Sounds ok...
It IS better to optimize when code is correct. And c++ takes the other
approach.
Is Pragma suppress part of the ARM or some extension?

"Marin David Condic" <marin.condic.auntie.spam@pacemicro.com> a �crit dans
le message news: 9j9vmh$c5p$1@nh.pace.co.uk...
> Oh, you'll pay - just not with CPU cycles. :-) Consider it this way: Its
> better to have the checks there by default (and by virtue of the
language -
> not by virtue of someone being "nice" and checking manually) and removing
> them from the small percentage of code that may be performance critical
(via
> pragma Supress) than to have them NOT there by default (and requiring that
> the programmers "roll their own", rather than have them by virtue of the
> language) and try to intersperse them in the code as you start discovering
> you have mysterious errors.
>
> I can't even begin to tell you how often my hide has been saved by having
> compiletime & runtime checks in place in Ada that exposed eggregious
errors
> that would have been very hard to track down without them. When I wear the
> embedded/realtime hat, I understand the desirability of getting rid of the
> checks (in some cases) but I'm glad they are there for me to decide to get
> rid of rather than have to insert them myself *after* I start discovering
> problems.
>
> Now if C++ had some sort of compile-time switch that could be thrown that
> said "Insert Checks", and the language actually had syntax/semantics that
> provided information to check (ranges on numbers, array constraints, etc.)
> then I'd think that C++ and Ada were at least equivalent in this area.
> However, since there aren't any activatable checks and the language
doesn't
> provide nearly as much safety information anyway, I find this a major
> liability.
>