**** BEGIN LOGGING AT Thu Jul 14 23:59:57 2005 Jul 15 04:03:45 Have a problem applying patch config-guess-uclibc.patch when building openslug - fails at hunk #5 (884). Is this a known problem or have I messed up something? Jul 15 04:22:29 bob_tm: when compiling what? Jul 15 04:24:47 and, are you building from the monotone repo, or the source tarball? Jul 15 04:45:52 NAiL: Mom - I'll get back to you soon Jul 15 04:53:58 <[g2]> DaKa2, hey Jul 15 04:54:26 hey hey Jul 15 04:54:50 <[g2]> how's that hot slug ? Jul 15 04:54:59 it works nicely naked Jul 15 04:55:12 <[g2]> *cool* Jul 15 04:55:28 Ill probably just drill up the holes a bit Jul 15 04:55:39 <[g2]> nod Jul 15 04:56:07 Im at work right now, playing with my slug3 Jul 15 04:56:21 <[g2]> neat Jul 15 04:56:42 <[g2]> any more luck with the native bb ? Jul 15 04:56:55 NAiL: gnu-config-0.1cvs20050715-r3 was the part that was to be compiled. Jul 15 04:57:15 didn't play any more with it, had to reinstall everything and by the time that was done I went to sleep Jul 15 04:57:23 bob_tm: aha. And you're building from the source tarball? Jul 15 04:57:34 ie. not monotone Jul 15 04:57:35 monotone make Jul 15 04:57:46 strange Jul 15 04:57:55 How so? Jul 15 04:58:13 (retrying now) Jul 15 04:58:23 Same same Jul 15 04:58:42 well.. they've probably done something that breaks the patch in CVS (Maybe even applied it ;) Jul 15 04:59:09 Is the patch one of "ours"? Jul 15 04:59:37 If so, I can look into it. Jul 15 05:00:16 Hang on a sec, gonna see if it happens to me too.. Jul 15 05:00:27 NOTE: Applying patch 'config-guess-uclibc.patch' Jul 15 05:00:27 NOTE: package gnu-config-0.1cvs20050713-r3: task do_patch: completed Jul 15 05:00:37 My cvs checkout is two days older though Jul 15 05:00:58 Yes - so you need the latest version for it to break Jul 15 05:01:53 I will take a look at the source and the patch Jul 15 05:05:51 <[g2]> DaKa2, have you gotten to this ? Jul 15 05:07:03 bob_tm: there was talk in #oe about gnu-config breaking. perhaps just set the CVSDATE back a couple of days ... Jul 15 05:07:39 rwhitby-away: Yes - I have the 0 Jul 15 05:07:48 <[g2]> checking build system type... Invalid configuration `armv5teb-linux': machine `armv5teb' not recognized Jul 15 05:07:48 <[g2]> configure: error: /bin/bash /home/tom/tmp/work/libtool-native-1.5.10-r1/libtool-1.5.10/config.sub armv5teb-linux failed Jul 15 05:07:48 <[g2]> FATAL: oe_runconf failed Jul 15 05:07:48 <[g2]> tom@apex:~$ armv5teb-linux- Jul 15 05:07:48 <[g2]> bash: armv5teb-linux-: command not found Jul 15 05:07:51 07/07 version - will try to use that Jul 15 05:08:04 bob_tm: 07/13 also worsk Jul 15 05:08:09 works :) Jul 15 05:08:13 :-) Jul 15 05:08:27 <[g2]> we should "nail" that config Jul 15 05:09:17 Yeah Jul 15 05:09:41 we had to catch up with recent changes in their cvs Jul 15 05:09:53 (gnu-config, that is) Jul 15 05:10:23 I thought we already had all cvs things nailed to either a tag or a date in nslu2.conf ... Jul 15 05:11:04 Hmm - we nailed gnu-config-native to 20050616 Jul 15 05:11:15 I guess we need to nail gnu-config too. Jul 15 05:11:39 <[g2]> might as well make it 07/13 Jul 15 05:11:46 rwhitby: I'm planning to put the monotone server from the net tomorrow Jul 15 05:11:51 I always get 0713 when compiling gnu-config now Jul 15 05:11:58 As far as I am concerned, we cannot have any packages in nslu2-linux which float from day to day in cvs. They have to be either tied to a tag or a date. Jul 15 05:12:01 rwhitby: in order to upgrade to 0.20 Jul 15 05:12:10 mickey|lunch: that's good to hear :-) Jul 15 05:12:11 rwhitby: it may well be monday or so until it's online again. so, just fyi. Jul 15 05:12:23 [g2]: yes, I got that on xinetd yesterday Jul 15 05:12:23 I'll write a mail to oe@hh.org now Jul 15 05:12:29 mickey|lunch: good luck :) Jul 15 05:12:46 mickey|lunch: no problem - we can still work on our server, and we'll upgrade that after you have finished yours Jul 15 05:13:13 <[g2]> DaKa2, did you get around it ? :) Jul 15 05:13:16 rwhitby: how is a specific package date nailed without modifying the bb? Jul 15 05:13:56 NAiL: we nail it in conf/machine/nslu2.conf for both unslung and openslug Jul 15 05:13:57 [g2]: I was working on it when my slug overheated.. Jul 15 05:14:40 NAiL: thanks. i hope it won't take much time :/ Jul 15 05:14:44 * mickey|lunch wanders home Jul 15 05:14:49 [g2]: I was thinking either fix the configure script, or make oe send it arm-unknown or whatever it wants Jul 15 05:14:58 rwhitby: Ok, I'll lock it to 0713 Jul 15 05:15:34 my other turboslug slug3, just freazed.. Jul 15 05:15:41 I was pushing it really hard Jul 15 05:15:54 NAiL: already done, pushing now Jul 15 05:16:00 rwhitby: ok Jul 15 05:16:12 <[g2]> DaKa2, you run everything off the net right ? Jul 15 05:16:20 yes Jul 15 05:16:42 Ill connect serial and try to replicate :-) Jul 15 05:16:51 <[g2]> I think you can drive your slugs harder that us disk based guys Jul 15 05:17:11 maybe still alive, just the phy or something dies Jul 15 05:17:18 Yeah, you get more b/w from the nic than usb Jul 15 05:27:48 <[g2]> DaKa2, did you try installing psyco ? Jul 15 05:27:59 no.. Jul 15 05:28:06 have you? Jul 15 05:28:18 <[g2]> no, but was *thinking* about it :) Jul 15 05:28:24 same here :-) Jul 15 05:28:59 <[g2]> I changed a line in openembedded/conf/nslu2.conf Jul 15 05:29:20 <[g2]> #TARGET_CC_ARCH = "-march=armv5te -mtune=xscale" Jul 15 05:29:20 <[g2]> TARGET_CC_ARCH = "-mtune=xscale" Jul 15 05:29:36 <[g2]> TARGET_ARCH = "armeb" Jul 15 05:29:51 <[g2]> the TARGET_ARCH is before those other lines Jul 15 05:29:51 [g2]: a tip on creating swapfile, dd if=/dev/zero