On 2015-10-14 14:26, Bob Hood wrote:
> On 10/14/2015 12:04 PM, Elvis Stansvik wrote:
>> I believe it is because QAbstractSocket::error is an overloaded
>> function, so it is ambiguous in your connect/disconnect calls. It
>> needs to be statically casted :/ See e.g:
>>
>> https://wiki.qt.io/New_Signal_Slot_Syntax#Overload
>
> Ah, that would be the problem, I guess. Thanks, Elvis. :)
>
> That would also explain why the old-style connects work.I can go back to that
> for this particular case. Assuming this is addressable, does anybody know if
> there are there plans to correct this new-style issue in a future release, or
> should I just plan on using old-style connects in these instances?
There's nothing to be "corrected"; if a signal or slot is overloaded,
you have to specify which one you mean. It's just not something that can
be avoided in the general case¹. Old style does that by giving the full
signature. With new style (and a C++11 compiler), you can use this:
template<class... Args, class T, class R>
auto qtOverload(R (T::*m)(Args...)) -> decltype(m)
{ return m; }
...to connect like so:
connect(mapper, qtOverload<int>(&QSignalMapper::mapped), ...);
(That is, in the <>'s you just need to give the list of parameter types.)
(¹ I seem to recall talk of making QAbstractSocket::error specifically
not overloaded, which would help your exact use case. In general,
however, the only other "solution" is to not ever overload signals or
slots.)
--
Matthew
_______________________________________________
Interest mailing list
Interest@qt-project.org
http://lists.qt-project.org/mailman/listinfo/interest
