**** BEGIN LOGGING AT Wed Dec 13 02:59:57 2006 Dec 13 07:39:02 good morning all Dec 13 07:45:23 morning Dec 13 08:20:24 hey koen Dec 13 08:21:18 hey likewise Dec 13 08:32:52 morning Dec 13 08:33:34 morning Dec 13 08:36:39 XorA: firefox/arm built Dec 13 08:38:17 * hrw|work is building opie-image/collie Dec 13 08:38:25 does it run ? Dec 13 08:38:25 :) Dec 13 08:38:38 * Genesis 's gnash build , run but no window create :) Dec 13 08:39:28 koen, : what about performance ? Dec 13 08:39:40 koen: just find that SIGILL now Dec 13 08:42:47 XorA: is is possible be to do 'mv firefox_2.0.bb mozilla-2.0.inc' and make a thunderbird, sunbird and minimo 2.0 .bb that only do 'require moz-2.0 ; --enable-app=?'? Dec 13 08:43:35 koen: no idea, I didnt do the original mozilla stuff, that would be nice though Dec 13 08:44:45 if the versions match we can have each bb generate libnspr libxpcom lib packages Dec 13 08:45:11 they will overwrite eachother, but that shouldn't matter Dec 13 08:45:46 koen: or a mozilla-2.0.bb that emits firefox, thunderbird, suncal + libs all modular packaged Dec 13 08:49:33 indeed Dec 13 08:49:46 --enable-application=suite iirc Dec 13 08:51:58 modular would also be nice for ipkg, as sometime installing firefox uses ALL ram and kills gpe Dec 13 08:55:34 * koen has 60MB swap for that Dec 13 08:55:58 and for plugins , should we install all of them in /opt/netscape/plugins or dynamically in a /usr/lib/ firefox-1.0.7/plugin ? Dec 13 08:56:43 Does anyone have "works-ok" experience with gdb on the host with gdbserver on the target, for arm targets? I am getting remote communication errors I cannot resolve. Dec 13 08:56:43 /opt should die Dec 13 08:57:04 Genesis: /usr/lib/mozilla/plugins/ Dec 13 08:57:54 never mind: seems like there was a dangling TCP connection timing out. It works now. Dec 13 09:01:06 XorA: I think the sigill you're seeing in ff is related to the one i'm seeing in minimo Dec 13 09:02:32 * koen adds a comment to #1600 Dec 13 09:03:04 XorA: try compiling that testcase (CFLAGS += "-lxpcomm_glue) Dec 13 09:05:13 hrw|work, : noticed Dec 13 09:05:36 Genesis: iirc all gecko based looks there Dec 13 09:05:39 koen: Im just making ANgstrom my main image on terrier Dec 13 09:06:09 XorA: cool Dec 13 09:06:55 and I might have sourced a cheap collie Dec 13 09:07:16 You're going to fix gcc to make eabi work on armv4? Dec 13 09:07:17 hrw|work, : and what do you think we should depend when compiling mozilla/firefox/iceeasel/whatever plugins ? Dec 13 09:07:28 koen: I thought it did Dec 13 09:07:42 hrw|work, : i try to introduice virtual/mozilla :) Dec 13 09:08:04 XorA: armv4*t* Dec 13 09:08:59 Genesis: debian has xulrunner for it Dec 13 09:09:06 XULRunner is a single "gecko runtime" that can be used to bootstrap Dec 13 09:09:06 multiple XUL + XPCOM applications that are as rich as Firefox and Dec 13 09:09:06 Thunderbird. Dec 13 09:11:17 but it's on the road Dec 13 09:11:36 i already develop on the xul framework , it's fun Dec 13 09:11:45 morning all Dec 13 09:11:54 ( svg + xul + js application for industrial supervision ) Dec 13 09:11:58 yo ade|desk Dec 13 09:15:14 hi ade|desk Dec 13 09:20:43 hrw|work: how is your business going, can we buy shares soon ? Dec 13 09:22:41 ade|desk: it is progressing Dec 13 09:30:47 morning all Dec 13 09:30:53 hey RP Dec 13 09:31:10 hey RP Dec 13 09:31:33 h Richard Dec 13 09:33:37 hi, all Dec 13 09:33:43 RP, gm! Dec 13 09:34:29 how could I add module package which needs to be compiled against some kernel? Dec 13 09:34:46 slapin_nb: Look at orinoco or hostap Dec 13 09:38:29 hello everyone ! Dec 13 09:40:40 it's getting better day after day ... Now INIT starts (doesn't make libc error anymore) but I get nothing on my serial port Dec 13 09:41:30 how can I tweak some config files from angstrom distribution (like inittab or some /dev files or udev config files ???) Dec 13 09:45:17 03hrw 07org.oe.dev * r03ee5c18... 10/ (1 packages/gnutls/gnutls.inc packages/gnutls/gnutls_1.6.0.bb): gnutls 1.6.0: package libgnutlsxx into separate package - close #1678 Dec 13 10:05:22 03hrw 07org.oe.oz354x * r096afe74... 10/ (15 files in 5 dirs): gnutls: added 1.6.0 from .dev, dropped older Dec 13 10:05:26 03hrw 07org.oe.oz354x * raf222d01... 10/ (4 files in 2 dirs): libgcrypt: update to 1.2.3 Dec 13 10:05:31 03hrw 07org.oe.oz354x * r88fe687a... 10/ (1 packages/libgcrypt/libgcrypt_1.2.3.bb): libgcrypt: update to 1.2.3 Dec 13 10:09:11 vlo, : look at packages/basefiles for example Dec 13 10:11:57 03hrw 07org.oe.oz354x * r33969de8... 10/ (1 packages/otpkeygen/otpkeygen_1.3.0.bb): otpkeygen: rename otpkeygen binary - close #1663 Dec 13 10:12:11 03hrw 07org.oe.dev * ra7185dca... 10/ (1 packages/otpkeygen/otpkeygen_1.3.0.bb): otpkeygen: rename otpkeygen binary - close #1663 Dec 13 10:14:46 Genesis: thank you Dec 13 10:17:34 NOTE: package opie-image-1.0-r21: task do_rootfs: started Dec 13 10:17:36 finally Dec 13 10:20:41 │-rw-r--r-- 1 hrw hrw 22152K Dec 13 11:46 opie-image-collie-20061213104152.rootfs.jffs2 � Dec 13 10:20:45 hi florian_kc Dec 13 10:23:02 hi all Dec 13 10:33:59 03hrw 07org.oe.dev * r745bb2e7... 