webrecorder · ikreymer · Jan 20, 2024 · Jan 4, 2024 · Jan 4, 2024 · Jan 6, 2024
diff --git a/dist/sw.js b/dist/sw.js
diff --git a/dist/wombat.js b/dist/wombat.js
diff --git a/dist/wombatWorkers.js b/dist/wombatWorkers.js
diff --git a/examples/live-proxy/index.html b/examples/live-proxy/index.html
@@ -18,6 +18,6 @@
 <script src="./index.js"></script>
 </head>
 <body>
-<iframe id="content"></iframe>
+<iframe sandbox="allow-downloads allow-modals allow-orientation-lock allow-pointer-lock allow-popups allow-popups-to-escape-sandbox allow-presentation allow-scripts allow-same-origin allow-forms" id="content"></iframe>
 </body>
 </html>
diff --git a/examples/live-proxy/index.js b/examples/live-proxy/index.js
@@ -5,9 +5,17 @@ class WabacLiveProxy
 {
   constructor() {
     this.url = "";
+    this.ts = "";
   }
 
   async init() {
+    window.addEventListener("load", () => {
+      const iframe = document.querySelector("#content");
+      if (iframe) {
+        iframe.addEventListener("load", () => this.onIframeLoad(iframe.contentWindow.location.href));
+      }
+    });
+
     const scope = "./";
 
     // also add inject of custom.js as a script into each replayed page
@@ -34,11 +42,12 @@ class WabacLiveProxy
       file: {"sourceUrl": `proxy:${proxyPrefix}`},
       skipExisting: false,
       extraConfig: {
-        "prefix": proxyPrefix, 
-        "isLive": true,
+        "prefix": proxyPrefix,
+        "isLive": false,
         "baseUrl": baseUrl.href,
         "baseUrlHashReplay": true,
-        "noPostToGet": true
+        "noPostToGet": true,
+        "archivePrefix": "https://web.archive.org/web/"
       },
     };
 
@@ -69,20 +78,31 @@ class WabacLiveProxy
     const m = window.location.hash.slice(1).match(/\/?(?:([\d]+)\/)?(.*)/);
 
     const url = m && m[2] || "https://example.com/";
+    const ts = m && m[1] || "";
 
     // don't change if same url
-    if (url === this.url) {
+    if (url === this.url && ts === this.ts) {
       return;
     }
 
-    let iframeUrl = `/w/liveproxy/mp_/${url}`;
+    let iframeUrl = ts ? `/w/liveproxy/${ts}mp_/${url}` : `/w/liveproxy/mp_/${url}`;
 
     const iframe = document.querySelector("#content");
     iframe.src = iframeUrl;
 
     this.url = url;
+    this.ts = ts;
+
+    window.location.hash = ts ? `#${ts}/${url}` : `#${url}`;
+  }
+
+  onIframeLoad(url) {
+    const m = url.match(/liveproxy\/([\d]+)?\w\w_\/(.*)/);
+
+    this.ts = m[1] || "";
+    this.url = m[2] || "";
 
-    window.location.hash = `#${url}`;
+    window.location.hash = this.ts ? `#${this.ts}/${this.url}` : `#${this.url}`;
   }
 }
 

diff --git a/package.json b/package.json
@@ -1,14 +1,14 @@
 {
   "name": "@webrecorder/wabac",
-  "version": "2.16.13",
+  "version": "2.17.0",
   "main": "index.js",
   "type": "module",
   "license": "AGPL-3.0-or-later",
   "dependencies": {
     "@peculiar/asn1-ecc": "^2.3.4",
     "@peculiar/asn1-schema": "^2.3.3",
     "@peculiar/x509": "^1.9.2",
-    "@webrecorder/wombat": "^3.6.1",
+    "@webrecorder/wombat": "^3.7.0",
     "acorn": "^8.10.0",
     "auto-js-ipfs": "^2.1.1",
     "base64-js": "^1.5.1",

diff --git a/src/liveproxy.js b/src/liveproxy.js
@@ -98,6 +98,7 @@ export class LiveProxy {
       date: new Date(),
       noRW: false,
       isLive: this.isLive,
+      archivePrefix: this.archivePrefix,
     });
 
     if (clonedResponse) {

diff --git a/src/response.js b/src/response.js
@@ -9,7 +9,7 @@ const decoder = new TextDecoder();
 class ArchiveResponse
 {
 
