I'm thinking exit() shouldn't be catchable to maintain status quo, and it
should be focused on the reason it was suggested(Unwinding stacks and
cleaning up memories instead of longjmp'ing to shutdown).

If there's any need to catch it's exception, that can be handled later
through maybe a RFC discussion.

This can be implemented directly without having any user land interaction
since the throwing and catching can't be caught by any user land
code(top-most hierarchy without possibility to be caught, which might
result in compile time error).

All the best

Reply via email to