 07:55:10 hrw: short story: build 'natively', put cpu-$arch.h into OE Feb 07 07:55:29 yep, that's about it Feb 07 07:55:29 sure Feb 07 07:55:37 if you look at the recipe you should have enough clues where to find that .h :) Feb 07 07:55:49 debian builds this version for arm, x86, x86-64, ppc, ia64 Feb 07 07:56:07 portable .net should be more portable although 0.7.4 is an old version and the dotgnu guys said they're working to make it even more portable Feb 07 07:56:13 and a new stable release is due out soon Feb 07 07:56:14 koen: did you removed mono from REMOVALS? Feb 07 07:56:33 mono was removed already, right? Feb 07 07:56:38 * hrw want battery API in kernel Feb 07 07:56:47 koen: ah.. right Feb 07 07:57:07 Hi Rod, Paul, Paul Feb 07 07:57:29 hey hrw Feb 07 07:57:38 hi hrw Feb 07 07:58:17 root@hx4700:/var/lib/ipkg# ipkg install mono Feb 07 07:58:18 Installing mono (1.2.2.1-r0) to root... Feb 07 07:58:18 Downloading http://www.angstrom-distribution.org/unstable/feed/armv5te/base/mono_1.2.2.1-r0_armv5te.ipk Feb 07 08:01:13 good morning! Feb 07 08:01:19 <_law_> koen, is there already a .bb for it? Feb 07 08:01:33 _law_: yes, scroll up Feb 07 08:02:57 koen: you're downloading to the platform you created the cpu-arm.h from, right ? :) Feb 07 08:03:10 yes Feb 07 08:03:35 aha Feb 07 08:03:56 Eblis: it doesn't have /usr/lib/mono/1.0/mscorlib.dll, is that correct? Feb 07 08:04:17 i'm away from the computer i built on, i can't check :( Feb 07 08:07:37 Hi! Feb 07 08:10:18 koen: mono-native is buildable only on ARM :( Feb 07 08:10:34 ah, hmm Feb 07 08:10:36 DEBUG: No providers of build target mono-native (for ['/home/hrw/devel/oe/org.openembedded/packages/mono/mono_1.2.2.1.bb']) Feb 07 08:10:48 DEBUG: skipping /home/hrw/devel/oe/org.openembedded/packages/mono/mono-native_1.2.2.1.bb Feb 07 08:10:59 COMPATIBLE_HOST = "*" on the -native .bb? Feb 07 08:11:12 or mono.inc Feb 07 08:11:19 I will handle it Feb 07 08:12:14 ooops ... then pnet will need that too :D ... sorry :( Feb 07 08:12:41 Eblis: wanna send a mail to openmoko-devel that mono is sort of working? Feb 07 08:12:51 yep Feb 07 08:12:54 :D Feb 07 08:13:31 i'll mention portable .net as well, if that's ok Feb 07 08:13:47 Eblis: http://rafb.net/p/zSOOzb75.html Feb 07 08:15:59 koen: pushed fix Feb 07 08:16:01 hehe Feb 07 08:16:20 I do not like EXTRA_OECONF vars in mono.inc Feb 07 08:16:24 uuh, just saw th error Feb 07 08:16:38 03hrw 07org.oe.dev * r762e8a29... 10/ (4 files in 2 dirs): mono: make mono-native buildable not only on ARM Feb 07 08:16:41 arm one override normal one instead of adding... Feb 07 08:17:34 koen: iirc msc should not have been built Feb 07 08:17:48 maybe you have the binary from your native build ? Feb 07 08:18:39 03koen 07org.oe.dev * r4b9a0b16... 10/ (1 packages/mono/mono_1.2.2.1.bb): mono: also package dlls needed for mcs Feb 07 08:21:28 koen: 1. use install instead of cp 2. nothing depends on mono-dll Feb 07 08:22:20 hrw: right, it isn't needed for mono, just for mcs Feb 07 08:22:28 and mcs will be split out later :) Feb 07 08:22:40 k Feb 07 08:24:27 ok. back to home work Feb 07 08:25:55 morning Feb 07 08:26:01 hey XorA Feb 07 08:27:02 koen: whats new this morning then Feb 07 08:27:30 XorA: a mono in OE that seems to be working, thanks to Eblis Feb 07 08:27:31 hey ssvb Feb 07 08:27:37 ssvb: good morning Feb 07 08:28:53 * rwhitby finally gets arm-kernel-shim to work, enabling arm kernels everywhere to be free of brain-dead bootloaders with bogus ATAGs. Feb 07 08:29:47 ah, cool Feb 07 08:30:13 I have a redboot script that pokes at various registers to do that on my ep93xx board Feb 07 08:30:56 just need one more round of clean-up with Marc Singer before I commit the changes. There's a hard-coded 'place-the-boot-section-after-the-rodata-cmdline' hack which I need to work out how to do in the ld script instead of with a hard-coded value on the ld command line Feb 07 08:31:46 (the shim has .text, then .rodata (the ATAG values and cmdline which are copied to the bootparam area), then a NOP before the kernel proper starts. Feb 07 08:32:08 it can even do the force-endian stuff too (removing the need for the devio magic in the build scripts) Feb 07 08:32:36 rwhitby: soounds pretty neat Feb 07 08:33:32 Marc Singer wrote it, I'm just integrating it. Feb 07 08:33:41 (and finding the bugs :-) Feb 07 08:38:48 APIC error on CPU0: 40(40) Feb 07 08:39:01 need to finally find out how to solve it Feb 07 08:41:56 NOTE: package mono-1.2.2.1-r1: task do_compile: started Feb 07 08:44:27 XorA, koen: good morning Feb 07 08:50:00 morning all ! Feb 07 08:50:20 hey ifaistos :) Feb 07 08:50:51 ~seen CoreDump|home Feb 07 08:51:15 coredump|home is currently on #gpe #oe #openmoko #openzaurus, last said: '~pester CoreDump|home to fix bug xyz'. Feb 07 08:51:17 CoreDump|home: seems to be MIA Feb 07 08:51:34 koen : Tried to build angstrom-gpe-image for efika last night. it fails at rxvt-unicode-7.9 Feb 07 08:51:49 Ifaistos: define fails Feb 07 08:51:56 yeah - I'm looking to port altboot to ixp4xx/nslu2, but need some assistance from CoreDump|home Feb 07 08:53:33 hrw: -> http://pastebin.ca/343825 Feb 07 08:54:16 Ifaistos: looks like no one built it before for ppc Feb 07 08:54:56 hrw: Are you sure ? i think koen did Feb 07 08:55:54 hrw: maybe the problem is here ?->ERROR: Nothing provides dependency rxvtlib Feb 07 08:57:32 hrw: this is what i get when i do a bitbake angtrom-gpe-image rxvt-unicode -c clean and try to rebuild Feb 07 08:57:53 hrw: btw how is 2.6.20 running for you? Feb 07 08:59:07 XorA: on desktop? nice Feb 07 08:59:53 hrw: not tried on arm yet? Feb 07 09:00:01 not Feb 07 09:00:36 hrw: I think i found something -> http://64.233.183.104/search?q=cache:Pdc3fIxPqZQJ:www.handhelds.org/hypermail/familiar-dev/current/1054.html+rxvtlib.h+error:+%23error+set+intp_t&hl=el&ct=clnk&cd=1&gl=gr&client=firefox-a Feb 07 09:08:47 hrw: with the change mentioned rxvt-unicode compiles Feb 07 09:12:34 how do we build the rest of the packages availiable for angstrom ? bootstrap-image and gpe-image build ok now Feb 07 09:31:02 03koen 07org.oe.dev * r8cd07540... 