This is the mail archive of the gcc-bugs@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [Bug c/16850] modulo - feature or bug?


For the scientific modulo the integer() function yields the lower
bound integer of the float argument. In C it appears that integer(f)
always equals integer(abs(f)). As a consequence, C modulo's integer
function does not fulfil the equation

  integer(f+n) = integer(f) + n
  n := any integer

for every f while n!=0. That's bad ground for algebra!

On the bright side, the equation


int(a/b + n) = int(a/b) + n

works fine for all a, b, n where a/b is defined in the
algebraic sense (and note we're talking integers here).

"Computer division" has nothing to do with algebraic
division (well, not very much, anyway).  Better get
used to it :-)


Segher



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]