**** BEGIN LOGGING AT Thu Oct 18 03:00:02 2012 Oct 18 07:59:15 morning all Oct 18 08:02:09 hi silvio Oct 18 08:11:55 hi woglinde Oct 18 08:46:44 morning all Oct 18 08:47:54 morning Oct 18 08:49:52 hi pb Oct 18 08:49:55 hi jama Oct 18 08:50:31 hm for bitbake would be cool to have a speration btw. parsing threads and execution threads Oct 18 08:50:39 woglinde: hey Oct 18 08:50:55 I might parse with 4 threads but buidling with only 2 Oct 18 08:50:57 he zecke Oct 18 08:57:26 hi bluelightning Oct 18 08:57:57 woglinde: I think that you can do that already Oct 18 08:58:19 woglinde: yep, BB_NUMBER_PARSE_THREADS Oct 18 09:08:11 oh fine Oct 18 09:08:19 thanks for the hint Oct 18 09:27:02 morning all Oct 18 09:35:32 hi pb Oct 18 09:57:02 hi pb Oct 18 10:15:06 hm deadlock in jamvm-initial when tried to compile jamvm-native, but second try worked Oct 18 10:27:04 does sstate-cache warrant fully reproductible builds? Oct 18 10:30:27 hrw: what do you mean by warrant? Oct 18 10:31:22 bluelightning: that build done from scratch == build done partially from sstate-cache (when it comes to results) Oct 18 10:32:01 yes, that should be the case, as long as none of the recipes you build touch the sysroot directly (which none ever should) Oct 18 10:32:29 cool Oct 18 10:33:02 time for coffee Oct 18 12:54:57 hrw: would you mind resending this with meta-oe prefix so it gets picked up? http://patchwork.openembedded.org/patch/30449/ Oct 18 12:57:32 sure, nearly forgot about that one Oct 18 12:58:21 I hit that issue recently, good that you've already got a patch for it :) Oct 18 12:58:31 * bluelightning is clearing out patchwork atm Oct 18 12:59:03 bluelightning: oe-core patchwork too? :D Oct 18 13:00:27 hrw: just the OE-Classic one for now Oct 18 13:00:54 I was trying to get a handle on any meta-networking patches still pending and noticed what a mess it was in :( Oct 18 13:03:54 sent Oct 18 13:08:34 thanks Oct 18 13:15:19 hmm. I changed gcc SRC_URI, bumped PR and fresh build got eglibc from sstate-cache instead of rebuild Oct 18 13:15:43 do_populate_lic_setscene (4906): log.do_populate_lic_setscene.4906 Oct 18 13:15:47 do_package_setscene (4904): log.do_package_setscene.4904 Oct 18 13:15:51 do_populate_sysroot_setscene (6735): log.do_populate_sysroot_setscene.6735 Oct 18 13:17:12 but maybe I am wrong Oct 18 13:25:14 nope Oct 18 13:25:27 eglibc was taken from sstate-cache while db was rebuilt Oct 18 13:25:49 hrw: had you previously done a build with that version of gcc? Oct 18 13:26:18 bluelightning: PV-PR? do not think so. Oct 18 13:26:27 hmm, that doesn't sound right then Oct 18 13:27:15 bumped PR just in case Oct 18 13:27:40 to r6.1 which I normally do not use Oct 18 13:28:13 0: eglibc-2.16-r18+svnr20393 do_populate_sysroot_setscene (pid 28613) Oct 18 13:28:14 1: eglibc-2.16-r18+svnr20393 do_package_setscene (pid 28614) Oct 18 13:28:25 3: eglibc-initial-2.16-r18+svnr20393 do_package_setscene (pid 28921) Oct 18 13:29:11 looks like repeat Oct 18 13:30:28 and this should not happen ;( Oct 18 13:30:53 my gcc update has many ICE related fixes so rebuild of everything is important Oct 18 13:31:26 gcc-linaro PV=4.7 PR=r1325426.1 and eglibc from sstate-cache as well Oct 18 13:31:31 oe-core ML? Oct 18 13:36:29 mail sent Oct 18 13:38:25 hrw: I usually see those setscene tasks run, but then real build performed later, isnt it the case for you too? Oct 18 13:38:59 JaMa: nope Oct 18 13:39:22 JaMa: eglibc was just package/sysroot unpacked and then db (which uses eglibc) got fully rebuilt Oct 18 13:43:07 anyway I keep that sstate-cache dir just in case Oct 18 13:52:50 bluelightning: I just resent the patch with signed off by but it looks a bit mangled Oct 18 13:53:11 bluelightning: what is the best way to rebase and add a message below the --? Oct 18 13:53:55 jackmitchell: not sure but I think you have to add those by hand afterwards Oct 18 13:54:14 jackmitchell: in this case it's fine to just send it without those comments since I know what's going on :) Oct 18 13:54:33 bluelightning: Ok, have a look and see if the patch is ok Oct 18 13:55:39 jackmitchell: I wasn't going to worry too much about it but since you ask it would be great if you could fix the indenting as well (both for EXTRA_OECMAKE and do_install_append) Oct 18 13:56:12 I did build-test the v2 patch and it built fine Oct 18 13:56:30 bluelightning: ok, I can see the issue with the _append, but how shoudl EOCMAKE be ? Oct 18 13:57:34 bluelightning: do we have 4 space tabs in _append? Oct 18 13:57:45 jackmitchell: er, ignore that one, I failed to factor in that my email client uses variable-width fonts :) Oct 18 14:02:49 jackmitchell: for shell functions I don't mind if it's four spaces or tabs as long as it's consisent, although strictly speaking it's