10/ (1 classes/gpephone.bbclass): gpephone.bbclass: start merging gpephone bits from http://linuxtogo.org/~florian/gpepe/ Feb 07 09:31:07 03koen 07org.oe.dev * rc2cdf755... 10/ (55 files in 8 dirs): Feb 07 09:31:07 gpephone: add recipes from http://linuxtogo.org/~florian/gpepe/ Feb 07 09:31:07 * They still need to be run through oe-stylize.py Feb 07 09:31:07 * things like linphone need to get merged with their original counterparts in OE Feb 07 09:37:12 03koen 07org.oe.dev * r4c3e3934... 10/ (1 conf/bitbake.conf): bitbake.conf: add GPEPHONEMIRROR Feb 07 09:38:33 ~seen krofton Feb 07 09:38:39 hmm - wonder why florian didn't respond to my linphone RFC last week ... Feb 07 09:38:47 ~seen crofton Feb 07 09:38:47 NDA? Feb 07 09:39:12 i haven't seen 'krofton', Ifaistos Feb 07 09:39:12 let's give him the benefit of the doubt :-) Feb 07 09:39:17 crofton is currently on #ol (7h 47m 55s) #oe (7h 47m 55s) #edev (7h 47m 55s) #elinux (7h 47m 55s), last said: 'good night'. Feb 07 09:42:08 03koen 07org.oe.dev * rf8e7501b... 10/ (6 files in 2 dirs): gpe-phone: fix some bugs Feb 07 09:58:39 03koen 07org.oe.dev * r477b5417... 10/ (3 files in 2 dirs): e2fsprogs 1.38: stage .pc files Feb 07 09:59:31 What happened to libqte-mt? What is the replacement for it? Feb 07 09:59:40 gtk+ Feb 07 10:01:53 koen: You are kidding. Feb 07 10:02:20 I thought opie was going to survive? Feb 07 10:03:02 psokolovsky_: What happened? opie dead? Feb 07 10:04:11 Laibsch, wazzup? Feb 07 10:04:56 koen: fancy making e2* alternatives now that busybox provides them? Feb 07 10:04:58 koen just suggested gtk+ as a replacement for libqte-mt which I need for one of my packages. I can only see that as a sign that opie has completely been removed from OE now? Feb 07 10:05:18 psokolovsky_: Is opie still alive in OE? Feb 07 10:05:27 Laibsch: opis is still there Feb 07 10:05:42 Laibsch, start with facts please. you need libqte-mt - use it. otherwise - wazzup? Feb 07 10:06:10 Laibsch, http://www.linuxtogo.org/gowiki/OpieWithAngstrom Feb 07 10:06:57 psokolovsky_: can you resolve opie-mediaplayer1 circular deps? Feb 07 10:07:15 hrw, that's in q Feb 07 10:07:16 ok Feb 07 10:07:28 psokolovsky_: OE tells me "ERROR: Nothing provides runtime dependency libqte-mt". This was previously in. So I inquired as to what the replacement was and koen answered gtk+ which I can only interpret as "opie is gone". Feb 07 10:07:29 hrw, do you have some hints/refs? Feb 07 10:07:46 Laibsch, ah, ok. because it's qte-mt Feb 07 10:08:05 psokolovsky_: So I ask again, if libqte-mt is no longer available what should I use instead? Feb 07 10:08:09 psokolovsky_: probably plugins should not depend on player itself Feb 07 10:08:10 Laibsch: you have to work out that Koen is out personal rabid GNU gtk+ advocate :-) Feb 07 10:08:16 psokolovsky_: OK, thanks for the info. Feb 07 10:08:23 s/ot/our/ Feb 07 10:08:23 fsck Feb 07 10:08:24 03koen 07org.oe.dev * rc7e3615f... 10/ (4 files in 2 dirs): gpephone: more bugfixes Feb 07 10:08:29 03koen 07org.oe.dev * r8ed317e6... 10/ (1 packages/gpephone/libim_0.1.bb): libim: disable parallel_make Feb 07 10:08:30 ~slap koen ;-) Feb 07 10:08:40 * ibot slaps koen ;-), keep your grubby fingers to yourself! Feb 07 10:09:47 ooooh, bitch fight :-) Feb 07 10:15:29 03koen 07org.oe.dev * r8643e683... 10/ (20 files in 2 dirs): gpe-phone: make everything fetch Feb 07 10:23:59 ~ping Crofton Feb 07 10:24:10 pong Crofton Feb 07 10:28:54 Ifaistos: Crofton is in US so he is sleeping now rather Feb 07 10:29:41 hrw: thanks. though he was in Europe also Feb 07 10:39:28 03koen 07org.oe.dev * rbe5bce99... 10/ (1 packages/linphone/linphone_1.6.0.bb): linphone 1.6.0: clean up recipe and mark hacks when found Feb 07 10:40:15 hi liam Feb 07 10:52:02 hey Marcin Feb 07 10:52:13 03xora 07org.oe.dev * r18cf39ef... 10/ (1 packages/linux/linux-rp_2.6.20.bb): linux-rp_2.6.20.bb : update to latest ASoC release Feb 07 11:00:33 koen: lakml has another post 'I have problem with eabi' - can you answer as OE/EABI master? Feb 07 11:10:16 is there a way to use version numbers in RDEPENDS in order to specify minumum/maximum run time dependency versions with oe/bitbake? Feb 07 11:12:49 pong Feb 07 11:12:59 hi Crofton Feb 07 11:13:36 gm Feb 07 11:13:43 I get up at a reasonable hour :) Feb 07 11:15:06 :) Feb 07 11:15:44 If I kept the hours some of you keep, we could never chat Feb 07 11:16:16 think about relocating to Europe ? Feb 07 11:16:24 heh Feb 07 11:16:27 not really Feb 07 11:16:46 I am originally from England Feb 07 11:17:00 came to US when I was 5 Feb 07 11:17:40 Does England count as part of Europe ? A lot of people would disagree with that :) Feb 07 11:17:46 Heh Feb 07 11:17:50 Even the English Feb 07 11:17:57 Ifaistos: only those who read Rupert Murdochs newspaper Feb 07 11:18:09 Let's just say, I am orginailly from a place closer to Europe than I am now Feb 07 11:18:11 koen: re: linphone. I see your pkg-config comment. Any hints which I can pass on to the author of the recipe? Feb 07 11:18:25 anyway Ifaistos you rang? Feb 07 11:18:37 anyway when you get you morning coffee i need some help with getting the efika going Feb 07 11:18:51 the kernel build with angstrom just reboots Feb 07 11:18:53 ok Feb 07 11:18:56 hmmm Feb 07 11:18:58 weird Feb 07 11:19:10 good point, coffee not ready yet Feb 07 11:19:40 rwhitby: easiest way is to see why it insist on a recent pkgconfig Feb 07 11:20:15 what gets used if that line is removed? Feb 07 11:22:07 the pkgconfig OE builds Feb 07 11:22:20 which is 0.15 + 0.19 m4 files Feb 07 11:22:48 so if configure passes it should be ok? Feb 07 11:23:05 yes Feb 07 11:24:02 ok, it fails when looking for SPEEX. Maybe I can set speex flags directly so as not to require pkgconfig. Feb 07 11:25:28 are lines like "export SPEEX_LIBS="-L${STAGING_LIBDIR} -lspeex" Feb 07 11:25:28 " acceptable alternatives? Feb 07 11:25:48 yes Feb 07 11:26:23 * koen heads to the uni Feb 07 11:29:36 Ifaistos, how are you trying to boot the kernel, via net or via disk/cf?? Feb 07 11:35:44 heh.. upgrading db to 0.32 takes time Feb 07 11:35:53 Crofton : usb flash stick Feb 07 11:37:29 hmm Feb 07 11:37:34 I haven't tried that :) Feb 07 11:37:36 03pfalcon 07org.oe.dev * r8b1c98ab... 10/ (1 packages/linux/linux-handhelds-2.6_cvs.bb): linux-handhelds-2.6 