From mboxrd@z Thu Jan  1 00:00:00 1970
From: Jakub Jelinek <jakub@redhat.com>
To: Andrew Cagney <ac131313@cygnus.com>
Cc: Ian Lance Taylor <ian@zembu.com>, Thiemo Seufer <ica2_ts@csv.ica.uni-stuttgart.de>, binutils@sources.redhat.com
Subject: Re: [PATCH] Include alloca.h for bfd
Date: Thu, 23 Aug 2001 11:49:00 -0000
Message-id: <20010823205212.J5765@sunsite.ms.mff.cuni.cz>
References: <20010821210909.E30301@rembrandt.csv.ica.uni-stuttgart.de> <sihev1180q.fsf@daffy.airs.com> <20010821234604.J30301@rembrandt.csv.ica.uni-stuttgart.de> <m3y9oaj0x7.fsf@north-pole.nickc.cambridge.redhat.com> <20010823174716.E3542@rembrandt.csv.ica.uni-stuttgart.de> <20010823180440.F5765@sunsite.ms.mff.cuni.cz> <20010823181513.F3542@rembrandt.csv.ica.uni-stuttgart.de> <3B8539CD.5060306@cygnus.com> <sibsl6wur2.fsf@daffy.airs.com> <3B854EA7.7020706@cygnus.com>
X-SW-Source: 2001-08/msg00548.html

On Thu, Aug 23, 2001 at 02:42:47PM -0400, Andrew Cagney wrote:
> > Hmm, shame you can't use asprintf().  It does everything you want
> >> except use bfd_malloc().
> > 
> > 
> > Now that you mention it, we could use asprintf there.  There is an
> > implementation in libiberty.  If it can't allocate memory, it sets the
> > buffer to NULL.
> > 
> > Doesn't matter much, though.
> 
> Yes.  Its the bit where bfd_malloc() sets the error message that is missing.

Guess
	      size_t len = strlen (h->root.string);
	      char *buf = (char *) bfd_malloc (len + sizeof ("__imp_"));
	      memcpy (buf, "__imp_", sizeof ("__imp_") - 1);
	      memcpy (buf + sizeof ("__imp_") - 1, h->root.string, len + 1);
              arh = archive_hash_lookup (&arsym_hash, buf, false, false);
	      free(buf);

will be faster anyway, so there is no need to look for asprintf for concat
of 2 strings together (well, the above would be even nicer with mempcpy).

	Jakub