Re: [PATCH v3 1/1] acl: enable ptest support

[Richard Purdie <richard.purdie@linuxfoundation.org>]

On Sun, 2014-01-26 at 15:50 +0800, Chong Lu wrote:
> Install acl test suite and run it as ptest.
> nfs test cases need depend on nfs service. So exclude them order to
> make ptest all pass.
> 
> Signed-off-by: Chong Lu <Chong.Lu@windriver.com>
> ---
>  meta/recipes-support/attr/acl.inc       | 14 +++++++++++++-
>  meta/recipes-support/attr/acl/run-ptest |  7 +++++++
>  2 files changed, 20 insertions(+), 1 deletion(-)
>  create mode 100644 meta/recipes-support/attr/acl/run-ptest
> 
> diff --git a/meta/recipes-support/attr/acl.inc b/meta/recipes-support/attr/acl.inc
> index a461232..04967a6 100644
> --- a/meta/recipes-support/attr/acl.inc
> +++ b/meta/recipes-support/attr/acl.inc
> @@ -9,7 +9,9 @@ LIC_FILES_CHKSUM = "file://doc/COPYING;md5=c781d70ed2b4d48995b790403217a249 \
>                      file://doc/COPYING.LGPL;md5=9e9a206917f8af112da634ce3ab41764"
>  
>  DEPENDS = "attr"
> -SRC_URI = "http://download.savannah.gnu.org/releases/acl/${BP}.src.tar.gz"
> +SRC_URI = "http://download.savannah.gnu.org/releases/acl/${BP}.src.tar.gz \
> +           file://run-ptest \
> +"
>  
>  require ea-acl.inc
>  
> @@ -24,3 +26,13 @@ do_install_append() {
>  	sed -i ${D}${libdir}/libacl.la -e \
>  	    s,^libdir=\'${base_libdir}\'$,libdir=\'${libdir}\',
>  }
> +
> +inherit ptest
> +
> +do_install_ptest() {
> +	tar -cf - test/ --exclude nfs | ( cd ${D}${PTEST_PATH} && tar -xf - )
> +	mkdir ${D}${PTEST_PATH}/include
> +	cp ${S}/include/builddefs ${S}/include/buildmacros ${S}/include/buildrules ${D}${PTEST_PATH}/include/
> +}
> +
> +RDEPENDS_${PN}-ptest = "bash coreutils perl perl-module-filehandle perl-module-getopt-std perl-module-posix shadow"

This set of dependencies is proving problematic. Firstly it is going to
increase build times as perl gets pulled in. coreutils also pulls in gmp
which doesn't appear to build on x32.

This patch is therefore why we keep seeing MUT builds of nightly-x32
fail.

http://autobuilder.yoctoproject.org/main/builders/nightly-x32/builds/9/steps/BuildImages/logs/stdio

No doubt we need to fix this, however it does also show how convoluted
our dependency chains are and how a small change like this, raises other
issues elsewhere.

I opened a bug about the gmp issue:

https://bugzilla.yoctoproject.org/show_bug.cgi?id=5783

Cheers,

Richard