P2321R2 zip: Make tuple, pair, and vector<bool>::reference indirectly_writable

[cpplearner]

This implements the uninteresting parts of P2321R2, namely changes to tuple, pair and vector<bool>::reference to make them usable as proxy references.

Partially addresses #2252.

[strega-nil-ms]

Alright, I'm about halfway through the review; I'll comment and submit now and continue later.

• tuple
• type_traits
• vector

[strega-nil-ms]

The rest looks reasonable to me; the only concern I have is using enable_if when requires would work.

[StephanTLavavej]

Because EDG doesn't fully support concepts yet, we use enable_if instead of requires, even in C++20/23 mode, when the code in question doesn't have an inherent dependency on concepts. Thus, it is proper to use enable_if within pair.

The eventual cleanup of this is tracked by #602.

[strega-nil-ms]

Given that we can't use requires here, this looks good to me; on a side note, it feels very weird to spend like 4 hours reviewing something and still not have any comments.

[strega-nil-ms]

This looks good to me; I had Billy look over my shoulder at it as well.

The only concern I had was adding a call to GetFileInformationByHandleEx, but was told it was inexpensive (and I noticed we're doing two calls anyways to get both basic and standard info), so I'm no longer concerned.

[StephanTLavavej]

@strega-nil-ms I think your last comment was meant for #2373?

[StephanTLavavej]

Thanks, this is great! 😻 I saw no issues with the extensive template machinery, so I went ahead and pushed changes for the minor issues and nitpicks I noticed (the concepts guard and the test.lst being notable although small). FYI @strega-nil-ms as you had previously approved.

[StephanTLavavej]

⚠️ Note to self:

I need to export the namespace-scope machinery being added here.

[StephanTLavavej]

I'm mirroring this to the MSVC-internal repo - please notify me if any further changes are pushed.

[StephanTLavavej]

Thanks for implementing this foundational machinery for zip! 🤐 😹 🚀