Path: utzoo!attcan!uunet!sugar!peter
From: peter@sugar.uu.net (Peter da Silva)
Newsgroups: comp.sys.atari.st
Subject: Re: Looking for an Evangelist
Message-ID: <3168@sugar.uu.net>
Date: 28 Dec 88 20:30:15 GMT
References: <681@stag.UUCP> <1271@atari.UUCP>
Organization: Sugar Land Unix - Houston, TX
Lines: 34

In article <1271@atari.UUCP>, apratt@atari.UUCP (Allan Pratt) writes:
> In article <681@stag.UUCP> to_stdnet@stag.UUCP writes:
> > [peter@sugar.uu.net (Peter da Silva) writes...]
> > 
> > > [TOS's] memory management has a fatal
> > > bug that forces everyone to roll their own. but I'll wager that if you free
> > > something twice in a complex 'C' program and let it run long enough it'll
> > > crash.

> I do not consider this a bug.  Who does? The fact that you attempted to
> free a block twice is a bug.

I agree with this emphatically. The original debate, which has been lost
in the mists of time, is that the Amiga memory allocation is flawed because
it doesn't watch for this bug. I was saying that the same thing probably
would happen on the ST under the same circumstances, and that this wasn't
a problem with the ST or the Amiga... but with the program in question.

> Secondly, the Malloc system call is not intended for general memory
> management: the overhead in memory and time for using it are
> prohibitive.

The problem is that most 'C' runtime environments deal with the problem
by Mallocing all available memory and doling it out. This is a problem
because it screws up when you try to call another program and the exec
call fails because there's no free memory.

I sympathise with you, by the way. I can see your frustration building up
as you try to keep the horde of PC hackers out of the innards of your box.
I suppose it's too late to file the serial numbers off OS-9 and call it
TOS-II.
-- 
Peter "Have you hugged your wolf today" da Silva  `-_-'  Hackercorp.
...texbell!sugar!peter, or peter@sugar.uu.net      'U`