**** BEGIN LOGGING AT Wed Apr 22 02:59:58 2015 Apr 22 03:25:01 kergoth: there ? Apr 22 03:25:26 kergoth: Can you think of a case where _append += can have unexpected results ? Apr 22 03:30:00 nope, the += is immediate, so it's no different than manually specifying the separator Apr 22 03:32:12 I mean if I use FOO_append_arm += "xx" Apr 22 03:32:50 and FOO_append_arm = " xx" Apr 22 03:42:00 i know Apr 22 03:42:24 no difference in behavior Apr 22 07:21:49 hey koen you on? Apr 22 07:25:12 yes Apr 22 07:25:47 just curious around the UBOOT_ARCH, does overriding it work? as in UBOOT_ARCH_ = "foo" Apr 22 07:26:44 * koen tries Apr 22 07:32:25 aargh Apr 22 07:32:34 * koen removes overlayed kernel-arch.bbclass Apr 22 07:38:06 koen: i think changing it to a ?= for the future is probably worth it. But i imagine getting that change backported would be a pain Apr 22 07:40:48 UBOOT_ARCH_aarch64 = "arm" in local.conf also seems to work Apr 22 07:41:08 I never realized that would take precedence over python methods Apr 22 07:41:30 well its not the python method that is setting it Apr 22 07:42:04 ah, right Apr 22 07:42:14 it's set using a method, not by the method Apr 22 08:06:16 morning all Apr 22 08:08:30 mornin Apr 22 08:18:43 hi Jin^eLD Apr 22 08:48:50 morning Apr 22 08:48:58 hi mago Apr 22 08:49:00 hi bl Apr 22 09:01:28 hi mago_, woglinde Apr 22 09:05:06 is there some way to add a string (like a localversion) to the name of the generated images? Apr 22 09:07:23 possibly like LOCALSTRING="mybuild" bitbake core-image-minimal would result in core-image-minimal-mybuild.ext3, for example Apr 22 09:08:10 LetoThe2nd: you could adjust IMAGE_BASENAME (the image file is named with IMAGE_NAME which by default is "${IMAGE_BASENAME}-${MACHINE}-${DATETIME}" Apr 22 09:09:30 LetoThe2nd: if you want LOCALSTRING to feed through into the environment you'd additionally need to add it to BB_ENV_EXTRAWHITE Apr 22 09:10:26 bluelightning: so if i do something like IMAGE_NAME = "${IMAGE_BASENAME}-${MACHINE}-${DATETIME}-${LOCALSTRING}" and add LOCALSTRING to BB_ENV_EXTRAWHITE it would probably do the trick? Apr 22 09:16:29 LetoThe2nd: if you want your local version to appear in the symlink then you'd need to modify IMAGE_BASENAME rather than IMAGE_NAME (or alternatively modify both IMAGE_LINK_NAME and IMAGE_NAME) Apr 22 09:21:38 bluelightning: ok, will look into it. thanks a lot! Apr 22 09:22:02 np Apr 22 11:20:55 bluelightning: if i set the nostamp flag on a function, shouldn't bitbake always re-run it? even though its in sstate cache? Apr 22 11:31:01 woglinde: your rootfs POSTPROCESS thing that adds build info into the filesystem, how do you make it re-run every time the manifest changes? i guess you'd have to make it always run, but how does that work when sstate-cache kicks in? Apr 22 11:32:54 (maybe changes to manifest file is a bad example, cause it will most likely trigger rebuilds of the rootfs in other ways. a better example is, how do you force bitbake to re-run your postprocess command for every build, so that it updates your UUID with the new build date?) Apr 22 11:41:58 mago_ ? it runs for each new image build Apr 22 11:42:33 mago_ our use case is to verify, which image is flashed, what we build some time ago and delcared as stable image Apr 22 15:17:59 is there a variable to filter out recipes for bitbake to avoid parsing? Apr 22 15:18:40 cbrake1: BBMASK Apr 22 15:18:56 bluelightning: thanks! Apr 22 17:09:46 I am trying to replace the SRC_URI from a recipe (tar.gz) with my copy (from git)....is there a better way to do this other than overriding SRC_URI completely? (there are a bunch of init scrips on SRC_URI that I would like to use too) Apr 22 17:16:19 adelcast1: which version of the build system are you using? Apr 22 17:17:12 dizzy Apr 22 17:18:58 adelcast1: ok, in dizzy and previous versions there isn't a better way, but in master / imminent 1.8 we have a new mechanism that does pretty much exactly this Apr 22 17:20:04 ah, great!, I can backport or wait (I think we are likely going to move to fido in a couple of months) Apr 22 17:20:22 any commits I should look at? Apr 22 17:20:52 well, it's actually quite a few commits Apr 22 17:21:44 I see, so I might be better of overriding for now, then switching once we get to fido Apr 22 17:22:01 which works for me Apr 22 17:23:19 mm, now I am adelcast1, I guess that's because I was logged in on a another computer as adelcast, hehe Apr 22 17:23:35 bluelightning: thanks for the insight Apr 22 17:29:30 no worries Apr 22 17:30:54 FWIW here's the section of the manual that talks about the new mechanism: http://www.yoctoproject.org/docs/latest/dev-manual/dev-manual.html#using-devtool-in-your-workflow Apr 22 17:31:30 i'll have to play with that, haven't had time to try it at all yet Apr 22 17:32:33 DOes it work without Poky? Apr 22 17:32:44 Crofton|work: absolutely, no poky dependencies at all Apr 22 17:33:02 anytime I see step 1 is clone the poky git repo, my stomach turns Apr 22 17:33:19 Crofton|work: yes, I should probably ask Scott to work that out of there Apr 22 17:36:16 I'd be less touchy except for finding a bug that made a poky like assumption and then being called an oe-classic user Apr 22 17:46:18 I need to play with it, but sounds like EXTERNALSRC would do the trick for me Apr 22 17:52:42 adelcast1: right, externalsrc.bbclass should; but in dizzy and earlier you don't get those extra files you mentioned Apr 22 17:53:06 adelcast1: however you should be able to just copy in the new externalsrc.bbclass from fido Apr 22 17:58:24 bluelightning: I see, for now I am going to override SRC_URI, but will add a task to move to externalsrc for that recipe, as part of our upgrade to fido work Apr 22 17:58:49 adelcast1: ok, great Apr 22 18:01:05 bbl Apr 22 21:45:40 i've configured a build with lots of parallel tasks, most are in do_fetch, and are failing, but my network appears fine Apr 22 21:45:59 is there some sort of way to up the number of allowed connections or something? Apr 22 22:08:17 yocto doesn't impose artificial connection restrictions. what are the error messages? Apr 22 22:11:40 kergoth: connection timed out in headers Apr 22 22:12:38 ooh, and i got a segmentation fault for "/usr/bin/env wget ..." Apr 22 22:13:27 i'll just keep running "bitbake -k -c fetchall" over and over until it's done :-) Apr 22 22:14:31 if that's happening under yocto, it'd almost certainly fail elsewhere as well, best to investigate hte sanityo f your network and tools on the build machine, as it's just running those **** ENDING LOGGING AT Thu Apr 23 02:59:58 2015