perf: mark move constructors and assignment operators as noexcept

[eric-wieser]

Detected by https://clang.llvm.org/extra/clang-tidy/checks/performance/noexcept-move-constructor.html.
This ensures constructions like std::vector<object_ref> call these operators instead of the copy ones, and do not do extra refcounting.

Note that optional and atomic need something more complex using noexcept(), as they are templated.

[leanprover-community-bot]

Mathlib CI status (docs):

• ❗ Batteries/Mathlib CI will not be attempted unless your PR branches off the nightly-with-mathlib branch. Try git rebase 4077bf2c05b74a2eaaf1db9be957236f29614f63 --onto 14ff08db6f651775ead432d367b6b083878bb0f9. You can force Mathlib CI using the force-mathlib-ci label. (2025-10-15 10:29:07)
• ❗ Batteries/Mathlib CI will not be attempted unless your PR branches off the nightly-with-mathlib branch. Try git rebase 4077bf2c05b74a2eaaf1db9be957236f29614f63 --onto efbbb0b230ce95653d25b59c83fd24a51a8bf363. You can force Mathlib CI using the force-mathlib-ci label. (2025-10-21 14:18:08)

[leanprover-bot]

Reference manual CI status:

• ❗ Reference manual CI will not be attempted unless your PR branches off the nightly-with-manual branch. Try git rebase 4077bf2c05b74a2eaaf1db9be957236f29614f63 --onto d3dda9f6d4428a906c096067ecb75e432afc4615. You can force reference manual CI using the force-manual-ci label. (2025-10-15 10:29:09)

[Kha]

@eric-wieser Is the stage0 change an oversight? We can't merge it the standard way unless it is reverted

[eric-wieser]

It was indeed an oversight, now reverted.

[eric-wieser]

Note that this and #10815 conflict in a way that git won't pick up; if either is merged, the other will need a small patch to be correct.

[Kha]

@eric-wieser Ok, feel free to update the other PR afterwards. Btw, I suspect a good amount of this code is dead; if you have a tool for that as well, let me know!