[OE-core] [PATCH] python2/3: Remove building host path from packages

Richard Purdie richard.purdie at linuxfoundation.org
Tue Apr 11 15:44:05 UTC 2017


On Tue, 2017-04-11 at 16:03 +0800, Li Zhou wrote:
> 
> Remove building host related path from python2 and python3 packages.
> 
> Signed-off-by: Li Zhou <li.zhou at windriver.com>
> ---
>  meta/recipes-devtools/python/python3_3.5.2.bb | 8 +++++++-
>  meta/recipes-devtools/python/python_2.7.13.bb | 6 ++++++
>  2 files changed, 13 insertions(+), 1 deletion(-)
> 
> diff --git a/meta/recipes-devtools/python/python3_3.5.2.bb b/meta/recipes-devtools/python/python3_3.5.2.bb
> index 2ff7c9e..ca8dfe3 100644
> --- a/meta/recipes-devtools/python/python3_3.5.2.bb
> +++ b/meta/recipes-devtools/python/python3_3.5.2.bb
> @@ -185,9 +185,15 @@ py_package_preprocess () {
>  	install -m 0644 ${B}/Makefile.orig ${PKGD}/${libdir}/python${PYTHON_MAJMIN}/config-${PYTHON_MAJMIN}${PYTHON_ABI}/Makefile
>  	# Remove references to buildmachine paths in target Makefile and _sysconfigdata
>  	sed -i -e 's:--sysroot=${STAGING_DIR_TARGET}::g' -e s:'--with-libtool-sysroot=${STAGING_DIR_TARGET}'::g \
> +		-e 's:${STAGING_DIR}/${MACHINE}${datadir}/pkgconfig:${datadir}/pkgconfig:g' \
> +		-e 's:${STAGING_DIR_NATIVE}::g' \
> +		-e 's:${STAGING_DIR_HOST}::g' \
> +		-e 's:${WORKDIR}::g' \
> +		-e 's:${TMPDIR}::g' \
>  		${PKGD}/${libdir}/python${PYTHON_MAJMIN}/config/Makefile \
>  		${PKGD}/${libdir}/python${PYTHON_MAJMIN}/config-${PYTHON_MAJMIN}${PYTHON_ABI}/Makefile \
> -		${PKGD}/${libdir}/python${PYTHON_MAJMIN}/_sysconfigdata.py
> +		${PKGD}/${libdir}/python${PYTHON_MAJMIN}/_sysconfigdata.py \
> +		${PKGD}/${bindir}/python${PYTHON_MAJMIN}${PYTHON_ABI}-config
>   }

This breaks our builds since it regresses 
$ oe-selftest -r sstatetests.SStateTests.test_sstate_sametune_samesigs
which is due to the use of MACHINE above. It also raises the question
about why we're fixing paths which do not exist in master?

This is the second patch in 24 hours from Wind River which is not
appropriate for master, not tested against it and actively breaks
things.

In case its not clear, the
path s:${STAGING_DIR}/${MACHINE}${datadir}/pkgconfig does not exist in
master after the RSS. You're therefore trying to get us to add complex
pointless broken code.

Yes, I'm not happy at all in case that isn't clear either as this is a
total waste of my time.

Richard



More information about the Openembedded-core mailing list