fast-xml-validator 1.4.2 → 1.4.3

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "fast-xml-validator",
-  "version": "1.4.2",
+  "version": "1.4.3",
   "description": "Validate XML for Syntax and Business rules without C/C++ based libraries",
   "main": "./lib/fxv.cjs",
   "type": "module",
@@ -21,7 +21,8 @@
   },
   "scripts": {
     "test": "c8 --reporter=lcov --reporter=text jasmine specs/*spec.js",
-    "bundle": "webpack --config webpack.cjs.config.js"
+    "bundle": "webpack --config webpack.cjs.config.js",
+    "codemap": "codebase-map scan && codebase-map format --format dsl --include \"src/**\" > .codebase-map.dsl"
   },
   "files": [
     "lib",
@@ -52,6 +53,7 @@
     "@types/node": "20",
     "babel-loader": "^8.2.2",
     "c8": "^10.1.3",
+    "codebase-map": "^0.6.1",
     "eslint": "^8.3.0",
     "jasmine": "^5.6.0",
     "typescript": "5",

package/src/SyntaxValidator.js

@@ -100,6 +100,24 @@ export default class SyntaxValidator {
       docTypeSeen = true;
     };
 
+    // Character data (text, entity references, CDATA) is only legal inside an
+    // element. At depth 0 it is "extra" data, whether before or after the root.
+    const checkTopLevelText = (pos) => {
+      if (tags.length === 0) {
+        throwError('InvalidXml', 'Extra text at the end', getLineNumberForPosition(xmlData, pos));
+      }
+    };
+
+    const checkCdataPlacement = (startBang) => {
+      if (tags.length === 0 && xmlData.startsWith('<![CDATA[', startBang)) {
+        throwError(
+          'InvalidXml',
+          'CDATA section is not allowed outside the root element.',
+          getLineNumberForPosition(xmlData, startBang)
+        );
+      }
+    };
+
     for (let i = 0; i < xmlData.length; i++) {
       if (xmlData[i] === '<' && xmlData[i + 1] === '?') {
         i += 2;
@@ -109,6 +127,7 @@ export default class SyntaxValidator {
         i++;
 
         if (xmlData[i] === '!') {
+          checkCdataPlacement(tagStartPos);
           try {
             checkDocTypePlacement(tagStartPos);
             i = commentCdataReader.read(xmlData, i, docTypeValidator);
@@ -214,14 +233,16 @@ export default class SyntaxValidator {
               throwError('InvalidXml', 'Multiple possible root nodes found.', getLineNumberForPosition(xmlData, i));
             } else if (options.unpairedTags.indexOf(tagName) !== -1) {
               skipMatcher.pop();
+              // An unpaired tag at depth 0 is a complete top-level element.
+              if (tags.length === 0) reachedRoot = true;
             } else if (skip) {
               // Consume the whole skipped subtree now, as opaque raw text,
-              // then continue scanning right after its closing tag.
+              // then fall through to the text scan below so that whatever
+              // follows the skipped tag's closing '>' (text, entities,
+              // comments, the next tag) is handled like after any other tag.
               i = RawSubtreeReader.consume(xmlData, tagName, i + 1);
               skipMatcher.pop();
               if (tags.length === 0) reachedRoot = true;
-              tagFound = true;
-              continue;
             } else {
               tags.push({ tagName, tagStartPos });
             }
@@ -235,6 +256,7 @@ export default class SyntaxValidator {
             if (xmlData[i] === '<') {
               if (xmlData[i + 1] === '!') {
                 i++;
+                checkCdataPlacement(i - 1);
                 try {
                   checkDocTypePlacement(i - 1);
                   i = commentCdataReader.read(xmlData, i, docTypeValidator);
@@ -251,6 +273,7 @@ export default class SyntaxValidator {
                 break;
               }
             } else if (xmlData[i] === '&') {
+              checkTopLevelText(i);
               const afterAmp = this._validateAmpersand(xmlData, i);
               if (afterAmp === -1)
                 throwError('InvalidChar', "char '&' is not expected.", getLineNumberForPosition(xmlData, i));
@@ -261,7 +284,7 @@ export default class SyntaxValidator {
               if (isIllegalControlCode(code)) {
                 throwError('IllegalCharacter', `Illegal control character ${ControlCharValidator.describe(code)} in document content.`, getLineNumberForPosition(xmlData, i));
               }
-              if (reachedRoot === true && tags.length === 0 && !isWhiteSpace(xmlData[i])) {
+              if (tags.length === 0 && !isWhiteSpace(xmlData[i])) {
                 throwError('InvalidXml', 'Extra text at the end', getLineNumberForPosition(xmlData, i));
               }
               if (checkTagValueSeq) {

