From: daniel.kruegler@googlemail.com
On 2013-01-21 04:16, Frank Birbacher wrote:
> Am 19.01.13 16:22, schrieb Daniel Krügler:
>> Theoretically this would be possible. I'm not so sure that this is
>> really a sufficient reason to extend the rules in that way,
>> though.
>
> Sometimes I wish I could move-initialize the captured values,
Agreed. This is one of the major complaints currently (following the
request for polymorphic lambdas) in regard to constraints on lambda
expressions. To my knowledge there is work on this to solve the problem
but I cannot find a proposal at the very moment.
> especially when working with std::function like in:
>
> typedef std::function Callback;
>
> Callback twice(Callback c)
> {
> //will copy "c", but I want to move it into the lambda:
> return [c](std::string const& s){
> c(s); c(s); // call twice
> };
> }
>
> The std::function does not do reference counting, does it? QOI?
I'm not aware of one (but I haven't looked into many). There is some
evidence that a reference-counted implementation would not be
conforming, because the difference would be user-observable in a way
that would point to a conflict with the wording. In [func.wrap.func.con]
p4 we have for the copy constructor:
function(const function& f);
[..]
4 Postconditions: !*this if !f; otherwise, *this targets a copy of
f.target().
and [func.wrap.func.targ] p3 provides a means to observe the address of
the contained target:
template T* target() noexcept;
template const T* target() const noexcept;
[..]
3 Returns: If target_type() == typeid(T) a pointer to the stored
function target; otherwise a null pointer.
The wording for std::function is intended to support the
short-object-optimization, thought. There are several places that point
to special guarantees in regard to function pointers and
reference_wrapper objects, such as
"shall not throw exceptions if f’s target is a callable object passed
via reference_wrapper or a function pointer."
and especially the note
[ Note: Implementations are encouraged to avoid the use of dynamically
allocated memory for small callable objects, for example, where f’s
target is an object holding only a pointer or reference to an object and
a member function pointer. —end note ]
I just notice that there are currently no no-throw-guarantees for
targets that are pointer to members, this looks like an oversight to me
and should be changed.
HTH & 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! ]
--- SoupGate-Win32 v1.05
* Origin: you cannot sedate... all the things you hate (1:229/2)
|
