In the past we have had a number of problems with Sun's optimizer
and our use of setjmp/longjmp to handle exceptions. The problem was
particularly nasty since many of the longjmps would work just fine, but
we would see crashes on some of them.
However, with the Sun 4.1.2 compiler, things seem to have improved. But how can I know for sure without testing every branch path? So: does anyone know if