mycroft@gnu.ai.mit.edu wrote:
> This works fine on other ports.  Furthermore, the first `1024' you
> pointed out is not a bug, and the other five, while they could in some
> random circumstance lose, are only used while building sh(1).
Ok, just looked suspicious from a grep

> This looks like a problem specific to the pc532.
cd /usr/src/lib/libc/arch ; ls
CVS/    i386/   m68k/   mips/   ns32k/

ns32k has a longer name than the others, & it gets repeated many times,
hence ns32k sees the problem before the i386 mainstream folk

--
Julian Stacey
<stacey@freefall.cdrom.com>	Tel. +49 89 268616  ( TZ=GMT+1 )
Alternates:	<julian@pcsbst.pcs.com>,<jhs@regent.e-technik.tu-muenchen.de>

------------------------------------------------------------------------------