@reteps/tree-sitter-htmlmustache 0.0.18

package/grammar.js

@@ -0,0 +1,426 @@
+/**
+ * @file HTML grammar for tree-sitter
+ * @author Max Brunsfeld <maxbrunsfeld@gmail.com>
+ * @author Amaan Qureshi <amaanq12@gmail.com>
+ * @license MIT
+ */
+
+/// <reference types="tree-sitter-cli/dsl" />
+// @ts-check
+
+module.exports = grammar({
+  name: 'htmlmustache',
+
+  extras: ($) => [$.html_comment, /\s+/],
+
+  externals: ($) => [
+    $._html_start_tag_name,
+    $._html_script_start_tag_name,
+    $._html_style_start_tag_name,
+    $._html_raw_start_tag_name,
+    $._html_end_tag_name,
+    $.html_erroneous_end_tag_name,
+    '/>',
+    $._html_implicit_end_tag,
+    $.html_raw_text,
+    $.html_comment,
+    // Mustache externals
+    $._mustache_start_tag_name,
+    $._mustache_end_tag_name,
+    $._mustache_erroneous_end_tag_name,
+    $._mustache_end_tag_html_implicit_end_tag,
+  ],
+
+  rules: {
+    document: ($) => repeat($._node),
+
+    html_doctype: ($) =>
+      seq('<!', alias($._html_doctype, 'doctype'), /[^>]+/, '>'),
+
+    _html_doctype: (_) => /[Dd][Oo][Cc][Tt][Yy][Pp][Ee]/,
+
+    _node: ($) => choice($._html_node, $._mustache_node),
+
+    _html_node: ($) =>
+      choice(
+        $.html_doctype,
+        $.html_entity,
+        $.html_element,
+        $.html_script_element,
+        $.html_style_element,
+        $.html_raw_element,
+        $.html_erroneous_end_tag,
+        $.text,
+        alias($._text_brace, $.text),
+        alias($._text_ampersand, $.text),
+      ),
+
+    _mustache_node: ($) =>
+      choice(
+        $.mustache_triple,
+        $.mustache_comment,
+        $.mustache_partial,
+        $.mustache_section,
+        $.mustache_inverted_section,
+        $.mustache_interpolation,
+      ),
+    // Mustache rules - order matters for parsing precedence
+    mustache_triple: ($) => seq('{{{', $._mustache_expression, '}}}'),
+
+    mustache_comment: ($) =>
+      seq('{{!', alias($._mustache_content, $.mustache_comment_content), '}}'),
+
+    _mustache_content: ($) => /[^}]+/,
+
+    mustache_partial: ($) =>
+      seq('{{>', alias($._mustache_content, $.mustache_partial_content), '}}'),
+
+    mustache_interpolation: ($) => seq('{{', $._mustache_expression, '}}'),
+
+    mustache_section: ($) =>
+      seq(
+        $.mustache_section_begin,
+        repeat($._node),
+        choice($.mustache_section_end, $.mustache_erroneous_section_end),
+      ),
+
+    mustache_section_begin: ($) =>
+      seq('{{#', alias($._mustache_start_tag_name, $.mustache_tag_name), '}}'),
+
+    mustache_section_end: ($) =>
+      seq('{{/', alias($._mustache_end_tag_name, $.mustache_tag_name), '}}'),
+
+    mustache_erroneous_section_end: ($) =>
+      seq(
+        '{{/',
+        alias(
+          $._mustache_erroneous_end_tag_name,
+          $.mustache_erroneous_tag_name,
+        ),
+        '}}',
+      ),
+
+    mustache_inverted_section: ($) =>
+      seq(
+        $.mustache_inverted_section_begin,
+        repeat($._node),
+        choice(
+          $.mustache_inverted_section_end,
+          $.mustache_erroneous_inverted_section_end,
+        ),
+      ),
+
+    mustache_inverted_section_begin: ($) =>
+      seq('{{^', alias($._mustache_start_tag_name, $.mustache_tag_name), '}}'),
+
+    mustache_inverted_section_end: ($) =>
+      seq('{{/', alias($._mustache_end_tag_name, $.mustache_tag_name), '}}'),
+
+    mustache_erroneous_inverted_section_end: ($) =>
+      seq(
+        '{{/',
+        alias(
+          $._mustache_erroneous_end_tag_name,
+          $.mustache_erroneous_tag_name,
+        ),
+        '}}',
+      ),
+
+    _mustache_expression: ($) =>
+      choice($.mustache_path_expression, $.mustache_identifier, '.'),
+
+    mustache_identifier: ($) => /[a-zA-Z0-9_-]+/,
+
+    mustache_path_expression: ($) =>
+      seq($.mustache_identifier, repeat1(seq('.', $.mustache_identifier))),
+
+    html_element: ($) =>
+      choice(
+        seq(
+          $.html_start_tag,
+          repeat($._node), // 0 or more times
+          choice(
+            $.html_end_tag,
+            $._html_implicit_end_tag,
+            alias(
+              $._mustache_end_tag_html_implicit_end_tag,
+              $.html_forced_end_tag,
+            ),
+          ),
+        ),
+        $.html_self_closing_tag,
+      ),
+
+    html_script_element: ($) =>
+      seq(
+        alias($.html_script_start_tag, $.html_start_tag),
+        optional($.html_raw_text),
+        $.html_end_tag,
+      ),
+
+    html_style_element: ($) =>
+      seq(
+        alias($.html_style_start_tag, $.html_start_tag),
+        optional($.html_raw_text),
+        $.html_end_tag,
+      ),
+
+    html_raw_element: ($) =>
+      seq(
+        alias($.html_raw_start_tag, $.html_start_tag),
+        optional($.html_raw_text),
+        $.html_end_tag,
+      ),
+
+    html_start_tag: ($) =>
+      seq(
+        '<',
+        alias($._html_start_tag_name, $.html_tag_name),
+        repeat($._attribute),
+        '>',
+      ),
+
+    html_script_start_tag: ($) =>
+      seq(
+        '<',
+        alias($._html_script_start_tag_name, $.html_tag_name),
+        repeat($._attribute),
+        '>',
+      ),
+
+    html_style_start_tag: ($) =>
+      seq(
+        '<',
+        alias($._html_style_start_tag_name, $.html_tag_name),
+        repeat($._attribute),
+        '>',
+      ),
+
+    html_raw_start_tag: ($) =>
+      seq(
+        '<',
+        alias($._html_raw_start_tag_name, $.html_tag_name),
+        repeat($._attribute),
+        '>',
+      ),
+
+    html_self_closing_tag: ($) =>
+      seq(
+        '<',
+        alias($._html_start_tag_name, $.html_tag_name),
+        repeat($._attribute),
+        '/>',
+      ),
+
+    html_end_tag: ($) =>
+      seq('</', alias($._html_end_tag_name, $.html_tag_name), '>'),
+
+    html_erroneous_end_tag: ($) =>
+      seq('</', $.html_erroneous_end_tag_name, '>'),
+
+    _attribute: ($) => choice($.mustache_attribute, $.html_attribute),
+    html_attribute: ($) =>
+      seq(
+        seq(
+          $.html_attribute_name,
+          optional(
+            seq(
+              '=',
+              choice(
+                $.html_attribute_value,
+                $.html_quoted_attribute_value,
+                $.mustache_interpolation,
+              ),
+            ),
+          ),
+        ),
+      ),
+
+    mustache_attribute: ($) =>
+      choice(
+        alias(
+          $.mustache_inverted_section_attribute,
+          $.mustache_inverted_section,
+        ),
+        alias($.mustache_section_attribute, $.mustache_section),
+      ),
+
+    mustache_inverted_section_attribute: ($) =>
+      seq(
+        $.mustache_inverted_section_begin,
+        repeat1($._attribute),
+        $.mustache_inverted_section_end,
+      ),
+    mustache_section_attribute: ($) =>
+      seq(
+        $.mustache_section_begin,
+        repeat1($._attribute),
+        $.mustache_section_end,
+      ),
+
+    html_attribute_name: (_) => /[^<>{}"'/=\s]+/,
+
+    html_attribute_value: (_) => /[^<>"{}'=\s]+/,
+
+    // An entity can be named, numeric (decimal), or numeric (hexacecimal). The
+    // longest entity name is 29 characters long, and the HTML spec says that
+    // no more will ever be added.
+    html_entity: (_) =>
+      /&(#([xX][0-9a-fA-F]{1,6}|[0-9]{1,5})|[A-Za-z]{1,30});?/,
+
+    _html_attribute_value_no_single_quote: ($) => /[^'{}]+/,
+    _html_attribute_value_no_double_quote: ($) => /[^"{}]+/,
+    // Single braces that aren't part of {{ or }}
+    _single_curly_brace: ($) => /[{}]/,
+    _attribute_value_no_double_quote: ($) =>
+      choice(
+        $._mustache_node,
+        alias($._html_attribute_value_no_single_quote, $.text),
+      ),
+    _attribute_value_no_single_quote: ($) =>
+      choice(
+        $._mustache_node,
+        alias($._html_attribute_value_no_double_quote, $.text),
+      ),
+    _mustache_section_no_single_quote: ($) =>
+      seq(
+        $.mustache_section_begin,
+        repeat(
+          alias(
+            $._attribute_value_no_single_quote,
+            $._mustache_section_content,
+          ),
+        ),
+        $.mustache_section_end,
+      ),
+    _mustache_section_no_double_quote: ($) =>
+      seq(
+        $.mustache_section_begin,
+        repeat(
+          alias(
+            $._attribute_value_no_double_quote,
+            '_mustache_section_content',
+          ),
+        ),
+        $.mustache_section_end,
+      ),
+    _mustache_inverted_section_no_single_quote: ($) =>
+      seq(
+        $.mustache_inverted_section_begin,
+        repeat(
+          alias(
+            $._attribute_value_no_single_quote,
+            $._mustache_inverted_section_content,
+          ),
+        ),
+        $.mustache_inverted_section_end,
+      ),
+    _mustache_inverted_section_no_double_quote: ($) =>
+      seq(
+        $.mustache_inverted_section_begin,
+        repeat(
+          alias(
+            $._attribute_value_no_double_quote,
+            $._mustache_inverted_section_content,
+          ),
+        ),
+        $.mustache_inverted_section_end,
+      ),
+    _mustache_comment_no_single_quote: ($) =>
+      seq(
+        '{{!',
+        alias(
+          $._html_attribute_value_no_single_quote,
+          $._mustache_comment_content,
+        ),
+        '}}',
+      ),
+    _mustache_comment_no_double_quote: ($) =>
+      seq(
+        '{{!',
+        alias(
+          $._html_attribute_value_no_double_quote,
+          $._mustache_comment_content,
+        ),
+        '}}',
+      ),
+    _mustache_partial_no_single_quote: ($) =>
+      seq(
+        '{{>',
+        alias(
+          $._html_attribute_value_no_single_quote,
+          $._mustache_partial_content,
+        ),
+        '}}',
+      ),
+    _mustache_partial_no_double_quote: ($) =>
+      seq(
+        '{{>',
+        alias(
+          $._html_attribute_value_no_double_quote,
+          $._mustache_partial_content,
+        ),
+        '}}',
+      ),
+    _mustache_node_no_single_quote: ($) =>
+      choice(
+        $.mustache_interpolation,
+        alias($._mustache_comment_no_single_quote, $.mustache_comment),
+        alias($._mustache_partial_no_single_quote, $.mustache_partial),
+        alias($._mustache_section_no_single_quote, $.mustache_section),
+        alias(
+          $._mustache_inverted_section_no_single_quote,
+          $.mustache_inverted_section,
+        ),
+      ),
+    _mustache_node_no_double_quote: ($) =>
+      choice(
+        $.mustache_interpolation,
+        alias($._mustache_comment_no_double_quote, $.mustache_comment),
+        alias($._mustache_partial_no_double_quote, $.mustache_partial),
+        alias($._mustache_section_no_double_quote, $.mustache_section),
+        alias(
+          $._mustache_inverted_section_no_double_quote,
+          $.mustache_inverted_section,
+        ),
+      ),
+
+    html_quoted_attribute_value: ($) =>
+      choice(
+        seq(
+          "'",
+          repeat(
+            choice(
+              alias(
+                $._html_attribute_value_no_single_quote,
+                $.html_attribute_value,
+              ),
+              $._mustache_node_no_single_quote,
+              alias($._single_curly_brace, $.html_attribute_value),
+            ),
+          ),
+          "'",
+        ),
+        seq(
+          '"',
+          repeat(
+            choice(
+              alias(
+                $._html_attribute_value_no_double_quote,
+                $.html_attribute_value,
+              ),
+              $._mustache_node_no_double_quote,
+              alias($._single_curly_brace, $.html_attribute_value),
+            ),
+          ),
+          '"',
+        ),
+      ),
+
+    // Text content - allows > (valid in HTML text) but excludes { and } for mustache handling
+    text: (_) => /[^<{}&\s]([^<{}&]*[^<{}&\s])?/,
+    // Single braces that aren't part of {{ or }} in text content
+    _text_brace: (_) => /[{}]/,
+    _text_ampersand: (_) => '&',
+  },
+});

package/package.json

@@ -0,0 +1,82 @@
+{
+  "name": "@reteps/tree-sitter-htmlmustache",
+  "version": "0.0.18",
+  "description": "HTML with Mustache/Handlebars template syntax grammar for tree-sitter",
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/reteps/tree-sitter-htmlmustache.git"
+  },
+  "license": "MIT",
+  "author": {
+    "name": "Peter Stenger"
+  },
+  "main": "bindings/node",
+  "bin": {
+    "htmlmustache": "cli/out/check.js"
+  },
+  "types": "bindings/node",
+  "keywords": [
+    "incremental",
+    "parsing",
+    "tree-sitter",
+    "html",
+    "mustache",
+    "handlebars"
+  ],
+  "files": [
+    "grammar.js",
+    "tree-sitter.json",
+    "binding.gyp",
+    "prebuilds/**",
+    "bindings/node/*",
+    "queries/*",
+    "src/**",
+    "*.wasm",
+    "cli/out/*"
+  ],
+  "dependencies": {
+    "chalk": "^4.1.2",
+    "node-addon-api": "^8.2.2",
+    "node-gyp-build": "^4.8.2"
+  },
+  "devDependencies": {
+    "@types/node": "^22.19.11",
+    "eslint": "^9.14.0",
+    "eslint-config-treesitter": "^1.0.2",
+    "prebuildify": "^6.0.1",
+    "prettier": "^3.8.1",
+    "tree-sitter": "^0.25.0",
+    "tree-sitter-cli": "^0.26.3",
+    "typescript": "^5.7.0",
+    "vitest": "^3.0.0"
+  },
+  "peerDependencies": {
+    "tree-sitter": "^0.25.0"
+  },
+  "peerDependenciesMeta": {
+    "tree-sitter": {
+      "optional": true
+    }
+  },
+  "pnpm": {
+    "onlyBuiltDependencies": [
+      "esbuild",
+      "tree-sitter",
+      "tree-sitter-cli"
+    ]
+  },
+  "scripts": {
+    "install": "node-gyp-build",
+    "build": "tree-sitter build --wasm",
+    "prepack": "tsc -p cli/tsconfig.json",
+    "build:cli": "tsc -p cli/tsconfig.json",
+    "check": "node cli/out/check.js check",
+    "lint": "eslint .",
+    "format": "prettier --write .",
+    "format:check": "prettier --check .",
+    "prestart": "tree-sitter build --wasm",
+    "start": "tree-sitter playground",
+    "test": "tree-sitter test && node --test bindings/node/*_test.js",
+    "test:cli": "vitest run --config cli/vitest.config.ts"
+  }
+}

package/queries/highlights.scm

@@ -0,0 +1,30 @@
+; HTML
+(html_tag_name) @tag
+(html_erroneous_end_tag_name) @tag.error
+(html_doctype) @constant
+(html_attribute_name) @attribute
+(html_attribute_value) @string
+(html_comment) @comment
+
+[
+  "<"
+  ">"
+  "</"
+  "/>"
+] @punctuation.bracket
+
+; Mustache
+(mustache_tag_name) @variable
+(mustache_identifier) @variable
+(mustache_comment) @comment
+
+[
+  "{{"
+  "}}"
+  "{{{"
+  "}}}"
+  "{{>"
+  "{{#"
+  "{{/"
+  "{{^"
+] @keyword

package/queries/injections.scm

@@ -0,0 +1,7 @@
+((html_script_element
+  (html_raw_text) @injection.content)
+ (#set! injection.language "javascript"))
+
+((html_style_element
+  (html_raw_text) @injection.content)
+ (#set! injection.language "css"))

package/src/custom_raw_tags.h

@@ -0,0 +1,3 @@
+// Custom raw text tags - content inside these elements is not parsed as HTML/Mustache.
+// Uncomment and edit to add custom raw tags:
+#define CUSTOM_RAW_TAGS "MARKDOWN"