🚀 go-pugleaf

I'm trying to track down a sporadic, but definitely repeatable bug
using Emacs 18.57 on an Ultrix 4.1 DECstation. Behaviour is simple to
describe: every so often (maybe once every 2-3 days), Emacs dumps core
(helpful huh ?).The same behaviour has also been seen in 18.55 but is
*much* less frequent.

Core dumps are not proving terribly helpful: gdb tries to tell me that
there are 179 (+/- 20) nested layers of calls to mark_object, with a
single call to mark_buffer about 40 levels deep. On the other hand,
dbx doesn't show the recursion, but still points to mark_object being
the last stack frame before abort() is called. I could spend a long
time tracing through the logic of gc and marking, but since I'm rather
new to the mips architecture, I thought I'd first see if anyone else
has come across this, or even better, has a fix for it.

Thanks in advance,
-- 
Paul Barton-Davis			<pauld@cs.washington.edu>
UW Computer Science Lab		``to shatter tradition makes us feel free''