On Fri, Jul 23, 2010 at 1:44 PM, Khem Raj <[email protected]> wrote: > On Fri, Jul 23, 2010 at 2:18 AM, Frans Meulenbroeks > <[email protected]> wrote: >> 2010/7/23 Richard Purdie <[email protected]> >> >>> On Fri, 2010-07-23 at 10:11 +0200, Koen Kooi wrote: >>> > -----BEGIN PGP SIGNED MESSAGE----- >>> > Hash: SHA1 >>> > >>> > On 23-07-10 10:02, Phil Blundell wrote: >>> > > On Fri, 2010-07-23 at 09:25 +0200, Koen Kooi wrote: >>> > >> There is a BIG problem with these patches, they break multimachine >>> builds. >>> > >> >>> > >> The previous situation had: >>> > >> >>> > >> cross/armv7a-angstrom-foo/usr/bin/ >>> > >> cross/armv5te-angstrom-foo/usr/bin/ >>> > >> etc >>> > >> >>> > >> The new situation has: >>> > >> >>> > >> x86_64-linux/usr/bin >>> > >> >>> > >> So all the toolchains get dropped into the *same* directory, which >>> > >> breaks horribly.
I have a half baked patch to get the above semantics like what poky did its not working and its getting uglier. I dont feel good about it. IMO having a new var like TARGET_SUB_ARCH point to sub-family (armv5te etc.) and later use it to form all prefixes etc would be cleaner and easier to implement and maintain. the limitation would be that you wont be able to have different version of toolchain components for two different machines based on same sub-arch e.g. (qemuarm and omap5912osk both are based on armv5te so one has to use same version of toolchain components for these kind of scenarios in a multimachine build setup. Therefore I would propose either we backout these patches for now or do single machine builds in meantime we develop the solution. -Khem _______________________________________________ Openembedded-devel mailing list [email protected] http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel
