Fixes #21062.

It doesn't have to be a literal memory allocation (ala malloc), e.g. it
can be in static memory, so saying "allocated" is mildly misleading.

Thanks to @mahkoh for pointing it out.

r? @steveklabnik

It returns `false`, not `None`.

Given that this is entirely internal, this enhancement isn't going to be needed. And if it is, we'll add it.

Closes #2741.

r? @alexcrichton

Fixes #22124

There are a number of holes that the stability lint did not previously cover,
including:

* Types
* Bounds on type parameters on functions and impls
* Where clauses
* Imports
* Patterns (structs and enums)

These holes have all been fixed by overriding the `visit_path` function on the
AST visitor instead of a few specialized cases. This change also necessitated a
few stability changes:

* The `collections::fmt` module is now stable (it was already supposed to be).
* The `thread_local::imp::Key` type is now stable (it was already supposed to
  be).
* The `std::rt::{begin_unwind, begin_unwind_fmt}` functions are now stable.
  These are required via the `panic!` macro.
* The `std::old_io::stdio::{println, println_args}` functions are now stable.
  These are required by the `print!` and `println!` macros.
* The `ops::{FnOnce, FnMut, Fn}` traits are now `#[stable]`. This is required to
  make bounds with these traits stable. Note that manual implementations of
  these traits are still gated by default, this stability only allows bounds
  such as `F: FnOnce()`.

Closes #8962
Closes #16360
Closes #20327

The live code analysis only visited the function's body when visiting a
method, and not the FnDecl and the generics, resulting in code to be
incorrectly marked as unused when it only appeared in the generics, the
arguments, or the return type, whereas the same code in non-method
functions was correctly detected as used. Fixes #20343.

Originally I just added a call to `walk_generics` and `walk_fndecl` alongside `walk_block` but then I noticed the `walk_method_helper` function did pretty much the same thing. The only difference is that it also calls `visit_mac`, but since this is not going to happen at this stage, I think it's ok. However let me know if this was not the right thing to do.

This renames the PrivateNoMangleFns lint to allow both to happen in a
single pass, since they do roughly the same work.

Closes #21856

Open questions:

[ ]: Do the tests actually pass (I'm running make check and running out the door now)
[ ]: Is the name of this lint ok. it seems to mostly be fine with [convention](https://github.com/rust-lang/rfcs/blob/cc53afbe5dea41e1f7d1c3dce71e013abe025211/text/0344-conventions-galore.md#lints)
[ ]: I'm not super thrilled about the warning text

r? @kmcallister (Shamelessly nominating because you were looking at my other ticket)

There are a number of holes that the stability lint did not previously cover,
including:

* Types
* Bounds on type parameters on functions and impls
* Where clauses
* Imports
* Patterns (structs and enums)

These holes have all been fixed by overriding the `visit_path` function on the
AST visitor instead of a few specialized cases. This change also necessitated a
few stability changes:

* The `collections::fmt` module is now stable (it was already supposed to be).
* The `thread_local::imp::Key` type is now stable (it was already supposed to
  be).
* The `std::rt::{begin_unwind, begin_unwind_fmt}` functions are now stable.
  These are required via the `panic!` macro.
* The `std::old_io::stdio::{println, println_args}` functions are now stable.
  These are required by the `print!` and `println!` macros.
* The `ops::{FnOnce, FnMut, Fn}` traits are now `#[stable]`. This is required to
  make bounds with these traits stable. Note that manual implementations of
  these traits are still gated by default, this stability only allows bounds
  such as `F: FnOnce()`.

Additionally, the compiler now has special logic to ignore its own generated
`__test` module for the `--test` harness in terms of stability.

Closes #8962
Closes #16360
Closes #20327

[breaking-change]

This is a resurrection and heavy revision/expansion of a PR that pcwalton did to resolve #8861.

The most relevant, user-visible semantic change is this: #[unsafe_destructor] is gone. Instead, if a type expression for some value has a destructor, then any lifetimes referenced within that type expression must strictly outlive the scope of the value.

See discussion on https://github.com/rust-lang/rfcs/pull/769

Feature gate `box` patterns.

Note that this adds a new feature gate, `box_patterns` specific to e.g. `let box i = ...`, while leaving  `box` expressions (alone) still guarded via the preexisting `box_syntax`.

just say "experimental."

It (1.) is invariant, (2.) must strictly outlive the arena itself,
(3.) constrains the inputs to the arena so that their borrows must
also strictly outlive the arena itself.

This implies that, for now, one can no longer have cross-references
between data allocated via the same `Arena` (even when the data is not
subject to the Drop Check rule).  Instead one must carry multiple
`Arena` instances, or (more commonly), use one or more `TypedArena`
instances with enums encoding the different variants of allocated
data.

Some compile-fail tests illustrated cases to be rejected by dropck,
including ones that check cyclic data cases designed to exposed bugs
if they are actually tricked into running by an unsound analysis.

E.g. these exposed bugs in earlier broken ways of handling `Vec<T>`.

(Note that all the uses of `unsafe_destructor` are just placating the
simple analysis used for that feature, which will eventually go away
once we have put the dropck through its paces.)

includes regression tests discovered during bootstrapping and tests of
cyclic structure that currently pass and are expected to continue
passing under the dropck rule.

(Note that all the uses of `unsafe_destructor` are just placating the
simple analysis used for that feature, which will eventually go away
once we have put the dropck through its paces.)

