**** BEGIN LOGGING AT Sat Feb 24 02:59:58 2007 Feb 24 19:14:30 andersee: hi Feb 24 19:14:37 andersee: i figured out which change is breaking grub Feb 24 19:14:40 nbd: hi Feb 24 19:14:43 nbd: oh? Feb 24 19:14:50 andersee: it's the uClibc_pthread.h change Feb 24 19:14:57 i have no clue why that happens, though... Feb 24 19:15:10 nbd: hmm, wonder what I broke Feb 24 19:15:38 if i revert that change and keep everything else, grub works Feb 24 19:18:26 nbd: reverting all of revison 17947 -- http://uclibc.org/cgi-bin/viewcvs.cgi?rev=17947&view=rev Feb 24 19:18:27 ? Feb 24 19:18:50 or just the uClibc_pthread.h change by itself? Feb 24 19:19:06 just uClibc_pthread.h by itself Feb 24 19:19:08 http://uclibc.org/cgi-bin/viewcvs.cgi/branches/uClibc_0_9_28/libc/sysdeps/linux/common/bits/uClibc_pthread.h?rev=17947&view=diff&r1=17947&r2=17946&p1=branches/uClibc_0_9_28/libc/sysdeps/linux/common/bits/uClibc_pthread.h&p2=/branches/uClibc_0_9_28/libc/sysdeps/linux/common/bits/uClibc_pthread.h Feb 24 19:19:23 i.e. the funky url above Feb 24 19:20:12 yes Feb 24 19:20:14 this change Feb 24 19:20:25 interesting Feb 24 19:20:52 I would not have expected that to cause problems Feb 24 19:20:59 me neither Feb 24 19:21:46 i did a diff between .2 and .1, threw it in patches/ and removed the changes file by file Feb 24 19:22:11 when removing the uClibc_pthread.h reverse change, segfaults started Feb 24 19:22:22 then i tested keeping only this change Feb 24 19:22:24 and it worked Feb 24 19:23:31 k Feb 24 19:24:54 any idea why this could be causing problems? Feb 24 19:25:26 my bet is on the weak_function bit Feb 24 19:25:35 yeah, that's what i thought as well Feb 24 19:25:53 anyway. i think i'll throw .2 in openwrt with just this part reverted Feb 24 19:25:58 if i don't find any other issues with it Feb 24 19:26:13 will let you know when any other bugs appear Feb 24 19:26:17 If you have a bit of time -- you could try putting the change back in, but simply remove the weak_function tags Feb 24 19:26:26 ok, will do Feb 24 19:26:46 I betcha it will work Feb 24 19:27:16 in which case, I'll do the same and push out 0.9.28.3 with that change in the near future Feb 24 19:28:26 testing... Feb 24 19:30:00 andersee: works Feb 24 19:30:49 nbd: that was a mighty fast rebuild/test cycle Feb 24 19:30:58 * andersee covets your build box Feb 24 19:31:13 i just rebuilt uclibc and grub, nothing else Feb 24 19:31:16 doesn't take that long... Feb 24 19:31:21 ahh, ok Feb 24 19:31:25 cool beans Feb 24 19:31:31 :) Feb 24 19:32:11 lemme drop that in, then do a full rebuild and test x86, ppc, etc Feb 24 19:32:21 to make sure that isn't going to blow up anything else Feb 24 19:32:22 ok Feb 24 19:40:13 nbd * r6352 /trunk/tools/firmware-utils/src/ptgen.c: fix cylinder rounding in ptgen Feb 24 19:41:15 nbd * r6353 /trunk/toolchain/uClibc/ (10 files in 2 dirs): upgrade to uClibc 0.9.28.2 Feb 24 19:41:34 nbd: if you could do a full openwrt build with this change and give it a try, that would be splendid Feb 24 19:42:22 nbd: would be nice to be certain everything is working before I push out the release Feb 24 19:44:49 ok Feb 24 19:45:57 andersee: i'll do a base system build now, and tell malbon to nuke the toolchain and staging dirs in the buildbot Feb 24 19:46:08 andersee: then we'll get some fresh builds for all our supported platforms soon Feb 24 19:46:20 nbd: excellent Feb 24 19:48:10 what i committed now is 0.9.28.2 with just weak_function removed in uClibc_pthread.h Feb 24 19:48:40 and some of our previous hacks Feb 24 19:52:02 sounds good Feb 24 19:52:23 if any of those hacks ought to go upstream, now would be a good time for it... Feb 24 19:52:32 just lemme know Feb 24 19:54:42 https://dev.openwrt.org/browser/trunk/toolchain/uClibc/patches - 150 helps with compiling on bsd, 160 fixes compile problems on ppc and 190 fixed some compile errors in user mode linux Feb 24 19:55:02 120 is just a local hack that you probably shouldn't merge Feb 24 19:57:43 andersee: credit for 150 goes to thorsten glaser from mirbsd, 160 and 190 were hacked up by me Feb 24 20:00:01 nbd: so bsd tar lacks --exclude Feb 24 20:00:02 ? Feb 24 20:00:09 yes, i think so Feb 24 20:00:24 bsd stuff is weird Feb 24 20:03:30 yep Feb 24 20:03:40 seems there is no --exclude in bsd's tar Feb 24 20:03:48 how annoying Feb 24 20:03:56 yeah Feb 24 20:05:06 I suppose we could cp everything to a tmp dir, remove the unwanted bits, then make the tarball Feb 24 20:05:10 sigh Feb 24 20:05:40 or do what we do when removing other things Feb 24 20:05:52 install them, then blow away the unwanted bits Feb 24 20:15:29 <[mbm]> tar xf file $(tar tf file | grep -v exclude) ... **** ENDING LOGGING AT Sun Feb 25 02:59:58 2007