**** BEGIN LOGGING AT Wed Jul 23 02:59:56 2008 Jul 23 07:19:19 03koen 07org.oe.dev * r4b92368c... 10/ (1 conf/machine/neuros-osd2.conf): neuros-osd2: try autoloading cmemk and dsplinkk modules Jul 23 07:48:24 morning Jul 23 08:07:23 bonsoir Jul 23 09:01:44 I'm having a problem with firefox 3.0 building Jul 23 09:01:53 its never built Jul 23 09:02:19 The error I'm getting is: *** No rule to make target `../protocol/jar/src/libnkjar_s.a', needed by `libnecko.so'. Stop. Jul 23 09:02:31 I did a google search and saw you asking the same question here Jul 23 09:02:43 did you ever found out about this? :) Jul 23 09:03:18 * XorA doesnt remeber ever asking that question Jul 23 09:03:41 firefox 3 I never built Jul 23 09:04:22 ok, thanks Jul 23 09:05:00 I guess you were building minimo then Jul 23 09:05:07 did you get it built? Jul 23 09:06:47 I just put together a rough ff3 .bb so the beagleboard guys could try and make it work Jul 23 09:08:52 ok Jul 23 09:09:21 but what about minimo? I think you got the same error building minimo as I'm getting with ff3 Jul 23 09:09:58 minimo built like 4 years ago, then mozilla moved things around and I never got it to build again Jul 23 09:11:24 RP: I just sent out another update to the toolchain build sequence patch anyone who could test it on any toolchain combination would be a great help Jul 23 09:11:26 ok, thanks a lot Jul 23 09:13:20 khem: I wlll get to it soon :) Jul 23 09:13:56 RP: I have tested many combos so far but its been only me Jul 23 09:14:30 khem: Different host systems adds a new dimension to that, certainly :) Jul 23 09:14:48 thats why Jul 23 09:15:02 RP: I have x86-64 here running linux/ubuntu Jul 23 09:15:57 khem: Poky's i586 targets build on i686 are usually "fun" for showing weird failures Jul 23 09:18:00 RP: cool, I am waiting on you Jul 23 09:18:25 khem: What I need to do it port that patch to poky, then I'll set it away on the laptop Jul 23 09:18:27 RP: I have done i586 builds here Jul 23 09:18:40 RP: ok Jul 23 09:19:18 is poky uptodate with OE wrt toolchains stuff Jul 23 09:20:26 khem: yes Jul 23 09:20:44 khem: In some ways its actually ahead :) Jul 23 09:20:51 I usually test in poky, then merge to OE Jul 23 09:21:09 but its in sync atm pretty much apart from some changes to the staging/cross layouts Jul 23 09:21:25 (cross layout change == CROSS_DIR redefined only) Jul 23 09:22:19 I see Jul 23 09:24:07 RP: while building uclibc I get this passed on by makefiles Jul 23 09:24:12 CC=ccache arm-angstrom-linux-uclibc-gcc -march=armv5te -mtune=arm926ej-s Jul 23 09:24:26 where is this string formed ? Jul 23 09:24:38 I think I need to put it in quote Jul 23 09:24:49 make is not liking it on older uclibcs Jul 23 09:25:15 I think that may come from bitbake.conf Jul 23 09:25:41 03koen 07org.oe.dev * r15f857e8... 10/ (1 classes/qmake.bbclass): qmake.bbclass: export qmakespec so we can use do_configure_prepend Jul 23 09:26:43 RP: hmmm Jul 23 09:27:06 RP: so I think it should quote them to be safe Jul 23 09:27:16 make CROSS=arm-angstrom-linux-uclibc- CC="ccache arm-angstrom-linux-uclibc-gcc -march=armv5te -mtune=arm926ej-s" -C utils Jul 23 09:27:20 works ok Jul 23 09:27:45 khem: I think that make commandline comes from somewhere in the uclibc recipes Jul 23 09:27:51 quoting there should be easier Jul 23 09:28:35 RP: it calls oe_runmake Jul 23 09:28:58 and not bare make Jul 23 09:29:17 oe_runmake utils Jul 23 09:29:37 oe_runmake PREFIX=${D} DEVEL_PREFIX=${prefix}/ RUNTIME_PREFIX=/ install_utils Jul 23 09:31:24 RP: I think its in base.bbclass Jul 23 09:33:58 RP: ah I can change EXTRA_OEMAKE in uclibc.inc Jul 23 09:34:03 that will fix it Jul 23 09:34:18 may be Jul 23 09:36:26 yes, that sounds right Jul 23 09:38:03 I changed it ti Jul 23 09:38:05 to Jul 23 09:38:08 EXTRA_OEMAKE = "${OEMAKE_NO_CC} CC="${CC}" HOSTCFLAGS="-I${STAGING_INCDIR_NATIVE}"" Jul 23 09:38:20 hopefully that fixes it Jul 23 09:38:37 I'm a little dubious about that quoting... Jul 23 09:38:45 me too Jul 23 09:38:48 using ' might be better Jul 23 09:38:56 that was being used before Jul 23 09:39:22 hmm :/ Jul 23 09:42:38 obviously did not work Jul 23 09:42:47 now i am trying this EXTRA_OEMAKE = "${OEMAKE_NO_CC} CC=\'${CC}\' HOSTCFLAGS=\'-I${STAGING_INCDIR_NATIVE}\'"\ Jul 23 09:43:40 khem: If thats really needed would should find out what on earth bitbake is doing with its quoting :/ Jul 23 09:44:17 all I want to do it when CC is passed down to make it should be quoted Jul 23 09:44:32 khem: yes... Jul 23 09:44:50 khem: What does the shell script that bitbake generates look like? Do the quotes make it to there? Jul 23 09:46:00 i've a pb with override netbase/files/interfaces with recent org.oe.dev http://pastebin.com/m7359bbb8 Jul 23 09:46:01 No it does not make it there Jul 23 09:46:15 have someone an idea of the pb ? Jul 23 09:46:30 i suspect some changes in bitbake i've missed Jul 23 09:47:41 ( it's the same recipe ) Jul 23 09:50:26 Genesis: offhand I don't know. You could compare OVERRIDES between the two? Jul 23 09:50:33 same Jul 23 09:50:44 khem: If it doesn't make it there, its a bitbake issue :/ Jul 23 09:50:52 khem: at least I'd suspect bitbake Jul 23 09:51:03 RP: yeah Jul 23 09:51:08 me too i should perharps downgrade my bitbake to test Jul 23 09:51:18 Genesis: are all the netbase files the same? Jul 23 09:51:24 yeap Jul 23 09:51:42 it's in my overlay Jul 23 09:52:02 be 100% sure its still using the one in your overlay Jul 23 09:52:16 if not it should not take opennao/interface Jul 23 09:52:22 maybe use -b with a full path just to check Jul 23 09:52:46 opennao/packages/netbase/netbase/ it's explicit Jul 23 09:52:50 Genesis: You could try downgrading bitbake. If that is the problem we need to know about it Jul 23 09:53:00 oki thx Jul 23 09:53:18 (i'm using git bitbake on this , 1.6 on the other side ) Jul 23 09:54:31 sorry svn://svn.berlios.de/bitbake/branches/bitbake-1.8 Jul 23 09:54:48 comapred to 1.6? Jul 23 09:58:06 RP: will u take a look at bitbake for this issue Jul 23 09:59:12 khem: Not soon. I'm try to test your patch as we speak Jul 23 09:59:35 RP: ok bitbake can wait Jul 23 09:59:44 * khem thinks of cloning RP Jul 23 10:00:31 there is a lot of use of EXTRA_OEMAKE Jul 23 10:00:42 and uclibc is no different Jul 23 10:01:17 no, its strange Jul 23 10:01:20 it demos here because there is another '-' which confuses make Jul 23 10:02:29 time to sleep now Jul 23 10:02:32 good night Jul 23 10:02:46 khem: 'night Jul 23 10:02:57 perhaps tomorrow I'll have some feedback :) Jul 23 10:05:51 03koen 07org.oe.dev * r4a4ba613... 10/ (8 files in 3 dirs): Jul 23 10:05:51 mythtv: <= 0.21 are qt3 based >= 0.22 are qt4 based, so move inherits around Jul 23 10:05:51 * update 0.21 to use the fixes branch Jul 23 10:18:45 03koen 07org.oe.dev * r0a219d89... 10/ (3 files in 2 dirs): mythtv: fix quoting for armv7a Jul 23 10:20:59 RP : i copied the bitbake i use on my old oe , i've the same pb , so i suspect bbclass , i'm on it :) Jul 23 10:23:58 Genesis: ok :) Jul 23 10:24:06 * RP is relived its not bitbake :) Jul 23 10:24:26 weird no one that use override for netbase has the same pb Jul 23 10:30:50 good morning Jul 23 10:35:46 RP: thx again for the advices about cmdline yesterday. Hopefully next month we'll see the long awaited initramfs-bootmenu-image. Jul 23 10:36:23 The 'u-boot-zaurus not passing ATAG' issue need more investigation, tough Jul 23 10:36:40 ant|work: np Jul 23 10:39:22 RP: do you know if mfd is upstreaming this merge window ? Jul 23 10:39:43 lrg: I know Samuel wanted to Jul 23 10:39:59 RP: I've got some other users Jul 23 10:39:59 lrg: I also know he's on holiday for the two weeks of the merge window without internet access Jul 23 10:40:22 Jolly good. Jul 23 10:40:38 It is new code so it should be possible for it to go in during rc1 I guess :/ Jul 23 10:41:32 broonie: I'm hoping thats what will happen Jul 23 10:42:04 RP: I've got a couple of PMICs users, I guess I'll convert them and go for the next merge window too. thanks Jul 23 10:43:26 morning Jul 23 10:44:11 hey Marcin Jul 23 10:50:19 hi lrg Jul 23 10:51:04 hey pb__ Jul 23 10:53:12 pb__: hey, thx you too. I guess I misinterpreted following line in u-boot's machine file +#define CONFIG_CMDLINE_TAG 1 /* enable passing of ATAGs */ Jul 23 11:18:36 pb__: hmm, no, I did not. Patch landed in u-dev tree (pxa255_idp.h) Jul 23 11:18:36 03koen 07org.oe.dev * re8326342... 