[boost] Re: geometry classes code [not so long ;)]

From: John Potter (jpotter_at_[hidden])
Date: 2000-01-02 18:22:01

Please excuse me for butting in.
 
On Sun, 2 Jan 2000, Dave Abrahams wrote:
 
> > What do you think about a chaining of member function calls in general? $
> > never heard anything about this subject from any person who can made me $
> > my mind, so I'll glad to hear something about it from members of this gr$
>
> It's unfamiliar to me at least. There is a reason that it is often
> recommended that operator+= should return a const reference instead of a
> full reference: for compatibility with the built-in types. IOW, you can't
> write: int x = 0; (x += 3)++

You can if you add a semicolon ;)  The built-in types produce lvalues.
This is undefined behavior of the form that I have never seen any
compiler do anything other than expected.  With my limited experience,
I see no reason to do anything else.  Chaining is also quite common.
You're just used to a different form.
 
    int x = 0;
    (x += 3)++;
    cout.operator<<(x).operator<<(endl);
 
The assignable requirements for container elements require a T&,
but I know of no place where a T const& would not work.  No desire
to open an old debate.  If it is policy, that is sufficient.
 
John