Another bug

Robert Graham Merkel rgmerk@mira.net
Wed, 20 Sep 2000 20:30:05 +1100


Robert Graham Merkel writes:
 > Conrad Canterford writes:
 >  > Robert Graham Merkel wrote:
 >  > > Conrad Canterford writes:
 >  > >  > There appears to be a threshold (I'm not sure yet whether it would be
 >  > >  > accounts or transactions), beyond which Gnucash goes into a processor
 >  > >  > bound loop from which it appears to never return (certainly didn't in
 >  > >  > over 15 minutes).
 >  > > This is likely to be a guile bug.  What version are you using?
 >  > 
 >  > Guile from (Redhat) RPM 1.3-10 which gives me libguile.so.4.0.0
 >  
 > OK, I've confirmed a problem exists (though I get different symptoms,
 > it just crashes for me).
 > 
 > Thanks, I'll let you know when I track the problem down.

Rob, I've got a bug that is manifesting itself in the transaction
report but appears to be a guile problem.  To reproduce, get a nice
large data file and do a transaction report that contains every
transaction in that data file.  Instant core dump.

I've tried altering the stack size and the maximum depth and it
appears to make no difference, but it's definitely a size thing, 
which tends to suggest it's not an error in the scheme code (nice thing about
scheme, it tends to follow the 0, 1 or infinitely-many principle).

Can you reproduce this?  Any ideas?  

------------------------------------------------------------
Robert Merkel	                           rgmerk@mira.net

------------------------------------------------------------