10/ (1 conf/distro/openzaurus-unstable.conf): openzaurus-unstable: fix binutils-cross-sdk (2.16) and wpa-supplicant (0.5.5) versions Dec 13 11:03:18 03hrw 07org.oe.dev * re9b48012... 10/ (1 packages/images/opie-image.bb): opie-image: removed familiar feed stuff Dec 13 11:03:33 03hrw 07org.oe.dev * r7f6e0e3f... 10/ (1 packages/tasks/task-base.bb): task-base: 2.4 machines need modutils-depmod Dec 13 11:04:48 03hrw 07org.oe.dev * rb36b81d6... 10/ (1 packages/tasks/task-opie.bb): task-opie: switch to bluez-utils due to fact that bluez-utils-nodbus got removed Dec 13 11:09:00 03hrw 07org.oe.dev * r8e3bc2d3... 10/ (1 packages/images/opie-image.bb): opie-image: install extra stuff on machines with more flash only. Supported are spitz/akita now - other must be added Dec 13 11:14:09 ~lart ipkg-make-index Dec 13 11:14:10 * ibot squishes ipkg-make-index like a bug Dec 13 11:14:55 XorA: ask psokolovsky for his python version Dec 13 11:15:04 that should be 6 times faster Dec 13 11:15:24 so you can hit the same bugs 6 times faster Dec 13 11:15:48 koen: actually this is an ipkg bug, it checks md5sum even if the package is installed as ipkg install blah.bb Dec 13 11:16:14 s|bb|ipk| Dec 13 11:16:21 koen: yeah that Dec 13 11:16:42 03hrw 07org.oe.dev * r21188e8c... 10/ (1 conf/distro/include/openzaurus.inc): openzaurus: we do not want ipsec Dec 13 11:17:39 who added dbus recipes? Dec 13 11:18:34 koen: dbus/ directory require CLEANING Dec 13 11:18:34 the original ones? Dec 13 11:18:36 dbus.inc Dec 13 11:18:41 creation etc Dec 13 11:18:48 I think pb_ Dec 13 11:18:57 koen: 0.9x 1.x are yours Dec 13 11:19:12 cp foo.old.bb foo.new.bb Dec 13 11:19:30 pb_ made the original ones, ph5 and I updated them Dec 13 11:19:36 koen: its shit not work Dec 13 11:19:46 I want to granulate dbus packaging Dec 13 11:19:56 but I *refuse* to do cleaning in it Dec 13 11:19:56 go ahead Dec 13 11:20:08 and without such clean I refuse to touch it Dec 13 11:20:17 6 recipes to change Dec 13 11:20:30 just beacuse someone decided to use cp instead of brain Dec 13 11:20:39 re Dec 13 11:21:06 -bash: brain: command not found Dec 13 11:22:42 just like I thought Dec 13 11:23:47 later all Dec 13 11:28:34 hi greentux Dec 13 13:19:30 in which file should "ideally" SERIAL_CONSOLE be defined ? Dec 13 13:20:33 I have it in my machine config file.. but don't take my word for "ideally", I just checked where other configs defined it and did the same Dec 13 13:20:51 * Jin^eLD is still an OE n00b Dec 13 13:20:58 hum, yes, it's in my machine config file too thx Jin Dec 13 13:21:19 * vlo is still an OE n00b too, but I'm trying to learn ;) Dec 13 13:21:25 yeah hehe, me too Dec 13 13:21:27 :) Dec 13 13:22:12 it also seems to make sense to have it there - after all it depends on your hardware - machine - which serial settings you have to use Dec 13 13:23:41 yeah exactly ... I confused with a CMDLINE_CONSOLE which I have in the kernel recipe Dec 13 13:26:11 hi paul Dec 13 13:26:21 Hi! Dec 13 13:27:09 hmm, any idea how to disable udev support, so it does not get into my rootfs? or does task-base automatically add this if I have kernel26 defined? Dec 13 13:27:18 Jin^eLD: auto Dec 13 13:27:31 hrw|work: huh? Dec 13 13:28:36 hrw|work: could you explain a little further? :) Dec 13 13:28:58 Jin^eLD: task-base automatically add udev if kernel26 is defined Dec 13 13:29:24 aah, ok.. so if I do not want it but want other kernel26 features - I would have to reproduce those features manually in my machine and distro configs and take out the kernel26 thing, right? Dec 13 13:29:35 udev requires tmpfs and tmpfs does not seem to work on my HW Dec 13 13:30:04 Jin^eLD: tmpfs need to be on in kernel Dec 13 13:30:28 it is, but then I can not boot because processes get killed with out of mem Dec 13 13:30:36 8mb ram is kind of... not enough I guess :) Dec 13 13:30:43 Jin^eLD: 8m ram... AU Dec 13 13:30:48 yes... :) Dec 13 13:31:05 I know.. but I am misusing some hardware from another project for my purposes Dec 13 13:31:24 and I am bad at soldering, so... :) Dec 13 13:33:19 btw, another question... can I point the SRC_URI to a file with an absolute location? usually some path is prepended to what I specify under "file://".. I'd like to have an absolute path there, just for testing my work-in-progress Dec 13 13:36:33 morning Dec 13 13:36:36 hi Dec 13 13:58:39 hey Dec 13 14:04:53 hey mickeyl Dec 13 14:06:00 koen, I received my email from fedex Dec 13 14:06:45 hi mickeyl Dec 13 14:07:37 Crofton: yay! Dec 13 14:15:04 psokolovsky_: check recent opie-image changes Dec 13 14:16:17 what's the best way to sync between my local svn rep and the OE one ? Dec 13 14:17:12 Ifaistos: not using svn would be the best option Dec 13 14:17:36 koen : unfortunatelly is not a option for us here :( Dec 13 14:18:30 