@docx4j/core-ts 0.4.0 → 0.5.0

package/dist/parts/XmlPart.mjs

@@ -25,11 +25,19 @@ const warnedPrefixes = new Set();
  * element, so the location is everything before the position or the colon - and `w:sz` must not be
  * cut at its own prefix separator, which a naive split on `:` does.
  */
+/** The runtime's located message: `w:sz/@w:val (line 12): reason`, the position optional. */
+const LOCATION_PREFIX = /^(.+?)(?: \(line \d+(?:, column \d+)?\))?: /;
 function locationOf(message, located) {
     var _a;
     if (!located)
         return undefined;
-    return (_a = /^(.+?)(?: \(line \d+(?:, column \d+)?\))?: /.exec(message)) === null || _a === void 0 ? void 0 : _a[1];
+    return (_a = LOCATION_PREFIX.exec(message)) === null || _a === void 0 ? void 0 : _a[1];
+}
+/** ` (line 12, column 9)`, or as much of it as is known. */
+function positionOf(line, column) {
+    if (line === undefined)
+        return '';
+    return column === undefined ? ` (line ${line})` : ` (line ${line}, column ${column})`;
 }
 /** The objects package's prefix table as XPath wants it, prefix to URI; built once. */
 const NAMESPACE_PREFIXES_BY_PREFIX = (() => {
@@ -48,6 +56,7 @@ const NAMESPACE_PREFIXES_BY_PREFIX = (() => {
 export class XmlPart extends Part {
     constructor(partName, contentType, relationshipType, rootName) {
         super(partName, contentType, relationshipType);
+        this.parseWarnings = [];
         this.rootName = rootName;
     }
     /** True once contents were unmarshalled or set; the part is then re-marshalled on save. */
@@ -65,6 +74,14 @@ export class XmlPart extends Part {
     get isUntouched() {
         return this.element === undefined && this.bytes === undefined;
     }
+    /**
+     * What the preprocessor reported the last time this part was parsed (CR-010 phase 1): an
+     * mc:AlternateContent whose branches' text differs (`mce.branches_differ`). Empty until the part
+     * is parsed, by `getContents()` or a private read, and again once its contents are set.
+     */
+    get warnings() {
+        return this.parseWarnings;
+    }
     /** The unmarshalled contents (docx4j getJaxbElement()); throws when not yet unmarshalled. */
     get contents() {
         if (this.element === undefined)
@@ -77,16 +94,16 @@ export class XmlPart extends Part {
     }
     /** Unmarshals on first call (docx4j getContents()). */
     async getContents() {
-        var _a, _b, _c;
+        var _a, _b;
         if (this.element === undefined) {
             const bytes = (_a = this.bytes) !== null && _a !== void 0 ? _a : (await this.loadSourceBytes());
             if (bytes === undefined)
                 throw new Docx4JException(`Part ${this.partName} has no content: set contents or load it from a container`);
-            const doc = parseXml(decodeXmlText(bytes));
+            const doc = this.parse(bytes);
             this.recordDeclarations(doc);
-            (_b = this.preprocessor) === null || _b === void 0 ? void 0 : _b.call(this, doc, this);
+            this.preprocess(doc);
             this.element = await this.unmarshal(doc);
-            (_c = this.rootName) !== null && _c !== void 0 ? _c : (this.rootName = { namespaceURI: this.element.name.namespaceURI, localPart: this.element.name.localPart });
+            (_b = this.rootName) !== null && _b !== void 0 ? _b : (this.rootName = { namespaceURI: this.element.name.namespaceURI, localPart: this.element.name.localPart });
             this.bytes = undefined;
         }
         return this.element.value;
@@ -101,17 +118,22 @@ export class XmlPart extends Part {
      * reads the styles and numbering parts of a document nobody has otherwise touched.
      */
     async readContents() {
-        var _a, _b;
+        var _a;
         if (this.element !== undefined)
             return this.element.value;
         const bytes = (_a = this.bytes) !== null && _a !== void 0 ? _a : (await this.loadSourceBytes());
         if (bytes === undefined)
             throw new Docx4JException(`Part ${this.partName} has no content: set contents or load it from a container`);
-        const doc = parseXml(decodeXmlText(bytes));
+        const doc = this.parse(bytes);
         this.recordDeclarations(doc);
-        (_b = this.preprocessor) === null || _b === void 0 ? void 0 : _b.call(this, doc, this);
+        this.preprocess(doc);
         return (await this.unmarshal(doc)).value;
     }
+    preprocess(doc) {
+        var _a;
+        const warnings = (_a = this.preprocessor) === null || _a === void 0 ? void 0 : _a.call(this, doc, this);
+        this.parseWarnings = warnings ? [...warnings] : [];
+    }
     /** Replaces the contents (docx4j setJaxbElement / setContents). The root name defaults to the part's. */
     setContents(value, rootName) {
         var _a, _b;
@@ -120,6 +142,7 @@ export class XmlPart extends Part {
             throw new Docx4JException(`Part ${this.partName}: no root element name; pass one`);
         this.element = { name: { namespaceURI: name.namespaceURI, localPart: name.localPart }, value };
         this.bytes = undefined;
+        this.parseWarnings = [];
     }
     /** Replaces the contents with raw XML bytes; they are unmarshalled on the next getContents(). */
     setBytes(bytes) {
@@ -183,28 +206,54 @@ export class XmlPart extends Part {
         this.declareIgnorablePrefixes(root);
         return root;
     }
+    /**
+     * Parses the part's bytes, and names the part when the XML is not well-formed. xmldom's error
+     * (Node) carries the line and column of the fault; a browser's says them in its text, which
+     * `parseXml` passes on.
+     */
+    parse(bytes) {
+        try {
+            return parseXml(decodeXmlText(bytes));
+        }
+        catch (cause) {
+            const at = cause === null || cause === void 0 ? void 0 : cause.locator;
+            const line = typeof (at === null || at === void 0 ? void 0 : at.lineNumber) === 'number' ? at.lineNumber : undefined;
+            const column = typeof (at === null || at === void 0 ? void 0 : at.columnNumber) === 'number' ? at.columnNumber : undefined;
+            const reason = (cause instanceof Error ? cause.message : String(cause)).replace(/^Malformed XML: /, '');
+            throw new PartUnmarshalException(`${this.partName}: Malformed XML${positionOf(line, column)}: ${reason}`, this.partName.name, { cause, line, column });
+        }
+    }
     /**
      * Unmarshals, and names the part when the runtime refuses. The runtime's own message says what is
      * wrong and, since `@docx4j/jsonix` 3.4.0, where - `w:sz/@w:val (line 12): Argument [NaN] must be
      * an integer` - but not which part it was reading, which is the first thing a person editing a
-     * part's XML needs (CR-002 section 27). The location it found is lifted onto the exception so a
-     * caller does not parse the message; `line` is undefined where the parser supplies no position,
-     * which is every browser (only xmldom carries a locator).
+     * part's XML needs (CR-002 section 27), nor the value it refused. The message names the part, and
+     * for an attribute gives the value as written and the column too:
+     * `/word/styles.xml: w:sz/@w:val="big" (line 12, column 9): Argument [NaN] must be an integer, ...`.
+     * The location it found is lifted onto the exception so a caller does not parse the message; `line`
+     * is undefined where the parser supplies no position, which is every browser (only xmldom carries a
+     * locator).
      */
     async unmarshal(doc) {
-        var _a, _b;
+        var _a, _b, _c, _d, _e, _f;
         try {
             return await unmarshalNode(doc);
         }
         catch (cause) {
             const located = cause;
             const message = typeof (located === null || located === void 0 ? void 0 : located.message) === 'string' ? located.message : String(cause);
-            throw new PartUnmarshalException(`${this.partName}: ${message}`, this.partName.name, {
-                cause,
-                location: locationOf(message, (located === null || located === void 0 ? void 0 : located.jsonixLocated) === true),
-                line: typeof ((_a = located === null || located === void 0 ? void 0 : located.node) === null || _a === void 0 ? void 0 : _a.lineNumber) === 'number' ? located.node.lineNumber : undefined,
-                column: typeof ((_b = located === null || located === void 0 ? void 0 : located.node) === null || _b === void 0 ? void 0 : _b.columnNumber) === 'number' ? located.node.columnNumber : undefined,
-            });
+            const location = locationOf(message, (located === null || located === void 0 ? void 0 : located.jsonixLocated) === true);
+            const line = typeof ((_a = located === null || located === void 0 ? void 0 : located.node) === null || _a === void 0 ? void 0 : _a.lineNumber) === 'number' ? located.node.lineNumber : undefined;
+            const column = typeof ((_b = located === null || located === void 0 ? void 0 : located.node) === null || _b === void 0 ? void 0 : _b.columnNumber) === 'number' ? located.node.columnNumber : undefined;
+            let detail = message;
+            if (location !== undefined) {
+                const reason = message.slice((_d = (_c = LOCATION_PREFIX.exec(message)) === null || _c === void 0 ? void 0 : _c[0].length) !== null && _d !== void 0 ? _d : 0);
+                const name = located === null || located === void 0 ? void 0 : located.attributeName;
+                const value = (name === null || name === void 0 ? void 0 : name.localPart) ? (_f = (_e = located === null || located === void 0 ? void 0 : located.node) === null || _e === void 0 ? void 0 : _e.getAttributeNS) === null || _f === void 0 ? void 0 : _f.call(_e, name.namespaceURI || null, name.localPart) : undefined;
+                const shown = typeof value === 'string' ? `=${JSON.stringify(value.length > 80 ? `${value.slice(0, 80)}...` : value)}` : '';
+                detail = `${location}${shown}${positionOf(line, column)}: ${reason}`;
+            }
+            throw new PartUnmarshalException(`${this.partName}: ${detail}`, this.partName.name, { cause, location, line, column });
         }
     }
     /** The source root element's namespace declarations, prefix to URI (`xmlns=` is the empty prefix). */

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@docx4j/core-ts",
-  "version": "0.4.0",
+  "version": "0.5.0",
   "description": "docx4j-core for TypeScript: Open Packaging (docx/pptx/xlsx zip and flat OPC), typed parts and relationships, style, numbering and font resolution over @docx4j/generated-objects-ts. Node, browsers, Word add-ins.",
   "type": "module",
   "main": "./dist/index.mjs",
@@ -83,7 +83,7 @@
   },
   "license": "Apache-2.0",
   "dependencies": {
-    "@docx4j/generated-objects-ts": "^0.3.1",
+    "@docx4j/generated-objects-ts": "^0.4.2",
     "fflate": "^0.8.2"
   },
   "devDependencies": {