[Larceny-users] Assertion failed in stats_stop_timer

[William D Clinger]

Sven wrote:
> >From time to time, I am seeing the following messages:
>
> Larceny Panic: Sys/stats.c;292: Assertion failed
>
> Is this my fault?

Probably not, but I can't tell for sure.  It appears
that the assertion is supposed to panic if an attempt
is made to stop a timer that isn't running.  (A panic
would also occur if too many timers are in use at
once, but I would think that would have been detected
by a different assertion first.)

It also appears to me that the assertion could fail
intermittently for timers that wrap around.  I wouldn't
expect that to happen very often, however, which may
explain why Lars apparently didn't worry about it when
he wrote the code; he probably thought of those timers
as a crude tool for his PhD research.

Another possibility is that some bug I don't know
about is setting the timer to 0.  It would help to
know whether and how you are trying to use the timers
(which may be used implicitly by run-benchmark, et
cetera).  In particular, are you using deeply nested
calls to run-benchmark or something similar?

Will