-  static fromResponse({url, response, date, noRW, isLive}) {
+  static fromResponse({url, response, date, noRW, isLive, archivePrefix}) {
     const payload = response.body ? new AsyncIterReader(response.body.getReader(), false) : null;
     const status = Number(response.headers.get("x-redirect-status") || response.status);
     const statusText = response.headers.get("x-redirect-statusText") || response.statusText;
@@ -21,6 +21,12 @@ class ArchiveResponse
       if (origLoc.startsWith(self.location.origin)) {
         origLoc = origLoc.slice(self.location.origin.length);
       }
+      if (archivePrefix && origLoc.startsWith(archivePrefix)) {
+        const inx = origLoc.indexOf("/http");
+        if (inx > 0) {
+          origLoc = origLoc.slice(inx + 1);
+        }
+      }
       headers.set("location", origLoc);
       headers.delete("x-orig-location");
       headers.delete("x-redirect-status");

diff --git a/src/rewrite/html.js b/src/rewrite/html.js
@@ -134,15 +134,17 @@ class HTMLRewriter
     const tagName = tag.tagName;
 
     for (let attr of tag.attrs) {
-      const name = attr.name;
-      const value = attr.value;
+      const name = attr.name || "";
+      const value = attr.value || "";
 
-      // js attrs
-      if (name.startsWith("on") && value.startsWith("javascript:") && name.slice(2, 3) != "-") {
+      // js attrs with javascript:
+      if (value.startsWith("javascript:")) {
         attr.value = "javascript:" + rewriter.rewriteJS(value.slice("javascript:".length), {inline: true});
-      }
+      } else if (name.startsWith("on") && name.slice(2, 3) != "-") {
+      // js attrs
+        attr.value = rewriter.rewriteJS(value, {inline: true});
       // css attrs
-      else if (name === "style") {
+      } else if (name === "style") {
         attr.value = rewriter.rewriteCSS(attr.value);
       }
 

diff --git a/src/rewrite/index.js b/src/rewrite/index.js
@@ -298,7 +298,7 @@ class Rewriter {
 
   // JS
   rewriteJS(text, opts) {
-    const noUrlProxyRewrite = opts && !opts.rewriteUrl && opts.isModule === undefined;
+    const noUrlProxyRewrite = opts && !opts.rewriteUrl && opts.isModule === undefined && !opts.inline;
     const dsRules = noUrlProxyRewrite ? baseRules : this.dsRules;
     const dsRewriter = dsRules.getRewriter(this.baseUrl);
 

diff --git a/src/rewrite/jsrewriter.js b/src/rewrite/jsrewriter.js
@@ -176,22 +176,35 @@ if (!self.__WB_pmw) { self.__WB_pmw = function(obj) { this.__WB_source = obj; re
     return false;
   }
 
-  parseLetConstGlobals(text) {
+  parseGlobals(text) {
     const res = acorn.parse(text, {ecmaVersion: "latest"});
 
     const names = [];
 
+    const excludeOverrides = new Set();
+
     for (const expr of res.body) {
       const { type, kind, declarations } = expr;
-      if (type === "VariableDeclaration" && (kind === "const" || kind === "let")) {
-        const decl = declarations.length && declarations[0];
-        if (decl && decl.type === "VariableDeclarator") {
-          const name = decl.id && decl.id.name;
-          names.push(`self.${name} = ${name};`);
+      if (type === "VariableDeclaration") {
+        for (const decl of declarations) {
+          if (decl && decl.type === "VariableDeclarator" && decl.id && decl.id.type === "Identifier") {
+            const name = decl.id.name;
+
+            if (GLOBAL_OVERRIDES.includes(name)) {
+              excludeOverrides.add(name);
+            } else if (kind === "const" || kind === "let") {
+              names.push(`self.${name} = ${name};`);
+            }
+          }
         }
       }
     }
 
+    if (excludeOverrides.size) {
+      const filteredGlobals = GLOBAL_OVERRIDES.filter(x => !excludeOverrides.has(x));
+      this.firstBuff = this.initLocalDecl(filteredGlobals);
+    }
+
     if (names.length) {
       return "\n" + names.join("\n");
     } else {
@@ -227,20 +240,23 @@ if (!self.__WB_pmw) { self.__WB_pmw = function(obj) { this.__WB_source = obj; re
 
     const wrapGlobals = GLOBALS_RX.exec(text);
 
