Re: move constructors likely a requirement

[John W. Eaton]

On 9/4/19 8:32 AM, Carlo De Falco wrote:
> > Il giorno 4 set 2019, alle ore 00:39, John W. Eaton <address@hidden> ha scritto:
> >
> > I pushed the following changes:
> >
> >   http://hg.savannah.gnu.org/hgweb/octave/rev/5bf76ab4cce3
> >   http://hg.savannah.gnu.org/hgweb/octave/rev/a2d3fa82b730
> >   http://hg.savannah.gnu.org/hgweb/octave/rev/fcaecdbc8d8a
>
> Hi,
>
> Since this changesets, when compiling with clang++ I see lot of warnings of the 
> form :
>
> octave/libinterp/parse-tree/pt-colon.cc:67:41: warning: moving a temporary 
> object prevents copy elision
>        [-Wpessimizing-move]
>              octave_value ov_increment = std::move (m_increment->evaluate (tw));
>
> AFAIU this warning means that, in this case, automatic return value 
> optimization performed by the compiler
> would result in code that is more efficient than that produced using move 
> constructor.
>
> I'm just reporting this in case it may be of interest, bu I don't know whether 
> it is worth making any changes to
> remove the warning.

It is of interest.  I'd like to avoid std::move when it isn't needed. 
Could you send a log of all the warnings?  I'm not seeing them with GCC. 
 If I understand correctly, -Wpessimizing-move is supposed to be 
implied when compiling C++ code with GCC and using -Wall.

jwe