Implement resolution of LWG-3449 for take_view

[Michael Schellenberger Costa (miscco)]

Addresses #1446

Fixes #1463.

[Casey Carter (CaseyCarter)]

Michael Schellenberger Costa (@miscco) Stephan T. Lavavej (@StephanTLavavej) FYI I pushed a commit to change the names _Counted_Iter and _Base_Ty to be consistent with our naming conventions (snakes and camels don't mix) and to correct the constraint on the new operator== overload.

[Stephan T. Lavavej (StephanTLavavej)]

Michael Schellenberger Costa (@miscco) Casey Carter (@CaseyCarter) Does this indicate that we're missing operator== test coverage?

[Casey Carter (CaseyCarter)]

Michael Schellenberger Costa (@miscco) Casey Carter (@CaseyCarter) Does this indicate that we're missing operator== test coverage?

Yes: the take_view test passes with the new operator== commented out. We need to investigate. (Maybe our test ranges all model simple-view?)

[Casey Carter (CaseyCarter)]

Maybe our test ranges all model simple-view?

Why yes, yes they do.

[Michael Schellenberger Costa (miscco)]

I would suggest that we merge this and then fix the simple_view issue globally.

I believe there will be sufficient breakage coming from it for many PRs

[Casey Carter (CaseyCarter)]

I would suggest that we merge this and then fix the simple_view issue globally.

I believe there will be sufficient breakage coming from it for many PRs

I agree; I've updated #1409 to include addressing this problem in all of the view tests. I'll move this to Final Review and we'll see if Stephan T. Lavavej (@StephanTLavavej) agrees with that strategy.

[Stephan T. Lavavej (StephanTLavavej)]

That strategy sounds good to me, thanks. I've pushed a merge to test this with 16.9 Preview 2 (no conflicts).

[Stephan T. Lavavej (StephanTLavavej)]

Thanks for implementing this LWG issue resolution and increasing the crystalline perfection of ranges! 💎 😸