**** BEGIN LOGGING AT Thu Jun 19 02:59:57 2008 Jun 19 04:46:07 * * OE Bug 4379 has been created by lynn.lin(AT)avocent.com Jun 19 04:46:09 * * missing some tasks Jun 19 04:46:11 * * http://bugs.openembedded.net/show_bug.cgi?id=4379 Jun 19 06:33:32 morning Jun 19 08:07:35 slightly OT, but does anyone know of a nice mtd-utils tutorial? I'm mainly interested in seeing the whole chain of stuff that one needs to do in order to store the EBS and such (via sumtool) to flash (StrataFlash NOR in my case) Jun 19 08:36:34 hiya Jun 19 08:39:02 i'm looking to crosscompile some software that has a traditional makefile for a gumstix openembedded system. i have the gumstix basic image all built up with openembedded already, so the cross-compiler and all of the other packages have been built. however i haven't had much success making a bitbake package to use the makefile of the software i want to build. i was wondering if i could get at the paths of the crosscompiler bins and lib Jun 19 08:39:02 s and includes myself and just do things the old fashioned way without bitbake? Jun 19 08:49:27 frozeneskimo, do you know where your "tmp" for bitbake is? Jun 19 08:49:33 czr_, yup Jun 19 08:49:43 look in tmp/cross/bin Jun 19 08:49:46 oh duh Jun 19 08:49:48 why didn't i see that Jun 19 08:49:55 * czr_ shrugs :-) Jun 19 08:50:00 thanks czr Jun 19 08:50:00 just started couple of days ago myself Jun 19 08:50:05 yeah Jun 19 08:50:11 frozeneskimo, any idea what I should pass to mount to mount a jffs2 filesystem properly? Jun 19 08:50:19 mm Jun 19 08:50:23 -t jffs2? Jun 19 08:50:23 trying to use /dev/mtd3 (where I wrote the image with flashcp), but it fails Jun 19 08:50:25 hehe Jun 19 08:50:30 well, obviously -t jffs2 :-) Jun 19 08:50:43 nah not sure, sorry Jun 19 08:50:47 np :-) Jun 19 08:50:48 just curious Jun 19 08:50:54 what hardware you working wtih? Jun 19 08:51:04 almost the same that you are (if you have gumstix) Jun 19 08:51:08 yeah Jun 19 08:51:08 pxa270 (toradex colibri) Jun 19 08:51:15 cool Jun 19 08:51:19 doesn't bitbake take Jun 19 08:51:26 an ungodly time to build the basic image? Jun 19 08:51:39 or is it just me Jun 19 08:51:40 everything is relative. also remember to set your TMPDIR in local.conf ;-) Jun 19 08:51:49 the first image takes some time indeed. Jun 19 08:51:54 i just remember the traditional gumstix builds Jun 19 08:52:02 after that I find it relatively fast, although I have a lot of memory. Jun 19 08:52:06 being much quicker Jun 19 08:53:38 but cool Jun 19 08:53:46 thanks for the tip and pointing to /tmp/cross Jun 19 08:53:50 hmm. maybe I'm supposed to mount /dev/mtdblockN instead Jun 19 08:54:00 np :-) Jun 19 08:54:49 well it works. but.. hmmhm. Jun 19 08:55:08 can anyone please verify that jffs2 should be mounted from /dev/mtdblockN instead of the /dev/mtdN devices? Jun 19 08:55:23 can you do that from within a gumstix oe system? Jun 19 08:55:31 i can check the mount on a working oe verdex Jun 19 08:55:54 I could imagine so, but not sure. I'm running a nfsrooted and tftp-loaded kernel on my board (non-gumstix) Jun 19 08:56:06 ill boot it up and check Jun 19 08:57:34 well, I'd assume that if anyone is running anything with jffs2, it'd be enough to check /proc/mounts Jun 19 08:59:14 hmm Jun 19 08:59:22 i dont think theres anything helpful here but Jun 19 08:59:29 two lines that refer to jffs2 are Jun 19 08:59:34 /dev/root / jffs2 rw,sync,noatime 0 0 Jun 19 08:59:51 hmm. what do you have in /proc/cmdline? Jun 19 08:59:58 /dev/root /dev/.static/dev jffs2 rw,sync 0 0 Jun 19 09:00:07 console=ttyS0,115200n8 root=1f01 rootfstype=jffs2 reboot=cold,hard Jun 19 09:00:17 