Crate normalize_syntax_rules 

Syntax-based linting with tree-sitter queries.

This crate provides:

• Rule loading from multiple sources (builtins, user global, project)
• Rule execution with combined query optimization
• Pluggable data sources for rule conditionals

Rule File Format

# ---
# id = "no-unwrap"
# severity = "warning"
# message = "Avoid unwrap() on user input"
# allow = ["**/tests/**"]
# requires = { "rust.edition" = ">=2024" }
# enabled = true  # set to false to disable a builtin
# fix = ""  # empty = delete match, or use "$capture" to substitute
# ---

(call_expression
  function: (field_expression
    field: (field_identifier) @method)
  (#eq? @method "unwrap")) @match

Structs

BuiltinRule

A builtin rule definition (id, content).

DebugFlags

Debug output categories.

EnvSource

Environment variable source.

Finding

A finding from running a rule.

GitSource

Git repository state source.

GoSource

Go project source - parses go.mod for module metadata.

PathSource

Path-based source for glob matching.

PythonSource

Python project source - parses pyproject.toml for project metadata.

Rule

A syntax rule definition.

RuleOverride

Per-rule configuration override.

RulesConfig

Configuration for syntax rules analysis. Maps rule ID to per-rule configuration. e.g., { “rust/unnecessary-let” = { severity = “warning” } }

RustSource

Rust project source - parses Cargo.toml for edition, resolver, etc.

SourceContext

Context passed to sources for evaluation.

SourceRegistry

Registry of all available rule sources.

TypeScriptSource

TypeScript/JavaScript project source - parses tsconfig.json and package.json.

Enums

Severity

Severity level for rule findings.

Constants

BUILTIN_RULES

All embedded builtin rules.

Traits

RuleSource

A source of data for rule conditionals.

Functions

apply_fixes

Apply fixes to findings, returning the number of files modified. Fixes are applied in reverse order within each file to preserve byte offsets.

builtin_registry

Create a registry with all built-in sources.

evaluate_predicates

Evaluate predicates for a match.

load_all_rules

Load all rules from all sources, merged by ID. Order: builtins → ~/.config/moss/rules/ → .normalize/rules/ Then applies config overrides (severity, disable).

parse_rule_content

Parse rule content string with TOML frontmatter.

run_rules

Run rules against files in a directory. Optimized: combines all rules into single query per grammar for single-traversal matching.