10/ (1 packages/lirc/lirc-modules_0.8.3+cvs20080713.bb): lirc modules: update to 0.8.3+cvs20080713 Jul 23 11:28:04 morning all. Jul 23 11:42:53 03koen 07org.oe.dev * ra57b205a... 10/ (1 packages/tcp-wrappers/tcp-wrappers_7.6.bb): tcp-wrappers: fix staging Jul 23 11:42:57 03koen 07org.oe.dev * r94a6fbb8... 10/ (1 packages/dsplink/codec-engine_2.10.bb): codec-engine: stage some more libsf Jul 23 11:43:11 hi thesing Jul 23 11:52:04 03koen 07org.oe.dev * r4acd9f12... 10/ (1 packages/mythtv/mythtv_0.21.bb): mythtv 0.21: fix depends for themes Jul 23 12:16:15 root Jul 23 12:16:37 03koen 07org.oe.dev * r5990e291... 10/ (1 packages/nfs-utils/nfs-utils_1.1.2.bb): nfs-utils: update to 1.1.2 Jul 23 12:27:57 RP : i found the pb , git doesn't copy empty directory as default, i wonder why netbase , sorry Jul 23 12:30:44 Genesis: ah, I'm pleased you found the problem! :) Jul 23 12:31:46 sorry Jul 23 12:41:15 shit, any form of rebuild on glibc seems to be broken Jul 23 12:47:53 Hello, I would like to know where can I get the list of packets supported in openembedded distribution ? I search on the wiki but i didn't find Jul 23 12:50:39 anyone know where limits.h gets messed up so glibc cant be rebuilt? Jul 23 12:51:58 Anyone been able to successfully build qmake2-native-2.10a? Jul 23 13:06:33 XorA: I think this happens when you just try to rebuild a recent glibc without the glibc-initial and intermediate stuff Jul 23 13:08:13 RP: happens everytime someone bumps glibc PR :-( Jul 23 13:08:31 XorA: which gcc? Jul 23 13:08:38 4.1.2 Jul 23 13:08:49 hmm, there goes a theory :) Jul 23 13:09:00 I merged OM with OE yesteday so shouldnt be anything different Jul 23 13:09:16 XorA: yes, poky sees the same problem Jul 23 13:09:25 I'd perhaps ask Khem to have a look at it and see if it happens with his toolchain changes Jul 23 13:09:34 if it does it needs fixing, if not, it will go away Jul 23 13:09:43 cool Jul 23 13:17:35 Hello, I would like to know where can I get the list of packets supported in openembedded distribution ? I search on the wiki but i didn't find Jul 23 13:17:40 anybody there ? Jul 23 13:18:23 Drakou: Packets? You mean packages? Jul 23 13:18:35 sorry packages Jul 23 13:18:39 yes Jul 23 13:18:46 Drakou: there is no list of supported packages. But most of the packages in the repository should work. Jul 23 13:19:20 and, hum, where can i browse the repository ? Jul 23 13:20:50 Drakou : i can help you in french if you want Jul 23 13:21:06 thank you Genesis Jul 23 13:21:21 Drakou: http://gitweb.openembedded.net/?p=org.openembedded.dev.git;a=tree;f=packages;h=66bf103642726610ae15fb4aa0c5e27da6134691;hb=HEAD Jul 23 13:21:48 i'm on colleague that comes from same school as you ,it's great scientist :) Jul 23 13:21:58 s/m on/have a Jul 23 14:47:54 Oke, can _anyone_ try to compile the qt-embedded package? Jul 23 14:47:56 *Ok Jul 23 14:48:11 Knowing if it's just me that's having the problem would help me tremendously Jul 23 15:09:57 hello, i've a problem: http://pastebin.com/m6eb358ae roughly it's a scons problem... Jul 23 15:10:56 i'm trying to build wesnoth.... Jul 23 15:17:42 is there a good overview of how 'overrides' work (if thats the term)? Specifically I'm interested in fully understanding what vars can be specified per machine, per architecture, etc (ie TARGET_FPU_arm = 'soft' to specify TARGET_FPU if arch is arm) Jul 23 15:19:15 arch, machine, distro, target os/sys Jul 23 15:20:29 is that the order of precedence as well? Jul 23 15:21:04 and package right? Jul 23 15:30:00 iirc it is in OE manual Jul 23 15:31:37 found a workaround... Jul 23 15:32:56 but now it fails at | Checking for Simple DirectMedia Layer library version >= 1.2.7... no but i have org.openembedded.dev/conf/distro/include/preferred-om-2008-versions.inc:PREFERRED_VERSION_libsdl-x11 ?