• Mikael Klasson <Mikael_member pathlink.com> Apr 13 2003
• "Walter" <walter digitalmars.com> Apr 13 2003
• Mikael Klasson <Mikael_member pathlink.com> Apr 13 2003
• "Walter" <walter digitalmars.com> Apr 13 2003

Ok, I'm out of ideas here.

fmul	st(0)
^
proth_sieve.cpp(407) : Warning 13: Illegal type/size of operands for the fmul
instruction

Why the warning? I haven't managed to assemble any fpu instruction that uses a
register explicitly.
The "warning" is a bit strange as well. If you look at the generated code, it
actually gets assembled to d8,08 == fmul   dword [eax]. That's quite wrong.

Regards,
Mikael

try:
    fmul    st, st(0)

"Mikael Klasson" <Mikael_member pathlink.com> wrote in message
news:b7cgn3$22uc$1 digitaldaemon.com...
 Ok, I'm out of ideas here.

 fmul st(0)
 ^
 proth_sieve.cpp(407) : Warning 13: Illegal type/size of operands for the


 instruction

 Why the warning? I haven't managed to assemble any fpu instruction that


 register explicitly.
 The "warning" is a bit strange as well. If you look at the generated code,


 actually gets assembled to d8,08 == fmul   dword [eax]. That's quite


 Regards,
 Mikael

try:
    fmul    st, st(0)


Oh, ok. Thanks for the quick reply! The support around here rocks. =)

Don't you think that's a bit of an odd syntax, though?

"Mikael Klasson" <Mikael_member pathlink.com> wrote in message
news:b7cko8$25a8$1 digitaldaemon.com...
try:
    fmul    st, st(0)


 Oh, ok. Thanks for the quick reply! The support around here rocks. =)

 Don't you think that's a bit of an odd syntax, though?


Intel/Microsoft never quite made up their mind how it should look in the
early days, and that confusion has carried forward. The inline assembler
does just one of the many variants.