Newsgroups: comp.lang.c
Path: utzoo!henry
From: henry@zoo.toronto.edu (Henry Spencer)
Subject: Re: alloca() portability
Message-ID: <1990Nov17.221822.25202@zoo.toronto.edu>
Organization: U of Toronto Zoology
References: <1990Nov09.233527.7489@chinet.chi.il.us> <27608@mimsy.umd.edu> <PAULB.90Nov12131357@hcx2.ssd.csd.harris.com> <8789@scolex.sco.COM>
Date: Sat, 17 Nov 90 22:18:22 GMT

In article <8789@scolex.sco.COM> seanf@sco.COM (Sean Fagan) writes:
>Lessee... we've had Chris and Doug denounce alloca(), but I haven't seen
>anything from Henry (yet?  I suspect he would also denounce it)...

Suspicion confirmed.  Alloca() is a kludge that should have died long ago.
There *is* a general problem with the need for various kinds of cleanup on
exiting a function, either by the normal route or by longjmp(), but alloca
solves only one special case of that, at the cost of major portability
problems and new opportunities for interesting bugs.  A proper fix is
impossible without language changes.
-- 
"I don't *want* to be normal!"         | Henry Spencer at U of Toronto Zoology
"Not to worry."                        |  henry@zoo.toronto.edu   utzoo!henry
