**** BEGIN LOGGING AT Mon Oct 15 02:59:57 2007 Oct 15 04:02:13 03rwhitby 07org.oe.dev * r08227e03... 10/ (8 files in 2 dirs): slugos-packages (and derivatives): inherit meta so all packages are built for the feeds Oct 15 04:29:02 03rwhitby 07org.oe.dev * r3156f282... 10/ (1 conf/distro/include/slugos.inc): slugos.inc: require sane srcdates and srcrevs Oct 15 05:45:55 03rwhitby 07org.oe.dev * rfc279ed8... 10/ (1 packages/meta/slugos-packages.bb): slugos-packages: Promoted some packages that now build, and demoted some that don't Oct 15 05:51:05 * * OE Bug 3156 has been created by nelson(AT)crynwr.com Oct 15 05:51:07 * * autotools_stage_all() tried to copy nonexistant files Oct 15 05:51:09 * * http://bugs.openembedded.org/show_bug.cgi?id=3156 Oct 15 05:51:46 cdbot2: 'bout time you noticed! Oct 15 07:15:06 good morning! Oct 15 07:16:38 hello Oct 15 07:17:30 hi nud..how's it going? Oct 15 07:17:39 not that well Oct 15 07:17:44 x_X Oct 15 07:17:49 that f*cking card makes me crazy Oct 15 07:18:05 what card? Oct 15 07:18:10 cm-x270 Oct 15 07:19:13 hehe Oct 15 07:28:30 gilligan_: did you make the script able to update the image ? Oct 15 07:28:35 * nud is dreaming Oct 15 07:29:45 did I make what? Oct 15 07:30:53 the python script able to update the nand image on the board Oct 15 07:32:02 armmon doesn't support flashing jffs2 images to nand, hence my script won't do that Oct 15 07:32:34 sniff Oct 15 07:32:36 * nud is crying Oct 15 07:44:37 uhhrm Oct 15 07:44:44 bitbake crashes one me ... Oct 15 07:44:46 Hmm... What is the correct usage of DEPENDS and RDEPENDS? If I understand http://www.openembedded.org/wiki/Bbfile correctly, a package that requires fx. libz.so.1 to run, should have zlib in RDEPENDS. Oct 15 07:44:59 File "", line 50, in check_sanityNameError: global name 'messags' is not defined Oct 15 07:45:09 But most likely it will also need it in DEPENDS. Oct 15 07:45:09 looks like a typo Oct 15 07:45:46 Do one of DEPENDS and RDPENDS by default include all from the other? Oct 15 07:52:31 Esben: librabry depends are calculated on the fly Oct 15 07:53:04 DEPENDS is for stuff it needs to build (e.g. zlib) and RDEPENDS if for runtime deps our shlibs code missed (e.g. iana-etc, terminfo, etc) Oct 15 07:53:13 koen: ok, then I must have stumbled upon a bug. Oct 15 07:53:25 that's not uncommon :) Oct 15 07:53:38 I just built an image with dropbear which did not include /usr/lib/libz.so.1 Oct 15 07:54:03 What is the correct way to fix it then (if not just adding a specific RDEPENDS)? Oct 15 07:58:59 rebuild dropbear Oct 15 08:01:30 I did, but did not help (before adding RDEPENDS="zlib") Oct 15 08:07:43 morning Oct 15 08:07:54 hi xora Oct 15 08:08:00 * nud has a "-lz not found" when compiling moc Oct 15 08:14:54 hey XorA Oct 15 08:16:08 morning Oct 15 08:18:53 hey hrw Oct 15 09:02:40 morning all Oct 15 09:16:05 * * OE Bug 3157 has been created by eha(AT)doredevelopment.dk Oct 15 09:16:07 * * Dropbear requires libz.so.1 Oct 15 09:16:09 * * http://bugs.openembedded.org/show_bug.cgi?id=3157 Oct 15 09:17:53 huh.. having problems with a clean compile ... Oct 15 09:17:56 cp: cannot stat `/home/pflug/projects/oedev/embedded/org.openembedded.dev/packages/base-files/base-files/angstrom/licenses/BSD': No such file or directoryNOTE: Task failed: NOTE: package base-files-3.0.14-r74: task do_unpack: failed Oct 15 09:29:51 bleh..was stupid vfat crap screwing things up Oct 15 09:38:10 lrg: yay! Oct 15 09:39:17 koen: better late than never :) Oct 15 09:40:12 wow, even with lcd Oct 15 09:43:51 " much of which involved a certain amount of ribbing of the Debian crew who didn’t look particularly amused, sipping on their half-pints of coke." Oct 15 09:44:02 jono's blog is funny :) Oct 15 09:46:01 He's on the A-list after all ;) Oct 15 09:54:47 hi all Oct 15 09:55:29 Esben : I am also seeing some strange dependency issues, but need to look into it with more detail Oct 15 09:55:58 Esben : quick solution to check is bitbake -g #foo Oct 15 09:56:10 Esben : and check the dep graphs Oct 15 09:59:17 moin Oct 15 10:08:48 03hrw 07org.oe.dev * r377affcc... 10/ (1 conf/distro/include/sane-srcrevs.inc): sane-srcrevs: bump openmoko-today2 Oct 15 10:08:54 03xora 07org.oe.dev * r394c920b... 10/ (3 files in 2 dirs): Oct 15 10:08:54 packages/gtk+/gtk+-fastscaling_2.10.14.bb : split out fast innaccurate Oct 15 10:08:54 scaling to a different gtk+ build. This makes more sense than a machine Oct 15 10:08:54 specific gtk+ build. Oct 15 10:08:59 03xora 07org.oe.dev * rc76fe542... 