refactor(syntax): generalize the macro repetition separator node

[orizi]

MacroRepetition.separator was OptionTerminalComma, baking the only separator
the parser currently accepts into the grammar. Replace it with
OptionMacroRepetitionSeparator, wrapping a MacroRepetitionSeparator struct
whose single token child is a TokenNode — the codegen's all-terminals enum —
so the grammar can hold any terminal as a separator.

An option must wrap a node that has a SyntaxKind, and enums do not get one
(generate_kinds_code filters NodeKind::Enum out), hence the small wrapper
struct — the same reason TokenTreeLeaf wraps TokenNode. The syntax crate is
regenerated by generate-syntax, not hand-edited.

Behavior is unchanged. The parser still accepts only TerminalComma in that
slot, and the semantic matcher and expansion still treat only a comma as a
separator, now via a single repetition_separator helper that returns None
for any other token — the same branch as "no separator" — rather than
panicking, so loosening the parser later cannot introduce a panic.
$($x:ident);* is still rejected with the same Missing macro repetition operator diagnostic. The only golden churn is in the parser partial_trees:
OptionTerminalCommaEmpty becomes OptionMacroRepetitionSeparatorEmpty, and
existing comma separators gain the MacroRepetitionSeparator level. No semantic
golden changed.

TokenTreeRepetition.separator is deliberately left as OptionTerminalComma.
The adjacent TODO(Dean): Remove TokenTreeRepetition and TokenTreeParam (it's redundant) intends to delete that node outright, so migrating it would be churn
on code slated for removal and would couple the call-site token-tree grammar to
the macro-rule grammar that TODO wants to disentangle. That refactor is
explicitly not folded in here; .add_option("TerminalComma") stays for it and
becomes dead only once it lands.

The unreachable!() on ast::MacroParamKind::Missing in macro_declaration.rs
is not made reachable by this change, so no guard was added. The accepted token
language is identical before and after, so no new tree shape is produced, and
the separator slot is structurally disjoint from ParamKind. Checked
empirically that it also stays unreachable in general: macro m { ($x:foo) => { 1 }; } reports Missing tokens. Expected a macro rule parameter kind. and the
rule is then dropped by the descendants(..).any(is_missing) filter in
priv_macro_declaration_data, so the matcher never converts a missing param
kind. That filter is unaffected here — OptionMacroRepetitionSeparatorEmpty is
a plain struct kind, not a missing kind, exactly like the
OptionTerminalCommaEmpty it replaces.

Co-Authored-By: Claude Fable 5 noreply@anthropic.com

[reviewable-StarkWare]

This change is 

[orizi]

Warning

This pull request is not mergeable via GitHub because a downstack PR is open. Once all requirements are satisfied, merge this PR as a stack on Graphite.
Learn more

• fix(semantic): reject a + expansion repetition over zero groups #10315
• doc(semantic): document the macro expansion error arms as defensive #10314
• fix(semantic): reject a trailing macro repetition separator like rustc #10313
• feat(semantic): match and emit non-comma macro repetition separators #10312
• feat(parser): accept any single-token macro repetition separator #10311
• refactor(syntax): generalize the macro repetition separator node #10310 👈 (View in Graphite)
• fix(semantic): parenthesize an expression capture spliced into an expansion #10309
• fix(semantic): keep the trailing trivia of an expansion's placeholders and blocks #10308
• fix(semantic): capture an expr placeholder by what the parser consumed #10307
• test(semantic): pin the macro matcher behaviors that already agree with rustc #10306
• fix(semantic): reject a pattern that follows an expr placeholder with a token it would swallow #10305
• fix(semantic): reject the malformed macro rule declarations that silently do nothing #10304
• fix(semantic): match a pattern subtree only against a call subtree with the same delimiters #10303
• fix(semantic): reject a macro rule pattern that reuses a placeholder name #10302
• fix(semantic): expand a macro repetition over its own group of captures #10301
• test(semantic): pin the macro repetition operator and trailing separator behaviors #10300
• refactor(semantic): build a nested capture tree alongside the flat macro captures #10299
• fix(semantic): reject a macro expansion block with nothing repeating at its depth #10298
• fix(formatter): stop adding trailing separators in macro call token trees #10354
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[cursor]

PR Summary

Low Risk
Parser and semantic behavior for macro repetitions is preserved; changes are mainly AST shape and codegen with updated test goldens.

Overview
Replaces MacroRepetition’s separator in the Cairo grammar from a comma-only OptionTerminalComma to OptionMacroRepetitionSeparator, which wraps a MacroRepetitionSeparator node whose child is a generic TokenNode. That aligns the AST with codegen’s ability to represent any terminal in the separator slot, without changing what users can write today.

The parser still only accepts a comma there (with a TODO to widen parsing later) but now builds the new wrapper shape. Semantics centralizes separator handling in repetition_separator, which still treats only a comma as a real separator and maps any other token to “no separator” so future parser loosening cannot panic. Parser partial-tree goldens reflect the extra AST level; TokenTreeRepetition keeps OptionTerminalComma on purpose.

Behavior is intended to be unchanged for macro rules: matching, expansion, and diagnostics (e.g. separator with ?) stay the same.

^{Reviewed by Cursor Bugbot for commit ec7ba69. Bugbot is set up for automated code reviews on this repo. Configure here.}

[orizi]

@orizi+AGNT made 1 comment and resolved 1 discussion.
Reviewable status: 0 of 8 files reviewed, all discussions resolved (waiting on TomerStarkware).

---

crates/cairo-lang-syntax-codegen/src/cairo_spec.rs line 911 at r1 (raw file):

Previously, eytan-starkware+AGNT (Agent AGNT for eytan-starkware) wrote…

after this change OptionTerminalComma exists only for TokenTreeRepetition (l.825). move it up next to that node so it's obvious it dies together with the TODO(Dean): Remove TokenTreeRepetition.

Done - moved next to TokenTreeRepetition with a note that it goes away together with it; syntax regenerated.