**** BEGIN LOGGING AT Sat Jul 01 02:59:56 2006 Jul 01 03:00:23 maybe you should try a different boot loader? Jul 01 03:00:34 see what grub-install to the CF card gives you Jul 01 03:00:44 lilo is ... picky. Jul 01 03:02:21 I've been trying grub Jul 01 03:02:30 I'll try grub-install just for kicks (been trying manual approach) Jul 01 03:03:27 maybe I need --force-lba Jul 01 03:04:06 is your device.map all ok? Jul 01 03:04:12 haven't set one up Jul 01 03:04:17 is that necessary? Jul 01 03:04:21 grub command-line uses it Jul 01 03:04:30 otherwise it looks at /proc/mtab Jul 01 03:04:39 I'm doing Jul 01 03:04:43 root (hd1,0) Jul 01 03:04:54 install /boot/grub/stage1 (hd1) /boot/grub/stage2 Jul 01 03:04:58 and that's it Jul 01 03:05:05 it gets grub installed. It just can't load the kernel Jul 01 03:05:13 what error? Jul 01 03:05:20 corrupt filesystem Jul 01 03:05:26 try running setup Jul 01 03:05:31 using a stage2 Jul 01 03:05:33 setup doesn't work Jul 01 03:05:37 ? Jul 01 03:05:41 at all? Jul 01 03:05:45 do I need a stage1.5 to get ext2 reading support Jul 01 03:05:48 as I remember no Jul 01 03:06:00 I don't think you do Jul 01 03:06:05 I meant Jul 01 03:06:10 using the grub command line Jul 01 03:06:14 and the command setup Jul 01 03:06:28 to force in a stage2, which may solve your problems Jul 01 03:06:48 root (hd1,0) Jul 01 03:06:50 setup (hd1) Jul 01 03:06:52 I've tried that Jul 01 03:06:59 --force-lba may fix it Jul 01 03:07:04 will research Jul 01 03:07:15 setup --force-lba (hd1) Jul 01 03:07:30 grub-install is a bash script and I'm reading through it to learn its secrets Jul 01 03:09:12 this may seem silly Jul 01 03:09:26 but did you put your partition so close to the front of the drive that it's actually in the MBR? Jul 01 03:09:48 you want the partition with the kernel to be at least 64k away from the drive's start Jul 01 03:09:49 no, I didn't Jul 01 03:09:55 ok then Jul 01 03:10:09 I think that grub ext2 reader is detecting that the end of the ext2 fs doesn't match up with the end of the partition, possibly Jul 01 03:10:20 because I"m writing the ext2fs straight out to a partition that's bigger Jul 01 03:10:23 but still, that should work... Jul 01 03:10:29 should, yes Jul 01 03:10:40 i booted the board Jul 01 03:10:46 typed root (hd0,0) at console Jul 01 03:10:52 sees it as ext2 Jul 01 03:11:03 oooh Jul 01 03:11:13 kernel /boot/bzImage worked Jul 01 03:11:17 well Jul 01 03:11:21 it gave me an error Jul 01 03:11:33 try using find Jul 01 03:11:36 but it actually did auto-complete on bzImage Jul 01 03:11:38 find bzImage Jul 01 03:11:47 or rather find /boot/bzImage Jul 01 03:11:51 see what it says Jul 01 03:12:07 maybe this flash is defective Jul 01 03:12:17 possible Jul 01 03:12:33 wouldn't be surprised Jul 01 03:12:56 it says: Jul 01 03:12:58 (hd0,0) Jul 01 03:13:02 and sits there.... Jul 01 03:13:26 heh Jul 01 03:13:48 how much does a 256MB CF cost? Jul 01 03:13:55 * Borealid sighs as both xtscal and xfbdev go down together in flames Jul 01 03:13:57 nothing Jul 01 03:14:06 I saw one for $10 the other day, no rebate Jul 01 03:14:09 k Jul 01 03:14:12 I'll go buy one Jul 01 03:14:18 that was a good deal, mind you Jul 01 03:14:23 but they're still cheap Jul 01 03:14:27 best buy/circuit city have em? Jul 01 03:14:31 yes Jul 01 03:14:36 they'll be $30 there Jul 01 03:14:36 gonna go on an errand Jul 01 03:14:40 urg Jul 01 03:14:54 ooh, I saw a 2GB SD card for $20 Jul 01 03:15:03 with a $10 mail-in rebate Jul 01 03:15:07 that was luscious Jul 01 03:15:27 oh my goodness Jul 01 03:15:34 ts_calibrate works Jul 01 03:15:37 after all that! Jul 01 03:15:43 all I had to do was not use xtscal Jul 01 03:15:46 Jul 01 03:15:47 ok, gotta play tickle monster, then best buy Jul 01 03:15:56 thanks for the companionship during this difficult time Jul 01 03:17:29 heh, np Jul 01 03:24:49 * Borealid slaps forehead as xclock fails to build Jul 01 05:46:03 mornin koen Jul 01 05:49:08 hey gremlin[it] Jul 01 05:50:01 koen u remeber the ep9307 board i talked about ? ... now i have one with a 7" wide screen LCD :) Jul 01 05:51:54 cool Jul 01 05:52:13 did you see that lennert pushed a lot of stuff upstream? Jul 01 05:53:02 i'll check this weekend ... i have to made some test on the 2.4 kernel code ... and then i'll move to 2.6 so i can be in sync :) Jul 01 05:53:06 gremlin[it]: http://www.freelists.org/archives/linux-cirrus/06-2006/msg00069.html Jul 01 05:55:15 good ... i'll have to test watchdog :) Jul 01 05:55:56 and i'll work for RS485, touch, raster (maybe also accelleration) ... Jul 01 05:56:45 i'm tring to figure out if 9307 can be a divx player ... the 7" colot TFT is perfect for that :) Jul 01 05:57:44 root@poodle: uname-a Jul 01 05:58:21 Linux poodle 2.6.17 #1 PREEMPT :) Jul 01 05:58:40 now I just need to get usb working ....;) Jul 01 06:07:08 RP: any progress on poodle audio? Jul 01 06:40:33 HEY gremlin[it] there you are Jul 01 06:40:59 hi emte Jul 01 06:41:11 only problem is that i've now forgotten all the questions i had for you... Jul 01 06:41:22 hahah :) Jul 01 06:41:37 hmm Jul 01 06:41:39 i was quite busy so didn't connect here Jul 01 06:41:52 about h3600 and battery ? Jul 01 06:41:57 always happens durring the summer Jul 01 06:42:01 yeah Jul 01 06:42:14 hmm Jul 01 06:42:38 i cant remember any off hand, altho can you send me the .config your using for 2.6 ? Jul 01 06:43:09 i am not finding any type of standard one arround Jul 01 06:43:48 standard about ? Jul 01 06:44:16 semi-default config for the h36 devices Jul 01 06:44:43 not sure which options i should be selecting or not Jul 01 06:47:22 ok Jul 01 06:49:39 thnks Jul 01 06:49:59 is the one present in oe ! Jul 01 06:50:21 ? Jul 01 06:50:29 hmmm Jul 01 06:50:34 i dont have one in my OE Jul 01 06:50:55 just the pxa and h39 stuff Jul 01 06:51:20 on the handhelds-sa-2.6-2.6.12-hh0+cvs Jul 01 06:52:30 yeah i dont have that ... Jul 01 06:52:46 ozfam or dev? Jul 01 06:52:51 sure u have ... arch/arm/config/h3600_defconfig Jul 01 06:53:33 dev Jul 01 06:53:34 ah yeah i have that :P Jul 01 06:53:51 i dont have a close to recent version of .dev Jul 01 06:54:15 ahhh i sync about every day Jul 01 06:54:43 yeah, i dont really have a reason to anymore and i dont want to upgrade monotone Jul 01 06:55:03 otherwise i cant pull the ozfam branch when i need to Jul 01 06:55:03 i'm still using 0.25 Jul 01 06:55:20 i cant pull :( Jul 01 06:55:28 it changes.. the end of the month i belive Jul 01 06:55:39 sorry ... i can't push ... mistake :) Jul 01 06:56:24 lol Jul 01 06:58:59 anyway .. i have to check but i think i have do some work forward in respect to the cvs kernel ... Jul 01 07:02:19 but i still miss some work on data excahnge between sa1110 and atmel micro ... Jul 01 07:03:58 hmm Jul 01 07:17:37 anybody out there building a 2.6.17 kernel from the oz3.5.4.1 branch? Jul 01 07:18:19 doubtful Jul 01 07:18:42 that is a stablization branch Jul 01 07:19:15 aka work out bugs and ship Jul 01 07:21:36 ok, I was just trying to minimize the headaches to get a working bootstrap for poodle Jul 01 07:22:25 going to a cutting edge kernel would increase headaches immensly Jul 01 07:22:59 you are aware of the kernel developer's shift in policy are you not? Jul 01 08:05:54 morning Jul 01 08:06:51 morning Jul 01 08:07:29 I managed to build a 2.6.17 image for poodle, but it had none of the usb modules :( Jul 01 08:07:53 almost done with second attempt... Jul 01 08:08:01 poor PII's Jul 01 08:08:44 emte: no, what shift was that? Jul 01 08:10:40 CoreDump|home: Any news from RP on poodle sound? Jul 01 08:11:53 none Jul 01 08:12:13 hmm...ok Jul 01 08:17:47 CoreDump|home: Do you know which modules 2.6 on poodle uses for usb? Jul 01 08:18:28 I don't sorry. I don't even know if USB is supposed to work at all Jul 01 08:19:57 Ok, I was going to try and test it... still kinda new to this thing. Jul 01 08:22:34 morning Jul 01 08:23:30 mikey|semifinal: moin Jul 01 08:27:02 03mickeyl 07org.oe.dev * r5ef996f8... 