@willbooster/tree-sitter-xml 0.0.0-stage → 1.0.0

package/package.json

@@ -1,6 +1,80 @@
 {
   "name": "@willbooster/tree-sitter-xml",
-  "version": "0.0.0-stage",
-  "stub": true,
-  "description": "Temporary package placeholder for staged publishing"
-}
+  "version": "1.0.0",
+  "description": "XML & DTD grammars for tree-sitter",
+  "keywords": [
+    "incremental",
+    "parsing",
+    "tree-sitter",
+    "xml",
+    "dtd"
+  ],
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/WillBooster/tree-sitter-xml.git"
+  },
+  "license": "MIT",
+  "author": "WillBooster Inc.",
+  "files": [
+    "common/common.mjs",
+    "xml/grammar.js",
+    "xml/package.json",
+    "dtd/grammar.js",
+    "dtd/package.json",
+    "xml/src/node-types.json",
+    "dtd/src/node-types.json",
+    "queries/**/*.scm",
+    "tree-sitter.json",
+    "*.wasm"
+  ],
+  "scripts": {
+    "build-wasm": "script/generate --build-wasm",
+    "build/ci": "bun run build-wasm",
+    "cleanup": "bun wb lint --fix --format",
+    "format": "bun wb lint --format",
+    "generate": "script/generate",
+    "lint": "bun wb lint",
+    "lint-fix": "bun wb lint --fix",
+    "prepare": "(bun lefthook install || true) && (test ! -f .lefthook/normalize-bun-lockfile.sh || bash .lefthook/normalize-bun-lockfile.sh)",
+    "release": "wb release",
+    "prestart": "bun run build-wasm",
+    "start": "script/tree-sitter playground",
+    "test": "bun wb test",
+    "test/ci": "bun wb test-on-ci",
+    "test/ci-setup": "playwright install --with-deps chromium",
+    "typecheck": "bun wb typecheck",
+    "verify": "bun wb verify",
+    "verify-full": "bun wb verify --full"
+  },
+  "devDependencies": {
+    "@tsconfig/bun": "1.0.11",
+    "@types/bun": "1.4.2",
+    "@vitest/browser-playwright": "5.0.3",
+    "@willbooster/oxfmt-config": "1.2.2",
+    "@willbooster/oxlint-config": "1.4.8",
+    "@willbooster/wb": "22.26.1",
+    "@willbooster/web-tree-sitter": "2.0.28",
+    "conventional-changelog-conventionalcommits": "10.4.0",
+    "lefthook": "2.2.1",
+    "oxfmt": "0.72.0",
+    "oxlint": "1.87.0",
+    "oxlint-tsgolint": "7.0.2004",
+    "playwright": "1.63.0",
+    "semantic-release": "25.0.9",
+    "sort-package-json": "4.0.0",
+    "vitest": "5.0.3",
+    "wrangler": "4.147.0"
+  },
+  "peerDependencies": {
+    "@willbooster/web-tree-sitter": "^2.0.0"
+  },
+  "peerDependenciesMeta": {
+    "@willbooster/web-tree-sitter": {
+      "optional": true
+    }
+  },
+  "publishConfig": {
+    "access": "public",
+    "registry": "https://registry.npmjs.org/"
+  }
+}

