**** BEGIN LOGGING AT Thu Dec 17 02:59:58 2015 Dec 17 07:45:21 good morning Dec 17 07:53:05 howdy mckoan Dec 17 08:33:40 hi Dec 17 12:39:39 Has anyone an advise on how i can initialize a plain apt sources.list with packaged installed on the embedded target. The goal is to use apt-get install git-core on debian repositories? Anyone? Dec 17 13:13:41 huh? you want a poky-generated image to use apt and pull from the debian repos? Dec 17 13:24:04 yes some people want to do this for comfort ... Dec 17 13:48:43 sounds like a very, very bad idea to me. Dec 17 13:49:45 frodo_yocto: if you want something that can go this route, i'd suggest to have a look at emdebian or linutronix embe rather. Dec 17 13:51:13 you'll have to replicate the exact linking method that debian has used, the versions, and the library packaging splits Dec 17 13:51:28 at which point you've just replicated debian Dec 17 13:51:29 in short, the complete ABI Dec 17 13:51:40 which interestingly is what meta-debian is all about Dec 17 13:51:44 its *absolutely mental* Dec 17 13:51:47 including all and every filesystem speciality. Dec 17 13:52:08 rburton: i'd suggest s/mental/metal/ and commence headbanging. Dec 17 13:57:13 Thanx for the infos !! Dec 17 14:33:11 would it be possible to move the initramfs .manifest generated during the build of a generated OS's rootfs which requires core-image-minimal? Dec 17 14:40:20 what do you mean by move? Dec 17 14:40:55 i guess you could write a task that runs after do_deploy before do_build to copy the file from deploydir to somewhere else Dec 17 14:42:30 I'll give it a shot, thanks Dec 17 17:45:48 * paulg_ idly wonders if anyone has brewed up a meta/conf/machine/include/tune-pa6t.inc for PA-Semi before... Dec 17 18:08:10 paulg_: check with montavista layers Dec 17 18:08:18 if there is one available in open Dec 17 18:12:16 khem`, reasonable idea ; will do a bit of googling later. Dec 17 18:13:38 if not, it doesn't look too hard to make one from scratch Dec 17 18:14:34 toying with the idea of reviving my PA-Semi as a testing platform for ppc64 and want a semi modern rootfs Dec 17 18:19:36 paulg_: yes thats good. Dec 17 18:19:49 ppc64 support should work well with oe-core Dec 17 18:20:00 there are other ppc64 SOCs Dec 17 18:20:06 which work Dec 17 18:21:26 not for any specific reason ; just one of those " 'cause I can" projects. :-) Dec 17 18:22:47 built for the old 7400 based sbc8641d and that worked well, but it is borderline ancient and only 1/4 the RAM of the PA-Semi IIRC. Dec 17 19:48:00 I'm a bit of a noob to Yocto and I'm stuck on trying to compile some vendor provided libraries to install on the image root. When I run oe_runmake it is not reading a preinst command in the Makefile which creates a build subdirectory for the libraries. When I run make from the devshell it compiles just fine. Does any know why the behavior is different using oe_runmake vs manually running make from the devshell, and perhaps provide some hints on Dec 17 19:48:01 how to debug? Dec 17 19:57:15 Hi, I would like to know if it is possible in Yocto 2.0 branch to override a machine configuration file that is in another layer. Dec 17 19:58:33 In particular, I need to make a small change in the meta-fsl-arm layer to ls1021atwr.conf in order to change a u-boot setting. Dec 17 19:59:05 I am trying to avoid making changes to any other layer besides my own Dec 17 20:27:37 rburton: next set ( after gcc 5.3 is merged ) I am preparing should have core-image-base, core-image-sato core-image-weston building and booting with musl Dec 17 20:35:54 nice Dec 17 20:58:39 what's the relationship between meta-freescale and meta-fsl-{arm,ppc} at this time? Dec 17 20:58:42 khem: nice Dec 17 21:14:28 kergoth: ppc and arm are two different groups AFAIK Dec 17 21:14:32 in fsl Dec 17 21:14:51 someone them even did not know that OE is already used in FSL in another grp Dec 17 21:15:00 atleast thats what my experience is from prior job Dec 17 21:15:06 as their customer Dec 17 21:15:43 Trusted git merge too much and a conflict got in accidentally Dec 17 21:15:46 hmm Dec 17 21:16:18 kergoth: one surprise I am seeign with musl is that the builds are lot faster for whole load Dec 17 21:16:34 somewhere between 15-20% Dec 17 21:17:12 I think the simple header include architecture of musl makes cpp faster compared to convoluted glibc inc mechanism Dec 17 21:30:20 khem: wow, that's a nice side benefit Dec 17 21:30:22 had no idea Dec 17 21:45:37 oh that's nice Dec 18 00:36:22 denix: with a fido core-image-minimal, on the BB X15 I have a lot of error about HDMI, pvrsrvinit fails (but it's a rev A1a, not sure if it's still supported) and on the OMAP5432 uEVM I hang after "Starting kernel", would you like an email on the ML ? Dec 18 00:37:51 Anarky: yes, please Dec 18 02:28:36 kergoth: I think in about years time most of software would build on musl Dec 18 02:29:09 for embedded use its a serious replacement for glibc **** ENDING LOGGING AT Fri Dec 18 02:59:59 2015