Well, if there is literally not a difference in a byte by byte comparison of the two compiled applications, then I have no idea. If you don't mind comparing the original version to a version with the decimal points, you could try the unix tool diff. For example if you navigate to a folder where two versions of vendetta client are stored, you could try the following:

diff vendettaclient.app/Contents/MacOS/vendettaclient vendettaclient2.app/Contents/MacOS/vendettaclient

If they are the same you will get no output, but if the compiler did something different and the files aren't identical, it will say "Binary files vendettaclient and vendettaclient differ".

However, I still think it's worth trying out the decimal point version on an intel mac to see if the problem is resolved.

Well, I guess I don't have much else to contribute. If you do plan on entering glOrtho manually, the following should do that for you:

GLfloat ortho[16]={2.,0.,0.,0.,0.,-2.,0.,0.,0.,0.,-1.,0.,-1.,1.,0.,1.};
glMultMatrixf(ortho);

Good luck.

Heh, thanks for the numbers. I'll probably just do that.

Ok, can someone check to see that glow is now working on macintel machines?

It's not. :(

Using the OpenGL trace, here is the point where you load the matrix:
glPushMatrix();
glLoadMatrixf({2, 0, 0, 0, 0, -2, 0, 0, 0, 0, -1, 0, -1, 1, 0, 1});
glMatrixMode(GL_MODELVIEW);

Looks great, and therefore looks like it wasn't the problem...

[Edit]
Removed big GL trace.

That all makes sense to me.

What does the glow effect look like now?

The glow effect looks the same as before. I'll give you a detailed description in case no one has yet:

When the glow does not intersect the bottom edge of the screen, you see nothing. When the glow does intersect the bottom edge of the screen, the two vertices that are below the screen are placed properly. However, the two vertices that are on the screen are locked in the upper left corner. The only time this changes is when the glow stradles the lower right corner of the screen. In that case the two lower vertices are still in the right spot, but the upper left vertex is at the upper left corner of the screen, and the upper right vertex is in the upper right corner of the screen. For that one particular example it results in the glow covering half of the screen (upper right triangle).

Is there anything in particular that you would want me to look for in the OpenGL trace? A question I have is which command defines where the glow (which I assume is a quad) is drawn on the screen?

same as before? Hmm. that doesn't make sense, unless the vertex shader is just not working correctly on that platform.

The glow is actually rendered to the screen right after 4 vertex program parameters are set to 0,0,0,0. The primitive type is a strip. It uses the same vertex program as the bluring passes use.