… changed"

…op order

The drop order of let and let else is supposed to be the same,
and in order to ensure this, the test checks that this holds for
the given list of cases.

The test also ensures that we drop the temporaries of the
condition before executing the else block.

We made the test matrix based so it can check all the possible
combinations and find out possible edge cases.

Add let else drop order tests

Add a systematic matrix based test that checks temporary drop order in various settings, `let-else-drop-order.rs`, as requested [here](rust-lang#93628 (comment)).

The drop order of let and let else is supposed to be the and in order to ensure this, the test checks that this holds for a number of cases.

The test also ensures that we drop the temporaries of the condition before executing the else block.

cc rust-lang#87335 tracking issue for `let else`

Add a Machine hook for inline assembly

I'm sketching out some support in Miri to "execute" inline assembly. I want this because there are codebases which have very simple inline assembly like hand-written syscall wrappers, and it would be nice to test such code without modification.

r? ``@oli-obk``

…shtriplett

Fix cleanup for uninitialized stdout

Fixes rust-lang#101375 by disabling buffering even if the buffer was not initialized yet.

…mention-r-a, r=Mark-Simulacrum

Revert "Mention rust-analyzer maintainers when `proc_macro` bridge is changed"

Reverts rust-lang#99183

rust-analyzer is now a subtree, and CI fails when the `proc_macro` bridge changes break our tests, so these notifications aren't needed anymore.

… r=spastorino

Point at type parameter in plain path expr

Slightly better error message for a kinda unique use case.

Don't duplicate file descriptors into stdio fds

Ensures that file descriptors are never duplicated into the stdio fds even if a stdio fd has been closed.

…torino

Remove generics_def_id_map from the resolver.

This is internal state for lowering.  This does not belong in the resolver.

r? ``@spastorino``