Introduce some URL.parse() usage in the code-base

This (fairly new) static method allows parsing URLs without having to wrap `new URL(...)` calls within `try...catch` blocks, thus simplifying the code; see https://developer.mozilla.org/en-US/docs/Web/API/URL/parse_static

For older browsers/environments the functionality will be polyfilled, but *only* in `legacy` builds, via `core-js`; see https://github.com/zloirock/core-js?tab=readme-ov-file#url-and-urlsearchparams

*Please note:* This is currently limited to the `src/`- and `web/`-folders, such that we don't break development/testing, since the functionality is not available in all Node.js versions that we support; see https://developer.mozilla.org/en-US/docs/Web/API/URL/parse_static#browser_compatibility

src/display/api.js

@@ -523,18 +523,20 @@ function getUrlProp(val) {
   if (val instanceof URL) {
     return val.href;
   }
-  try {
-    // The full path is required in the 'url' field.
-    return new URL(val, window.location).href;
-  } catch {
+  if (typeof val === "string") {
     if (
       typeof PDFJSDev !== "undefined" &&
       PDFJSDev.test("GENERIC") &&
-      isNodeJS &&
-      typeof val === "string"
+      isNodeJS
     ) {
       return val; // Use the url as-is in Node.js environments.
     }
+
+    // The full path is required in the 'url' field.
+    const url = URL.parse(val, window.location);
+    if (url) {
+      return url.href;
+    }
   }
   throw new Error(
     "Invalid PDF url data: " +
@@ -2080,14 +2082,9 @@ class PDFWorker {
       // Check if URLs have the same origin. For non-HTTP based URLs, returns
       // false.
       this._isSameOrigin = (baseUrl, otherUrl) => {
-        let base;
-        try {
-          base = new URL(baseUrl);
-          if (!base.origin || base.origin === "null") {
-            return false; // non-HTTP url
-          }
-        } catch {
-          return false;
+        const base = URL.parse(baseUrl);
+        if (!base?.origin || base.origin === "null") {
+          return false; // non-HTTP url
         }
         const other = new URL(otherUrl, base);
         return base.origin === other.origin;
@@ -2200,7 +2197,7 @@ class PDFWorker {
       if (
         typeof PDFJSDev !== "undefined" &&
         PDFJSDev.test("GENERIC") &&
-        !PDFWorker._isSameOrigin(window.location.href, workerSrc)
+        !PDFWorker._isSameOrigin(window.location, workerSrc)
       ) {
         workerSrc = PDFWorker._createCDNWrapper(
           new URL(workerSrc, window.location).href

src/display/display_utils.js

@@ -402,13 +402,9 @@ function isValidFetchUrl(url, baseUrl) {
   if (typeof PDFJSDev !== "undefined" && PDFJSDev.test("MOZCENTRAL")) {
     throw new Error("Not implemented: isValidFetchUrl");
   }
-  try {
-    const { protocol } = baseUrl ? new URL(url, baseUrl) : new URL(url);
-    // The Fetch API only supports the http/https protocols, and not file/ftp.
-    return protocol === "http:" || protocol === "https:";
-  } catch {
-    return false; // `new URL()` will throw on incorrect data.
-  }
+  const res = baseUrl ? URL.parse(url, baseUrl) : URL.parse(url);
+  // The Fetch API only supports the http/https protocols, and not file/ftp.
+  return res?.protocol === "http:" || res?.protocol === "https:";
 }
 
 /**

src/display/network_utils.js

@@ -33,13 +33,8 @@ function createHeaders(isHttp, httpHeaders) {
 }
 
 function getResponseOrigin(url) {
-  try {
-    return new URL(url).origin;
-  } catch {
-    // `new URL()` will throw on incorrect data.
-  }
   // Notably, null is distinct from "null" string (e.g. from file:-URLs).
-  return null;
+  return URL.parse(url)?.origin ?? null;
 }
 
 function validateRangeRequestCapabilities({

src/shared/util.js

@@ -412,35 +412,28 @@ function createValidAbsoluteUrl(url, baseUrl = null, options = null) {
   if (!url) {
     return null;
   }
-  try {
-    if (options && typeof url === "string") {
-      // Let URLs beginning with "www." default to using the "http://" protocol.
-      if (options.addDefaultProtocol && url.startsWith("www.")) {
-        const dots = url.match(/\./g);
-        // Avoid accidentally matching a *relative* URL pointing to a file named
-        // e.g. "www.pdf" or similar.
-        if (dots?.length >= 2) {
-          url = `http://${url}`;
-        }
-      }
-
-      // According to ISO 32000-1:2008, section 12.6.4.7, URIs should be encoded
-      // in 7-bit ASCII. Some bad PDFs use UTF-8 encoding; see bug 1122280.
-      if (options.tryConvertEncoding) {
-        try {
-          url = stringToUTF8String(url);
-        } catch {}
+  if (options && typeof url === "string") {
+    // Let URLs beginning with "www." default to using the "http://" protocol.
+    if (options.addDefaultProtocol && url.startsWith("www.")) {
+      const dots = url.match(/\./g);
+      // Avoid accidentally matching a *relative* URL pointing to a file named
+      // e.g. "www.pdf" or similar.
+      if (dots?.length >= 2) {
+        url = `http://${url}`;
       }
     }
 
-    const absoluteUrl = baseUrl ? new URL(url, baseUrl) : new URL(url);
-    if (_isValidProtocol(absoluteUrl)) {
-      return absoluteUrl;
+    // According to ISO 32000-1:2008, section 12.6.4.7, URIs should be encoded
+    // in 7-bit ASCII. Some bad PDFs use UTF-8 encoding; see bug 1122280.
+    if (options.tryConvertEncoding) {
+      try {
+        url = stringToUTF8String(url);
+      } catch {}
     }
-  } catch {
-    /* `new URL()` will throw on incorrect data. */
   }
-  return null;
+
+  const absoluteUrl = baseUrl ? URL.parse(url, baseUrl) : URL.parse(url);
+  return _isValidProtocol(absoluteUrl) ? absoluteUrl : null;
 }
 
 function shadow(obj, prop, value, nonSerializable = false) {