**** BEGIN LOGGING AT Wed Jul 13 23:59:57 2005 Jul 14 06:29:54 * [g2] notes bb running on Fatslug -- NOTE: Handling BitBake files: \ (0054/2545) [ 2 %] Jul 14 06:30:02 <[g2]> that was for mickeyl Jul 14 06:31:08 heh, that's pretty cool Jul 14 06:31:14 it'll probably take ages to parse though Jul 14 06:34:56 <[g2]> NOTE: Handling BitBake files: | (0191/2545) [ 7 Jul 14 06:34:59 <[g2]> % Jul 14 06:35:15 <[g2]> our openslug symlinks are only 600ish Jul 14 06:35:32 <[g2]> probably 3 or 4 not counting tools Jul 14 06:39:52 it's annoying how .bb-files are almost identical to gentoo ebuild files Jul 14 06:42:08 hopefully, portage will support cross-building soonish Jul 14 06:52:07 <[g2]> mickeyl, know anything about INHIBIT_DEFAULT_DEPS? Jul 14 06:52:35 <[g2]> I guess I should ask in #oe Jul 14 06:58:16 ya Jul 14 06:58:37 that's for (the rare case) where the default deps like patcher, glibc and a toolchain shouldn't be added Jul 14 06:58:49 i.e. the deps that come from base.bbclass Jul 14 06:59:05 <[g2]> right and in this case that's appropriate Jul 14 06:59:13 <[g2]> the native envirn is all installed Jul 14 06:59:33 <[g2]> lord knows we don't want to rebuild it again :) Jul 14 07:00:36 <[g2]> NOTE: Handling BitBake files: | (0353/0637) [55 %] Jul 14 07:00:43 <[g2]> heh Jul 14 07:00:50 I'm impressed Jul 14 07:00:56 I though it'd take longer Jul 14 07:01:09 Well... it probably will when it really starts to swap? ;) Jul 14 07:01:26 <[g2]> actually we haven't started yet Jul 14 07:01:32 <[g2]> and have plenty to go Jul 14 07:01:36 <[g2]> memory Jul 14 07:01:54 col Jul 14 07:01:56 cool Jul 14 07:02:49 <[g2]> Ok over 400 done and 9.6MB free yet Jul 14 07:03:58 <[g2]> and with ipkg -i :)))) Jul 14 07:06:08 ipkg install bitbake :-) Jul 14 07:06:13 now some symlinks Jul 14 07:08:34 ugh... Anyone that can help me with Makefiles a tad? Jul 14 07:09:00 The upload-openembedded-cross target in the master makefile just gives me "nothing to be done for..." Jul 14 07:12:36 <[g2]> NOTE: package quilt-native-0.39: started Jul 14 07:12:36 <[g2]> NOTE: package quilt-native-0.39-r0: task do_fetch: started Jul 14 07:12:36 <[g2]> NOTE: fetch http://www.oesources.org/source/current/quilt_savannah.nongnu.org_VER_0_39_.tar.gz Jul 14 07:12:36 <[g2]> wget: invalid option -- t Jul 14 07:12:36 <[g2]> BusyBox v1.00 (2005.07.02-01:41+0000) multi-call binary Jul 14 07:13:02 damned busybox Jul 14 07:13:24 not that -t is needed Jul 14 07:13:40 <[g2]> well that's just the test Jul 14 07:13:49 <[g2]> looks like the concept will fly Jul 14 07:14:00 <[g2]> lots of stuff to be worked out Jul 14 07:15:21 uhm.. Jul 14 07:16:08 slug@THG-SLUG1:~/openslug$ bb Jul 14 07:16:10 bash: /usr/bin/bitbake: /home/slug/openslug/tmp/staging/i686-linux/bin/python: bad interpreter: No such file or directory Jul 14 07:17:04 stupid cross build.. :-) Jul 14 07:17:14 <[g2]> nod. Jul 14 07:17:28 <[g2]> you're gonna need bash to locally Jul 14 07:17:47 ? Jul 14 07:18:22 <[g2]> the busybox sh script doesn't support the environ vars enough Jul 14 07:18:30 <[g2]> for . setup-env Jul 14 07:18:32 ah, Im already using bash Jul 14 07:18:49 first thing I ipkg on a newly installed slug Jul 14 07:18:53 <[g2]> well there ya go :) Jul 14 07:19:59 uhm. python glob module? Jul 14 07:20:27 <[g2]> Ok.. I build and tested my python natively Jul 14 07:20:45 ahh, that could explain a lot.. Jul 14 07:20:53 <[g2]> So I've get most everything, ditto perl Jul 14 07:20:58 slug@THG-SLUG1:~/openslug$ bb Jul 14 07:20:59 Traceback (most recent call last): Jul 14 07:20:59 File "/usr/bin/bitbake", line 21, in ? Jul 14 07:20:59 import sys, os, getopt, glob, copy, os.path, re Jul 14 07:20:59 ImportError: No module named glob Jul 14 07:21:14 <[g2]> nod. Jul 14 07:21:29 <[g2]> I ran into the same thing trying to run Zope/etc ... Jul 14 07:21:45 dammit.. Jul 14 07:22:02 <[g2]> this is exactly why we need to Jul 14 07:22:10 <[g2]> a) get this things packaged up