[libcxx-commits] [PATCH] D117608: [libcxx][test] view_interface need not derive from view_base

[Casey Carter via Phabricator via libcxx-commits]

CaseyCarter added inline comments.


================
Comment at: libcxx/test/std/ranges/range.utility/view.interface/view.interface.pass.cpp:35
 
-static_assert(std::derived_from<std::ranges::view_interface<Empty>, std::ranges::view_base>);
+LIBCPP_STATIC_ASSERT(std::derived_from<std::ranges::view_interface<Empty>, std::ranges::view_base>);
 
----------------
Quuxplusone wrote:
> jloser wrote:
> > Quuxplusone wrote:
> > > This is late-breaking C++20, right?
> > > If so, just delete this line (and the following blank line): there's no point testing something that we (1) shouldn't do and (2) presumably plan to change.
> > I'd prefer one of two different options:
> > 1) Make this `LIBCPP_STATIC_ASSERT` as the current `view_interface` class template does indeed inherit from `view_base`.
> > 2) Implement LWG3549 and make it so that `view_interface` class template no longer derives from `view_base` and then kill this line in the test as you've done.
> > 
> > Any appetite for implementing LWG3549? If not, I volunteer.
> > Any appetite for implementing LWG3549? If not, I volunteer.
> 
> I suspect I speak for us all when I say: go for it! :)
> (I still don't think that ought to block Casey's landing //this// one-line PR, though, so I'm being the second libc++ accepter.)
I'm not interested in implementing LWG3549 (for a second time =)), I'm only motivated to have the test not apply non-standard requirements when `stdlib=msvc`. How about if we split the work: I'll remove this test line, and you can implement LWG3549?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D117608/new/

https://reviews.llvm.org/D117608