package/src/fxv.d.ts

@@ -1,19 +1,18 @@
-
 export type docTypeOptions = {
   /**
-   * maximum number of entity present in DOCTYPE 
-   * 
+   * maximum number of entity present in DOCTYPE
+   *
    * Defaults to `Infinity`
    */
   maxEntityCount?: number;
 
   /**
-   * maximum size of each entity present in DOCTYPE 
-   * 
+   * maximum size of each entity present in DOCTYPE
+   *
    * Defaults to `Infinity`
    */
   maxEntitySize?: number;
-}
+};
 
 export type invalidCharSequenceOptions = {
   /**
@@ -59,14 +58,14 @@ export type xmlDeclarationOptions = {
 export type validationOptions = {
   /**
    * Whether to allow attributes without value
-   * 
+   *
    * Defaults to `false`
    */
   allowBooleanAttributes?: boolean;
 
   /**
    * List of tags without closing tags
-   * 
+   *
    * Defaults to `[]`
    */
   unpairedTags?: string[];
@@ -79,10 +78,13 @@ export type validationOptions = {
   /** Controls handling of the `<?xml ... ?>` prolog. */
   xmlDeclaraion?: xmlDeclarationOptions;
 
+  /** Correctly-spelled alias of `xmlDeclaraion`. */
+  xmlDeclaration?: xmlDeclarationOptions;
+
   /**
    * Allow more than one root-level element (as if wrapped in a virtual root).
    *
-   * Defaults to `false`.
+   * Defaults to `true`.
    */
   multipleRoots?: boolean;
 
@@ -95,22 +97,46 @@ export type validationOptions = {
   skipTags?: string[];
 };
 
-export type ValidationError = {
-  err: {
-    code: string;
-    msg: string,
-    line: number,
-    col: number
-  };
-};
+/**
+ * Known error codes. The union stays open (`string & {}`) so new codes
+ * added in the future don't break consumers' exhaustive checks.
+ */
+export type ValidationErrorCode =
+  | 'InvalidXml'
+  | 'InvalidTag'
+  | 'InvalidAttr'
+  | 'InvalidChar'
+  | 'InvalidDocType'
+  | 'IllegalCharacter'
+  | 'UnquotedAttributeValue'
+  | (string & {});
+
+/**
+ * Thrown by `SyntaxValidator#validate` on the first well-formedness problem.
+ */
+export class ValidationError extends Error {
+  constructor(message: string, code: ValidationErrorCode, line: number, col: number);
+  code: ValidationErrorCode;
+  line: number;
+  col: number;
+}
 
 export class SyntaxValidator {
   constructor(options?: validationOptions);
+
+  /**
+   * @returns `true` when the document is well-formed.
+   * @throws {ValidationError} on the first problem found.
+   */
   validate(xmlData: string): true;
+
+  /**
+   * @returns `true` when the document is well-formed.
+   * @throws {ValidationError} on the first problem found.
+   */
   static validate(xmlData: string, options?: validationOptions): true;
 }
 
-
 export interface BusinessRulesValidatorOptions {
   unknownAllow?: boolean;
   boolean?: string[];
@@ -122,7 +148,7 @@ export type ValidationFailure = {
   value?: string;
   actual?: number | string;
   expected?: number | string;
-}
+};
 
 export type CustomValidatorFn = (
   value: string,
@@ -140,7 +166,7 @@ export class BusinessRulesValidator {
    *
    * @example
    * ```ts
-   * const validator = new Validator(`
+   * const validator = new BusinessRulesValidator(`
    *   <students>
    *     <student repeatable minOccurs="1">
    *       <email unique="true" nillable="false"></email>