Jul 14 07:22:16 <[g2]> b) fix the repo Jul 14 07:22:28 <[g2]> s/this/these Jul 14 07:22:35 ahh, glob is part of python-shell Jul 14 07:22:46 ipkg install python-shell Jul 14 07:24:00 and you need python-lang too Jul 14 07:24:31 that should be added to bitbake.bb in RDEPENDS probably, maybe.. Jul 14 07:24:41 <[g2]> this gets into the whole packaging/distro issue I was trying to mention earlier Jul 14 07:24:48 yup.. Jul 14 07:25:19 and python-pickle... Jul 14 07:25:31 Im starting to think ipkg install python-* Jul 14 07:25:49 <[g2]> even when you do that you'll still be missing stuff Jul 14 07:25:57 probably Jul 14 07:26:22 <[g2]> for certain things like Zope nod... been there done that Jul 14 07:26:29 <[g2]> dunno about bb's req's Jul 14 07:27:01 and python-io... Jul 14 07:27:05 cant be any more now Jul 14 07:27:19 <[g2]> well there are 40+ Jul 14 07:27:26 :) Jul 14 07:28:33 python-textutils... Jul 14 07:28:36 woho! Jul 14 07:28:49 no more depends Jul 14 07:29:18 "NOTE: Psyco JIT Compiler (http://psyco.sf.net) not available. Install it to increase performance." <-- does that make a big difference? Jul 14 07:29:43 <[g2]> it does but you probably don't have the memory for it :) Jul 14 07:29:49 <[g2]> nor do I Jul 14 07:30:08 are you running eith 64? Jul 14 07:30:10 with* Jul 14 07:30:14 <[g2]> :) Jul 14 07:30:22 <[g2]> FatTurbo Jul 14 07:30:26 :) Jul 14 07:30:49 NOTE: Parsing finished. 0 cached, 1 parsed, 0 skipped, 0 masked. Jul 14 07:30:50 NOTE: Building provider hash: [####################] (100%) Jul 14 07:30:50 NOTE: build 199912310354: started Jul 14 07:31:11 ok, then comes all the stuff needed to build.. Jul 14 07:31:33 ERROR: Nothing provides ipkg-utils-native Jul 14 07:31:37 <[g2]> right we need to figure out the INHERIT_NO_DEPS stuff Jul 14 07:32:17 <[g2]> make that INHIBIT_DEFAULT_DEPS Jul 14 07:32:39 how about just a dummy package that provides all we already have? Jul 14 07:33:01 there is something like Provides: in bb-files? Jul 14 07:33:02 <[g2]> possible Jul 14 07:33:21 <[g2]> but you've got the whole tool-chain issue etc... Jul 14 07:34:13 fun... Jul 14 07:37:07 [g2]: How do I figure out what kind of chips are on a ram stick? Jul 14 07:37:31 <[g2]> look at the chips Jul 14 07:37:57 <[g2]> look up the part number Jul 14 07:38:01 aha Jul 14 07:44:20 [g2]: how about ASSUME_PROVIDED ? Jul 14 07:45:19 <[g2]> 100%[====================================================================================================>] 294,966 339.93K/s Jul 14 07:45:19 <[g2]> 10:43:40 (338.83 KB/s) - `/home/tom/downloads/quilt_savannah.nongnu.org_VER_0_39_.tar.gz' saved [294966/294966] Jul 14 07:45:19 <[g2]> NOTE: Fetched quilt_savannah.nongnu.org_VER_0_39_.tar.gz from tarball stash, skipping checkout Jul 14 07:45:19 <[g2]> NOTE: package quilt-native-0.39-r0: task do_fetch: completed Jul 14 07:45:19 <[g2]> NOTE: package quilt-native-0.39-r0: task do_unpack: started Jul 14 07:45:21 <[g2]> NOTE: Unpacking /home/tom/downloads/quilt_savannah.nongnu.org_VER_0_39_.tar.gz to /home/tom/tmp/work/quilt-native-0.39-r0/ Jul 14 07:45:24 <[g2]> tar: unrecognized option `--no-same-owner' Jul 14 07:47:25 <[g2]> the plot sickens :) Jul 14 07:47:30 <[g2]> NOTE: package quilt-native-0.39-r0: task do_patch: started Jul 14 07:47:30 <[g2]> NOTE: Applying patch 'install.patch' Jul 14 07:47:30 <[g2]> ERROR: function do_patchcmd failed Jul 14 07:47:38 :) Jul 14 07:48:05 well.. Im stuck at python... Jul 14 07:48:22 <[g2]> really ? NO ? Jul 14 07:48:26 :) Jul 14 07:48:27 * [g2] is shocked Jul 14 07:48:37 Im not.. Jul 14 07:49:58 Well, I get to: "NOTE: package quilt-native-0.39-r0: task do_fetch: started" Jul 14 07:50:17 <[g2]> and you don't have the real wget installed Jul 14 07:50:35 <[g2]> you need that and the real tar to do the extract Jul 14 07:50:47 actully, I get a 30 line callback of python.. Jul 14 07:50:58 traceback* Jul 14 07:52:42 AttributeError: 'module' object has no attribute 