hmm. let me check :-) Jun 19 09:00:45 yeah, that's /dev/mtdblock1 Jun 19 09:00:48 thanks frozeneskimo Jun 19 09:00:54 np Jun 19 09:00:56 (root=1f01) Jun 19 09:01:00 yeah i figured Jun 19 09:01:19 let me know if you need anything else, that system will be up for a bit anyway Jun 19 09:01:36 frozeneskimo, I will, although I'm still trying to learn the whole OE setup and mtd-utils Jun 19 09:02:05 frozeneskimo, btw, did you do any jtag stuff with the verdex? Jun 19 09:02:19 not yet Jun 19 09:02:29 frozeneskimo, planning to? Jun 19 09:02:31 but the uboot got screwed up on two other vedex's Jun 19 09:02:39 * czr_ nods Jun 19 09:02:44 and i have a wiggler and the board and everything i just havent gotten around to reflashing yet Jun 19 09:02:52 ah. regular wiggler? Jun 19 09:03:05 yeah just a regular one built on a perf board hehe Jun 19 09:03:10 I had grief with amontec's usb jtagkeys, but I haven't really found out what the issue was (was using openocd) Jun 19 09:03:15 well a clone i mean Jun 19 09:03:18 * czr_ nods Jun 19 09:04:05 frozeneskimo, do you have any idea wether gumstix is going to push their stuff upstream to OE at some point? Jun 19 09:04:22 hmm no clue Jun 19 09:04:32 but i'd imagine they'd want to Jun 19 09:04:48 just because its all consolidated and less to worry about i guess? Jun 19 09:06:04 but yeah i've no clue to be honest Jun 19 09:08:23 morning Jun 19 09:08:29 morning hrw Jun 19 09:10:11 frozeneskimo, would make it easier to compare the various configs across different pxa270 devices Jun 19 09:11:01 yup true Jun 19 09:12:00 * czr_ shudders at the thought of developing board support for pxa320s Jun 19 09:12:12 there's scarcely any foss stuff for those. Jun 19 09:12:18 hehe Jun 19 09:13:55 cbrake, is there a reason why the -miwxxt tuning is disabled for cm-x270 in .stable? Jun 19 09:14:10 (no, I haven't tried enabling it yet, just curious) Jun 19 09:15:00 czr_: segfaults Jun 19 09:15:32 hrw, ah, that old lame excuse :-) Jun 19 09:15:43 czr_: and nearly no speedup Jun 19 09:15:56 czr_: we enable iwmmxt for mplayer as this show some speedup Jun 19 09:15:58 yeah, was wondering whether there'd be any Jun 19 09:16:32 I read somewhere that gcc doesn't yet know how to vectorize for the iwmmxt coproc anyway. gas supports it (so I hear), but there was something missing still. Jun 19 09:19:57 good morning Jun 19 09:20:02 mornink Jun 19 09:21:04 anyone here who knows how sumtool works? (mtd-utils) Jun 19 09:22:12 anybody know somekind of cflag that would prevent including posix1 limits/minimum-values like _POSIX_PATH_MAX? Jun 19 09:23:02 frozeneskimo, why do you want to do that? Jun 19 09:23:22 oh i don't Jun 19 09:23:27 (I'm not aware of any cflag for that specifically) Jun 19 09:23:29 im trying to find a reason that it's not including that Jun 19 09:23:47 are you using the correct include path? Jun 19 09:23:51 yeah Jun 19 09:24:21 .../gumstix-oe/tmp/staging/arm-angstrom-linux-gnueabi/include Jun 19 09:24:22 try running the cross gcc with -E and inspect what happens to the point of code that includes the limits.h Jun 19 09:24:30 ok Jun 19 09:24:38 -E only runs the cpp on the input Jun 19 09:25:57 righto Jun 19 09:29:49 hmm. mounting jffs2 takes ages (2 seconds). adding summary info doesn't seem to make it any faster (seems slower infact) Jun 19 09:32:00 czr_: .jffs2 is the summary version Jun 19 09:32:19 czr_: erm forget that wrong channel Jun 19 09:32:21 XorA|gone, hmm? I built my own jffs-image Jun 19 09:32:28 :-) Jun 19 09:32:48 make a massive difference on some machine Jun 19 09:33:17 hmm. not on this one. although I'm thinking that maybe the kernel