Mark constructors that do not throw exceptions as noexcept

[AZero13 (AZero13)]

By ensuring we strength each move constructor with noexcept as we can, we can avoid needless exception checking and handling.

Had to open a new PR because GitHub wouldn't let me reopen the last one because it turns out editing commit history when the PR is closed is not good.

[Stephan T. Lavavej (StephanTLavavej)]

We talked about this at the weekly maintainer meeting and decided to close this PR without merging. We've repeatedly asked to avoid force-pushing changes, and multiple people have noted that many of these functions are physically allocating memory (e.g. hash_map is backed by list which dynamically allocates a sentinel node in all modes). We're sorry, but simply resubmitting the same PR (#3278, #3565) and gradually trimming down the list of edits as people raise issues, rather than verifying each carefully, is not a productive use of the maintainer team's very limited review capacity.

[AZero13 (AZero13)]

We talked about this at the weekly maintainer meeting and decided to close this PR without merging. We've repeatedly asked to avoid force-pushing changes, and multiple people have noted that many of these functions are physically allocating memory (e.g. hash_map is backed by list which dynamically allocates a sentinel node in all modes). We're sorry, but simply resubmitting the same PR (#3278, #3565) and gradually trimming down the list of edits as people raise issues, rather than verifying each carefully, is not a productive use of the maintainer team's very limited review capacity.

That isn't what happened. What happened was I closed the original PR to do some research on the code I was modifying and then made those changes. Once those changes were made, GitHub would not allow me to reopen. I should have marked it as a draft. There is nothing I am trying to hide. It was an honest Git mistake.