+    if (opts.inline) {
+      newText = newText.replace(/\n/g, " ") ;
+    }
+
     if (wrapGlobals) {
       let hoistGlobals = "";
       if (newText) {
         try {
-          hoistGlobals = this.parseLetConstGlobals(newText);
+          hoistGlobals = this.parseGlobals(newText);
         } catch (e) {
-          console.warn("acorn parsing failed on: " + newText);
+          console.warn(`acorn parsing failed, script len ${newText.length}`);
         }
       }
       newText = this.firstBuff + newText + hoistGlobals + this.lastBuff;
-    }
-
-    if (opts && opts.inline) {
-      newText = newText.replace(/\n/g, " ") ;
+      if (opts.inline) {
+        newText = newText.replace(/\n/g, " ") ;
+      }
     }
 
     return newText;

diff --git a/test/rewriteHTML.js b/test/rewriteHTML.js
@@ -49,6 +49,10 @@ ${text}
 
 }
 
+function wrapScriptInline(text) {
+  return wrapScript(text).replace(/\n/g, " ").replace(/["]/g, "&quot;").replace(/[&][&]/g, "&amp;&amp;");
+}
+
 
 function wrapScriptModule(text) {
   return `\
@@ -275,6 +279,12 @@ test("srcset", rewriteHtml,
   "<img srcset=\"\">",
   "<img srcset=\"\">");
 
+// background attr
+test("background", rewriteHtml,
+  "<td background=\"https://example.com/\"></td>",
+  "<td background=\"http://localhost:8080/prefix/20201226101010mp_/https://example.com/\"></td>"
+);
+
 // SCRIPT Tag
 // pywb diff: no script url rewriting!
 test("script proxy wrapped", rewriteHtml,
@@ -289,6 +299,17 @@ test("script not wrapped", rewriteHtml,
   {useBaseRules: true}
 );
 
+// inline attr rewrite
+test("inline attr rewrite", rewriteHtml,
+  "<body onload=\"window.location.href = '/path.html'\"></body>",
+  `<body onload="${wrapScriptInline("window.location.href = '/path.html'")}"></body>`
+);
+
+test("inline attr rewrite with javascript: prefix", rewriteHtml,
+  "<body onload=\"javascript:window.location.href = '/path.html'\"></body>",
+  `<body onload="javascript:${wrapScriptInline("window.location.href = '/path.html'")}"></body>`
+);
+
 // no rewriting if no props
 test("script", rewriteHtml,
   "<script>var foo = \"http://example.com/a/b/c.html\"</script>",

diff --git a/yarn.lock b/yarn.lock
@@ -398,10 +398,10 @@
   resolved "https://registry.yarnpkg.com/@webpack-cli/serve/-/serve-1.5.2.tgz#ea584b637ff63c5a477f6f21604b5a205b72c9ec"
   integrity sha512-vgJ5OLWadI8aKjDlOH3rb+dYyPd2GTZuQC/Tihjct6F9GpXGZINo3Y/IVuZVTM1eDQB+/AOsjPUWH/WySDaXvw==
 
-"@webrecorder/wombat@^3.6.1":
-  version "3.6.1"
-  resolved "https://registry.yarnpkg.com/@webrecorder/wombat/-/wombat-3.6.1.tgz#55e23a2c0df28ad8e6d67b87fffd913a101f6982"
-  integrity sha512-yJ5NBf/ERCBz07cFA4phIQCpj6U/28HWFk67V+yg3S3GSAAmlFP8dC9g1z+JO77PKuAe/qlrzO6CcS0VyUr5Ww==
+"@webrecorder/wombat@^3.7.0":
+  version "3.7.0"
+  resolved "https://registry.yarnpkg.com/@webrecorder/wombat/-/wombat-3.7.0.tgz#809e47c62b6acd6d5435708909d0d88c1090b0c3"
+  integrity sha512-xSDjuywIp8tmjIW5RqgeISgXhabqbbt1LcgrgBLUdqz1JHzJTGJ/OjmiNgV8bT+f6vtIbL+mtU8qxc86n/BgtA==
   dependencies:
     warcio "^2.2.0"