crash page on prompt dialog loop to continue: (#929)

- if a page is stuck in a window.alert / window.prompt loop, showing >10
or more consecutive dialogs (unrelated to unloading), call Page.crash()
to more quickly move on to next page, as not much else can be done.
- add exception handling in dialog accept/dismiss to avoid crawler crash
- fixes #926

Author: ikreymer

src/crawler.ts

@@ -47,6 +47,7 @@ import {
   ExitCodes,
   InterruptReason,
   BxFunctionBindings,
+  MAX_JS_DIALOG_PER_PAGE,
 } from "./util/constants.js";
 
 import { AdBlockRules, BlockRuleDecl, BlockRules } from "./util/blockrules.js";
@@ -874,30 +875,49 @@ self.__bx_behaviors.selectMainBehavior();
       await this.browser.addInitScript(page, initScript);
     }
 
+    let dialogCount = 0;
+
     // Handle JS dialogs:
     // - Ensure off-page navigation is canceled while behavior is running
     // - dismiss close all other dialogs if not blocking unload
     page.on("dialog", async (dialog) => {
       let accepted = true;
-      if (dialog.type() === "beforeunload") {
-        if (opts.pageBlockUnload) {
-          accepted = false;
-          await dialog.dismiss();
+      let msg = {};
+      try {
+        if (dialog.type() === "beforeunload") {
+          if (opts.pageBlockUnload) {
+            accepted = false;
+          }
         } else {
+          // other JS dialog, just dismiss
+          accepted = false;
+          if (dialogCount >= MAX_JS_DIALOG_PER_PAGE) {
+            // dialog likely in a loop, need to crash page to avoid being stuck
+            logger.error(
+              "JS Dialog appears to be in a loop, crashing page to continue",
+            );
+            await cdp.send("Page.crash");
+            return;
+          }
+          dialogCount++;
+        }
+        msg = {
+          accepted,
+          blockingUnload: opts.pageBlockUnload,
+          message: dialog.message(),
+          type: dialog.type(),
+          page: page.url(),
+          workerid,
+        };
+        if (accepted) {
           await dialog.accept();
+        } else {
+          await dialog.dismiss();
         }
-      } else {
-        // other JS dialog, just dismiss
-        await dialog.dismiss();
+        logger.debug("JS Dialog", msg);
+      } catch (e) {
+        logger.warn("JS Dialog Error", { ...msg, ...formatErr(e) });
       }
-      logger.debug("JS Dialog", {
-        accepted,
-        blockingUnload: opts.pageBlockUnload,
-        message: dialog.message(),
-        type: dialog.type(),
-        page: page.url(),
-        workerid,
-      });
     });
 
     // only add if running with autoclick behavior

src/util/constants.ts

@@ -43,6 +43,9 @@ export const SITEMAP_INITIAL_FETCH_TIMEOUT_SECS = 30;
 
 export const ROBOTS_CACHE_LIMIT = 100;
 
+// max JS dialogs (alert/prompt) to allow per page
+export const MAX_JS_DIALOG_PER_PAGE = 10;
+
 export type ExtractSelector = {
   selector: string;
   extract: string;