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

On Thu, 28 Sep 2017 23:43:47 +0900
Kohei Takahashi <[email protected]> wrote:

> Hi all,
> 
> I propose deprecating fusion.tuple, it makes fusion.vector
> implementation complex.
> 
> fusion.tuple provides TR1 style std::tuple (experimental)
> implementation but we already have truly std::tuple, and some other
> TR1-ism feature, especially Boost.TR1, had been removed or deprecated.
> 
> I did grep and got file list which include fusion.tuple:
> 
>     hana/benchmark/including/fusion.erb.cpp
>     hana/example/ext/boost/fusion/tuple.cpp
>     hana/include/boost/hana/ext/boost/fusion/tuple.hpp
>     hana/include/boost/hana/ext/boost/fusion.hpp
>     hana/test/_include/auto/README.md
>     hana/test/ext/boost/fusion/tuple/auto/_specs.hpp
>     hana/test/ext/boost/fusion/tuple.cpp
>     math/include/boost/math/tools/tuple.hpp
>     math/test/test_root_finding_concepts.cpp
>     phoenix/test/scope/let_tests.cpp
>     phoenix/test/scope/let_tests_113.cpp
>     phoenix/test/scope/let_tests_113a.cpp
>     phoenix/test/scope/let_tests_157.cpp
>     phoenix/test/scope/let_tests_157a.cpp
>     phoenix/test/scope/let_tests_rest.cpp
>     phoenix/test/scope/more_let_tests.cpp
>     phoenix/test/scope/more_let_tests1.cpp
>     phoenix/test/scope/more_let_tests2.cpp
>     phoenix/test/scope/more_let_tests2a.cpp
>     phoenix/test/scope/more_let_tests2b.cpp
>     process/include/boost/process/detail/execute_impl.hpp
>     proto/example/mini_lambda.cpp
>     proto/test/examples.cpp
>     proto/test/lambda.cpp
>     proto/test/make_expr.cpp
>     spirit/test/qi/regression_fusion_proto_spirit.cpp
>     spirit/test/qi/regression_transform_assignment.cpp
> 
> I didn't check how they used it, but it seems only testing or
> adapting. So, I think the deprecation is not so big deal.
> 
> Any thought?
> 
> Thanks,
> Kohei Takahashi
> 
> 

How does `fusion::tuple` make `fusion::vector` implementation more
complex? Are you referring to the C++03 or C++11 version?

For C++11 `fusion::tuple` I noticed that if the assignment from fusion
sequences was removed (its constructor currently does not allow
conversion from sequences and `std::tuple` has no such requirement),
then the headers automatically included could be reduced. So
`fusion::tuple` _could_ be a faster compiling sequence for when the
conversion properties of `fusion::vector` was not needed. Of course,
since most have a good implementation of `std::tuple` perhaps this
unnecessary? I would like to point out that recent tweaks to
`fusion::vector` have [improved performance over libstdc++ `std::tuple`
in construction][0][1].

Anyway, the bigger issue with deprecation will be on users of Boost.
Hopefully the fix would be easy - just change to another sequence type.
I'm not aware of an situations where this is _not true_ (i.e. where
deprecation would cause major code re-write for Boost users).

Lee

[0] http://metaben.ch/hetero/clang++-4.0/make/index.html
[1] http://metaben.ch/hetero/g++-7/make/index.html

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