incr.comp.: Use 128bit SipHash for fingerprinting

This PR switches incr. comp. result fingerprinting from 128 bit BLAKE2 to 128 bit SipHash. When we started using BLAKE2 for fingerprinting, the 128 bit version of SipHash was still experimental. Now that it isn't anymore we should be able to get a nice performance boost without significantly increasing collision probability.

~~I'm going to start a try-build for this, so we can gauge the performance impact before merging (hence the `WIP` in the title).~~

EDIT: Performance improvements look as expected. Tests seem to be passing.

Fixes #41215.

Mark block exits as reachable if the block can break.

This only happens when desugaring `catch` expressions for now, but regular blocks (in HIR) can be broken from - respect that when doing reachability analysis.

Fixes #45124.

Refactor `ensure` and `try_get_with`

There was a bit of code shared between `try_get_with` and `ensure`, after I
added `ensure`. I refactored that shared code into a query-agnostic method
called `read_node_index`.

The new method `read_node_index` will attempt to find the node
index (`DepNodeIndex`) of a query. When `read_node_index` finds the
`DepNodeIndex`, it marks the current query as a reader of the node it's
requesting the index of.

This is used by `try_get_with` and `ensure` as it elides the unimportant (to
them) details of if the query is invalidated by previous changed computation (Red)
or new and if they had to mark the query green. For both `try_get_with` and
`ensure`, they just need to know if they can lookup the results or have to
reevaluate.

@nikomatsakis this is the [refactor we discussed](https://github.com/rust-lang/rust/pull/45228#discussion_r144577392) in the comment thread of #45228

Save the highlighted item when switching tab

To be merged after #45281.

r? @rust-lang/docs

Issue 44986/fix windows ui path

#44968

Rollup of 8 pull requests

- Successful merges: #45343, #45349, #45352, #45374, #45375, #45376, #45377, #45382
- Failed merges:

rustbuild: fix dist in debug mode

In debug mode, the artifacts are placed in "debug", so don't hardcode "release" and use our helper to get the right directory name

Fix typos in librustc/ty/README.md

Fix typos in src/librustc/README.md

Allow passing a path with tilde to the configure script

Currently `./configure --local-rust-root=~/.cargo --enable-local-rebuild` fails with
 ```
Exception: no cargo executable found at `~/.cargo//bin/cargo`
```

Remove --enable-rustbuild config option from an example

`--enable-rustbuild` is no longer recognized by the configure script. Also I think we should use `./configure` and `./x.py` instead of `configure` and `../x.py`.

test: Update Emscripten failures/passing

All tests should now have annotation for *why* they're ignored on emscripten. A
few tests no longer need such an annotation as well!

Closes #41299

added examples of closures for str::find

This is an attempt to fix https://github.com/rust-lang/rust/issues/45327

r? @steveklabnik

Make Solaris builder compatible with Solaris 10 retry

Unfortunately, #45255 does not quite cut it,
so use a different approach to have Solaris 10 compatibility
by tricking libbacktrace's autoconf tests.
The sysroot download routine is slightly changed, too.

Backport ThinLTO LLVM 5 fixes

This makes building nightly more convenient on Arch.

code suggestions for non-shorthand field pattern, no-mangle lints

continuing in the spirit of #44942

![moar_lint_suggestions](https://user-images.githubusercontent.com/1076988/31485011-3b20cc80-aee7-11e7-993d-81267ab77732.png)

r? @estebank

Signed-off-by: NMarc-Antoine Perennou <Marc-Antoine@Perennou.com>

Ref: https://github.com/llvm-mirror/llvm/commit/ccb80b9c0f60f33780e5e29bf66a87bb56968b99

Rollup of 10 pull requests

- Successful merges: #44138, #45082, #45098, #45181, #45217, #45281, #45325, #45326, #45340, #45354
- Failed merges:

rustdoc: add a primitive page for "unit"

In `src/libstd/primitive_docs.rs`, a `#[doc(primitive = "unit")]`
section has sat long neglected.  This patch teaches rustdoc to recognize
"unit", and steals its trait implementations away from the tuple page.

Docs: a LocalKey might start in the Valid state

Add a comment to the docs for `LocalKey::state` explaining that some keys might skip the `Uninitialized` state and start in the `Valid` state.

cc #27716

r? @alexcrichton

Bump the minimum LLVM to 3.9

Old LLVM bugs are reportedly cropping up harder, but 3.9 seems to be OK.

Fixes #45277.

Generate FirstStatementIndex using newtype_index macro

Save selected search tab

Fixes #45278.

r? @rust-lang/docs

Fix out of date unstable book entries for `alloc_*` features.

The `alloc_jemalloc` crate does not provide a type to use with `#[global_allocator]`, and (according to Alex) `extern crate alloc_jemalloc;` alone became a no-op when `#[global_allocator]` was introduced.

List of all lang items in unstable book.

I don't know how to link to the source code correctly, so I just put the information aside the lang item name.

Documenting the process for when rustfmt/rls break

**DO NOT MERGE YET**

I'm documenting what to do when rustfmt or rls break because of your changes. I'm currently going through this and will keep adding more as I figure out what all the steps are. This first commit is based on @alexcrichton's [comment on my original PR](https://github.com/rust-lang/rust/pull/44766#issuecomment-332653525).

[Rendered](https://github.com/sunjay/rust/blob/breakingrustfmtrls/CONTRIBUTING.md#breaking-tools-built-with-the-compiler)

Reviews are welcome, but as I mentioned, I will be revising this as I go.

Mention Clone and refs in --explain E0382

I followed the discussion in #42446 and came up with these additions.

- Mention references before going into traits. They're probably more likely solutions.
- Mention `Clone` before `Copy`. Cloning has wider applicability and `#derive[Copy, Clone]` makes more sense after learning about `Clone`.

The language is not great, any suggestions there would be appreciated ✨

Deprecate several flags in rustdoc

Part of #44136

cc @rust-lang/dev-tools @rust-lang/docs

This is a very basic PR to start deprecating some flags; `rustdoc` doesn't really have fancy output options like `rustc` does, so I went with `eprintln!`. Happy to change it if people feel that's not appropriate.

Also, I have no idea if we can or should write tests here, so I didn't try. If someone feels strongly about it, then let's do it, but given that the only outcome here is a side effect...

Avoid unnecessary allocas for indirect function arguments

The extra alloca was only necessary because it made LLVM implicitly
handle the necessary deref to get to the actual value. The same happens
for indirect arguments that have the byval attribute. But the Rust ABI
does not use the byval attribute and so we need to manually add the
deref operation to the debuginfo.

The extra alloca was only necessary because it made LLVM implicitly
handle the necessary deref to get to the actual value. The same happens
for indirect arguments that have the byval attribute. But the Rust ABI
does not use the byval attribute and so we need to manually add the
deref operation to the debuginfo.

remove or encapsulate the remaining non-query data in tcx

I wound up removing the existing cache around inhabitedness since it didn't seem to be adding much value. I reworked const rvalue promotion, but not that much (i.e., I did not split the computation into bits, as @EddyB had tossed out as a suggestion). But it's now demand driven, at least.

cc @michaelwoerister -- see the `forbid_reads` change in last commit

r? @EddyB -- since the trickiest of this PR is the work on const rvalue promotion

cc #44137

In `src/libstd/primitive_docs.rs`, a `#[doc(primitive = "unit")]`
section has sat long neglected.  This patch teaches rustdoc to recognize
"unit", and steals its trait implementations away from the tuple page.