[Xcb] unit test patch, 2

[Trevor Woerner]

I'll prepare a new patch when I'm home and settled, but here's a few
quick comments.

On 9/29/05, Jamey Sharp <jamey at minilop.net> wrote:
>  AC_OUTPUT([Makefile src/Makefile xcb.pc])
> +
> +if test x$HAVE_CHECK = xtrue; then
> +       AC_OUTPUT(tests/Makefile)
> +fi
>
> I'm not sure you're allowed to have multiple AC_OUTPUT statements, or to
> enclose AC_OUTPUT in a shell if statement. Can't you just
> unconditionally output tests/Makefile? Your top-level Makefile.am seems
> to be expecting it to exist anyway.

Good point, I don't know if multiple AC_OUTPUT's are allowed, but they
work for the versions I use... actually just checking the info page:
"Call this macro once, at the end of `configure.ac'". So there it is.
I guess I'll unconditionally include the tests/Makefile output but
make sure tests/Makefile will be sane both with and without CHECK.

> -                   -o $@ c-client.xsl $<
> +                   -o $@ $(srcdir)/c-client.xsl $<
>
> Excellent catch, but please submit this sort of thing as a separate
> patch.

Whoops! Actually I had already submitted that as a patch a while back
but it hasn't been committed either. So when I did my "cvs diff -u" it
ended up in this in the new patch as well.

> Commented out stuff shouldn't go in CVS.

Right, sorry. My grand evil scheme is to eventually add optional
doxygen support and those are some of the lines I use to support it.
Muahaahah!

> Also, there's no ChangeLog entry here. I hate ChangeLogs, but we have
> one, so would you please add an appropriate entry?

Sure, no problem.

> Otherwise, this patch looks completely unobjectionable. Would you fix
> these things and re-post it?

As soon as I can.