package/queries/dtd/highlights.scm

@@ -0,0 +1,129 @@
+;; XML declaration
+
+"xml" @keyword
+
+[ "version" "encoding" ] @property
+
+(EncName) @string.special
+
+(VersionNum) @number
+
+;; Processing instructions
+
+(PI) @embedded
+
+(PI (PITarget) @keyword)
+
+;; Element declaration
+
+(elementdecl
+  "ELEMENT" @keyword
+  (Name) @tag)
+
+(contentspec
+  (_ (Name) @property))
+
+"#PCDATA" @type.builtin
+
+[ "EMPTY" "ANY" ] @string.special.symbol
+
+[ "*" "?" "+" ] @operator
+
+;; Entity declaration
+
+(GEDecl
+  "ENTITY" @keyword
+  (Name) @constant)
+
+(GEDecl (EntityValue) @string)
+
+(NDataDecl
+  "NDATA" @keyword
+  (Name) @label)
+
+;; Parsed entity declaration
+
+(PEDecl
+  "ENTITY" @keyword
+  "%" @operator
+  (Name) @constant)
+
+(PEDecl (EntityValue) @string)
+
+;; Notation declaration
+
+(NotationDecl
+  "NOTATION" @keyword
+  (Name) @constant)
+
+(NotationDecl
+  (ExternalID
+    (SystemLiteral (URI) @string.special)))
+
+;; Attlist declaration
+
+(AttlistDecl
+  "ATTLIST" @keyword
+  (Name) @tag)
+
+(AttDef (Name) @property)
+
+(AttDef (Enumeration (Nmtoken) @string))
+
+(DefaultDecl (AttValue) @string)
+
+[
+  (StringType)
+  (TokenizedType)
+] @type.builtin
+
+(NotationType "NOTATION" @type.builtin)
+
+[
+  "#REQUIRED"
+  "#IMPLIED"
+  "#FIXED"
+] @attribute
+
+;; Entities
+
+(EntityRef) @constant
+
+((EntityRef) @constant.builtin
+ (#any-of? @constant.builtin
+   "&" "<" ">" """ "'"))
+
+(CharRef) @constant
+
+(PEReference) @constant
+
+;; External references
+
+[ "PUBLIC" "SYSTEM" ] @keyword
+
+(PubidLiteral) @string.special
+
+(SystemLiteral (URI) @markup.link)
+
+;; Delimiters & punctuation
+
+[
+  "<?" "?>"
+  "<!" ">"
+  "<![" "]]>"
+] @punctuation.delimiter
+
+[ "(" ")" "[" ] @punctuation.bracket
+
+[ "\"" "'" ] @punctuation.delimiter
+
+[ "," "|" "=" ] @operator
+
+;; Misc
+
+; FIXME: enable when ts test is fixed
+; [ "INCLUDE" "IGNORE" ] @keyword
+
+(Comment) @comment
+
+(ERROR) @error

package/queries/xml/highlights.scm

@@ -0,0 +1,168 @@
+;; XML declaration
+
+"xml" @keyword
+
+[ "version" "encoding" "standalone" ] @property
+
+(EncName) @string.special
+
+(VersionNum) @number
+
+[ "yes" "no" ] @boolean
+
+;; Processing instructions
+
+(PI) @embedded
+
+(PI (PITarget) @keyword)
+
+;; Element declaration
+
+(elementdecl
+  "ELEMENT" @keyword
+  (Name) @tag)
+
+(contentspec
+  (_ (Name) @property))
+
+"#PCDATA" @type.builtin
+
+[ "EMPTY" "ANY" ] @string.special.symbol
+
+[ "*" "?" "+" ] @operator
+
+;; Entity declaration
+
+(GEDecl
+  "ENTITY" @keyword
+  (Name) @constant)
+
+(GEDecl (EntityValue) @string)
+
+(NDataDecl
+  "NDATA" @keyword
+  (Name) @label)
+
+;; Parsed entity declaration
+
+(PEDecl
+  "ENTITY" @keyword
+  "%" @operator
+  (Name) @constant)
+
+(PEDecl (EntityValue) @string)
+
+;; Notation declaration
+
+(NotationDecl
+  "NOTATION" @keyword
+  (Name) @constant)
+
+(NotationDecl
+  (ExternalID
+    (SystemLiteral (URI) @string.special)))
+
+;; Attlist declaration
+
+(AttlistDecl
+  "ATTLIST" @keyword
+  (Name) @tag)
+
+(AttDef (Name) @property)
+
+(AttDef (Enumeration (Nmtoken) @string))
+
+(DefaultDecl (AttValue) @string)
+
+[
+  (StringType)
+  (TokenizedType)
+] @type.builtin
+
+(NotationType "NOTATION" @type.builtin)
+
+[
+  "#REQUIRED"
+  "#IMPLIED"
+  "#FIXED"
+] @attribute
+
+;; Entities
+
+(EntityRef) @constant
+
+((EntityRef) @constant.builtin
+ (#any-of? @constant.builtin
+   "&" "<" ">" """ "'"))
+
+(CharRef) @constant
+
+(PEReference) @constant
+
+;; External references
+
+[ "PUBLIC" "SYSTEM" ] @keyword
+
+(PubidLiteral) @string.special
+
+(SystemLiteral (URI) @markup.link)
+
+;; Processing instructions
+
+(XmlModelPI "xml-model" @keyword)
+
+(StyleSheetPI "xml-stylesheet" @keyword)
+
+(PseudoAtt (Name) @property)
+
+(PseudoAtt (PseudoAttValue) @string)
+
+;; Doctype declaration
+
+(doctypedecl "DOCTYPE" @keyword)
+
+(doctypedecl (Name) @type)
+
+;; Tags
+
+(STag (Name) @tag)
+
+(ETag (Name) @tag)
+
+(EmptyElemTag (Name) @tag)
+
+;; Attributes
+
+(Attribute (Name) @property)
+
+(Attribute (AttValue) @string)
+
+;; Delimiters & punctuation
+
+[
+ "<?" "?>"
+ "<!" "]]>"
+ "<" ">"
+ "</" "/>"
+] @punctuation.delimiter
+
+[ "(" ")" "[" "]" ] @punctuation.bracket
+
+[ "\"" "'" ] @punctuation.delimiter
+
+[ "," "|" "=" ] @operator
+
+;; Text
+
+(CharData) @markup
+
+(CDSect
+  (CDStart) @markup.heading
+  (CData) @markup.raw
+  "]]>" @markup.heading)
+
+;; Misc
+
+(Comment) @comment
+
+(ERROR) @error

package/tree-sitter.json

@@ -0,0 +1,61 @@
+{
+  "$schema": "https://tree-sitter.github.io/tree-sitter/assets/schemas/config.schema.json",
+  "grammars": [
+    {
+      "name": "xml",
+      "camelcase": "XML",
+      "scope": "source.xml",
+      "path": "xml",
+      "external-files": [
+        "common/scanner.h"
+      ],
+      "file-types": [
+        "xml",
+        "svg",
+        "xsd",
+        "xslt",
+        "xsl",
+        "rng"
+      ],
+      "highlights": "queries/xml/highlights.scm",
+      "injection-regex": "^(xml|svg|xsd|xslt|xsl|rng)$"
+    },
+    {
+      "name": "dtd",
+      "camelcase": "DTD",
+      "scope": "source.dtd",
+      "path": "dtd",
+      "external-files": [
+        "common/scanner.h"
+      ],
+      "file-types": [
+        "dtd"
+      ],
+      "highlights": "queries/dtd/highlights.scm",
+      "injection-regex": "^dtd$"
+    }
+  ],
+  "metadata": {
+    "version": "1.0.0",
+    "license": "MIT",
+    "description": "XML & DTD grammars for tree-sitter",
+    "authors": [
+      {
+        "name": "WillBooster Inc."
+      }
+    ],
+    "links": {
+      "repository": "https://github.com/WillBooster/tree-sitter-xml"
+    }
+  },
+  "bindings": {
+    "c": false,
+    "go": false,
+    "java": false,
+    "node": false,
+    "python": false,
+    "rust": true,
+    "swift": false,
+    "zig": false
+  }
+}

package/xml/grammar.js

@@ -0,0 +1,196 @@
+/**
+ * @file Tree-sitter grammar definition for XML
+ * @author ObserverOfTime
+ * @license MIT
+ */
+
+/// <reference types="tree-sitter-cli/dsl" />
+// @ts-check
+
+import * as c from '../common/common.mjs';
+
+const O = optional;
+
+export default grammar({
+  name: 'xml',
+
+  externals: $ => [
+    // DTD
+    $.PITarget,
+    $._pi_content,
+    $.Comment,
+
+    // XML
+    $.CharData,
+    $.CData,
+    'xml-model',
+    'xml-stylesheet',
+    $._start_tag_name,
+    $._end_tag_name,
+    $._erroneous_end_name,
+    '/>',
+  ],
+
+  extras: _ => [],
+
+  supertypes: $ => [
+    $._markupdecl,
+    $._AttType,
+    $._EnumeratedType,
+    $._EntityDecl,
+    $._Reference,
+  ],
+
+  conflicts: $ => [
+    [$.AttlistDecl, $.AttDef]
+  ],
+
+  word: $ => $.Name,
+
+  rules: {
+    document: $ => prec(2, seq(
+      O($._S),
+      O($.prolog),
+      field('root', $.element),
+      repeat($._Misc),
+    )),
+
+    prolog: $ => choice(
+      seq(
+        $.XMLDecl,
+        repeat($._Misc),
+      ),
+      seq(
+        O($.XMLDecl),
+        repeat($._Misc),
+        $.doctypedecl,
+        repeat($._Misc),
+      ),
+      repeat1($._Misc)
+    ),
+
+    _Misc: $ => choice(
+      $.PI,
+      $.StyleSheetPI,
+      $.XmlModelPI,
+      $.Comment,
+      $._S
+    ),
+
+    XMLDecl: $ => seq(
+      '<?',
+      'xml',
+      $._VersionInfo,
+      O($._EncodingDecl),
+      O($._SDDecl),
+      O($._S),
+      '?>'
+    ),
+
+    _SDDecl: $ => seq(
+      $._S,
+      'standalone',
+      $._Eq,
+      c.str(choice('yes', 'no'))
+    ),
+
+    doctypedecl: $ => seq(
+      '<!',
+      'DOCTYPE',
+      $._S,
+      $.Name,
+      O(seq($._S, $.ExternalID)),
+      O($._S),
+      O(seq(
+        '[',
+        O(choice(
+          $._intSubset,
+          $._S
+        )),
+        ']',
+        O($._S)
+      )),
+      '>'
+    ),
+
+    _intSubset: $ => c.rseq1(
+      O($._S),
+      $._markupdecl,
+      $._DeclSep
+    ),
+
+    element: $ => choice(
+      $.EmptyElemTag,
+      seq($.STag, O($.content), $.ETag)
+    ),
+
+    EmptyElemTag: $ => seq(
+      '<',
+      alias($._start_tag_name, $.Name),
+      c.rseq($._S, $.Attribute),
+      O($._S),
+      '/>'
+    ),
+
+    Attribute: $ => seq($.Name, $._Eq, $.AttValue),
+
+    STag: $ => seq(
+      '<',
+      alias($._start_tag_name, $.Name),
+      c.rseq($._S, $.Attribute),
+      O($._S),
+      '>'
+    ),
+
+    ETag: $ => seq('</', alias($._end_tag_name, $.Name), O($._S), '>'),
+
+    _ErroneousETag: $ => seq(
+      '</',
+      alias($._erroneous_end_name, $.ERROR),
+      O($._S),
+      '>',
+    ),
+
+    content: $ => repeat1(
+      choice(
+        $.CharData,
+        $.element,
+        $._Reference,
+        $.CDSect,
+        $.PI,
+        $.Comment
+      )
+    ),
+
+    CDSect: $ => prec.left(
+      seq($.CDStart, optional($.CData), ']]>')
+    ),
+
+    CDStart: _ => seq('<![', 'CDATA', '['),
+
+    StyleSheetPI: $ => seq(
+      '<?',
+      'xml-stylesheet',
+      c.rseq($._S, $.PseudoAtt),
+      O($._S),
+      '?>'
+    ),
+
+    XmlModelPI: $ => seq(
+      '<?',
+      'xml-model',
+      c.rseq($._S, $.PseudoAtt),
+      O($._S),
+      '?>'
+    ),
+
+    PseudoAtt: $ => seq($.Name, $._Eq, $.PseudoAttValue),
+
+    PseudoAttValue: $ => choice(
+      c.att_value($, '"'),
+      c.att_value($, "'")
+    ),
+
+    ...c.rules
+  }
+});

package/xml/package.json

@@ -0,0 +1,4 @@
+{
+  "private": true,
+  "type": "module"
+}