Re: [boost] Improving the assignment operators of various Boost types

Subject: Re: [boost] Improving the assignment operators of various Boost types
From: Niels Dekker - mail address until 2008-12-31 (nd_mail_address_valid_until_2008-12-31_at_[hidden])
Date: 2008-09-11 13:01:52

David Abrahams wrote:
> Niels: are you planning to pursue tickets for the classes with
> implicit assignment operators and the others I mentioned in
> response to your original post?

Honestly I haven't yet looked into those. But my idea was to start off 
by submitting tickets to get that "low-hanging fruit", for those cases 
that just need a few small changes in their existing assignment 
operators.  I'd rather wait and see if those simple cases are going well 
(especially at the regression site). But feel free to create tickets for 
those implicit assignment operators if you can't wait any longer.  :-)

I've just submitted a ticket to Douglas, for boost::function: 
https://svn.boost.org/trac/boost/ticket/2319  boost::function's copy 
assignment operator is kinda interesting, because it currently does 
copy-and-swap, while it still doesn't provide the strong guarantee.  The 
"function_assignment.patch" that I attached to #2319 should 
significantly reduce the chance of having the assignment fail, but it 
still won't provide the strong guarantee. I'm afraid that's the price of 
small-object optimization.

Do you think that it would be okay in C++0X to have operator= overloaded 
for by-value (copy-assignment) and by-rvalue-reference 
(move-assignment)?  Or would it be preferable in C++0X to have a 
const-reference as copy-assignment argument?

Kind regards,

  Niels