feat: Comprehensive YARA language support upgrade (Grammar, Outlines, Folding, and Highlighting)

[binjo]

Title

feat: Comprehensive YARA language support upgrade (Grammar, Outlines, Folding, and Highlighting)

Description

This Pull Request delivers a massive upgrade to the yara.zed extension, bringing it to full feature-parity with advanced editors. It replaces the basic syntax highlighter with a robust, syntactically complete YARA development environment by upgrading the tree-sitter grammar, adding structured navigation queries, patching edge-case syntax bugs, and optimizing syntax highlighting.

---

Key Enhancements & Changes since Fork

1. ⚙️ Parser & Grammar Upgrade

• Namespace & Member Access Support: Replaced the incomplete egibs/tree-sitter-yara parser with celstur/tree-sitter-yara. This adds support for standard module namespaces, dot member accesses, and native function calls (e.g., pe.is_dll(), math.entropy()), resolving syntax-highlighting breakage on complex rules.
• Anonymous Variable Indexing Bug Fix: Resolved an issue where the parser choked on array-style index lookups on anonymous offsets (@) and lengths (!) (e.g., @[j] or ![j]), producing a syntax error block.

Upstream Grammar Patch details for grammar.js:

To resolve this inside the tree-sitter parser, grammar.js was modified to make bracketed index sequences optional on anonymous offsets and lengths:

string_offset: ($) =>
  seq(
    token(/@[a-zA-Z0-9_]*/),
    optional(seq($._lbrack, $._numeric_expression, $._rbrack))
  ),
string_length: ($) =>
  seq(
    token(/![a-zA-Z0-9_]*/),
    optional(seq($._lbrack, $._numeric_expression, $._rbrack))
  ),

The parser was then re-generated (npx tree-sitter generate) to update src/parser.c to fully support modern, complex rule indexing.

2. 🎨 Enhanced Syntax Highlighting (highlights.scm)

• Mapped standard library module namespaces (like pe, math, elf) cleanly to @type.
• Mapped members/fields to @function to represent them as callable components rather than plain strings.
• Added native #match? regular expression queries to highlight standard built-in functions dynamically (e.g. uint16, uint32be, etc.).
• Stripped literal operators like "!", "#" and "@" from query delimiters to prevent tree-sitter parser load rejections.

3. 📁 Structural Editor Features (Emacs yara-mode Parity)

• Document Outline (outline.scm): Added outline matches to display defined rule headers in Zed's Symbol/Outline navigation search (Ctrl/Cmd + Shift + O).
• Smart Auto-Indents (indents.scm): Structured and aligned block-indentation scopes for rule_body, meta_section, strings_section, and condition_section.
• Code Folding (folds.scm): Enabled vertical collapsing/folding on both rule-level blocks and individual sections (meta, strings, condition) to make analyzing large files easy.

---

Verification and Testing

All layers of this upgrade have been tested and verified inside Zed:

1. Parser & Highlights: Verified syntax rendering and AST accuracy across standard and complex rule datasets.
2. Editor Queries: Successfully verified rule folding, auto-indentation on bracket insertion, and rule outline indexing.