**** BEGIN LOGGING AT Tue Sep 21 02:59:57 2010 Sep 21 03:06:18 where is the file that makes this file "backfire/staging_dir/host/bin/autoreconf" Sep 21 03:06:53 :/ that file breaks building when moving the topdir to a new location Sep 21 03:07:28 because that file has direct locations Sep 21 07:08:26 hello Sep 21 07:42:02 ping xMff Sep 21 07:57:49 build #97 of ifxmips is complete: Exception [exception failed slave lost shell_10 compile_12] Build details are at http://tksite.gotdns.org:8010/builders/ifxmips/builds/97 Sep 21 07:57:50 build #94 of adm5120 is complete: Exception [exception failed slave lost shell_10 compile_12] Build details are at http://tksite.gotdns.org:8010/builders/adm5120/builds/94 Sep 21 07:57:52 build #91 of kirkwood is complete: Exception [exception failed slave lost shell_10 compile_12] Build details are at http://tksite.gotdns.org:8010/builders/kirkwood/builds/91 Sep 21 09:10:46 are there plans for backfire/package/busybox to be upgraded to 1.17.2 (stable) ? Sep 21 11:51:25 nbd * r23107 /branches/backfire/package/ (53 files in 3 dirs): mac80211: merge version 2010-09-14 from trunk Sep 21 14:03:44 xMff: http://openwrt.pastebin.ca/1945808 works for me. I have tested with extroot in overlay and in image. The kernel modules can be in image or modules. block-mount still needs to be modified to be able to load from overlay though Sep 21 14:04:11 xMff: let me know your comments, please Sep 21 14:04:51 ah btw, the line 74 is leftover debugging code Sep 21 14:05:40 ah, yeah, I thought it was useful for know what was happening when I tested it...have to remember to remove it Sep 21 14:06:59 the er_load_modules can bwe simplified I suppose Sep 21 14:07:52 xMff: I'm not sure how to get the correct order on the modules any simpler Sep 21 14:08:20 xMff: it's got to interface to numbered lists Sep 21 14:08:25 *interleave Sep 21 14:08:30 *two Sep 21 14:11:32 http://openwrt.pastebin.ca/1945816 - less the debug print Sep 21 14:12:13 doh! that's wrong Sep 21 14:13:51 http://openwrt.pastebin.ca/1945821 should do it Sep 21 14:16:45 getting block-mount so it can be a module will require adding something like pi_include to /etc/functions, and also, a function like pi_include, but for all the shell scripts in a directory (basically test if the directory exists and include any shell scripts in it) Sep 21 14:18:03 just let pi_include do a for x in $1; do ...; done if [ -d $1 ] Sep 21 14:18:41 * [ -d $1 ] || [ -d /tmp/overlay/$1 ] Sep 21 14:20:37 hmmm....ok, so there's only the one function, just does the include of a directory if it's directory instead of file that's passed in Sep 21 14:21:11 as for the interleaving of modulesi n er_load_modules Sep 21 14:21:16 you could do the ln -sf trick Sep 21 14:21:28 doh!, right Sep 21 14:22:08 ln -sf /etc/modules.d/ /tmp/overlay/etc/modules.d/ /tmp/something/ Sep 21 14:23:14 ln -sf /etc/modules.d/* /tmp/overlay/etc/modules.d/* /tmp/something/ Sep 21 14:24:18 what did you use? maybe we should use a hidden dir in tmp, and subdirs of the same dir? (in the interest of cleanliness)? Sep 21 14:24:36 /tmp/preinit-hook-merge Sep 21 14:27:52 xMff: hmmm...what do you think of /tmp/.preinit-merge/{hook,extroot_modules}? Sep 21 14:28:00 sure Sep 21 14:30:01 or do you think it really matters? Sep 21 14:30:47 no, because its removed right after Sep 21 14:35:30 http://openwrt.pastebin.ca/1945831 ? Sep 21 14:44:02 and with the pi_include change: http://openwrt.pastebin.ca/1945839 Sep 21 14:44:30 looks okay Sep 21 14:45:21 there seems to be a stray "fi" in l159 Sep 21 14:46:15 I see it Sep 21 14:47:26 http://openwrt.pastebin.ca/1945840, and need to change block-mount Sep 21 14:51:55 this one looks almost golden Sep 21 14:53:08 http://openwrt.pastebin.ca/1945842 ; then I just need to find time to test Sep 21 14:53:28 line 85-92 looks like a dupe of pi_include Sep 21 14:53:56 one's in /etc/functions the other is in /lib/functions/boot.sh Sep 21 14:53:56 111-118 in latest Sep 21 14:54:16 hmm Sep 21 14:54:17 preinit doesn't use /etc/function.sh Sep 21 14:54:32 then /etc/functionsh should not need it Sep 21 14:54:49 xMff: mount.sh is used outside of preinit Sep 21 14:55:13 yeah but ouside of üpreinit there should be no need for /tmp/overlay/ access Sep 21 14:55:16 hmmm....don't need the /tmp/overlay stuff for that one though Sep 21 14:55:17 right Sep 21 14:56:53 http://openwrt.pastebin.ca/1945846 ? Sep 21 14:57:10 yes Sep 21 14:57:43 wait Sep 21 14:57:50 "done" in l223 loks wrong Sep 21 14:58:08 xMff: oops, you're right Sep 21 14:58:15 the cat above too Sep 21 14:59:03 I'd say "sed 's/^\([^#].*\)/insmod \.\/\1.ko/' /tmp/extroot_modules/modules.d/* | sh 2>&- || :" Sep 21 14:59:34 or better "sed -e 's/^\([^#].*\)/insmod \.\/\1.ko/' /tmp/extroot_modules/modules.d/* | sh 2>&- || :" Sep 21 15:01:23 http://openwrt.pastebin.ca/1945849 Sep 21 15:01:45 yes Sep 21 15:03:44 ok, I have to go now and work after, but so if you have a chance to test today let me know, otherwise I'll probably do it tomorrow. Got a deadline coming up. Sep 21 15:03:57 kk, ty so far Sep 21 18:21:32 nbd: my x86 image compiled with gcc-4.4.1+cs+graphite seems to boot fine Sep 21 18:21:41 cool Sep 21 18:22:06 going to sysupgrade my ipv6 router in datacenter with it now Sep 21 18:22:15 and see if I have any issues there Sep 21 18:22:45 unfortunately I forgot my 2nd rspro at work again, so unable to test on ar71xx yet Sep 21 22:07:25 heil * r23108 /packages/net/haproxy/ (3 files in 2 dirs): Sep 21 22:07:25 [packages] haproxy: re-enable polling on listener() when memory shortage occours after accept Sep 21 22:07:25 - thanks to Willy Tarreau **** ENDING LOGGING AT Wed Sep 22 02:59:57 2010