10/ (5 files in 4 dirs): Oct 15 10:08:59 distro/openmoko.conf : make gta01/02 equivalent machines for packages Oct 15 10:08:59 where the resulting packages are in fact identical. Oct 15 10:09:05 03xora 07org.oe.dev * rdaa2781e... 10/ (1 packages/ipkg/ipkg-collateral.bb): Oct 15 10:09:05 packages/ipkg/ipkg-collateral.bb : remove hard setting of PACKAGE_ARCH to Oct 15 10:09:05 allow the machine overrides in SRU_URI stuff to work correctly Oct 15 10:22:56 morning all, who can I ask about tslibs-test Oct 15 10:23:47 don't ask to ask, ask Oct 15 10:26:56 hello koen, I got some strange output of ts_* tools on OM2007.2 both on neo and qemu here http://claesmogren.com/~cm/neo/ts_harvest.out raw from neo (CM gratefully did the test), and here my "analysis" http://img81.imageshack.us/img81/4930/tsharvestgraphkg8.png Oct 15 10:27:33 looks like, X measured as Y and inverted, Y measured as X and multiplied by 2. looks as typo of some sort Oct 15 10:30:08 I don't really know, is it bug in neo patch for touchscreen driver or ts_harvest behave strange. Oct 15 10:30:28 or normal behaviour even Oct 15 10:30:46 you used ts_print or ts_print_raw to get data? Oct 15 10:31:01 hrw, ts_harvest Oct 15 10:31:24 hm. do not know that one ;( Oct 15 10:31:45 it defines points and ask to pin it Oct 15 10:31:59 or touch if you want Oct 15 10:35:02 I see - long time it takes to make full run. Oct 15 10:35:13 hrw: ts_harvest is a ts-print-raw tutorial Oct 15 10:35:13 just tried on neo Oct 15 10:35:46 a Oct 15 10:35:58 hi greentux Oct 15 10:36:05 hrw, imagine me doing it in qemu :-) Oct 15 10:36:21 in qemu it is easier Oct 15 10:37:04 a little, I got a a mouse Oct 15 10:47:13 22 degrees outside and I have a cold... Oct 15 10:54:11 ~lart 2.6 for generating big kernels Oct 15 10:54:11 * ibot pulls out a ClueBat (tm) and thwaps 2.6 for generating big kernels Oct 15 10:54:15 ~curse sharp Oct 15 10:54:16 May the fleas of a thousand camels infest your most sensitive regions, sharp ! Oct 15 10:57:00 hi hrw/koen Oct 15 10:57:01 03xora 07org.oe.dev * r47430349... 10/ (6 files in 4 dirs): Oct 15 10:57:01 packages/usbpath/* : add usbpath library and add is as DEPENDS to Oct 15 10:57:01 dfu-util Oct 15 10:57:06 03xora 07org.oe.dev * r3ce0d02c... 10/ (4 files in 2 dirs): conf/distro/*srcrev* : add usbpath Oct 15 10:57:10 03koen 07org.oe.dev * r7de13708... 10/ (4 files in 4 dirs): adzs-bf548-ezlite: add machine and defconfig Oct 15 11:02:05 * * OE Bug 3157 has been RESOLVED (INVALID) by Oct 15 11:02:07 * *  Dropbear requires libz.so.1 Oct 15 11:02:09 * * http://bugs.openembedded.org/show_bug.cgi?id=3157 Oct 15 11:45:38 hi all Oct 15 11:47:27 ~lart psokolovsky for his defconfigman generated configs without TS support Oct 15 11:47:27 * ibot resizes psokolovsky's terminal to 40x24 for his defconfigman generated configs without TS support Oct 15 11:52:43 hi florian Oct 15 11:53:37 koen : can you try to add boa to an image (i.e minimal-image) and see if it builds for you Oct 15 11:56:07 * koen boycots boa Oct 15 11:56:14 I can add cherokee and try :) Oct 15 11:56:49 koen : there seems to be a dependency issue that manifests with boa Oct 15 11:57:43 koen : boa rdepends to mime-support but mime-support does not get build Oct 15 11:59:00 steliosk, I can try that Oct 15 11:59:52 Crofron|home: cool. let me know if you do it Oct 15 12:00:58 Crofron|home: just make sure that mime-support is not allready build Oct 15 12:01:42 I will wipe tmp Oct 15 12:02:50 add boa to IMAGE_INSTALL? Oct 15 12:03:04 yes Oct 15 12:03:48 bother Oct 15 12:03:56 I seem to have deleleted build Oct 15 12:05:29 koen: how your autobuilder failed to build nmap 4.20 when it is not in repo? Oct 15 12:06:22 local diff Oct 15 12:06:34 ESP ... Oct 15 12:10:14 koen: so can I close 3148 3149 as invalid then? Oct 15 12:16:48 steliosk, we really need a script to add each package to minimal-image and build Oct 15 12:16:55 to look for broken dep's Oct 15 12:16:57 :) Oct 15 12:17:53 Crofron|home: how was the slogan during OEDEM... Don't be an elitist ? :) Oct 15 12:18:01 hrw: yes Oct 15 12:18:25 don't be so elitsit Oct 15 12:19:11 in the meantime, I can burn spare cycles on suspect packages Oct 15 12:19:21 ~emulate steliosk Oct 15 12:19:46 hrw : :) Oct 15 12:20:25 Crofton|home: for f in oe/packages/*/*.bb:do PKGNAME=`mangle-somehow-to-get-only-name'; rm -rf tmp; bitbake $PKGNAME>log-$PKGNAME;done Oct 15 12:20:44 Crofton|home: require really fast machine and eons to finish but its doable ;) Oct 15 12:20:57 the mangle bit is a pain Oct 15 12:20:58 Crofton|home: for f in oe/packages/*/*.bb:do PKGNAME=`mangle-somehow-to-get-only-name $f'; rm -rf tmp; bitbake $PKGNAME>log-$PKGNAME;done Oct 15 12:21:20 I have duall cores at home and school Oct 15 12:21:31 I only really need to use one at any given time Oct 15 12:21:39 so run autobuilds or tests on teh other Oct 15 12:21:40 Crofton|home: each time you will build toolchain... Oct 15 12:21:43 yes Oct 15 12:22:08 I need to look at u-boot source some Oct 15 12:22:17 and edit thesis.tex :) Oct 15 12:22:36 thesis.tex meets the size and weight requirements Oct 15 12:22:59 Crofton|home: for f in oe/packages/*/*.bb:do PKGNAME=`echo $f|cut -f1 -d"_"'; rm -rf tmp; bitbake $PKGNAME>log-$PKGNAME;done Oct 15 12:23:02 and I have a very terse style, but there is no extraneous information to skip :) Oct 15 12:23:03 hrw,Crofton|home : or... get SDK working :) Oct 15 12:23:09 Crofton|home: for f in oe/packages/*/*.bb:do PKGNAME=`echo $f|cut -f1 -d"_"`; rm -rf tmp; bitbake $PKGNAME>log-$PKGNAME;done Oct 15 12:23:12 now better Oct 15 12:23:17 steliosk: or packaged-staging Oct 15 12:23:21 I suspect the SDK will not heklp with build depends Oct 15 12:23:46 Crofron|home: no. but it will save you the time to rebuild the toolchain Oct 15 12:24:03 ok Oct 15 12:24:54 * steliosk looks at his todo list and starts crying Oct 15 12:28:13 hi cbrake Oct 15 12:28:19 hrw,Crofton|home : Ans still with this you are not going to find all dependency problems. The boa RDEPENDS to mime-support, and does not get build Oct 15 12:28:51 so the problem does not appear until runtime? Oct 15 12:28:59 lets see what builds Oct 15 12:29:12 and you can ask me to look at things, likely tomorrow :) Oct 15 12:29:38 Crofron|home: Well the image fails to build, because the package is missing. Build the package by hand and the image builds ok Oct 15 12:30:10 haven't actually tested though to see that it boots Oct 15 12:34:05 XorA: you were talking about packages were configure doesn't match configure.ac, nmap is one Oct 15 12:34:50 koen: heh heh, there are numerous Oct 15 12:37:12 steliosk: I use boa in my image and it builds with no problems -- distro=angstrom-2007, uclibc Oct 15 12:38:10 sakoman : Yes it builds ok, but if you add it as a dependency to an image then mime-support does not get build and image build fails Oct 15 12:38:41 sakoman: http://www.cherokee-project.com/benchmarks Oct 15 12:39:07 koen, do you own stock in cherokee :) Oct 15 12:39:33 no, I met alvaro in mexico and had some beers :) Oct 15 12:39:40 hehe Oct 15 12:39:43 even better :) Oct 15 12:39:59 koen : So your opinion is bias ;) Oct 15 12:40:01 beer is a powerful marketing tool Oct 15 12:41:13 steliosk: boa is in my image recipe as a IMAGE_INSTALL Oct 15 12:42:36 sakoman : Then its very probable that my local build is warped... Oct 15 12:43:15 koen: I've looked at (& tried) all of the embedded web servers Oct 15 12:43:46 boa has this advantage: Oct 15 12:43:48 root@sakoman-gumstix-custom:~$ ls -l /usr/sbin/boa Oct 15 12:43:49 -rwxr-xr-x 1 root root 64264 Dec 31 1969 /usr/sbin/boa Oct 15 12:44:09 And it doesn't pull in a bunch of libs Oct 15 12:45:01 boa is the webserver of choice for most of the embedded systems with minimal resources Oct 15 12:45:08 and works well Oct 15 12:45:30 steliosk: exactly -- it's tiny and it works Oct 15 13:23:02 hi likewise Oct 15 13:23:15 hi woglinde et al Oct 15 13:36:19 what's a 'fresh branch'? Oct 15 13:36:45 ah, the usual hh.org kernel breakage Oct 15 13:40:00 hrw: I'll try to look at the cm-x270 bug you pointed out some time this week. Oct 15 13:40:13 thx Oct 15 13:42:42 Good day all! Oct 15 13:43:09 I have some questions regarding uboot, subsequent to a good chat with Crofton, and an earlier one with hrw. Oct 15 13:43:27 I posted a bug recently due to compression.... Oct 15 13:44:12 gm Oct 15 13:44:25 did printascii go anywhere? Oct 15 13:44:35 Crofton pointed out this was a kernel issue, and that agrees with a tip from Koen. So I'm going down the path of figuring out what's wrong in teh decomp.. Oct 15 13:44:53 * Crofton|home I am still not 100% certain Oct 15 13:44:57 Crofton: thanks, no - i will be starting on that today. Oct 15 13:45:06 it depends on what the kernel assumes :) Oct 15 13:45:22 Crofton|home: no prob - I think your observation was dead on! Oct 15 13:45:27 my copy of the kernel book is at school Oct 15 13:45:47 I will try and make today u-boot day Oct 15 13:45:58 I have another project that needs some u-boot time also Oct 15 13:46:00 Crofton|home: which book (need more clues)? Oct 15 13:46:51 OS my question is.... If U-boot now jumps to kernel and decompresses.... Can we skip U-boot altogether in a production system??? Oct 15 13:47:34 basically, you want a super light wieght boot loader Oct 15 13:47:51 I suspect u-boot helps you with field upgrades though Oct 15 13:48:16 http://www.amazon.com/Linux-Kernel-Development-Novell-Press/dp/0672327201/ Oct 15 13:48:31 russf: I would left u-boot just in case. Oct 15 13:49:00 I suspect everyone has this idea at one point in their life Oct 15 13:49:04 Crofton|home: Yes. loghtweight is good. U-boot has tftp boot capability, which is attractive. What I'm trying to find out is what the Kernel sets up compared with u-boot. Oct 15 13:49:25 It's not a single decision, the need chagnes over time... Oct 15 13:49:44 but what does U-boot do for me that the kernel does not. And vice-versa Oct 15 13:49:52 thanks for book tip! Oct 15 13:50:11 u-boot/board/kb9202 Oct 15 13:50:23 is the unique stuff Oct 15 13:50:32 there will also be some