'createCopy' Jul 14 07:53:01 that makes me understand whats wrong... not Jul 14 07:55:41 <[g2]> ohhh Jul 14 07:55:46 <[g2]> oooooh Jul 14 07:55:51 <[g2]> aaaaahhhhh Jul 14 07:55:59 something works? Jul 14 07:56:06 <[g2]> NOTE: package quilt-native-0.39-r0: task do_configure: completed Jul 14 07:56:06 <[g2]> NOTE: package quilt-native-0.39-r0: task do_compile: started Jul 14 07:56:31 <[g2]> NOTE: package quilt-native-0.39: completed Jul 14 07:56:50 <[g2]> followed by: Jul 14 07:57:04 <[g2]> ERROR: Nothing provides m4-native Jul 14 07:57:04 <[g2]> ERROR: dependency m4-native (for autoconf-native) not satisfied Jul 14 07:57:25 just add that to ASSUME_PROVIDED Jul 14 07:57:30 you have m4 installed right? Jul 14 07:57:44 <[g2]> think so Jul 14 07:58:09 <[g2]> in the pp ? Jul 14 07:58:11 <[g2]> bb ? Jul 14 07:58:17 I added in local.conf Jul 14 07:58:46 <[g2]> pastebin ? Jul 14 07:58:54 <[g2]> or here Jul 14 07:58:58 but you could add it in dists/openslug-native.conf or something... Jul 14 07:59:12 what? ASSUME_PROVIDED = "m4-native" Jul 14 07:59:25 <[g2]> ok Jul 14 08:00:16 so, I guess I need native python.. Jul 14 08:00:24 anything to think about? Jul 14 08:00:43 <[g2]> well I'm no python/perl/apache/zope expert Jul 14 08:01:04 well, you have python native built? Jul 14 08:01:20 <[g2]> yes Jul 14 08:01:39 just ./configure and make, or did you have to change anything? Jul 14 08:01:41 <[g2]> it's a few hours on a fatturb Jul 14 08:01:48 damn... Jul 14 08:02:09 <[g2]> yep that's it Jul 14 08:02:18 <[g2]> and make test Jul 14 08:02:23 <[g2]> and make install Jul 14 08:02:47 <[g2]> that's why I'd like to get that packaged up Jul 14 08:03:08 <[g2]> it's probably 6-8 hours on a non-fat slug Jul 14 08:03:31 well, with a working python I could make a native package, probably :-) Jul 14 08:03:51 <[g2]> nod. Even I just made one :) Jul 14 08:04:21 <[g2]> well not the full package yet Jul 14 08:04:38 <[g2]> but it's on its way Jul 14 08:07:15 <[g2]> hmm... I wonder where getopt comes from :) Jul 14 08:08:01 python-core? Jul 14 08:08:37 getopt as in the shell command? Jul 14 08:08:49 <[g2]> probaly Jul 14 08:08:59 util-linux normally Jul 14 08:09:29 <[g2]> kolla_, thx Jul 14 08:34:57 <[g2]> any quilters around ? Jul 14 08:35:25 my mom does quilts. Does that count? Jul 14 08:35:42 :) Jul 14 08:35:47 <[g2]> if she does OE Jul 14 08:37:03 nah, the old fashion kind Jul 14 08:37:15 <[g2]> Can you see the end of http://pastebin.ca/17698 Jul 14 08:41:25 Well, does /home/tom/tmp/work/autoconf-native-2.59-r3/autoconf-2.59/lib/autoconf/general.m4 exist? Jul 14 08:44:13 <[g2]> :) Jul 14 08:44:16 <[g2]> ls -l /home/tom/tmp/work/autoconf-native-2.59-r3/autoconf-2.59/lib/autoconf/general.m4 Jul 14 08:44:16 <[g2]> ls: /home/tom/tmp/work/autoconf-native-2.59-r3/autoconf-2.59/lib/autoconf/general.m4: No such file or directory Jul 14 08:44:25 that explains alot.. Jul 14 08:44:43 did it unpack correctly? Jul 14 08:44:53 do you have /home/tom/tmp/work/autoconf-native-2.59-r3/autoconf-2.59 ? Jul 14 08:45:20 <[g2]> ls /home/tom/tmp/work/autoconf-native-2.59-r3/autoconf-2.59/ Jul 14 08:45:20 <[g2]> patches Jul 14 08:45:27 :) Jul 14 08:45:39 it didnt unpack it there then Jul 14 08:45:41 :-) Jul 14 08:52:32 compiling python took 9+33 minutes, only make test and make install left Jul 14 09:05:42 <[g2]> DaKa2, OK... that was interesting Jul 14 09:06:28 <[g2]> you'll need to ipkg --force-overwrite install bzip2 Jul 14 09:06:59 <[g2]> the native bzip2 wasn't enought Jul 14 09:07:03 <[g2]> NOTE: Unpacking /home/tom/downloads/autoconf-2.59.tar.bz2 to /home/tom/tmp/work/autoconf-native-2.59-r3/ Jul 14 09:13:45 <[g2]> Daka ipkg install cvs is required for the ipkg-utils Jul 14 09:17:04 <[g2]> DaKa2, around >? Jul 14 09:40:09 yup.. Jul 14 09:40:26 dammit.. make test still not done.. Jul 14 09:40:53 <[g2]> DaKa2, you saw the success in #nslu2-linux ? Jul 14 09:41:31 hm.. now I did Jul 14 09:47:18 test test_random failed -- Traceback (most recent call last): Jul 14 09:47:18 File "/home/root/Python-2.4.1/Lib/test/test_random.py", line 313, in test_referenceImplementation Jul 14 09:47:21 self.assertAlmostEqual(a,e,places=14) Jul 14 09:47:24 AssertionError: 0.92848331726782152 != 0.92848331693227715 within 14 places Jul 14 09:47:38 uhm.. that is kind of strange.. Jul 14 09:47:38 normal Jul 14 09:47:41 ok.. Jul 14 09:47:48 * [g2] is too slow Jul 14 09:47:50 no hwfp Jul 14 09:48:08 ahh Jul 14 09:48:10 soft-float doesn't have the same precision Jul 14 09:48:23 <[g2]> I didn't understand the mime and other test failing though Jul 14 09:48:38 yeah I haven't looked into those Jul 14 09:48:38 I know why Jul 14 09:48:41 I think Jul 14 09:48:56 <[g2]> 244 tests OK. Jul 14 09:48:56 <[g2]> 3 tests failed: Jul 14 09:48:56 <[g2]> test_mimetools test_random test_urllib2 Jul 14 09:49:10 add your self to /etc/hosts Jul 14 09:49:17 maybe.. Jul 14 09:49:21 have to test Jul 14 09:49:30 can I rerun just one test? Jul 14 09:49:53 (still not finished with the first testrun, started an hour ago..) Jul 14 09:50:15 <[g2]> FatTurbo time: Jul 14 09:50:18 <[g2]> Command exited with non-zero status 2 Jul 14 09:50:19 <[g2]> real 40m 44.58s Jul 14 09:50:19 <[g2]> user 35m 46.78s Jul 14 09:50:19 <[g2]> sys 0m 54.07s Jul 14 09:50:40 <[g2]> probably around 90 minutes Jul 14 09:50:42 for make test? Jul 14 09:50:50 Im on test_xrange Jul 14 09:50:52 so.. Jul 14 09:51:15 configure: Jul 14 09:51:15 real 9m25.509s Jul 14 09:51:15 user 3m44.490s Jul 14 09:51:15 sys 3m8.780s Jul 14 09:51:15 make: Jul 14 09:51:17 real 33m16.414s Jul 14 09:51:20 user 31m9.370s Jul 14 09:51:22 sys 1m20.320s Jul 14 09:51:47 <[g2]> are you running nfs mount ? Jul 14 09:51:53 yup Jul 14 09:51:54 make test: Jul 14 09:51:54 real 59m10.773s Jul 14 09:51:54 user 36m45.660s Jul 14 09:51:54 sys 2m53.670s Jul 14 09:52:02 with nfs swap Jul 14 09:52:03 <[g2]> net I/O is probably faster Jul 14 09:52:14 <[g2]> cool. Jul 14 09:52:49 <[g2]> did you following the wiki for setting up the nfs swap ? Jul 14 09:52:53 yup Jul 14 09:52:59 <[g2]> COOL! Jul 14 09:53:01 and made my own startscript for it Jul 14 09:53:23 <[g2]> did you add it to the wiki ? Jul 14 09:53:38 hm, no, but probably should Jul 14 09:53:43 <[g2]> :) Jul 14 09:54:03 <[g2]> have you tuned nfs at all ? Jul 14 09:54:09 nope.. Jul 14 09:54:13 probably should.... Jul 14 09:54:29 <[g2]> me too, if I only knew how :) Jul 14 09:55:38 DaKa2, damn, you're doing native builds all over nfs including swap? Jul 14 09:55:45 yes Jul 14 09:55:47 nice Jul 14 09:55:52 yes :-) Jul 14 09:57:37 <[g2]> Hey I wanna see he startup script.... I'd like see if he can mount multiple slugs on the root Jul 14 09:57:49 <[g2]> s/he/his Jul 14 09:58:04 ? Jul 14 09:58:36 <[g2]> yeah, multiple slugs nfs mount the same root Jul 14 10:01:32 I guess there are a hundred better ways to do it... Jul 14 10:03:12 [g2]: updated.. but, I think I might have done some things differently Jul 14 10:04:40 <[g2]> DaKa2, THX Jul 14 10:04:57 update-rc.d is fantastic Jul 14 10:06:33 <[g2]> looks good Jul 14 10:07:29 how do I run just one test of python? Jul 14 10:08:23 <[g2]> dunno Jul 14 10:08:43 <[g2]> the tests run twice though Jul 14 10:08:57 $(TESTPYTHON) $(TESTPROG) $(TESTOPTS) Jul 14 10:08:57 <[g2]> once as a script and once compiled Jul 14 10:13:43 test_mimetools Jul 14 10:13:44 1 test OK. Jul 14 10:13:57 only two not working then Jul 14 10:14:17 test_urllib2 Jul 14 10:14:17 1 test OK. Jul 14 10:14:22 only one not working then Jul 14 10:14:42 <[g2]> DaKa2, you add the host name to /etc/hosts ? Jul 14 10:14:46 yup Jul 14 10:14:50 same problem with saned Jul 14 10:14:55 <[g2]> cool Jul 14 10:15:05 it needs the ip and host in /etc/hosts Jul 14 10:15:40 the only problem left is because of no hardware floating point Jul 14 10:16:08 turnup should add that line in /etc/hosts Jul 14 10:16:34 <[g2]> only for statics Jul 14 10:16:47 on dhcp it