koen : we used svn/trac extensivly, even for windows (don't laught) project Dec 13 14:20:30 trac also has a monotone plugin Dec 13 14:21:50 koen : so in short... no easy way to do it ? Dec 13 14:22:02 correct Dec 13 14:23:04 koen : btw did you get the Efika board ? Mine is stuck (as expected) in sutoms Dec 13 14:23:17 koen : sutoms=customs Dec 13 14:23:43 ~laft fedex Dec 13 14:23:50 ~lart fedex Dec 13 14:23:53 * ibot takes large quantities of Krispy Kream donuts and stuffs them one after another down fedex's throat until fedex puts on 150lbs Dec 13 14:24:09 Ifaistos: http://dominion.kabel.utwente.nl/koen/cms/another-new-toy-genisi-efika Dec 13 14:24:41 * XorA has a toolchain issue that looks easy to fix, but might have ramafications Dec 13 14:25:00 I want to set this in libstdc++.la instead of what is actually there Dec 13 14:25:27 dependency_libs=' -L/home/gg/zaurus/build-terrier-ang/tmp/cross/lib/gcc/${TARGET_SYS} -lgcc_s -lc -lm' Dec 13 14:26:56 obviously without the stuff about my home directory Dec 13 14:31:27 XorA: take a look at lib-package.bbclass Dec 13 14:33:15 koen: thats got nothing to do with it Dec 13 14:33:47 *lightbulb slowly turning on* Dec 13 14:34:34 koen: look at you tmp/cross/${TARGET_SYS}/lib/libstdc++.la file and tell me whats slightly wrong about it :-) Dec 13 14:35:58 -L/data/build/koen/OE/build/tmp/angstrom/work/ppc603e-angstrom-linux/gcc-cross-4.1.1-r8/gcc-4.1.1/build.i686-linux.powerpc-angstrom-linux/powerpc-angstrom-linux/libstdc++-v3/src Dec 13 14:36:58 koen: the main offender is one that ends up in tmp/staging/${BUILD_SYS}/lib Dec 13 14:37:18 XorA: could you send an RFC to the list? Dec 13 14:37:42 koen: will do Dec 13 14:37:56 hey zecke Dec 13 14:38:11 zecke: are you going to do a presentation about QA at fosdem? Dec 13 14:38:35 zecke: insane.bbclass and fail-fast would be pretty sexy subjects for the embedded room Dec 13 14:38:57 when is the deadline? Dec 13 14:40:35 zecke: 14/01/2007 Dec 13 14:41:16 okay, sounds managable Dec 13 14:41:20 03freyther 07org.oe.oz354x * r639934fb... 10/ (1 classes/tinderclient.bbclass): copied the tinderclient.bbclass from .dev to oz354x Dec 13 14:43:17 zecke: thx Dec 13 14:43:47 this shity corperate firewall loves to misbehave... Dec 13 14:44:38 * koen looks at http://linuxdevices.com/news/NS2209350555.html Dec 13 14:45:07 koen: I will scan the ALP flyer from ITU Hong Kong Dec 13 14:45:11 koen: it is so funny Dec 13 14:45:37 I bet Dec 13 14:45:43 basicly they found out that Gtk2.6 is faster than Gtk2.8 Dec 13 14:45:52 does OE build a gdb for the target that can run on the build system ? Dec 13 14:45:59 a non eixsting gdk2.6 directFB port performs worse than kdrive Dec 13 14:46:06 vlo: bitbake gdb Dec 13 14:46:17 and after MVista has created Linux, Access improved Gtk+ and pango Dec 13 14:46:23 vlo: and gdb-cross iirc Dec 13 14:47:04 hrw: thank you Dec 13 14:49:11 np Dec 13 14:53:38 koen: sent Dec 13 14:53:41 hi Crofton Dec 13 14:53:53 gm Dec 13 14:55:04 yay!!! I just converted my wrtsl54gs into a brick ... Dec 13 14:55:39 chouimat: oops Dec 13 14:55:59 cya tomorrow Dec 13 14:56:22 XorA: not really a brick I forgot to enable boot_wait :( Dec 13 14:57:10 chouimat: a bit hard to reflash without it Dec 13 14:57:41 hrw|work: I know ... I will have to install a jtag port ... Dec 13 14:59:26 can't wait for ALP to actually release their platform Dec 13 14:59:30 sounds promising Dec 13 14:59:37 s/ALP/Access/ Dec 13 15:05:55 Laibsc1: ping Dec 13 15:06:44 hrw|work: pong Dec 13 15:07:11 Laibsch: how much stuff is needed to get japanese in qt/e? I want to have it in OZ 3.5.5 Dec 13 15:10:36 hrw|work, so regarding opie-image.bb changes: why did you remove testing for flash size and replaced with adhoc machine overrides? Dec 13 15:13:02 psokolovsky_: it was 'if flashsize==24 then add stuff' Dec 13 15:13:10 hrw|work: You mean what packages? Either nunome or qpobox (+poboxserver) plus dependency for Japanese font. That should be it. Not too much. But nunome and qpobox do not compile ATM. I tried stripping -fnortti from QMAKE CXXFLAGS but that was not sufficient it seems. Dec 13 15:13:21 psokolovsky_: most of machines does not have exactly 24 set Dec 13 15:13:42 hrw|work: If you pick just one input method, I think qpobox would be the choice. Dec 13 15:13:53 Laibsch: they need nkf? Dec 13 15:13:56 hrw|work, ok, so let's have <, >, etc conditions? Dec 13 15:14:26 psokolovsky_: feel free to improve - I needed it for collie testing Dec 13 15:14:43 hrw|work, ok Dec 13 15:15:00 hrw|work: I never did. But there is an nkf bb file in openzaurus-ja. nkf is a kanji converter. I am not exactly sure what it is used for and when it would be necessary. Dec 13 15:15:18 ok Dec 13 15:16:20 NOTE: package nunome-1.0.1-r1: task do_build: completed Dec 13 15:16:22 .oz354x Dec 13 15:16:39 hrw|work: According to bug 1362 I did get nkf to compile. Dec 13 15:17:00 hrw|work: Cool! Still waiting for recompilation to finish here. Dec 