It seems that for the question /Library/UCSB/Stewart5_3_4/Stewart5_3_4_28.pg in the problem library, the random seed 1157 always results in "Can't generate enough valid points for comparison". Is this a bug? The formula is defined for all positive values of x.
In reply to Michael Shulman
Re: One random seed can't generate enough comparison points
by Michael Shulman -
If no one knows, can anyone at least give me some help for debugging? E.g. how can I find out what comparison points it is trying to generate?
In reply to Michael Shulman
Re: One random seed can't generate enough comparison points
by Davide Cervone -
The default range of values is -2 to 2 for MathObjects (and -1 to 1 for the traditional checkers), and so it is possible for all the random points to be negative. The problem needs to be modified to restrict to positive values.
You can add diagnostics=>1 to the cmp() call in order to get debugging output about the points used and their values, but this only works when sufficient points actually are found.
In any case, these issues are better handled via the WeBWorK Bugzilla site rather than this question-answer site.