of=swapfile bs=1M seek=128 count=0 Jul 15 05:29:59 <[g2]> heh Jul 15 05:30:00 hm, it works? Jul 15 05:30:06 <[g2]> testing now Jul 15 05:30:14 <[g2]> well know in 5-10 Jul 15 05:30:17 <[g2]> minutes Jul 15 05:30:40 [g2]: check with jbowler-zzz - there was a reason he put in armv5te, but I don't recall what it was Jul 15 05:31:00 <[g2]> rwhitby, this is just on the native build Jul 15 05:32:23 <[g2]> I'd imagine we probably want to install our native compilers as the armv5teb Jul 15 05:33:10 <[g2]> then we wouldn't have go change it, but right now there's a mismatch and I don't even know if this will fix it Jul 15 05:33:24 <[g2]> but it's a good point to mention anyway Jul 15 05:33:50 yeah, just mentioning it so it didn't get pushed Jul 15 05:34:12 <[g2]> this is just testing on the native builds Jul 15 05:34:32 <[g2]> there will be oe changes that will need to be discussed Jul 15 05:35:06 <[g2]> I'm not even sure I'm building the right package Jul 15 05:35:14 <[g2]> I'm building perl-native Jul 15 05:35:27 <[g2]> which is actually perl armeb Jul 15 05:35:51 <[g2]> I don't know if anyone other than the x86 crowd does native Jul 15 05:36:17 <[g2]> unless is on that box like ppc Jul 15 05:40:28 dammit.. I disslike not beeing able to crach my slug when I want Jul 15 05:40:57 <[g2]> I'd hate that too! Jul 15 05:41:23 I just managed... Jul 15 05:41:47 dammit.. kernel problem.. with nfs probably Jul 15 05:41:49 so, no heat Jul 15 05:42:46 DAMMIT.. my serial terminal doesn't buffer everything.. Jul 15 05:44:44 oh well, just have to crash it again Jul 15 05:54:51 dammit... last time I got 5 oops, this time just an irq26: nobody cared Jul 15 05:57:04 finally Jul 15 05:57:10 * NAiL replaced his router Jul 15 05:57:22 Now I can reflash my slug without the entire network going down :) Jul 15 05:57:32 Now all I need is a root disk of somesort Jul 15 05:58:07 anyone good at this kind of stuff: http://david.thg.se/saker/dammit-050715 ? Jul 15 05:59:02 address b824 should be a 01 instead of a 00 .... Jul 15 05:59:32 * NAiL assumes rwhitby is joking Jul 15 06:00:22 * DaKa2 too Jul 15 06:00:49 <[g2]> DaKa2, it crashed on that ? Jul 15 06:01:14 well, the first part is partial, the seccond is complete, but something completly different Jul 15 06:01:21 but still complete crash Jul 15 06:01:32 serial and network not working to access it Jul 15 06:01:53 that was with heavy network and memory access Jul 15 06:03:07 <[g2]> that's with 2.6.12.2 right Jul 15 06:03:11 yup Jul 15 06:03:32 <[g2]> you don't have an ethernet-to-usb adapter do you ? Jul 15 06:03:41 at home... Jul 15 06:03:44 not here Jul 15 06:04:23 <[g2]> you know about lak right ? Jul 15 06:04:29 ? Jul 15 06:04:29 <[g2]> linux-arm-kernel ml Jul 15 06:04:39 oh.. Jul 15 06:04:40 no Jul 15 06:04:55 <[g2]> yeah bunch of kernel hacks Jul 15 06:05:18 <[g2]> normally we could post the ooops there Jul 15 06:05:33 hm.. Im trying to replicate this with lots of cooling on the chips.. Jul 15 06:05:37 <[g2]> however, the kernel is tainted with the Intel driver which is closed source Jul 15 06:06:06 <[g2]> this may be an NFS or interrupt thing Jul 15 06:06:32 <[g2]> and if you don't load the ixp modules and run with the usb-ethernet then the kernel isn't "tainted" Jul 15 06:06:42 ahh Jul 15 06:06:45 <[g2]> I did that with the irq26 thing Jul 15 06:06:55 <[g2]> failed with/without ixp Jul 15 06:07:07 well, I cannot get it to crash with cooling... Jul 15 06:07:16 Ill try a bit more Jul 15 06:07:27 <[g2]> got a fan going ? Jul 15 06:07:30 yup Jul 15 06:13:02 I really cannot get it to die with a fan Jul 15 06:15:07 night all Jul 15 06:15:14 nite Jul 15 08:19:05 [g2]: where on earth does armv5teb come from? I thought I had suggested that on the OE mailing list but not actually implemented it. Jul 15 08:20:16 armv5teb is actually the correct PACKAGE_ARCH and TARGET_ARCH, but to change it is a big change (all the feeds will be invalid) Jul 15 08:20:17 TARGET_CC_ARCH = "-march=armv5te -mtune=xscale" in openembedded/conf/machine/nslu2.conf Jul 15 08:20:31 (no b) Jul 15 08:20:32 DaKa2 armv5te B Jul 15 08:20:48 And TARGET_CC_ARCH is just the compiler options. Jul 15 08:20:50 any performance increased possible? :) Jul 15 08:21:14 hm.. oh.. Jul 15 08:22:05 With the compiler -march? Jul 15 08:24:23 VoodooZ_work: the more specific the information the compiler gets the better it can do, the default is something like -march=armv4 Jul 15 08:24:39 But it depends on how the compiler is patched... Jul 15 08:25:03 mickeyl: ping Jul 15 08:25:12 I see. Me like Jul 15 08:25:39 mine already has TARGET_CC_ARCH = "-march=armv5te -mtune=xscale" Jul 15 08:25:39 PACKAGE_ARCH = "armeb" Jul 15 08:25:59 That's what's in monotone Jul 15 08:26:20 <[g2]> tom@apex:~$ gcc -dumpmachine Jul 15 08:26:20 <[g2]> armeb-linux Jul 15 08:26:24 'armeb' is a weird non-specific description Jul 15 08:26:37 ha Jul 15 08:26:57 [g2]: where does 'armv5teb' come from? Jul 15 08:26:59 good to see that you're on top of things. Watching like an eagle! :) Jul 15 08:27:00 <[g2]> that's the way the compiler is getting built by OE Jul 15 08:27:18 <[g2]> jbowler, that's a great question Jul 15 08:27:45 ?, yes.... and the answer is? Jul 15 08:28:04 <[g2]> I see a bunch of stuff like Jul 15 08:28:21 <[g2]> + oenote Executing autoreconf --verbose --install --force --exclude=autopoint -I /home/tom/tmp/work/db3-native-3.2.9-r1/db-3.2.9/dist/aclocal/ -I /home/tom/tmp/staging/armv5teb-linux/share/aclocal-1.9 -I /home/tom/tmp/staging/armv5teb-linux/share/aclocal Jul 15 08:28:21 <[g2]> + echo NOTE: Executing autoreconf --verbose --install --force --exclude=autopoint -I /home/tom/tmp/work/db3-native-3.2.9-r1/db-3.2.9/dist/aclocal/ -I /home/tom/tmp/staging/armv5teb-linux/share/aclocal-1.9 -I /home/tom/tmp/staging/armv5teb-linux/share/aclocal Jul 15 08:28:57 <[g2]> at one point configure was dying due to a difference between armeb-linux and armv5teb-linux Jul 15 08:29:29 Staging!! ?? Jul 15 08:29:38 did you manage to fix that [g2]? Jul 15 08:29:52 [g2]: fix your TARGET_OS Jul 15 08:30:04 <[g2]> DaKa2, no Jul 15 08:30:12 :( Jul 15 08:30:25 <[g2]> DaKa2, you're getting the same thing too right ? Jul 15 08:30:41 Ah ha! It's those squirrelly Linux guys. Jul 15 08:30:52 Try 'uname -a' Jul 15 08:31:15 <[g2]> uname -a Jul 15 08:31:15 <[g2]> Linux apex 2.6.11.2 #1 Thu Jul 7 21:48:02 EDT 2005 armv5teb unknown unknown GNU/Linux Jul 15 08:31:51 Well, great minds think alike (remainder of quote deleted) - if they came up with armv5teb and I did too it must be right ;-) Jul 15 08:32:14 DaKa2: just override TARGET_OS Jul 15 08:32:50 Like I said, if someone wants to spend a couple of weeks fixing up the feeds it would be better as armv5teb, but... Jul 15 08:34:31 TARGET_OS = armv5teb ? Jul 15 08:35:05 Um, oh, not TARGET_OS Jul 15 08:35:19 Whatever it is that sets that element of the 'staging' directory. Jul 15 08:35:20 target_os = linux.. Jul 15 08:35:47 It's in bitbake.conf (and it's whatever is currently using 'uname') Jul 15 08:36:00 ah Jul 15 08:36:27 BUILD_ARCH? Jul 15 08:37:07 I think so Jul 15 08:37:33 You can zap it in local.conf Jul 15 08:37:36 BUILD_ARCH = armeb? Jul 15 08:38:19 I think that will work Jul 15 08:38:28 hm, Ill try Jul 15 08:38:30 What I don't understand though is why it causes the build to fail. Jul 15 08:38:48 After all, it's just a path - but that's what [g2] reported as being the problem... Jul 15 08:39:11 the configure doesn't know what armv5teb is, its not the path that is the problem Jul 15 08:39:18 actually the configure should be fixed Jul 15 08:39:28 but there are lots of them Jul 15 08:39:37 Oh, well, that's different to the thing [g2] identified Jul 15 08:39:46 <[g2]> DaKa2, right I think configure is passed the wrong mojo Jul 15 08:40:03 yup, oe gives it armv5teb-linux Jul 15 08:40:10 and it doesn't know what that is Jul 15 08:40:11 <[g2]> and I think that get pass all the way through Jul 15 08:40:28 ? if you want to fix it, find out where it is coming from... Jul 15 08:40:30 yup, so.. juts make oe send something all configure knows Jul 15 08:40:42 its probably BUILD_ARCH Jul 15 08:40:51 I don't think so... Jul 15 08:41:09 <[g2]> BUILD_ARCH = "armeb" doesn't seem to change it Jul 15 08:41:13 It might be, but I suspect configure runs uname itself Jul 15 08:41:16 hm.. Jul 15 08:41:16 <[g2]> in local.conf Jul 15 08:41:39 well, configure on its own makes it, its OE passing on armv5teb-linux Jul 15 08:41:48 You will end up doing something which can more simply be done with a single command: Jul 15 08:42:11 find / -type f -exec sed -i 's/armv5te/armeb' {} \; Jul 15 08:42:43 (And that is not a suggestion) Jul 15 08:42:53 <[g2]> runnign Jul 15 08:42:58 <[g2]> DOH! Jul 15 08:43:06 <[g2]> jk Jul 15 08:43:07 upslug Jul 15 08:43:07 <[g2]> :) Jul 15 08:43:21 <[g2]> gotcha Jul 15 08:43:24 :) Jul 15 08:43:53 <[g2]> Ok.... I'd just replace uname Jul 15 08:45:59 It's important to find out where it comes from. The architecture information is essential to configure, if we get it wrong then we will have no end of problems porting packages. Jul 15 08:46:14 well, the problem seems to be that it gives the configure: Jul 15 08:46:14 --build=armv5teb-linux \ Jul 15 08:46:14 --host=armv5teb-linux \ Jul 15 08:46:14 --target=armv5teb-linux \ Jul 15 08:46:29 and that is not needed when compiling nativly Jul 15 08:46:57 Ok, that's a -native .bb Jul 15 08:47:23 That should be coming from $(BUILD_SYS), which is changed if $(BUILD_ARCH) is changed. Jul 15 08:48:00 So a change in local.conf should fix it. Jul 15 08:48:20 <[g2]> I did change the local.conf BUILD_ARCH Jul 15 08:48:27 A change in local.conf will fix it. Jul 15 08:49:05 And it is very very broken without that change... Jul 15 08:49:57 Oh, ok, it may not be that bad - I was looking at line 376 of bitbake.conf, but we don't have any conf/build/* Jul 15 08:51:02 ... that said creating conf/build/armeb-linux.conf could be helpful for a native build environment Jul 15 08:52:05 <[g2]> Ok... somehow I lost the _ARCH part in the BUILD_ARCH Jul 15 08:52:44 Did you clear tmp? Jul 15 08:53:43 [g2]: do you have that list of changes needed to setup the native build? Jul 15 08:57:15 <[g2]> jbowler, did a couple of minutes ago Jul 15 08:57:19 <[g2]> it's building Jul 15 08:57:29 <[g2]> DaKa2, yes i've to a full list of changes Jul 15 08:57:44 wiki? Jul 15 08:57:44 <[g2]> I've got Jul 15 08:57:53 <[g2]> not yet Jul 15 08:58:10 somewhere I can get it? :-) Jul 15 08:58:13 <[g2]> the wiki was updated yesterday with the native compile stuff Jul 15 08:58:39 <[g2]> ok... maybe my local file was Jul 15 08:58:59 <[g2]> checking the wiki Jul 15 09:10:17 <[g2]> DaKa2, I'm diff'ing now between the wiki an my stuff Jul 15 09:10:37 ah, good Jul 15 09:10:53 I managed to break my nfs root completly yesterday so I have to redo everything Jul 15 09:11:22 experimenting with bitbake builds as root is maybe not that smart :-) Jul 15 09:11:41 :-\ Jul 15 09:11:51 <[g2]> DaKa2, .... Jul 15 09:12:06 ok, this time I added a user fo it, ok? :-) Jul 15 09:12:08 <[g2]> you know I just switched to a normal user in the last few days :) Jul 15 09:12:31 I always do things as a normal user on linux, but not the nslu2 :-) Jul 15 09:12:37 haha.. [g2] never broke