[MPlayer-dev-eng] [PATCH] modify X11 lib path search for mixed 32/64 bit case

[Ivan Kalvachev]

On 6/1/05, Reimar Döffinger <Reimar.Doeffinger at stud.uni-karlsruhe.de> wrote:
> Hi,
> next try...
> 
> On Wed, Jun 01, 2005 at 02:50:25PM +0300, Ivan Kalvachev wrote:
> > 1. If user gives x11libdir and the autodetection fails then the we
> > should keep the given by user dir. (it have the power of force, or it
> > would have the power of force if x11=yes)
> 
> Huh? When a a list of possible X11 directories is specified,
> autodetection is done on these, and if it fails it will use the last
> one.
> 
> > I guess that adding _x11_paths and keeping  _ld_x11 would do the trick, but
> > actually handling _x11=no case could be better:
> > +      break
> > +    fi
> > +  done
> > +else
> > +  I=$_x11_paths
> 
> Why setting I here? It is not used at all with --disable-x11, and even
> in the other cases only for the status message...

Oopsss, my bad.
The problem I am thinking of, is when --enable-x11 is used to force
x11 and skip autodetection.
as we have _x11=yes, we will try to find the X11 dir. So we will have
autodetection of directory.  If it cannot find directory it will
disable x11 build, so in effect we overide the force.

So I guess that if developer forces detection then we should use the
library path without checking it (it could be the reason detection to
fail)

Let me try again

@@somewhere
 echocheck "X11"
 if test "$_x11" != no ; then
+  if test "$_x11" = yes && test -n "$_x11_paths" ; then
+       $I=_x11_paths 
+  else
@@do_detection here...
+  fi
 fi
 if test "$_x11" = yes ; then
   _def_x11='#define HAVE_X11 1'
@@end

The rest is OK.

   Ivan Kalvachev
  iive