Unstable items used in a macro expansion will now always trigger
stability warnings, *unless* the unstable items are directly inside a
macro marked with `#[allow_internal_unstable]`. IOW, the compiler warns
unless the span of the unstable item is a subspan of the definition of a
macro marked with that attribute.

E.g.

    #[allow_internal_unstable]
    macro_rules! foo {
        ($e: expr) => {{
            $e;
            unstable(); // no warning
            only_called_by_foo!();
        }}
    }

    macro_rules! only_called_by_foo {
        () => { unstable() } // warning
    }

    foo!(unstable()) // warning

The unstable inside `foo` is fine, due to the attribute. But the
`unstable` inside `only_called_by_foo` is not, since that macro doesn't
have the attribute, and the `unstable` passed into `foo` is also not
fine since it isn't contained in the macro itself (that is, even though
it is only used directly in the macro).

In the process this makes the stability tracking much more precise,
e.g. previously `println!("{}", unstable())` got no warning, but now it
does. As such, this is a bug fix that may cause [breaking-change]s.

The attribute is definitely feature gated, since it explicitly allows
side-stepping the feature gating system.

The API this exposes is a little strange (being attached to `static`s),
so it makes sense to conservatively feature gate it. If it is highly
popular, it is possible to reverse this gating.

This changes the type of some public constants/statics in libunicode.
Notably some `&'static &'static [(char, char)]` have changed
to `&'static [(char, char)]`. The regexp crate seems to be the
sole user of these, yet this is technically a [breaking-change]

Fix #22173

they are used without a feature-gate. This is both kinder to existing
code and should make it easier to land this PR, since we don't
have to catch EVERY SINGLE SUFFIX.

 - We shouldn't be using `check_name` here at all
 - `contains_name(ref_slice(foo), bar)` is redundant, `contains_name` just iterates over its first arg and calls `check_name`
 - match would be better than a bunch of ifs

The other cases: `concat_idents!`, `log_syntax!`, and `trace_macros!`,
(these macros, with `asm!`, are handled (eagerly) in feature_gate.rs).

Added a second note about keeping the reference.md list up-to-date to
the bottom of the list, since not everyone (including me) reads the
big comment at the top of it.  :)

Ensured that the feature gate list in reference.md is kept in
alphabetical order.

Rollup merge (373cbab5) of PR #20723
accidentally reverted a portion of commit
8327bcc1 which shifted
`slicing_syntax` from Active to Accepted.

just say "experimental."

See RFC 320, "Non-zeroing dynamic drops."

Fix #22173

[breaking-change]

Switch feature-gate checker from `box_syntax` to `box_patterns` when
visiting a pattern.

(Having to opt into both `box_syntax` and `box_patterns` seemed
unnecessary.)

[breaking-change]

    #[plugin] #[no_link] extern crate bleh;

becomes a crate attribute

    #![plugin(bleh)]

The feature gate is still required.

It's almost never correct to link a plugin into the resulting library /
executable, because it will bring all of libsyntax and librustc with it.
However if you really want this behavior, you can get it with a separate
`extern crate` item in addition to the `plugin` attribute.

Fixes #21043.
Fixes #20769.

[breaking-change]

Fixes #21833.

[breaking-change]

The 'stable_features' lint helps people progress from unstable to
stable Rust by telling them when they no longer need a `feature`
attribute because upstream Rust has declared it stable.

This compares to the existing 'unstable_features', which is used
to implement feature staging, and triggers on *any* use
of `#[feature]`.

The spelling corrections were made in both documentation comments and
regular comments.

Also some tidying up of a bunch of crate attributes

Does a sanity check of the version numbers.