Discard postfix completion if the next_non_trivia_sibling of dot_token
is an ELSE_KW.

Looks like this is a native encoding for Emacs at least!

minor: sync from downstream

…indivisble-expr, r=Veykril

fix: Don't trigger postfix completion in `if` block which has an `else` block

Fix rust-lang#14096

…verflow, r=Veykril

fix: Don't expand macros in the same expansion tree after overflow

This patch fixes 2 bugs:

- In `Expander::enter_expand_id()` (and in code paths it's called), we never check whether we've reached the recursion limit. Although it hasn't been reported as far as I'm aware, this may cause hangs or stack overflows if some malformed attribute macro is used on associated items.
- We keep expansion even when recursion limit is reached. Take the following for example:

  ```rust
  macro_rules! foo { () => {{ foo!(); foo!(); }} }
  fn main() { foo!(); }
  ```

  We keep expanding the first `foo!()` in each expansion and would reach the limit at some point, *after which* we would try expanding the second `foo!()` in each expansion until it hits the limit again. This will (by default) lead to ~2^128 expansions.

  This is essentially what's happening in rust-lang#14074. Unlike rustc, we don't just stop expanding macros when we fail as long as it produces some tokens so that we can provide completions and other services in incomplete macro calls.

This patch provides a method that takes care of recursion depths (`Expander::within_limit()`) and stops macro expansions in the whole macro expansion tree once it detects recursion depth overflow. To be honest, I'm not really satisfied with this fix because it can still be used in unintended ways to bypass overflow checks, and I'm still seeking ways such that misuses are caught by the compiler by leveraging types or something.

Fixes rust-lang#14074

Co-authored-by: Stig Brautaset <stig@brautaset.org>

Support UTF-32 position encoding

Looks like this is a native encoding for Emacs at least!

…, r=Veykril

fix: Search raw identifiers without prefix

When we find references/usages of a raw identifier, we should disregard `r#` prefix because there are keywords one can use without the prefix in earlier editions (see rust-lang#13034; this bug is actually fallout from the PR). `name`, the text we're searching for, has already been stripped of the prefix, but the text of nodes we compare it to hasn't been.

The second commit is strictly refactoring, I can remove it if it's not much of value.

internal: Improve parser recovery for delimited lists

Closes rust-lang/rust-analyzer#11188, rust-lang/rust-analyzer#10410, rust-lang/rust-analyzer#10173

Should probably be merged after the stable release as this might get the parser stuck if I missed something

Don't assume VSCode internal commands in the server

Bump lsp-types

…nly of `None` args

internal: Don't allocate the generic_args vec in `hir_def::Path` if it consists only of `None` args

Saves roughly 5mb on self

internal: Enable smallvec's `union` feature

Trigger call info for more completions of signature having things

Replace some often empty `Vec`s with boxed slices

Add v7 metadata support to rust-analyzer

Paired to rust-lang#101550

minor: Fix a pair of typos

internal: Don't reconstruct ref match completion in to_proto manually

cc rust-lang/rust-analyzer#12571

Adjust binding mode inlay hints to render better with @ patterns

fix: Bring back hovering call parens for return type info

…-snippet-cap-check, r=Veykril

Conditionally add snippet marker.

Fixes rust-lang#14161.

Heya, I just added the code with some tests, but not sure if it's *the way* to do it -- I didn't refactor existing methods for the `check` test method, but added another calling layer.

internal: Make CompletionItem more POD-like

fix link

Fixes rust-lang/rust-analyzer#14170