The rtio interface is a thin low-level interface over the I/O subsystems, and
the `Path` type is a little too high-level for this interface.

Closes #14675 (rustc: Encode argument names for traits)
Closes #14681 (rustc: Avoid UB with signed division/remainder)
Closes #14682 (librustc: Update AutoObject adjustment in writeback.)
Closes #14683 (Avoid 16-byte filenames in rlibs)
Closes #14687 (rustdoc: Inline static documentation across crates)
Closes #14689 (Remove reference to ~str in documentation)
Closes #14692 (Rename Iterator::len to count)
Closes #14693 (Implement Eq for HashSet and HashMap)
Closes #14699 (url: encode small bytes correctly.)
Closes #14700 (rustdoc: Submit examples to play.rust-lang.org)
Closes #14701 (mk: Run doc tests with --cfg dox)
Closes #14710 (rustc: Preserve reachable extern fns with LTO)
Closes #14711 (Removing unused wrapper to libc::close.)

This grows a new option inside of rustdoc to add the ability to submit examples
to an external website. If the `--markdown-playground-url` command line option
or crate doc attribute `html_playground_url` is present, then examples will have
a button on hover to submit the code to the playground specified.

This commit enables submission of example code to play.rust-lang.org. The code
submitted is that which is tested by rustdoc, not necessarily the exact code
shown in the example.

Closes #14654

These were only used for the markdown tests, and there's no reason they should
be distinct from the other tests.

All rust functions are internal implementation details with respect to the ABI
exposed by crates, but extern fns are public components of the ABI and shouldn't
be stripped. This commit serializes reachable extern fns to metadata, so when
LTO is performed all of their symbols are not stripped.

Closes #14500

There were a few examples in the macros::builtin module that weren't being run
because they were being #[cfg]'d out.

Closes #14697

Previously, bytes less than 16 would be encoded as %X, rather than %XX,
since the output width was left to be automatic.

Also fix documentation references to PartialEq.

This commit carries out the request from issue #14678:

> The method `Iterator::len()` is surprising, as all the other uses of
> `len()` do not consume the value. `len()` would make more sense to be
> called `count()`, but that would collide with the current
> `Iterator::count(|T| -> bool) -> unit` method. That method, however, is
> a bit redundant, and can be easily replaced with
> `iter.filter(|x| x < 5).count()`.
> After this change, we could then define the `len()` method
> on `iter::ExactSize`.

Closes #14678.

[breaking-change]

In addition to avoiding 16-byte filenames with bytecode files, this commit also
avoids 16-byte filenames with object files pulled in from native libraries.

LLDB contains a bug that makes it crash if an archive it reads
contains a file the name of which is exactly 16 bytes long. This
bug recently has made it impossible to debug Rust applications with
LLDB because some standard libraries triggered it indirectly:
For rlibs, rustc includes the LLVM bytecode in the archive, giving
it the extension ".bc.deflate". For liballoc (for example) this
results in the 16 character filename "alloc.bc.deflate", which is
bad.

This commit replaces the ".bc.deflate" suffix with
".bytecode.deflate" which itself is already longer than 16 bytes,
thus making sure that the bug won't be run into anymore.

The bug could still be run into with 14 character filenames because
then the .o files will trigger it. However, this is much more rare
and working around it would introduce more complexity than necessary
at the moment. It can always be done later on, if the need arises.

Fixes #14356.

Division and remainder by 0 are undefined behavior, and are detected at runtime.
This commit adds support for ensuring that MIN / -1 is also checked for at
runtime, as this would cause signed overflow, or undefined behvaior.

Closes #8460

This ensures that rustdoc can properly document inlined traits across crates.

Closes #14670

Separate out initial refactorings for PR #14604

(Fixes #14603)

Also updated/corrected a few other details in the tuple and struct sections.

I tried to split up the less mechanical changes into separate commits so they are easier to review. One thing I'm not quite sure of is whether `MoveReason` should just be replaced with `move_data::MoveKind`.

When converting check_loans to use ExprUseVisitor I encountered a few
issues where the wrong number of errors were being emitted for multiple
closure captures, but there is no existing test for this.

Currently it is not possible to distinguish moves caused by captures
in the ExprUseVisitor interface. Since check_Loans needs to make that
distinction for generating good diagnostics, this is necessary for
check_loans to switch to ExprUseVisitor.

This isn't necessary right now, but check_loans needs to be able to
distinguish between initialization and writes in the ExprUseVisitor
mutate callback.

Currently mem_categorization categorizes an AutoObject adjustment the
same as the original expression. This can cause two moves to be
generated for the same underlying expression. Currently this isn't a
problem in practice, since check_loans doesn't rely on ExprUseVisitor.

Refactor a number of functions in check_loans to take node IDs and spans
rather than taking expressions directly. Also rename some variables to
make them less ambiguous.

This is the first step towards using ExprUseVisitor in check_loans, as
now some of the interfaces more closely match those used in
ExprUseVisitor.

There's no need to distribute these ABI helpers for tests with the standard rust
distribution they're only needed for our tests.

Closes #2665

 Remove a warning (./configure: 1140: [: unexpected operator) when built with --disable-manage-submodules