supposed to be tabs for shell functions Oct 18 14:02:52 bluelightning: ah man, this is infuriating, I wish git send-email had an easy way to preview your email before sending.. Oct 18 14:03:05 jackmitchell: it does, --confirm=always Oct 18 14:03:52 jackmitchel use format-patch with coverletter Oct 18 14:08:22 bluelightning: ok, how does that last one look? Oct 18 14:08:31 bluelightning: v5 :( Oct 18 14:10:08 jackmitchell: looks good at a glance Oct 18 15:05:07 hm. /etc/hosts does not contain entry for /etc/hostname Oct 18 15:06:09 that'd be a problem, hostname -s won't return anything useful, eh? Oct 18 15:06:26 mysql then complain at start: Oct 18 15:06:39 WARNING: The host 'genericarmv8' could not be looked up with resolveip. This probably means that your libc libraries are not 100 % compatible with this binary MySQL version. The MySQL daemon, mysqld, should work normally with the exception that host name resolving will not work. This means that you should use IP addresses instead of hostnames when specifying MySQL privileges ! Oct 18 15:07:31 root@genericarmv8:~# hostname -s Oct 18 15:07:31 hostname: genericarmv8: Host name lookup failure Oct 18 15:07:50 /etc/hosts is in base-files, /etc/hostname is in net-base Oct 18 15:09:20 hrw: could handle it in a postinst maybe Oct 18 15:10:27 bluelightning: sed -i -e "s/localhost$/localhost `cat /etc/hostname`/g" /etc/hosts kind of? Oct 18 15:10:52 yeah, something like that I guess Oct 18 15:10:56 or just: echo "127.0.0.1 `cat /etc/hostname`" Oct 18 15:11:28 it has to not break for upgrades though of course Oct 18 15:12:35 or just: echo "127.0.0.1 `cat /etc/hostname`" >>/etc/hosts - kind of ugly ;( Oct 18 15:15:38 that wouldn't work on upgrade though Oct 18 15:15:59 also would need to be ${D}${sysconfdir}/hostname I think Oct 18 15:17:21 and ${D}${sysconfdir}/hosts# Oct 18 15:17:25 er s/#// Oct 18 15:28:34 sure Oct 18 15:41:15 morning Oct 18 15:56:35 exit Oct 18 16:20:06 ok, how do I get sftp-server into my image? I have IMAGE_FEATURES += "openssh-sftp-server", but it's still missing from the resulting image Oct 18 16:20:54 hollisb: that's a package name, so you want to add it to IMAGE_INSTALL Oct 18 16:21:45 $#@! Oct 18 16:22:34 I wasted many hours on that last night. :( thanks Oct 18 16:24:45 hollisb: I think for 1.4 we will add a check to ensure that people don't put invalid things in IMAGE_FEATURES Oct 18 16:24:56 that would be awesome Oct 18 16:53:31 bye Oct 18 16:55:21 hrw: cya Oct 18 18:08:31 do we have a recipe for sshfs anywhere? Oct 18 18:51:15 JaMa, exactly my thought opn the ATUOREV comment :) Oct 18 18:53:35 good :) Oct 18 19:05:01 where is this meta-networking people have been talking about? Oct 18 19:06:01 Crofton: in meta-openembedded repo Oct 18 19:07:00 dont' think there's much in that one yet, but it's at least a place to put such things Oct 18 19:07:03 heh Oct 18 19:07:04 hmm Oct 18 19:10:03 ok, it seems like there was a lot of talk for something with not many recipes Oct 18 19:10:11 trying to figure out where sshfs should live Oct 18 19:13:18 meta-network Oct 18 19:14:36 depends on fuse though, whish is meta-oe Oct 18 19:16:16 hm right Oct 18 19:16:55 doesn't seem unreasonable that meta-networking might require meta-oe, the meta-oe layers are all intertwined anyway Oct 18 19:17:07 try doing a bitbake world with meta-oe but without meta-efl, etc Oct 18 19:17:11 blows up Oct 18 19:17:13 heh Oct 18 19:18:41 not good Oct 18 19:18:53 looks like jama have to clean it up Oct 18 19:19:30 yeah, I am just trying to think things through Oct 18 19:23:10 kergoth: do you remember what depends on meta-efl in meta-oe? Oct 18 19:23:30 not offhand, will take a closer look. was trying to get a world build going Oct 18 19:26:09 JaMa: mcnavi needs ecore Oct 18 19:26:51 kergoth: ok I'll move mcnavi to meta-efl then Oct 18 19:27:03 thx for report Oct 18 19:27:26 np Oct 18 19:31:07 kergoth: btw what do you think about world-image? Oct 18 19:31:30 kergoth: to test not only build of world, but also installability in do_rootfs? Oct 18 19:36:46 hmm, external-sourcery-toolchain mjust be msising something from eglibc, a world build wants to build both Oct 18 19:36:49 but i have no idea why Oct 18 19:36:51 * kergoth rolls eyes Oct 18 19:37:04 hehehe Oct 18 19:50:38 kergoth bitbake -DDD ? Oct 18 19:50:48 should tell you why to build boths Oct 18 22:53:56 pb_ if the patch you just sent is for meta-oe, I suspect patchwork will miss it Oct 18 22:54:17 nope Oct 18 22:54:22 I am wrong, it is there Oct 18 22:55:49 is there a video transcoder readily available? the ffmpeg recipe doesn't seem to end up packaging the ffmpeg binary Oct 18 22:56:11 actually wait, that recipe is a yocto thing. sorry :) **** ENDING LOGGING AT Fri Oct 19 03:00:01 2012