cvs: Bump to the latest release, 2.6.19-hh10. Feb 07 11:38:23 can you try booting by loading kernel via tftp? Feb 07 11:38:42 Crofton : this is what i am setting up as we speak :) Feb 07 11:38:47 ah Feb 07 11:38:50 good plan :) Feb 07 11:39:37 Crofton : i tried also to boot the bplan kernel i get the same output so probably is something i am not doing right Feb 07 11:39:44 I've been busy with other stuff, going to replave my hardrive with a micro drive next Feb 07 11:39:46 yeah Feb 07 11:40:00 it is pretty easy, once you get the hang of it Feb 07 11:43:24 Hmm - linphone configure fails normally, but works if export PKG_CONFIG=${STAGING_BINDIR_NATIVE}/pkg-config Feb 07 11:43:24 is used. Feb 07 11:43:51 why wouldn't that value be automatically passed to configure when inherit pkgconfig is used? Feb 07 12:08:12 re Feb 07 12:11:11 03pfalcon 07org.oe.dev * r53df1a1e... 10/ (1 packages/linux/linux-handhelds-2.6_2.6.19-hh9.bb): (log message trimmed) Feb 07 12:11:11 linux-handhelds-2.6 2.6.19-hh9: Add new release. Feb 07 12:11:11 * Changes since -hh8: Feb 07 12:11:11 1. Tree cleanup, get rid of unused arch/arm/plat-ipaq/, etc. Feb 07 12:11:11 2. Added PXA serial hooks setters. Feb 07 12:11:13 3. Revamped asus620 port. Feb 07 12:11:15 4. Cleanup of rx3000 port. Feb 07 12:11:17 03pfalcon 07org.oe.dev * rdb5f6796... 10/ (1 packages/linux/linux-handhelds-2.6_2.6.19-hh10.bb): (log message trimmed) Feb 07 12:11:22 linux-handhelds-2.6 2.6.19-hh10: Add new release. Feb 07 12:11:24 * Changes since -hh9: Feb 07 12:11:26 1. Added aximx50 port. Feb 07 12:11:28 2. IR and battery classdev support for h4000. Feb 07 12:11:30 3. Battery classdev support for htcapache. Feb 07 12:11:32 4. Generic TS drivers for ads7846 & ad7877 chips (used for h4000 & htcapache). Feb 07 12:11:36 03pfalcon 07org.oe.dev * rbbfe9271... 10/ (1 packages/linux/linux-handhelds-2.6_2.6.19-hh7.bb): linux-handhelds-2.6 2.6.19-hh7: Drop old version. Feb 07 12:14:01 Crofton : manage to boot the kernel from the usb stick also. it seems i had set the wrong root device and not used console=ttyPSC0 to see it. so it was rebooting after 180 secs Feb 07 12:14:13 ah Feb 07 12:14:48 ~lart me Feb 07 12:14:48 * ibot DoSes hrw Feb 07 12:15:11 that rebooting after 180 secs can be pretty confusing at first :) Feb 07 12:15:24 Crofton : Are you running angstrom on the board with keyboard,screen etc or just booting in console mode Feb 07 12:15:29 psokolovsky__: do we have aximx50 machine config? Feb 07 12:15:45 I have a usb hish speed card in my PCI slot Feb 07 12:15:50 I login via ssh :) Feb 07 12:16:02 hrw, nope yet, folks should submit support soon Feb 07 12:16:29 I'll get a picture when I get it running with microdrive and picopsu Feb 07 12:17:38 Crofton : i think i'll stick to the network boot... it's faster when changing things... Feb 07 12:17:47 yeah Feb 07 12:19:32 I use mine to demonstrate software radio on a small platform Feb 07 12:21:14 Crofton : cool ! I will run asterisk and hopefully g729 on it Feb 07 12:25:05 neat Feb 07 12:25:30 I need a small display/keypad for it so I could make an internet radio for my house :) Feb 07 12:28:02 I get an error on mtn pull now Feb 07 12:28:05 mtn: multiple update candidates: Feb 07 12:28:05 mtn: 531b56411734b2c857cc1bcb737fcccef9833ab1 pfalcon@openembedded.org 2007-02-07T12:41:01 Feb 07 12:28:08 mtn: f70da3815672ac605b525b2d9d2ee990d53d421f rwhitby@nslu2-linux.org 2007-02-06T10:23:52 Feb 07 12:29:04 * CM tries with the latest Feb 07 12:29:47 CM: an error on pull, or multiple heads on update? Feb 07 12:30:04 (the former would be a problem, the latter is normal for a distributed SCM) Feb 07 12:30:46 Latter.. Feb 07 12:30:59 I'm a real newbie on monotone Feb 07 12:31:03 fixed Feb 07 12:35:13 I've started having a strange problem with bitbake. On occassion it crashes and gives the following error: OSError: [Errno 2] No such file or directory: './sbin/.debug' Feb 07 12:35:13 I can post the full error report if needed. Feb 07 12:35:33 are you using symlinks ? :) Feb 07 12:35:39 I rerun the exact command and then bitbake works just fine. It is very random but happens several times.. Feb 07 12:36:05 Eblis yes, my local repository and source directory are symbolic links. Feb 07 12:36:33 there are warnings everywhere saying you should not use symlinks :) Feb 07 12:39:08 how can I set SRC_URI = "svn://.......;rev=${SRCDATE} so that when I 'require' this file in another recipie and override SRCDATE, it passes down to the required file? Feb 07 12:39:25 is that possible? or must I also override SRC_URI in the new recipie? Feb 07 12:41:48 set SRCDATE then require other recipe Feb 07 12:41:59 order matter Feb 07 12:42:14 hrw: ah, thanks Feb 07 12:42:30 Eblis, ok thanks.. I have been using them for almost a year, I wonder why it just now started to mis-behave. Feb 07 12:42:33 oh... thats what I did Feb 07 12:42:49 tkp: ;rev= should work Feb 07 12:43:06 cu Feb 07 12:43:46 koen|away: I need to be able to override the rev number tho Feb 07 12:44:19 hmm Feb 07 12:44:33 there aren't any really clean ways to do that Feb 07 12:45:05 like: http://rafb.net/p/yI3UbI12.html Feb 07 12:45:07 oh Feb 07 12:45:27 ( where the top file is settopcore.inc) Feb 07 12:46:34 http://rafb.net/p/yI3UbI12.html Feb 07 12:46:53 that way you can override MYREV in local.conf Feb 07 12:46:58 pb_, so it got further this time, but still hangs on boot :( Feb 07 12:47:26 I get to: "Freeing init memory: 80K" Feb 07 12:48:36 mccarthy: drat. does it say anything about an initial console and not being able to open it? Feb 07 12:48:51 koen|away: you pasted back the same code? Feb 07 12:50:15 pb_, I see nothing that looks like an error, really. As for a console, the only note I get is: "Console: colour dummy device 80x30" Feb 07 12:50:31 that is odd Feb 07 12:50:42 what's the kernel command line that it reports? Feb 07 12:50:55 (should be near the top of the output) Feb 07 12:51:12 I think it gets truncated (because of my minicom width), but what I can see says: Feb 07 12:51:19 Kernel command line: mtdparts=ipaq:0x00040000@0x00000000(bootldr)ro,0x01FC0000@ Feb 07 12:51:25 ah Feb 07 12:51:29 yeah, that does look truncated Feb 07 