doesn't support summaries (not OE kernel, yet) Jun 19 09:33:28 but it's 2.6.20, and summary support went into mainline ages ago (pre .20) Jun 19 09:33:28 :-D Jun 19 09:33:35 so I'm wondering.. Jun 19 09:33:50 I dont much of the details Jun 19 09:34:08 on the zaurus, summary never seemed to speed up boot, on gta02 device, it saves masses of time Jun 19 09:34:13 doesn't seem that many peiople do. mtd-utils and that stuff doesn't have too much odcumentation floating around Jun 19 09:34:17 maybe it depends on CPU and NAND read speed Jun 19 09:34:19 do they have NAND? Jun 19 09:34:25 cause I'm using NOR for this device Jun 19 09:34:48 NOR is faster to read isnt it? Jun 19 09:34:48 and I guess for NOR it wouldn't make a big difference sense, since they support random access sub-page-reads. at least I'm hoping that they do. Jun 19 09:35:05 kind of. depends. I guess :-) Jun 19 09:35:11 * czr_ is no expert Jun 19 09:43:03 ah. Jun 19 09:43:24 found the kernel config for the kernel that is running.. "# CONFIG_JFFS2_SUMMARY is not set" Jun 19 09:43:31 * czr_ curses into the general direction of austria and germany Jun 19 09:47:49 odd Jun 19 09:47:57 is JFFS2 suitable for NAND? should i be using yffs instead? Jun 19 09:48:00 i ended up just defining it manually and everything built Jun 19 09:50:47 kario, I guess the main grief with jffs2 is its scalability problems. so for large nand, probably not the best choice. although it's been used for ages.. Jun 19 09:51:07 I haven't used yaffs yet. there's also ubifs which is coming along. Jun 19 09:52:27 hi lrg Jun 19 09:57:54 hrw: I built tosa console-image (2.6.24 kernel) from stable branch, to trouble-shoot the unbootable problem. I replaced with zImage built from dev Jun 19 09:58:38 hrw: it boots, but showing "Could not load/lib/modules/2.6.23/modules.dep" Jun 19 09:58:42 you mean ?yaffs2 Jun 19 09:59:04 hrw: where does 2.6.23 come from? Jun 19 10:00:15 hey pb_ Jun 19 10:01:52 xjqian: "uname -a"? Jun 19 10:03:19 aha: 2.6.23 Jun 19 10:03:51 now you know answer? Jun 19 10:04:08 hrw: no, but checking my dev branch set-up Jun 19 10:07:04 methril, yes Jun 19 10:07:31 hrw: still no idea how 2.6.23 gets there. Jun 19 10:08:06 no idea either Jun 19 10:08:09 hrw: 2.6.24 was built from dev branch, weird Jun 19 10:08:13 hmm. is anyone aware of an uboot tutorial? Jun 19 10:08:25 me Jun 19 10:08:59 methril, hmm. riddle me this, does uboot decompress the kernel or the kernel decompresses it? as I understand, the kernel needs to be in 'uImage' format in order to load in uboot, right? Jun 19 10:08:59 see #u-boot channel or http://www.denx.de/wiki/UBoot/WebHome Jun 19 10:09:05 ah, thanks Jun 19 10:09:19 right Jun 19 10:10:19 the kernel is compressed in flash, it gives the data and decompress it Jun 19 10:10:40 you could store the kernel decompressed, and the boot up time is shorter Jun 19 10:10:57 but needs more flash space :) Jun 19 10:11:05 so, u-boot loads the compressed image into RAM, then jumps to the kernel, which is actually stub code that then decompresses the kernel into another location in RAM? Jun 19 10:11:34 yeah, I know these things in the x86/x86-64-world, but not in embedded. hence the silly sounding questions :-) Jun 19 10:11:57 hrw: my mistake, never mind. But still trying to figure out why the kernel built from stable fails to boot Jun 19 10:12:06 czr_ yes Jun 19 10:12:30 methril, excellent. thanks. any hints on how to generate an uImage formatted kernel? I'm assuming make bzImage isn't the way to go :-) Jun 19 10:12:47 czr_: look at packages/linux/linux.inc Jun 19 10:12:55 it's a start Jun 19 10:13:02 thanks Jun 19 10:13:21 