[MLton-devel] Optimizing garbage collection efficiency paper

Stephen Weeks MLton@mlton.org
Fri, 3 Jan 2003 12:36:14 -0800

> re MLton being safe-for-space, the only counter-example I know of (and it
> really depends on the precise definition of safe-for-space) is the one we
> talked about: if a function never returns except by raising an exception then
> the stack frames from the top down to the first exception handler should be
> dead.  

True enough, although I think that falls outside most people's
definition.  There is also the related example of when a function
never returns or raises, then the entire stack below it is dead.  This
is actually relevant when it comes to signal handling or multithreaded
programs, where I could imagine a thread entering an infinite loop
waiting for something to happen.

Both of those have been for sometime (and may always be :-) and on the
todo list.

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