Handles e.g. `impl<T> Drop for Vec<T>` as parametric: If `T` does not
have any drop code that could read from borrowed data of lifetime `'a`,
then we infer that the drop code for `Vec<T>` also cannot read from
borrowed data of lifetime `'a`, and therefore we do not need to inject
the SafeDestructor constraint for it.

Notably, this enables us to continue storing cyclic structure, without
any `unsafe` code, in `Vec`, without allowing (unsound) destructors on
such cyclic data. (Later commits have tests illustrating these two
cases in run-pass and compile-fail, respectively.)

(This is "Condition (B.)" in Drop-Check rule described in RFC 769.)

Port `core::ptr::Unique` to have `PhantomData`. Add `PhantomData` to
`TypedArena` and `Vec` as well.

As a drive-by, switch `ptr::Unique` from a tuple-struct to a struct
with fields.

Namely: `collections` (used in `dlist.rs`), `syntax`, `rustc`,
`rustc_typeck`, `rustc_trans`, and `rustdoc`.

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]

`make docs` fails when (xe)latex is not installed. The output is pretty weird, looks like it's doing some `eval` tricks but something is blank:

    /bin/sh: -output-directory=.: command not found
    /home/tim/dev/rust/rust/mk/docs.mk:220: recipe for target 'doc/reference.pdf' failed

I have neither latex or xelatex installed. It seems like the following snippet is meant for me, but the logic is backwards:

    ifeq ($(CFG_XELATEX),)
        CFG_LATEX := $(CFG_XELATEX)
        XELATEX = 1
      else
        $(info cfg: no xelatex found, disabling LaTeX docs)
        NO_PDF_DOCS = 1
    endif

I verified with:

    $ make doc/reference.pdf CFG_XELATEX=/bin/foo
    cfg: no xelatex found, disabling LaTeX docs

Fixes #22124

Largely adapted from pcwalton's original branch, with following
notable modifications:

Use `regionck::type_must_outlive` to generate `SafeDestructor`
constraints.  (this plugged some soundness holes in the analysis).

Avoid exponential time blowup on compile-fail/huge-struct.rs by
keeping the breadcrumbs until end of traversal.

Avoid premature return from regionck::visit_expr.

Factored drop-checking code out into dropck module.

Added `SafeDestructor` to enum `SubregionOrigin` (for error reporting).

----

Since this imposes restrictions on the lifetimes used in types with
destructors, this is a (wait for it)

[breaking-change]

immediately surrounding a node that is a terminating_scope
(e.g. statements, looping forms) during which the destructors run (the
destructors for temporaries from the execution of that node, that is).

Introduced DestructionScopeData newtype wrapper around ast::NodeId, to
preserve invariant that FreeRegion and ScopeChain::BlockScope carry
destruction scopes (rather than arbitrary CodeExtents).

Insert DestructionScope and block Remainder into enclosing CodeExtents
hierarchy.

Add more doc for DestructionScope, complete with ASCII art.

Switch to constructing DestructionScope rather than Misc in a number
of places, mostly related to `ty::ReFree` creation, and use
destruction-scopes of node-ids at various calls to
liberate_late_bound_regions.

middle::resolve_lifetime: Map BlockScope to DestructionScope in `fn resolve_free_lifetime`.

Add the InnermostDeclaringBlock and InnermostEnclosingExpr enums that
are my attempt to clarify the region::Context structure, and that
later commmts build upon.

Improve the debug output for `CodeExtent` attached to `ty::Region::ReScope`.

Loosened an assertion in `rustc_trans::trans::cleanup` to account for
`DestructionScope`.  (Perhaps this should just be switched entirely
over to `DestructionScope`, rather than allowing for either `Misc` or
`DestructionScope`.)

----

Even though the DestructionScope is new, this particular commit should
not actually change the semantics of any current code.

(My fix to for-loops (21984) did not deal with similar problems in
if-let expressions, so those binding shifts stay.)

Updates to the bison grammar:

* Fixes to range syntax - allow `expr[..]`, and fix precedence to allow `for _ in i.. { }`
* Allow "extern crate" in stmts
* Add qualified path expressions (`<TYPE as TRAIT_REF>::item`)

Fixes #21062.

I was working on adding examples to the documentation in `std::num::Float`. I got to `exp2`, which says "Returns 2 raised to the power of the number, `2^(self)`."

So I tried running this code:

```
use std::num::Float;

#[test]
fn test_exp2() {
    assert_eq!(32.0, 5.0.exp2());
}
```

and it resulted in a failure of `(left: `32`, right: `148.413159`)`. That 148.413159 is the value for e^5, which is `exp()`, not `exp2()`.

Sure enough, `exp2` is calling `exp` and shouldn't be, looks like a copy-paste error. 

I haven't added any tests for this since it's unlikely to break again, but I will happily do so if people think that would be a good idea. The doc examples are coming :)

I scanned through the other functions in these files for similar sorts of errors and didn't notice any.

It doesn't have to be a literal memory allocation (ala malloc), e.g. it
can be in static memory, so saying "allocated" is mildly misleading.

Bison grammar: refactor index expressions to use ordinary expressions as indexes, remove index_expr and expr_norange

As the function comment already says, the types generated in the
foreign_signture function don't necessarily match the types used for a
corresponding rust function. Therefore we can't just use these types to
guide the translation of the wrapper function that bridges between the
external ABI and the rust ABI. Instead, we can query LLVM about the
types used in the rust function and use those to generate an appropriate
wrapper.

Fixes #21454