On Wed, 21 Dec 2005, Peter W. Draper wrote:
> finally found the Solaris bug! It was a buffer overrun in HDSObject.c, the
> array of pointers passed to datGetVC needs to be one larger than the number
> requested (the final pointer is set to NULL to terminate).
Crap. I was worried that that would happen. Maybe I should remove that and
only require an array of the declared size of the HDS object (and not
NULL terminate on the basis thatthe caller now knows how many elements
there are. (datGet1C has the same issue)
--
Tim Jenness
JAC software
http://www.jach.hawaii.edu/~timj
|