Until now, x86_64-pc-windows-gnu linked `rsbegin.o` and `rsend.o` just
like i686-pc-windows-gnu, even though they were no-ops for it.
This was likely done for the simplicity back when it was introduced.

Today the things are different and these startup/end objects harm other
features, like `build-std`. Given the demotion of i686-pc-windows-gnu
from tier 1, there is no point in hurting x86_64-pc-windows-gnu,
which remains a tier 1.

The files are still shipped in case downstream crates expect them, as in
case of the unmaintained `xargo`.

Use llfn_attrs_from_instance() to generate the attributes for the
allocator shim. This ensures that we generate all the usual
attributes (and don't get to find out one-by-one that a certain
attribute is important for a certain target). Additionally this
will enable emitting the allocator-specific attributes (not
included here).

This change is quite awkward because the allocator shim uses
SimpleCx, while llfn_attrs_from_instance uses CodegenCx. I've
switched it to use SimpleCx plus tcx/sess arguments where necessary.
If there's a simpler way to do this, I'd love to know about it...

Eliminate a use of `map_or` in favor of a match.

Inline some variable definitions that don't add clarity, and that
prevent short-circuiting.

I.e. do not mark them as used, or non-speculative loaded, or similar.
Previously they were sometimes finalized during early resolution, causing issues like rust-lang#144793 (comment).

resolve: Do not finalize shadowed bindings

I.e. do not mark them as used, or non-speculatively loaded, or similar.
Previously they were sometimes finalized during early resolution, causing issues like rust-lang#144793 (comment).

… r=jackh726

Demote both armebv7r-none-* targets.

OK, slightly more controversial than rust-lang#146520 and rust-lang#146522  - I'd like to drop the bare-metal **big-endian** Armv7-R targets down to Tier 3.

The reason is simple - we cannot test them in https://github.com/rust-embedded/cortex-ar/. This because QEMU support for Big Endian Armv7-R is broken. I tried quite hard, but all the strings I printed with semihosting came out byte swapped (or "etybawa depp") because of how QEMU kludges the access to memory in big-endian mode.

The target also has only a single maintainer. Although, if ```@chrisnc``` wants to put up a case for keeping it at Tier 2 though, I'm happy to hear it!

This PR wil be rebased once rust-lang#146419 completes the queue.

…, r=petrochenkov

port `#[debugger_visualizer]` to the new attribute system

… r=petrochenkov

Stop linking rs{begin,end} objects on x86_64-*-windows-gnu

Until now, x86_64-pc-windows-gnu linked `rsbegin.o` and `rsend.o` just like i686-pc-windows-gnu, even though they were no-ops for it. This was likely done for the simplicity back when it was introduced.
Today the things are different and these startup/end objects harm other features, like `build-std`. Given the demotion of i686-pc-windows-gnu from tier 1, there is no point in hurting x86_64-pc-windows-gnu, which remains a tier 1.

The files are still shipped in case downstream crates expect them, as in case of the unmaintained `xargo`.

Fixes rust-lang#146739

…=jackh726

Use standard attribute logic for allocator shim

Use llfn_attrs_from_instance() to generate the attributes for the allocator shim. This ensures that we generate all the usual attributes (and don't get to find out one-by-one that a certain attribute is important for a certain target). Additionally this will enable emitting the allocator-specific attributes (not included here).

This change is quite awkward because the allocator shim uses SimpleCx, while llfn_attrs_from_instance uses CodegenCx. I've switched it to use SimpleCx plus tcx/sess arguments where necessary. If there's a simpler way to do this, I'd love to know about it...

…ang, r=petrochenkov

Reduce some uses of `LegacyBang`

- **Switch `dummy_bang` from `LegacyBang` to `Bang`**
- **mbe: Switch dummy extension used for errors from `LegacyBang` to `Bang`**

fix doc comments to be more standard

…or, r=lolbinarycat

Add new `tyalias` intra-doc link disambiguator

Fixes rust-lang#146855.

Alternative to rust-lang#146866.

This adds support for a new disambiguator: `tyalias`. I think it's common enough to have type aliases nowaday, so no reason to not be able to have a disambiguator for them.

cc `@fmease`
r? lolbinarycat

mbe: Simplify check_redundant_vis_repetition

Eliminate a use of `map_or` in favor of a match.

Inline some variable definitions that don't add clarity, and that
prevent short-circuiting.

Rename `rust.use-lld` to `rust.bootstrap-override-lld`

First part of rust-lang#146640. The old option is kept for backwards compatibility, we can remove it in ~6 months, as usually.

I'm not sure if the bootstrap prefix is ideal, after all we have a bunch of other configs that only affect bootstrap's behavior and not the built artifacts. Maybe `build.override-lld`? But I don't think it matters that much, as long as it's clear that it is an override, and how does it differ from `rust.lld`.

r? ```@jieyouxu```

…s, r=Kobzol

Ignore more failing ui tests for GCC backend

Needed for rust-lang#146414.

r? `@Kobzol`