Path: utzoo!utgpu!watserv1!watmath!att!ima!ima.isc.com
From: karl@ima.isc.com (Karl Heuer)
Newsgroups: comp.lang.c
Subject: Re: alloca() portability
Keywords: portability, alloca
Message-ID: <5237@ima.ima.isc.com>
Date: 5 Nov 90 06:08:49 GMT
References: <9032@ncar.ucar.edu> <499@taumet.com>
Sender: karl@ima.ima.isc.com
Reply-To: karl@ima.isc.com (Karl Heuer)
Distribution: na
Organization: Interactive Systems
Lines: 14
Supersedes: <1990Nov04.213253.4801@dirtydog.ima.isc.com>

In article <499@taumet.com> steve@taumet.com (Stephen Clamage) writes:
>russ@groucho.ucar.edu (Russ Rew) writes:
>>[Are there any real implementations where Doug Gwyn's alloca() fails?]
>There is more to it than the target machine.  Some compilers get a
>performance win by omitting frame pointers ...

Irrelevant.  The Gwyn alloca() doesn't use the algorithm of moving the frame
pointer and jumping back to the caller; it's written in C, and I've used it on
systems where the assembly version can't work for the reasons stated.

In response to the original question: it might conceivably fail on a system
that does strict pointer checking (Sabre).

Karl W. Z. Heuer (karl@ima.isc.com or uunet!ima!karl), The Walking Lint