From: carroll@m.cs.uiuc.edu
Date: 19 Aug 90 15:58 CDT
Newsgroups: comp.unix.i386
Subject: Re: Orphaned Response
Message-ID: <70400014@m.cs.uiuc.edu>
Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!wuarchive!brutus.cs.uiuc.edu!ux1.cso.uiuc.edu!ux1.cso.uiuc.edu!m.cs.uiuc.edu!carroll
Nf-ID: #R:<26BEEBD0:2:m.cs.uiuc.edu:70400014:000:1063
Nf-From: m.cs.uiuc.edu!carroll    Aug 19 15:58:00 1990
References: <2@<26BEEBD0>


/* Written 11:26 am  Aug  7, 1990 by chip@tct.uucp in m.cs.uiuc.edu:comp.unix.i386 */
/* ---------- "Bug fix for GDB 3.5 with COFF" ---------- */
This patch fixes a bug that shows up in GDB 3.5 when reading COFF
files.  The code that reads enum members doesn't stop when it should,
thus consuming symbols that should be read elsewhere.  In my system,
it showed up when debugging gdb failed; gdb complained because the
.bb/.eb (begin block/end block) symbols didn't nest properly.
Chip Salzenberg at ComDev/TCT     <chip@tct.uucp>, <uunet!ateng!tct!chip>
/* End of text from m.cs.uiuc.edu:comp.unix.i386 */
I've experienced the same problem under ISC 2.0.2, using GCC 1.37. I've
seen underruns as well as overruns for this (I put in a similar fix, but had
it generate a warning message). It seems that it is not a GDB bug, but a
problem with the binary itself. A count is included in addition to the EOS
mark, and that count is wrong sometimes. I don't use GAS, and assembler output
from GCC seems to be correct. I'm not an expert on this, so I could be
mistaken.