NetNews Usenet Archive 1992 #30/NN_1992_30.iso/spool/gnu/misc/discuss/4033

Text File | 1992-12-17 | 1.4 KB | 45 lines

Path: sparky!uunet!olivea!sgigate!sgi!quasar.mti.sgi.com!davea
From: [e-mail] (David B.Anderson)
Newsgroups: gnu.misc.discuss
Subject: Re: alloca (was: makefile for gnu make on SGI)
Message-ID: <[e-mail]>
Date: 17 Dec 92 17:23:47 GMT
Sender: [e-mail]
Organization: Silicon Graphics, Inc.  Mountain View, CA
Lines: 34

In article <[e-mail]> [e-mail] (Steve Bacher) writes:
>In article <[e-mail]>,
>[e-mail] (David B.Anderson) writes:
[ ]
>>Brendan is quite right.  And it _is_ probably simplest to use gcc alloca.c
[ ]
>I weighed grabbing alloca.c out of the gcc or emacs distribution.  But
>this alloca requires a function xmalloc() to be defined.  Why not just
[ ]
>Do you have to change xmalloc to malloc, or define xmalloc to be
>malloc in the compile?

This question has arisen several times in the last year because the
source to alloca.c does not describe xmalloc.  xmalloc is simply a malloc
wrapper that guarantees it will never return 0.

#include <stdio.h>
#include <malloc.h>
#include <stdlib.h>
void *xmalloc(unsigned size)
{
  void *val;

  val = malloc(size);
  if(val == 0) {
    fprintf(stderr, "Oops! malloc on %lu bytes returned 0.\n",
    (unsigned long)size);
    exit(EXIT_FAILURE);
  }
  return val;
}

Regards,
[ David B. Anderson             [phone]             [e-mail] ]