[oe] [meta-filesystems][PATCH 1/1] aufs-util: add new recipe

[Paul Eggleton]

On Saturday 18 January 2014 12:29:40 Koen Kooi wrote:
> Paul Eggleton schreef op 17-01-14 18:36:
> /recipes-utils/aufs-util/aufs-util_git.bb
> b/meta-filesystems/recipes-utils/aufs-util/aufs-util_git.bb
> 
> > new file mode 100644 index 0000000..cea3636 --- /dev/null +++
> > b/meta-filesystems/recipes-utils/aufs-util/aufs-util_git.bb @@ -0,0 +1,48
> > @@ +SUMMARY = "Tools for managing AUFS mounts" +SECTION = "base"
> > +HOMEPAGE = "http://http://aufs.sourceforge.net/"

Heh, noone noticed this is broken, will fix.

> > +SRCREV = "f29056fe396d56fc2a06a96312feabaebbe14c59" 
> > +SRC_URI =
> > "git://git.code.sf.net/p/aufs/aufs-util;protocol=git;branch=aufs3.x-rcN \
> > +          file://aufs-util-don-t-strip-executables.patch \ 
> > +          file://include/linux/aufs_type.h \ 
> > +          file://aufs-util-add-tool-concept-to-Makefile-for-cross-com.patch 
> > \ 
> > +"
> > +
> > +PV = "3.x-rcN+git${SRCPV}"
> 
> This is a new low for upstream versioning :/ I've looked at the git repo for
> aufs-util and it seems 3.x-rcN is identical to 3.9, could we use that for
> the aufs-utils recipe to avoid using that horrid, horrid PV above?

Right, this is upstream's fault. I did just look, and you're correct, although 
they aren't the same hash, the 3.9 and 3.x-rcN branches aren't materially 
different, so we could use 3.9 instead.

> > + +S = "${WORKDIR}/git/" + +do_configure_append () { +	if [ "${PN}" !=
> > "aufs-util-native" ]; then +		for i in ver c2sh c2tmac; do +	cp
> > ${STAGING_BINDIR_NATIVE}/aufs-util-${PV}/$i ./ +		done +	fi +}
> 
> Change to do_configure_append_class-target ?

Indeed, that would be cleaner.

> And you'll probably want a
> 
> RRECOMMENDS_${PN} = "kernel-module-aufs"
> 
> here to drag in the aufs modules.

I guess so...

Cheers,
Paul

-- 

Paul Eggleton
Intel Open Source Technology Centre