Re: [Mesa3d-dev] Bug in z-depth

[Allen A. <ak...@po...>]

On Thu, Oct 05, 2000 at 02:36:30PM -0700, Michael Vance wrote:
| On Thu, Oct 05, 2000 at 11:43:10AM -0600, Brian Paul wrote:
| 
| > glXChooseVisual() really is poorly designed.  Serious apps should
| > avoid it and use a better visual selection method.
| 
| Such as?

Since 1994 I've been using the visinfo package that I wrote while at
SGI.  Among other things, it's used in isfast and glean, and it had
some influence on Mark Kilgard's design for the display initialization
string in GLUT.

With visinfo, you write short Boolean expressions involving attribute
variables and C operators that say exactly what visual attributes you
want, plus sort keys that determine which visual is returned first if
more than one meets the selection criteria.  That way you don't run
into the ambiguities that cause problems for glXChooseVisual(). 
Examples:

	"max rgb, max z"

		Deepest color buffer, deepest depth buffer -- roughly
		what glXChooseVisual is supposed to return if you
		specify nonzero color and depth attributes.

	"max rgb, z == 16, db"

		Deepest color buffer; depth buffer must be exactly
		16 bits; double-buffered.

	"max rgb, min z"

		Deepest color buffer; shallowest available nonzero
		depth buffer.

	"max rgba, fast, conformant, accumrgba >= 2*rgba"

		Deepest RGBA that's hardware accelerated, passes
		the OpenGL conformance tests, and has an accumulation
		buffer at least twice as deep as the color buffer.

	"r%2 || g%2 || b%2, !s"

		Color buffer in which at least one channel has an
		odd number of bits; no stencil planes.

And so on.

This has enough expressive power that I've never felt the need for
anything else, and it's been ported to most systems running OpenGL so
it has relatively good availability.  However, you can always roll
your own using glXGetConfig() or other window-system-dependent query
function..

Allen