[meta-intel] [PATCH] lms7: re-write do_unpack to fix warning

[Darren Hart]

On 4/10/15 7:47 AM, Saul Wold wrote:
> Since the primary tarball unpacks into a non-standard directory create
> a do_unpack that sets and resets  to first unpack the primary tarball
> and then unpacks the actual src tarball into the correct directory.
> 
> This is similar to lsof which has a dual tarball also.
> 
> Fixes this WARNING:
> WARNING: /srv/hdd/intel/test/tmp/work/corei7-64-poky-linux/lms7/7.1.20-r0/lms-7.1.20 ('S') doesn't exist, please set 'S' to a proper value
> 
> [YOCTO #7551]
> 
> Signed-off-by: Saul Wold <sgw at linux.intel.com>
> ---
>  common/recipes-bsp/amt/lms7_7.1.20.bb | 22 ++++++++++++++++++----
>  1 file changed, 18 insertions(+), 4 deletions(-)
> 
> diff --git a/common/recipes-bsp/amt/lms7_7.1.20.bb b/common/recipes-bsp/amt/lms7_7.1.20.bb
> index acc08c1..c2ab955 100644
> --- a/common/recipes-bsp/amt/lms7_7.1.20.bb
> +++ b/common/recipes-bsp/amt/lms7_7.1.20.bb
> @@ -12,8 +12,12 @@ SRC_URI = "http://software.intel.com/sites/default/files/m/4/e/a/9/b/37962-${BPN
>             file://atnetworktool-printf-fix.patch \
>             file://readlink-declaration.patch"
>  
> +LOCALSRC = "file://${WORKDIR}/outputdir/${BPN}-${PV}-${PV_SUB}.tar.gz"
> +
>  COMPATIBLE_HOST = '(i.86|x86_64).*-linux'
>  
> +
> +

This appears to be whitespace noise added, not a huge deal, but
something to be mindful of.

>  LIC_FILES_CHKSUM = "file://COPYING;md5=7264184cf88d9f27b719a9656255b47b"
>  
>  SRC_URI[md5sum] = "687b76e78bfdbcf567c0e842c1fe240a"
> @@ -24,12 +28,22 @@ inherit autotools update-rc.d
>  INITSCRIPT_NAME = "lms7"
>  INITSCRIPT_PARAMS = "defaults"
>  
> -do_unpack2() {
> -	cd ${WORKDIR}
> -	tar -xvzf ${WORKDIR}/outputdir/${BPN}-${PV}-${PV_SUB}.tar.gz
> +python do_unpack() {
> +    s = d.getVar('S', True)
> +    d.setVar('S', '${WORKDIR}/outputdir')
> +
> +    bb.warn ("Unpacking with S = %s" % d.getVar('S', True))

WARN is generally reserved for unexpected or potentially problematic
situations. In this case it's behaving as expected, so .info() would
appear to be a more appropriate test. Anytime I see a warn I would
expect something needs to be fixed - and this will warn twice everytime
under normal operation. Correct?

> +    bb.build.exec_func('base_do_unpack', d)
> +    
> +    # temorarily change SRC_URI for unpack
> +    src_uri = d.getVar('SRC_URI')
> +    d.setVar('SRC_URI', '${LOCALSRC}')
> +    d.setVar('S', s)
> +    bb.warn ("Unpacking2 with S = %s" % d.getVar('S', True))
> +    bb.build.exec_func('base_do_unpack', d)
> +    d.setVar('SRC_URI', src_uri)
>  }
>  
> -addtask unpack2 after do_unpack before do_patch
>  
>  do_install_append () {
>  	mv ${D}/${sbindir}/lms ${D}/${sbindir}/lms7
> 

-- 
Darren Hart
Intel Open Source Technology Center