processor specific stuff Oct 15 13:50:45 ok... Oct 15 13:50:56 so it made it upstream. I had not looked. Duh! Oct 15 13:51:06 (does anyone use this? : http://www.amazon.com/Embedded-Linux-Primer-Real-World-Development/dp/0131679848/ref=pd_sim_b_shvl_title_3/105-0148085-0999606) Oct 15 13:52:38 http://wiki.buici.com/wiki/Apex_Bootloader Oct 15 13:53:02 u-boot/cpu/arm920t Oct 15 13:54:57 apex looks interesting. Oct 15 13:55:16 I think we'll prolly end up with a bootloader and perhaps 2 images, like cbrake Oct 15 13:55:22 's suggestion. Oct 15 13:57:59 prpplague in #edev is an apex fan Oct 15 13:58:22 we should create an OE reading list Oct 15 13:58:30 nslu2 people often use apex iirc Oct 15 14:00:52 A reading list would be very useful. User manual is a good place for that. Oct 15 14:00:57 Or is that a sore topic? Oct 15 14:01:16 Crofotn|home : http://www.powerdeveloper.org/forums/viewtopic.php?p=9313#9313 Oct 15 14:02:02 Crofotn|home : Do you have time to work on new Efika ? I can ask them to send us one Oct 15 14:02:41 Crofton|home : Do you have time to work on new Efika ? I can ask them to send us one Oct 15 14:02:51 let me think :) Oct 15 14:03:11 if you have time to think, then you have time :) Oct 15 14:03:11 Crofton|home: where is that file you found? in OE, or on the kb disk? Oct 15 14:03:33 in the u-boot source Oct 15 14:03:53 I am looking in git Oct 15 14:04:00 but 1.2 should be similar Oct 15 14:04:21 thanks - got it... Oct 15 14:05:16 steliosk, do you have a link? Oct 15 14:05:18 bbiab Oct 15 14:09:05 * * OE Bug 3092 has been RESOLVED (INVALID) by Oct 15 14:09:07 * *  Same problem like Bug 2289 Oct 15 14:09:09 * * http://bugs.openembedded.org/show_bug.cgi?id=3092 Oct 15 14:09:17 * * OE Bug 3149 has been RESOLVED (INVALID) by Oct 15 14:09:19 * *  nmap-4.20-r0-do_configure Oct 15 14:09:21 * * http://bugs.openembedded.org/show_bug.cgi?id=3149 Oct 15 14:09:29 * * OE Bug 3148 has been RESOLVED (INVALID) by Oct 15 14:09:31 * *  nmap-4.20-r0-do_fetch Oct 15 14:09:33 * * http://bugs.openembedded.org/show_bug.cgi?id=3148 Oct 15 14:12:25 koen: you were right, it was a problem in autotools.bbclass Oct 15 14:32:15 koen : is the OE User Manual available in a single (editable) file ? Oct 15 14:35:41 steliosk: the 8314e and 8315e comes with VOIP ports. Oct 15 14:36:15 steliosk: I am interested in these boards also. Oct 15 14:39:48 likewise : Haven't looked at them Oct 15 14:41:40 steliosk: it's in the org.openembedded.documentation branch Oct 15 14:42:25 * Crofton|home needs to chat up a tech writer .... Oct 15 14:43:53 koen: darling, looks like u`re so in love with me so you just can't stand my presence.. it's ok my love, i forgive you Oct 15 14:44:04 morning Oct 15 14:51:05 hi chouimat Oct 15 14:55:37 XorA: Did try that xserver patch I posted with 1.4 btw? Oct 15 14:55:47 RP: not yet Oct 15 15:01:51 03koen 07org.oe.dev * r7c68840f... 10/ (4 files in 3 dirs): linux-libc-headers: add 0.23 Oct 15 15:02:27 gah Oct 15 15:02:29 could not connect to server (http://svn.openmoko.org) Oct 15 15:03:00 bye Oct 15 15:04:23 steliosk, minimal-image + boa builds for me Oct 15 15:12:15 does anyone use the preferred versions X11R7 inc files from oe? to build xserver-kdrive ???? Oct 15 15:21:40 hi likewise. Oct 15 15:22:18 Hi HopsNBarley Oct 15 15:26:41 Crofton|home : Thanks. so my local build needs rebuilding Oct 15 15:27:01 it is likely Oct 15 15:27:23 * Crofton|home has been watching the chumby magic 8 ball application Oct 15 15:27:58 gilligan_: did you play with your board's serial port ? Oct 15 15:32:32 "play with it" ? Oct 15 15:32:36 use it Oct 15 15:33:04 there is some random crap on ttyS1 here Oct 15 15:33:18 I think this is the kernel login console, with another bitrate Oct 15 15:35:52 this would explain why it doesn't appear on ttyS0 Oct 15 15:35:56 cbrake: there ? Oct 15 15:36:05 nud: hello Oct 15 15:36:24 cbrake: ^^^^^ Oct 15 15:38:17 does anyone remember what's the bitbake command to pre-fetch all source files ? Oct 15 15:45:31 -c fetch #image-foo does not work :( Oct 15 15:46:47 nud: I've not noticed any random chars on ttyS1 Oct 15 15:47:19 nud: the only thing I've observed is the login does not work with my ucblic minimal images, but does work with glibc/console-image Oct 15 15:47:26 I just saw my /etc/inittab was referring ttyS1 instead of ttyS0 Oct 15 15:47:27 nud: have not had time to debug the uclibc issue Oct 15 15:47:56 nud: have you seen the latest changes in cm-x270.conf for setting the serial console? Oct 15 15:48:12 Crofton: Is there a quick way to turn kernel compression OFF during a build? Oct 15 15:48:44 nud: CMX270_CONSOLE_SERIAL_PORT Oct 15 15:50:30 gah..does anyone here have experience with setting up bugzilla? Oct 15 15:51:53 gilligan_: not since I've started using TRAC Oct 15 15:52:24 cbrake: I didn't but this is an old image Oct 15 15:52:38 cbrake: I'm