10/packages/zile/zile_2.2.15.bb: zile 2.2.15 add latest version, patches no longer necessary Jul 01 08:31:57 morning mickey|semifinal Jul 01 08:41:01 * CoreDump|home just got an email about another OZ mirror offer: "No charge to you guys, obviously, and I'd only start Jul 01 08:41:01 raising eyebrows at the 1-1.5 TB mark bandwidth wise. " Jul 01 08:42:07 heh, cool Jul 01 08:43:10 in regards to mirrors we are set with 3-5 machines Jul 01 08:46:00 Whoooooo!!!! Jul 01 08:46:00 I can ping my poodle over USB on 2.6.17...:) Jul 01 08:46:06 root@poodle:~# uname -a Jul 01 08:46:08 Linux poodle 2.6.17 #1 PREEMPT Thu Jun 29 20:33:24 PDT 2006 armv5tel unknown Jul 01 08:46:10 hvontres: excellent! Jul 01 08:46:10 over ssh...:) Jul 01 08:46:16 CoreDump|home: I guess USB works in 2.6:) Jul 01 08:46:20 =) Jul 01 08:46:32 did you add an modules to defconfig? Jul 01 08:50:08 no, I just extracted the contents from modules-2.6.17-poodle.tgz Jul 01 08:50:30 and then did an "insmod g_ether.ko" from the right directory Jul 01 08:51:15 I was using RP's bb files from the .dev branch, not sure which rev tho Jul 01 08:51:24 nice Jul 01 08:51:49 Now, for my next trick... Jul 01 08:51:58 there are very few bugs left w/ poodle 2.6. It's shame that RP is so busy ATM :( Jul 01 08:52:31 I will try to see if I can't get kexec and alboot to dual-boot back to a 2.4 setup for normal usage Jul 01 08:52:46 Any hints? Jul 01 08:53:14 2.6 is known the "reconfigure" your flash partitions in a way that makes 2.4 oops... Jul 01 08:53:53 well, I have set up a loop image of my old root fs, will that help? Jul 01 08:54:03 it should, yes Jul 01 08:54:21 ohhh, I see, you can't get to /dev/mtdblk2 and 3 right? Jul 01 08:54:37 you can try =) but usually the kernel oopses Jul 01 08:55:05 hmmmm... I guess I'll be flashing back and forth then... Jul 01 08:55:26 I got kinda used to having my Poodle be my "My-Pod" Jul 01 08:55:28 you can edit your loop-file to not mount /home and you should be fine Jul 01 08:55:53 That's a thought.. Jul 01 08:56:08 so the zImage should also be on SD then? Jul 01 08:56:58 no, after kexec it doesn't matter where the zImage is Jul 01 08:57:32 and since kexec runs in 2.6, that should be fine. Jul 01 08:57:42 indeed Jul 01 08:58:31 Ok, maybe I'll try tomorrow. At least I can get my Z on the net from the cradel now. Jul 01 08:59:00 The old 2.4 kernel wouldn't work with Fedora....:( (MTU problems) Jul 01 08:59:23 2:15 am ---- time to go to sleep. Jul 01 08:59:32 see ya'll later Jul 01 08:59:33 n8 hvontres Jul 01 09:05:19 BTW: the drive light in 2.6 is really cool :) Jul 01 09:06:06 what is the best way to change CCFLAGS for configure? I have a configure (for libgcrypt) that fails because of overagressive optimisation Jul 01 09:07:54 eFfeM: fails as in miscompilation? Jul 01 09:08:22 * CoreDump|home is off to work Jul 01 09:08:54 zecke: configure tries to determine the size of long log, it does so by creating a static array of 256 elements, then uses size to determine the size (and divide by 256 using awk) Jul 01 09:09:09 hehe Jul 01 09:09:21 issue is that with gcc 4.1.1. with -funit-at-a-time and -Os the array is optimised away .... Jul 01 09:09:33 eFfeM: what about putting the size into the site script? Jul 01 09:09:43 site script ? Jul 01 09:10:22 check sites/ Jul 01 09:11:23 you lost me, I have no sites dir (I assume that is what you mean) Jul 01 09:11:46 oh, you mean site Jul 01 09:12:31 zecke the size info is there, but apparently it is not used Jul 01 09:12:44 because they invented there own check Jul 01 09:13:49 actually I was thinking of adding something like CFLAGS="" in EXTRA_OECONF, but not sure if that is going to work Jul 01 09:14:42 eFfeM: no it won't :) Jul 01 09:14:55 feared that (as they are probably additive) Jul 01 09:15:00 eFfeM: you can just set CFLAGS yourself (in the bbfile), or fix the configure check :) Jul 01 09:15:34 if I set the CFLAGS in the bb file that will also affect the compile, doesn't it Jul 01 09:16:09 I could fix the check in configure, but actually the configure is generated using automake etc and that is where I get lost Jul 01 09:18:16 is this the bad guy (from configure.ac): AC_CHECK_SIZEOF(unsigned long long, 0) ?? Jul 01 09:19:01 ah okay, I just remember Jul 01 09:19:08 this size of test, is from 'us' Jul 01 09:19:25 ok Jul 01 09:19:28 we could sanitize CFLAGS within this test Jul 01 09:19:38 but talk to kergoth or pb_ they are more clueful Jul 01 09:22:18 the issue is that eventually I end up with a config.h that says: Jul 01 09:22:18 #define SIZEOF_UNSIGNED_LONG 0 Jul 01 09:22:18 #define SIZEOF_UNSIGNED_LONG_LONG 0 Jul 01 09:22:31 which unsurprisingly does not work Jul 01 09:23:44 well do not use that -f optimisation Jul 01 09:24:27 zecke: yeah, but how ??? Jul 01 09:25:14 I can add CFLAGS= whatever to the bb file, but that will also affect the package Jul 01 09:25:22 a) update the autoconf patch b) set custom CFLAGS in do_configure (not too good) Jul 01 09:25:24 I'd rather only change the conf Jul 01 09:25:42 CFLAGS_do_configure = "MOO" <- but that is bad Jul 01 09:26:00 there is no autoconf patch and I am not too good in autoconf/automake/configure/... Jul 01 09:26:16 eFfeM: there is no patch to autoconf in OE? If you say that Jul 01 09:26:22 need to reboot Jul 01 09:27:03 zecke, not for this package ... Jul 01 09:27:13 cu after the reboot Jul 01 09:33:18 hi zecke Jul 01 09:33:47 re Jul 01 09:34:49 zecke what did you mean with autoconf patch, there is none in this package and my automake skills are not really good enough to know what I have to change Jul 01 09:36:56 http://www.openembedded.org/repo/org.openembedded.dev/packages/autoconf/autoconf-2.59/sizeof_types.patch Jul 01 09:37:09 but 1) find out why the site cache is not working Jul 01 09:37:42 now I try to get the akademy paper done Jul 01 09:38:08 zecke: tnx for the info and good luck with your paper! Jul 01 09:38:17 didn't know this url Jul 01 09:39:04 I think I can fix the patch Jul 01 09:39:16 awesome :} Jul 01 09:39:24 make it work with BSD size as well ;) Jul 01 09:39:38 lol, i don't think i have bsd size Jul 01 09:42:08 zecke, before making a bad change that affects everyone: my proposal is to change the patch by replacing Jul 01 09:42:08 [static const $1 x[[256]];])], Jul 01 09:42:10 with Jul 01 09:42:19 [static const $1 x[[256]];dummy(f)])], Jul 01 09:42:35 actually [static const $1 x[[256]];dummy(f);])], Jul 01 09:42:38 forgot the ; Jul 01 09:43:12 aarg: dummy(x) of course Jul 01 09:43:20 hehe Jul 01 09:43:28 this will cause a reference to x which defeats the optimizer Jul 01 09:43:30 ok? Jul 01 09:43:45 zecke: hello Jul 01 09:43:45 will it always use gcc for it? Jul 01 09:43:50 or will it use g++ as well? Jul 01 09:44:21 zecke, no idea, this is beyond the scope of my knowledge Jul 01 09:44:23 eFfeM: if it is using g++, it will complain about the lack of void dummy(...) Jul 01 09:44:28 hehe mine as well Jul 01 09:44:45 pb_: which compiler wil AC_TRY_COMPILE use? Jul 01 09:45:12 ${CC}, I think Jul 01 09:45:27 it's pretty easy to inspect either the .m4 file or the generated configure script to find out for sure Jul 01 09:45:48 or even (horror) the manual Jul 01 09:45:57 eFfeM: as pb_ didn't explode yet, I think this change would be okay Jul 01 09:46:10 lol Jul 01 09:46:56 the configure script I am using uses gcc (armeb-linux-gcc or so) Jul 01 09:47:47 don't know what a good way is to insert a #include otherwise I would replace dummy(x) with a library function Jul 01 09:48:20 no, dummy is fine Jul 01 09:48:39 maybe extern void dummy($1); but as it is using gcc it will compile anyway Jul 01 09:48:56 yeah, that could be nice Jul 01 09:49:05 better fix it fully Jul 01 09:50:25 ok, I ended up with: Jul 01 09:50:26 [extern void dummy($1); static const $1 x[[256]]; dummy(x);])] Jul 01 09:50:54 will test this and if it works for libgcrypt (where I faced the problem), I'll commit the change Jul 01 09:51:01 zecke: pb_ ok? Jul 01 09:51:51 Is sh part of the kernel image meta compile or part of gpe and opie images? Jul 01 09:53:23 lindenle: task-bootstrap should build busybox Jul 01 09:53:33 oh and sh is in busybox Jul 01 09:53:35 which should have sh builtin Jul 01 09:53:36 ok I forgot thanks Jul 01 09:56:51 pb_ there is another issue I'd like your advice on. I moved gcc for nslu2 to 4.1.1 and it works perfectly for normal arm code, but for thumb it gives bad asm Jul 01 09:57:13 saw your name while googling for solutions so thought you might have some suggestions Jul 01 09:57:28 eg I get: /tmp/cclUfDLr.s:42: Error: lo register required -- `ldr sl,.L_GET_GOT' Jul 01 09:57:40 which to me seems that the insn generator or the optimiser is no good Jul 01 09:58:11 this worked on 3.4.4. and I saw no patch missing that would deal with that Jul 01 09:58:47 zecke, pb_ the autoconf patch worked for me, will commit Jul 01 09:58:53 okay Jul 01 09:59:14 eFfeM: If things break, I will tell everyone you are guilty Jul 01 09:59:28 oh, thank you .... Jul 01 10:00:02 zecke: do you know if it is sh or ash? Jul 01 10:00:23 that's why I asked first. If a fix by any reason breaks a single