Re: [bug-grep] length of dec. representation of a number

[Paul Eggert]

Stepan Kasal <address@hidden> writes:

>   I'm going to check in the following small patch.

How about if you use the standard gnulib macro for this instead of
reinventing the wheel?  Here it is, taken from inttostr.h:

/* Upper bound on the string length of an integer converted to string.
   302 / 1000 is ceil (log10 (2.0)).  Subtract 1 for the sign bit;
   add 1 for integer division truncation; add 1 more for a minus sign.  */
#define INT_STRLEN_BOUND(t) ((sizeof (t) * CHAR_BIT - 1) * 302 / 1000 + 2)

#define INT_BUFSIZE_BOUND(t) (INT_STRLEN_BOUND (t) + 1)



Julian Foad <address@hidden> writes:

> Certainly "CHAR_BIT" was inappropriate and misleading

No, it's appropriate.  For example, sizeof (int) * CHAR_BIT is an
upper bound on the number of value bits in an int.  (On some
architectures, some of the int bits might not contribute to the value;
but you can't have more value bits than total bits.)

On some machines, CHAR_BIT is 32 and sizeof (int) == 1.  Admittedly
these are rare as GNU porting targets (they're typically DSPs or
something like that) but if it's easy to port to them we might as well
do so.