experiencing crashes with the latest image Oct 15 15:52:50 cbrake: the fact that I have exactly 0 experience with apache makes this whole thing rather annoying... :) Oct 15 15:53:08 at ld-linux level (it doesn't like my debugging symbols) and at the pwc driver level (for .22 and .23) Oct 15 15:53:13 this is unusual Oct 15 15:53:39 steliosk: -c fetchall Oct 15 15:53:58 * stappers_ is following http://www.openembedded.org/wiki/GettingStarted and stuck at the monotone pull, Oct 15 15:54:14 mtn: misuse: database /usr/src/opembed/OE.mtn is laid out according to an old schema Oct 15 15:54:41 my monotone is 0.36 (as in Debian unstable ) Oct 15 15:54:45 also, gdb get killed Oct 15 15:56:21 How to get a OE.mtn with new schema? Oct 15 15:56:52 stappers_: Run the command to upgrade the database (which it prints in the diagnostic). Oct 15 15:57:48 stappers_: http://www.openembedded.org/snapshots/OE-this-is-for-mtn-0.36.mtn.bz2 Oct 15 15:58:28 * stappers_ wgets the new mtn.bz thank you XorA Oct 15 15:58:33 the "If you have troubles in building monotone, there are two ways out:" section in GettingStarted tells you what to do Oct 15 15:59:46 koen : thanks Oct 15 16:00:06 time to go home... cu all tommorow Oct 15 16:00:23 koen: that section makes bugger all sense at all Oct 15 16:00:30 cu steliosk Oct 15 16:01:17 XorA: the GettingStarted needs to get pruned again Oct 15 16:01:30 people keep adding crap to it that makes it incomprehensible Oct 15 16:02:26 cbrake: I'll try to provide you the bitbake files I use for my stuff Oct 15 16:02:34 so you can check by yourself Oct 15 16:03:10 koen, are there any problems with compiling python extensions on 64bit for 32bit (i.e. arm) Oct 15 16:06:38 koen, i'll be back later Oct 15 16:06:47 how to get a wiki account? Oct 15 16:07:12 stappers_: you dont need an account Oct 15 16:15:02 cbrake: CMX270_CONSOLE_SERIAL_PORT <-- is it to be put in local.conf ? Oct 15 16:16:27 nud: yes Oct 15 16:17:36 Shoragan: hrw|gone is the python 64->32 bit expert Oct 15 16:20:31 RP: http://tuxrecife.blogspot.com/2007/10/xomap-3x-more-fast.html Oct 15 16:21:52 koen: Nice although out devices didn't seem to use that data path irrc? Oct 15 16:21:54 iirc? Oct 15 16:34:06 cbrake: here is the faulty bakefile Oct 15 16:34:59 cbrake: http://rafb.net/p/WRzW6C57.html Oct 15 16:35:27 the vidtest stuff is crashing when used in conjunction with the pwc cam driver (it didn't before) Oct 15 16:35:46 I'll try to get you the "crashing .debug" bitbake file with an example program as well asap Oct 15 16:41:46 03rwhitby 07org.oe.dev * rc8012b91... 10/ (3 files in 2 dirs): slugos-{packages,native}: Fixed building empty packages Oct 15 16:57:12 good evening Oct 15 17:27:08 Crofton|home: This looks like a fix for the u-boot issue - seems like it may have been offered to upstream, but was rejected. Who knows. This is my starting point. I'll keep you up to date. I should really get on with the compression issue, but I'm in this for another hour.... Oct 15 17:33:40 what looks like a fix? Oct 15 17:34:07 if you have an issue, I can try and create a patch we can use in oe for the short term Oct 15 17:34:23 I am actually practicing my u-boot skillz at the moment Oct 15 17:43:45 OE fails to build on a powerpc host -- did anybody notice that? Oct 15 17:47:46 Crofton|home: DUH! http://www.nabble.com/Compile-Error-:-kb9202-board!-t3371246.html SORRY! Oct 15 17:47:58 doesn't koen build on ppc? Oct 15 17:48:21 Crofton|home: note the final comment about ( or ) Oct 15 17:48:38 where should I apply these patches?? Oct 15 17:49:53 Crofton|home: In my temp tree I see kb9202.h under a u-boot-1.2.--r**2** dir Oct 15 17:50:16 basically, in the u-boot dir do patch -p1 < fileWithPath in ti Oct 15 17:50:28 Crofton|home: but also under git in an openmoko ... Oct 15 17:51:13 both the openmoko gits are r0, and r1. So did the r2 come from a patch file? Oct 15 17:51:30 we need to get the path into a file and add it to u-boot bb file for the kb9202 machine Oct 15 17:51:42 In .../tmp Oct 15 17:51:44 ? Oct 15 17:51:45 no Oct 15 17:51:50 need to edit the bb file Oct 15 17:51:53 right, that was my though. Oct 15 17:51:54 t Oct 15 17:51:55 ok Oct 15 17:52:03 it may be easiest to do this outside OE atm Oct 15 17:52:09 so create a patch in the .bb tree. Oct 15 17:52:12 ok. Oct 15 17:52:24 not really ... Oct 15 17:52:42 Perhaps I'll kluge that to a nice patch file and use it under tmp for a bit. Then decide the details later. Oct 15 17:52:48 grr, I am in the middle of something here Oct 15 17:52:54 kludge around a bit Oct 15 17:53:08 maybe try apllying by hand to make sure your patch file is good Oct 15 17:53:11 OK. No prob. I'll stumble on for a bit... and catch me l8r Oct 15 17:53:14 right. Oct 15 17:53:18 I know what to do, just short on time :) Oct 15 17:53:38 I do too, for the basic. You can help me get into OE when it