Support TfHashAppend leveraging of std::hash for unspecialized types

[nvmkuruc]

Description of Change(s)

A C++17 friendly implementation of leveraging std::hash in TfHashAppend had been left commented out. During the removal of boost::hash, specializations were added (#2298, #2304) for some types that explicitly forwarded the result of std::hash as their TfHashAppend implementation.

Now that OpenUSD requires C++17, TfHashAppend can hash any type that supports std::hash and those specializations can be removed. To avoid stdext::hash_value from being prioritized over std::hash on some platforms, the lookup order is reordered so that std::hash is second and hash_value is last.

This also removes SFINAE for the std::vector specialization in favor of a static_assert guard against using AppendContiguous on std::vector<bool>.

Fixes Issue(s)

• I have verified that all unit tests pass with the proposed changes

• I have submitted a signed Contributor License Agreement

[jesschimein]

Filed as internal issue #USD-8911