runaway gnatsds

[Joseph S. Myers]

On Thu, 5 Apr 2001, Jason Molenda wrote:

> It seems unlikely that there is a valid reason for gnatsd to use
> two or three hours of CPU time...  I tried attaching to one of them
> and got a BS stacktrace -

Could these be related to the GNATS database corruption?  Does gnatsd
handle corrupted PRs any better than GNATSweb (which seems to truncate
corrupted lines in the PRs after NUL characters, making me wonder whether
any of the corrupted PRs are getting further and less visibly corrupted by
people editing them)?

http://sources.redhat.com/ml/overseers/2001-q2/msg00013.html
http://gcc.gnu.org/ml/gcc/2001-04/msg00071.html

-- 
Joseph S. Myers
jsm28@cam.ac.uk