= "1.2.9" and DEPENDS += "libsdl-image libsdl-mixer libsdl-net libsdl-ttf zlib-native"...mabe the zlib is wrong...must check Jul 23 15:33:47 hrw, you mean this? http://bec-systems.com/oe/html/special_overrides.html (empty stub) :P Jul 23 15:37:11 and i have 1.2.9 inside ./staging/arm-angstrom-linux-gnueabi/shlibs/libsdl-x11.ver Jul 23 15:37:26 what should i do now? Jul 23 15:38:24 ~oedoc Jul 23 15:38:33 ~oe manual Jul 23 15:38:38 ~oemanual Jul 23 15:38:40 argh. Jul 23 15:38:42 heh Jul 23 15:38:51 now you see my issue... Jul 23 15:39:11 tharvey: http://bec-systems.com/oe/html/index.html Jul 23 15:39:33 hrw, yes, but I haven't found a description of variable overrides there Jul 23 15:40:09 hm. I thought that it was documented Jul 23 15:40:21 conveniently its a stub Jul 23 15:40:44 its a feature of bitbake right? looking through the bitbake manual I haven't found it yet either Jul 23 15:49:20 how would I disable task-base-bluetooth? seems that task-base.bb adds it to PACKAGES without looking at MACHINE_FEATURES or DISTRO_FEATURES Jul 23 15:49:55 but machine has pci/pcmcia/usbhost? Jul 23 15:50:40 yes... I see that if one of those exists its also added... can I still kill it somehow? Jul 23 15:50:53 edit task-base Jul 23 15:51:54 was afraid you were going to say that... I'm trying to make minimal changes - theres no variable that can used to remove it from a list? Jul 23 15:54:07 I do have my own distro... and I've not set bluetooth in DISTRO_FEATURES so I'm not sure why ADD_BT would be getting set Jul 23 15:54:30 anyone that can help me? Jul 23 15:55:16 here's my debug output: http://rafb.net/p/P02xaY83.html Jul 23 15:56:29 Gnutoo, looks like you need sdl > 1.2.7? Jul 23 15:56:58 tharvey, the problem is that i have it... Jul 23 15:57:12 tharvey, i have the 1.2.9 Jul 23 15:57:27 ./staging/arm-angstrom-linux-gnueabi/shlibs/libsdl-x11.ver says 1.2.9 Jul 23 15:57:52 so perhaps scons is looking in the wrong place Jul 23 15:58:12 not real familiar with scons - should that prefix be '/usr'? Jul 23 15:58:17 tharvey, yes but how do i make it look in the good place and where is the good place... Jul 23 15:58:20 or something relative... Jul 23 15:59:26 tharvey, don't know...i just exported theses 2 variables in order to make it work: http://rafb.net/p/Kg97eo26.html Jul 23 16:01:22 Gnutoo, the wesnoth_svn package I see in my oe tree uses autotools - so your trying to perhaps fix the recipe b/c the source now uses scons instead? Jul 23 16:02:01 tharvey, they will drop autotools so they will only use scons or cmake... Jul 23 16:03:15 scons suxx Jul 23 16:03:22 cmake is better Jul 23 16:03:39 ah ok...so i'd better use cmake? Jul 23 16:03:40 Gnutoo, is there reason to believe that the scons class works in oe? - I only see 2 recipes that inherit it Jul 23 16:04:15 tharvey, i compiled another recipe and it worked... Jul 23 16:04:19 scons was working Jul 23 16:04:19 g Jul 23 16:04:31 cmake 2.6.0 has cross-compilation support which works Jul 23 16:05:11 perhaps the scons class doesn't work with libsdl or something... Jul 23 16:05:34 seems to me that its probably not setting up all the include dirs right Jul 23 16:06:03 ah: i have to use scons because of that: "but for the moment we are using SCons. When and if a CMake recipe is written we will do a technical comparison and make a final decision." (from http://www.wesnoth.org/wiki/CompilingWesnoth) Jul 23 16:06:57 i could also use autotools but i would need to patch a makefile in src/ after it has been generated...is it possible? Jul 23 16:06:58 Gnutoo, look at classes/scons.bbclass - see if there is perhaps something else that should be on the compile statement Jul 23 16:07:36 Gnutoo, figure out how to build it by hand, referencing the proper staging dirs, then look at how your cmdline differs from what would occur from scons.bbclass::scons_do_compile() Jul 23 16:08:01 SCons doesn't entirely do cross-compilation. Jul 23 16:14:28 bye guys Jul 23 16:16:20 tharvey, my prefix is /usr...is it normal? Jul 23 16:16:46 (prefix="${prefix}") Jul 23 16:18:34 Gnutoo, not that familiar with scons but I think thats wrong - typically prefix is where to install stuff, and you want it relative not /usr - the fact that there are no include dirs specified with relative paths seems wrong too Jul 23 16:18:54 Gnutoo, like broonie said I guess scons doesn't entirely do cross-compilation Jul 23 16:19:02 ok Jul 23 16:20:04 Gnutoo, I would start asking on a scons channel or forum - figure out if/how you can build your app manually with scons and then work back from there by fixing scons.bbclass Jul 23 16:20:16 assuming its broken... I may not know what I'm talking about... Jul 23 16:20:24 tharvey, ah...i have found a switch: sdldir