then you could use mkimage, but are in oe (transparent) Jun 19 10:13:50 well, I'm not using oe kernel yet. I'm trying to live with the vendor provided one since it has board support. Jun 19 10:14:06 longer term plan is to integrate the necessary stuff into oe-driven process, but not quite yet. Jun 19 10:14:10 czr_: should be easy to put your kernel in OE Jun 19 10:14:36 XorA, easy is relative, just used oe for 3 days now ;-) Jun 19 10:14:41 as i see OE, if it works out of it, it works in OE Jun 19 10:15:02 XorA is a OE master, for him is easy :) Jun 19 10:15:44 mkimage is the tool that generates uImage formatted kernels? Jun 19 10:16:10 czr_: u-boot can load compressed kernel, unpack it and boot. or can give decompression task for loaded kernel Jun 19 10:16:19 hrw, which is better? Jun 19 10:16:33 * XorA is just a drunkard who staggered into the channel one day and they couldnt turf me out Jun 19 10:16:34 no idea - I am not uboot master Jun 19 10:16:43 ah well i feel refreshed, everything finally worked out after all Jun 19 10:16:47 I'm thinking of trying the lzma patches at some point, so I'm thinking I should move to the kernel decompression because of that. Jun 19 10:16:54 use the kernel decrompessor Jun 19 10:16:56 czr_ it depends Jun 19 10:16:57 u-boots code sucks Jun 19 10:19:18 hmm. is xuImage the correct KERNEL_IMAGETYPE? Jun 19 10:19:26 ah sorry, misread the script ;-) Jun 19 10:19:27 nm Jun 19 10:19:52 are you going to use lzma and you want to keep the kernel decompressed? Jun 19 10:20:03 if you want to save space, compress it too Jun 19 10:20:25 methril, lzma while on flash. uncompressed while in ram (obviously) Jun 19 10:20:43 ah! ok Jun 19 10:20:46 :) Jun 19 10:20:55 although I haven't yet decided on the lzma bit. depends on many things that I haven't tested yet. Jun 19 10:21:14 so I'ms starting with regular zlib. Jun 19 10:22:51 hmm. uImage leads to uncompressed image. what other choices are there? reading linux.inc it seems that if KERNEL_IMAGETYPE is anything else, it will be an compressed one Jun 19 10:24:58 if test -e arch/${ARCH}/boot/compressed/vmlinux ; then Jun 19 10:25:04 it hinges on that if statement Jun 19 10:25:30 ah yes. misread that too. Jun 19 10:34:15 XorA, is there any reason why one would want u-boot to decompress the kernel? I'm just thinking that it seems like unnecesary code in u-boot. Jun 19 10:34:35 czr_: havent found one yet, but maybe there is somewhere :-) Jun 19 10:34:41 * czr_ nods Jun 19 10:34:48 czr_: u-boot can boot non-linux too probably Jun 19 10:34:58 ah. that would be a reason. Jun 19 10:35:04 it can Jun 19 10:35:08 * XorA has used u-boot to boot u-boot Jun 19 10:35:40 hahaha Jun 19 10:35:59 you suggest another embedded boot loader? Jun 19 10:36:40 hmm. does u-boot support jffs2? Jun 19 10:36:47 (or have a driver for it) Jun 19 10:36:57 it's a boot loader, not an OS Jun 19 10:37:07 well, it has ext2 support. Jun 19 10:37:11 optional but still. Jun 19 10:37:54 yes, and jffs2 and VFAT Jun 19 10:37:54 good. more options then. Jun 19 10:37:54 I haven't really decided how to load the kernel yet, hence the pondering. Jun 19 10:38:18 for SMC /MMC cards it's usefull to have VFAT support (for testing) Jun 19 10:38:35 well, I'm testing using tftp/nfs anyway, much more convenient. Jun 19 10:38:49 but I guess for boards without ethernet, mmc would be the way to go Jun 19 10:39:33 hmm. otoh, mounting jffs2 will probably take ages with u-boot. /me sighs. Jun 19 10:40:23 i don't know i'll never test Jun 19 11:32:46 hi sebastian Jun 19 11:33:01 hi woglinde Jun 19 11:35:37 hi cedric Jun 19 11:38:25 bye Jun 19 11:39:50 ciao hrw