Subj : Re: Any help on why?
To   : borland.public.cpp.borlandcpp
From : Bob Gonder
Date : Thu Jul 08 2004 11:52 am

Wayne A. King wrote:

>On Wed, 07 Jul 2004 13:45:43 -0700, Bob Gonder <notbg@notmindspring.invalid>
>wrote:
>
>>It also looks like you are using Windows Target, not DOS, as your
>>floats are 64 bits.
>
>Floats are 32 bits in both DOS (16-bit) and Windows (32-bit).
>Doubles are 64 bits in both DOS (16-bit) and Windows (32-bit).

Then why does the compiler (BCB and apparently BC5.02) push 8 bytes
when passing a float?
 fld    dword ptr[esp]
	yep, 32bit float
 add    esp,FFFFFFF8
	make room for 8 bytes
 fstp   qword ptr[esp]
	put 64 bits on the stack!

Maybe all floats are passed as doubles? Everywhere?
Is this Standard?

.