Jul 23 16:20:44 what dir should i put? Jul 23 16:20:56 switch for scons? Jul 23 16:21:00 yes Jul 23 16:21:30 you would put the dir you found in staging that it installed to Jul 23 16:22:36 edit scons_do_compile in scons.bbclass and add that switch... to do it correctly you'll want to use vars that describe the location of the staging area versions etc - I'm not the one to ask there as I'm still learning oe Jul 23 16:23:15 ok Jul 23 16:23:18 thanks a lot Jul 23 16:27:54 the problem is that in /oetmp/rootfs/usr/include i have nothing...and i suspect that the rootfs is only used by images... Jul 23 16:39:29 hello, is there anyone that knows well oe: what should i set the prefix to...in order to make it find sdl... Jul 23 16:40:06 i've tried ${prefix} and ${D}${prefix} Jul 23 16:40:10 and it doesn't work Jul 23 16:50:11 Gnutoo, rootfs/usr/include would be for runtime builds... your trying to build something prior to runtime - you should end up with libs in rootfs but not includes - for pointing scons to your sdllib it would point somewhere in build/tmp/staging Jul 23 16:51:06 Gnutoo, I doubt 'prefix' is what you need to set - typically thats where to install things to after they are built - your issue is that scons can't locate the libs, you said you found that you can provide sdldir for that... Jul 23 16:52:09 wow...staging...thanks a lot... Jul 23 16:52:30 i will look if there is a switch for the include dir Jul 23 17:09:13 i've tried sdldir="${prefix}",sdldir="/home/gnutoo/oetmp/work/armv4t-angstrom-linux-gnueabi/libsdl-x11-1.2.9-r5/install/libsdl-x11/usr/,sdldir="/home/gnutoo/oetmp/work/armv4t-angstrom-linux-gnueabi/libsdl-x11-1.2.9-r5/install/libsdl-x11/usr/lib...none work... Jul 23 17:13:54 ah i have forgetten the \ for the ${prefix} ...i'll retry Jul 23 17:21:00 doesn't work Jul 23 17:27:53 what is strange is that it does a do_configure(enven if i put a blank do_configure with a oewarn it doesn't execute it) Jul 23 17:28:00 tharvey, ^^^ Jul 23 17:29:03 Gnutoo, can you build it manually outside of OE via scons and proper reference to sdllib? Jul 23 17:29:18 be back in a few Jul 23 17:29:29 cmake is really starting to annoy me Jul 23 17:29:40 Gnutoo: Are you sure that SCons is paying a blind bit of notice to what you're doing? Jul 23 17:30:34 kergoth`work: cmake always annoys me :P Jul 23 17:30:46 first time messing with it Jul 23 17:30:49 seems so messy Jul 23 17:31:04 broonie, evrytime i bitbake -c clean wesnoth then run bitbake wesnoth twice because the second time i have something like: | Saved options: default_targets = 'test,wesnoth,wesnothd', lowmem = True, prefix = '/usr', python = False, gui = 'tiny', sdldir = '/home/gnutoo/oetmp/work/armv4t-angstrom-linux-gnueabi/libsdl-x11-1.2.9-r5/install/libsdl-x11/usr/' Jul 23 17:31:41 or mabe i should use automake instead of scons... Jul 23 17:32:06 If wesnoth has an automake based build system that would be vastly easier to cross build with, yes. Jul 23 17:32:42 it has but it will soon be deprecated... Jul 23 17:33:33 i come back soon Jul 23 17:48:45 can we have multiple do_patch() that runs at different times? Jul 23 17:50:35 broonie, ^^^ Jul 23 17:54:53 Gnutoo: do you need to patch something after unpacking an inner zip or something like that? Jul 23 17:55:21 rschuster, no i need to patch after autoconf and automake.... Jul 23 17:55:35 but i also want to patch before... Jul 23 17:56:24 Gnutoo: patching after do_configure isn't possible (yet) Jul 23 17:56:42 Gnutoo: perhaps you can use a little bit of sed magic Jul 23 17:56:51 you really dont want to do that. patching generated files is not a good thing Jul 23 17:56:56 rschuster, yes i can use sed Jul 23 17:56:58 fix it correctly Jul 23 17:57:50 Gnutoo: btw. what are you trying to do? building the wesnoth_svn recipe we have in OE or a different version? Jul 23 17:57:58 kergoth`work, ok but here's my problem....i have some /usr/include...but they are not in Makefile.am...there is a directive in Makefile.in that generate it(because it's the default) then it's inside Makefile Jul 23 17:58:07 I wanted to look at providing a recipe for 1.4.3 ... Jul 23 17:58:38 rschuster, a svn version...but mabe i should do the 1.4.3 recipie Jul 23 17:58:41 Gnutoo: I'd