should add the name tolocalhost Jul 14 10:16:52 to localhost* Jul 14 10:17:07 ie, 127.0.0.1 localhost THG-SLUG1 Jul 14 10:17:25 then it still works, I think Jul 14 10:17:30 better test that.. Jul 14 10:17:48 yep Jul 14 10:17:50 works too Jul 14 10:18:29 <[g2]> what's the cmd line for a test ? Jul 14 10:18:37 ./python Lib/test/regrtest.py -l test_urllib2 Jul 14 10:19:20 ./python Lib/test/regrtest.py -l test_urllib2 test_mimetools test_random Jul 14 10:19:22 works too Jul 14 10:19:35 2 tests OK. Jul 14 10:19:35 1 test failed: Jul 14 10:19:35 test_random Jul 14 10:20:42 <[g2]> excellent! Jul 14 10:20:54 :) Jul 14 10:25:53 make install: Jul 14 10:25:53 real 5m50.955s Jul 14 10:25:53 user 3m3.540s Jul 14 10:25:53 sys 1m27.980s Jul 14 10:26:45 lets see now.. hm.. still same problem.. better try reinstalling bitbake Jul 14 10:31:25 woho! I think.. Jul 14 10:35:09 haha, guess why my do_patch didnt work... Jul 14 10:35:21 ipkg install patch :) Jul 14 10:35:31 hm.. not Jul 14 10:35:49 <[g2]> symlinks for patch util-linx Jul 14 10:36:18 ahh, good Jul 14 10:36:27 <[g2]> ipkg install bash || exit 1 Jul 14 10:36:27 <[g2]> ipkg install --force-overwrite wget || exit 1 Jul 14 10:36:27 <[g2]> ipkg install --force-overwrite tar || exit 1 Jul 14 10:36:27 <[g2]> ipkg install patch || exit 1 Jul 14 10:36:27 <[g2]> ipkg install --force-overwrite util-linux || exit 1 Jul 14 10:36:28 <[g2]> ipkg install --force-overwrite bzip2 || exit 1 Jul 14 10:36:30 <[g2]> ipkg install cvs Jul 14 10:36:44 <[g2]> ln -s armeb-linux-ranlib ranlib Jul 14 10:36:53 <[g2]> last one from /usr/bin Jul 14 10:37:33 this busybox thing should really be fixed, so it uses alternatives for everything Jul 14 10:37:57 <[g2]> exactly that's why I mentioned the distro thing Jul 14 10:38:32 <[g2]> we should size some stuff and maybe even put it on the flash Jul 14 10:38:54 <[g2]> we've gotta split stuff between core and other Jul 14 10:39:14 or just use /etc/alternatives, so you can start with busybox and replace what you want Jul 14 10:39:26 without -force-overwrite Jul 14 10:39:55 Cannot find package patch. Jul 14 10:39:56 Cannot find package util-linux. Jul 14 10:40:04 <[g2]> symlinks Jul 14 10:40:05 do I have to build them? Jul 14 10:40:08 dammit Jul 14 10:40:27 <[g2]> just go in and add them in to symlinks Jul 14 10:40:50 yup, you said that, and i missed it :-) Jul 14 10:42:09 <[g2]> so how does the update-alternatives work ? Jul 14 10:42:27 the packages installs wget.busybox and wget.real Jul 14 10:42:39 and update-alöternatives links the one with higest priority Jul 14 10:42:46 alternatives* Jul 14 10:43:25 <[g2]> so ipkg somehow understands that ? Jul 14 10:43:31 yes Jul 14 10:43:42 they run update-alternatives in postinst Jul 14 10:43:51 with some parameter Jul 14 10:44:08 well, I havn't tested on openslug Jul 14 10:44:17 <[g2]> so we need to add postinst stages ? Jul 14 10:44:20 but thats how it works in debian Jul 14 10:44:32 dont know, oe may have some magic for it Jul 14 10:44:43 like inherit update-rc.d Jul 14 10:44:48 <[g2]> Ok... we'll have to figure that out Jul 14 10:46:11 its already used for modprobe Jul 14 10:46:21 ls /usr/lib/ipkg/alternatives/ Jul 14 10:46:26 and other things Jul 14 10:47:45 pkg_postinst_ipkg () { Jul 14 10:47:45 update-alternatives --install ${bindir}/ipkg ipkg ${bindir}/ipkg-cl 100 Jul 14 10:47:45 } Jul 14 10:47:49 pkg_postrm_ipkg () { Jul 14 10:47:50 #!