package that is bad, but if it breaks all that is very bad Jul 01 10:00:45 the more impact a change has the more careful I try to be Jul 01 10:02:31 shall I write in the commit message "approved by zecke" :-) Jul 01 10:03:25 eFfeM: which size was taken for long long? Jul 01 10:03:50 8 Jul 01 10:04:27 #define SIZEOF_UNSIGNED_LONG_LONG 8 Jul 01 10:06:28 8 byte? my first computer didn't even have 8 byte... in total ;) Jul 01 10:06:45 eFfeM: and did it check for other sizes as well? Jul 01 10:12:08 zecke it did check for uns int, uns long uns long long uns short, all those sizes are ok Jul 01 10:12:28 zecke I'm sure your first compu had more than 8 bytes. Jul 01 10:12:53 actually the smallest I worked with is probably 128 bytes (ram that is) Jul 01 10:16:00 eFfeM: got me :) Jul 01 10:17:29 eFfeM: what sort of assembler problems does it produce? Jul 01 10:19:12 oh, right, that PIC thing Jul 01 10:19:17 that's odd, I thought it was fixed months ago Jul 01 10:19:46 and, for that matter, I thought it was broken in 3.4.4. Are you sure that thumb worked for shared libraries in that version? Jul 01 10:20:24 pb_ there are a few other error messages, but pastebin is not very cooperative today Jul 01 10:20:43 we have a uclibc distro, that one used to work but does not compile now Jul 01 10:21:36 actually I am not actively using the thumb set, but when I advanced gcc from 3.4.4 to 4.1.1. things broke for this variant Jul 01 10:22:52 pb_: pastebin is not working for me, but this is what I get (and apologies for pasting 4 lines) Jul 01 10:22:53 /tmp/cclUfDLr.s:42: Error: lo register required -- `ldr sl,.L_GET_GOT' Jul 01 10:22:53 /tmp/cclUfDLr.s:43: Error: dest and source1 must be the same register -- `add sl,pc,sl' Jul 01 10:22:53 /tmp/cclUfDLr.s:48: Error: invalid base register in load/store -- `ldr r4,[sl,r4]' Jul 01 10:22:53 /tmp/cclUfDLr.s:54: Error: garbage following instruction -- `add sp,sp,r4,lsl#2'/tmp/cclUfDLr.s:60: Error: garbage following instruction -- `add r3,r2,r1,lsl#2'/tmp/cclUfDLr.s:66: Error: invalid base register in load/store -- `ldr r0,[sl,r0]' Jul 01 10:23:16 it is in uclibc where it tries to compile ldso.c Jul 01 10:23:33 lunch is ready, back in 30 minutes or so Jul 01 10:39:47 back Jul 01 10:40:42 pb_ do you have an idea on how to fix or where to find a patch Jul 01 10:42:34 03eFfeM 07org.oe.dev * r27ae8c38... 10/packages/autoconf/ (autoconf-2.59/sizeof_types.patch autoconf_2.59.bb): Jul 01 10:42:34 autoconf: changed the sizeof patch a little bit; Jul 01 10:42:34 gcc 4.1.1 optimisations would cause this to fail Jul 01 10:42:34 this patch makes sure x is referenced and cannot be optimised away Jul 01 10:44:31 finally got through to pastebin: this http://pastebin.com/737673 gives some more context (eg compiler flags) Jul 01 10:56:57 eFfeM: I think you'd need to talk to the uclibc people about that. Jul 01 10:58:12 ok, tnx Jul 01 11:08:50 D'oh! Jul 01 11:09:00 Folks, have a look at qt4.2-tp1 Jul 01 11:09:10 mickey|semifinal: hehe Jul 01 11:09:13 The GraphicsView is - u n b e l i e v a b l e - Jul 01 11:09:25 mickey|semifinal: Nothing new :) Jul 01 11:09:25 * mickey|semifinal impressed Jul 01 11:09:29 mickey|semifinal: glib integration Jul 01 11:09:42 mickey|semifinal: QtDBUS Jul 01 11:10:34 mickey|semifinal: need to test performance on ARM :) Jul 01 11:11:13 03eFfeM 07org.oe.dev * re45e18cd... 10/packages/cyrus-sasl/ (cyrus-sasl-2.1.19/client.c.patch cyrus-sasl_2.1.19.bb): Jul 01 11:11:13 cyrus-sasl: removed static for local decl of a var mentioned in .h as Jul 01 11:11:13 gcc 4.1.1. does not like that Jul 01 11:15:03 mickey|semifinal: TT has other tough stuff in the pipeline Jul 01 11:15:33 mickey|semifinal: e.g. a Qt based terminal server system. That will kick the crap out of NX/X on low latency links Jul 01 11:15:39 (at least this is my impression) Jul 01 11:15:56 too bad that they blow up on Qtopia Jul 01 11:16:26 yeah Jul 01 11:16:35 for desktop i can't recommend anything else Jul 01 11:16:45 for embedded i don't think i can recommend it Jul 01 11:16:54 :( Jul 01 11:17:15 any suggestions on a distro to start with for a PXA270 platform? Do any of the zaurus machine/distros combinations build in dev? Jul 01 11:17:32 I used to used familiar-unstable ... Jul 01 11:17:58 cbrake: use openzaurus-unstable (for soft float) and it 'should' compile Jul 01 11:18:22 zecke: I just tried that. Just a min ... Jul 01 11:18:40 gpe-image for openzaurus unstable built for me yesterday Jul 01 11:19:17 DISTRO = "generic" Jul 01 11:19:23 should be the best option Jul 01 11:19:53 hi mickey|semifinal Jul 01 11:19:58 hey phil Jul 01 11:20:10 XorA|gone: zecke: I must have my machine set up incorrectly then. Jul 01 11:20:25 XorA|gone: what zaruii have PXA270's in them? Jul 01 11:20:31 cbrake: boxer? Jul 01 11:20:45 All cxx00 Jul 01 11:21:09 zecke, mickey|thesis ok, I'll give one of those a try --thx Jul 01 12:13:53 How do I reload the keymap after I change it? Jul 01 12:13:58 do I need to reboot? Jul 01 13:09:21 konq-emb is driving me nutters Jul 01 13:09:33 hehe Jul 01 13:09:58 that kde-config issue makes no sense, in either direction Jul 01 13:13:10 * chouimat is back. Jul 01 13:13:55 happy Canada Day everybody Jul 01 13:14:13 Is that today? OMG cant believe I forgot :P Jul 01 13:15:28 Founding day or ? Jul 01 13:15:57 How do I change the keymap? Jul 01 13:17:49 loadkeys? Jul 01 13:20:46 Is there good reason to recompile tslib anew every time? Does it need to be the CVS version? Jul 01 13:21:09 Laibsch, no Jul 01 13:21:23 set a cvsdate Jul 01 13:21:29 Laibsch: set a SRCDATE Jul 01 13:21:59 Laibsch: "send a patch" (not meant to sound mean, strange, angry) :) Jul 01 13:23:09 zecke: ;-) Jul 01 13:24:45 Laibsch: :) Jul 01 13:32:17 Where is the best place to set SRCDATE? In conf/local.conf? Jul 01 13:33:22 Laibsch, distro file or local.conf Jul 01 13:33:32 If you got your own distro file use that, otherwise local.conf Jul 01 13:33:36 Thank you. Jul 01 13:33:40 np Jul 01 13:33:56 I do not have my own distro (yet ;-)) Jul 01 13:34:09 what machine are you working against? Jul 01 13:34:30 collie Jul 01 13:35:35 one of the zaurus right? Jul 01 13:36:30 SL-5500, correct. Jul 01 13:37:07 I might start compiling stuff for the WRT54GL (But I guess only ipk is possible, not the complete boot image) Jul 01 13:37:39 Laibsch, there is a a couple of meta packages that builds complete boot images Jul 01 13:38:01 "bitbake bootstrap-image opie-image" Jul 01 13:38:26 That leads me to a question concerning GPL and publication of compiled stuff. Does any of you do that outside of the big projects? How do you handle the requirement to publish source? Do you keep a backup of the exact sources you used? Seems like a bit of a hassle. Jul 01 13:38:55 Kristoffer: The problem appears that the linksys router cannot be flashed in that way. Jul 01 13:39:57 Laibsch, I currently got a member looking into what requirements the GPL states, all sources is saved when compiling so its more of a space issue Jul 01 13:40:02 Laibsch, aha, understand Jul 01 13:41:10 yes, of course it is a space issue. but it is also a question of how to handle this. Having stuff and finding it are not the same ;-) Jul 01 13:42:27 Will probobly host them on the site, but I would rather not Jul 01 13:55:44 INAL strictly speaking the source should be in the same place Jul 01 13:56:20 this means you should put GNU make, GNU gcc... sourcecode of all used host utilities + BB Files + BB Patches + Sources into the feed directory :) Jul 01 13:56:30 I have never seen such thing though :) Jul 01 14:00:36 zecke, got a quick question for ya if you got time Jul 01 14:01:13 scummvm doesnt like 4.1.1, so thinking about setting CCSUFFIX, but the trouble is that should we force 3.4.4 version or detect if theyve used CCSUFFIX somewhere else? Jul 01 14:01:56 Or perhaps set different for different distros?..want to create as little of a hassle as possible Jul 01 14:02:06 Kristoffer: hmm Jul 01 14:02:41 Kristoffer: this is a short coming of OE. The proper fix is to make scummvm work, or mask it out Jul 01 14:06:57 Good evening Jul 01 14:07:08 morning sirfred Jul 01 14:07:23 ~ugt Jul 01 14:07:34 well, ugt is Universial Greeting Time. Created in #mipslinux, it is a rule that states that whenever somebody enters an IRC channel it is always morning, and it is always late when the person leaves. The local time of any other people in the channel, including the greeter, is irrelevant. Jul 01 14:07:45 eFfeM: :) Jul 01 14:08:45 zecke, how about if I destroy the "check for 