Merge pull request #19493 from Snuffleupagus/URL-parse

Introduce some `URL.parse()` usage in the code-base
mozilla · Feb 21, 2025 · e3ea926 · e3ea926
2 parents 34ef74c + c2e3330
commit e3ea926
Show file tree

Hide file tree

Showing 5 changed files with 52 additions and 72 deletions.
diff --git a/src/display/api.js b/src/display/api.js
@@ -525,18 +525,20 @@ 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: " +
@@ -2082,14 +2084,9 @@ 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;
@@ -2202,7 +2199,7 @@ 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

diff --git a/src/display/display_utils.js b/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:";
 }
 
 /**

diff --git a/src/display/network_utils.js b/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({

diff --git a/src/shared/util.js b/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) {

diff --git a/web/app.js b/web/app.js
@@ -2274,35 +2274,34 @@ if (typeof PDFJSDev === "undefined" || !PDFJSDev.test("MOZCENTRAL")) {
 }
 
 if (typeof PDFJSDev === "undefined" || PDFJSDev.test("GENERIC")) {
-  const HOSTED_VIEWER_ORIGINS = [
+  const HOSTED_VIEWER_ORIGINS = new Set([
     "null",
     "http://mozilla.github.io",
     "https://mozilla.github.io",
-  ];
+  ]);
   // eslint-disable-next-line no-var
   var validateFileURL = function (file) {
     if (!file) {
       return;
     }
-    try {
-      const viewerOrigin = new URL(window.location.href).origin || "null";
-      if (HOSTED_VIEWER_ORIGINS.includes(viewerOrigin)) {
-        // Hosted or local viewer, allow for any file locations
-        return;
-      }
-      const fileOrigin = new URL(file, window.location.href).origin;
-      // Removing of the following line will not guarantee that the viewer will
-      // start accepting URLs from foreign origin -- CORS headers on the remote
-      // server must be properly configured.
-      if (fileOrigin !== viewerOrigin) {
-        throw new Error("file origin does not match viewer's");
-      }
-    } catch (ex) {
-      PDFViewerApplication._documentError("pdfjs-loading-error", {
-        message: ex.message,
-      });
-      throw ex;
+    const viewerOrigin = URL.parse(window.location)?.origin || "null";
+    if (HOSTED_VIEWER_ORIGINS.has(viewerOrigin)) {
+      // Hosted or local viewer, allow for any file locations
+      return;
+    }
+    const fileOrigin = URL.parse(file, window.location)?.origin;
+    if (fileOrigin === viewerOrigin) {
+      return;
     }
+    const ex = new Error("file origin does not match viewer's");
+
+    PDFViewerApplication._documentError("pdfjs-loading-error", {
+      message: ex.message,
+    });
+    // Removing of the following line will not guarantee that the viewer will
+    // start accepting URLs from foreign origin -- CORS headers on the remote
+    // server must be properly configured.
+    throw ex;
   };
 
   // eslint-disable-next-line no-var