Path: utzoo!attcan!telly!lethe!torsqnt!news-server.csri.toronto.edu!rutgers!mit-eddie!snorkelwacker.mit.edu!apple!usc!cs.utexas.edu!uunet!mcsun!ukc!slxsys!ibmpcug!mantis!tony
From: tony@mantis.co.uk (Tony Lezard)
Newsgroups: comp.lang.c
Subject: Re: Finding NULL byte in a long
Message-ID: <ic46T1w163w@mantis.co.uk>
Date: 14 Dec 90 11:34:41 GMT
References: <1990Dec13.153552.1862@craycos.com>
Organization: Mantis Consultants, Cambridge. UK.
Lines: 23

pmk@craycos.com (Peter Klausler) writes:

>Finding the *position* of the null byte is a little trickier. It helps a lot
>if your machine has a "leading zero count"/"find first set bit" instruction
>and your compiler lets you use it.

Assuming you're using 2's compliment arithmetic, there's always a "find
first set bit" instruction. For any x, the expression x & -x yields all zeros
except for a 1 in the position of the first set bit from the right.

For example, if x = ...000010011001000
               -x = ...111101100111000
        => x & -x = ...000000000001000
                                  ^
                                  `-- Voila!

I don't know how well known this is. There are even occasions where it is
actually quite useful!

==========================================================================
Tony Lezard.  E-mail: tony@mantis.UUCP, Snail-mail: Mantis Consultants,
Unit 56, St. John's Innovation Centre, Cambridge, CB4 4WS, United Kingdom.
Most appropriate anagram of name: Lazy Rodent.