We haven't had a __cpp_concepts since the TS specified its value as 201507 - which GCC defines. I've been making up values since then for various compilers which cmcstl2 supports to various degrees: * clang-concepts defines __cpp_concepts to 201707L - the date of the Toronto merge - to indicate support for concepts as merged. * MSVC uses 201811L - the date of the P1084 merge - to indicate that it supports the above plus P1084's changes to return-type-requirements. (Unfortunately MSVC does not yet support the changes in P1141 which were merged at the same meeting.)
I'd appreciate a bump for the changes in Cologne. P1452 seems significant enough to me, but regardless it would be nice to have a way to discriminate between what MSVC will support in 16.3 and C++20 concepts' final form. On Mon, Sep 16, 2019 at 8:26 AM John Spicer <[email protected]> wrote: > Yes, I think we should have one. > > John. > > On Sep 14, 2019, at 4:07 PM, Barry Revzin <[email protected]> wrote: > > On Sat, Sep 14, 2019, 2:38 PM Jason Merrill via Core < > [email protected]> wrote: > >> This doesn't seem to have been addressed at Cologne. I don't think >> any of the changes at Cologne (conditionally trivial, unconstrained >> TTP, dropping return-type requirements) require a version bump, does >> anyone else? >> >> So I'd stick with the 201811 value. >> >> Jason >> > > Wait, do we even have a __cpp_concepts? We have a __cpp_lib_concepts (... > that def needs to be bumped to 201907), but not one for the language. We > should add one... > > Barry > >> _______________________________________________ > Features mailing list > [email protected] > http://www.open-std.org/mailman/listinfo/features > > > _______________________________________________ > Features mailing list > [email protected] > http://www.open-std.org/mailman/listinfo/features >
_______________________________________________ Features mailing list [email protected] http://www.open-std.org/mailman/listinfo/features