12:51:35 try turning on line wrap in minicom Feb 07 12:51:40 oh Feb 07 12:51:48 didn't know tht was an option ... ;) Feb 07 12:52:06 or try again after you widen the window :) Feb 07 12:52:11 heh Feb 07 12:52:22 tkp: http://rafb.net/paste/p/b1nedh16.html Feb 07 12:52:24 minicom -w, or ctrl-a w when it's running Feb 07 12:52:52 thanks Feb 07 12:52:53 koen|away: broken link! Feb 07 12:52:57 booting again .... Feb 07 12:53:08 grrr Feb 07 12:53:13 tkp: http://rafb.net/p/b1nedh16.html Feb 07 12:53:16 Kernel command line: mtdparts=ipaq:0x00040000@0x00000000(bootldr)ro,0x01FC0000@0 Feb 07 12:53:16 x00040000(root) root=/dev/mtdblock1 noinitrd console=ttyS0,115200 rootfstype=jf Feb 07 12:53:16 fs2 Feb 07 12:53:18 ta Feb 07 12:53:25 * koen|away stabs ubuntu for breaking text select in firefox Feb 07 12:53:30 (damn line wrap) ;) Feb 07 12:53:49 mccarthy: that all looks basically okay. try putting "init=/bin/sh sh" on the end and see if anything changes Feb 07 12:53:59 ok Feb 07 12:54:11 why use MYREV rather than SRCDATE? I thought thats what SRCDATE was for Feb 07 12:54:38 SRCDATE is a *date* Feb 07 12:54:45 not a revision number Feb 07 12:54:54 pb_, hmm. Can I do that in the bootldr? Feb 07 12:54:55 ah ok Feb 07 12:55:17 pb_, when I do params show linuxargs I get something different than above ... Feb 07 12:55:36 mccarthy: 'params eval' Feb 07 12:55:46 koen|away, after I change it? Feb 07 12:55:52 before Feb 07 12:55:59 that will make it read /boot/params Feb 07 12:55:59 koen|away, thanks Feb 07 12:57:14 pb_, no joy Feb 07 12:58:03 * koen|away heads to the station Feb 07 13:08:30 mccarthy: drat, still the same? Feb 07 13:08:43 how strange Feb 07 13:11:42 same olde Feb 07 13:13:40 but the prebuilt images do boot, right? Feb 07 13:14:09 yes Feb 07 13:14:45 I made both a gpe-image and a bootstrap-image so I am trying the bootstrap now Feb 07 13:15:39 then I will install one of the testing images and see if I can build from there (I am not sure that they have lzrsz or whatever it is called though so getting ipks onto it will be painful) Feb 07 13:15:45 might be worth looking in your tmp/rootfs directory to see if there are any obviously missing files (compared to what's in the prebuilt version) Feb 07 13:16:03 i.e. that you have /bin/busybox, /bin/sh, /lib/libc.so.6 and that kind of thing Feb 07 13:16:15 pb_, ok Feb 07 13:29:02 has anyone had success (and do I need to) build at76c503a-modules for the h5000? Feb 07 13:29:40 it looks like the build dies do to some missing include file (although it is not identified, it simply says lots of stuff is not defined) Feb 07 13:30:08 the Makefile patch gets rejected, but it does not look like it would affect include files ... Feb 07 13:54:36 what is the best way to tell kdrive not to power the display off after 10 minutes? Feb 07 13:55:50 cbrake: I remember that the xset utility had some screensaver / dpms options.. but I have no idea if that applies to kdrive Feb 07 13:59:02 Jin^eLD: I've spent some time with xset and have not been able to make it turn off the blanking. I'll give it another go .. Feb 07 14:00:22 I just remember it worked for me on PC, but that was quite some time ago.. maybe kdrive has something different Feb 07 14:07:04 how do you handle a package that already has an _ in the name? (doesn't OE think that it is part of the version number then?) Feb 07 14:13:21 re Feb 07 14:16:03 mccarthy: recipe renamed to not have _ Feb 07 14:16:33 Laibsch: did you get my message about do_stage and kpengine? Feb 07 14:18:25 hrw, but then when the source is unpacked bitbake looks in the wrong dir (I realize that this is a dumb euestion)? Feb 07 14:18:32 question even Feb 07 14:19:23 S="${WORKDIR}/source-location/" in recipe Feb 07 14:19:54 hrw, cool. thanks Feb 07 14:24:07 cbrake: I had a similar problem with kdrive... did you work out how to alter the screensaver options? Feb 07 14:26:10 koen|gprs_: No. What channel? mail? IRC? Feb 07 14:26:41 ~lart mtn 0.32 migration Feb 07 14:26:41 * ibot hits mtn 0.32 migration with an anvil and laughs with a contralto voice ... Haha Ha HA Ha Feb 07 14:26:48 morning Feb 07 14:26:56 good morning Feb 07 14:27:21 tkp: testing "xset s off" now Feb 07 14:27:45 tkp: also did a "xset -dpms" for good measure. Feb 07 14:28:59 natezg: saw an add for the iPac 9302 in LJ. Is that the board you are working on? Feb 07 14:29:30 damn. still lots of include-related errors Feb 07 14:31:08 tkp: argh, shut off after 10 minutes :-( Feb 07 14:31:14 doh! Feb 07 14:31:44 tkp: guess its time to strace xset and see what it is doing ... Feb 07 14:32:26 Laibsch: irc, but my connection is flaky Feb 07 14:32:37 Laibsch: kpengine is missing a do_stage, so you can't buid other programs against it Feb 07 14:33:01 OK Feb 07 14:33:23 This is new to me but I guess that is what I was trying to do with the native package? Feb 07 14:33:35 koen|gprs_: Would an empty do_stage suffice? Feb 07 14:34:23 do_stage() { Feb 07 14:34:27 autotools_stage_all Feb 07 14:34:28 } Feb 07 14:34:31 ping XorA Feb 07 14:36:37 koen|gprs_: Thanks a lot for the hint. I will update the bb file. do_stage is important for any kind of lib package? Feb 07 14:37:53 tkp: xset appears to be openning a socket at /tmp/.X11-unix/X0, and then doing some stuff. Time to look at the code ... Feb 07 14:38:50 cbrake: which server are you using (Xfbdev, Xvesa or one of the more hardware specific ones)? Feb 07 14:39:11 Laibsch: it is, otherwise it a bit useless to DEPEND on it :) Feb 07 14:39:11 tkp: Xfbdev Feb 07 14:40:16 koen|gprs_: kpengine is a small binary, not a library. only RDEPENDs are useful. Feb 07 14:40:27 ah Feb 07 14:41:36 koen|gprs_: I was planning to RDEPEND on it. But I guess this issue is still a bit foggy for me. Feb 07 14:41:37 what is the difference between RDEPENDS and RRECOMMENDS? Feb 07 14:41:54 cbrake: MUST and SHOULD Feb 07 14:42:27 cbrake: RDEPENDS makes your package manager error out if it doesn't find the package, RRECOMMENDS doesn't Feb 07 14:42:30 cbrake: Feb 07 14:42:53 that's why you shouldn't RDEPEND on kernel-modules, since someone might have them built-in Feb 07 14:43:45 koen|gprs_: makes sense -- thanks! Feb 07 14:43:55 cbrake: Interestingly, I