be willing to bet its in Makefile.am, just as a @foo@ variable, and you're missing it. Jul 23 17:59:03 for example, you never want a -I@includedir@ in a package you're crosscompiling Jul 23 17:59:44 Gnutoo: the SVN recipe we have in OE is not very good Jul 23 17:59:53 rschuster, ok Jul 23 18:00:11 but was compiling a newer svn version...the last one... Jul 23 18:00:16 ok i change to 1.4.3 Jul 23 18:00:24 Gnutoo: I looked at how debian splitted the packages and wanted to provide something similar (wesnoth-server, wesnoth, wesnoth-data, wesnoth-httt, etc) Jul 23 18:00:34 ok Jul 23 18:01:09 Gnutoo: furthermore you can just to "apt-get source wesnoth" and see how they got it compile :) Jul 23 18:01:33 rschuster, but they don't cross-compile it no? Jul 23 18:01:52 * rschuster gets information on how to build a certain package from Debian all day :$ Jul 23 18:01:57 and i don't have a debian machine but i've got an ubuntu machine.... Jul 23 18:02:54 Gnutoo: no they do not cross-compile but reading their 'rules' file usually helps a lot Jul 23 18:03:08 ok Jul 23 18:03:08 Gnutoo: ubuntu is debian unstable + some extras Jul 23 18:04:05 Gnutoo: additionally you can always manually downloads the source packages from pages like http://packages.debian.org/sid/wesnoth Jul 23 18:04:26 by the way how do i get a package that is on sourceforge ? Jul 23 18:04:32 Gnutoo: look they have 1.5.1 in experimental: http://packages.debian.org/experimental/wesnoth Jul 23 18:04:55 SRC_URI = "${SOURCEFORGE_MIRROR}//" Jul 23 18:05:06 grep for SOURCEFORGE_MIRROR :) Jul 23 18:06:26 thanks,i grepped for 'sourceforge' and not SOURCEFORGE_MIRROR Jul 23 18:09:48 Gnutoo: "grep -i" is your friend :) Jul 23 18:09:58 ignores case Jul 23 18:10:17 yes...i know but i thought it was something like sourceforge:// Jul 23 18:15:01 i'll eat...i'll be back soon Jul 23 18:20:03 fyi, the reason i couldnt fetch anything successfully the other day was my .wgetrc Jul 23 18:20:14 the oe calls to wget obey it.. there's no choice there, wget doesnt seem to have an option to do otherwise Jul 23 18:20:22 although.. Jul 23 18:59:10 03koen 07org.oe.dev * r87bf384f... 10/ (5 files in 3 dirs): linux omap2 git: refresh mru's clock patches Jul 23 19:20:04 Gnutoo, so you ditched scons and are now tweaking or writing autotool files for your recipe? Jul 23 19:20:54 I'm trying to write autotools files for a simple app that doesn't use autotools (and needs to as it builds a small C app to determine the size of a pointer) Jul 23 19:21:13 I'm not autotools savy though, so I'm going through a learning curve Jul 23 19:24:17 tharvey, yes i am now using autotools.... Jul 23 19:33:10 tharvey, in the release there is even no autoconf....there is only configure and make Jul 23 19:41:58 tharvey mabyee a simple Makefile is enough Jul 23 19:42:06 not always autotools are useful Jul 23 20:05:59 tharvey, now it says: | checking for Boost headers version >= 1.33...but boost just get installed as a depedency Jul 23 20:07:52 http://pastebin.com/m2e15e71a Jul 23 20:08:08 hm??? Jul 23 20:08:43 hm you try to fix/add battel of wesnoth? Jul 23 20:08:47 battle Jul 23 20:09:05 yes Jul 23 20:09:11 version 1.4.3 Jul 23 20:09:47 could you please pastebin config.log which is in the topsourcedir Jul 23 20:12:24 woglinde, http://www.pastebin.ca/1081129 Jul 23 20:14:38 CROSS COMPILE Badness: /usr/include in INCLUDEPATH: /usr/include Jul 23 20:15:29 woglinde, ah ok...i should fix that... Jul 23 20:15:42 could you check the configure.in/ac if this came from there? Jul 23 20:16:41 ok i'll try to fix it and report if i have success or problems Jul 23 20:20:02 woglinde, in the compilation logs i have: --includedir=/usr/include --oldincludedir=/usr/include Jul 23 20:20:17 woglinde, do you have a suggestion for dealing with the fact that the app I'm doing a recipe for compiles an app to determine the size of a pointer? thats a clear use for autotools right? autotools seems overly complicated though.... Jul 23 20:21:05 * * OE Bug 4452 has been created by clarson(AT)kergoth.com Jul 23 20:21:07 * * Fetchers not run with a clean environment causes wget to obey user wgetrc. Jul 23 20:21:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4452 Jul 23 20:21:52 