Path: chuka.playstation.co.uk!news
From: laura smith <laura_smith@interactive.sony.com>
Newsgroups: scee.yaroze.programming.gnu_compiler
Subject: Re: Relatively Important(!) : Buggy gcc
Date: Tue, 29 Apr 1997 15:43:37 +0100
Organization: Sony Computer Entertainment Europe
Lines: 41
Message-ID: <33660919.23E2@interactive.sony.com>
References: <336635d6.3891193@news.playstation.co.uk> <336516C4.2958@micronetics.com> <3366dd38.1265914@news.playstation.co.uk>
NNTP-Posting-Host: 194.203.13.10
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Mailer: Mozilla 3.01 (Win95; I)

Alex Amsel wrote:
> 
> On Mon, 28 Apr 1997 22:29:40 +0100, Jim <jim@micronetics.com> did
> quoth at me:
> 
> >padding the structure out.  The problem is caused by XNeg, YNeg both
> >being a byte each and unaligning the whole structure.
> 
> Funny thing is that it happened when both were UWORDS as well. I got
> Mark to double check this, but again it was late at night sooo....
> 
> We seem to usually get problems with arrays, but then that doesn't
> explain bug 1. I'll have to start using gdb I htink, but since I don't
> have time for yaroze coding this week it'll have to wait.
> 
> * Alex Amsel * Into Beyond Web Design & JAVA Programming *
> * http://www.intobeyond.com  *  WWFC Utter Rubbish 1996-7*
> MM - "Steve Corica is every bit as good as that Kinkladze"


Apologies for these peculiar bugs.

Cannot offer real insight/explanations, definitely agree
that the problem is with padding; I can offer a way
to avoid coming up against this:
namely, never both with non-integer (ie 32bit long) types
in structures. I have never encountered the bug types you
report, and I nearly always use ints; 
if memory is genuinely getting short, use the different bits
of ints, eg with accessor macros.

Since the trouble seems to arise with arrays of non-simple-int
type, avoid them.

Sorry if this sounds didactic or frivolous,
but I'd far rather quickly work out how to avoid a class of bugs
than spend the time tracing it to its (compiler?)
basis and then fixing it.


Lewis