**** BEGIN LOGGING AT Mon Jun 01 02:59:59 2015 Jun 01 08:11:51 good morning Jun 01 08:13:54 howdy Jun 01 08:14:48 hi mckoan and pompomjuice Jun 01 08:43:59 morning all Jun 01 08:44:30 hi bl Jun 01 08:45:52 hi woglinde Jun 01 09:28:49 morning Jun 01 09:29:34 is it just me or the perl-module-cgi (defined in meta/recipes-devtools/perl/perl-rdepends_5.20.0.inc) deps are broken ? Jun 01 09:29:59 use if $] >= 5.019, 'deprecate'; Jun 01 09:30:25 and of course, it doesnt work at runtime as it cant find the deprecate module Jun 01 09:32:49 bencoh: I guess there's a missing dependency; perl-module-deprecate should exist, so at least as a workaround you should be able to install that Jun 01 09:33:11 yeah, it's defined in the .inc as well Jun 01 09:35:00 I've just upgraded my OE layers from may-2014 to may-2015, so pretty much everything broke ... just wanted to make sure this one wasnt my fault ;) Jun 01 13:29:09 I would like to be the first to welcome our Intel Friends back to the ARM business. Jun 01 13:29:59 Crofton|work: and their first move would be squeeze atom core instead of arm one? Jun 01 13:30:16 hrw hm as FPGA? Jun 01 13:30:19 dunno, they have to be careful not to alienate existing customers Jun 01 13:30:45 woglinde: as part of cpu+fpga combo? Jun 01 13:31:02 Crofton|work: migration plans etc... Jun 01 13:31:23 * Crofton|work uses Xilinx, so we are OK Jun 01 13:32:14 ;D Jun 01 13:32:31 * hrw moved to atom for small devices Jun 01 13:32:55 hm.. even my android tablet is atom... Jun 01 13:34:00 what is good is that we will be able to reduce the number of layers :) Jun 01 13:34:08 avago/brcm Jun 01 13:34:11 nxf/fsl Jun 01 13:34:16 nxp/fsl Jun 01 13:34:22 and intel/altera Jun 01 13:34:29 hmm Jun 01 13:34:50 abelloni: since when amount of layers is an issue? if they are properly done then they should not conflict Jun 01 13:34:58 just BBLAYERS gets longer Jun 01 13:35:21 I can just imagine comments submiting ARM support to meta-intel :) Jun 01 13:36:57 :)) Jun 01 13:38:28 'oh, no... intel is doing arm again...' Jun 01 13:38:48 'strongarm, xscale and now what?' Jun 01 13:39:05 'hey, intel: where is sa1100 support for 4.0 kernel?' Jun 01 13:40:55 cortex-a53 in stratix 10? :) Jun 01 13:44:03 cortex-a72, now with a floppy disk controller attached Jun 01 15:12:08 I'm looking into how bitbake determines which tasks need to be (re)run. According to the documentation, a checksum for each task is generated based on its "inputs," but I haven't found any additional details on how this is done. Can anyone point me to additional documentation or the source code that pertains to this? Jun 01 15:23:36 metadata checksumming is handled in bitbake, see lib/bb/siggen.py, lib/bb/codeparser.py. for checksumming of file:// file content, see lib/bb/checksum.py Jun 01 15:23:55 to see how bitbake uses the checksums to implement shared state / binary caching / task acceleration, see sstate.bbclass and staging.bbclass Jun 01 15:24:07 or rather, how bitbake+oe uses Jun 01 15:26:50 great, thanks! Jun 01 15:27:36 additional variable dependencies can be explicitly specified with the 'vardeps' flag, for cases where bitbake cannot automatically determine them, due to e.g. dynamically constructed variable names in python code Jun 01 15:27:45 deps can be excluded via vardepsexclude flag Jun 01 15:30:24 http://www.yoctoproject.org/docs/current/ref-manual/ref-manual.html#shared-state-cache Jun 01 15:31:03 nice Jun 01 15:31:11 didn't realize it was covered that well Jun 01 15:31:51 hmm, the debugging section could use some improvement, i think. no mention of bitbake-whatchanged, richard's sigwalker, or bitbake -S printdiff, or whatnot Jun 01 15:35:23 That section in the manual is what prompted my question. I'm interested in how exactly oe/bb generates the checksums/signatures for tasks. I've been looking in lib/bb so the files mentioned are helpful Jun 01 15:35:51 ah Jun 01 17:29:22 This is weird. Building from oe_classic, m4-native fails to build, something to do with gets. So I do a .bbappend and a patch (mostly taking cues from http://arago-project.org/pipermail/meta-arago/2013-January/000671.html ). But now when i bitbake, it's telling me do_install id failing because.... "make: *** No rule to make target `install'. Stop." Weirder still, when I bitbake -c devshell so I can look at the Makefile, all the Jun 01 17:29:22 directories except temp are empty. Jun 01 17:37:02 accidentally use = instead of += on SRC_URI or something? Jun 01 17:37:03 hmm Jun 01 17:37:12 will double check Jun 01 17:37:34 FILESEXTRAPATHS_prepend := "${THISDIR}/${PN}" Jun 01 17:37:35 PR_append ="-r1" Jun 01 17:37:35 SRC_URI += " file://remove-gets.patch " Jun 01 22:17:57 I've made an .mk file, but the install target commands seem to be ignored. Anybody know why? http://pastebin.com/uZBJfmrF Jun 01 22:25:09 KobeBryant: a .mk file is for buildroot, this channel is for OpenEmbedded Jun 02 02:18:33 darknighte / Crofton - who do i ask to get a yocto/oe speaker at open source conference in honduras? **** ENDING LOGGING AT Tue Jun 02 02:59:58 2015