anything? Jul 15 09:12:49 <[g2]> I had built python and perl with tests as normal user Jul 15 09:13:15 <[g2]> I break stuff all the time, but not that Jul 15 09:13:36 <[g2]> it's actually surprising I don't break *more* stuff Jul 15 09:13:41 :) Jul 15 09:14:11 Im make installing python from last days build tree, should work Jul 15 09:14:29 nice with only one test failed Jul 15 09:15:32 couldn't that floating point thing be fixed? or would that need to much cpu? Jul 15 09:15:37 <[g2]> now would be a great time to resolve two issue Jul 15 09:15:49 <[g2]> 1) /etc/update-alternatives Jul 15 09:16:04 <[g2]> 2) whether perl is really needed in the default install Jul 15 09:16:26 3) adding hostname to /etc/hosts Jul 15 09:16:42 <[g2]> 3) is easy Jul 15 09:16:50 yup, but still needs to be done Jul 15 09:17:43 <[g2]> jbowler, could you do me a favor and append the host name to /etc/hosts if dhcp and create a line if static in turnup ? Jul 15 09:18:27 <[g2]> meaning if dhcp 127.... localhost thishost Jul 15 09:21:05 I don't understand, but if you want to modify /etc/hosts you need to do it in a way such that it doesn't happen on every run of the dhcp script (because of rewriting the flash). There's already a shell function in there to help with this. Jul 15 09:21:30 no, just on turnup init Jul 15 09:21:50 <[g2]> yeah just on turnup init Jul 15 09:22:05 <[g2]> DaKa2, http://www.nslu2-linux.org/wiki/HowTo/OpenSlugDevInst Jul 15 09:22:40 Enter a slugbug describing the problem... Jul 15 09:22:41 nice Jul 15 09:24:45 <[g2]> heh... I guess that was a no Jul 15 09:31:14 <[g2]> DaKa2, in /usr/sbin/turnup in init_network() between lines 750-760 it's determined whether it's static or dhcp Jul 15 09:31:38 <[g2]> I'll get find my keys and get setup to change the repo Jul 15 09:31:49 :) Jul 15 09:32:12 <[g2]> I'll put the sed's in there for changing the /etc/hosts file Jul 15 09:32:33 nice, that fixes lots of problems Jul 15 09:32:52 <[g2]> what does it fix besides the two python tests ? Jul 15 09:32:58 many networking things need that Jul 15 09:33:02 saned for one Jul 15 09:33:13 I had to note that in the sane howto Jul 15 09:33:29 <[g2]> I'll have to check where the host name actually gets initialized Jul 15 09:34:57 * [g2] bops over to other slug to test turnup init change Jul 15 09:35:07 :) Jul 15 09:37:10 hm.. python didn't like just make install from a previously compiled tree on another box.. Jul 15 09:38:02 <[g2]> we need to find a python expert Jul 15 09:38:39 <[g2]> so we can just fix the oe version to be as adequate as the natively compiled version Jul 15 09:38:56 python is one of those things I've been thinking to look at, but never really did Jul 15 09:39:24 <[g2]> there's a lot of cool stuff out there Jul 15 09:41:12 yeah, and too little time Jul 15 09:51:35 <[g2]> DaKa2, getting close the static case just about works Jul 15 09:51:47 :) Jul 15 09:52:12 I dont think that the alternatives has to be that much harder Jul 15 09:52:24 update-alternatives that is.. Jul 15 09:52:53 <[g2]> many of these things are not hard Jul 15 09:53:06 <[g2]> it just takes a little while to grasp them Jul 15 09:53:19 just add a do_install_append() that renames the files, and the postinst and postrm things Jul 15 09:53:22 yes Jul 15 09:53:23 exactly Jul 15 09:53:58 maybe should be a nice idea to talk to mickeyl, since I guess busybox is an important package :-) Jul 15 09:54:05 <[g2]> the beauty of working together is progress can roll *way* faster because there tremendous talent around Jul 15 09:54:24 <[g2]> we build a custom busybox in OpenSlug Jul 15 09:54:33 <[g2]> the busybox config is pass in Jul 15 09:54:34 ahh, then things are allright Jul 15 09:54:54 then we can do what we want :-))) Jul 15 09:54:58 <[g2]> I don't think things have changed that much Jul 15 09:55:02 <[g2]> in a year Jul 15 09:55:32 <[g2]> well I've been wanting to have a chat with the OpenSlug stakeholders about direction Jul 15 09:55:52 <[g2]> we're going through some growing pains right now which is totally normal Jul 15 09:56:22 yup Jul 15 09:56:33 just need to take care of them correctly Jul 15 09:56:42 if not, things could go bad Jul 15 09:56:57 <[g2]> how so ? Jul 15 09:57:46 well, I dont really have a good example Jul 15 09:58:04 <[g2]> cat /etc/hosts Jul 15 09:58:04 <[g2]> 127.0.0.1 localhost.localdomain localhost Jul 15 09:58:04 <[g2]> 192.168.123.9 Turbo4 Jul 15 09:58:10 :-) Jul 15 09:58:26 <[g2]> it has a bug Jul 15 09:59:03 <[g2]> right now I'm appending the line to the file Jul 15 09:59:18 <[g2]> I really should sed the line based on the IP addr Jul 15 09:59:23 <[g2]> and possibly the name Jul 15 09:59:43 <[g2]> but it only happens if you turnup init more than once Jul 15 09:59:51 yup.. but who does that? Jul 15 09:59:57 well, maybe someone.. Jul 15 10:00:22 <[g2]> IMHO not having that stuff in there in the first place is worse Jul 15 10:00:32 <[g2]> I'll just add a little echo in the about that Jul 15 10:00:49 its better than nothing Jul 15 10:01:08 <[g2]> no... on second thought I'll just right a slugbug about it :) Jul 15 10:01:18 the only thing different on openslug in busybox is: Jul 15 10:01:18 # This disables the syslog startup links in openslug (see openslug-init) Jul 15 10:01:18 INITSCRIPT_PARAMS_${PN}_openslug = "start 20 ." Jul 15 10:01:39 but we could add a do_install_append_openslug probably Jul 15 10:01:44 and the other things Jul 15 10:03:22 oh, and a differen defconfig Jul 15 10:04:02 <[g2]> DaKa2, longer-term meaning couple weeks/months what do you want to do with OpenSlug ? Jul 15 10:04:19 dont know really.. Jul 15 10:04:27 it already does all I