Ok, so I'm wrong. Stroustrup gives such an example on p. 83 {Sect 4.9.4, 3rd
ed} and says "int x = x;" is not illegal, just silly. He does, however,
state that "a good compiler will warn if a variable is used before it has
been set" so I suppose this means he [Stroustup] must not consider that GCC
version a "good compiler." :)

I would've thought the scope started following the assignment, but that's
not the case. I can't imaging a useful reason for allowing an uninitialzied
reference, but I didn't write it so maybe there is something I've
overlooked. Still...

Ciao,
Dee

-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] Behalf Of Paul Grenyer
Sent: Monday, March 22, 2004 2:58 PM
To: [EMAIL PROTECTED]
Subject: Re: [OT] gcc & QT (was Re: RE: [msvc] using bit fields)


Hi

> I maintain that the compiler should have, at the very least, supplied a
> warning message regarding the use of an uninitialized object reference.
For
> that matter, as I understand the standard, the 'cname' variable should not
> even exist until after the semicolon on that line (i.e. should not be
> considered in scope and referenceable).

Is that in the standard?

If not, can I recommend joining your local standards committee and trying to
change it?

I find it quite rewarding.

Regards
Paul

Paul Grenyer
Email: [EMAIL PROTECTED]
Web: http://www.paulgrenyer.co.uk

Have you met Aeryn: http://www.paulgrenyer.co.uk/aeryn/?
Version 0.3.0 beta now ready for download


_______________________________________________
msvc mailing list
[EMAIL PROTECTED]
See http://beginthread.com/mailman/listinfo/msvc_beginthread.com for
subscription changes, and list archive.



Reply via email to