Jun 19 11:46:46 gm mickeyl Jun 19 11:46:55 good morning Crofton! Jun 19 11:47:21 How's the eV? Jun 19 11:49:13 hi crofton Jun 19 11:49:16 good question Jun 19 11:49:18 ~seen rschuster Jun 19 11:49:22 rschuster was last seen on IRC in channel #oe, 8d 16h 3m 47s ago, saying: 'kario: it is also available freely (CC-BY-SA) as PDF'. Jun 19 11:49:37 hmm, I was hoping that was an easy question :) Jun 19 11:49:51 iirc rschuster has received the signed documents and needs to or already submit(ted) them to the authorities Jun 19 11:50:08 we chose him since he's in berlin most of the time Jun 19 11:50:11 crofton *g* Jun 19 11:51:26 * Crofton|work grumbles Jun 19 11:51:34 berlios is refusing connectons Jun 19 11:51:48 crofton this happend often Jun 19 11:51:54 yeah Jun 19 11:52:07 hm okay than we have to wait half year for the e.v. Jun 19 11:59:11 mickeyl, rschuster had some networks problems and was offline in the last days Jun 19 11:59:24 ok, thanks for the hint Jun 19 11:59:34 looking forward to seeing you in Bonn Jun 19 11:59:50 anything you want me to talk about in my talk? Jun 19 11:59:54 Jalimo? ;) Jun 19 12:00:21 I'm also awaiting the froscon ;-) Jun 19 12:00:49 jalimo would of course be good, but the middleware api would be good too! Jun 19 12:06:33 *nod* I'll talk more about the API Jun 19 12:06:45 will present some funny examples Jun 19 12:06:49 sms echo server or something Jun 19 12:07:19 ah, great. Jun 19 12:07:42 did you already do the announced release? Jun 19 12:08:11 yep Jun 19 12:08:23 two days ago milestone 1 was released Jun 19 12:08:25 barebones Jun 19 12:08:33 but works fine as a simple phone :) Jun 19 12:08:45 which is quite an achievement :D Jun 19 12:09:00 from there we will grow that fast Jun 19 12:11:57 where is it? Is there an image available? Jun 19 12:13:35 http://downloads.openmoko.org/framework/milestone1/ Jun 19 12:13:49 thx Jun 19 12:13:57 !oebug 4343 Jun 19 12:14:02 * * Bug 4343, Status: NEW, Created: 2008-05-30 15:52 Jun 19 12:14:04 * * mdarland(AT)pager.net: syslogd init script fails Jun 19 12:14:05 Has anyone experienced this? Jun 19 12:14:06 * * http://bugs.openembedded.net/show_bug.cgi?id=4343 Jun 19 13:05:21 re Jun 19 13:06:29 re hrw Jun 19 13:12:51 greetins....im trying to build for chinook and fakeroot_1.9.4.tar.gz is failing to be located....if i manually download it im not sure what to edit to get the process rolling again...any clues? Jun 19 13:20:28 zedstar put it in your archive dir and that should be okay Jun 19 13:20:42 you could too touch the stamp_file for it Jun 19 13:21:52 something like touch OEDIRPATH/stamps/i686-linux/fakeroot.do_fetch Jun 19 13:22:05 ups Jun 19 13:22:28 fakeroot-1.4.9-r0.do_fetch Jun 19 13:25:19 thanks will have a go...is there a simple way to just edit the broken url to the tarball or the repositories are not changeable on a package basis Jun 19 13:25:34 nope Jun 19 13:25:40 you have to edit the .bb file Jun 19 13:44:18 thanks woglinde that seems to do the trick Jun 19 13:45:08 no prob Jun 19 14:15:41 good morning Jun 19 14:15:51 hi sakoman Jun 19 14:16:10 Is anyone else seeing build issues with libsdl-x11? Jun 19 14:16:27 sakoman hm Jun 19 14:16:29 | oe_libinstall: cd src Jun 19 14:16:30 | oe_libinstall: install -d /home/sakoman/source/omap3-oe/tmp/staging/arm-angstrom-linux-gnueabi/usr/lib/ Jun 19 14:16:30 | FATAL: oe_libinstall: unable to locate shared library Jun 19 14:16:42 I am on the way to update sdl to the latest Jun 19 14:17:03 but havent found the time yet to commit my stuff Jun 19 14:17:15 Ah, then you might see the issue! Jun 19 14:17:20 hm and yes I think i fixed that localy Jun 19 