runs. Oct 15 17:53:40 Ok? Oct 15 17:53:43 wirelessdreamer: http://oe.pastebin.com/m1255755d Oct 15 17:53:46 right Oct 15 17:53:55 enjoy your typing... Oct 15 17:53:58 if it works, I'll try sending upstream also Oct 15 17:54:55 hi crofton Oct 15 17:55:15 Crofton: cool.. thx. I'll keep u in the loop Oct 15 17:56:16 gm woglinde Oct 15 17:56:37 I have a git tree for u-boot pulled atm Oct 15 17:56:49 * Crofton wonders if the patch is already in Oct 15 17:57:28 nope, not in got Oct 15 17:57:31 er git Oct 15 17:57:57 ~emulate crofton Oct 15 17:57:58 * ibot don't be so elitist! Oct 15 17:58:15 RP: the n800 is one of our devices, right? :) Oct 15 17:59:10 koen: hi. wirelessdreamer just said you build on a ppc host. do you get this, too? http://oe.pastebin.com/m1255755d Oct 15 17:59:32 I did, till I donated my powerbook to my girlfriend Oct 15 17:59:40 steliosk is our ppc expert Oct 15 17:59:56 rofl Oct 15 18:00:41 two hours idle.. Oct 15 18:18:03 Crofton|home: Not surprised in not in - very messy patch - yuk! Oct 15 18:18:18 lol Oct 15 18:18:23 cobol for eclipse Oct 15 18:18:28 RP: but AIUI jitblt would make that patch obsolete Oct 15 18:18:34 Have ruber gloves on and doing cleanup. Oct 15 18:19:50 Crofton: going for a quick dinner. Working direct on sources, since that email patch had 2 much garbage and choked patch. Oct 15 18:22:56 that is what I was afraid iff Oct 15 18:26:04 wtf micrsoft will patent handhelds-ui? Oct 15 18:26:43 XorA: http://blogs.gnome.org/dcbw/2007/10/15/networkmanager-07-is-the-new-chuck-norris/ Oct 15 18:31:07 re network manager: as mexican subtitles say, "¡Santo Dios!" Oct 15 18:34:04 * * OE Bug 3158 has been created by ar(AT)oszine.de Oct 15 18:34:06 * * Can't build linux-rp_2.6.22.bb Oct 15 18:34:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3158 Oct 15 18:36:27 any boot from nand with uboot gurus about? Oct 15 18:36:46 "gurus" plural?? Oct 15 18:36:53 guru will do Oct 15 18:37:57 I have a u-bbot in my DV board that has a possibly working nand system Oct 15 18:38:34 kewl Oct 15 18:50:10 uh!!! who added the empty-dir packages/libmulti? Oct 15 18:57:14 koen: I don't think it even helps the n800 if you have subpixel rendering turned on or something Oct 15 18:57:32 koen: At least in the tests Ross did, poky saw no benefit Oct 15 18:57:55 a Oct 15 18:57:57 a Oct 15 18:58:00 oops Oct 15 18:58:12 pig on the keyboard? Oct 15 18:58:19 it would explain why I saw no difference on my ipaqs with the same colour space Oct 15 18:58:23 Crofton: sneeze Oct 15 18:58:26 koen, I have u-boot going with NAND suport for the SFF SDR Oct 15 18:58:32 now to try a kernel Oct 15 18:58:36 Crofton: cool Oct 15 18:58:46 yeah Oct 15 18:59:08 so far I am pleased with the DV support in u-boot git Oct 15 18:59:22 do any of the media players in the oe database have good playlist management support for movies and music? Oct 15 19:02:05 gah Oct 15 19:02:18 linux-bin r3527 works, 3528 doesnt' Oct 15 19:10:46 hi khem Oct 15 19:19:32 Crofton: Back. No pressure. Still hand fettling that file. Oct 15 19:19:44 grrrr glib-2.0_2.12.12 don't build with uclibc ... Oct 15 19:21:07 gremlin hm hu why? Oct 15 19:21:48 gremlin will test myself Oct 15 19:23:15 woglinde: howdy Oct 15 19:28:22 ok, just made a good start with a bitbake file for Ruby. Oct 15 19:28:32 now I need to pass extra options to configure Oct 15 19:28:53 kero you dont know how? Oct 15 19:29:09 any thoughts on a nand erase faliure, MTD get bad blork failed Oct 15 19:29:21 yeah I think there is a typo in u-boot Oct 15 19:30:29 mm, from other .bb files, EXTRA_OECONF Oct 15 19:30:37 kero right Oct 15 19:32:04 and for now let's switch off rm_work :) Oct 15 19:34:23 * Kero got pretty far on the first few runs. impressive! Oct 15 19:35:02 * cbrake can't get any pointer input working with kdrive 1.4 Oct 15 19:35:08 anyone else using kdrive 1.4? Oct 15 19:36:19 * pH5 lifts hand Oct 15 19:36:36 you have to specify the mouse input driver. -mouse tslib for me Oct 15 19:38:26 hi lrg Oct 15 19:38:42 hey woglinde Oct 15 19:38:44 pH5: great, touch how works -- USB mouse still does not -- can you have both? Oct 15 19:38:45 got everything, pretty much. now I only need to copy contents-per-package from my own files. someone did a good job with all automagic configuration for OpenEmbedded! cool! Oct 15 19:39:24 kero as long as the software uses sane autoconf Oct 15 19:39:26 cbrake: I think multiple input devices should be possible, but I didn't try that yet. Oct 15 19:39:29 but there arent much Oct 15 19:40:03 woglinde: ruby compiles on many platforms. it's about as sane as you get, I guess. Oct 15 19:40:33 kero *g* I never said ruby has bad usage of autotools Oct 15 19:41:05 woglinde: I've crosscompiled ruby for familiar years ago. and applied that on the openembedded tree recently. Oct 15 19:41:32 I'm impressed that the scripts do what I had