really need Jul 15 10:04:49 but I would like to see some cleanup Jul 15 10:05:00 that is, the update-alternatives stuff and things like that Jul 15 10:05:08 actually getting running packages in the feed Jul 15 10:05:22 <[g2]> yeah Jul 15 10:05:34 <[g2]> productization Jul 15 10:05:39 exactly Jul 15 10:05:51 <[g2]> ok we're on the same wave length Jul 15 10:05:55 I want something working, and seeing a release would be nice Jul 15 10:06:28 <[g2]> a release that bundles all the fixed in ? Jul 15 10:07:01 well, when everuthing is working, seeing a 2.2 final would be nice Jul 15 10:07:08 and then a new 2.3 beta starting Jul 15 10:09:54 <[g2]> well that's the root of one of the problems we are having Jul 15 10:10:43 <[g2]> currently there's an unstable which is a developer version and some ppl would like stable stuff to be more stable Jul 15 10:14:03 <[g2]> Ok.. static is working fine Jul 15 10:14:08 <[g2]> now on to dynamic Jul 15 10:16:49 <[g2]> Hey beewoolie-afk ! Jul 15 10:19:14 <[g2]> DaKa2, beewoolie is our bootloader expert he wrote APEX Jul 15 10:19:34 ah Jul 15 10:19:43 <[g2]> the Redboot replacement bootloader I've been running with Jul 15 10:19:53 I was thinking about trying that, but decided not to Jul 15 10:20:07 <[g2]> really how come ? Jul 15 10:20:26 no jtag, and I really dont want to mess up.. Jul 15 10:20:45 <[g2]> reasonable for now Jul 15 10:20:59 chicken! ;) Jul 15 10:21:02 ... Jul 15 10:21:11 <[g2]> VoodooZ_work, are you running APEX ? Jul 15 10:21:53 no. :( Jul 15 10:21:56 but I want to. Jul 15 10:22:06 If somebody holds my hand :) Jul 15 10:22:17 haha Jul 15 10:22:41 well, I just need to find a jtag, and test it Jul 15 10:22:52 It's all about speed for me. so if I can reduce this horribly long ~60 seconds boot I will do it. Jul 15 10:23:15 Flash now, worry about JTAG later.... :) Jul 15 10:23:20 ... Jul 15 10:26:41 wow! It's amazing how much a good hand optimize loop will squeeze out. Jul 15 10:27:06 I went from 11fps to 20fps by unrolling the inner loop of my erosion function. yipee! Jul 15 10:27:13 wow Jul 15 10:27:41 VoodooZ_work: the locate fire thingy? Jul 15 10:27:52 I'm not surprised too, 4 for loops nested. 320x240x3x3. Just unrolling the last x3 did that! Jul 15 10:28:14 DaKa2, not specifically, but I might use it for that. Jul 15 10:28:19 ah Jul 15 10:28:25 just fooling around with basic vision. Jul 15 10:28:45 I got a nice motion detector working too. Jul 15 10:29:08 thats what I was planning to do.. when my bot actually can move Jul 15 10:29:10 not that useful but could be used to terrorize my cat! :) Jul 15 10:29:17 :) Jul 15 10:29:40 The intel appnote section had a real good article on writting efficient C code for the ARM. It was a good refresher. Jul 15 10:30:05 DaKa2, let me know if I can help you in any way. Jul 15 10:30:42 [g2]: yo g Jul 15 10:31:04 <[g2]> hey big dog what's up! Jul 15 10:31:07 well, I need to finish the control logic and the other half of the mosfets for the motorcontroller, then its juts software left Jul 15 10:31:41 DaKa2: BTW, you can try APEX without flashing it. Jul 15 10:31:47 <[g2]> beewoolie-afk, wanna hear something funny ? Jul 15 10:31:49 DaKa2, cool. The hardware was the toughest part for me as I'm not as mechanically gifted as some of my friends. Jul 15 10:31:56 [g2]: tickle me. Jul 15 10:32:04 beewoolie-afk: oh? Jul 15 10:32:19 <[g2]> DaKa2 and I are running bitbake on the slug! Jul 15 10:32:23 DaKa2: Yeah. You can load it into RAM with Redboot and see it work. Jul 15 10:32:37 * beewoolie-afk is rotfl Jul 15 10:32:41 I guess I could try that too. Jul 15 10:32:56 There's help on this http://wiki.buici.com Jul 15 10:33:10 DaKa2, Now I'm starting to look into hough transform for extracting line segment from an image so I can use it to align with walls or finding landmarks. Jul 15 10:33:12 well, still might fuckup when reflashing... but might actually try that some time Jul 15 10:33:32 <[g2]> beewoolie-afk it'll actually solve our native packaging issues Jul 15 10:33:43 [g2]: How is that? Jul 15 10:33:43 VoodooZ_work: ahh.. I guess I could use some help on the software part, I understood about 0 of that Jul 15 10:33:47 <[g2]> once we get it all sorted out Jul 15 10:34:08 <[g2]> because we've natively built and run ipkgs Jul 15 10:34:36 DaKa2, don't worry, I didn't know any of it 2 months ago but I learned quite a big toying with it. Jul 15 10:34:52 <[g2]> there's some auto* stuff to work out and then we'll be building big packages Jul 15 10:34:55 I've never gotten this native build issue. I've been cross-building for months. It's hard to setup because some package maintainers are lazy, but it tends to work. Jul 15 10:35:29 <[g2]> Ok... mister "Pls update the stable feed" .... :) Jul 15 10:35:48 beewoolie-afk, let us know when you get a working cross-built perl and emacs Jul 15 10:35:51 And how is that a native build issue? Jul 15 10:36:19 jacques: OK. I'll do. IIRC, perl is pretty bad because it uses it's own script. Jul 15 10:36:59 I've built a lot of other packages, tho. Samba and nfs-utils should be fine cross built. At least, I've build samba without difficulty. Jul 15 10:37:22 <[g2]> samba and nfs-utils are built cross Jul 15 10:37:55 so, what's the issue with the feed... ;-) Jul 15 10:38:08 apache, perl and python doesnt build cross, any more? Jul 15 10:38:31 <[g2]> python mostly works cross Jul 15 10:38:39 apache should work, but noone has done it.. Jul 15 10:38:56 <[g2]> php Jul 15 10:39:13 isn't apache kinda big for the slug? Jul 15 10:39:24 <[g2]> runs fine for me on the fatslug Jul 15 10:39:36 <[g2]> with php5 Jul 15 10:39:40 <[g2]> and