14:17:33 let me try Jun 19 14:18:14 So sad when it it configures, compiles, and packages, and then can't stage :-( Jun 19 14:18:41 sorry 400 tasks to complete Jun 19 14:18:59 so it will last a while Jun 19 14:19:02 no problem, I'm patient ;-) Jun 19 14:21:38 args Jun 19 14:21:43 libtool-cross fails Jun 19 14:21:59 woglinde, how? Jun 19 14:22:12 install: cannot stat `arm-angstrom-linux-uclibc-libtool Jun 19 14:24:53 hi kgilmer Jun 19 14:26:46 good morning woglinde :) Jun 19 14:26:56 hm seems the prefix.patch dont do for cross Jun 19 14:27:03 libtool is there as libtool Jun 19 14:27:09 working on a recipe for Ant and Concierge Jun 19 14:28:42 kgilmer classpath cacao and co makes a lot of problems Jun 19 14:29:13 woglinde, like what? Jun 19 14:31:58 hm with newer autotools and cacao version it dont works anymore as rschuster has suggest it Jun 19 14:34:05 ic. i've read on some various blogs that there is some complexity in the various bootstrapping phases with cacao Jun 19 14:34:46 i'm still working on porting our application layer to Poky. once I feel good about that I'd like to get more involved with cacao/jamvm/classpath Jun 19 14:36:03 kgilmer are you using qt directly or qtopia? Jun 19 14:36:32 args I cant find the error in the libtool building Jun 19 14:36:35 we were using qt directly...it's still up in the air what we will be using. Jun 19 14:36:59 OE/Poky give us a lot more options than LTIB did (out of the box). Jun 19 14:37:24 ltib? Jun 19 14:38:37 http://savannah.nongnu.org/projects/ltib Jun 19 14:38:48 its a build system from freescale. Jun 19 14:39:06 an interesting approach and it worked well for us, but we've kind of out grown it. Jun 19 14:39:07 croton what was the updates from poky to libtool? Jun 19 14:40:23 are okay Jun 19 14:40:31 looked it up myself Jun 19 14:40:43 hm strange Jun 19 14:42:08 kgilmer, thanks Jun 19 14:48:29 crofton is libtool-cross-2.2.4-r23 okay at your side? Jun 19 14:48:42 it seems to be Jun 19 14:48:46 I ma glibc though Jun 19 14:48:57 here it ignors the host variable Jun 19 14:49:01 somehow Jun 19 14:49:16 so libtool is libtool Jun 19 14:49:29 :) Jun 19 14:49:33 hm I will delete all stuff Jun 19 14:49:38 and try again Jun 19 14:49:52 sakoman sorry it will last longer than I tought Jun 19 14:50:19 woglinde: OK, I'll be patient Jun 19 14:50:39 have other work I can do :-) Jun 19 17:03:35 hm Jun 19 17:03:44 now libtool-cross worked Jun 19 17:03:45 okay Jun 19 17:08:35 bye Jun 19 17:27:03 http://pastebin.com/m20e16bdb Jun 19 17:27:12 can anyone look at that and tell me how I might fix it Jun 19 17:32:38 anyone? Jun 19 17:36:24 sorry, over my head :( Jun 19 17:38:23 is anyone having any luck with berlios svn today? Jun 19 17:40:10 I can't get a new copy of bitbake Jun 19 17:44:07 Crofton|work: http seems to work, http://svn.berlios.de/svnroot/repos/bitbake/trunk/ Jun 19 17:44:43 hmmm Jun 19 17:44:52 I wonder if they turned off svn:// Jun 19 17:48:38 morning Jun 19 17:50:13 Crofton: yeah, I had trouble with svn:// access at berlios the other day Jun 19 17:54:24 anyone here with a NSLU2 and OpenSlug and can help me make a working bb file for APCUPSD? Jun 19 17:55:32 can you pastebin the error? Jun 19 17:56:13 yes, but i have a feeling the help provided won't help me, i'm too much of a newb Jun 19 17:58:34 * cbrake wonders what mdev + upstart would do to boot time ... Jun 19 18:04:28 Crofton: are you still here? Jun 19 18:05:04 apcupsd_3.14.4.bb: http://pastebin.ca/1051254 Jun 19 18:05:04 log.do_compile: http://pastebin.ca/1051257 Jun 19 18:05:04 log.do_configure: http://pastebin.ca/1051258 Jun 19 18:05:04 run.do_compile: http://pastebin.ca/1051260 Jun 19 18:05:04 run.do_configure: http://pastebin.ca/1051261 Jun 19 18:05:24 re Jun 19 18:05:43 sakoman why work on libsdl-x11 is okay its build here Jun 19 18:08:28 hi woglinde Jun 19 18:08:37 the problem is shown in log.do_compile: http://pastebin.ca/1051257 Jun 19 18:09:28 hi oh5 Jun 19 18:09:31 ups ph5 Jun 19 18:09:47 ph5 I could sent you my gettext-0.17 stuff if you want Jun 19 18:10:08 Crofton: =( Jun 19 18:10:31 Crofton: i can read, i don't know how to fix it Jun 19 18:11:18 woglinde: that'd be cool. on what distro/arch do you have it working? Jun 19 18:11:19 hm slugos use such an old linker? Jun 19 18:11:30 I know there is a problem, I know what log the resulting error is shown in, I know the problem comes from passing linker arguments to ld and not to gcc (I think), again, I don't know how to fix it Jun 19 18:12:32 scant try ti use newer binutils Jun 19 18:12:56 why do you think that will help? AND how do I use the newer binutils? Jun 19 18:13:24 because a newer linker knows the -rpath-link? Jun 19 18:13:50 is that a question or a statement? Jun 19 18:14:44 scant: you could try something like EXTRA_OEMAKE = "'LD=${CC}'" Jun 19 18:15:04 pH5: you're saying, just include that line and it should work? Jun 19 18:15:22 ph5 why should that help? Jun 19 18:15:28 gpsdrive/gpsdrive-pda_2.10pre2.bb does something similar Jun 19 18:15:40 woglinde: because ld doesn't recognize the -l option Jun 19 18:15:43 gcc will Jun 19 18:16:02 (and strip it and pass the -rpath-link through to ld) Jun 19 18:16:10 please, someone a long time ago, i've been at this for MONTHS, said the problem is coming from the scripts apcupsd provides, but working with the maintainer of apcupsd doesn't seem like a viable option, i don't have the skills necessary to relay the information for them to help Jun 19 18:16:12 args right sorry Jun 19 18:16:16 woglinde: not sure I understand. are you saying that current .dev works for you? Jun 19 18:16:16 scant sorry Jun 19 18:16:30 sakoman no my changes on libsdl-x11 works Jun 19 18:16:36 however, someone said months ago that apcupsd is passing linker args to LD and not through gcc and that's the problem Jun 19 18:16:44 scant ph5 is right and I was really wrong Jun 19 18:17:00 ah, OK. when will you be checking in the changes? Jun 19 18:17:15 hm, let me try it, if it's that simple, i'm very upset it tooks months for someone to say that, if you look in the channel logs, you'll see i've asked this question more than 10 times for more than 5 months Jun 19 18:19:32 pH5: that got me past the compile, YEY!!!!!!!! Jun 19 18:20:25 pH5: I'm getting this error now: http://pastebin.ca/1051282 Jun 19 18:22:10 scant: I guess your toolchain is broken, strip should be part of binutils. Jun 19 18:22:43 pH5: u sure? Jun 19 18:23:01 pH5: I need a NSLU2 expert to chime in Jun 19 18:23:16 sakoman not now I will update libsdl-qpe too Jun 19 18:23:46 woglinde: is there an easy workaround for the staging issue I'm seeing? Jun 19 18:24:07 If not, I will wait for your checkin Jun 19 18:25:31 sakoman ah hm Jun 19 18:25:37 pH5: does gpsdrive use binutils strip? Jun 19 18:25:57 it has to do with FILES_${PN} Jun 19 18:26:00 hm Jun 19 18:27:21 scant: sure? no, it's a guess. do you have the strip tool in your tmp/cross/bin directory? Jun 19 18:28:08 actually, I think packages should not strip binaries during install at all - isn't this supposed to be done during packaging so the debug info can be extracted for the -dbg packages? Jun 19 18:32:48 can anyone look at this: http://pastebin.com/m20e16bdb and let me know what I might be able to do to fix? Jun 19 18:35:25