13 15:17:10 hrw|work: What changes were necessary? Dec 13 15:17:59 hrw|work: I guess 3.5.5 will not be available for collie? Can I help you test with qemu? Dec 13 15:18:24 3.5.5 will be 2.4/collie and 2.6/rest Dec 13 15:19:47 OK, nice. Dec 13 15:20:12 nkf is in OE dev tree, BTW. Just checking who commited it about three weeks ago. Dec 13 15:29:10 we need customizable operators for base_conditional Dec 13 15:29:32 base_conditional( lefthandside, operator, righthandside, on-true, on-false) Dec 13 15:29:41 >= makes much more sense for the flash check Dec 13 15:30:01 mickeyl: collie/dev does not fit with opie image Dec 13 15:30:15 *nod* Dec 13 15:30:21 we knew that time would come Dec 13 15:30:26 will check gpe-image Dec 13 15:30:28 collie only altboot + sd now Dec 13 15:30:39 NOTE: multiple providers are available (linux-rp, hostap-modules); Dec 13 15:30:39 NOTE: consider defining a PREFERRED_PROVIDER to match runtime hostap-modules Dec 13 15:30:53 ~lart lack of PREFERRED_RPROVIDER Dec 13 15:30:53 * ibot beats lack of PREFERRED_RPROVIDER senseless with a 50lb Unix manual Dec 13 15:31:57 ls Dec 13 15:32:06 ls: command not found Dec 13 15:32:15 focus! Dec 13 15:32:27 do what I mean, not what I type Dec 13 15:32:32 hehe Dec 13 15:32:48 I am sending BPSK across the lab Dec 13 15:33:00 sounds scary Dec 13 15:33:05 ls: too many pr0n in this directory ... Dec 13 15:33:05 binary phase shift keying? Dec 13 15:33:10 heh chouimat Dec 13 15:33:14 yep Dec 13 15:33:15 mickeyl: :) Dec 13 15:33:26 so... should i buy that MBP this year or next year? Dec 13 15:33:31 * mickeyl ponders re. taxes Dec 13 15:33:36 I would be doig from my OSK, but my USB1.1 fixes are sucking in the usrp Dec 13 15:35:10 mickeyl: I suspect it will take a while for apple to release a new model Dec 13 15:35:55 koen: i agree. might not make sense to wait much longer. just a question whether my earnings next year will be up to par with this year or not Dec 13 15:36:25 earnings... Dec 13 15:36:30 ~lart my customer Dec 13 15:36:31 * ibot beats my customer into protomatter with the andromeda galaxy Dec 13 15:37:53 koen: actually my plan is to retire my windows machine once the MBP is there... Dec 13 15:38:05 cultural shock! :D Dec 13 15:39:54 hrw|work: I should have a 5000 for testing from next week Dec 13 15:41:01 XorA: plan to port asoc for it? :D Dec 13 15:41:24 hrw|work: nah, but useful for testing for moany collie users ~:-) Dec 13 15:43:19 list Dec 13 15:43:54 | /bin/sh: ../src/util/makekeys: cannot execute binary file Dec 13 15:43:55 argh Dec 13 15:44:13 hrw|work: koen's fix broke it. please revert the "fix" Dec 13 15:44:20 i have no OE machine atm. Dec 13 15:44:30 or try Dec 13 15:44:35 mickeyl: he 'fixed' libx11, I have problem with diet-x11 Dec 13 15:44:41 oh Dec 13 15:44:46 didn't build that in a while Dec 13 15:44:52 perhaps do the same as in libx11 Dec 13 15:44:55 oe_runmake etc. Dec 13 15:51:01 mickey|bbl: Are you Freiberufler? In that case buy it in January and if earnings really are lower next year use an Ansparabschreibung to defer the payment of taxes from this year into something in the future. I thought you finished your thesis this year? Still expect to make bigger bucks this year than next? With openmoko and everything??? Dec 13 15:55:53 can anyone suggest a good browser for embedded devices, preferably with flash support? Dec 13 15:58:27 how could I change a package source quickly for debug purpose and to make it rebuild without having to make a .patch file and thus changing the recipe ? Dec 13 15:59:50 bye Dec 13 15:59:59 well, if the source is downloaded from your local server you could do that Dec 13 16:00:27 by cleaning the package and manually removing the downloaded sources + md5 file Dec 13 16:02:33 vlo: bitbake -b recipe.bb -f -ccompile is not enough? Dec 13 16:04:27 03justinp 07org.oe.dev * r526df197... 10/ (1 conf/distro/openzaurus-unstable.conf): openzaurus-unstable: virtual/libx11 provider should be overridable Dec 13 16:09:02 vlo: You want to download from a different URI? Dec 13 16:10:10 vlo: What is the problem with changing the bitbake recipe? Do you know quilt? I use it for such purposes. Dec 13 16:10:23 no, for something I've hacked locally Dec 13 16:11:09 hrw|work: qpobox RDEPENDS on poboxserver which again DEPENDS on nkf-native. So, all these are needed but OE should take care of it automatically. Dec 13 16:11:15 in the c code Dec 13 16:11:55 vlo: So you changed the bitbake recipe to point to your local tar ball or not? Dec 13 16:13:20 Laibsch: no, I would like to hack the bitbake uncompressed tar ball and take the compilation only at the do_compile of the recipe Dec 13 16:13:53 I guess that's exactly what hrw showed with bitbake -b recipe.bb -f -ccompile Dec 13 16:14:14 vlo: try this: "file://localdir/*" - worked for me with all files in this dir Dec 13 16:20:20 btw, what is quilt ? I don't know this ... Dec 13 16:21:18 ok, man quilt did the answer Dec 13 16:38:55 cu Dec 13 16:49:19 * mrz80 is away: Metaphysically I'm still here; physically, I'm elsewhere Dec 13 16:54:11 03koen 07org.oe.dev * r7dc5c614... 