openssl Jul 15 10:40:07 <[g2]> not that I've really stress tested it Jul 15 10:40:31 <[g2]> but loading up all the man pages in https worked fine has was surprisingly snappy Jul 15 10:42:36 and how much ram does the fatslug have... [rhetoric alert] Jul 15 10:42:46 <[g2]> 64 Jul 15 10:43:07 BTW, has anyone else noticed the file size issue that I found with samba? Jul 15 10:43:27 Very large files appear to display the wrong file size when viewed via samba. Jul 15 10:49:00 that seems strange.. Jul 15 10:49:15 <-- not using samba on the slug Jul 15 10:49:45 <--- need food, brb Jul 15 10:51:02 <[g2]> NAiL, ping Jul 15 10:55:29 WEE :-) things build Jul 15 11:07:10 and, things build with lots of stuff in ASSUME_PROVIDED :) Jul 15 11:11:09 hi, someone there? Jul 15 11:11:37 <[g2]> DaKa2, what do you have building ? Jul 15 11:11:47 <[g2]> foreal, what's up Jul 15 11:12:08 hey i got a problem with a second ethernet dev on my nslu2 Jul 15 11:12:14 I just build random stuff to test what works Jul 15 11:12:22 can someone help me Jul 15 11:12:33 <[g2]> unslung or openslug ? Jul 15 11:12:36 openslug Jul 15 11:12:44 <[g2]> which version ? Jul 15 11:12:54 <[g2]> natively built or binary release ? Jul 15 11:13:12 2 Jul 15 11:13:29 <[g2]> DaKa2, what's your ASSUME_PROVIDED line ? Jul 15 11:13:32 binary Jul 15 11:13:41 <[g2]> ok 2.0 binary Jul 15 11:13:44 yes Jul 15 11:13:52 [g2]: Jul 15 11:13:53 ASSUME_PROVIDED = "libtool-cross libtool-native automake-native autoconf-native \ Jul 15 11:13:56 quilt-native ipkg-utils-native gnu-config-native \ Jul 15 11:13:58 virtual/armeb-linux-gcc virtual/libc update-rc.d" Jul 15 11:14:13 but Im having libtool problems with compiling jpeg.. so Jul 15 11:14:28 dont we all love libtool :-) Jul 15 11:14:56 i have a linksys usb200m usb to etherne adapter, an i installed the kernel module for usbnet. Jul 15 11:14:59 <[g2]> what's the error AM_PROG_... something Jul 15 11:15:37 AM_PROG? Jul 15 11:16:37 everything looks good, but i can not ping eth2. Jul 15 11:16:51 [g2]: ipkg install libtool quilt ipkg-utils to make that ASSUME_PROVIDED work btw Jul 15 11:17:25 <[g2]> is libtool 1.5.8 in the build ? Jul 15 11:17:33 <[g2]> that's what I'm running natively Jul 15 11:18:01 libtool_1.5.10-r2_armeb.ipk Jul 15 11:18:02 <[g2]> foreal, ifconfig eth0 down Jul 15 11:18:12 <[g2]> then try to ping on eth1 intf Jul 15 11:18:34 <[g2]> foreal, are you connected via serial ? Jul 15 11:18:45 no im connected via eth1 Jul 15 11:19:04 ssh session with putty Jul 15 11:19:05 <[g2]> eth0 Jul 15 11:19:12 yeah i mean Jul 15 11:19:22 <[g2]> and you can't ping on eth1 Jul 15 11:19:42 <[g2]> eth2 Jul 15 11:20:13 <[g2]> eth2 and eth0 are on the same lan segment I'm guessing Jul 15 11:20:37 <[g2]> ping eth2's IP addr and do a couple ifconfig Jul 15 11:20:45 <[g2]> ifconfig eth2 Jul 15 11:20:59 <[g2]> and tell me if the packet counts are incrementing Jul 15 11:21:12 k sec Jul 15 11:21:49 <[g2]> DaKa2, Jul 15 11:21:52 <[g2]> configure.in:208: error: possibly undefined macro: AM_PROG_LIBTOOL Jul 15 11:21:52 <[g2]> If this token and others are legitimate, please use m4_pattern_allow. Jul 15 11:21:52 <[g2]> See the Autoconf documentation. Jul 15 11:21:52 <[g2]> autoreconf: /usr/bin/autoconf failed with exit status: 1 Jul 15 11:22:08 uhm.. Jul 15 11:22:13 what program? Jul 15 11:22:23 <[g2]> perl-native Jul 15 11:22:36 <[g2]> building db3-native Jul 15 11:23:12 strange.. Jul 15 11:23:37 so i got eht0 connected with my lanswitch, an eth2 is also connected with my switch. i can ping both ipaddresses, but if i disconnect eth0, i cant ping to eth2 anymore. Jul 15 11:25:05 they both in teh same lan segment Jul 15 11:25:40 eth0 Link encap:Ethernet HWaddr 00:0F:66:7C:04:EA Jul 15 11:25:40 inet addr:192.168.96.85 Bcast:192.168.96.255 Mask:255.255.255.0 Jul 15 11:25:40 UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 Jul 15 11:25:40 RX packets:531 errors:0 dropped:0 overruns:0 frame:0 Jul 15 11:25:40 TX packets:603 errors:0 dropped:0 overruns:0 carrier:0 Jul 15 11:25:41 collisions:0 txqueuelen:256 Jul 15 11:25:43 RX bytes:57262 (55.9 KiB) TX bytes:65341 (63.8 KiB) Jul 15 11:25:45 eth2 Link encap:Ethernet HWaddr 00:10:60:84:B0:0D Jul 15 11:25:49 UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 Jul 15 11:25:51 RX packets:0 errors:0 dropped:0 overruns:0 frame:0 Jul 15 11:25:53 TX packets:0 errors:0 dropped:0 overruns:0 carrier:0 Jul 15 11:25:55 collisions:0 txqueuelen:1000 Jul 15 11:25:57 RX bytes:0 (0.0 B) TX bytes:0 (0.0 B) Jul 15 11:26:08 you have no ip on eth2? Jul 15 11:26:14 yeah i see Jul 15 11:26:15 sec Jul 15 11:26:24 eth2 Link encap:Ethernet HWaddr 00:10:60:84:B0:0D Jul 15 11:26:24 inet addr:192.168.96.86 Bcast:192.168.96.255 Mask:255.255.255.0 Jul 15 11:26:24 UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 Jul 15 11:26:24 RX packets:0 errors:0 dropped:0 overruns:0 frame:0 Jul 15 11:26:25 TX packets:0 errors:0 dropped:0 overruns:0 carrier:0 Jul 15 11:26:27 collisions:0 txqueuelen:1000 Jul 15 11:26:29 RX bytes:0 (0.0 B) TX bytes:0 (0.0 B) Jul 15 11:26:58 seems a bit strange with 0 RX and TX packages.. Jul 15 11:27:08 <[g2]> no they are all going out eth0 Jul 15 11:27:16 i think so Jul 15 11:27:26 but damn why Jul 15 11:27:33 <[g2]> ifconfig eth2 192.168.123.1 Jul 15 11:27:37 <[g2]> ifconfig eth2 192.168.123.1 up Jul 15 11:28:17 eth2 Link encap:Ethernet HWaddr 00:10:60:84:B0:0D Jul 15 11:28:17 inet addr:192.168.123.1 Bcast:192.168.123.255 Mask:255.255.255.0 Jul 15 11:28:17 UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 Jul 15 11:28:17 RX packets:0 errors:0 dropped:0 overruns:0 frame:0 Jul 15 11:28:18 TX packets:0 errors:0 