[libcxx-commits] [PATCH] D107584: [libc++][modularisation] Split up <concepts> into granular headers.

Christopher Di Bella via Phabricator via libcxx-commits libcxx-commits at lists.llvm.org
Thu Aug 5 21:10:16 PDT 2021


cjdb added inline comments.


================
Comment at: libcxx/include/__concepts/invocable.h:34-35
+
+template<class _Fn, class... _Args>
+concept regular_invocable = invocable<_Fn, _Args...>;
+
----------------
Quuxplusone wrote:
> cjdb wrote:
> > Quuxplusone wrote:
> > > cjdb wrote:
> > > > It's not immediately clear that `regular_invocable` lives in `invocable.h`. Please either:
> > > > 
> > > > * split the file up further
> > > > * consolidate with the other callable concepts and call the header `callable.h`
> > > > 
> > > > I'm indifferent to the chosen approach, but please be consistent in the approach you choose.
> > > I prefer not to split `invocable` from `regular_invocable`, because they are (as far as libc++ is concerned) equivalent.
> > > However, it's good to split this concept from `predicate` and `relation` (and the other concepts equivalent to `relation`) because //this// file doesn't depend on `__boolean_testable` but //those// concepts do.
> > :shrug: I guess the same argument that I made for `equality_comparable.h` vs `equality_comparable_with.h` could be made here re not splitting further. Though, I'm not sure why you're opposed to `__boolean_testable` being in this header. Ditto for `object.h`, but we should keep the conversation in one place.
> `__boolean_testable` is an ancestor of both `equality_comparable` and `predicate`.
Yes, but `__boolean_testable` has its own header. I'm trying to work out why you're opposed to combining the all the callables into a single header, and you pointed at `__boolean_testable`.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D107584/new/

https://reviews.llvm.org/D107584



More information about the libcxx-commits mailing list