Re: [RFC] C++ vs forced unwinding

[Mike Stump <mrs at apple dot com>]

On Monday, March 31, 2003, at 06:12 PM, Richard Henderson wrote:
> To refresh your memory, forced unwinding is a not-an-exception
> that runs only cleanups, but can't be caught.

Hum...  I'd almost like to see the complete description of this...  It 
differs from the model we worked out years ago for threads and EH.  In 
that scheme, cancellation was just a throw of a pthread_terminate type 
value.  The problem I see is that in C++, catch (...) { ... } can be a 
cleanup, but, if it is, or not, cannot be known by the compiler.  So, 
when you say, cancellation runs cleanups, this means, in C++, that at 
least some catch (...) { } things are run.  :-(

With the cancellation being part of EH, the interactions with C++ are 
natural and obvious.

Also, does the design cover Ada as well?

If C++ was well represented in the designed of this feature, ignore 
me...