[Larceny-users] Assertion failed in stats_stop_timer

Felix Klock felixluser at pnkfx.org
Mon Feb 19 18:34:01 EST 2007


On Feb 18, 2007, at 2:33 PM, Sven.Hartrumpf at FernUni-Hagen.de wrote:

> Hi all.
>
>> From time to time, I am seeing the following messages:
>
> Larceny Panic: Sys/stats.c;292: Assertion failed
>
> Is this my fault?
>
> Current VSize was 1494756KB.
> Larceny v0.93.2 "(development)" (Feb 16 2007 13:46:06,  
> precise:Linux:unified)

 From this version header, I infer that you are using a development  
version of Larceny off the svn-trunk and that you built the  
development version yourself.

If this is true, then you might consider adding a variation of the  
failing assertion to the function stats_start_timer in your copy of  
src/Rts/Sys/stats.c, and then rebuild the runtime.  That way you can  
at least isolate for us whether this is a problem where  
osdep_realclock() or osdep_cpuclock() is returning 0, or if its a  
problem where the timer state is being corrupted by something else.

(I am not sure whether osdep_cpuclock and osdep_realclock are  
supposed to be allowed to return 0.  Their specifications in osdep.h  
explicitly say in a FIXME that they will overflow after 49.7 days;  
this to me indicates that clients of the current interface should be  
prepared for any value, including 0.  But this 49.7 day figure would  
also seem to indicate that this should not be what is causing Sven's  
problem. . .)

If I manage to replicate your problem on our end, then I will do this  
sort of sleuthing myself.  But I do not know if I will be able to  
replicate the problem on our end.  You sound like you can reproduce  
the problem, so it might be easiest to track down the problem on your  
end.

-Felix




More information about the Larceny-users mailing list