Problems with Tao-Framework and Cobra

[johannes]

here are the atachments good luck fixing the problems or telling me what I did wrong..

[Charles]

Bugs squashed. New code added to the regression test suite.

Note that instead of "1.0 to float", you can use "1.0f". Also, if you say "640.0f / 480.0f" then the expression will be float which eliminates another cast.

Here is the code:

def main is shared
        evt as Sdl.SDL_Event
        Sdl.sdl_Init(Sdl.sdl_INIT_EVERYTHING)
        surfacePtr = Sdl.sdl_SetVideoMode(640,480,32,Sdl.sdl_HWSURFACE|Sdl.sdl_DOUBLEBUF|Sdl.sdl_ANYFORMAT|Sdl.sdl_OPENGL)
        
        h = 640.0f / 480.0f
        Gl.glViewport(0, 0, 640, 480)
        Gl.glMatrixMode(Gl.gl_PROJECTION)
        Gl.glLoadIdentity
        Gl.glFrustum(-1.0f, 1.0f, -h, h, 5.0f, 60.0f)
        Gl.glMatrixMode(Gl.gl_MODELVIEW)
        Gl.glLoadIdentity
        Gl.glTranslatef(0.0 to Single, 0.0 to Single, -40.0 to Single)
        while true
            Sdl.sdl_PollEvent(out evt)
            branch evt.type
                on Sdl.sdl_QUIT
                    break

The SDL API is strange in that the members of classes use the prefixes like "GL" and "SDL_" that I'm sure were needed in C, but aren't needed in C# or Cobra. It might be interesting to write a Cobra program that reads the DLLs, scans the members and spews out a new wrapper class with better names.

Future plans for Cobra that you might be interested in:

• Allow the "f" suffix even without the fractional part. In other words, recognize "0f" as well as "0.0f"
• Add float32 and float64 types. "float" becomes an alias to "float64". Literals can have suffixes like "-49.5f32".
• Provide a "-number:float64" option that changes the default fractional type between float32, float64 and decimal.

[Charles]

I've implemented the first two bullets above. So you can refer to types 'float32' and 'float64'. The type 'float' is equivalent to 'float64'.

You can now suffix literals with 'f', 'f32' and 'f64'. You can also use underscores before those suffixes. Examples would be:

1f
5.0_f32
10f32