Re: exiting a bloc

[Philip Anderson <phil.anderson@amsjv.com>]

Ted Dennison wrote:
>  
> I know I pretty much follow that rule myself. But I'm not sure I'd like to
> see it elevated to Law. I do occasionaly like to use "declare" for sizing
> arrays (particularly strings) dynamicly on the stack. I think I've also used
> them in the "accept" alternatives of server tasks or in branches of large
> command-processing case statements to declare local variables that are
> unique to that alternative.
> 
> About the only other reason I ever use a declare is to declare local
> variables that are used in the exception code of a block's exception
> handler. It'd be a bit excessive to create a new routine just because the
> exception handler needs a variable. It makes sense to limit the variable's
> scope to the block, since that's the only place its used and the block is
> already there.

I'd agree with that in general, but it's a not unreasonable restriction
to impose if restrictions are being imposed.

Declare blocks are excluded in SPARK, and it makes sense there; most of
the situations where you and I might use it are not applicable in SPARK
anyway - no dynamic arrays, tasks or exceptions.  Branches of case
statements could arise, but it would seem more readable to call a
procedure if one is that complicated.


-- 
hwyl/cheers,
Philip Anderson
Alenia Marconi Systems
Cwmbr�n, Cymru/Wales