Alexandre Julliard wrote:
>
> I'd prefer if we could abstract parts of the current checks and put
> that into aclocal. For instance many tests do something like
>
> AC_CHECK_HEADER(xx.h)
> if ac_cv_header_xx = yes
> then
> AC_CHECK_LIB(xx)
> fi
Actually, now that I look at it, I'm a little confused by this usage in
the first place. If the test is that simple, wouldn't it be better to
use the form:
And in the case of an if/then/else/fi:
AC_CHECK_HEADERS(xx.h, AC_CHECK_LIB(xx),
[ echo "Failed..."
exit 1 ]
)
Is there some reason this was avoided? I noticed a few of the header
tests only fire off if _all_ headers are present; this presents a bit of
a wrinkle. However, three of the four tests that do this have
X11/Xlib.h as the first of two header files. These three tests could be
nested inside a AC_CHECK_HEADERS(X11/Xlib.h, ...) call. The fourth
case, for the OpenGL stuff, is more complex, and may have to be left as
is.
I'll throw a patch together.... (c:
John
--
[EMAIL PROTECTED] http://www.gnome.org
[EMAIL PROTECTED] http://www.worldforge.org
