28 Jan
2005
28 Jan
'05
3:47 p.m.
Sérgio Vale e Pace wrote:
not exactlly, sometimes you get the 1 difference correctly so, you will introduce an error for such cases, and on some compillers (bcc) for some cases you get an error because it don´t know which version to instantiate Sqrt<N,X,XP1,XP1> or Sqrt<N,X,X,XP1>
XP1 should never be equal to X, so how could there be an ambiguity between the instantiations? Also, even when the difference of 1 is arrived at correctly, isn't it reasonable to assume that the lesser of the two numbers is the floor of the square root? -Jason