From: nmm1 on
In article <i2hr8t$av9$2(a)speranza.aioe.org>,
Dragan Milenkovic <dragan(a)plusplus.rs> wrote:
>>
>> That misses the critical point. The library does not specify enough
>> behaviour in most exceptional cases, which means that it is impossible
>> to write portable or reliable recovery code. As I posted, it could
>> do a lot better, WITHOUT comprimising efficiency, but the design,
>> specification and implementation efforts are comparable with what
>> has already been done.
>
>Are you sure that it is impossible? Or is it just not trivial?

I am sure that it is impossible, both theoretically and practically.

> From what I understand on this topic is that the main reason for
>all of this is that the Standard does not want to force a specific
>implementation and representation. However, I don't have a strong
>opinion whether this is good or bad, only that it is fairly
>reasonable.

That is correct. The principle is good, but I will leave others to
argue over whether C++ does it well.

>Of course, wording in the Standars is a horror... :-/

True, but it's a hell of a lot better than the average for modern
standards!

And, to answer another question of yours, the reason that you can't
use exception handlers in many embedded programs is that you can't
guarantee to recover completely. As I say in another response,
that could be enabled, but it would be a major job and is not
necessarily worth attempting.


Regards,
Nick Maclaren.

--
[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]