[oe] [meta-java][PATCH 0/7] Allow build for multiple ARCH in the same builddir

[Max Krummenacher]

Hi Otavio

2015-02-17 0:13 GMT+01:00 Otavio Salvador <otavio at ossystems.com.br>:
> Hello Max,
>
> First I want to Thank you for all the hard work you've been putting in
> the meta-java lately. This is an important step on the right direction
> and fixes several things a lot the process...
Welcome.

>
> On Mon, Feb 16, 2015 at 7:55 PM, Max Krummenacher <max.oss.09 at gmail.com> wrote:
>> Building a recipe which provides packages for allarch and ${TUNE_PKGARCH}
>> fails if one builds in the same build directory for different TUNE_PKGARCHs.
>>
>> An example is rhino which fails as follows when one builds first for
>> armv7at2hf-vfp and then for armv7at2hf-vfp-neon.
>>     ERROR: The recipe rhino is trying to install files into a
>>     shared area when those files already exist. Those files and their
>>     manifest location are:
>>        /.../ipk/all/librhino-java_1.7r4-r0_all.ipk
>>     Matched in manifest-armv7at2hf-vfp-rhino-package_write_ipk
>>
>> The change in package arch is done in java-library.bbclass. Most of the
>> users of java-library.bbclass don't install any files in the ${TUNE_PKGARCH}
>> packages or install shell scripts which may also be deployed in an allarch
>> package.
>>
>> Most of the recipes including java-library.bbclass no longer build with newer
>> layer revisions.
>> This patchset first update the recipes to build again, changes java-library
>> build all packages in allarch by default and recipes which must have some
>> of their packages in ${TUNE_PKGARCH} to have all of their recipes in
>> ${TUNE_PKGARCH}.
>
> All this is perfect. I agree with all here and it does move things on
> a saner way.
Great.

>
>> Finaly some of the recipes I tried did not build before and after the proposed
>> changes. I document these in README.nonworking. I don't have the time to work
>> on these and I guess no one has been interested in years.
>
> This is the only thing I believe we can improve. We could use
> PNBLACKLIST as done in meta-oe and put a short reasoning on why it
> fails. I prefer this mostly because people don't read README often and
> will end reporting it is failing...
>
> What do you think?
I have a look into the PNBLACKLIST syntax and send a V2 of the patchset.

>
> --
> Otavio Salvador                             O.S. Systems
> http://www.ossystems.com.br        http://code.ossystems.com.br
> Mobile: +55 (53) 9981-7854            Mobile: +1 (347) 903-9750
> --
> _______________________________________________
> Openembedded-devel mailing list
> Openembedded-devel at lists.openembedded.org
> http://lists.openembedded.org/mailman/listinfo/openembedded-devel

Regards
Max