On Fri, Nov 21, 2014 at 17:21:07 -0700, Eric Blake wrote:
On 11/21/2014 01:41 PM, Laine Stump wrote:
> On 11/21/2014 12:20 PM, Jiri Denemark wrote:
>> With this change, any patch declared in libvirt.spec with Patch[0-9]* is
>> automatically applied in %prep. Unlike with the standard %patch[0-9]*,
>> patches are applied with "git am" to avoid some unexpected results.
>> However, as a result of this, all patches must be in the right format
>> for "git am" to be able to apply them; they should ideally be
generated
>> from git using "git format-patch".
>
> I've tried this out in the netcf specfile and it works well there too.
> So a functional ACK from me. Someone else may have comments on the finer
> details.
>
> (The one thing I wonder about is this - once this patch is in, git will
> be required for a build even if there are no patches beyond the original
> tarball. I don't know if this concerns anyone or not...)
Doesn't bother me :) Anyone liable to be developing rpms can be assumed
to know about git by now.
>> +if [ $COUNT -gt 0 ]; then
>> + xargs git am <$PATCHLIST || exit 1
>
> This is the first I noticed it doesn't use git am -3 (which I always
> use), but since that option would require the upstream git history to be
> available, I guess that is pointless here anyway :-)
Also, _not_ using -3 forces the patch application to be a bit stricter
(nothing to 3-way fuzz against), so it is actually desirable as proof
that the patches were generated correctly.
I'm not seeing any problems in the patch either, and it's had some
runtime testing in downstream RHEL builds already, so ACK.
Pushed, thanks.
Jirka