don't get the screen blanking on my TV out.. only on the main monitor Feb 07 14:44:06 koen|gprs_, so bb will try to build all the packages in the RRECOMMENDS list (task-base.bb) . Is there away (other than making a new task) to tell bb not to build the RRECOMMENDS for a task? Feb 07 14:44:26 Gerrath_: there isn't Feb 07 14:44:44 koen|gprs_, ok thanks. Feb 07 14:48:23 koen|gprs_, that would be a nice feature (in my opionion) to be able to in the FEATURES (MACHINE_FEATURES, DISTRO_FEATURES...) list to disable RRECOMMENDS: example:(DISTRO_FEATURES. = alsa_NORRECOMMENDS) Feb 07 14:48:44 koen|gprs_, of course me being able to spell would be a nice feature too :-p Feb 07 14:51:33 tkp: just had a chat with Jim Gettys on #xorg. He said that after the last X apps, the server is reset. Feb 07 14:51:48 tkp: so you need to start your app, and then run xset. Feb 07 14:51:58 tkp: trying now Feb 07 14:52:15 cbrake: oh, interesting... lets see then! ;) Feb 07 14:53:51 tkp: run "xset q". I think the critical parameter is the Screen Saver:timout Feb 07 14:54:38 tkp: I only have one full screen app in my system, so I was running xset, and then my app. When the last X app starts, the Xserver gets reset. Feb 07 14:54:51 tkp: so, xset exits and is the last app, resetting the xserver. Feb 07 14:55:04 right Feb 07 14:56:07 tkp: may also be able to use the "-s 0" argument to Xfbdev Feb 07 15:09:30 Hi. We have a problem determining the correct SECTION for kpengine. This is a helper program that gets fed stroke information and produces a list of Kanji out of it. There is libs/inputmethods, but kpengine technically is not a library... Feb 07 15:11:23 little problem with busybox bb file... Feb 07 15:11:51 it specifically sets CROSS_COMPILER_PREFIX = ${TARGET_PREFIX} Feb 07 15:12:07 well this is wrong if (like me) you are using an external toolchain Feb 07 15:12:11 koen|away: http://www.openembedded.org/viewmtn/revision.psp?id=bde874d0d89e56263d17c6a6b5cc1478e83d5d2a will that error fix itself? Feb 07 15:13:11 since my toolchain tools are not named ${TARGET_PREFIX}gcc etc Feb 07 15:15:48 pH5: To pick up your question. I just fixed uim.inc and anthy*.bb from /libs/inputmethod to libs/inputmethods as suggested on http://www.openembedded.org/user-manual&dpage=section_variable. But I think neither of them is a true lib either. Maybe they all should be moved to just "inputmethods"? hrw, any comment? Feb 07 15:19:00 Laibsch: x11/inputmethods or opie/inputmethods rather Feb 07 15:19:29 hrw: what about helper programs that could be used by both x11/gpe and opie frontends? Feb 07 15:20:19 base/inputmethods? libs/inputmethods? Feb 07 15:21:22 hrw: is it ok to put a binary like kpengine in the libs/inputmethods section? Feb 07 15:22:13 or inputmethods itself Feb 07 15:22:38 I think that is the best option. Feb 07 15:22:55 It is not base and it is not a lib. Feb 07 15:23:40 pH5: I guess the common stuff to anthy and uim should be moved from libs as well, right? Feb 07 15:24:27 hrw, Laibsch: agreed, just libanthy (which is built by the anthy recipe) really belongs in libs/inputmethods. Feb 07 15:24:44 same for libuim Feb 07 15:25:14 uim-gtk2.0 should be in x11/inputmethods then. Feb 07 15:31:57 ok Feb 07 15:39:51 OK, I will try to make the changes. Feb 07 15:40:41 Laibsch: great Feb 07 15:41:35 hrw: Do you allow commits to .documentation? I would like to fix some issues mentioned on http://www.openembedded.org/user-manual&dpage=section_variable and then update the page. Feb 07 15:43:38 Laibsch: fix, commit, pester to review Feb 07 15:44:01 hrw: OK Feb 07 15:44:52 Is there some kind of "accepted rule of thumb" when it makes sense to keep even a large number of packages in a single bb file rather than creating individual bb files for them? I was wondering about uim and now kpengine in the past. I moved kpengine out of kanjipad because a) the GUI stuff does not build and b) it can be used by other programs. Feb 07 15:45:34 I wonder if others have thoughts on this they'd like to share? How did you guys handle such situations? What criteria are reasonable? Feb 07 15:46:44 Laibsch: I see no reason to split .bb files for a single source package. Feb 07 15:46:56 Laibsch: if can be build in one go and this not add extra deps then one per many Feb 07 15:47:33 if some parts are common use and few subpackages are rare but need many extra deps then split to common ones + extra ones Feb 07 15:48:11 Laibsch: look at kdepimpi - nearly all is built from kdepimpi but pwmpi is separated because need to link with stdc++ Feb 07 15:48:14 brb Feb 07 15:53:25 damn Feb 07 15:53:52 I can boot the testing image from Angstrom, but it has no fb support ... Feb 07 16:03:41 YEEES! after 1.5 weeks I finally have a working mipsel build / rootfs Feb 07 16:03:46 :) Feb 07 16:04:46 mmc: Add support for SDHC cards Feb 07 16:04:50 yay Feb 07 16:05:01 Jin^eLD: Whooooo :) now to make it do something usefull :) Feb 07 16:05:19 Angstrom on me WRT54G :-) Feb 07 16:05:29 XorA: ;D Feb 07 16:05:47 yeah.. indeed :) Feb 07 16:06:59 I wish the vendor would supply a newer kernel, that was what caused the fight.. newer glibc with nptl did not play together with it Feb 07 16:08:51 Jin^eLD: heh... It seems these days if you want it done right, do it yourself ;) Feb 07 16:10:01 hvontres|poodle: yes, you are absolutely right Feb 07 16:10:17 grrrrr... http://www.openembedded.org/viewmtn/ comes back with a "database locked error" Feb 07 16:10:31 well, my problem in that regard is, that I am not a kernel hacker, I mostly do app-level coding, not to good with lowlevel hardware stuff Feb 07 16:10:51 so I am kind of stuck with the kernel, I would also lose vendors support if I port to a newer kernel Feb 07 16:10:55 so crappy situation Feb 07 16:10:56 :) Feb 07 16:11:35 Jin^eLD: yeah, and not just with software :( I am a mechanical engineer and it is amazing how inept some vendors can be and how stupid they think the customer must be to not notice the crappy parts they send :( Feb 07 16:12:51 that's bad... seems everone wants to get out their stuff as cheap as possible Feb 07 16:14:20 well. I am really thankfull there is OE :) Feb 07 16:14:47 else the "doing