On Thu, Feb 06, 2014 at 04:30:52PM -0700, Eric Blake wrote:
On 02/04/2014 06:51 AM, Daniel P. Berrange wrote:
> The maint branches will often have out of date copyright headers
> so we must skip the 'sc_copyright_check' rule there.
Is that the only rule, or are there others? But yeah, that's definitely
the most annoying one.
>
> Signed-off-by: Daniel P. Berrange <berrange(a)redhat.com>
> ---
> cfg.mk | 11 +++++++++++
> 1 file changed, 11 insertions(+)
>
> Perhaps there's a better way to detect -maint branches than the
> logic I tried.
>
> diff --git a/cfg.mk b/cfg.mk
> index 207dfeb..47c7798 100644
> --- a/cfg.mk
> +++ b/cfg.mk
> @@ -88,6 +88,17 @@ else
> distdir: sc_vulnerable_makefile_CVE-2012-3386.z
> endif
>
> +GIT_BRANCH := $(shell if test -d $(srcdir)/.git; then \
> + cd $(srcdir) && \
> + (git status -s -b | grep '\#\#' | \
> + sed -e 's/.*-maint/-maint/'); fi)
I'm still thinking about this one...
I think the point is that we need to go to the root directory of the
repository if we're in a submodule and this made me realize my
root-searching one-liner in my rebuild script doesn't work for
submodules, although it would be nicer to use it:
TOPLEVEL=$(git rev-parse --show-toplevel 2>/dev/null || echo '.')
Martin
> +
> +# We fully expect -maint branches to have out of date
> +# copyright dates, so we must skip that check
> +ifeq ($(GIT_BRANCH),-maint)
> + local-checks-to-skip += sc_copyright_check
> +endif
But this part looks fine, once the GIT_BRANCH rule is optimized.
> +
> # Files that should never cause syntax check failures.
> VC_LIST_ALWAYS_EXCLUDE_REGEX = \
> (^(HACKING|docs/(news\.html\.in|.*\.patch))|\.po)$$
>
--
Eric Blake eblake redhat com +1-919-301-3266
Libvirt virtualization library
http://libvirt.org
--
libvir-list mailing list
libvir-list(a)redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list