to find out by hand, and more. Oct 15 19:43:18 I had multiple packages from one source; so does oe. Oct 15 19:43:42 mm, I bet oe can set depends per package, too? Oct 15 19:44:40 kero yes Oct 15 19:44:57 look at packages/midpath Oct 15 19:45:12 DEPENDS_package and/or RDEPENDS_package it seems Oct 15 19:46:15 woglinde: nice set of example files. Oct 15 19:46:53 cbrake: can you "kill -USR1" the Xfbdev process twice and get it back working, or does it hang for you on 'resume'? Oct 15 19:47:16 kero hm yes, but my problem is that I dont know how to make serval configs from one source Oct 15 19:47:24 RP: thanks for your hints about the kdrive 1.4 resume issue, but your patch does not help. Oct 15 19:48:15 pH5: I have not tried kdrive 1.4 with resume Oct 15 19:48:27 woglinde: like PACKAGES = "a b c" \n FILES_a = "${bindir}/a.exe" ... ? Oct 15 19:48:31 pH5: system I'm working on does not suspend/resume Oct 15 19:48:41 or do you mean something else? Oct 15 19:49:24 pH5: ok, it was worth a shot I guess :/ Oct 15 19:49:28 kero CONFFILES_${PN} bla Oct 15 19:49:37 hm maybe rp knows Oct 15 19:49:44 pH5: I've just replied to your led mail btw, still catchiing up after OEDEM :/ Oct 15 19:50:11 rp how can I made serval packages with same CONFFILES_${PN} from one spurce? Oct 15 19:50:20 source Oct 15 19:50:21 cbrake: you don't need to suspend. just send a USR1 signal to the Xfbdev process to have it let go of console/input devices and send USR1 again to have it enable them again. Oct 15 19:50:30 hit hangs here during LinuxKeyboardEnable Oct 15 19:50:34 s/hit/it/ Oct 15 19:50:52 RP: no problem, I'm only just answering to your kdrive 1.4 mail :) Oct 15 19:51:36 woglinde: I don't understand the question exactly :/ Oct 15 19:51:51 rp I have one source Oct 15 19:52:08 from the source I generate many packages Oct 15 19:52:24 some of the packages share the same config-file but with other options Oct 15 19:52:58 so I used CONFFILES = bla to asked the user at install time which config he will use Oct 15 19:54:16 but with one source I seems impossible Oct 15 19:54:26 woglinde: This is tricky since the files should all end up with the same name, right? Oct 15 19:54:31 rp right Oct 15 19:55:04 woglinde: I think the only way would be to patch the way OE's packaging classes work Oct 15 19:55:20 okay Oct 15 19:55:47 for now I unpack the source in every package Oct 15 19:58:16 RP: k, thanks for the clarification about the led naming scheme. Oct 15 19:58:52 woglinde: now I understand what you mean. seems to create conflicts, indeed. Why do you want it? Oct 15 19:59:23 kero with midpath you can use serval backends for graphic or sound Oct 15 19:59:37 and you can install them without conflicts Oct 15 19:59:56 is that in subdirs? or always a single package? Oct 15 20:00:10 mostly one lib Oct 15 20:03:51 pH5: trying ... Oct 15 20:07:19 w/ shlibdeps, bitbake is smarter in finding dependencies than I am. Oct 15 20:11:30 pH5: now I'm left with a defunct Xfbdev process ... Oct 15 20:18:28 cbrake: heh 0wned :P Oct 15 20:18:45 so it's not something specific to my setup but a real kdrive bug. Oct 15 20:19:49 pH5: ok, I tried again after a reboot Oct 15 20:20:02 pH5: first kill -USR1, the screen goes black Oct 15 20:20:28 pH5: the 2nd, I see parts of my app return on the screen, but thats about it Oct 15 20:20:46 pH5: processes are still running Oct 15 20:21:07 ok, same here. it magically works again if I just remove the read() call from LinuxKeyboardEnable() in hw/linux/keyboard.c Oct 15 20:21:09 I'm sure I've seen something like this with earlier kdrives too... Oct 15 20:21:14 Just very intermitently... Oct 15 20:21:48 pH5: and its opened for nonblocking IO? Oct 15 20:22:24 RP: yes, that's why I'm so confused Oct 15 20:24:45 hm, that DEPENDS_irb1.8 is not used in the ./control file in the resulting ipkg. Oct 15 20:25:16 Kero: It needs to be RDEPENDS_irb1.8 Oct 15 20:25:38 well, I'm not sure about the 1.8 bit, that looks suspect too Oct 15 20:25:51 works for FILES_irb1.8 Oct 15 20:26:36 Kero: I guess it works for glib-2.0 too :) Oct 15 20:27:00 Kero: It just struck me as a little odd at first glance... Oct 15 20:27:07 yeah, me too. Oct 15 20:27:26 but I would've screamed earlier this sevening, had it not worked :) Oct 15 20:28:11 can I tell bitbake -c to only re-install and re-package? Oct 15 20:28:46 per _package_ DEPENDS are bogues Oct 15 20:29:04 you probably want RDEPENDS Oct 15 20:32:47 koen is there a reason why libiconv isnt build with -Os? Oct 15 20:36:56 gremlin[it] glib-2.0 builds here fine with uclibc Oct 15 20:37:27 RDEPENDS works. Oct 15 20:37:49 woglinde ok ... so as i suppose i have to wipe and rebuild all ... :) Oct 15 20:37:58 ah Oct 15 20:38:11 gremlin can you show me the error? Oct 15 20:55:50 03likewise 07org.oe.dev * rcfbfea81... 