RE: [Boost-Users] boost tokenizer

From: Jim.Hyslop (jim.hyslop_at_[hidden])
Date: 2002-01-17 09:54:04

Peter Dimov [SMTP:pdimov_at_[hidden]] wrote:
> From: "hicks" <hicks_at_[hidden]>
> > Do you know why the compiler places precedence on that interpretation,
> > rather than on one giving an  instatiation of a local variable?
> 
> The C++ standard says that a function declaration takes precedence. Don't
> ask me why. :-)
(At the risk of sliding seriously off topic...)

I suspect it's something along these lines. Given this function:

void f()
{
   int WhatIsThis();
}

Is "WhatIsThis" intended to be a declaration of a function, or is it
intended to be a definition and default-initialization of an int variable? A
C++ compiler cannot know, so the C++Standards committee had several choices:

1) require the compiler to flag the statement as an ambiguous statement
2) require the compiler to interpret the statement as a declaration of a
function
3) require the compiler to interpret the statement as a definition of an int

If we examine the same syntax in C, there is no question what it is:
'WhatIsThis' is a declaration of a function. That effectively eliminates #3
from consideration because you don't want one statement to have different
meanings in C and in C++.

I suspect the Committee then either flipped a coin (or some used other
democratic means, such as debating and voting on the issue, thumb-wrestling,
or playing paper-rock-scissors) to decide between #1 and #2, or choose to
adapt the well-worn guideline "When in doubt, do as the ints do" to "When in
doubt, do as C does".

-- 
Jim