Re: Compiling problem on AIX 4.2.1 with gcc

From: Michael V. Pelletier <mikep@dont-contact.us>
Date: Tue, 20 Apr 1999 18:22:33 -0700 (PDT)

On Wed, 21 Apr 1999, Chris Wedgwood wrote:

> This is a lie... I fixed this and verified it works quickly and for
> the most part harmlessly to 600K FDs (slight overshoot of two here,
> not sure why, I think it because the starting point was poorly
> chosen).

It works best if the starting point is a power of two, since we're dealing
with integer values and division by two via the bit-shift-right operator.
On five occasions, the bit shift pushed off a one value in the 2^0
position of the number, losing the 0.5 and leading to the 60002 result.
The better starting value would be 524288, also known as 2^19.

        -Mike Pelletier.
Received on Tue Jul 29 2003 - 13:15:57 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:12:05 MST