10/ (1 packages/xorg-lib/libx11_X11R7.1-1.0.1.bb): libx11: revert last commit till someone comes up with a proper fix Dec 13 17:05:21 hmm this is not recursive and does not include subdirs? FILES_${PN} = "${bindir} ${datadir}/mediatomb/*" Dec 13 17:05:36 still have the "installed but unpackaged files" message Dec 13 17:10:19 hmm it seems I generally fail to get it right with the FILES_ thing Dec 13 17:14:12 Hello, I wanted to know how can I obtain a full Ruby package including the complete standard library Dec 13 17:14:34 The OpenEmbedded ones seem to be missing, at least, the curses part Dec 13 17:16:46 I've set up an OpenEmbedded environment, but although the ruby source code includes curses, the ipk produced by bitbake doesn't Dec 13 17:26:24 RDEPENDS += "ncurses" in the rby bb file? Dec 13 17:32:03 it's not really ncurses, just curses. And the file is in Ruby sourcecode, so I don't think it's a dependency... Dec 13 17:33:49 I don't know why the directory lib/ext/curses of the source is not compiled by bitbake Dec 13 17:35:10 Could it be some configure options? Dec 13 17:39:16 Jin^eLD: FILES is a python glob, so foo/* only includes files, while foo would include everything Dec 13 17:45:06 elKano_: check the ruby bb file and see what its configure options are. Either it's disabling curses or not enabling it. It also may be patched out if there are patches. Dec 13 17:45:57 no patches or options it seems Dec 13 17:46:36 so either a) you need to explicitly enable it or b) some internal check fails and ruby disables it Dec 13 17:47:04 My EFIKA is in Memphis! Dec 13 17:47:11 hmm Dec 13 17:47:21 I don't think you need to explicitly enable it (I always have had it in other compilations I've done), so it could be the internal check Dec 13 17:47:22 maybe we need to rebuild it Dec 13 17:47:26 arg Dec 13 17:47:28 Crofton: and it has been in san antonia Dec 13 17:47:41 :) Dec 13 17:47:43 last night Dec 13 17:48:02 it took ~1 day to get from memphis to my doorstep Dec 13 17:48:03 and that's San Antonio :) Dec 13 17:48:11 :) Dec 13 17:48:18 ferigners Dec 13 17:48:45 I am going to Canada tomorrow, need to line up someone to check my doorstep tomorrow Dec 13 17:49:15 most likely someone to sign a "P Balister" Dec 13 17:49:20 heh Dec 13 17:49:27 Usually they just leave it Dec 13 17:49:31 yeah Dec 13 17:49:44 can you build an image in OE yet? Dec 13 17:49:55 even if it says "SIGNATURE REQUIRED, DO NOT LEAVE AT NEIGHBOURS" Dec 13 17:50:01 Crofton: sort of Dec 13 17:50:01 heh :) Dec 13 17:50:22 Crofton: usbutils doesn't build, so I locally remove usbhost from machine features Dec 13 17:50:28 and I edited madwifi Dec 13 17:50:29 I need to find a power supply, and I have a build machine at home Dec 13 17:50:51 wadwifi for a usb nic? Dec 13 17:51:25 MACHINE_FEATURES = "pci" + DISTRO_FEATURES = "wifi" Dec 13 17:51:53 madwifi needs something like export ARCH_powerpc = powerpc Dec 13 17:51:58 I guess ndiswrapper would be a disaster :) Dec 13 17:52:05 (OE sets ppc by default) Dec 13 17:52:11 ops Dec 13 17:52:47 the basic stuff is already in the angstrom feeds Dec 13 17:52:47 I just the basic stuff working and usb Dec 13 17:52:47 then I will be stuck at my usb full speed and usrp sucks problem I am already having with the OSK Dec 13 17:52:49 which is progress :) Dec 13 17:52:49 i'm sure somebody (else) has tried building binutil-2.16 with since the migration to the multiple site/* cache files... ??? Dec 13 17:53:11 HopsNBarley: i tried 2.17.50.0.6 and that failed on libiberty Dec 13 17:53:25 ah, same boat my friend. i've narrowed it down a bit. Dec 13 17:53:39 I added the needed tests to the site file, but that doesn't seem to work Dec 13 17:53:58 the run.do_configure script does not copy the site file(s) to config.cache at all. Dec 13 17:54:39 where should i look for this ,siteinfo.bbclass? Dec 13 17:55:08 that's because the libiberty configure gets run in oe_runmake Dec 13 17:55:19 so you probably need to copy them by hand Dec 13 17:55:40 of course ,but it used to do it, when there was a single file. i'm diffing run.do_config's... Dec 13 17:56:03 it won't show up in do_configure Dec 13 17:56:36 EXTRA_OEMAKE = "configure-build-libiberty all" Dec 13 17:56:52 about two months ago, it did show up in run.do_configure. i'm looking at it. Dec 13 17:56:53 the libiberty configure gets run in do_compile Dec 13 17:57:05 koen: I see, thanks Dec 13 17:57:26 i understand last two lines. Dec 13 18:00:24 OK, I looked a bit more into Ruby compilation, and it seems to actually require ncurses, so I added the RDEPENDS line Crofton mentionned earlier; I am trying to re-bitbake ruby to see if this does it Dec 13 18:01:23 koen, separately, i've got some perl ppc diffs, if yours isn't building. Dec 13 18:01:36 indeed it isn't Dec 13 18:01:43 i'll push