yourself" part would really be a horror scenario Feb 07 16:14:59 so thanks to all of you guys :) Feb 07 16:34:37 monotone 0.32 migration took 4.5h here Feb 07 16:59:38 pH5|hrw: Is poboxserver a real lib? I doubt it since there is no do_stage. Feb 07 17:01:57 Laibsch: not a lib. it is a server binary. Feb 07 17:02:40 So, similar to kpengine which is not a daemon, though, right? Feb 07 17:02:50 I will move it to "inputmethods", too. Feb 07 17:03:37 Laibsch: right. Feb 07 17:03:39 PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND Feb 07 17:03:39 23457 hrw 26 10 948m 716m 3124 D 0.2 76.2 1:00.09 cc1plus Feb 07 17:03:49 masacra.. Feb 07 17:04:36 hrw: what in the world are you compiling there ? Feb 07 17:05:27 CIA-19: Are you alive? Feb 07 17:05:42 kick CIA-19 Feb 07 17:06:00 * Laibsch feels bad for the mistreatment of such a nice servant Feb 07 17:06:17 But it seems he is asleep at the wheel. Feb 07 17:06:44 Laibsch: he had a long night with miss CIA-28 ;) Feb 07 17:22:33 cu Feb 07 17:22:51 Wouldn't it make sense to strip "native" from ${S} for native package? I doubt any source code will ever ship in ${P}/ plus ${PN}-native-${PV}/. What do you think? What save us from resetting ${S} manually a lot of times Feb 07 17:22:56 see you hrw|gone. Feb 07 17:23:05 s/What/Would/ Feb 07 17:36:41 I have put together a meta file following these instructions: http://www.openembedded.org/user-manual&dpage=commonuse_new_image Feb 07 17:37:43 I then bitbake metafile-name and it builds all of the packages but does not create a tmp/rootfs or the jffs2 filesystem image (which I have stated in IMAGE_FSTYPES = "jffs2 tar" Feb 07 17:38:18 what am I missing? This same meta file worked fine several months ago but does not seem to work with the latest OE. Feb 07 17:42:18 Gerrath_, meta recipe supposed to build old packages, not image Feb 07 17:43:00 psokolovsky, it use to work, I'm taking it with the task-base system it no longer does? Feb 07 17:44:37 Gerrath_, I have no idea if it worked, but it's not supposed to be that way. or rather, all stuff in mainline OE follows meta-* va *-image pattern. Feb 07 17:47:07 psokolovsky, so I would set up my tasks then add in any extra packages with MACHINE_EXTRA_RDEPENDS then do a bitbake bootstrap-image? Feb 07 17:49:40 Gerrath_, yes, if you're satisfied with what bootstrap-image does for you Feb 07 17:50:58 psokolovsky, I just used that one as an example but basically I need to find the closest *-image for my needs, add in the tasks that meet my minimum requiremetns with out adding too much extra then bridge the remaining gap with MACHINE_EXTRA_RDEPENDS. Feb 07 17:51:19 psokolovsky, if I'm trying to keep a lean system :-) Feb 07 17:52:11 my build is missing the kernel modules (/lib/modules is empty) Feb 07 17:52:22 what flags set and unset this in the configs? Feb 07 17:52:44 I'm building a bootstrap image Feb 07 17:57:15 natezg: add a kernel module to RDEPENDS, and everything should get populated. Feb 07 17:58:12 natezg: I think you also need to add "kernel" to redpends, or module autoloading won't work properly. Feb 07 17:59:45 cbrake: I'll give that a shot, also where does kernel-module-xxxx get interpreted? where is this recipe? Feb 07 18:00:25 natezg: when you build the kernel, any modules are automatically packaged into individual ipkg packages. Feb 07 18:01:08 natezg: so build the kernel, see what what kernel module packages get populated in tmp/deploy/ipk/ ... Feb 07 18:01:29 natezg: you can then add these packages to RDEPENDS to get them included in the image, or install them after the fact with ipkg. Feb 07 18:01:33 I can see it's building them, just not copying them into the build for some reason, I'll try adding to RDEPENDS Feb 07 18:02:07 natezg: they will probably need to end up in tmp/deploy/ipk/.. for this to work. Feb 07 18:02:17 kk Feb 07 18:02:35 natezg: do you have anything in tmp/deploy/ipk? Feb 07 18:04:21 kernel-modules_2.6.19-r1_IPAC9302.ipk Feb 07 18:04:34 I have bitbaked an Angstrom image that does not seem to include all of the kernel modules that are actually built ... where does this get fixed? Feb 07 18:04:53 (incidently the Angstrom testing images have the same issue) Feb 07 18:05:17 hey koen|gprs__ :) Feb 07 18:05:29 koen|away, any chance you can answer for me? Feb 07 18:05:52 (don't even know if you heard the question, with all the joining ...) Feb 07 18:06:12 mccarthy:I'm dealing with the same issue, cbrake just addressed it Feb 07 18:06:36 mccarthy:or at least a very similar one Feb 07 18:07:28 re Feb 07 18:07:37 mccarthy: not without knowing the question Feb 07 18:07:50 natezg, doh! I didn't even scroll up (who would have thought you just asked that ...) Feb 07 18:08:04 koen|away, I have bitbaked an Angstrom image that does not seem to include all of the kernel modules that are actually built ... where does this get fixed? Feb 07 18:08:30 mccarthy: in the $machine.conf, $distro.conf or task-base.bb Feb 07 18:09:05 I think were is the wrong question, how would be a better one Feb 07 18:09:17 koen|gprs__: Just for fun, have your Controls teacher cover resonant systems...they make for really amusing models :) Feb 07 18:09:21 cbrake recommends adding to RDEPENDS, I"m experimenting Feb 07 18:10:20 were->where Feb 07 18:10:46 natezg: easiest place is local.conf::DISTRO_EXTRA_RDEPENDS Feb 07 18:11:24 I RDEPENDS += "kernel" in the distro conf, that sound kosher? Feb 07 18:11:46 natezg: probably not because RDEPENDS defined in distro.conf is global. Feb 07 18:12:09 o...., better change that Feb 07 18:12:27 natezg: local.conf + distro.conf + machine.conf + form a bunch of global variables that apply to all packages. Feb 07 18:13:04 natezg: see openembedded/conf/bitbake.conf Feb 07 18:14:01 still it would probably be better to putDISTRO_EXTRA_RDEPENDS in my distro's .conf rather than local.conf right? Feb 07 18:14:44 natezg: it will have the same effect, so it really depends how you like to organize things. Feb 07 18:15:00 03rwhitby 07org.oe.dev * r34e476df... 10/ (1 conf/distro/include/angstrom.inc): angstrom.inc: override DISTRO_FEATURES for the nslu2le machine to get the rootfs under 6.5MB Feb 07 18:15:04 03Laibsch 07org.oe.dev * rbd59bf13... 