@rgrove/parse-xml 4.0.1 → 4.1.0

package/src/lib/XmlDocumentType.ts

@@ -0,0 +1,67 @@
+import { XmlNode } from './XmlNode.js';
+
+/**
+ * A document type declaration within an XML document.
+ *
+ * @example
+ *
+ * ```xml
+ * <!DOCTYPE kittens [
+ *   <!ELEMENT kittens (#PCDATA)>
+ * ]>
+ * ```
+ */
+export class XmlDocumentType extends XmlNode {
+  /**
+   * Name of the root element described by this document type declaration.
+   */
+  name: string;
+
+  /**
+   * Public identifier of the external subset of this document type declaration,
+   * or `null` if no public identifier was present.
+   */
+  publicId: string | null;
+
+  /**
+   * System identifier of the external subset of this document type declaration,
+   * or `null` if no system identifier was present.
+   */
+  systemId: string | null;
+
+  /**
+   * Internal subset of this document type declaration, or `null` if no internal
+   * subset was present.
+   */
+  internalSubset: string | null;
+
+  constructor(
+    name: string,
+    publicId?: string,
+    systemId?: string,
+    internalSubset?: string,
+  ) {
+    super();
+    this.name = name;
+    this.publicId = publicId ?? null;
+    this.systemId = systemId ?? null;
+    this.internalSubset = internalSubset ?? null;
+  }
+
+  override get type() {
+    return XmlNode.TYPE_DOCUMENT_TYPE;
+  }
+
+  override toJSON() {
+    let json = XmlNode.prototype.toJSON.call(this);
+    json.name = this.name;
+
+    for (let key of ['publicId', 'systemId', 'internalSubset'] as const) {
+      if (this[key] !== null) {
+        json[key] = this[key];
+      }
+    }
+
+    return json;
+  }
+}

package/src/lib/XmlError.ts

@@ -0,0 +1,80 @@
+/**
+ * An error that occurred while parsing XML.
+ */
+export class XmlError extends Error {
+  /**
+   * Character column at which this error occurred (1-based).
+   */
+  readonly column: number;
+
+  /**
+   * Short excerpt from the input string that contains the problem.
+   */
+  readonly excerpt: string;
+
+  /**
+   * Line number at which this error occurred (1-based).
+   */
+  readonly line: number;
+
+  /**
+   * Character position at which this error occurred relative to the beginning
+   * of the input (0-based).
+   */
+  readonly pos: number;
+
+  constructor(
+    message: string,
+    charIndex: number,
+    xml: string,
+  ) {
+    let column = 1;
+    let excerpt = '';
+    let line = 1;
+
+    // Find the line and column where the error occurred.
+    for (let i = 0; i < charIndex; ++i) {
+      let char = xml[i];
+
+      if (char === '\n') {
+        column = 1;
+        excerpt = '';
+        line += 1;
+      } else {
+        column += 1;
+        excerpt += char;
+      }
+    }
+
+    let eol = xml.indexOf('\n', charIndex);
+
+    excerpt += eol === -1
+      ? xml.slice(charIndex)
+      : xml.slice(charIndex, eol);
+
+    let excerptStart = 0;
+
+    // Keep the excerpt below 50 chars, but always keep the error position in
+    // view.
+    if (excerpt.length > 50) {
+      if (column < 40) {
+        excerpt = excerpt.slice(0, 50);
+      } else {
+        excerptStart = column - 20;
+        excerpt = excerpt.slice(excerptStart, column + 30);
+      }
+    }
+
+    super(
+      `${message} (line ${line}, column ${column})\n`
+        + `  ${excerpt}\n`
+        + ' '.repeat(column - excerptStart + 1) + '^\n',
+    );
+
+    this.column = column;
+    this.excerpt = excerpt;
+    this.line = line;
+    this.name = 'XmlError';
+    this.pos = charIndex;
+  }
+}

package/src/lib/XmlNode.ts

@@ -21,6 +21,11 @@ export class XmlNode {
    */
   static readonly TYPE_DOCUMENT = 'document';
 
+  /**
+   * Type value for an `XmlDocumentType` node.
+   */
+  static readonly TYPE_DOCUMENT_TYPE = 'doctype';
+
   /**
    * Type value for an `XmlElement` node.
    */
@@ -36,11 +41,28 @@ export class XmlNode {
    */
   static readonly TYPE_TEXT = 'text';
 
+  /**
+   * Type value for an `XmlDeclaration` node.
+   */
+  static readonly TYPE_XML_DECLARATION = 'xmldecl';
+
   /**
    * Parent node of this node, or `null` if this node has no parent.
    */
   parent: XmlDocument | XmlElement | null = null;
 
+  /**
+   * Starting byte offset of this node in the original XML string, or `-1` if
+   * the offset is unknown.
+   */
+  start = -1;
+
+  /**
+   * Ending byte offset of this node in the original XML string, or `-1` if the
+   * offset is unknown.
+   */
+  end = -1;
+
   /**
    * Document that contains this node, or `null` if this node is not associated
    * with a document.
@@ -50,10 +72,13 @@ export class XmlNode {
   }
 
   /**
-   * Whether this node is the root node of the document.
+   * Whether this node is the root node of the document (also known as the
+   * document element).
    */
   get isRootNode(): boolean {
-    return this.parent !== null && this.parent === this.document;
+    return this.parent !== null
+      && this.parent === this.document
+      && this.type === XmlNode.TYPE_ELEMENT;
   }
 
   /**
@@ -68,7 +93,7 @@ export class XmlNode {
    * @see https://www.w3.org/TR/2008/REC-xml-20081126/#sec-white-space
    */
   get preserveWhitespace(): boolean {
-    return Boolean(this.parent?.preserveWhitespace);
+    return !!this.parent?.preserveWhitespace;
   }
 
   /**
@@ -102,6 +127,11 @@ export class XmlNode {
       json.preserveWhitespace = true;
     }
 
+    if (this.start !== -1) {
+      json.start = this.start;
+      json.end = this.end;
+    }
+
     return json;
   }
 }

package/src/lib/syntax.ts

@@ -20,7 +20,7 @@ export const attValueCharSingleQuote = /[^'&<]+/y;
  *
  * @see https://www.w3.org/TR/2008/REC-xml-20081126/#AVNormalize
  */
-export const attValueNormalizedWhitespace = /[\t\n]/g;
+export const attValueNormalizedWhitespace = /\r\n|[\n\r\t]/g;
 
 /**
  * Regular expression that matches one or more characters that signal the end of