Signed-off-by: Tarek <[email protected]>

…rder_fields

Migrate `reorder_fields` Assist to Use `SyntaxFactory`

minor: Bump xshell

They were converted to LSP position using the current file's line index, which is obviously wrong.

fix: Fix related documents diagnostics

No functional change (yet).

…ndirect

Overhaul the `-l` option parser (for linking to native libs)

The current parser for `-l` options has accumulated over time, making it hard to follow. This PR tries to clean it up in several ways.

Key changes:
- This code now gets its own submodule, to slightly reduce clutter in `rustc_session::config`.
- Cleaner division between iterating over multiple `-l` options, and processing each individual one.
- Separate “split” step that breaks up the value string into `[KIND[:MODIFIERS]=]NAME[:NEW_NAME]`, but leaves parsing/validating those parts to later steps.
  - This step also gets its own (disposable) unit test, to make sure it works as expected.
- A context struct reduces the burden of parameter passing, and makes it easier to write error messages that adapt to nightly/stable compilers.
- Fewer calls to `nightly_options` helper functions, because at this point we can get the same information from `UnstableOptions` and `UnstableFeatures` (which are downstream of earlier calls to those helper functions).

There should be no overall change in compiler behaviour.

…r-errors

rename rustc_const_stable_intrinsic -> rustc_intrinsic_const_stable_indirect

In rust-lang#120370 this name caused confusion as the author thought the intrinsic was stable. So let's try a different name...

If we can land this before the beta cutoff we can avoid needing `cfg(bootstrap)` for this. ;)
Cc ``@compiler-errors`` ``@saethlin``

Document alternatives to `static mut`

In rust-lang#133143 I've noticed alternatives to `static mut` aren't documented anywhere.

Subtree update of `rust-analyzer`

r? ``@ghost``