10/ (6 files in 4 dirs): Feb 07 18:15:04 kanjipad: Add kpengine module which can be used by a number of packages. Feb 07 18:15:04 Thanks to pH5 for doing most of the preliminary work. Feb 07 18:15:09 03rwhitby 07org.oe.dev * r68ffb089... 10/ (7 files in 3 dirs): arm-kernel-shim: Updated to latest upstream, and fixed the .boot section placement with a temporary Makefile ld command-line hack. Feb 07 18:15:13 03rwhitby 07org.oe.dev * r1e9ea583... 10/ (6 files in 4 dirs): ixp4xx-kernel: Replaced nslu2 and nas100d cmdline hack patches with arm-kernel-shim. Feb 07 18:15:17 03florian 07org.oe.dev * r38430e5d... 10/ (5 files in 3 dirs): e2fsprogs-libs: Add 1.39, take care of pkgconfig files. Feb 07 18:15:21 03florian 07org.oe.dev * rceb628b9... 10/ (4 files in 3 dirs): phoneserver: Add patch to use a different port. Feb 07 18:15:29 03rwhitby 07org.oe.dev * ra4be82b8... 10/ (1 packages/linphone/linphone_1.6.0.bb): linphone (1.6.0): Use the OE-built pkg-config Feb 07 18:15:36 hmm, well I wast the kernel to addition to be specfic to the distro, as this is what it will effect Feb 07 18:15:37 03rwhitby 07org.oe.dev * r865584df... 10/ (1 conf/distro/include/slugos.inc): slugos: bump version to 4.3 (tested BE) Feb 07 18:15:45 03rwhitby 07org.oe.dev * r5c2c70f1... 10/ (1 packages/arm-kernel-shim/files/place-boot-section.patch): arm-kernel-shim: Add missing patch file Feb 07 18:15:54 natezg: Gerrath_ and I have been hashing this over -- it seems like a good way to create custom images is define a -image.bb file that DEPENDS on task-base and other packages/meta packages you want in them image. Feb 07 18:16:30 that sounds logical, yea Feb 07 18:17:05 natezg: but, I'm still using DISTRO_EXTRA_RDEPENDS for most of my stuff. Feb 07 18:18:05 natezg: I try to minimize my changes to .conf so I can easily track changes in OE, but depending on your needs, a custom distro may be needed as well. Feb 07 18:18:46 creating new image recipies seems wrong though, thinking about it Feb 07 18:19:12 The idea of the boostrap-image, I thought, was to create the minimum image for a distibution Feb 07 18:19:49 what that minimum image is , is supposed to be determined by the distro's .conf, correct? Feb 07 18:19:57 hi all Feb 07 18:19:59 right Feb 07 18:20:02 natezg: good point Feb 07 18:20:11 ~seen zecke Feb 07 18:20:14 within the boundaries of said -image.bb Feb 07 18:20:20 zecke was last seen on IRC in channel #oe, 1d 21h 15m 56s ago, saying: 'later'. Feb 07 18:20:48 in that case, using flags in the distro.conf to add to the distro would make sense, but how is that done? should I be appending directly to RDEPENDS? Feb 07 18:20:50 natezg: thinking more, you probably want .conf to drive inclusion of kernel modules in the image as they are really machine specific. Feb 07 18:21:02 natezg: that is really a machine specific thing. Feb 07 18:21:18 cbrake:maybe so..... Feb 07 18:21:32 if you are in doubt where to put it, imagine building the distro for a completely different machine Feb 07 18:21:46 say, a playstation or cellphone Feb 07 18:22:14 well, in that case I would want it to include the kernel modules in general, i.e. the depmod files and all that Feb 07 18:22:22 probably at the distribution level Feb 07 18:22:24 natezg: one distro should work on many machines, so you need MACHINE_EXTRA_RDEPENDS in .conf Feb 07 18:22:25 but which ones..... Feb 07 18:22:33 true Feb 07 18:23:03 I'll try MACHINE_EXTRA_RDEPENDS = "kernel" see if that gets me anywere Feb 07 18:23:51 MACHINE_EXTRA_RDEPENDS += Feb 07 18:23:58 natezg: including one kernel module ipk will pull in dependencies for depmod, modprobe, etc -- I think. Or most distros migh include modutils, etc. Feb 07 18:24:30 module-init-tools :) Feb 07 18:24:53 koen:isn't that brought in by task-base? Feb 07 18:25:03 it should be Feb 07 18:25:26 but I'm not seeing anything in the "/lib/modules" directory Feb 07 18:26:01 yet I do have modprobe Feb 07 18:26:09 natezg: try an actual kernel module like kernel-module-snd Feb 07 18:26:15 kk Feb 07 18:26:44 extra kernel modules are generally included in $machine.conf via MACHINE_EXTRA_RRECOMMENDS Feb 07 18:26:46 cbrake:where are these names coming from, are they ipkg names? Feb 07 18:26:57 natezg: yes Feb 07 18:27:52 natezg: compulab-pxa270/tmp/deploy/glibc/ipk/kernel-module-snd_2.6.16-r13_compulab-pxa270.ipk Feb 07 18:28:01 wierd: I just tried to do a bitbake bootstrap-image and I get the following error: ERROR: Cannot satisfy the following dependencies for task-base: Feb 07 18:28:01 | base-passwd module-init-tools-depmod portmap udev-utils Feb 07 18:28:12 but all those packages are built. Feb 07 18:28:30 cbrake:by that logic I shuld be able to MACHINE_EXTRA_RRECOMMENDS = "kernel_2.6.19-r1_IPAC9302" Feb 07 18:28:36 which is a generated ipkg Feb 07 18:28:59 natezg: no, you never include version in the package name. Feb 07 18:29:30 cbrake: so just MACHINE_EXTRA_RRECOMMENDS = "kernel-modules" then Feb 07 18:29:43 cbrake: for kernel-modules_2.6.19-r1_IPAC9302.ipk Feb 07 18:30:18 that will include *all* modules that are built Feb 07 18:30:41 fine with me at the moment Feb 07 18:30:47 as I'm controlling that Feb 07 18:31:18 but in keeping with the OE system you would probably build everything and just include what you actually want for the machine I'm assuming Feb 07 18:31:30 right Feb 07 18:31:37 very kewl, thank you Feb 07 18:31:52 MACHINE_EXTRA_RRECOMMENDS is what is needed to get the machine in a reasonable state Feb 07 18:32:17 koen: RR = reasonable recommends? Feb 07 18:32:18 e.g. framebuffer, ide, usbhost, etc Feb 07 18:32:25 ah Feb 07 18:32:31 PR = package revision Feb 07 18:32:41 RRECOMMENDS = Runtime Recommends Feb 07 18:32:56 it's what gets put in the Recommends: fields of ipkg, debs and rpms Feb 07 18:33:13 ll Feb 07 18:33:15 kk Feb 07 18:33:21 I mean... Feb 07 18:35:47 Gerrath_: seems like I ran into something similiar once, but can't remember what the solution was ... Feb 07 18:37:05 cbrake, its reall strange.. if I do a clean on base-passwd then rebuild it the build crashes during the install but if I re-run the build in passes the second time. I repeated this several times. I