Re: [fusion] Proposal for deprecating fusion.tuple.

On Fri, 29 Sep 2017 11:18:35 -0700
chris beck <[email protected]> wrote:

> > I propose deprecating fusion.tuple, it makes fusion.vector
> > implementation complex.  
> 
> If fusion::tuple is deprecated in favor of std::tuple, would this
> also mean that boost::fusion would require C++11-conforming standard
> library to compile?

No, I _think_ what the OP noticed was that users of `fusion::tuple`
could move to `fusion::vector`. There are some subtle differences, but
the difference is unlikely to matter in most (all?) cases. This could
save on some maintenance work in Fusion.

> Right now, unless I'm mistaken, fusion works fine even on e.g. gcc
> 4.5, and there are many C++03 etc. projects that depend on it.
> 
> What is the gain that you hope to achieve here, improved compilation
> times? How can moving to std::tuple accomplish that, since presumably
> the implementation quality of std::tuple will vary greatly from
> compiler to compiler? Wouldn't it be simpler to just try to make
> fusion::tuple faster in the perceived problematic cases?

The OP was primarily concerned about maintaining `fusion::tuple`. I
unfortunately injected a performance angle, and hopefully my next
response to Kohei will make it more clear why.

> > Hence, why not replace fusion::tuple with hana tuple?  
> 
> And hana is also far more demanding than fusion in terms of compiler
> support. For instance it doesn't work at all on MSVC yet, and so is
> totally inappropriate for some cross-platform projects.
> 

Agreed, `hana::tuple` is not appropriate for many situations since it
requires a conforming C++14 compiler which is not always available. I
do not expect the removal of a working C++03 `fusion::vector`.

Lee

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot