Html5 Laravel Package

Getting Started

Install via Composer: composer require masterminds/html5. The primary entry point is the Masterminds\HTML5 class. Start by parsing a fragment or full HTML document into a DOMDocument using loadHTML($html), then serialize back with saveHTML($dom) or save($dom, 'file.html'). For typical use cases (e.g., sanitizing user input, transforming HTML, preprocessing templates), this high-level API is sufficient—no need to touch lower-level parsers or events unless building a custom tool.

Implementation Patterns

• DOM manipulation pipelines: Parse → modify DOM (e.g., via DOMXPath or DOMDocument methods) → serialize. Ideal for cleaning HTML from WYSIWYG editors or extracting structured data.
• Template preprocessing: Parse mixed PHP/HTML files (thanks to embedded PI support) to extract or transform HTML before rendering.
• HTML sanitization: Combine with a DOM walker to strip unsafe tags/attributes (e.g., <script>, on* handlers) before output.
• Custom token handling: Implement EventHandlerInterface for SAX-like streaming parsing (e.g., content extraction, analytics, or SEO meta extraction) without building a full DOM.
• Namespace-aware parsing: Use xmlNamespaces: true and/or implicitNamespaces for SVG/MathML embedded in HTML (e.g., rendering dynamic charts or diagrams).
• Custom serialization rules: Extend OutputRules for tailored output (e.g., minification, pretty-printing with custom indentation, or canonicalization).

Gotchas and Tips

• Attribute name validation: PHP’s DOM ignores non-XML 1.0-compliant attribute names (e.g., containing :, .). Namespaced attributes must be declared explicitly or via implicitNamespaces—otherwise they’re silently dropped.
• Optional end tags: Since v2.9.0, table elements (<td>, <tr>, <th>, <tbody>, etc.) respect HTML5’s optional end tags, which affects DOM structure—be cautious when serializing to match original source.
• Void vs empty elements: Inline <svg>/<math> tags are serialized as void (<svg/>) unless they have content (2.10.0+). Use innerHTML-style logic carefully if relying on empty-string children.
• Entity encoding: Set encode_entities: true only when required (e.g., for email templates), as it aggressively encodes characters (e.g., é → é). Default encodes only what’s necessary.
• No built-in validation: The parser auto-corrects malformed HTML but doesn’t validate against the spec. Don’t rely on it for HTML5 compliance checks—use a linter instead.
• Reentrancy limitations: The parser is not re-entrant; avoid modifying input mid-parse. For large documents, consider chunked parsing at logical boundaries.
• PHP version quirks: The package dropped ext-libxml and ctype dependencies—ensure no legacy assumptions about those in your codebase. Works across PHP 5.3+, but latest releases target PHP 8.0+.
• Performance tuning: For batch parsing, instantiate HTML5 once and reuse it—costly setup (tokenization tables, rules) occurs on construction. Prefetch loadHTML()/saveHTML() for predictable latency.