Step 2a: Tree builder — core scaffold and basic insertion modes

[thomasnemer]

Parent: #21

Goal

Implement the TreeBuilder scaffold and the basic insertion modes that handle a typical HTML document: Initial, BeforeHtml, BeforeHead, InHead, AfterHead, InBody (basic elements only), Text, AfterBody, AfterAfterBody. No adoption agency, no tables, no templates.

File Changes

• crates/ie-html/src/tree_builder.rs — rewrite with TreeBuilder struct
• crates/ie-html/src/insertion_mode.rs — new file, mode enum + basic modes
• crates/ie-html/src/lib.rs — update parse() signature

Implementation

TreeBuilder struct

• Document, tokenizer, insertion_mode, open_elements stack
• head_pointer, form_pointer, frameset_ok
• ParseResult { document, errors, style_elements, link_stylesheets }

Core operations

• insert_element, insert_character, insert_comment
• generate_implied_end_tags, close_element
• Switch tokenizer state for script/style/title/textarea

Insertion modes (basic)

• Initial: doctype handling
• BeforeHtml: insert
• BeforeHead: insert
• InHead: <title>, <style>, <script>, , ,
• AfterHead: insert
• InBody: basic block/inline/void elements,

  closing, heading nesting

• Text: raw text for style/script
• AfterBody, AfterAfterBody

Tests

• Minimal document → correct structure
• Implicit elements:

  text → html/head/body/p created

• Void elements:
  ,
• Style/link extraction
• Nested elements:

  text

• Heading and paragraph handling

Acceptance Criteria

• cargo test -p ie-html passes
• Clippy clean
• parse("

  Hello

  ") produces correct DOM