4.x" part and then set a CCSUFFIX for distro specific? Jul 01 14:09:17 zecke, that way all gcc-3.x distros/users wont notice it and all others can set ccsuffix Jul 01 14:09:44 zecke what is the exact problem with scummvm? Jul 01 14:09:55 eFfeM, scummvm doesnt like gcc 4.x Jul 01 14:09:59 or want Jul 01 14:10:10 i've been working on patching other packages to work with 4.1.1 Jul 01 14:10:34 hmm, you just said something interesting, it just says that its not supported, not that it would work Jul 01 14:10:45 Kristoffer: generally this gives an error (e.g. external variable redeclared as static is one that I often see) Jul 01 14:11:09 Kristoffer: that is more serious Jul 01 14:12:15 eFfeM, so whats the best approach, setting ccsuffix and going with 3.x or forcing 4.x build/ Jul 01 14:13:12 ehm, I have never used ccsuffix, nor do I have any idea what this package does. Jul 01 14:13:43 I probably would force a 4.x build and see if it works and if so probably come up with a patch to change the test Jul 01 14:14:04 ccsuffix just adds suffix to CC.. making it use gcc-XXX instead of gcc Jul 01 14:14:21 eFfeM, oki will try that first then Jul 01 14:14:54 note that I am not an expert, but my general philosophy is to create as little exceptions as possible Jul 01 14:15:25 eFfeM: that matches mine Jul 01 14:16:04 Kristoffer: scummvm is C++, it is not fun to mix c++ libraries compiled with different versions of gcc Jul 01 14:16:07 So its basicly better to patch it to ignore gcc version and wether or not 4.1.1 works, fix that seperate later Jul 01 14:16:08 yeah there are a lot more goodies in that bag like "if it ain't broken, don't fix it" .... Jul 01 14:16:54 Kristoffer: I would say: patch locally, build, test and if it works make a nice patch and commit it Jul 01 14:17:08 And if it doesnt work? Jul 01 14:17:18 just checked what scummvm means, it definitely looks nice but it is not for nslu2 people Jul 01 14:17:35 if it doesn't work you have a bigger challenge (to find out why it is not working). Jul 01 14:18:00 If there is a good reason why this check is there I would expect it to be documented Jul 01 14:18:07 yeah, but im just trying to figure out if the "accept gcc>4.x" patch is needed in all scenarios Jul 01 14:18:32 no idea Jul 01 14:22:20 eFfeM, while looking at the configure script it noticed Jul 01 14:22:39 it doesnt say 4.x is bad Jul 01 14:23:05 it basicly just cant get the version out of gcc, due to changes "-1" is not working "should be "-n 1" Jul 01 14:23:14 so fixing the configure might be good for all scenarios Jul 01 14:25:48 Kristoffer: agree Jul 01 14:28:22 anything wrong with anoncvs.handhelds.org? I cannot download the source for libopie2 from anywhere. Jul 01 14:31:26 Laibsch: ask france :) Jul 01 14:35:19 Just browsed through the ChangeLog for bitbake and found in the first line that apparently it can send stuff to oe.pastebin.com? How? bitbake --help does not tell. Jul 01 14:35:40 zecke: I suspect the SRCDATE screwing up something. Jul 01 14:35:42 the interactive shell can Jul 01 14:36:06 OK, I do not use that (yet). Should I? Jul 01 14:38:10 no need to Jul 01 14:38:15 no need to use it Jul 01 14:41:23 03eFfeM 07org.oe.dev * r9c951cf8... 10/packages/reiser4progs/ (4 files in 2 dirs): reiser4progs: added patches to let it compile with gcc 4.1.1 Jul 01 14:41:46 OK Jul 01 14:42:14 Should I really put the SRCDATE stuff in build/conf/local.conf or is bitbake/conf/bitbake.conf better? Jul 01 14:42:56 local.conf Jul 01 14:43:11 Laibsch: normally the distribution you use sets a SRCDATE Jul 01 14:43:22 e.g. like angstrom is doing Jul 01 14:43:28 if they don't, put it into local.conf Jul 01 14:43:55 Laibsch: e.g. SRCDATE_tslib = "THEDATE" allows you to set the version for just one package Jul 01 14:51:31 hi koen Jul 01 14:51:47 been enjoying the nice weather ? Jul 01 14:53:28 ~weather eheh Jul 01 14:53:32 Eindhoven, Netherlands; (EHEH) 51-27N 005-25E 28M; last updated: 2006.07.01 1425 UTC; Dew Point: 53 F (12 C); Pressure (altimeter): 30.18 in. Hg (1022 hPa); Relative Humidity: 34%; Sky conditions: mostly clear; Temperature: 84 F (29 C); Visibility: greater than 7 mile(s); Wind: Variable at 3 MPH (3 KT) Jul 01 14:57:24 hey eFfeM Jul 01 14:57:46 I picked some stuff at my parents today Jul 01 14:58:01 and installed openslug 3.10 on my fathers slug as well :) Jul 01 14:58:06 hey, great Jul 01 14:58:17 actually you should teach him to do that himself .... Jul 01 14:58:44 he did all the steps, I just pointed at all the relevant wiki pages Jul 01 14:58:58 ah, ok Jul 01 14:59:27 did you also learn him to update the wiki pages with all things that were not correct ? Jul 01 14:59:32 heh Jul 01 14:59:36 one step at a time Jul 01 14:59:41 lol Jul 01 15:00:03 it took me a while to explain that 'optware' and 'unslung' howtos can be ignored Jul 01 15:03:02 btw, can you have a short look at ircp_0.3.bb ? configure does not work for me as it cannot find obex.h, but the file exists in staging/armeb-linux/include/openobex/obex.h Jul 01 15:03:19 do I need an EXTRA_OECONF with --prefix ?? Jul 01 15:04:00 koen, unrelated: has the nslu2 tinderbox problem been solved. Saw that the link is still there Jul 01 15:04:54 but didn't see the nslu2 builds on ewi Jul 01 15:05:15 and apparently nudi is out of disk space, but I cannot fix that as I do not have access Jul 01 15:05:19 zecke: OK, thanks. Seems much better to do it for just one package. Jul 01 15:11:24 koen, nevermind, added EXTRA_OECONF="--prefix=${D}" to the bb file and it worked Jul 01 15:15:21 has anyone built oe on mac-osx? Jul 01 15:16:47 * chouimat is away: LUNCH Jul 01 15:16:54 if this isnt the right channel for my question then please point me to the right place Jul 01 15:17:21 the wiki for building on osx is empty :( Jul 01 15:21:41 poushag, right channel Jul 01 15:22:34 thanks for the response kristoffer Jul 01 15:22:54 no probzie :D this channel is more or less active, usually more during nights Jul 01 15:24:59 USA timezone nights or EU nights? Jul 01 15:25:13 03kristoffer 07org.oe.dev * rd50b000e... 10/packages/scummvm/ (4 files in 2 dirs): Jul 01 15:25:13 scummvm/files/gcc-4.x.x-accept.patch: Makes scummvm approve of gcc 4.x.x Jul 01 15:25:13 * Adds gcc 4.x.x to accept list Jul 01 15:25:13 scummvm/files/sh3-linux-new-arch.patch: Makes scummvm know of sh3 Jul 01 15:25:13 * Adds linux-sh3 to the configure script of known archs Jul 01 15:25:13 scummvm/files/tail-obselete-fix.patch: sets -n 1 instead of -1 Jul 01 15:25:15 * Sets tail -n 1 instead of -1, kept for future reference. Jul 01 15:28:02 Kristoffer: wanna try to add scummvm 0.8.x to OE? Jul 01 15:28:15 or are you too busy playing beneath a steel sky ;) Jul 01 15:28:26 koen, hehe Jul 01 15:28:39 koen, will look at it :) Jul 01 15:28:41 poushag: we need to migrate wiki content Jul 01 15:28:54 poushag: it does not work, unless you turn your system into a GNU Jul 01 15:30:15 zecke, have you tried it? or do you have a url to a helpful site? Jul 01 15:30:53 poushag: I'm on OSX and FreeBSD, and it is currently not possible Jul 01 15:31:04 poushag: I think it would require a month of work to get it working Jul 01 15:33:30 so it sounds like the only other way to do this on the mac would be some kind of virtual machine or maybe a dual boot system (in order to run a full linux distro) Jul 01 15:34:18 or ssh :) Jul 01 15:35:13 well all my linux boxes are servers for my work = forget ssh Jul 01 15:35:37 and since my mac is ppc a vm is probably not realistic Jul 01 15:36:02 i dont even know what linux distro i could use on a ppc mac Jul 01 15:36:46 the vm might be ok if the linux distro is built for ppc Jul 01 15:39:26 where are the sourceforge mirrors set? Jul 01 15:40:16 Kristoffer: bitbake.conf Jul 01 15:40:20 Kristoffer: and base.bbclass Jul 01 15:40:34 the sources for scummvm isnt on heanet..the filename is correct Jul 01 15:40:39 its on prdownloads Jul 01 15:41:14 best way to fix? Jul 01 15:43:30 zecke, thx for the input - will reconsider feasibility of this...but it doesnt look very good Jul 01 15:45:32 Laibsch: ?? Jul 01 15:53:09 I was wondering if there are problems with anoncvs.handhelds.org and asked here. I cannot compile libopie2 because fetching the sources fails. Jul 01 15:53:26 zecke suggested to ask you, france. Jul 01 16:00:55 Laibsch: checking... Jul 01 16:02:17 Laibsch: I see 3 or 4 people download stuff, try again while I watch the logs. Jul 01 16:02:49 france: I think it looks more like some problem with my HD again. http://rafb.net/paste/results/G1X8hX83.html Jul 01 16:03:05 I guess I will need to e2fsck again. Jul 01 16:04:07 Laibsch: sounds like it. Jul 01 16:04:45 Laibsch: good luck on your fdisk quest. I know it can be painful. Jul 01 16:05:18 bb in a while, lunch time here. Jul 01 16:27:41 03kristoffer 07org.oe.dev * r95fa8952... 