10/ (3 files in 3 dirs): mpd: Added sane revision. Oct 15 21:02:07 ynezz: the "svn co -r ${AUTOREV} ..." problem you had some days ago ... i'm now hitting it too. turns out that the data.getVar("SRCREV", d, 0) is called with 0 there for a reason, the comments just before that indicate a nasty hack. which fails. Oct 15 21:02:51 :) Oct 15 21:05:19 Crofton|home: Will try building next... diffs so far... http://pastebin.com/m1c7764e5 Oct 15 21:06:09 If you set the path to include tmp/cross/bin (not quite the right path) Oct 15 21:06:17 you can use the cross compiler from OE Oct 15 21:06:38 you will also need to adjust the CROSS_COMPILE variable in uboot Oct 15 21:06:44 I need to go home now Oct 15 21:07:09 I'll try and spend some cycles on this tomorrow, time permitting Oct 15 21:07:11 bye crofton Oct 15 21:07:15 gn Oct 15 21:08:17 bye Crofton. Thanks! Oct 15 21:08:32 Henryk: I've been running into AUTOREV problems as well ... Oct 15 21:09:13 the problem is line 76 and 77: http://svn.berlios.de/wsvn/bitbake/trunk/bitbake/lib/bb/fetch/svn.py?op=blame&rev=0&sc=0 Oct 15 21:09:37 When SRCREV="${AUTOREV}" this fails Oct 15 21:09:50 where are all those .debug directories coming rom, suddenly? Oct 15 21:10:05 kero from the scripts Oct 15 21:11:27 and why do the scripts complain about them now... (non debug package contains.debug ...) Oct 15 21:12:01 I've been thinking about that for the last 15 minutes and can't come up with a correct solution that wouldn't include some slightly heavy bitbake modification. (e.g. a expand() version which only expands once, or only expands variables but not python code, would be nice. or even better: expand everything except the following variables/function calls) Oct 15 21:14:36 hmm, or maybe ... Oct 15 21:16:02 RP: why is line 77/78 necessary at all? why not just ud.revision = data.expand(rev, d)? Oct 15 21:16:04 kero make a sepperate -dbg packge and include the .debug/ stuff there Oct 15 21:17:11 woglinde: but why did bitbake not have problems before? I've only added packages with files... Oct 15 21:17:35 well, and directories. Oct 15 21:17:41 kero normaly bb trys to build the -dbg on it owns guesses but that fails sometime Oct 15 21:18:38 hmf Oct 15 21:18:53 the -dbg is not the same as -dev is it? Oct 15 21:20:01 right Oct 15 21:20:18 dev should contain headers and libs to link against e.g .a Oct 15 21:21:13 Crofton|home: well, it builds.... Oct 15 21:23:14 cool Oct 15 21:23:25 does it boot? Oct 15 21:26:31 U-boot it boots Wahhoo! Oct 15 21:27:05 how about the kernel, has this solved the problem Oct 15 21:27:11 exactly as in the pastebin page - it works! OK.... Calming down a bit, it ran, and it loaded a kernel! Oct 15 21:27:24 with the kernel decompressor Oct 15 21:27:27 now I need to try latest kernel. Oct 15 21:27:32 ok Oct 15 21:34:28 woglinde: ok, tha work-around works (of course). redistributing files over packages a bit (made some errors...) Oct 15 21:34:37 so how do I ge the result into OE? Oct 15 21:34:53 as in, the .bb file int the official distro? Oct 15 22:01:21 bye Oct 15 22:23:25 Henryk, cbrake_away: I've posted on the OE list about this in the past Oct 15 22:23:58 Henryk: The problem is we need SRCREV = "45" to work too Oct 15 22:24:43 hmm, data.expand would just leave "45" alone, wouldn't it? Oct 15 22:26:14 Crofton|home: ;) Oct 15 22:26:38 Crofton|home: muahhah! Oct 15 22:35:52 Henryk: hmm... Oct 15 22:44:22 Henryk: I don't remember why that code ended up like it did to be honest. There was a good reason for it... Oct 15 22:44:43 Crofton|home: Final score... The updated u-boot works. But this has not changed the kernel decompress behavio(u)r, which still gives "invalid compressed format (err=2)". I'm not surprised by that - we (and koen) thought it was kernel related. Oct 15 22:45:00 RP: hmm, true. (also the comment doesn't fit the code exactly) Oct 15 22:45:24 Crofton|home: I'll look at the kernel tomorrow afternoon. Oct 15 22:51:57 russf: what is the problem with decompressing kernel Oct 15 22:52:28 khem: not sure yet - could be just me... Oct 15 22:52:44 doing a quick test at the moment... Oct 15 22:52:53 are you using uimage ? Oct 15 22:52:58 y Oct 15 22:53:23 ok and uboot is not loading it properly ? Oct 15 22:53:27 just updated to latest u-boot - no diff. Oct 15 22:54:08 well, i (we) think the kernel starts decomp then fails. Let me get back to you.. I'll know more in 10 mins Oct 15 22:54:08 russf: what is your u-boot env Oct 15 22:54:20 oh i see Oct 15 22:54:26 sorry - I'll get back soon. Oct 15 22:54:26 ok Oct 15 22:54:47 how do I prevent bibake from exporing CFLAGS and CPPFLAGS ? Oct 15 22:54:55 +t+t Oct 15 22:58:35 CFLAGS="" or CFAGS= Oct 16 01:35:26 mtn: 9.7 M | 1.8 M | 8332/8332 | 2046/2046 =) Oct 16 01:35:55 2046? Good grief! Oct 16 01:36:16 took forever Oct 16 01:36:47 WOW Oct 16 01:40:00 * CoreDump kicks off an angstrom build **** ENDING LOGGING AT Tue Oct 16 02:59:57 2007