[MLton-devel] implementing raise and the exception stack

Henry Cejtin henry@sourcelight.com
Thu, 2 Jan 2003 15:54:40 -0600

My impression is that in C++ the fancy stack-walking makes exceptions very
expensive.  One thing that I vaguely recall is some kind of trick so that
different ranges of the program counter in some way.
The slowness of handing doesn't matter much to C++ or Java since exceptions
really do tend to be some kind of error or rare event.  It would matter more
for ML.  Still, it would be interesting to play around.

This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
MLton-devel mailing list