🚀 go-pugleaf

In article <1990Mar28.135444.12665@granite.cr.bull.com> kittlitz@granite.cr.bull.com (Edward N. Kittlitz) writes:
>I have experienced the same problem with ESIX while trying to debug
>emacs problems. I believe I got around the problem by compiling the
>offending module using the supplied cc.

Minor permutations of symbol order make the bug appear and disappear.  I was
able to circumvent it in gcc by moving the failing enum definition from the
middle of the source file to the beginning.

>I suspect the bug lies in some common ATT source.

The problem is pretty clearly in the 386/ix linker, which I believe 386/ix
took unmodified from the common AT&T distribution.  (Since ISC did the port
for AT&T, you could equally say that AT&T uses ISC's linker.  You get the
idea.)

As far as I can tell, if you apply the patch I sent, gdb can read any symbol
table that the linker puts out.
-- 
John R. Levine, Segue Software, POB 349, Cambridge MA 02238, +1 617 864 9650
johnl@esegue.segue.boston.ma.us, {ima|lotus|spdcc}!esegue!johnl
"Now, we are all jelly doughnuts."