On 12 Mai, 00:40, Pete Becker <p...@[EMAIL PROTECTED]
> wrote:
> On 2008-05-10 10:07:54 -0400, marlow.and...@[EMAIL PROTECTED]
said:
> > Er, that doesn't sound like an extension to me. It sounds like a bug.
> > Allowing stuff to compile that shouldn't compile an' all that. Unless,
> > of course, it is do***ented in which case it's a feature.
>
> Note that the C++ standard does not have a notion of "shouldn't
> compile". The requirement is that for ill-formed code the compiler must
> issue a diagnostic. Once it's done that, the standard does not impose
> any requirements. That's the standard's mechanism for conforming
> extensions: issue a diagnostic, then do whatever your extension is.
As I read the definition of "ill-formed" (as the negation of
well-formed), not all ill-formed programs need to (or can) be
diagnosed. This is so, because well-formedness also requires
fulfillment of the One Definition Rule, and not all parts of
the ODR need to be diagnosed (e.g. duplicate, but not "equivalent"
definitions of an entity in different translation units).
Greetings from Bremen,
Daniel Krügler
--
[ See http://www.gotw.ca/resources/clcm.htm
for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]
|