/bin/sh Jul 14 10:47:50 update-alternatives --remove ipkg ${bindir}/ipkg-cl Jul 14 10:47:50 } Jul 14 10:48:40 someone should probably make something that lists things conflicting Jul 14 10:48:53 <[g2]> right Jul 14 10:48:54 there is a list of all files in the packages somewhere I think Jul 14 10:49:23 since its used in ipkg I think it should work in all of OE Jul 14 10:49:37 <[g2]> good find! Jul 14 10:49:50 :) Jul 14 10:49:59 <[g2]> I still need to figure exactly the mechanism to use it Jul 14 10:50:29 yes Jul 14 10:50:51 because --force-overwrite is plain wrong Jul 14 10:51:02 * [g2] hides :) Jul 14 10:51:06 :) Jul 14 10:52:07 I dont think there should be any packages providing the same files not conflicting with eachother Jul 14 10:52:16 but thats jut me Jul 14 10:52:52 <[g2]> I wouldn't go that far but it's a guiding principle Jul 14 10:53:11 exactly Jul 14 10:53:58 Im a debian user since, I think 1.3, started using it for real with 2.0 Jul 14 10:54:22 Im a bit influenced by their policy Jul 14 10:54:38 <[g2]> then why are you running new sw ? :) Jul 14 10:54:59 Im running debian unstable... Jul 14 10:55:00 <[g2]> you should be building with 2.95 Jul 14 10:55:12 :) Jul 14 10:55:28 <[g2]> did you stay up all night ? Jul 14 10:55:54 no, got to sleep at around 3:30, got up 9:30 Jul 14 10:56:02 <[g2]> cool Jul 14 10:56:04 (19:55 now) Jul 14 10:59:30 [g2]: btw, gcc version 4.0.1 (Debian 4.0.1-2), too old? :-) Jul 14 11:00:43 woho!, building with bb works perfect Jul 14 11:01:01 ok, not Jul 14 11:01:03 <[g2]> hey they've just been getting in the swing of things lately Jul 14 11:01:05 but still, almost Jul 14 11:01:19 <[g2]> hey 95% congrats! Jul 14 11:01:27 <[g2]> what's the almost ? Jul 14 11:01:44 forgot to symlink ipkg-native Jul 14 11:02:04 <[g2]> on the target Jul 14 11:02:25 Im running with my own ow-native-symlinks Jul 14 11:02:30 oe* Jul 14 11:03:03 <[g2]> I added a INHIBIT_DEFAULT_DEPS = "1" to the lsof bb file Jul 14 11:03:33 hm.. Jul 14 11:04:01 well, it wanted ipkg-build.. but I guess ipkg install ipkg-utils would fix that Jul 14 11:04:43 if that was corssbuilt that is... Jul 14 11:05:18 <[g2]> you need that in the symlinks Jul 14 11:05:27 well, no, dont think so Jul 14 11:05:39 building it cross, and ipkg install it instead Jul 14 11:05:56 <[g2]> ok... Jul 14 11:06:03 and add it to ASSUME_PROVIDED Jul 14 11:06:06 just testing.. Jul 14 11:06:45 probably could do that with quilt too Jul 14 11:07:05 <[g2]> probably Jul 14 11:07:25 ERROR: Cannot satisfy the following dependencies for ipkg-utils: Jul 14 11:07:25 python Jul 14 11:07:28 grrr... Jul 14 11:07:55 -force-depends Jul 14 11:07:58 yep Jul 14 11:08:38 Package ipkg-utils wants to install file /usr/bin/ipkg-link Jul 14 11:08:38 But that file is already provided by package ipkg-link Jul 14 11:08:41 GRRR Jul 14 11:08:44 -force-overwrite Jul 14 11:08:54 heh Jul 14 11:09:04 -do-what-I-tell-you Jul 14 11:09:06 <[g2]> how about update-alternative :) Jul 14 11:09:19 <[g2]> or just building it natively :) Jul 14 11:09:43 hehe, if it doesnt work, maybe Jul 14 11:09:56 <[g2]> I know that works :) Jul 14 11:10:08 and soon I will know if this works Jul 14 11:10:11 :) Jul 14 11:11:13 quilt wants diffstat... Jul 14 11:11:25 I wonder what diffstat is going to make me symlink and bb... Jul 14 11:16:16 WOHO!, it works Jul 14 11:16:55 <[g2]> CONGRATS! what did you build ? Jul 14 11:16:57 rm -rf tmp/* && time bitbake vlan Jul 14 11:17:02 lets see.. Jul 14 11:17:20 real 0m 33.59s Jul 14 11:17:20 user 0m 26.37s Jul 14 11:17:20 sys 0m 4.64s Jul 14 11:17:24 33 seconds :-) Jul 14 11:17:29 from a clean tmp Jul 14 11:18:01 <[g2]> Cool! Jul 14 11:18:17 (guess why I want to ipkg install as much a possible) Jul 14 11:18:55 ipkg install tmp/deploy/ipk/vlan_1.8-r1_armeb.ipk works Jul 14 11:19:00 <[g2]> because eventually the native environ can go way Jul 14 11:19:01 and the program works Jul 14 11:19:13 <[g2]> go away Jul 14 11:19:15 :) Jul 14 11:20:57 so, time to try something that inherits autotools Jul 14 11:21:08 xinetd Jul 14 11:25:21 'bb -cclean vlan && time bitbake vlan' on my crossbuild takes real 0m22.192s Jul 14 11:26:46 dammit.. xinetd: Jul 14 11:26:49 checking build system type... Invalid configuration `armv5teb-linux': machine `armv5teb' not recognized Jul 14 11:33:04 why is oe telling it that.. stupid... Jul 14 11:35:25 <[g2]> hmm... I'm not getting that Jul 14 11:35:41 on xinetd? Jul 14 11:36:07 <[g2]> I'm trying to build perl-native Jul 14 11:36:32 if you could try xinetd when you get time, it would be great Jul 14 11:36:41 so I can see if its oe, xinetd, or me Jul 14 11:37:05 and since I made xinetd, I guess its oe or me :-) Jul 14 11:37:12 xinetd.bb* Jul 14 11:37:22 <[g2]> what do I need for xinetd ? Jul 14 11:37:31 xinetd Jul 14 11:37:40 building I dont know Jul 14 11:37:47 <[g2]> that's the only one ? Jul 14 11:38:05 probably want lots of -natives Jul 14 11:38:42 libtool-native automake-native autoconf-native gnu-config-native update-rc.d Jul 14 11:38:54 Im running with: Jul 14 11:38:55 ASSUME_PROVIDED = "libtool-native automake-native autoconf-native \ Jul 14 11:38:55 quilt-native ipkg-utils-native gnu-config-native \ Jul 14 11:38:55 virtual/armeb-linux-gcc virtual/libc update-rc.d" Jul 14 11:39:22 <[g2]> update-rc.d doesn't that pull in all basefiles ? Jul 14 11:39:34 hm.. dont think so.. ... Jul 14 11:39:44 Im using the local one.. Jul 14 11:40:10 (again, guess why Im using packages and not compile everything native) :-) Jul 14 11:47:22 <[g2]> checking build system type... Invalid configuration `armv5teb-linux': machine `armv5teb' not recognized Jul 14 11:47:29 yup... Jul 14 11:47:54 <[g2]> that could be an assumes issue no ? Jul 14 11:47:58 if you run ./configure yourself it works, I guess oe gives it armv5teb and it doesnt know Jul 14 11:48:17 could be Jul 14 11:49:22 Im just linking in everything and will try a bit more nativly Jul 14 11:50:04 otherwise its just patch the configure so it knows what armv5teb is Jul 14 11:53:01 hm.. I guess it is xinetd thats broken.. Jul 14 12:21:34 dammit.. guess I have to build perl nativly too: Jul 14 12:21:35 /usr/bin/perl: symbol lookup error: /usr/lib/libperl.so.5: undefined symbol: __eqdf2 Jul 14 12:21:39 configure: error: Perl 5.005_03 or better is required Jul 14 12:22:08 that has always been the case Jul 14 12:22:14 <[g2]> remember I said perl built but didn't run Jul 14 12:22:26 ahh, missed that.. Jul 14 12:22:30 or, forgot :-) Jul 14 12:27:49 I dunno why perl is even in openslug-packages - it only causes confusion Jul 14 12:28:04 <[g2]> we should remove it for now Jul 14 12:28:04 I guess it's a dep for some things Jul 14 12:28:09 (just thought of that) Jul 14 12:28:14 <[g2]> yeah Jul 14 12:28:33 <[g2]> I'm trying to get a real ipk built now Jul 14 12:28:41 :-) nice Jul 14 12:29:04 <[g2]> but perl needs autoreconf which needs perl :) Jul 14 12:29:25 perl needs perl to build? didn't used to Jul 14 12:29:58 <[g2]> no perl depends on autoreconf that needs perl Jul 14 12:30:14 <[g2]> well the way OE builds perl Jul 14 12:30:33 oh Jul 14 13:29:41 <[g2]> DaKa2, around ? Jul 14 13:29:45 yep Jul 14 13:30:08 <[g2]> ok i'm back to what I was doing 12 hours ago what was reflashing the slug Jul 14 13:30:26 :) Jul 14 13:30:35 testing that everything works? Jul 14 13:30:50 <[g2]> on the linksys web GUI the login cgi just sets up a session cookie or something similar with the admin/admin login ? Jul 14 13:31:14 uhm.. no idea.. used it once... Jul 14 13:31:30 but probably Jul 14 13:31:30 <[g2]> to reflash to openslug eh :) Jul 14 13:31:41 <[g2]> oh... thx Jul 14 13:32:35 it probably sets up a session with an ID, that the server knows belongs to a logged in admin user Jul 14 13:33:42 <[g2]> so if I sniffed the ID I could hi-jack the session ? Jul 14 13:33:57 <[g2]> or is it a cookie-id Jul 14 13:34:30 <[g2]> I'm thinking about the best way to setup a web intf for openslug Jul 14 13:34:50 that would be using sessions Jul 14 13:34:56 <[g2]> I'd probably just use https with some for of auth Jul 14 13:35:00 <[g2]> form Jul 14 13:35:02 saves an id as a cookie Jul 14 13:35:04 yup Jul 14 13:35:07 https Jul 14 13:35:10 php? Jul 14 13:35:14 or just cgi? Jul 14 13:35:25 <[g2]> the auth ? Jul 14 13:35:30 the thing Jul 14 13:35:35 the interface Jul 14 13:36:15 <[g2]> the things that does the auth ? Jul 14 13:36:20 <[g2]> thing Jul 14 13:36:23 the entire web interface