Date: Thu, 29 Jan 1998 11:18:14 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: Nate Eldredge <eldredge AT ap DOT net>
cc: cterry AT plinet DOT com, djgpp AT delorie DOT com
Subject: Re: djgpp/rhide acting strange
In-Reply-To: <199801282323.PAA15405@adit.ap.net>
Message-ID: <Pine.SUN.3.91.980129111756.6822M-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Precedence: bulk


On Wed, 28 Jan 1998, Nate Eldredge wrote:

> My preference is to compile without optimizations, get it to work (perhaps
> with the debugger), bang on it for a while, then compile with optimizations
> and test some more before releasing. I think that's the best trade-off, but
> of course for anyone, YMMV.

Well, my mileage varies.  I was doing the same as you suggest
initially.  But then I noticed that most of the testing and debugging
is done on the non-optimized version; by the time I got to debug the
optimized one, I was usually bored or sick and tired, and tended to
skip some of the tests.  So now I begin the debugging with the
optimized version and only go to unoptimized one if I can't find the
bug otherwise.

Besides, turning on optimizations will make GCC check the source for
additional problems, like variables that are used before they are
initialized, so at least -O is mandatory, I think.