had a directory with a symbolic link and thought that was the issue so I removed it but no-go. Feb 07 18:38:02 Gerrath_: I've noticed some strange problems in bitbake interactive. Have you tried this outside bitbake interactive? Feb 07 18:38:33 cbrake, thats what I'm running "bitbake base-passwd" Feb 07 18:39:18 I think I might blitz everything since I have been having several strange problems and just do a fresh build. Feb 07 19:01:09 I have the rt kernel running on the OSK :) Feb 07 19:01:13 has the divide by zero error on shutdown in /etc/init.d/rc been fixed? Feb 07 19:03:22 cbrake: there was a 'fix' somewhere in a bugreport Feb 07 19:03:24 haven't seen a proper patch yet Feb 07 19:05:18 koen|away, in the h5000.conf the kernel-modules-h5400-lcd is part of the MACHINE_EXTRA_RRECOMMENDS, but it is still not put into the image?! Feb 07 19:05:35 (although I can find the ipkg in the ipk directory) Feb 07 19:05:42 any ideas? Feb 07 19:06:40 koen|gprs_: I've been looking at the bug reports. Looks to me like there is a typo in /etc/init.d/rc Feb 07 19:07:42 koen|gprs_: numsteps="1" should be num_steps="1" Feb 07 19:08:31 koen|gprs_: unless you object, I'll push a fix once I figure out who provides the file. Feb 07 19:08:48 cbrake: iirc initscript in the files/angstrom dir Feb 07 19:08:48 and I don't object Feb 07 19:09:00 mccarthy: weird Feb 07 19:09:42 like I said earlier (but maybe before you were fully joined) I think the same thing is happending in the testing builds of Angstrom Feb 07 19:09:53 (since no lcd modules are in the gpe-image) Feb 07 19:11:04 I see the problem.... Feb 07 19:11:43 * koen|gprs_ pushes a fix Feb 07 19:11:56 for me or for cbrake ? Feb 07 19:12:36 you Feb 07 19:12:36 (now if I could only get my home-grown images to get past the INIT stage (hangs)) Feb 07 19:12:41 koen|away, cool Feb 07 19:12:46 (and the gang) Feb 07 19:16:00 03koen 07org.oe.dev * r400e309c... 10/ (1 conf/machine/h5000.conf): h5000.conf: another hyphen versus underscore problem.... Feb 07 19:16:07 03cbrake 07org.oe.dev * r918e51da... 10/ (1 packages/sysvinit/sysvinit/angstrom/rc): sysvinit: fix divide by 0 error on shutdown in /etc/init.d/rc Feb 07 19:16:25 mtn: 1.5 M | 192.1 k | 69/75 | 1/1 Feb 07 19:16:29 * koen|gprs_ stabs gprs Feb 07 19:17:09 cbrake: don't forget to bump PR for sysvinit Feb 07 19:19:58 koen|gprs_: oops, will do ... Feb 07 19:20:14 hmm, this means my efika should reboot properly :) Feb 07 19:20:45 Crofton: seems to be working for me :-) Feb 07 19:24:20 cbrake, updated OE, did a clean on all the troubled packages and now boot-strap-image works. Feb 07 19:24:58 Gerrath_: :-) Feb 07 19:29:45 03cbrake 07org.oe.dev * r23dabd48... 10/ (1 packages/sysvinit/sysvinit_2.86.bb): sysvinit: bump PR for divide by zero fix in /etc/init.d/rc Feb 07 20:00:36 03shanevolpe 07org.oe.dev * r3f262d1f... 10/ (6 files in 3 dirs): (log message trimmed) Feb 07 20:00:36 gdb-cross: fix issue where patch has already been applied to 6.6 Feb 07 20:00:36 gdb-cross version 6.6 fails the patch process because the sim-install.patch Feb 07 20:00:36 tries to apply patches to Makefile-common which have already been merged into Feb 07 20:00:36 gdm mainstream. Feb 07 20:00:36 Attached is a patch for OE that seperates the two patches in sim-install.patch Feb 07 20:00:40 so that gdb-cross 6.6 only applies the first one while gdb-cross 6.3 (which Feb 07 20:10:19 ~lart psplash for hanging on h3900 (ladnscape LCD) Feb 07 20:10:19 * ibot whacks psplash with a giant beaver's tail for hanging on h3900 (ladnscape LCD) Feb 07 20:12:14 just got back from lunch Feb 07 20:12:31 was there a bug preventing the modules from being installed or was that something else? Feb 07 20:12:50 something else Feb 07 20:12:54 k Feb 07 20:13:06 where 'something else' means 'koen made another set of typos' Feb 07 20:13:14 ah Feb 07 20:13:24 updating Feb 07 20:24:28 updated OE gives me:ERROR: Could not include required file ../meta/nylon-feed.inc while parsing /home/nathan/lab/IPAC-9302/openembedded/IPAC9302/packages/images/nylon-image-extended.bb Feb 07 20:24:49 ah, nvmd part of my custom stuff sorry Feb 07 20:38:06 if I need a perl module, and a .bb file isn't available for it, what other options are there for building it? Feb 07 20:40:54 I'm trying to build libxml-parser-perl and it's failing Feb 07 20:41:02 re Feb 07 20:41:58 xumbi8: create recipe for perl module which you need Feb 07 20:43:33 hrw: ok, I'd need to make my own .bb file? I didn't know if there might be other ipk repositories out there Feb 07 20:49:10 xumbi8: if there is no recipe for that module in .dev then there is no such recipe Feb 07 20:52:04 hi philippe Feb 07 20:59:09 hi phil Feb 07 21:02:06 hi hrw Feb 07 21:02:09 grr, stupid ups Feb 07 21:02:28 * pb__ at the office reconnecting power to the network equipment Feb 07 21:03:01 * hrw fetching amd64 debs from incomgin.debian.org Feb 07 21:47:47 cu Feb 07 23:23:52 does anybody have a clue what could cause this? | /home/fille/oe/build/tmp/work/arm-oabi-angstrom-linux/libgpepimc-0.8+svn20070207-r0/temp/run.do_configure.7132: line 201: syntax error near unexpected token `(' Feb 07 23:32:59 hi, all! Feb 07 23:38:57 howdy Feb 07 23:39:15 my build is still not including the kernel modules Feb 07 23:39:32 does anyone know where this copy occurs in the OE system so I can debug it? Feb 08 00:17:52 03pfalcon 07org.oe.dev * r2db23066... 10/ (4 files in 3 dirs): Feb 08 00:17:52 opie-taskbar: init.d/opie: explicitly shut down psplash on opie start, plus Feb 08 00:17:52 allow it some time for that. Otherwise, there's race between psplash and Feb 08 00:17:52 opie on some machines (seen on h3900), leading to opie hang. With this change, Feb 08 00:17:53 opie *could* run on VT 2, but it appears that opie's GUI manager has off-by-one Feb 08 00:17:55 error, causing it to not use last screen scanline, so old content is seen Feb 08 00:17:57 there, which is unaesthetic. So, keep running opie on VT 3, which is empty. Feb 08 00:17:59 03pfalcon 07org.oe.dev * r064f6cf6... 10/ (3 files in 3 dirs): libopie2: Add patch to properly handle all various native LCD orientations of ipaqs. **** ENDING LOGGING AT Thu Feb 08 02:59:58 2007