dropped:0 overruns:0 carrier:0 Jul 15 11:28:20 collisions:0 txqueuelen:1000 Jul 15 11:28:20 <[g2]> ping 192.168.123.255 Jul 15 11:28:22 RX bytes:0 (0.0 B) TX bytes:0 (0.0 B) Jul 15 11:28:51 [g2]: what automake version do you have? Jul 15 11:29:05 and autoconf Jul 15 11:29:22 <[g2]> 1.9.3 and 2.5.9 Jul 15 11:29:45 uhm.. then things should work.. Jul 15 11:30:36 <[g2]> ain't that always the case :) Jul 15 11:30:46 <[g2]> things *should* work :) Jul 15 11:31:00 :) Jul 15 11:31:02 <[g2]> foreal, wb Jul 15 11:31:17 <[g2]> so do you see stuff going out eth2 ? Jul 15 11:31:23 yeah Jul 15 11:31:28 it works Jul 15 11:31:51 eth2 Link encap:Ethernet HWaddr 00:10:60:84:B0:0D Jul 15 11:31:52 inet addr:192.168.123.1 Bcast:192.168.123.255 Mask:255.255.255.0 Jul 15 11:31:52 UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 Jul 15 11:31:52 RX packets:117 errors:0 dropped:0 overruns:0 frame:0 Jul 15 11:31:52 TX packets:10 errors:0 dropped:0 overruns:0 carrier:0 Jul 15 11:31:52 collisions:0 txqueuelen:1000 Jul 15 11:31:54 RX bytes:11575 (11.3 KiB) TX bytes:676 (676.0 B) Jul 15 11:32:32 <[g2]> ok... so do you understand that it was just using the eth0 as the default Jul 15 11:33:03 * [g2] did a similar thing when I first set it up Jul 15 11:33:16 :) Jul 15 11:33:52 <[g2]> I wound up using spliting the Class C into 0-127 and 128-255 just on the slug device Jul 15 11:34:16 <[g2]> with the approp netmask Jul 15 11:34:46 <[g2]> so foreal what are you going to do with the second intf ? Jul 15 11:35:15 i try to use it as a router, next step pppoe over that intf Jul 15 11:35:39 <[g2]> that should be pretty straight forward Jul 15 11:36:09 <[g2]> I think rwhitby did that with ppp running over a serial upling Jul 15 11:36:17 <[g2]> upling heh.... uplink Jul 15 11:36:41 <[g2]> foreal, you know about the wiki right ? Jul 15 11:37:19 yeah Jul 15 11:37:36 <[g2]> http://www.nslu2-linux.org/wiki/Unslung/PPPoE Jul 15 11:38:09 <[g2]> I guess that's acutally written for OpenSlug Jul 15 11:38:14 ah, nice thats what im lookin fo Jul 15 11:38:28 somehow my slug is damed slow right now.. compiling perl in the background while building other stuff with bitbake Jul 15 11:38:51 <[g2]> ouch... Jul 15 11:39:03 <[g2]> DaKa2, is one swapping machine Jul 15 11:39:10 yep Jul 15 11:39:16 <[g2]> you need another slug pal :) Jul 15 11:39:23 I have three Jul 15 11:39:26 wow Jul 15 11:39:27 slugarmy! Jul 15 11:39:31 lol Jul 15 11:39:46 <[g2]> but I'll bet you are trying to lock the cpu again Jul 15 11:39:59 <[g2]> 3 that's nothing man Jul 15 11:40:16 once things are working Im going to buy a few for work Jul 15 11:40:31 setting up a few scanner servers Jul 15 11:40:44 <[g2]> you'll probably like my HW Jul 15 11:40:59 yup.. I know I will :-) Jul 15 11:41:36 <[g2]> hmmm... distcc with 50 boxes ... Jul 15 11:41:45 <[g2]> feeding into the gigE backbone Jul 15 11:41:53 <[g2]> fun... Jul 15 11:41:55 :) Jul 15 11:42:30 I have 700 PCs on a 1gbit backbone with 100mbit to each.. :-) Jul 15 11:43:35 ARGH, stupid jpeg_6b.bb... Jul 15 11:43:42 EXTRA_OEMAKE='"LIBTOOL=${STAGING_BINDIR}/${HOST_SYS}-libtool"' Jul 15 11:43:52 THATS NOT WHERE MY LIBTOOL IS! Jul 15 11:44:10 I wonder if more packages do that.. Jul 15 11:45:45 hey wha about the turboslugs? http://www.nslu2-linux.org/wiki/HowTo/OverClockTheSlug who tryed that before? Jul 15 11:46:23 all of us? Jul 15 11:46:29 ok Jul 15 11:46:39 :) Jul 15 11:46:51 you're just restoring the cpu to its rated speed Jul 15 11:47:05 no heatproblems? Jul 15 11:47:38 thats one of my next points Jul 15 11:47:48 <[g2]> DaKa2, 700 PCs ??? Man that's some heat Jul 15 11:47:52 DaKa2 states that he has had one lockup, but ambient temp for him was 30-31 C Jul 15 11:48:14 [g2]: hehe, well, its at work, all mine :-))) MUAHAHAHA Jul 15 11:48:28 with *normal* ambient temps, I wouldn't think heat would be an issue Jul 15 11:48:32 ByronT: actually, a few more, a small fan fixed that Jul 15 11:48:47 <[g2]> DaKa2, server farm ? Jul 15 11:48:49 DaKa2, really - I only saw that once... Jul 15 11:49:00 no thats ok, i try that. :) Jul 15 11:49:12 [g2]: no, all the computers in the building :-) still all mine :-)))) Jul 15 11:49:20 where is your slug located anyways? over a stove? Jul 15 11:49:25 ;) Jul 15 11:49:28 ByronT: I had a few more some hours ago Jul 15 11:49:38 <[g2]> in the middle of 700 PCs :) Jul 15 11:49:40 well, my apartment is that hot.... Jul 15 11:49:45 I HATE THE SUMMER Jul 15 11:49:48 lol Jul 15 11:49:49 ah... Jul 15 11:50:05 my computer area at home is in front of an A/C register... Jul 15 11:50:30 armeb-linux-libtool: compile: unable to infer tagged configuration Jul 15 11:50:30 armeb-linux-libtool: compile: specify a tag with `--tag' Jul 15 11:50:35 dammit.. Jul 15 11:50:47 <[g2]> I think I'm getting that too Jul 15 11:50:52 ambient temp is somewhere between 16-18 C Jul 15 11:50:59 GRRR Jul 15 11:51:06 thats what I want Jul 15 11:51:14 excuse me... (bad temp conversion) 18-20 C Jul 15 11:51:27 Im moving into the serverroom here... I cant stand the heat Jul 15 11:51:35 <[g2]> DaKa2, the funny thing is the command line from the same dir works! Jul 15 11:51:52 [g2]: dammit... thats not funny... Jul 15 11:52:29 <[g2]> in a preverse way it is Jul 15 11:53:09 hm.. doesnt work for me.. Jul 15 11:53:19 still same error Jul 15 11:53:19 <[g2]> from the command line ? Jul 15 11:53:23 yes Jul 15 11:53:25 /home/slug/openslug/tmp/staging/armeb-linux/bin/armeb-linux-libtool --mode=compile armeb-linux-gcc -march=armv5te -mtune=xscale -I/home/slug/openslug/tmp/staging/armeb-linux/include -fexpensive-optimizations -fomit-frame-pointer -frename-registers -O2 -D_REENTRANT -I/home/slug/openslug/tmp/staging/armeb-linux/include -I. -I/home/slug/openslug/tmp/work/jpeg-6b-r0/jpeg-6b -c /home/slug/openslug/tmp/work/jpeg-6b-r0/jpeg-6b/jcapimin.c Jul 15 11:54:01 <[g2]> what are you trying to build ? Jul 15 11:54:12 random stuff :-) Jul 15 11:54:17 this is jpeg Jul 15 11:54:26 <[g2]> does it have deps ? Jul 15 11:54:38 but it could be because I skipped all the -natives Jul 15 11:54:43 nope Jul 15 11:56:15 <[g2]> ERROR: Nothing provides libtool-cross Jul 15 11:57:18 ahh, thats in my ASSUME_PROVIDED Jul 15 11:57:26 could be the problem.. Jul 15 11:57:49 <[g2]> I just ASSUME_PROVIDED it Jul 15 11:58:07 but I dont want to crossbuild.. stupid thing... Jul 15 11:59:33 <[g2]> + make LIBTOOL=/home/tom/tmp/staging/armeb-linux/bin/armeb-linux-libtool Jul 15 11:59:33 <[g2]> /home/tom/tmp/staging/armeb-linux/bin/armeb-linux-libtool --mode=compile armeb-linux-gcc -march=armv5te -mtune=xscale -I/home/tom/tmp/staging/armeb-linux/include -fexpensive-optimizations -fomit-frame-pointer -frename-registers -O2 -D_REENTRANT -I/home/tom/tmp/staging/armeb-linux/include -I. -I/home/tom/tmp/work/jpeg-6b-r0/jpeg-6b -c /home/tom/tmp/work/jpeg-6b-r0/jpeg-6b/jcapimin.c Jul 15 11:59:33 <[g2]> make: /home/tom/tmp/staging/armeb-linux/bin/armeb-linux-libtool: Command not found Jul 15 11:59:33 <[g2]> make: *** [jcapimin.lo] Error 127 Jul 15 11:59:35 <[g2]> + die oe_runmake failed Jul 15 12:00:02 <[g2]> lemme put a symlink in for armeb-linux-libtool Jul 15 12:00:09 right.. I # OE_something in jpeg_6b.bb Jul 15 12:00:33 <[g2]> does you compile fail ? Jul 15 12:00:38 lets see Jul 15 12:00:42 <[g2]> for the make after it ? Jul 15 12:01:26 <[g2]> s/for/or Jul 15 12:01:41 <[g2]> heh .. you Jul 15 12:01:53 its compiling right now Jul 15 12:02:02 <[g2]> with the symlink in there ? Jul 15 12:02:20 I had one earlier, now I modified the bb-file Jul 15 12:02:39 <[g2]> to run libtool Jul 15 12:02:46 commented out EXTRA_OEMAKE Jul 15 12:03:04 #EXTRA_OEMAKE='"LIBTOOL=${STAGING_BINDIR}/${HOST_SYS}-libtool"' Jul 15 12:03:13 that is just stupid... Jul 15 12:03:53 damn slow.. but it is compiling perl at the same time Jul 15 12:05:41 seems to work Jul 15 12:06:24 we should really make this work in the monotone repo.. like, make a new dist openslug-native or something Jul 15 12:06:42 so we can make changes for native building Jul 15 12:06:52 not that jpeg actually needs native.. Jul 15 12:06:53 <[g2]> absolutely Jul 15 12:07:09 <[g2]> but I still trying to understand what works and what doesn't Jul 15 12:07:29 same here Jul 15 12:07:30 <[g2]> interestingly trying bb jpeg-native runs into the same issue Jul 15 12:07:38 :) Jul 15 12:08:06 <[g2]> but let me try the symlink there Jul 15 12:08:20 do that, I dont think it works, but maybe Jul 15 12:09:02 how can a library for decoding jpeg take so much time to compile... Jul 15 12:13:12 whoops... this is my non turboslug.. I knew I should have added them to dns... Jul 15 12:13:15 oh well.. Jul 15 12:13:43 then Im the first with a running bitbake on a normal slug :-) Jul 15 12:16:08 <[g2]> DaKa2, which file are you up to ? Jul 15 12:16:21 compiling? Jul 15 12:16:28 <[g2]> yeah Jul 15 12:16:30 armeb-linux-gcc -march=armv5te -mtune=xscale -I/home/slug/openslug/tmp/staging/armeb-linux/include -fexpensive-optimizations -fomit-frame-pointer -frename-registers -O2 -D_REENTRANT -I/home/slug/openslug/tmp/staging/armeb-linux/include -I. -I/home/slug/openslug/tmp/work/jpeg-6b-r0/jpeg-6b -c -fPIC -DPIC /home/slug/openslug/tmp/work/jpeg-6b-r0/jpeg-6b/transupp.c -o .libs/transupp.lo Jul 15 12:17:41 NOTE: package jpeg-6b-r0: task do_compile: completed Jul 15 12:19:05 NOTE: package jpeg-6b: completed Jul 15 12:22:18 DaKa2: that '-O2' should be '-Os' Jul 15 12:22:30 dammit.. Jul 15 12:22:34 <[g2]> ok I'm up to transupp on jpeg-native Jul 15 12:22:36 :-) Jul 15 12:22:50 <[g2]> NOTE: build 200507151513: completed Jul 15 12:22:50 <[g2]> Build statistics: Jul 15 12:23:13 <[g2]> NOTE: package jpeg-native-6b-r0: task do_build: completed Jul 15 12:23:18 :) Jul 15 12:23:58 <[g2]> ok... so there's still somthing with the autotool Jul 15 12:24:17 Some of the makefiles kill the CFLAGS settings - highly annoying. Jul 15 12:25:05 [g2]: what package? Jul 15 12:25:27 <[g2]> do a db3-native Jul 15 12:25:49 Ill try that Jul 15 12:26:41 damn this thing is slow Jul 15 12:27:00 but starting again on my turbo at home would take to long Jul 15 12:27:48 ah, dammit.. my perl is not done yet Jul 15 12:27:51 so db3 fails Jul 15 12:28:07 <[g2]> nod Jul 15 12:28:35 <[g2]> you'll also have to switch /usr/bin/perl to be a symlink to the newly minted one Jul 15 12:29:31 hm.. I thought this would install it there.. Jul 15 12:32:13 <[g2]> I think the default is /usr/local/bin Jul 15 12:33:39 hm.. Jul 15 12:33:43 oh well, Ill just see Jul 15 12:34:14 (in a few hours) Jul 15 12:34:26 default is indeed /usr/local/bin Jul 15 12:34:34 I usually just symlink Jul 15 12:35:19 Hum. The default .conf files do not set SELECTED_OPTIMIZATION="-Os" Jul 15 12:36:07 <[g2]> there are lots of fish to fry Jul 15 12:38:19 <[g2]> DaKa2, ok... so what do you think about an gnu package that installs a full native development environment and all the gnu core stuff Jul 15 12:38:53 well, why not Jul 15 12:40:13 slugbug 199 Jul 15 12:40:59 hm.. Jul 15 12:50:10 <[g2]> well gotta run Jul 15 12:50:22 <[g2]> see you guys tomorrow Jul 15 13:08:03 is it already tomorrow, I must have fallen asleep :-) Jul 15 13:08:14 <[g2]> heh **** ENDING LOGGING AT Fri Jul 15 18:44:43 2005 **** BEGIN LOGGING AT Fri Jul 15 18:44:52 2005 **** ENDING LOGGING AT Fri Jul 15 23:59:56 2005