10/packages/scummvm/ (scummvm_0.9.0.bb scummvm.inc): Jul 01 16:27:41 scummvm/scummvm_0.9.0: Addition of new scummvm file Jul 01 16:27:41 * New scummvm file (upgrade from 0.6.1b) Jul 01 16:27:41 scummvm/scummvm.inc: Change of sourceforge mirror Jul 01 16:27:41 * Change of sourceforge mirror, heanet.dl didnt have scummvm. Jul 01 16:27:45 03kristoffer 07org.oe.dev * rda04e847... 10/packages/scummvm/files/sh3-arch-0.9.0+.patch: Jul 01 16:27:46 scummvm/files/sh3-arch-0.9.0+.patch: Patch for sh3 to > 0.9.0 scummvm Jul 01 16:27:48 * Addition of sh3 arch, should work for after also. Jul 01 16:28:07 calling it a night Jul 01 16:30:06 morning all :) Jul 01 16:31:52 hi Jul 01 16:36:25 still up? Jul 01 16:36:35 nevermind...:) Jul 01 16:37:28 Anybode know how to force OE to use the installed version of a program rahter than building it from scratch? Jul 01 16:37:32 hvontres|home: eh, it's 7pm here ;) Jul 01 16:38:02 that would be ASSUME_PROVIDED Jul 01 16:38:13 check local.conf, there should be examples Jul 01 16:38:34 CoreDump|home:thanks. Jul 01 16:38:39 np Jul 01 16:38:48 hows your poodle feeling w/ 2.6? Jul 01 16:39:31 CoreDump|home:I am trying to rebuild the bootstrap to include the usb stuff by default. Jul 01 16:39:54 I tried using the poodle.conf from .dev but it did not like me :( Jul 01 16:41:55 so I added the usb related RDEPENDS instead. Jul 01 16:42:57 CoreDump|home: How much longer does an opie-image take compared to a bootstrap image? Jul 01 16:43:18 To compile, that is Jul 01 16:43:45 don't know really. I very rarely build normal opie or gpe images Jul 01 16:44:00 but fwiw, hentges GPE + hentges Opie = ~8hrs ;) Jul 01 16:44:27 On what kind of Machine? Jul 01 16:45:56 P4 2.8 HT w/ 2G of RAM running off a RAID10 Jul 01 16:46:43 ugg...so my Poor Dual PII 333 with 512MB Ram....... an eternety.. Jul 01 16:47:17 hvontres|home: well normal opie images won't take nearly as long Jul 01 16:48:45 anyone ever tried to compile madwifi with gcc 4.1.1 ? Jul 01 16:49:00 I get errors about -mapcs-32 and -malignment-traps Jul 01 16:49:16 as I don't know where they are for I am a little bit reluctant to remove them Jul 01 16:59:09 CoreDump|home:so this should work, right: ASSUME_PROVIDED += "virtual/qemu-native" Jul 01 17:00:48 looks about right Jul 01 17:05:06 With a new OE.db is it still monotone 0.25-2 wich is preferd? Someone mension to my that you are moving over to 0.27, but is that version supported now Jul 01 17:06:05 CoreDump|home: hey, nice to catch you online Jul 01 17:06:53 hi there JustinP Jul 01 17:07:02 I'm having some strange problems with altboot Jul 01 17:07:19 it started with a previous installation that worked fine forever and now it's happening on a fresh one Jul 01 17:07:35 choosing normal boot says ERROR then drop to the failsafe shell Jul 01 17:07:44 ouch Jul 01 17:07:50 manually running "exec /sbin/init.sysvinit 5" works fine Jul 01 17:07:57 any ideas? Jul 01 17:08:04 let me see Jul 01 17:08:04 I was also seeing some other strange behavior Jul 01 17:08:15 2 directories showing up in /etc/altboot-menu Jul 01 17:08:21 "chown" and "root.roor" Jul 01 17:08:25 "root.root" Jul 01 17:08:31 in both my old and new installations.... Jul 01 17:09:15 JustinP: that's a bug in populate-volatile which has since been fixed Jul 01 17:09:24 normal boot is just doing: Jul 01 17:09:27 run_module() { Jul 01 17:09:27 exec $REAL_INIT "$INIT_RUNLEVEL" Jul 01 17:09:27 exit 0 Jul 01 17:09:27 } Jul 01 17:09:33 yeah, I know Jul 01 17:09:38 I wonder how that can error out Jul 01 17:09:41 which is why I was very confused Jul 01 17:09:45 what's the exact error? Jul 01 17:09:50 no error Jul 01 17:09:56 just ERROR: ....failsafe... Jul 01 17:09:57 just "ERROR"? Jul 01 17:09:59 hold on, I'll boot Jul 01 17:11:26 "ERROR: WARNING: Using the failsafe shell" Jul 01 17:11:39 "/bin/sh: can't access tty; job control turned off" Jul 01 17:11:46 (I assume the second line is normal...?) Jul 01 17:12:08 it worked a few times just fine....but started bailing lately Jul 01 17:12:18 it is Jul 01 17:12:28 will have a look Jul 01 17:12:36 ./00-Default run Jul 01 17:12:49 "exec: 14: : Permission denied" Jul 01 17:13:09 . $1/$file run "$file" >"$OUT_TTY" Jul 01 17:13:09 die "WARNING: Using failsafe shell" >"$OUT_TTY" Jul 01 17:13:19 that's a fall-through Jul 01 17:13:32 JustinP: that's the problem Jul 01 17:13:50 but the script is sourcing, not directly running the module Jul 01 17:15:05 eek Jul 01 17:15:24 running ". ./00-Default run "00-Default"" causes my Z to lock up Jul 01 17:15:24 03eFfeM 07org.oe.dev * ra5eefc5c... 10/packages/mutt/ (mutt-1.5.4i/keymap.h.patch mutt_1.5.4i.bb mutt_1.5.9i.bb): Jul 01 17:15:24 mutt: added #include in keymap.h to make it compile Jul 01 17:15:24 also added PR in both bb files Jul 01 17:15:29 03eFfeM 07org.oe.dev * r8fd42d6b... 10/packages/mdadm/ (mdadm.inc mdadm_1.12.0.bb): Jul 01 17:15:29 mdadm: added CFLAGS += "-fno-strict-aliasing" to avoid build errors Jul 01 17:15:29 a clean solution would be to fix the error in the source, but Jul 01 17:15:29 that is more intrusive Jul 01 17:15:33 03eFfeM 07org.oe.dev * rf83035de... 10/packages/ircp/ircp_0.3.bb: ircp: added EXTRA_OECONF="--prefix=${D}" to bb file to make it work Jul 01 17:15:37 03eFfeM 07org.oe.dev * r5308023f... 10/packages/radlib/ (5 files in 2 dirs): radlib: removed /usr/include and friends from include path Jul 01 17:15:41 03eFfeM 07org.oe.dev * ra2b34b40... 10/packages/wview/ (5 files in 2 dirs): wview: removed /usr/include from list of include dirs Jul 01 17:15:51 JustinP: it should only lock up the shell since it's "exec"'ing init Jul 01 17:15:51 could someone else send the monotone conversion reminder? Jul 01 17:16:04 cursor stopped blinking Jul 01 17:16:16 switching the battery switch didn't turn off Jul 01 17:16:21 it was a hard lock Jul 01 17:18:34 pff Jul 01 17:18:43 of course running it from shell doesn't work Jul 01 17:18:53 the vars aren't set Jul 01 17:19:20 are the vars in altboot-2.6.conf correct? Jul 01 17:19:29 checking that now Jul 01 17:19:40 don't know why they wouldn't be.... Jul 01 17:21:42 00-Default isn't even being run... Jul 01 17:23:04 hmmm...mine is different Jul 01 17:23:12 it explicitly has /dev/tty0 Jul 01 17:23:18 this is an oz3541 build Jul 01 17:24:00 ~lart ewi some more Jul 01 17:24:00 * ibot gets a hotmal account and SPAMs ewi some more Jul 01 17:24:01 just changed a few packages to add --prefix=${D} to EXTRA_OECONF, but shouldn't this be part of the default settings? Jul 01 17:24:34 umm....that's wrong, I think.... Jul 01 17:24:51 ir perhaps I have my vars mized up Jul 01 17:24:55 which one is ${D}? Jul 01 17:24:55 adding it or that it is not there ? Jul 01 17:24:58 CoreDump|home: aha! Jul 01 17:25:08 CoreDump|home: I uncommented the "echo [$file]" Jul 01 17:25:14 CoreDump|home: [00-Default Jul 01 17:25:25 Advanced/70-install-tgz] Jul 01 17:25:27 ehm, I copied from somewhere and that used ${D}, added that and things compiled Jul 01 17:25:33 er Jul 01 17:25:41 eFfeM: what does ${D} have in it? Jul 01 17:25:50 eFfeM: host or target? Jul 01 17:26:13 eFfeM: if it's host, I'm pretty sure that's going to introduce host paths on the target... Jul 01 17:26:47 host Jul 01 17:27:08 then I don't think that's right Jul 01 17:27:16 --prefix should be set to the target installation dir Jul 01 17:27:27 and other things should take care or the host path munging Jul 01 17:27:30 (at least IIRC) Jul 01 17:27:43 yup Jul 01 17:28:05 otherwise you let bitbake lose into your system and bad things happen Jul 01 17:28:21 I checked the configure log and you are right, it calls configure with --prefix=/usr Jul 01 17:28:24 emte: you sure you don't have that backwards? Jul 01 17:29:27 i was refering to eFfeM injecting host enviroment paths into bitbake Jul 01 17:29:36 what should I use instead? Jul 01 17:29:58 i suppose i should type more clearly Jul 01 17:30:12 $D does not exist in my world and I was trying to fix an issue where the gcc 4.1.1. complained about /usr/include being in the path Jul 01 17:30:16 for cross compilation Jul 01 17:30:17 pb_: I'm german, but please don't hurt me Jul 01 17:30:22 emte: I assumed that it was host as in staging, not host as in actual host.... Jul 01 17:30:29 eFfeM: yeah, I see that one too Jul 01 17:30:37 