Constructs a TerminatorKind::SwitchInt for an equivalent conditional true-false branch.

This ends up not really mattering because we generate a plain conditional branch in LLVM either
way.

Previously it used to build a switch in a way that didn’t preserve the invariat of SwitchInt. Now
it builds it in an optimal way too, where otherwise branch becomes all the branches which did not
have partial variant drops.

First example of optimisation that applies to many more cases than originally.

Also use a Cow to avoid full Vec for all SwitchInts

This removes another special case of Switch by replacing it with the more general SwitchInt. While
this is more clunky currently, there’s no reason we can’t make it nice (and efficient) to use.

Previously AdtDef variants contained ConstInt for each discriminant, which did not really reflect
the actual type of the discriminants. Moving the type into AdtDef allows to easily put the type
into metadata and also saves bytes from ConstVal overhead for each discriminant.

Also arguably the code is cleaner now :)

Update set operations documentation

Reminding people of set terminology.

Rollup of 6 pull requests

- Successful merges: #39587, #39674, #39693, #39700, #39705, #39707
- Failed merges:

change span_notes to notes in E0368/E0369

Fixes #39650.

All the uses of `span_note` in these errors were reusing the same span as the original error, which causes unnecessary repetition.

For an example, see the changes to [src/test/ui/span/issue-39018.stderr](https://github.com/rust-lang/rust/pull/39707/files?diff=unified#diff-46336f62958fdb34233db414cb9914a1R4).

r? @jonathandturner

name anonymous fn parameters in libcore traits

This follows the discussion in rust-lang/rfcs#1685. The patch gives names to anonymous parameters in libcore traits. It would have two benefits I can think of: firstly it would provide names to tools that can use the names from the traits, and secondly core/std can serve as an example when writing traits; this change helps by not encouraging the use of anonymous parameters.

Adding compile fail test for const_indexing feature

First attempt at contributing to rust. Picked up an easy feature to test.

Issue #39059
r? @est31

driver: restore partially deleted comment

Fixes #39689. Comment was lost in commit 811b8747 (#31916 @nagisa).

parser: fix ICE when parsing token trees after an error

Fixes #39388, fixes #39616.
r? @nrc

rustbuild: support setting verbosity in config.toml

Most if not all the configuration is settable trhough config.toml but the verbosity isn't yet.

This avoids having to pass -v to x.py on each command if you want verbosity to be always on.

Reminding people of set terminology.

Rollup of 6 pull requests

- Successful merges: #39604, #39619, #39670, #39678, #39682, #39683
- Failed merges:

Fix unaligned load in librustc_metadata::index.

The derived `Clone` impl contains UB and will be unsafe when we fix https://github.com/rust-lang/rust/issues/27060. See [this comment](https://github.com/rust-lang/rust/issues/27060#issuecomment-278617096) for more context.

r? @bluss

Fix unsafe unaligned loads in test.

r? @EddyB
cc @Aatch @nikomatsakis

The `#[derive(PartialEq, Debug)]` impls on a packed struct contain undefined behaviour. Both generated impls take references to unaligned fields, which will fail to compile once we correctly treat that as unsafe (see https://github.com/rust-lang/rust/issues/27060).

This UB was found by running the test under [Miri](https://github.com/solson/miri/) which rejects these unsafe unaligned loads. 😄

Here's a simpler example:

```rust
struct Packed {
    a: u8,
    b: u64,
}
```

It expands to:

```rust
    fn fmt(&self, __arg_0: &mut ::std::fmt::Formatter) -> ::std::fmt::Result {
        match *self {
            Packed { a: ref __self_0_0, b: ref __self_0_1 } => { // BAD: these patterns are unsafe
                let mut builder = __arg_0.debug_struct("Packed");
                let _ = builder.field("a", &&(*__self_0_0));
                let _ = builder.field("b", &&(*__self_0_1));
                builder.finish()
            }
        }
    }
```

and

```rust
    fn eq(&self, __arg_0: &Packed) -> bool {
        match *__arg_0 {
            Packed { a: ref __self_1_0, b: ref __self_1_1 } => // BAD: these patterns are unsafe
            match *self {
                Packed { a: ref __self_0_0, b: ref __self_0_1 } => // BAD: these patterns are unsafe
                true && (*__self_0_0) == (*__self_1_0) &&
                    (*__self_0_1) == (*__self_1_1),
            },
        }
    }
```

Exclude top-level macro expansions from source location override.

It occurred to me that a simple heuristic can address the issue #36382: any macros that expand into items (including `include!()`) don't need to be stepped over because there's not code to step through above a function scope level.

r? @michaelwoerister

Fix mailmap for @gifnksm

cc rust-lang-nursery/thanks#71

Choose different name for metadata obj-file to avoid clashes with user-chosen names.

Fixes #39585 and probably https://github.com/rust-lang/rust/issues/39508.
Incremental compilation assigns different names to obj-files than regular compilation. If a crate is called "metadata" this can lead to a clash between the root module's obj-file and the obj-file containing crate-metadata. This PR assigns a name to the metadata obj-file that cannot clash with other obj-file because it contains a `.` which is not allowed in a Rust module identifier.

r? @alexcrichton

cc @nikomatsakis

Unignore u128 test for stage 0,1

Even more SNAP cleanup.

Follow-up of #39519.

Sorry, I didn't check twice.

Emit DW_AT_main_subprogram

This changes rustc to emit DW_AT_main_subprogram on the "main" program.
This lets gdb suitably stop at the user's main in response to
"start" (rather than the library's main, which is what happens
currently).

Fixes #32620
r? michaelwoerister

Rollup of 5 pull requests

- Successful merges: #39595, #39601, #39602, #39615, #39647
- Failed merges:

build std for sparc64-unknown-linux-gnu

r? @alexcrichton

panicking / unwinding is broken (#39646) but with std available at least people
will be able to debug that issue on real hardware

Improve format float

* Move float into mod float like in test
* Add more tests for f64 f32, lower exp, upper exp, which can come if handy in the future if we want refactor further
* Use `assert_eq` for clearer error messages

Fix ICE when accessing mutably an immutable enum

Fix #39544.