[DK22]

I saw that "fst" and "fstp" instructions in .exe used in this way (example):

CODE

fstp    [esp+0C8h+var_B8]

Would this right?

CODE

float fCoordX;
float flt_0125 = 0.125;
...
__asm
{
    mov ax, wordCoordX
    movzx eax, ax
    mov fCoordX, eax
    fild fCoordX
    fmul ds:flt_0125
    fstp fCoordX
}

In other words, I have 2-bytes coord, stored in wordCoordX. I need to transform it to 4-bytes float.

This post has been edited by DK22: Nov 13 2011, 09:08 AM

[Silent]

Yeah, that looks proper, but why not just typecast it in C++? Typecasting should produce similar results.

CODE

float fCoordX = (float)wordCoordX;

This post has been edited by Silent: Nov 13 2011, 11:57 AM

[DK22]

I tried this. But then I had lags...

I insert my code to CAutomobile__preRender, when I used this

CODE

float fCoordX = (float)wordCoordX;

then, all automobiles was without wheels (o_O) Idk why...

Also I wanted to ask about CGeneral__GetAtanOfXY...
What is the 1st param and second? X Y or Y X ?
Cause on Cpp it is

CODE

Angle = atan2(coordY, coordX)

And dont ask why I dont use cpp's function... I have problems with compiling...

This post has been edited by DK22: Nov 13 2011, 12:45 PM

[Deji]

Well, the easiest way to test atan is to do "atan2(0.0, 1.0)" and see what angle it produces, if the result is 90 degrees then the it is atan2(coordy, coordx)...

[Wesser]

You can avoid the first line by overwriting the second one with this:

CODE

movzx   eax, wordCoordX

FPU atan computes st(1) / st(0), so I guess it is atan2(x, y), unlike ANSI built-in function.