more accurate comment for the x==1.0 hack

This commit is contained in:
Jeff Bezanson 2013-12-14 15:49:02 -05:00
parent 9bc257a323
commit 89d232d114

View file

@ -125,8 +125,10 @@ __ieee754_exp(double x) /* default IEEE double exp */
if(x < u_threshold) return twom1000*twom1000; /* underflow */ if(x < u_threshold) return twom1000*twom1000; /* underflow */
} }
/* by pure bad luck, this implementation's error for x==1.0 is /* this implementation gives 2.7182818284590455 for exp(1.0),
unusually large, although still within 1 ulp. */ which is well within the allowable error. however,
2.718281828459045 is closer to the true value so we prefer that
answer, given that 1.0 is such an important argument value. */
if (x == 1.0) if (x == 1.0)
return 2.718281828459045235360; return 2.718281828459045235360;