Jul 14 13:36:33 php or something else? Jul 14 13:37:09 <[g2]> don't you need a web server in front of the php/cgi ? Jul 14 13:37:17 <[g2]> unless it's perl Jul 14 13:37:18 well, yes Jul 14 13:37:24 but thats less important Jul 14 13:37:28 <[g2]> or python Jul 14 13:37:44 you are going to do something without a web server? Jul 14 13:39:24 <[g2]> I was thinking thttpd or apache Jul 14 13:39:27 <[g2]> with PHP Jul 14 13:39:34 <[g2]> but I don't really know Jul 14 13:39:46 <[g2]> seems a little overkill but it was running Jul 14 13:39:58 ahh, thats what I wanted to know :-) Jul 14 13:41:00 I would check disk and memory rquierments for thttpd and apache and see how much differenc there is Jul 14 13:41:28 <[g2]> I think thttpd is smaller Jul 14 13:41:29 using apache would be better, but if that brings down the machine.. well, not good Jul 14 13:41:41 <[g2]> it's runs fine on 64MB Jul 14 13:41:45 <[g2]> and a HD Jul 14 13:41:47 Im kindof sure it is smaller Jul 14 13:42:03 but the question is how much Jul 14 13:42:12 <[g2]> I was surfing all around the appache docs in https and it was very quick Jul 14 13:42:24 <[g2]> I was surprised Jul 14 13:42:46 its a really nice machine this thing Jul 14 13:42:56 and apache is really good Jul 14 13:43:15 <[g2]> I was running 2.0 with PHP5 Jul 14 13:43:22 might be overkill for just a small one user interface... but still Jul 14 13:43:50 <[g2]> having all those "howtos" on the web helps Jul 14 13:44:03 <[g2]> but it is a whole lot to understand Jul 14 13:44:14 yes Jul 14 13:46:00 * [g2] has a *ton* to learn there though Jul 14 13:48:55 well, start learning :-) Jul 14 13:49:09 hm.. my turboslug just died... Jul 14 13:49:16 stone dead freeze Jul 14 13:49:18 <[g2]> died ? Jul 14 13:49:38 well, rebooting now Jul 14 13:49:43 you have serial? Jul 14 13:49:57 not here.. Jul 14 13:50:09 well.. it will probably work after a reboot Jul 14 13:50:31 I just opened it, and the cpu is hot... really hot.. Jul 14 13:50:51 ambient temp? Jul 14 13:51:04 31degC Jul 14 13:51:08 I've had it freeze on me once, when I had two stacked on top of each other and the ambient temp was >30C. Jul 14 13:51:49 oh well, Ill put a heatsink on it Jul 14 13:51:58 and drill a few more holes Jul 14 13:53:16 ah well, it reboots ok Jul 14 13:54:37 <[g2]> hey.. you're working it too hard Jul 14 13:55:23 died on perl build.. Jul 14 14:09:05 <[g2]> perl-native or perl Jul 14 14:09:16 <[g2]> in bb ? Jul 14 14:09:36 no.. perl native not bb Jul 14 14:10:01 I was wondering why it was taking so long on a single file Jul 14 14:10:36 <[g2]> how big is your swap ? Jul 14 14:10:44 128mb Jul 14 14:11:35 <[g2]> hmmm I've got 128 with 64 onboard Jul 14 14:11:54 <[g2]> I wonder if the nfs was getting taxed Jul 14 14:12:05 ? Jul 14 14:12:32 it died on the third file it tried to build Jul 14 14:12:56 <[g2]> I mean because you were swapping heavily over NFS the chip got hot Jul 14 14:13:01 <[g2]> that's odd Jul 14 14:13:36 well, its outside its casing, and we'll see how it goes Jul 14 14:13:51 <[g2]> naked slug Jul 14 14:14:17 heh Jul 14 14:14:20 :) Jul 14 14:14:22 barenekkid slug Jul 14 14:15:05 now is the time I should have had my serial cable.. Jul 14 14:15:19 it doesn't want to boot.. Jul 14 14:15:31 or well, it mounts nfs Jul 14 14:15:40 creates .recovery and does nothing else Jul 14 14:15:54 if I reboot it goes on but without nfs because of .recovery Jul 14 14:16:02 strang Jul 14 14:16:04 strange Jul 14 14:16:07 if I remove it it still doesn't work Jul 14 14:16:08 <[g2]> which kernel ? Jul 14 14:16:16 hm.. 2.6.12.2 Jul 14 14:16:34 I could have messed up some important files with all the building Jul 14 14:16:49 oh well, clean nfs, woho... Jul 14 14:17:01 Im saving this one for later **** BEGIN LOGGING AT Thu Jul 14 14:33:08 2005 Jul 14 14:39:45 <[g2]> see you guys later Jul 14 15:05:06 oh well.. should probably sleep a bit too.. bbl **** ENDING LOGGING AT Thu Jul 14 23:59:56 2005