By using 'impl trait', it's possible to create a self-referential
type as follows:

fn foo() -> impl Copy { foo }

This is a function which returns itself.
Normally, the signature of this function would be impossible
to write - it would look like 'fn foo() -> fn() -> fn() ...'
e.g. a function which returns a function, which returns a function...

Using 'impl trait' allows us to avoid writing this infinitely long
type. While it's useless for practical purposes, it does compile and run

However, issues arise when we try to generate llvm debuginfo for such a
type. All 'impl trait' types (e.g. ty::Opaque) are resolved when we
generate debuginfo, which can lead to us recursing back to the original
'fn' type when we try to process its return type.

To resolve this, I've modified debuginfo generation to account for these
kinds of weird types. Unfortunately, there's no 'correct' debuginfo that
we can generate - 'impl trait' does not exist in debuginfo, and this
kind of recursive type is impossible to directly represent.

To ensure that we emit *something*, this commit emits dummy
debuginfo/type names whenever it encounters a self-reference. In
practice, this should never happen - it's just to ensure that we can
emit some kind of debuginfo, even if it's not particularly meaningful

Fixes rust-lang#58463

This commit removes the check that disallows the `#[non_exhaustive]`
attribute from being placed on enum variants and removes the associated
tests.

Further, this commit lowers the visibility of enum variant constructors
when the variant is marked as non-exhaustive.

This commit updates the unstable book and diagnostics to reflect that
the `#[non_exhaustive]` attribute is now available for enum variants.

This commit adds support for non-exhaustive enum variants in rustdoc,
extending the existing support for non-exhaustive enums and structs.

…unused attribute.

…es not flag them as unused.

…_{start,end}.

…builds.

Wrapped lines earlier such that it is more coherent with the rest of the text.

Co-Authored-By: DevQps <[email protected]>

Removed duplicate line.

Co-Authored-By: DevQps <[email protected]>

Calling `Read::read` or `Write::write` without checking the returned
`usize` value is almost always an error.  Example code in the
documentation should demonstrate how to use the return value correctly.
Otherwise, people might copy the example code thinking that it is okay
to "fire and forget" these methods.

This was incorrectly copypasta'd from RepeatWith.

…, r=petrochenkov,QuietMisdreavus

RFC 2008: Enum Variants

Part of rust-lang#44109. See [Zulip topic](https://rust-lang.zulipchat.com/#narrow/stream/131828-t-compiler/topic/rfc-2008/near/132663140) for previous discussion.

r? @petrochenkov
cc @nikomatsakis

…=oli-obk

Fix stack overflow when generating debuginfo for 'recursive' type

By using 'impl trait', it's possible to create a self-referential
type as follows:

fn foo() -> impl Copy { foo }

This is a function which returns itself.
Normally, the signature of this function would be impossible
to write - it would look like 'fn foo() -> fn() -> fn() ...'
e.g. a function which returns a function, which returns a function...

Using 'impl trait' allows us to avoid writing this infinitely long
type. While it's useless for practical purposes, it does compile and run

However, issues arise when we try to generate llvm debuginfo for such a
type. All 'impl trait' types (e.g. ty::Opaque) are resolved when we
generate debuginfo, which can lead to us recursing back to the original
'fn' type when we try to process its return type.

To resolve this, I've modified debuginfo generation to account for these
kinds of weird types. Unfortunately, there's no 'correct' debuginfo that
we can generate - 'impl trait' does not exist in debuginfo, and this
kind of recursive type is impossible to directly represent.

To ensure that we emit *something*, this commit emits dummy
debuginfo/type names whenever it encounters a self-reference. In
practice, this should never happen - it's just to ensure that we can
emit some kind of debuginfo, even if it's not particularly meaningful

Fixes rust-lang#58463

…rochenkov

Recover from parse error in tuple syntax

…d, r=davidtwco

Account for short-hand field syntax when suggesting borrow

Fix rust-lang#52965.

…=alexcrichton

Fix broken download link in the armhf-gnu image

Thanks to @johnterickson for pointing this out!

r? @alexcrichton

implement `AsRawFd` for stdio locks

cc rust-lang/rfcs#2074.

…, r=petrochenkov

Whitelist some rustc attrs

These rustc attrs are used within libcore, and were causing failures when one mixed incremental compilation with bootstrapping (due to a default of `-D warnings` when bootstrapping).

Fix rust-lang#59523
Fix rust-lang#59524

Cc rust-lang#58633

…ntril

Improve the dbg! macro docs

# Description

As stated has been discussed in rust-lang#58383 the docs do not clearly state why it is useful to have the option to use `dbg!` in release builds as well. This PR should change that.

closes rust-lang#58383

In doc examples, don't ignore read/write results

Calling `Read::read` or `Write::write` without checking the returned `usize` value is almost always an error.  Example code in the documentation should demonstrate how to use the return value correctly.  Otherwise, people might copy the example code thinking that it is okay to "fire and forget" these methods.

…r=GuillaumeGomez

rustdoc: collapse blanket impls in the same way as normal impls

If the rustdoc setting _Auto-hide trait implementations documentation_ is activated (on by default), normal trait implementations are collapsed by default.

Blanket impls on the other hand are not collapsed. I'm not sure whether this is intended, but considering that the blanket impls for `From`, `Into`, `TryFrom`, ... are on every type, it would reduce the documentation bloat if these would also be collapsed when the setting is active.

(I'm not really familiar with the codebase and therefore just copied the code for the normal impl collapsing, but I could deduplicate it into a method, of course, too.)

Fix OnceWith docstring.

This was incorrectly copypasta'd from RepeatWith.