Until `external_traits` is cleaned up (i.e., no longer behind a
`RefCell`), `DocVisitor` will have to `take` `external_traits` -- just
like `DocFolder` -- to prevent `RefCell` runtime errors.

`DocFolder` allows transforming the docs, accomplished by making its
methods take and return types by-value. However, several of the rustdoc
`DocFolder` impls only *visit* the docs; they don't change anything.
Passing around types by-value is thus unnecessary, confusing, and
potentially inefficient for those impls.

`DocVisitor` is very similar to `DocFolder`, except that its methods
take shared references and return nothing (i.e., the unit type). This
should both be more efficient and make the code clearer.

There is an additional reason to add `DocVisitor`, too. As part of my
cleanup of `external_traits`, I'm planning to add a `fn cache(&mut self)
-> &mut Cache` method to `DocFolder` so that `external_traits` can be
retrieved explicitly from the `Cache`, rather than implicitly via
`Crate.external_traits` (which is an `Rc<RefCell<...>>`). However, some
of the `DocFolder` impls that could be turned into `DocVisitor` impls
only have a shared reference to the `Cache`, because they are used
during rendering. (They have to access the `Cache` via
`html::render::Context.shared.cache`, which involves an `Rc`.)

Since `DocVisitor` does not mutate any of the types it's visiting, its
equivalent `cache()` method will only need a shared reference to the
`Cache`, avoiding the problem described above.

This simplifies the code and future-proofs it against changes to
`Variant`.

Also, contrary to the comment, the clone is not that small, since
`Variant` contains `Item`s, which are quite large when you factor in
both stack- and heap-allocated memory.

It turned out to cause (minor) perf regressions.

The only bit failing was the module, so change that before removing the
`span` field.

It can be calculated on-demand even without a TyCtxt.

This also changed `from_item_kind` to take a whole item, which avoids
having to add more and more parameters.

I'm wondering if it was originally there so that we could `take` the
module which enables `after_krate` to take an `&Crate`. However, the two
impls of `after_krate` only use `Crate.name`, so we can pass just the
name instead.

There's no `Variant`, so it seems silly to have `Kind`.

It had exactly one field and no special behavior, so there was no point.

fixes clippy warnings of type:
match_like_matches_macro
or_fun_call
op_ref
needless_return
let_and_return
single_char_add_str
useless_format
unnecessary_sort_by
match_ref_pats
redundant_field_names

remove unused return type of dropck::check_drop_obligations()
don't wrap return type in Option in get_macro_by_def_id() since we would always return Some(..)
remove redundant return type of back::write::optimize()
don't Option-wrap return type of compute_type_parameters() since we always return Some(..)
don't return empty Result in assemble_generator_candidates()
don't return empty Result in assemble_closure_candidates()
don't return empty result in assemble_fn_pointer_candidates()
don't return empty result in assemble_candidates_from_impls()
don't return empty result in assemble_candidates_from_auto_impls()
don't return emtpy result in assemble_candidates_for_trait_alias()
don't return empty result in assemble_builtin_bound_candidates()
don't return empty results in assemble_extension_candidates_for_traits_in_scope() and assemble_extension_candidates_for_trait()
remove redundant wrapping of return type of StripItem::strip() since it always returns Some(..)
remove unused return type of assemble_extension_candidates_for_all_traits()

This brings the size of `Item` from

```
[src/librustdoc/lib.rs:103] std::mem::size_of::<Item>() = 680
```

to

```
[src/librustdoc/lib.rs:103] std::mem::size_of::<Item>() = 280
```

This gets rid of a bunch of `unwrap()`s and makes it a little more clear
what's going on.

Originally I wanted to make `fold_item` non-nullable too, which would
have been a lot nicer to work with, but unfortunately `stripper` does
actually return `None` in some places. I might make a follow-up moving
stripper to be special and not a pass so that passes can be
non-nullable.

This gives warnings about dead code.

 #58972 ignored extern_traits because before #65983 was fixed, they
would always fail to resolve, giving spurious warnings.
This undoes that change, so extern traits are now seen by the
`collect_intra_doc_links` pass. There are also some minor changes in
librustdoc/fold.rs to avoid borrowing the extern_traits RefCell more
than once at a time.

This drops the parking_lot dependency; the ReentrantMutex type appeared
to be unused (at least, no compilation failures occurred).

This is technically a possible change in behavior of its users, as
lock() would wait on other threads releasing their guards, but since we
didn't actually remove any threading or such in this code, it appears
that we never used that behavior (the behavior change is only noticeable
if the type previously was used in two threads, in a single thread
ReentrantMutex is useless).

constraints:

- clean/inline.rs needs this map to fill in traits when inlining
- fold.rs needs this map to allow passes to fold trait items
- html/render.rs needs this map to seed the Cache.traits map of all
  known traits

The first two are the real problem, since `DocFolder` only operates on
`clean::Crate` but `clean/inline.rs` only sees the `DocContext`. The
introduction of early passes means that these two now exist at the same
time, so they need to share ownership of the map. Even better, the use
of `Crate` in a rustc thread pool means that it needs to be Sync, so it
can't use `Lrc<Lock>` to manually activate thread-safety.

`parking_lot` is reused from elsewhere in the tree to allow use of its
`ReentrantMutex`, as the relevant parts of rustdoc are still
single-threaded and this allows for easier use in that context.

This commit upgrades the main LLVM submodule to LLVM's current master branch.
The LLD submodule is updated in tandem as well as compiler-builtins.

Along the way support was also added for LLVM 7's new features. This primarily
includes the support for custom section concatenation natively in LLD so we now
add wasm custom sections in LLVM IR rather than having custom support in rustc
itself for doing so.

Some other miscellaneous changes are:

* We now pass `--gc-sections` to `wasm-ld`
* The optimization level is now passed to `wasm-ld`
* A `--stack-first` option is passed to LLD to have stack overflow always cause
  a trap instead of corrupting static data
* The wasm target for LLVM switched to `wasm32-unknown-unknown`.
* The syntax for aligned pointers has changed in LLVM IR and tests are updated
  to reflect this.
* The `thumbv6m-none-eabi` target is disabled due to an [LLVM bug][llbug]

Nowadays we've been mostly only upgrading whenever there's a major release of
LLVM but enough changes have been happening on the wasm target that there's been
growing motivation for quite some time now to upgrade out version of LLD. To
upgrade LLD, however, we need to upgrade LLVM to avoid needing to build yet
another version of LLVM on the builders.

The revision of LLVM in use here is arbitrarily chosen. We will likely need to
continue to update it over time if and when we discover bugs. Once LLVM 7 is
fully released we can switch to that channel as well.

[llbug]: https://bugs.llvm.org/show_bug.cgi?id=37382

Like #43008 (f6689991), but _much more aggressive_.

Instead of hardcoding knowledge about the strip-private pass into the
rendering process we represent (some) stripped items as `ItemEnum::StrippedItem`.

Rustdoc will, for example, generate redirect pages for public items
contained in private modules which have been re-exported to somewhere
externally reachable - this will now not only work for the `strip-private`
pass, but for other passes as well, such as the `strip-hidden` pass.

This ensures that all external traits are run through the same filters that the
rest of the AST goes through, stripping hidden function as necessary.

Closes #13698