Union fields have the ParamEnv of the union.

* Handle cycles in `needs_drop` correctly
* Normalize types when computing `needs_drop`
* Move queries from rustc to rustc_ty

* Handle arrays with const-generic lengths
* Use closure for repeated code.

Currently, every iteration of the main loop in `generic_extension`
instantiates a `Parser`, which is expensive because `Parser` is a large
type. Many of those instantiations are only used immutably, particularly
for simple-but-repetitive macros of the sort seen in `html5ever` and PR
68836.

This commit initializes a single "base" parser outside the loop, and
then uses `Cow` to avoid cloning it except for the mutating iterations.
This speeds up `html5ever` runs by up to 15%.

The previous commit wrapped `Parser` within a `Cow` for the hot macro
parsing path. As a result, there's no need for the `Cow` within
`Directory`, because it lies within `Parser`.

This is a small win, because `Failure` is much more common than
`Success`.

Co-Authored-By: varkor <[email protected]>

This makes it more ergonomic to create a dataflow engine and obviates
the need to pick between `new_gen_kill` and `new_generic`.

Co-Authored-By: Mazdak Farrokhzad <[email protected]>
Co-Authored-By: Tim Robinson <[email protected]>
Co-Authored-By: Peter Todd <[email protected]>
Co-Authored-By: Dylan DPC <[email protected]>

Added dyn and true keyword docs

r? @Centril

…ls, r=pnkfelix

Migrate borrowck dataflow impls to new framework

This uses rust-lang#65672 to implement the dataflow analyses needed by borrowck. These include all the `InitializedPlaces` analyses as well as `Borrows`. Additionally, this PR contains several independent changes around the dataflow API which improve performance and make it more ergonomic.

* An optimization that inhibits the caching of block transfer functions for acyclic MIR (~0.3% I-CNT savings).
* A `ResultsVisitor` for dataflow results that is more efficient than `ResultsCursor` when we have to visit every statement unconditionally (~0.3% I-CNT savings).
* An `into_engine` method on `Analysis` that selects the appropriate `Engine` constructor.
* A `contains` method for `ResultsCursor` as a shorthand for `.get().contains()`.
* A `find_descendants` helper on `MovePath` that replaces `has_any_child_of` on the old `FlowsAtLocation`

These changes made porting the dataflow analyses much easier. Finally, this PR removes some now-unused code in `dataflow/at_location.rs` and elsewhere.

You can view the perf results for the final version of this PR [here](https://perf.rust-lang.org/compare.html?start=29b854fb741809c29764e33fc17c32ba9c6523ba&end=6e516c1410c18cfe4eb6d030a39fdb73c8d8a4fe). Here's an example of the graphviz diagrams that are generated for the `MaybeInitializedPlaces` analysis.

![image](https://user-images.githubusercontent.com/29463364/72846117-c3e97d80-3c54-11ea-8171-3d48981c9ddd.png)

Improve `ty.needs_drop`

* Handle cycles in `needs_drop` correctly
* Normalize types when computing `needs_drop`
* Move queries from rustc to rustc_ty
* Avoid query in simple cases

reopens rust-lang#65918

…=petrochenkov

Hasten macro parsing

r? @eddyb

…enkov

On mismatched argument count point at arguments

Add missing `_zeroed` varants to `AllocRef`

The majority of the allocator wg has decided to add the missing `_zeroed` variants to `AllocRef`:

> these should be added since they can be efficiently implemented with the `mremap` system call on Linux. `mremap` allows you to move/grow/shrink a memory mapping, and any new pages added for growth are guaranteed to be zeroed.
>
> If `AllocRef` does not have these methods then the user will have to manually write zeroes to the added memory since the API makes no guarantees on their contents.

For the full discussion please see rust-lang/wg-allocators#14.

This PR provides default implementations for `realloc_zeroed`, `alloc_excess_zeroed`, `realloc_excess_zeroed`, and `grow_in_place_zeroed`.

r? @Amanieu

…t-lang#69017, r=petrochenkov

ICE in nightly-2020-02-08: handle TerminatorKind::Yield in librustc_mir::transform::promote_consts::Validator method

IR: rust-lang#69017
regressed commit: rust-lang@f8fd462
Source: https://play.rust-lang.org/?version=nightly&mode=debug&edition=2018&gist=55e65a869e1f5fef64cc4462b1a5a087

Addresses ICE reported in rust-lang#69017 by handling `TerminatorKind::Yield` in https://github.com/rust-lang/rust/blob/4d1241f5158ffd66730e094d8f199ed654ed52ae/src/librustc_mir/transform/promote_consts.rs#L465-L468.

<details><summary>Nightly build</summary>
<p>

```
$ cargo +nightly build
Compiling yielder v0.1.0 (/Users/chris/Desktop/tests/rustlang-tests/yielder)
error: internal compiler error: src/librustc_mir/transform/promote_consts.rs:467: _1 = suspend(move _21) -> [resume: bb2, drop: bb3] not promotable
 --> src/main.rs:8:27
  |
8 |         println!("-> {}", yield);
  |                           ^^^^^

thread 'rustc' panicked at 'Box<Any>', <::std::macros::panic macros>:2:4
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

note: the compiler unexpectedly panicked. this is a bug.

note: we would appreciate a bug report: https://github.com/rust-lang/rust/blob/master/CONTRIBUTING.md#bug-reports

note: rustc 1.43.0-nightly (71c7e14 2020-02-09) running on x86_64-apple-darwin

note: compiler flags: -C debuginfo=2 -C incremental --crate-type bin

note: some of the compiler flags provided by cargo are hidden

error: aborting due to previous error

error: could not compile `yielder`.

To learn more, run the command again with --verbose.
```

</p>
</details>

<details><summary>Stage 1 dev build</summary>
<p>

```
$ cargo +stage1 build
Compiling yielder v0.1.0 (/Users/chris/Desktop/tests/rustlang-tests/yielder)
warning: function is never used: `gen`
 --> src/main.rs:6:4
  |
6 | fn gen() -> impl Generator<usize> {
  |    ^^^
  |
  = note: `#[warn(dead_code)]` on by default

    Finished dev [unoptimized + debuginfo] target(s) in 0.53s
```

</p>
</details>

@jonas-schievink @oli-obk

Preparation for allocator aware `Box`

This cleans up the `Box` code a bit, and uses `Box::from_raw(ptr)` instead of `Box(ptr)`.
Additionally, `box_free` and `exchange_malloc` now uses the `AllocRef` trait and a comment was added on how `box_free` is tied to `Box`.

This a preparation for an upcoming PR, which makes `Box` aware of an allocator.

r? @Amanieu

Add self to .mailmap

Fix outdated doc comment.

r? @RalfJung