Re: 1.3.10 release candidates

[Owen Taylor <otaylor redhat com>]

Hans Breuer <hans breuer org> writes:

> At 17:53 26.10.01 +0200, Sven Neumann wrote:
> >
> > [...]
> >
> >pango builds fine, so I guess it counts as a warning. Is Win32 really
> >using @ in function names? /me shudders. Perhaps there's a different
> >function name that could be used for this test instead?!
> >
> These are autogenerated names which give some link time safety
> for the calling convention __stdcall (Callee clears the stack)
> @n 'n' is the stack size in bytes. If the caller and the callee
> don't aggree on the stack size real bad things will happen ...
> 
> I guess because auto*(what?) needs to look into the lib files
> and M$ has:
> 
> #define WINAPI __stdcall
> 
> there isn't much to do against (beside not using auto* :-)

I believe autoconf-2.5x can handle the @ here fine.

For now, I've gone with the following patch.

Regards,
                                        Owen

Index: configure.in
===================================================================
RCS file: /cvs/gnome/pango/configure.in,v
retrieving revision 1.80
diff -u -p -r1.80 configure.in
--- configure.in	2001/10/25 18:56:05	1.80
+++ configure.in	2001/10/26 16:45:21
@@ -153,7 +153,12 @@ AM_CONDITIONAL(HAVE_XFT, $have_xft) 
 have_win32=false
 WIN32_LIBS=""
 WIN32_CFLAGS=""
-AC_CHECK_LIB(gdi32, GetTextMetricsA 8, have_win32=true, : )
+# The following doesn't work with autoconf-2.13, so we check $host instead
+# AC_CHECK_LIB(gdi32, GetTextMetricsA 8, have_win32=true, : )
+case "$host" in
+  *-*-mingw*|*-*-cygwin*) have_win32=true ;;
+esac
+
 
 if test $have_win32 = true; then
   WIN32_LIBS="-lgdi32"