[PATCH util-macros] XORG_GIT_MODULE_VERSION: writes module version in xorg-git-version.h

[Peter Hutterer]

On Thu, Apr 22, 2010 at 01:32:24PM -0400, Gaetan Nadon wrote:
> On Wed, 2010-04-21 at 06:20 -0700, Dan Nicholson wrote:
> 
> > On Tue, Apr 20, 2010 at 11:06 PM, Peter Hutterer
> > <peter.hutterer at who-t.net> wrote:
> > >
> > > would also be good checking through Matthias' script from radeonhd for any
> > > obvious use-cases that this patch doesn't cover. It certainly seems more
> > > complex.
> > 
> > Another way to handle this would be distributing a script from the
> > macro. So, instead of substituting a huge make command, the macro
> > would install a script which you'd just have to distribute from the
> > Makefile. Then it would be a lot easier to have a complex script. Just
> > a scratch implementation:
> > 
> 
> Some afterthoughts. Unless there is big demand for such a functionality,
> the server should just copy the radeonhd script and be done with it.

++ for not rewriting code that's already tested.

Cheers,
  Peter

> This does not close the door to a project wide solution should the
> demand be there. On the other hand, if we push for a project wide
> solution without a good user base, it will be hard to change while
> maintaining backward compatibility. I know copying is not the best reuse
> method, but in small dose it is acceptable. Consider it a "trial"
> version.
> 
> 
> > AC_DEFUN([XORG_GIT_MODULE_VERSION],
> > [# use m4 so it only happens when you run autoconf
> > m4_syscmd([install -m755 /usr/share/xorg-macros/xorg-git-version.sh "$srcdir"])
> > AC_SUBST([xorg_git_version_sh], ['$(top_srcdir)/xorg-git-version.sh'])
> > ])
> > 
> > module's configure.ac:
> > XORG_GIT_MODULE_VERSION
> > 
> > module's Makefile.am:
> > EXTRA_DIST = $(xorg_git_version_sh)
> > xorg-git-version.h:
> >     $(SHELL) $(xorg_git_version_sh) $@
> > DISTCLEANFILES = xorg-git-version.h
> > .PHONY: xorg-git-version.h
> > 
> > And then you could put any fanciness you want in the script. The only
> > minor worry I'd have is that the script wouldn't have all the make
> > variables available to it like $(top_srcdir) since automake
> > intentionally unexports the variables.
> > 
> > --
> > Dan