eFfeM: /usr/include shouldn't be in the include path Jul 01 17:30:41 JustinP, depends on the enviroment variable Jul 01 17:30:58 eFfeM: --prefix=/usr is 'right' Jul 01 17:31:05 eFfeM: this is meant for the to be installed files Jul 01 17:31:25 does it then work as I expect it to be ? Jul 01 17:31:36 eg saw this: unzip/unzip_552.bb: mv ${D}${prefix}/man/* ${D}${mandir} Jul 01 17:32:24 actually I peeked at groff when I made the patch and groff says: Jul 01 17:32:25 groff/groff_1.19.1.bb:EXTRA_OECONF="--prefix=${D} Jul 01 17:32:39 CoreDump|home: in my system, 1 is 00-Default and 11 is Advanced/70-install-tgz..... Jul 01 17:32:39 the prefix is also used for the include path Jul 01 17:33:02 eFfeM: that is bad then Jul 01 17:33:05 JustinP, ${D}${sysconfdir} seems used alot in the kernel bb files Jul 01 17:33:22 eFfeM: wait a second, I need to see if I talk garbage once again Jul 01 17:33:32 zecke: from configure --help: Jul 01 17:33:32 --includedir=DIR C header files [PREFIX/include] Jul 01 17:33:43 *that* caused me to change prefix Jul 01 17:33:56 eFfeM: for the to be installed headers :) Jul 01 17:34:34 ehm, wouldn't the installed headers and the used headers reside in the same place ... Jul 01 17:35:07 JustinP: do you have filenames w/ space in altboot-menu? Jul 01 17:35:07 eFfeM: hmm, not for cross compiling, not when creating toolchains Jul 01 17:35:54 zecke, you lost me Jul 01 17:36:28 anyway, I'll happily change things, if I only knew how Jul 01 17:36:55 eFfeM: I have lost myself Jul 01 17:37:03 lol Jul 01 17:37:06 eFfeM: if --prefix is used for includes then it's a bug in the package's configure script Jul 01 17:37:18 eFfeM: --prefix is supposed to be only for install paths Jul 01 17:37:20 glad to hear I am not the only one who has this Jul 01 17:37:21 eFfeM: --prefix, --includedir are all for 'where to install X' Jul 01 17:37:36 zecke: from configure --help: Jul 01 17:37:40 --includedir=DIR C header files [PREFIX/include] Jul 01 17:37:55 to me this suggests this is for referral also Jul 01 17:38:09 eFfeM: if that package is actually using prefix for where to include from it is, as I said before, wrong" Jul 01 17:38:17 eFfeM: it should be patched and the patch sent upstream Jul 01 17:38:25 (or my favorite hack....sed) Jul 01 17:38:30 CoreDump|home: no Jul 01 17:38:36 eFfeM: this is a false assumption :) Jul 01 17:38:41 JustinP: I'll happily patch, if I know what to patch it into Jul 01 17:38:44 eFfeM: and a bug :) Jul 01 17:38:58 eFfeM: change the Makefile.am Jul 01 17:38:58 JustinP, QT and a few other cross plat systems use that flag Jul 01 17:39:05 its not uncommin Jul 01 17:39:10 uncommon* Jul 01 17:39:18 --includedir is not what I'm saying the problem is.... Jul 01 17:39:28 eFfeM: I suggest you set --includedir instead of --prefix Jul 01 17:39:37 usually used in cases where you have multiple versions of something installed in different places Jul 01 17:40:10 JustinP: again: according to configure --help, --includedir == PREFIX/include ! Jul 01 17:40:17 CoreDump|home: it has to be the setting of $file...I'm not sure what that sed command does, but $file has both entries after that Jul 01 17:40:26 eFfeM: that's a default, yes. Override it Jul 01 17:40:34 to what ? Jul 01 17:40:45 eFfeM: if --includedir is used for includes when compiling and not for an install dir, that it Jul 01 17:40:54 eFfeM: to the staging include dir Jul 01 17:41:03 yup Jul 01 17:41:08 ${STAGING_DIR}/include ? Jul 01 17:41:20 there's a var for it I think... Jul 01 17:41:34 eFfeM: yes, see CFLAGS and CXXFLAGS for --help as well :) Jul 01 17:41:45 ${STAGING_INCDIR} Jul 01 17:41:47 so I happened to look at a bad example! Jul 01 17:41:54 yep :-) Jul 01 17:42:25 JustinP: yeah i remember such a bug in a different part. Give me a minute to prep an ipk Jul 01 17:42:42 what is preferred: --includedir or patching Makefile.am Jul 01 17:43:04 CoreDump|home: ok. Remember, I'm using .oz354x version...perhaps this is already fixed in .dev? I can build my own .dev version and test i fyou like Jul 01 17:43:23 yes, testing _0.0.0 from dev would be ideal Jul 01 17:43:31 CoreDump|home: 0.0.0? ok Jul 01 17:43:36 03eFfeM 07org.oe.dev * r98febddf... 10/packages/xinetd/ (xinetd/service.c.patch xinetd_2.3.13.bb): xinetd: slightly changed code to make it compilable with gcc 4.1.1 Jul 01 17:43:40 03eFfeM 07org.oe.dev * rbbe2ca73... 10/packages/musicpd/ (mpd-alsa_0.11.5.bb mpd_0.11.5.bb): mpd: added --prefix to remove /usr/include Jul 01 17:43:58 ok this one was still in the pipeline Jul 01 17:44:07 brb Jul 01 17:45:15 * JustinP plugs in his new wireless card Jul 01 17:45:24 ^_^ my Z is so much more useful with wireless Jul 01 17:45:44 :) Jul 01 17:48:08 JustinP: wireless rocks Jul 01 17:48:36 Im writing this on a notebook hooked w/ WLAN, watching DVB-T =D Jul 01 17:50:05 well, my laptop is on wireless as well Jul 01 17:50:08 argh Jul 01 17:50:12 wireless on my Z not working... Jul 01 17:50:21 and it worked so well the first time...WTF Jul 01 17:54:17 damn it! Jul 01 18:00:17 ~lart ipkg Jul 01 18:00:17 * ibot burns ipkg to a crisp with a laser Jul 01 18:01:05 * JustinP gets out a knife and starts trimming Jul 01 18:02:25 heh Jul 01 18:03:26 heh Jul 01 18:03:51 I discovered a distro that doesn't use correct dependencies in their packages Jul 01 18:03:56 *drumroll* Jul 01 18:03:56 it should not be reading the lists of packages it *could* install when I give it a file to install....not until it needs them.. Jul 01 18:04:01 Fedora Core Jul 01 18:04:07 really Jul 01 18:04:24 what do they use? Jul 01 18:04:35 yum install autoconf -> autoreconf -i -> segfault Jul 01 18:04:47 yum update pulls in a new glibc -> it works Jul 01 18:04:55 bloody rpm crap Jul 01 18:05:12 * JustinP hates rpm Jul 01 18:05:28 ~hail aptitude and debian thouroughness Jul 01 18:05:29 * ibot bows down to aptitude and debian thouroughness and chants, "I'M NOT WORTHY!!" Jul 01 18:05:37 indeed Jul 01 18:05:40 of course what I hate most is that we had RHEL systems...but no support, so I had to install FC rpms on RHEL...that was a nightmare Jul 01 18:05:50 thank God I got them to switch to FC instead at least Jul 01 18:06:26 ~hail gentoo and build-time dependencies Jul 01 18:06:28 * ibot bows down to gentoo and build-time dependencies and chants, "I'M NOT WORTHY!!" Jul 01 18:16:00 WTF Jul 01 18:16:14 it's even more screwed up...I can't get it to install a local ipk over the feed one! Jul 01 18:17:14 JustinP: yeah, ipkg is a pain at times Jul 01 18:17:25 s/at times/usually Jul 01 18:17:49 son of a.... Jul 01 18:17:54 it refuses to install a local copy Jul 01 18:18:34 JustinP: you'll have to delete the feed list Jul 01 18:18:38 I did Jul 01 18:18:50 ahem, that should have worked Jul 01 18:18:53 it says altboot isn't available from any confiured source Jul 01 18:19:32 "no such file or directory" Jul 01 18:19:33 WTF Jul 01 18:19:45 wrong ARCH of the ipk maybe? Jul 01 18:21:42 really doubt that Jul 01 18:22:13 why do things always have to be broken? Jul 01 18:25:13 because none really bother to the the god ugly ipkg Jul 01 18:30:06 * JustinP finally gets altboot-0.0.0 installed Jul 01 18:36:01 CoreDump|home: -0.0.0 from .dev works Jul 01 18:37:14 excellent Jul 01 18:41:24 yeah ipkg needs some love Jul 01 18:41:49 JustinP: your ipkg is too old, that bug was fixed a while ago Jul 01 18:42:40 it needs filtering abilities i think when parsing the feed list Jul 01 18:43:17 and maybe a centeral "feed aquisition" capability Jul 01 18:43:22 central* Jul 01 18:44:08 where it could pickup valid feed entries and automatically add it to it's list from a single general url Jul 01 18:44:16 ipkg needs to be rewritten using all the knowledge we gathered with OE Jul 01 18:44:18 emte, JustinP, changed --prefix=${D} into --includedir=${STAGING_INCDIR} in EXTRA_OECONF, but that does not really fly Jul 01 18:44:25 configure for ircp says: Jul 01 18:44:25 checking for openobex/obex.h... no Jul 01 18:44:26 configure: error: OpenOBEX header files not found Jul 01 18:44:48 eFfeM, try ${D}${STAGING_INCDIR} Jul 01 18:44:58 er wait that doesnt make sense Jul 01 18:45:00 ah, ok Jul 01 18:45:34 eFfeM: both ${D} and ${STAGING_INCDIR} look not appropriate Jul 01 18:45:36 you need ${TARGET}${STAGING_INCDIR} i think it is Jul 01 18:45:58 emte: TARGET + STAGING_INCDIR Jul 01 18:46:09 isnt that what i just wrote? Jul 01 18:46:17 emte: I missed the ??? Jul 01 18:46:37 I have missed most of the conversation, but this just looks plain wrong Jul 01 18:46:59 it is plain wrong Jul 01 18:47:01 i've only read parts of it, had to drop my wife off at work Jul 01 18:47:09 staging_incdir is an absolute path to the includes Jul 01 18:47:16 prepending it with stuff is madness Jul 01 18:47:19 03coredump 07org.oe.oz354x * r2d35331d... 10/packages/samba/samba-essential_3.0.20.bb: samba: Add missing directory /etc/samba/private Jul 01 18:47:19 zecke I got a few packages that complain with gcc about being cross and having /usr/include Jul 01 18:47:23 03coredump 07org.oe.dev * r4ef7d0c1... 