in a few... Dec 13 18:02:17 I got a NOTE on bitbake Ruby: Multiple libraries (libncurses.so.5, libpanel.so.5, libmenu.so.5, libform.so.5) found and LEAD_SONAME not defined Dec 13 18:02:54 libncurses.so.5?!?!?! Dec 13 18:03:05 doesn't that conflict with the 'real' ncurses lib? Dec 13 18:03:20 my problem with unpackaged files seems to be a different one.. seems that in the image directory the files got installed not to image/share/packge but image/home/path/to/oe/build/directory/etc... it may have something to do with the Makefile,allsthough I think I am doing it correctly Dec 13 18:03:25 i presume he got the note from teh ncurses build that was built due to the dependency Dec 13 18:03:31 and just wasnt verbose enough in his report Dec 13 18:03:34 koen: I don't actually understand the note Dec 13 18:03:36 heh Dec 13 18:03:41 could anyone take a look at that: http://pastebin.ca/277948 Dec 13 18:03:52 hey kergoth Dec 13 18:03:55 hey Dec 13 18:03:57 how goes it? Dec 13 18:03:57 hey kergoth, you did the whole perl thing, right? Dec 13 18:04:04 * kergoth hides Dec 13 18:04:07 if I use the makefile as in the post, the stuff gets installed in the wrong directory, allthough that would be the correct automake style as far as I know Dec 13 18:04:15 elKano_: OE can do debian-like shlib renaming and wants to know what lib to use to rename the package Dec 13 18:04:27 if I use the stuff that is commented out - then it works correctly, but that is a hack and afaik is discouraged Dec 13 18:04:32 kergoth, no no, i owe you for figuring that out. thanks! Dec 13 18:04:37 elKano_: e.g. ncurses_5.0.ipk -> libncurses5_5.0.ipk Dec 13 18:04:51 i had a hand in the perl build, but it wasnt me that did most of it, it was someone thats no longer around Dec 13 18:05:01 hmmm, i wonder Dec 13 18:05:03 ~perl's buildsystem Dec 13 18:05:10 wow, I should really look into OE inner workings Dec 13 18:05:12 extra, extra, read all about it, perl's buildsystem is the devil! Dec 13 18:05:17 hehe Dec 13 18:05:18 ~botsnack Dec 13 18:05:19 :), kergoth Dec 13 18:05:40 hm, i still need to go make that oz dns change before i forget about it Dec 13 18:05:46 kergoth: did you see hrw's mail Dec 13 18:05:48 heh Dec 13 18:05:52 * koen <- too slow Dec 13 18:05:54 :) Dec 13 18:07:41 koen: I still don't get it, why does it mix libmenu, libncurses, etc, in the naming of libcurses? and, may this be the issue with Ruby curses? Dec 13 18:08:20 what is there to not get? some packages have more than one lib in them. big deal. Dec 13 18:08:47 oh, ok, now I get it Dec 13 18:08:58 koen, think i got binutils building. Dec 13 18:09:06 HopsNBarley: yay! Dec 13 18:16:00 koen: any plan for adding sh-common to powerpc-linux? i'm testing my build both ways, if curious. Dec 13 18:16:49 isn't sh super-h? Dec 13 18:17:59 ? Dec 13 18:19:03 sh-common is the sitefile for sh3 and sh4 Dec 13 18:19:29 sh3 and sh4 are so similar that it got merged into a common site file Dec 13 18:20:10 So, everything else failing, I am going to try and set the '--with-ncurses-dir' option in EXTRA_OECONF; could anybody point me in the right direction about the dir to give it? Dec 13 18:21:13 ${STAGING_LIBDIR} for the libs Dec 13 18:21:23 cool, thanks :) Dec 13 18:26:22 so.. anyone has an idea why OE puts stuff defined as below into a wrong location in the image directory? Dec 13 18:26:23 # Dec 13 18:26:23 imagesdir = $(sharedir)/web/icons/nanotree/images Dec 13 18:26:23 # Dec 13 18:26:25 images_DATA = $(_DIST_FILES_IMAGES) Dec 13 18:26:39 afaik this is the correct way of installing data stuff Dec 13 18:26:48 with the _DATA thingie.. Dec 13 18:27:06 maybe the makefile doesn't honour DESTDIR too much Dec 13 18:27:12 if I use a custom script via install-data-local: it works Dec 13 18:27:31 sharedir is defined as sharedir = $(DESTDIR)@datadir@/@PACKAGE@ Dec 13 18:28:01 and if I use this same variable via a install-data-local: script it is installed correctly Dec 13 18:28:19 that's the whole thing http://pastebin.ca/277948 Dec 13 18:28:34 the stuff that is commented out would work, but is not the correct aproach I think... Dec 13 18:48:08 koen : Got the icecc.bbclass working :) Dec 13 18:48:51 koen: now on the same page re: sh-common. Dec 13 18:54:22 koen: how's this look: http://pastebin.ca/278007 Dec 13 18:58:19 HopsNBarley: looks ok, it could use a comment explaining why Dec 13 19:00:06 okay. should be okay for all flavors of ixp4xx??? Dec 13 19:04:28 koen, hi! mreimer suggested that you're using eabi. Could you (or any eabi user here, with kernel >= 2.6.18, glibc >= 2.4) please confirm or discredit that bug: http://bugzilla.handhelds.org/show_bug.cgi?id=1773 -- it work for me using old ABI, and I'm curious if it also happens with EABI Dec 13 19:18:13 hi kergoth, mreimer, woglinde Dec 13 19:18:20 hi pb_! Dec 13 19:34:50 hi pb Dec 13 19:42:55 good evening everyone Dec 13 19:52:27 * mrz80 is back (gone 03:03:09) Dec 13 19:55:16 I've got a weird thing