**** BEGIN LOGGING AT Mon Apr 21 02:59:58 2014 Apr 21 10:40:33 khem: hi, patchwork is currently failing with Exception Value: (1040, 'Too many connections') Apr 21 11:16:11 anyone want to comment on this before I submit it? Apr 21 11:16:13 https://github.com/balister/meta-oe/blob/zeroc-ice-work/meta-oe/recipes-connectivity/zeroc-ice/zeroc-ice_3.5.1.bb Apr 21 11:16:22 yeah it is scary and likely has order issues Apr 21 11:22:21 FILESEXTRAPATHS_prepend := "${THISDIR}/zeroc-ice-3.5.1:" Apr 21 11:22:48 isn't BPN-PV in the default already? Apr 21 11:22:56 not sure Apr 21 11:23:07 the recipe evolved some :) Apr 21 11:23:13 from an inc and a native Apr 21 11:23:29 I can check if it builds wo Apr 21 11:23:39 I need to wack the icedatadir line also Apr 21 11:24:37 believe it or not, you can cross compile gnuradio from an sdk with ice support now Apr 21 11:39:10 how does the order of the elements look? Apr 21 11:41:20 looks OK to me Apr 21 11:41:36 run it through oe-stylize and see if there are huge differences Apr 21 11:41:52 keep in mind that oe-stylize is a guideline, it gets it very wrong sometimes Apr 21 11:42:24 I was sort of figuring oe-stylize might have issues Apr 21 11:42:29 I'll try it Apr 21 13:19:57 anyone know what image fs type to use for a uramdisk.image.gz file? Apr 21 13:21:36 Crofton|work: cpio should work Apr 21 13:21:44 didn't Apr 21 13:25:43 ERROR: cannot map 'allarch' to a linux kernel architecture Apr 21 13:26:57 I do not suggest adding image_types_uboot to user classes Apr 21 13:28:13 cpio.gz didn't trying cpio Apr 21 13:28:20 u-boot with CONFIG_RAW_INITRD and make cpio.gz from OE Apr 21 13:55:50 hmm, why isn't that set Apr 21 13:59:00 it doesnt have to be, just saves having to ubootize the image Apr 21 14:03:54 I am confused why I cna't uboot-ize it Apr 21 14:06:15 I cant actually remember if I have ever done that Apr 21 14:06:42 make: *** No rule to make target `/usr/local/oecore-x86_64/sysroots/x86_64-oesdk-linux/usr/lib/armv7a-vfp-neon-oe-linux-gnueabi/gcc/arm-oe-linux-gnueabi/4.8.1/include/stdbool.h', needed by `include/autoconf.mk'. Stop. Apr 21 14:06:47 funny sdk issue Apr 21 14:07:19 fixed Apr 21 14:07:32 \o/ Apr 21 14:07:52 grr Apr 21 14:08:04 arm-oe-linux-gnueabi-ld.bfd -Wl,-O1 -Wl,--hash-style=gnu -Wl,--as-needed -r -o built-in.o api.o api_display.o api_net.o api_platform-arm.o api_storage.o Apr 21 14:08:04 arm-oe-linux-gnueabi-ld.bfd: unrecognized option '-Wl,-O1' Apr 21 14:08:10 why did I delete old sdk Apr 21 14:30:31 excessive optimism, Crofton|work? Apr 21 14:31:04 well, do not source the sdk env file and build u-boot Apr 21 14:31:15 jsut setting path and cross_compile aorks Apr 21 14:32:26 I'm still stuck on an SDK for the NSLU2 that dates back to OE classic... not modern or shiny, but it always works! Apr 21 14:39:32 Crofton|work: OE LDFLAGS are meant only for gcc/g++ as linked, not ld Apr 21 14:39:44 if you want to use ld, sed out the '-Wl,' bit Apr 21 14:40:01 koen, something in th esdk env file caused the troble Apr 21 14:40:56 I've been seeing that type of error a lot on IRC lately Apr 21 14:41:17 and every time I see it I point out the LDFLAGS thing Apr 21 14:41:42 could always pass in CCLD as LD :) Apr 21 14:41:43 * kergoth yawns Apr 21 14:42:08 kergoth: indeed Apr 21 14:57:50 kergoth, I'm trying to make a ramdisk for u-boot Apr 21 14:57:52 and failing Apr 21 14:58:00 what is the modern way of doing this Apr 21 14:58:14 I see this image-types_uboot class Apr 21 14:58:28 but if I add to to SURE_CLASSES I get weird erros Apr 21 14:59:32 USEr_CLASSES :) Apr 21 15:06:13 ERROR: cannot map 'allarch' to a linux kernel architecture Apr 21 15:06:24 which was reported arond 2010 Apr 21 15:09:07 so that recipe isn't allarch, right? Apr 21 15:09:56 the build works unless I incldue the image_tyes_uboot class Apr 21 15:10:44 ah Apr 21 15:10:45 so Apr 21 15:10:49 trying to build an image and ubootize it Apr 21 15:11:02 modern uboot have an option that doesn't force you to mkimage the ramdisk Apr 21 15:11:21 zimage support should trigger that Apr 21 15:11:49 ask tartarus or rcn_ee, they know the name of that option Apr 21 17:14:24 khem: ka6sox: wiki is also down due to mysql connections limit Apr 21 18:19:52 maybe someone has some experience with this.. Apr 21 18:20:14 we need to modify the target of an update-alternative from /usr/sbin to /sbin, for something we're trying to run before /usr is mounted.. Apr 21 18:20:45 but we get an error from the post-installs crip tthat the link is already setup to point to /usr/sbin.. is there a way to tell update-alternative to -change- the link target for a given name? or do we need to come up with a new alternative name or? Apr 21 18:22:52 could always use —remove + —install, no? any gien altenrative name has any number of possible paths. if one of those paths is no longer valid, you should remove it, i would think Apr 21 18:22:53 * kergoth shrugs Apr 21 18:23:16 thats the part I'm confused by (and stillt rying to figure out) Apr 21 18:23:40 is there an existing entry that is still there, preventing this.. or is something retained in the update-laternative "database" saying these are the only valid link targets Apr 21 18:24:01 we're trying to update via a package feed.. so it's causing issues.. Apr 21 18:27:18 ...that and I think there is a bug in the update-alternatives.bbclass.. if one of a list of alternatives fails to appy.. the post installs cript finishes based on only the last itme.. Apr 21 18:27:23 we might want to enable set -e :P Apr 21 18:29:58 ya.. on the removal of the link (if it's the last one), it does seem to clear the entry from the DB Apr 21 19:38:44 Hi everyone, I have a skeleton for rootfs. It has 15 folders and many files and sub-folder. What is the best way to populate the rootfs? Apr 21 22:01:16 JaMa: whats the error output you see Apr 21 22:27:38 I have two svn repos for a single recipe can SRC_URI handle it ? Apr 22 00:19:48 Hi, every time I build a kernel using bitbake, the older kernel Image files are deleted from the deploy dir. Is there any way of avoiding this? **** ENDING LOGGING AT Tue Apr 22 02:59:58 2014