Closes #15506.

prefer `Deref` over `DerefMut` in all other circumstances.

Because the compiler now prefers `Deref`, this can break code that
looked like:

    let mut foo = bar.borrow_mut();
    (*foo).call_something_that_requires_mutable_self();

Replace this code with:

    let mut foo = bar.baz();
    (&mut *foo).call_something_that_requires_mutable_self();

Closes #12825.

[breaking-change]

r? @nikomatsakis

This bound is already implicit through the AnyPrivate trait,
but since it is not explicit, you still have to write Box<Any + 'static>,
even though Any can only be 'static.

Introducing the 'static bound here makes this bound explicit, making
Box<Any> legal.

This breaks code that looks like:

    match foo {
        1..3 => { ... }
    }

Instead, write:

    match foo {
        1...3 => { ... }
    }

Closes #17295.

r? @nick29581 

This bound is already implicit through the AnyPrivate trait,
but since it is not explicit, you still have to write Box<Any + 'static>,
even though Any can only be 'static.

Introducing the 'static bound here makes this bound explicit, making
Box<Any> legal.

Continuation of https://github.com/rust-lang/rust/pull/16741.

prefer `Deref` over `DerefMut` in all other circumstances.

Closes #12825.

closes #17478

I swear this is the last step. p=1 please so i can get the bots changed over before the test suite regresses again.

It turns out that adding new diagnostics is causing link failures
in runpass-full-deps tests. Further investigation pending.

Modify ast::ExprMatch to include a new value of type ast::MatchSource,
making it easy to tell whether the match was written literally or
produced via desugaring. This allows us to customize error messages
appropriately.

I don't know exactly what made this regress so badly…maybe it was all the lifetimes that @EddyB landed. According to `git blame` this is all somewhat old code. Regardless this is an enormous improvement in compile times.

r? @brson 

This breaks code that looks like:

    match foo {
        1..3 => { ... }
    }

Instead, write:

    match foo {
        1...3 => { ... }
    }

Closes #17295.

[breaking-change]

4x improvement in pre-trans compile time for rustc.

This makes the windows `make dist` target start producing binary tarballs, and tweaks install.sh so they work, in preparation for working on a combined Rust+Cargo installer.

This is the bare minimum to stop using split stacks on Windows, fixing https://github.com/rust-lang/rust/issues/13259 and #14742, by turning on stack probes for all functions and disabling compiler and runtime support for split stacks on Windows.

It does not restore the out-of-stack error message, which requires more runtime work.

This includes a test that the Windows TCB is no longer being clobbered, but the out-of-stack test itself is pretty weak, only testing that the program exits abnormally, not that it isn't writing to bogus memory, so I haven't truly verified that this is providing the safety we claim.

A more complete solution is in https://github.com/rust-lang/rust/pull/16388, which has some unresolved issues yet.

cc @Zoxc @klutzy @vadimcn

closes #17392 

r? @nikomatsakis 

Kind of a first draft because I'm not sure if this is the right approach. I believe the general idea of giving an error rather than an ICE in obligation_for_builtin_bound is right, but not sure about returning an Option, etc.

Also, could probably have a better error message.

closes #17392

Makes it work on windows

We may use these for creating combined rust/cargo installers

Add deallocate calls to the realloc-16687.rs test.

Fix #17303.

closes #17478