In article 94Jul7200831@physics16.berkeley.edu, matt@physics16.berkeley.edu (Matt Austern) writes:
>In article <JASON.94Jul7162555@deneb.cygnus.com> jason@cygnus.com (Jason Merrill) writes:
>
>> It is part of the current WP.  On the other hand, there is a proposal to be
>> considered at the X3J16 meeting next week which would remove all exception
>> specifications from the language; I don't expect it to pass, but I could be
>> wrong.
>
>I wouldn't mind that: as presently defined, I don't think that
>exceptiomn specifications are very useful.  It's a good idea, but only
>if it's done right---and doing it right, in my opinion, would require
>a major redesign of much of the language.  I wouldn't be sorry if the
>exception specifications we've currently got disappeared.
>
>--
>       Matt Austern                       "Se non e vero, e ben trovato"

I agree about that. I remember when learning about exceptionhandling.
My first thought of how to use it was handling "out of memory" when
objects were being constructed. Using it as a part of the objects
implementation, (using the exception semantics to solve functionallity),
never occured to me because I never really felt sure about what
was going to happen when I throwed the exception! (I'm the kind of
programmer who likes to know exactly what is happening and
obviously dont like compilergenerated features).


Fredrik Jonsson.