10/packages/samba/samba-essential_3.0.20.bb: samba: Add missing directory /etc/samba/private Jul 01 18:47:53 I added --prefix={D} in EXTRA_OECONF and was told that this was wrong Jul 01 18:47:55 CIA-9 courtesy poushag Jul 01 18:47:59 eFfeM, that sounds like a broken enviroment Jul 01 18:48:11 more than OE/bitbake Jul 01 18:48:34 check your $PATH $BBPATH.. etc Jul 01 18:48:39 with --prefix={D} it at least works Jul 01 18:48:55 emte what would you want me to check / where should I look for ? Jul 01 18:49:10 that your not inheritiing your host system paths Jul 01 18:49:29 aside from /sbin and /usr/bin Jul 01 18:49:34 koen: my ipkg is from an rc4 image.... Jul 01 18:49:52 BBPATH=/home/frans/OpenSlugHead/openslug:/home/frans/OpenSlugHead/openslug/openembedded:/home/frans/OpenSlugHead/openslug/bitbake Jul 01 18:50:01 PATH=/home/frans/OpenSlugHead/openslug/bitbake/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/bin/X11:/usr/games Jul 01 18:50:12 JustinP: like I said, too old Jul 01 18:50:22 JustinP: OE is 8 or 9 versions behind Jul 01 18:50:26 yeah, eFfeM try cleaning out $PATH Jul 01 18:50:39 then perhaps OE should be updated Jul 01 18:50:52 down to jsut your bitbake/bin, sbin, and usr/bin Jul 01 18:50:54 sorry emte, but I still do not understand what the problem is and why cleaning the PATH would help here Jul 01 18:50:55 * JustinP goes to so some other work Jul 01 18:51:02 JustinP: feel free to test .163 and add it Jul 01 18:51:15 eFfeM, eliminates the posibilities that your using the wrong tools Jul 01 18:51:17 cleaning $PATH wouldn't help Jul 01 18:51:23 koen, my idea Jul 01 18:51:34 since OE isn't completely self-hosting Jul 01 18:51:49 there is alwas a chance you could have another cross gcc on your system that your calling by mistake Jul 01 18:52:39 if there is it would reside, most likly in /usr/local or /usr/share Jul 01 18:53:10 i've had the issue in the past, not recently mind you Jul 01 18:54:25 can someone first explain me why --prefix=${D} is no good ? Jul 01 18:54:40 because I still do not understand that Jul 01 18:55:04 valid headers are oar supposedly only in the satging directories Jul 01 18:55:13 and at least that works and the --includedir=${STAGING_INCDIR) doesn't! Jul 01 18:55:32 * emte changes keysets to help spelling Jul 01 18:56:15 D = "${WORKDIR}/image" Jul 01 18:56:35 I doubt that header would be hiding there during configure Jul 01 18:57:04 unless he has mangled it to point elsewhere Jul 01 18:57:33 D stands for Destination/Deploy does it not? Jul 01 18:57:52 no Jul 01 18:58:02 it stands for what I pasted Jul 01 18:58:06 see conf/bitbake.conf Jul 01 18:58:08 but without --prefix it tries to grab from /usr/include (at least configure complains that it cannot find openobex/obex.h) Jul 01 18:58:47 eFfeM: sounds like the configure script needs to get fixed Jul 01 18:59:19 * koen turns his attention back to dalgliesh Jul 01 18:59:40 somoen probably hardcoded the default Jul 01 18:59:53 which means it needs to be patched Jul 01 19:00:39 eFfeM, take a quick look in ytour staging dirs Jul 01 19:00:47 does obex.h exist there? Jul 01 19:10:08 sorry was called away Jul 01 19:10:15 emte, yes it exists there Jul 01 19:11:01 /home/frans/OpenSlugHead/openslug/tmp/staging/armeb-linux/include/openobex/obex.h Jul 01 19:13:05 eFfeM: see the openembedded.org front page Jul 01 19:13:33 eFfeM: Let us start easy: You recognize why including from /usr/include is bad? Jul 01 19:14:17 sure, don't want to mix target headers and host headers Jul 01 19:14:59 eFfeM: good. --prefix is for the application to know where to install it to Jul 01 19:15:23 eFfeM: e.g. applications like GPE * use --prefix (specially share dir) to compile where to look for resources like images Jul 01 19:15:24 yeah, understood that from koen Jul 01 19:15:31 eFfeM: ah :) Jul 01 19:15:43 okay then I'm out of the loop :) Jul 01 19:15:56 OT: Could someone please send the monotone reminder? Jul 01 19:15:58 hey, I still don't know how to fix this Jul 01 19:16:16 btw also do not understand "see the openembedded.org front page" Jul 01 19:16:37 what did you want to say with that or what item did you want to bring to my attention Jul 01 19:16:58 NOTE: package fluxbox-1.0rc: completed Jul 01 19:17:03 eFfeM: we are talking about OpenObex here? Jul 01 19:17:06 lol? Could it be really that easy Jul 01 19:17:23 zecke, yes Jul 01 19:17:42 eFfeM: okay Jul 01 19:18:02 I am trying to get ircp to configure, this one depends on openobex Jul 01 19:18:04 eFfeM: http://www.openembedded.org/repo/org.openembedded.dev/packages/openobex/openobex_1.2.bb this one? Jul 01 19:18:42 no: http://www.openembedded.org/repo/org.openembedded.dev/packages/ircp/ircp_0.3.bb Jul 01 19:18:45 eFfeM: okay then fix their Makefile.am :) (Can't compile ATM, I'm occupied with windows) Jul 01 19:18:48 ah okay Jul 01 19:19:02 i added EXTRA_OECONF="--prefix=${D}" Jul 01 19:19:08 eFfeM: okay that is definately wrong Jul 01 19:19:39 can you build it with gcc 4.1.1 without that line (or what should it contain?) Jul 01 19:20:04 my local copy now has EXTRA_OECONF="--includedir=${STAGING_INCDIR}" based on the suggestions made before, but does not configure (can't find obex.h) Jul 01 19:20:08 eFfeM: this means it gets installed into /home/effem/work/image Jul 01 19:20:15 and the --prefix one would. sigh Jul 01 19:20:16 eFfeM: remove that line Jul 01 19:20:50 eFfeM: I'm trying to think what the guy wanted to get done :} Jul 01 19:21:10 eFfeM: either ircp is awfully broken, or 'DESTDIR' is broken within ircp Jul 01 19:21:17 (sorry you are not progressing) Jul 01 19:21:31 1.) this EXTRA_OECONF line can only be wrong Jul 01 19:21:39 zecke removed the line, did a bb -cclean ircp; bb ircp and configure fails: Jul 01 19:21:49 checking openobex/obex.h usability... no Jul 01 19:21:49 checking openobex/obex.h presence... no Jul 01 19:21:49 checking for openobex/obex.h... no Jul 01 19:21:49 configure: error: OpenOBEX header files not found Jul 01 19:22:15 check where it looked for openobex Jul 01 19:23:13 hmm. this does not look good Jul 01 19:23:17 some snippets from configure Jul 01 19:23:28 if (test "${prefix}" = "NONE"); then Jul 01 19:23:29 openobex_prefix=${ac_default_prefix} Jul 01 19:23:29 else Jul 01 19:23:29 openobex_prefix=${prefix} Jul 01 19:23:29 fi Jul 01 19:23:46 test -d "${openobex_prefix}/include" && OPENOBEX_CFLAGS="$OPENOBEX_CFLAGS -I${openobex_prefix}/include" Jul 01 19:24:08 *lutscher* (not you effeM) Jul 01 19:24:29 zecke, nevermind, i don't even know what lutscher means Jul 01 19:24:57 it has many meanings Jul 01 19:25:44 eFfeM: option one is, force ircp to trust you openobex is there Jul 01 19:27:41 dang. tried this: EXTRA_OECONF="--prefix=${STAGING_DIR} ; this also works Jul 01 19:28:24 zecke: actually I would force it but I see similar problems with mpd (although there patching makefile.am might fix it because there it was not configure causing the problem) Jul 01 19:29:03 why not rebind in do_configure ? Jul 01 19:29:18 as a temp workaround Jul 01 19:29:45 $OPENOBEX_PREFIX = ${STAGING_INCDIR} Jul 01 19:30:20 should be STAGING_DIR as the test adds /include Jul 01 19:30:41 yeah, but you get what i mean Jul 01 19:30:48 what is wrong with EXTRA_OECONF="--prefix=${STAGING_DIR} Jul 01 19:31:13 nothing should be Jul 01 19:31:28 it normally works fine unless the configure/makefile is hardcoded Jul 01 19:32:42 maybe you should use -i mode to find out what its actually doing or enable debug mode for bitbake Jul 01 19:32:47 I'm quite surprised that without the EXTRA_OECONF it does not pick the file from /home/frans/OpenSlugHead/openslug/tmp/staging/armeb-linux/include/openobex/obex.h Jul 01 19:33:23 apart from within the openobex work dir this is the only obex.h I have on my system (just did a find on / as root) Jul 01 19:33:56 that is because its using a non-standard script Jul 01 19:34:19 its implicetly using your host system's includde dir without using --prefix Jul 01 19:34:30 include* Jul 01 19:35:05 somehow its finding soemthing somewhere Jul 01 19:35:09 is the