Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!utfyzx!sq!msb
From: msb@sq.UUCP
Newsgroups: comp.arch,comp.lang.c
Subject: Re: String Handling -- strcpy() and shared memory
Message-ID: <1987Apr20.125657.6693@sq.uucp>
Date: Mon, 20-Apr-87 12:56:57 EST
Article-I.D.: sq.1987Apr20.125657.6693
Posted: Mon Apr 20 12:56:57 1987
Date-Received: Wed, 22-Apr-87 03:26:48 EST
References: <15915@sun.uucp> <1987Apr1.121330.15976@sq.uucp> <573@csun.UUCP>
Reply-To: msb@sq.UUCP (Mark Brader)
Organization: SoftQuad Inc., Toronto
Lines: 31
Xref: utgpu comp.arch:971 comp.lang.c:1736
Checksum: 38254

Mike Stump (aeusemrs@csun.UUCP) writes:

> Well, not to pick on you, but, have you ever done ANY programming using shared
> memory, succesfully, and bug free?

Just enough to be aware of the problems involved.

> Tell me, how would you do something like a
> strcpy in such an environment?

	char *strcpy (to, from)
	register char *to, *from;
	{
		char *start = to;
		while ((*to++ = *from++)) != 0)
			;
		return start;
	}

But this would be with the understanding that a copy of a string in the
process of being changed might produce garbage; any process using strcpy()
to lengthen a string would be expected to plant a suitably located NUL first.
(Not necessarily at strlen(from)+to; at to+sizeof(tobuffer)-1 would do.)
If this isn't sufficient protection, then semaphores or the like become
appropriate.

> You should just say:  ``No one should use shared memory, unless
> he/she knows what they are doing''

Well, I can certainly agree with that.
Mark Brader