Re: [gil] Phasing out support C++11 and switching to C++17

From: Emil Dotchevski (emildotchevski_at_[hidden])
Date: 2020-10-26 19:38:37

On Mon, Oct 26, 2020 at 8:38 AM Mateusz Loskot via Boost <
boost_at_[hidden]> wrote:
>
> On Mon, 26 Oct 2020 at 15:53, Robert Ramey via Boost
> <boost_at_[hidden]> wrote:
> > On 10/26/20 2:37 AM, Mateusz Loskot via Boost wrote:
> > >
> > > We're discussing [1] to start switching GIL to C++17
> > > in the near future and, following Peter Dimov's policy
> > > on phasing out C++XY support [2], we'd like to:
> > >
> > > 1. Declare C++11 support deprecated in 1.75 (December 2020)
> > > 2. Drop C++11 support in 1.77 (?)  (August 2021)
> > >
> > > Could anyone help us validate and confirm that
> > > it's a good plan or are we overlooking any policies
> > > from Boost's release perspective?
> > >
> > > [1] https://lists.boost.org/boost-gil/2020/10/0465.php
> > > [2] https://pdimov.github.io/articles/phasing_out_cxx03.html
> >
> > Just out of curiosity, what new features and/or benefits (if any) will
> > moving to conformance with C++17 add?
>
> e.g. make the library sexier for contributors who actually
> expressed interest in developing features for the library
> and who wish to use `-std=c++17` at least
> when they develop pull requests for the library.

Really, the stated reason for dropping support for C++11 is that it will
make the library "sexier"?

I understand that this would enable GIL to be compiled as a standalone
library, and that's a good thing, but it would be better to provide that
functionality and not break what currently works. Or at least take a stab
at it and see how bad it is to support both. After all, Boost provides
excellent support for workarounds and compiler compatibility.