EXTRA_OECONF="--prefix=${STAGING_DIR}" ok ? Jul 01 19:35:13 eFfeM: NO Jul 01 19:35:18 it is still not OKAY Jul 01 19:35:26 sigh Jul 01 19:35:30 lost Jul 01 19:35:34 in reality the configure/makefiles need to be patched Jul 01 19:36:07 eFfeM: Would you want GPE to search for your icons in /home/effe/foo/tmp/staging/arm-linux or /usr/share/? Jul 01 19:36:22 eFfeM: if you want the first, use the prefix like you want Jul 01 19:37:04 eFfeM: instead of searching a workaround, the author of the autoconf/Makefile needs to be tortured Jul 01 19:37:28 lol Jul 01 19:37:30 eFfeM: --prefix is for installing, not for searching headers Jul 01 19:37:40 ok Jul 01 19:37:54 eFfeM: People abusing it would kill kittens as well Jul 01 19:38:01 and I don't like these kind of people Jul 01 19:38:33 zecke, emte have a look at http://www.rafb.net/paste/results/1Ommpz22.html Jul 01 19:38:38 this is the complete configure.in Jul 01 19:38:59 the section you pasted here already tells me whats happening Jul 01 19:39:14 eFfeM: http://www.gnu.org/software/autoconf/manual/autoconf.pdf 4.1.2 Jul 01 19:39:19 4.12 even Jul 01 19:39:27 because -prefix is not defined it is useing your host's settings Jul 01 19:40:10 it is avoiding the OE/bitbake settings because its explicitly using OPENOBEX_PREFIX set to your host Jul 01 19:40:28 eFfeM: is there a m4 directory? Jul 01 19:40:47 leaving to emte to continue to speak Jul 01 19:41:06 thats a lot of tos Jul 01 19:41:07 zecke, yes there is an m4 dir but it is empy Jul 01 19:41:20 there is also an empty acinclude.m4 Jul 01 19:42:06 eFfeM: so openobex itself has a broken m4 file Jul 01 19:42:39 dunno, actually i have no idea how this m4 stuff really works Jul 01 19:42:47 or tehy dont ship it for starnge reasons Jul 01 19:42:52 strange* Jul 01 19:43:25 openobex also has an empty m4 dir Jul 01 19:43:30 eFfeM: me neither :) (I would not be able to write makefiles or m4/autoconf) Jul 01 19:43:49 but there is an openobex.m4 Jul 01 19:43:50 eFfeM: greo your staging directory for AM_PATH_OPENOBEX Jul 01 19:44:49 ./armeb-linux/share/aclocal/openobex.m4:AC_DEFUN([AM_PATH_OPENOBEX], [ Jul 01 19:45:34 this has the lines I pasted before Jul 01 19:46:17 there is also this: AC_CHECK_HEADER(openobex/obex.h,, AC_MSG_ERROR(OpenOBEX header files not found)) Jul 01 19:47:07 http://www.rafb.net/paste/results/2HseMq26.html is what you pasted earlier and causing the issue Jul 01 19:47:43 yes, it comes from this m4 file and I now understand what is wrong, but still not how to fix it Jul 01 19:48:14 there are many approaches Jul 01 19:48:24 some better than others Jul 01 19:49:30 lol, did you take philosophy 101 recently ? Jul 01 19:49:33 if you dont want to patch you could try setting ac_default_prefix to the correct variable Jul 01 19:49:45 no it's worse Jul 01 19:49:51 i took sociology Jul 01 19:49:58 wauw Jul 01 19:50:04 i'm an ee Jul 01 19:50:15 yeah thats what i am doing now Jul 01 19:50:54 sociology mangles your brain, i'd reccomend avoiding it Jul 01 19:51:01 i'm inclined to go for your do_configure patch with $OPENOBEX_PREFIX = ${STAGING_DIR} Jul 01 19:51:30 thats not correct in retrosepct Jul 01 19:51:37 oh sh*t Jul 01 19:51:39 bind ac_default_prefix Jul 01 19:51:44 that was the one I understood most Jul 01 19:51:56 in do_configure Jul 01 19:51:57 instead of openobex_prefix Jul 01 19:51:58 ? Jul 01 19:52:02 yeah Jul 01 19:52:04 upper case I assume Jul 01 19:52:21 its lowercase in what you pasted Jul 01 19:52:56 if you bind openobex_prefix in do_configure it will probably be recast Jul 01 19:53:39 if you use ac_default_prefix it should hold througout the compile providing something else does not recast it Jul 01 19:53:59 this is the main reason why patching is a better solution Jul 01 19:54:23 this ok: Jul 01 19:54:24 do_configure() { Jul 01 19:54:24 ac_default_refix = ${STAGING_DIR} Jul 01 19:54:24 oe_runconf Jul 01 19:54:24 } Jul 01 19:54:26 ? Jul 01 19:54:32 prefix that is Jul 01 19:54:49 or should there be a $ in front ? Jul 01 19:55:14 not sure, its python Jul 01 19:55:18 and i forget Jul 01 19:55:58 without $ it doesn't work Jul 01 19:56:15 with neither Jul 01 19:56:22 i guess thats the answer Jul 01 19:56:44 look at some otehr packages Jul 01 19:58:06 ok. i don't know, giving up for now Jul 01 19:58:43 THATS NO FUN Jul 01 19:58:47 bad caps Jul 01 19:59:19 i have no clue on how to proceed and I am starving for a beer Jul 01 19:59:32 mmmm beer Jul 01 19:59:40 actually I already spent way more time on this than I wanted Jul 01 20:00:02 myself i find patching faster and easier ... Jul 01 20:00:24 only because i have to drag out my python book every time i play with the functions Jul 01 20:03:31 cu later Jul 01 20:07:13 yummy yummy Jul 01 20:07:28 oh no Jul 01 20:07:33 :) Jul 01 20:07:43 :P Jul 01 20:09:29 10 revs, someone been busy pushing Jul 01 20:09:58 Could any arm guy try and build scummvm Jul 01 20:10:09 I had no luck, but superH isnt exactly standard Jul 01 20:35:29 ~lart brazilian footbal team Jul 01 20:35:29 * ibot teaches brazilian footbal team the basics, including how to RTM Jul 01 20:35:36 HAHA Jul 01 20:35:46 hehe Jul 01 20:36:25 ~hail ibot Jul 01 20:36:27 * ibot bows down to ibot and chants, "I'M NOT WORTHY!!" Jul 01 20:36:51 very humble ibot Jul 01 20:43:24 yeah Jul 01 20:43:31 thejapa, you getting far yet? Jul 01 20:43:48 any deve stuff to try? Jul 01 20:43:55 not as far as I wanted. Jul 01 20:44:01 nothing good to test yet Jul 01 20:44:36 I've just begun writing the opie pim text backend, using existing xml/vcard Jul 01 20:44:52 it's more of an exercise than some real thing Jul 01 20:45:12 ah Jul 01 20:47:05 i've either overstimated my skills or understimated system complexity. I thought it would be a lot easier, it turns out I need to be a lot more careful than... well. windows dev. Jul 01 20:47:06 duh Jul 01 20:48:50 this is my first try into "open source development", so far, I think I just needed more time to devote, unfortunately, I've been very busy with university and my work. Jul 01 20:49:48 now that I'm into it, though, I'm not willing to turn back. :) Jul 01 20:54:24 the console output when compiling a 2.6 kernel always looks odd to me Jul 01 20:55:03 Kristoffer: why did you remove the tremor.patch from scummvm? Jul 01 20:55:19 koen, only from the last release Jul 01 20:55:30 the code has changed alot, so it wasnt needed/didnt work Jul 01 20:55:46 Kristoffer: http://www.openembedded.org/viewmtn/getdiff.py?id1=889a2048fa859be73c244aeb31e982b29d624283&id2=95fa8952511a57ba0ef9058fbbb57e4af4d93cf6&fname=packages/scummvm/scummvm.inc Jul 01 20:55:54 you removed it in the .inc Jul 01 20:56:08 didnt I add it specificly to 0.6.1b then? Jul 01 20:56:20 not afaics Jul 01 20:56:31 ah, missed that then..will add it for 0.6.1b Jul 01 20:56:44 0.6.0 didnt use the .inc file I believe Jul 01 20:56:48 thanks Jul 01 20:57:10 thanks for pointing it out, 0.9.0 doesnt build for me, but probobly for any arm Jul 01 20:57:12 make clean Jul 01 21:02:38 03kristoffer 07org.oe.dev * r7adb1db7... 10/packages/scummvm/scummvm_0.6.1b.bb: Jul 01 21:02:38 scummvm/scummvm_0.6.1b.bb: Addition of tremor patch Jul 01 21:02:38 * Addition of tremor patch to 0.6.1b as this was removed from the Jul 01 21:02:38 include file (no longer needed for new versions) Jul 01 21:11:23 evening Jul 01 21:40:11 reenoo: hey, I'm glad you survived Jul 01 21:40:27 yo zecke Jul 01 21:40:28 heh Jul 01 21:40:36 survived robocup you mean? Jul 01 21:49:30 Kristoffer: scummvm 0.9.0 builds fine here for arm Jul 01 21:49:41 aquadran, good to know thx Jul 01 21:50:00 should be abit better than 0.6.1b :D Jul 01 21:50:38 reenoo: yes Jul 01 21:50:42 more than a bit :) Jul 01 21:52:42 zecke: yah. I've actually spent most time in the lab or manning our booth at the AI conference. only managed to see the finals on sunday as far as robocup itself is concerned. Jul 01 22:07:04 * chouimat is away: f00d Jul 01 22:10:07 OT: Any win32 guy around? What is the tool that is like 'ldd'? Jul 01 22:10:34 zecke: btw... if you were there I failed to find you, sorry :/ Jul 01 22:11:08 didn't make it :( Jul 01 22:11:35 ah, ok Jul 01 22:12:23 too bad. it was a great event. wished I had been able to spend more time there myself... Jul 01 22:12:23 03koen 07org.oe.dev * r44f2e80d... 10/packages/scummvm/scummvm_0.9.0.bb: scummvm 0.9.0: depend on libvorbis since the tremor patch hasn't been forward ported yet Jul 01 22:55:56 * Ken|JLime runs around screaming "The Beeellls, The Beells..." Jul 01 23:07:45 nite **** ENDING LOGGING AT Sun Jul 02 02:59:57 2006