Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utcs!mnetor!seismo!topaz!gaynor
From: gaynor@topaz.UUCP
Newsgroups: net.lang.c
Subject: Re: Precedent for use of =
Message-ID: <5330@topaz.RUTGERS.EDU>
Date: Sun, 13-Jul-86 02:01:01 EDT
Article-I.D.: topaz.5330
Posted: Sun Jul 13 02:01:01 1986
Date-Received: Sun, 13-Jul-86 06:55:19 EDT
References: <2323@umcp-cs.UUCP> <514@cbmvax.cbmvax.cbm.UUCP> <5057@sun.uucp>
Organization: Rutgers Univ., New Brunswick, N.J.
Lines: 36

In article <5057@sun.uucp>, guy@sun.uucp (Guy Harris) writes:
> If your compiler can't generate equally efficient code for
> 
> 	a = b;
> 	if (a != 0)
> 
> and
> 
> 	if ((a = b) != 0)
> 
> on any machine other than an incredibly weird one, then your compiler needs
> to be improved.

Ideally, degrees of optimization should be available, offering
trade-offs between code-speed vs code-space vs compilation-time vs
code-size, and optional code generation, to name the biggies.  For
example, when debugging syntax errors, why waste time generating code
when all you want to do is check the syntax?  When verifying a
program's correctness, your not going to need really good code, just
something done quick-n-dirty (to save on system resources).  For
production, the trade-off of code-size vs code-speed rears its ugly
head (code-size is very important when, say, you try to port gnu-emacs
to a small computer).  I reiterate, this is the ideal situation.


                         _   /|
                         \`o_O'
                           ( )   Aachk! Phft!   
                            U

Disclaimer: The opinions and/or information and/or code expressed
            here were generated by this characature, stolen from
            Dave Rasmussen, to which I have taken the liberty of
            adding ears.  So don't look to me for a disclaimer!

Silver  {...!topaz!gaynor}