[Libva] [PATCH 2/2] Fix error caused by 'make distclean' with wayland disabled

[Xiang, Haihao]

On Thu, 2015-12-10 at 20:31 +0000, Emil Velikov wrote:
> On 10 December 2015 at 19:58, Xiang, Haihao <haihao.xiang at intel.com>
> wrote:
> > This is based on Mesa's 426a23af147720ae3b89995ffee792a29e8ae2db
> > 
> > Signed-off-by: Xiang, Haihao <haihao.xiang at intel.com>
> > ---
> >  configure.ac                        |   7 +-
> >  va/wayland/Makefile.am              |   9 ++-
> >  va/wayland/protocol/Makefile.am     |   1 -
> >  va/wayland/protocol/wayland-drm.xml | 155 ------------------------
> > ------------
> >  va/wayland/wayland-drm.xml          | 155
> > ++++++++++++++++++++++++++++++++++++
> >  5 files changed, 167 insertions(+), 160 deletions(-)
> >  delete mode 100644 va/wayland/protocol/wayland-drm.xml
> >  create mode 100644 va/wayland/wayland-drm.xml
> > 
> > diff --git a/configure.ac b/configure.ac
> > index b54eafe..6270d35 100644
> > --- a/configure.ac
> > +++ b/configure.ac
> > @@ -273,9 +273,10 @@ if test "$enable_wayland" = "yes"; then
> >      PKG_CHECK_MODULES([WAYLAND], [wayland-client >=
> > wayland_api_version],
> >          [USE_WAYLAND="yes"], [:])
> >      if test "$USE_WAYLAND" = "yes"; then
> > -        m4_ifdef([WAYLAND_SCANNER_RULES],
> > -            [WAYLAND_SCANNER_RULES(['$(top_builddir)/va/wayland/pr
> > otocol'])],
> > -            [wayland_scanner_rules="";
> > AC_SUBST(wayland_scanner_rules)])
> > +
> > +        WAYLAND_PREFIX=`$PKG_CONFIG --variable=prefix wayland-
> > client`
> > +        AC_PATH_PROG([WAYLAND_SCANNER], [wayland-scanner],,
> > +                     [${WAYLAND_PREFIX}/bin$PATH_SEPARATOR$PATH])
> > 
> >          AC_DEFINE([HAVE_VA_WAYLAND], [1],
> >                    [Defined to 1 if VA/Wayland API is built])
> > diff --git a/va/wayland/Makefile.am b/va/wayland/Makefile.am
> > index ca6c594..7a724ae 100644
> > --- a/va/wayland/Makefile.am
> > +++ b/va/wayland/Makefile.am
> > @@ -60,7 +60,14 @@ noinst_HEADERS                       =
> > $(source_h_priv)
> > 
> >  # Wayland protocol
> >  va_wayland_drm.c: $(protocol_source_h)
> > - at wayland_scanner_rules@
> > +%-protocol.c : %.xml
> > +       $(AM_V_GEN)$(WAYLAND_SCANNER) code < $< > $@
> > +
> > +%-server-protocol.h : %.xml
> > +       $(AM_V_GEN)$(WAYLAND_SCANNER) server-header < $< > $@
> > +
> > +%-client-protocol.h : %.xml
> > +       $(AM_V_GEN)$(WAYLAND_SCANNER) client-header < $< > $@
> > 
> Hmm all three ? Afaics only the client header is needed. Speaking of
> which one shouldn't really have it in git. Can we remove it, please ?

Yes, only the client header is needed. I can remove it in a new version
of patch.


> 
> > diff --git a/va/wayland/protocol/Makefile.am
> > b/va/wayland/protocol/Makefile.am
> > index 614d8a4..e984e0f 100644
> > --- a/va/wayland/protocol/Makefile.am
> > +++ b/va/wayland/protocol/Makefile.am
> > @@ -21,7 +21,6 @@
> >  # SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
> > 
> >  EXTRA_DIST = \
> > -       wayland-drm.xml         \
> You still need to ship this file in the tarball. Although better move
> the hunk to va/wayland/Makefile.am and nuke this makefile.

wayland-drm.xml is moved to the upper directory. I will add it
in va/wayland/Makefile.am

> 
> 
> > diff --git a/va/wayland/protocol/wayland-drm.xml
> > b/va/wayland/protocol/wayland-drm.xml
> > deleted file mode 100644
> > index 265d4f8..0000000
> > --- a/va/wayland/protocol/wayland-drm.xml
> > +++ /dev/null
> 
> > diff --git a/va/wayland/wayland-drm.xml b/va/wayland/wayland-
> > drm.xml
> > new file mode 100644
> > index 0000000..265d4f8
> > --- /dev/null
> > +++ b/va/wayland/wayland-drm.xml
> 
> To avoid this unnecessary noise (and make the patch more readable)
> you
> can use -M when generating the patch.
> ie. git format-patch/send-email -M ...

Thanks for your suggestion.

> 
> Mildly related: I'm wondering if one shouldn't use "the source" of
> this protocol rather than coping it around. If I have to guess I'll
> point fingers at wayland-protocols ;-)
> 
> And let's not forget that intel-driver could use a similar patch :-P

I will check intel-driver.


> 
> -Emil