with my Angstrom build: I have a glibc 2.5 and all my executables are linked against glibc 2.4 ? Dec 13 19:55:46 glibc 2.5 probably still uses the glibc 2.4 symbols Dec 13 19:56:54 mmmhhh, ok Dec 13 20:00:24 so I will have to find out why init hangs at its very beginning Dec 13 20:01:32 hi zecke Dec 13 20:01:54 *scared* Dec 13 20:03:15 what are you scared of? Dec 13 20:03:52 a test :) Dec 13 20:04:06 zecke: how's bitbake-parser coming? Dec 13 20:04:11 * kergoth 's outta the loop Dec 13 20:04:40 kergoth: pending 'rewrite' to use the visitor pattern Dec 13 20:05:04 kergoth: but it should be 'complete' and one should be able to hook it into bitbake... Dec 13 20:05:17 completion is scheduled for christmas Dec 13 20:05:41 back to my notes :) Dec 13 20:05:44 bye Dec 13 21:21:06 hi Dec 13 21:24:23 ERROR: No providers of runtime build target jhead (for ['/home/hrw/devel/oe/org.openembedded.dev/packages/gallery/gallery_1.5.5.bb']) Dec 13 21:24:37 HopsNBarley: you added gallery? Dec 13 21:36:33 NOTE: Preparing Runqueue Dec 13 21:36:33 NOTE: Executing runqueue Dec 13 21:36:39 which line to change? Dec 13 21:37:01 s/Runqueue/runqueue or viceversa? Dec 13 21:37:23 i'd say both lowercase, personally. *shrug* Dec 13 21:37:36 so do I Dec 13 21:38:54 ~lart koen^Wour dbus maintainer for DEF_PREF = -1 for release version Dec 13 21:38:55 * ibot takes koen^Wour dbus maintainer for DEF_PREF = -1 to the vet for a "special" visit for release version Dec 13 21:39:45 hrw: yeah. isn't jhead in there? Dec 13 21:39:57 I do not like to work as vacuum cleaner just to get something working... (dbus) Dec 13 21:40:16 HopsNBarley: my fetch task failed on jhead or it is not in Oe Dec 13 21:40:38 it is not in OE - just checked Dec 13 21:40:55 one moment (-; Dec 13 21:41:25 my improved fetch.sh is doing nice job Dec 13 21:43:45 added fetcher to tinderbox Dec 13 21:44:13 round 4000 recipes Dec 13 21:44:47 koen: ping? Dec 13 21:45:06 03awilcox 07org.oe.dev * r2b68c3f3... 10/ (3 files in 3 dirs): Dec 13 21:45:06 perl: addition of powerpc-linux config.sh file and bitbake hooks to use Dec 13 21:45:06 it. Dec 13 21:46:04 hrw: pong Dec 13 21:46:15 koen: why dbus 1.0.1 is not default? Dec 13 21:46:31 http://ewi546.ewi.utwente.nl/tinderbox/showbuilds.pl?tree=OpenEmbeddedBuilds - hrw-source-mirror-fetcher added Dec 13 21:46:44 1) because the initscript changed, which needs testing Dec 13 21:46:56 2) because it still needs the no-segfault hack Dec 13 21:47:07 3) because dbus-launch is still broken Dec 13 21:47:24 4) OE lacks goats to sacrifice to the dbus gods Dec 13 21:47:47 anyone play with initng? Dec 13 21:47:57 kergoth: I moved to upstart Dec 13 21:49:03 kergoth: used initng in past Dec 13 21:49:42 interesting, not looked at upstart before Dec 13 21:49:43 how is it? Dec 13 21:50:27 kergoth: no boot messages at all, quite fast (but not compared it with timer), good replacement for sysvinit Dec 13 21:50:33 nice Dec 13 21:50:47 no problem at all on debian sid amd64 Dec 13 21:51:29 bitbake world is slow with trunk (but it has to be due to lot of recipes and tasks) Dec 13 21:52:18 koen: 2) or 3) has bugreport iirc - what about others? Dec 13 21:53:03 if someone feels it has had enough testing, feel free to remove the -1 Dec 13 21:53:15 I'm not the dbus maintainer :) Dec 13 21:53:35 | /home/hrw/devel/build/collie-dev/tmp/cross/lib/gcc/arm-linux/3.4.4/../../../../arm-linux/bin/ld: cannot find -lXt Dec 13 21:53:36 (I did propose a cleanup a while ago) Dec 13 21:53:39 NOTE: package osb-nrcore-0.5.0+cvs20050430-r1: task do_compile: failed Dec 13 21:53:50 lack of depends? Dec 13 21:54:13 looks like it Dec 13 21:54:42 hrw: here ya go... Dec 13 21:55:19 03awilcox 07org.oe.dev * r1fc9d550... 10/ (4 files in 3 dirs): jhead: addition of jhead recipe Dec 13 21:55:34 gracias Dec 13 21:55:45 hrw: what hardware are you going to run gallery on? Dec 13 21:56:25 HopsNBarley: none - I'm fetching all OE sources Dec 13 21:57:48 ~lart gtk-webcore 'maintainer' for not unified recipes Dec 13 21:57:48 * ibot chops gtk-webcore 'maintainer' in half with a free AOL CD for not unified recipes Dec 13 21:58:49 for next bugdays I suggest unifing recipes instead of fixing bugs Dec 13 22:01:26 hi Tim Dec 13 22:02:20 HopsNBarley: did you push the binutils stuff yet? Dec 13 22:02:28 just about to. Dec 13 22:02:39 03koen 07org.oe.dev * r184d60b9... 10/ (22 files in 6 dirs): Dec 13 22:02:39 dbus: make Marcin happy: Dec 13 22:02:39 * create a dbus.inc Dec 13 22:02:39 * remove outdated dbi Dec 13 22:02:40 * remove DEFAULT_PREFERENCE from 1.0.1 Dec 13 22:02:42 * dbus-glib still has both versions since 0.72 doesn't build for me Dec 13 22:02:43 i was adding the comment (-; Dec 13 22:05:27 hmmm Dec 13 22:05:33 binutils could use a cleanup as well Dec 13 22:06:19 03awilcox 07org.oe.dev * r8229c859... 10/ (1 packages/binutils/binutils.inc): binutils: be sure and prime the config.cache for libiberty Dec 13 22:06:40 koen, pushed. Dec 13 22:07:14 HopsNBarley: you forgot to sync before your commit Dec 13 22:07:56 oh crap. Dec 13 22:07:56 np Dec 13 22:07:56 * koen merges Dec 13 22:09:28 alias mtpush="mtn pull && mtn merge && mtn update && mtn push" Dec 13 22:13:37 handy combo Dec 13 22:14:01 hrw: could you send an rfc about unifying recipes? Dec 13 22:14:19 HopsNBarley: does usbutils build for you? Dec 13 22:16:14 I'm still on the "L's", lets advance a bit... Dec 13 22:16:18 koen: when will get time Dec 13 22:16:28 hrw: What do you mean unifying recipes? Dec 13 22:16:48 Laibsch: packages/pcmciautils/ Dec 13 22:16:55 it is good exampl Dec 13 22:17:12 koen: yep Dec 13 22:17:16 Laibsch: if you have more then 1 recipe for PN then create PN.inc with common stuff Dec 13 22:17:42 3.6G www/sources Dec 13 22:18:13 hrw: I understand. I'll keep it in mind when I come across such things. Dec 13 22:18:45 Laibsch: otherwise it is more work when you want to fix something Dec 13 22:19:01 Of course, redundant and inconsistent stuff. Dec 13 22:19:04 HopsNBarley: D'oh, I meant libusb Dec 13 22:19:20 03koen 07org.oe.dev * r0a20426c... 10/ (1 packages/binutils/binutils_2.17.50.0.5.bb): Dec 13 22:19:20 binutils 2.17.50.0.5: add site prime fix. Dec 13 22:19:20 *mumbles something about unifying recipes* Dec 13 22:20:22 koen, yeah, libusb too. Dec 13 22:20:31 hmmm Dec 13 22:20:40 bombs out with tons of c++ errors over here Dec 13 22:20:47 want the config log or anything? Dec 13 22:21:01 back in a sec... Dec 13 22:22:28 could it be that "mtn log $file" is REALLY slow? Like taking half a day and more on a slower computer? Dec 13 22:22:53 Laibsch: monotone 0.32 will improve it Dec 13 22:22:57 a LOT Dec 13 22:23:09 nice. looking forward to it. Dec 13 22:23:13 I already have 0.31+ on one machine and mtn log flies Dec 13 22:26:09 HopsNBarley: I have a fix for libusb :) Dec 13 22:26:12 * koen hugs sed Dec 13 22:27:57 koen, what was wrong? Dec 13 22:28:03 hrm Dec 13 22:28:14 my sed fix just completes my hack Dec 13 22:28:17 | usbpp.cpp:(.text+0xc68): undefined reference to `std::basic_string, std::allocator >::~basic_string()@@GLIBCXX_3.4' Dec 13 22:28:32 koen: link with g++? Dec 13 22:28:44 what compiler? I'm using 4.1.1 Dec 13 22:28:48 4.1.1 Dec 13 22:29:12 03hrw 07org.oe.oz354x * rdbbca732... 10/ (9 files in 5 dirs): acpid: dropped - hardware does not support it Dec 13 22:29:20 03hrw 07org.oe.dev * r3b92a7d1... 10/ (5 files in 2 dirs): osb-nrcore: unified recipes, added missing libxt dependency Dec 13 22:29:31 03hrw 07org.oe.dev * r8f363fda... 10/ (1 packages/ekg/ekg_1.5.bb packages/ekg/ekg_1.6.bb): ekg: updated to 1.6 Dec 13 22:30:00 * koen tried to remember how to get OE to link with g++ Dec 13 22:30:02 someone use pmount? Dec 13 22:30:10 koen: CC=g++? Dec 13 22:31:11 http://ewi546.ewi.utwente.nl/tinderbox/showlog.pl?machine_id=172&logfile=20061213231239.log - trunk logs looks nice indented ;D Dec 13 22:31:16 ah, the dd-wrt dudes finally sent my POE thingo Dec 13 22:31:55 NOTE: Handling BitBake files: \ (2537/3975) [63 %]ERROR: Could not include required file dbus.inc while parsing /home/hrw/devel/oe/org.openembedded.dev/packages/dbus/dbus_1.0.1.bb Dec 13 22:31:58 koen: ? Dec 13 22:32:03 hrw: heh, did zecke try to modulate a dvb signal over the logs? Dec 13 22:32:46 akita/celinux-test fetching now.. Dec 13 22:33:04 koen: will you mtn add/push missing file? Dec 13 22:33:18 hrw: already did Dec 13 22:33:22 thx Dec 13 22:33:44 so my next MACHINE run will fetch it probably. Dec 13 22:34:09 hrw: You have really stripped opie. Dec 13 22:34:21 goxboxlive: ? Dec 13 22:35:25 03koen 07org.oe.dev * rab4307bb... 10/ (1 packages/dbus/dbus.inc): dbus.inc: add it Dec 13 22:35:26 After the changes in oe.dev. There are alot of apps that is not installed. It's good i think. The image has becomed almost 6MB smaller Dec 13 22:35:44 hrw: CC=g++ doesn't work, since not everything is c++ :( Dec 13 22:36:45 goxboxlive: opie-image change you mean? Dec 13 22:37:50 anyone played with jlibtool? Dec 13 22:37:55 koen: on which machine you are building angstrom images? ewi or ltg? Dec 13 22:37:59 (C libtool replacement, its in the apr sources) Dec 13 22:38:10 apr build with regular libtool = 1 minute, 10 seconds. apr build with jlibtool = 27 seconds Dec 13 22:38:12 kergoth: nope Dec 13 22:38:17 hrw: Yes, it where alot smaller. I havent booted it yet, just took a look inthe folders. Seems like opie-wellenreiter, and opie-mediaplyer is missing. I'll get back to it when i have booted it. Dec 13 22:38:38 goxboxlive: how big / you have? Dec 13 22:39:16 hrw: Mine where 18-19MB but now it is just 13MB. Dec 13 22:40:08 hrw: ewi Dec 13 22:40:31 goxboxlive: I do not ask about image sizes but / partition Dec 13 22:41:27 hrw: I am running from my SD card, so it is 256MB. Dec 13 22:41:52 We dont have a bootloader yet, no one has the currage to test it :-) Dec 13 22:46:07 'night all Dec 14 00:37:42 EFIKA has left Memphis **** ENDING LOGGING AT Thu Dec 14 02:59:57 2006