first bug against oe in the bugzilla ever. go me. Jul 23 20:21:54 and i don't find anything in configure.ac...i'll pastebin it Jul 23 20:22:00 kergoth hehe Jul 23 20:22:45 gnutoo hm this seems fairly false, can you provide the .bb? Jul 23 20:23:00 http://www.pastebin.ca/1081140 Jul 23 20:23:06 ok i'll pastebin the bb Jul 23 20:23:51 here it is: http://pastebin.com/m2e4383d3 Jul 23 20:24:08 it's a quite simple bb file Jul 23 20:28:29 in Makefile.in there is things such as includedir = @includedir@ Jul 23 20:28:41 s/is/are Jul 23 20:31:07 sheesh, glibc locale generation sure takes a while Jul 23 20:31:35 autotools seems very complicated... can anyone point me to a 'modern' howto for a simple c-app? I keep finding howtos that seem to be very dated Jul 23 20:32:08 tharvey: you probably want http://sources.redhat.com/autobook/ Jul 23 20:32:17 there are multiple example projects, and an online book Jul 23 20:33:00 cool thx... I'll chew on that for a bit Jul 23 20:33:30 tharvey: http://bec-systems.com/web/content/view/95/9/ Jul 23 20:33:32 still a couple years old (2006 was the last update), but its a starting point Jul 23 20:33:59 cbrake, yes, I looked at your example but it didn't quite work for me Jul 23 20:34:40 I hope I'm going about this right - I'm not sure I know how to actually use autotools to determine the size of a pointer on the target Jul 23 20:36:06 tharvey: http://www.gnu.org/software/autoconf/manual/html_node/Generic-Compiler-Characteristics.html#Generic-Compiler-Characteristics Jul 23 20:36:22 woglinde, ^^^ Jul 23 20:36:39 tharvey: the autoconf/automake manuals are quite thorough. read up Jul 23 20:36:47 once you've got the basics, anyway :) Jul 23 20:37:23 cbrake, with your simple example running autoreconf complains about some things missing Jul 23 20:37:42 tharvey: that should prrobably be "autoreconf -i" Jul 23 20:37:54 tharvey: I'll change the article Jul 23 20:39:21 cbrake, that helps but not completely http://www.pastebin.org/56477 Jul 23 20:40:21 tharvey: I usually just create those files -- not sure if there is a switch to bypass those Jul 23 20:40:25 kergoth`work, ah.. so the macro's will add defs to config.h - thx Jul 23 20:41:03 they set shell variables you can use in the configure script, also Jul 23 20:41:54 "This macro now works even when cross-compiling. The unused argument was used when cross-compiling." huh, nice to see they support that themselves now, no need for our old patches to the core m4 files Jul 23 20:42:27 cbrake, ah I see - note that in your article too if you will Jul 23 20:42:51 kergoth`work: you mentioned cmake a while back -- so you still prefer autotools over cmake? Jul 23 20:42:58 tharvey, you don't need autoconf for the size of a pointer Jul 23 20:43:13 hi diego Jul 23 20:43:23 hi woglinde Jul 23 20:43:24 havent used it enough to judge yet, my instincts tell me they both suck, but.. :) Jul 23 20:43:36 deigo tommorrow I think apr will compile again Jul 23 20:43:45 then with version 1.3.2 Jul 23 20:43:48 O_O nice Jul 23 20:43:59 diego, to determine the size of a pointer on the target? what do you suggest? Jul 23 20:44:05 have you found the problem? Jul 23 20:44:11 woglinde, what should i do now? Jul 23 20:44:22 tharvey, do it in the C code Jul 23 20:45:00 diego, my point is its someone else's app - I'm just trying to write a recipe for it - not fix the app Jul 23 20:45:09 something like sizeof(*int) Jul 23 20:45:21 diego, honestly I don't know what they are even using it for... Jul 23 20:46:14 less you use autoconf, more happier you are :D Jul 23 20:46:23 tharvey: done, thanks for the feedback Jul 23 20:46:57 gnutoo are there Makefile.am? Jul 23 20:47:11 woglinde, yes there is Jul 23 20:47:12 then Makefile.in are generated Jul 23 20:47:13 tharvey, btw you can fix C code using patch Jul 23 20:47:16 diego, your point is well taken, I'll look into it - perhaps a simple patch to the code would do better than patching in missing autotools files Jul 23 20:47:22 heh... yup Jul 23 20:47:28 beat me to it Jul 23 20:47:43 woglinde, yes normally Makefile.am->Makefile.in Jul 23 20:47:59 wait a moment Jul 23 20:48:06 I will give it a try here Jul 23 20:49:18 woglinde, ok thanks a lot Jul 23 20:50:33 okay will last while