DDRAW example bltfast arguments question

Hi,

I'm a newbie in asm & windows-programming and I'm analyzing the ddraw example to learn something about it.

But I discovered something I don't understand:

Code:

comcall DDSBack,BltFast,0,0,[DDSPicture],rect,DDBLTFAST_SRCCOLORKEY

This generates:

Code:

...
PUSH_imm8 0x1
PUSH_imm32 0x4498cb
PUSH RM 0x4020c8
PUSH_imm8 0x0
PUSH_imm8 0x0
MOV EAX, [0x4020c4]
PUSH EAX 
MOV EAX, [ EAX ]
...

before calling the bltfast...

In a later execution

Code:

comcall DDSBack,BltFast,288,200[DDSPicture],rect,DDBLTFAST_SRCCOLORKEY

which generates:

Code:

PUSH_imm8 0x1
PUSH_imm32 0x4498cb
PUSH RM 0x4020c8
PUSH_imm32 0xc8
PUSH_imm32 0x120
MOV EAX, [0x4020c4]
PUSH EAX

The API itself is defined as:

Code:

HRESULT BltFast(
  DWORD dwX,                            
  DWORD dwY,                            
  LPDIRECTDRAWSURFACE7 lpDDSrcSurface,  
  LPRECT lpSrcRect,                     
  DWORD dwTrans                         
);

So... my question is:
Why is it ok to give PUSH imm8 to the dwX & dwY in the first call and later PUSH imm32 ?
Why doesn't fasm generate imm32 pushes in the first place ?
How does windows decide to use the unspecified(?) "imm8" version of bltfast instead of the "imm32" version ?
Where did I miss something ? (I discovered that the only relevant differences could be in the EBX & EDX registers but I can't figure out what)

Could someone of you Assembler-Gods or Demi-Gods help me ?

Thanks,
Sephiroth

PS: I'm using the fasmw151 example of DDRAW but discovered the same thing in the latest release too.