993 – incorrect ABI documentation for float parameters

If the last function parameter is a float it is passed via the stack and not 
via EAX.
# The last parameter is passed in EAX rather than being pushed on
# the stack if the following conditions are met:
#
# * It fits in EAX.
# * It is not a 3 byte struct.
missing condition:
# * not a float/ifloat
# void test(float f){
# assert(16.5 == f);
# }
# 
# void main(){
# float f = 16.5;
# int i = *cast(int*)&f;
# version(bug){
# asm{
# mov EAX, i;
# call test;
# }
# }else{
# asm{
# push i;
# call test;
# pop i;
# }
# }
# }
dmd test.d && ./test && echo OK
> OK
dmd -version=bug test.d && ./test && echo OK
> Error: AssertError Failure bug.d(2)

Nitpick: that "pop i" after the call in the else clause shouldn't be there. D functions pop their own arguments, except in case of varargs.

Fixed DMD 1.009