floor(-0.5)

[Francois Grieu]

In ISO/IEC 9899:1999, it is unambiguous that floor(-0.5)
has the value -1 (converted to double).

Was this the case in earlier C standards?

Is anyone aware of an implementaton that has it wrong ?

Françosi Grieu

 

[Keith Thompson]

In ISO/IEC 9899:1999, it is unambiguous that floor(-0.5)
has the value -1 (converted to double).

Was this the case in earlier C standards?

C90 says:

The floor function returns the largest integral value not greater
than x, expressed as a double.

C99 says:

The floor functions return |_ x _|, expressed as a floating-point number.

where "|_" and "_|" are my crude rendition of the mathematical symbol
for the floor function.

Is anyone aware of an implementaton that has it wrong ?

I'm not -- and given that the function is called "floor", getting it
wrong (say, by rounding towards 0.0) would be a pretty dumb mistake.

 

[Francois Grieu]

C90 says:

The floor function returns the largest integral value not greater
than x, expressed as a double.

Very clear indeed.

C99 says:

The floor functions return |_ x _|, expressed as a floating-point number.

where "|_" and "_|" are .. the mathematical symbol for the floor function.

And |_ x _| is defined in ?3.19 as
... "the greatest integer less than or equal to x.
Example: |_ 2.4 _| is 2, |_ -2.4 _| is -3."


Thanks for the C90 quote !

Francois Grieu