Jul 23 20:50:39 x is compiling Jul 23 20:50:53 ok Jul 23 20:51:06 tharvey, autoconf is usually more complex than required, so i prefer avoid it when i can Jul 23 20:52:50 and usually i use this, when i have to: http://sources.redhat.com/autobook/autobook/autobook_toc.html Jul 23 20:55:35 diego hm this outdated a bit Jul 23 20:55:51 I use directly the lates doku in autoconf automake and libtool Jul 23 20:59:00 hm the sdl deps of wesnoth reminds me that a have some work on the sdl bb to finish too Jul 23 21:00:40 ok Jul 23 21:02:18 *g* Jul 23 21:02:25 goodnight Jul 23 21:02:26 nono what is in oe now Jul 23 21:02:27 wirks Jul 23 21:02:30 nite diego Jul 23 21:02:40 600 taks to go Jul 23 21:02:42 tasks Jul 23 21:03:09 hm have to finish some documentation for 2nd last course of the summer term Jul 23 21:06:26 woglinde: i think with the latest pull, my libsdl problems got resolved, if you remember me complaining about that last week sometime Jul 23 21:10:12 kdub not really Jul 23 21:10:40 heh, well i did something to get it working. magic, maybe :D Jul 23 21:10:42 my brain has a small bandtape Jul 23 21:11:01 kdub what was the spultion? Jul 23 21:12:19 updating the mtn base i guess. i may have also changed things from DISTRO=angstrom2008.1 to DISTRO=openmoko Jul 23 21:12:48 now its failing on something python related... darn Jul 23 21:13:28 woglinde, i will try to sleep...do you want my mail(i'll add #oe to my the list of channels that are joined at startup) Jul 23 21:14:16 woglinde, how many tasks left? Jul 23 21:15:00 this time it looks like a 404 error from openmoko.org though, so i hope its just server downtime... Jul 23 21:16:43 400 Jul 23 21:16:45 re flo Jul 23 21:17:04 wahaha tcl is now build Jul 23 21:18:52 re Jul 23 21:19:12 woglinde, ok do you need my mail...and thanks a lot for your help... Jul 23 21:20:44 gnutoo hm not really Jul 23 21:20:47 if it works Jul 23 21:20:51 I will check it in Jul 23 21:20:55 woglinde, ok thanks a lot Jul 23 21:21:14 hm err Jul 23 21:21:17 if it compiles Jul 23 21:21:23 and produce some packages Jul 23 21:21:31 testing is than up to you Jul 23 21:21:32 *g* Jul 23 21:23:19 woglinde, ok...thanks i'll go Jul 23 21:24:23 bye Jul 23 21:25:33 hmm, why is DEPLOY_DIR_PSTAGE = in the bbclass, not ?= Jul 23 21:32:23 rp may knows Jul 23 21:43:03 shouldnt the -native packaged staging packages be in a more general area of the pstage deploy dir, since they're not bound to a specific target? Jul 23 21:44:59 hm time for the pokerchannel Jul 23 21:56:35 * Jay7 is preparing for linuxfest 10.0 :) Jul 23 21:59:53 anyone have any suggested low-level CPU/FPU/Memory type benchmark apps that would be suited to OE targets? Jul 23 22:00:39 memtester? Jul 23 22:00:54 cpuburn Jul 23 22:01:43 stream Jul 23 22:01:44 ramspeed Jul 23 22:01:56 those are more tests and loaders... I'm looking for something that could give you relative performance indexes that you can compare on other achs Jul 23 22:01:59 archs Jul 23 22:05:23 good evening Jul 23 22:05:34 (morning UGT) Jul 23 22:11:07 CosmicPenguin, those look interesting... have you used them on embedded systems across multiple architectures before? Jul 23 22:12:05 nope Jul 23 22:38:19 I have a fix for h2200-bootloader.bb, should I post it for review or just commit to .dev? Jul 23 22:38:51 I think this should be discussed before the commit Jul 23 22:39:13 ok, I will prepare a patch Jul 23 22:41:33 woglinde, i tried to sleep and it didn't work...did it build? Jul 23 22:41:46 nope Jul 23 22:41:52 sdl needs some love Jul 23 22:41:55 with libtool2 Jul 23 22:42:04 but I am working on it Jul 23 22:42:10 woglinde, wow...thanks a lot Jul 23 22:54:44 gnutoo wuah wesnoth is 145 mb Jul 23 22:55:38 woglinde, yes it is Jul 23 22:55:48 is it a problem? Jul 23 22:56:05 (other than a long download?) Jul 23 23:53:00 XorA|gone: RP this rebuildability problem is solved now Jul 23 23:53:19 thats was one factor that prompted me to fix the build sequence Jul 24 00:05:00 kehm cool Jul 24 00:05:29 * woglinde waits any second for the commit Jul 24 00:19:26 'night Jul 24 01:25:04 tharvey: I have used mediabench testsuite Jul 24 01:25:36 tharvey: err mediabench benchmark I meant **** ENDING LOGGING AT Thu Jul 24 02:59:56 2008