single-file-cli 2.0.21 → 2.0.23

package/README.MD

@@ -69,7 +69,7 @@ Make sure Chrome or a Chromium-based browser is installed in the default folder.
   
     ```sh
     npm install "single-file-cli"
-    cd single-file-cli-master
+    cd node_modules/single-file-cli
     ```
 
     You can also install SingleFile globally with `-g` when running `npm install`.

package/deno.json

@@ -1,13 +1,13 @@
 {
   "name": "@single-file/single-file-cli",
-  "version": "2.0.21",
+  "version": "2.0.23",
   "description": "SingleFile CLI",
   "nodeModulesDir": true,
   "exports": {
     ".": "./single-file-cli-api.js"
   },
   "imports": {
-    "simple-cdp": "jsr:@simple-cdp/simple-cdp@^1.8.4",
+    "simple-cdp": "jsr:@simple-cdp/simple-cdp@^1.8.5",
     "path": "jsr:@std/path@^0.220.1"
   }
 }

package/deno.lock

@@ -2,15 +2,15 @@
   "version": "3",
   "packages": {
     "specifiers": {
-      "jsr:@simple-cdp/simple-cdp@^1.8.4": "jsr:@simple-cdp/simple-cdp@1.8.4",
+      "jsr:@simple-cdp/simple-cdp@^1.8.5": "jsr:@simple-cdp/simple-cdp@1.8.5",
       "jsr:@std/assert@^0.220.1": "jsr:@std/assert@0.220.1",
       "jsr:@std/path@^0.220.1": "jsr:@std/path@0.220.1",
-      "npm:simple-cdp@^1.8.4": "npm:simple-cdp@1.8.4",
+      "npm:simple-cdp@^1.8.5": "npm:simple-cdp@1.8.5",
       "npm:ws@^8.16.0": "npm:ws@8.16.0"
     },
     "jsr": {
-      "@simple-cdp/simple-cdp@1.8.4": {
-        "integrity": "5e31ab90a552f9f1cf84e6cf0454d21c5b9dc1d89f9c1f9a750772b47bad9a4e"
+      "@simple-cdp/simple-cdp@1.8.5": {
+        "integrity": "7386ca561cd6293bb3115eadf14b8c98d1c41c35a69c79774ea70e5101159413"
       },
       "@std/assert@0.220.1": {
         "integrity": "88710d54f3afdd7a5761e7805abba1f56cd14e4b212feffeb3e73a9f77482425"
@@ -23,8 +23,8 @@
       }
     },
     "npm": {
-      "simple-cdp@1.8.4": {
-        "integrity": "sha512-fsvMUyz13vzX8nxSPwJfdxdROGxRDdTAFX7XitVpRsKwClowsyL3NZ6kR4vR5tZfxbz31p3uoM14eFUvhaWisA==",
+      "simple-cdp@1.8.5": {
+        "integrity": "sha512-bpXoeCk0mwFrr+8C8vVkhCug5h4HO49GI+MZdnyLQcFVKzTlokrimTjEoc69X2Y2n7mtjzDHTriq6B/kgUeUcA==",
         "dependencies": {}
       },
       "ws@8.16.0": {
@@ -36,12 +36,12 @@
   "remote": {},
   "workspace": {
     "dependencies": [
-      "jsr:@simple-cdp/simple-cdp@^1.8.4",
+      "jsr:@simple-cdp/simple-cdp@^1.8.5",
       "jsr:@std/path@^0.220.1"
     ],
     "packageJson": {
       "dependencies": [
-        "npm:simple-cdp@^1.8.4",
+        "npm:simple-cdp@^1.8.5",
         "npm:ws@^8.16.0"
       ]
     }

package/lib/cdp-client.js

@@ -28,7 +28,7 @@ import { getScriptSource, getHookScriptSource } from "./single-file-script.js";
 import { CDP } from "simple-cdp";
 
 const LOAD_TIMEOUT_ERROR = "ERR_LOAD_TIMEOUT";
-const INTERACTIVE_TIME_STATE = "InteractiveTime";
+const DEFAULT_NETWORK_STATE = "networkIdle";
 const NETWORK_STATES = ["InteractiveTime", "networkIdle", "networkAlmostIdle", "load", "DOMContentLoaded"];
 const MINIMIZED_WINDOW_STATE = "minimized";
 const SINGLE_FILE_WORLD_NAME = "singlefile";
@@ -37,10 +37,10 @@ const EMPTY_PAGE_URL = "about:blank";
 export { initialize, getPageData, closeBrowser };
 
 async function initialize(options) {
-	if (!options.browserRemoteDebuggingUrl) {
-		await launchBrowser(getBrowserOptions(options));
+	if (options.browserRemoteDebuggingURL) {
+		CDP.remoteDebuggingUrl = options.browserRemoteDebuggingURL;
 	} else {
-		CDP.remoteDebuggingUrl = options.browserRemoteDebuggingUrl;
+		await launchBrowser(getBrowserOptions(options));
 	}
 }
 
@@ -50,7 +50,7 @@ async function getPageData(options) {
 		targetInfo = await CDP.createTarget(EMPTY_PAGE_URL);
 		const { Browser, Security, Page, Emulation, Fetch, Network, Runtime, Debugger } = new CDP(targetInfo);
 		if (options.browserStartMinimized) {
-			const { windowId, bounds } = await Browser.getWindowForTarget({ targetId: targetInfo.targetId });
+			const { windowId, bounds } = await Browser.getWindowForTarget({ targetId: targetInfo.id });
 			if (bounds.windowState !== MINIMIZED_WINDOW_STATE) {
 				await Browser.setWindowBounds({ windowId, bounds: { windowState: MINIMIZED_WINDOW_STATE } });
 			}
@@ -64,7 +64,7 @@ async function getPageData(options) {
 		if (options.browserMobileEmulation) {
 			await Emulation.setDeviceMetricsOverride({ mobile: true });
 		}
-		if (options.httpProxyServer && options.httpProxyUsername) {
+		if (options.httpProxyUsername) {
 			await Fetch.enable({ handleAuthRequests: true });
 			Fetch.addEventListener("authRequired", ({ params }) => Fetch.continueWithAuth({
 				requestId: params.requestId,
@@ -194,12 +194,14 @@ function getTopFrameContextId({ Page, Runtime }, options) {
 							contextId = contextIds[contextIdIndex];
 						}
 						contextIdIndex++;
-					} while (contextId === undefined && contextIdIndex < contextIds.length);
+					} while (contextId === undefined && contextIdIndex < contextIds.length && !navigationSignal.aborted);
 				}
-				if (contextId === undefined) {
-					reject(new Error("Execution context not found"));
-				} else {
-					resolve(contextId);
+				if (!navigationSignal.aborted) {
+					if (contextId === undefined) {
+						reject(new Error("Execution context not found"));
+					} else {
+						resolve(contextId);
+					}
 				}
 			}
 		}
@@ -219,20 +221,20 @@ function getTopFrameContextId({ Page, Runtime }, options) {
 				contextIds.splice(index, 1);
 			}
 		}
+	});
 
-		async function testExecutionContext(contextId) {
-			try {
-				const { result } = await Runtime.evaluate({
-					expression: "singlefile !== undefined",
-					contextId
-				});
-				return result.value === true;
-			} catch (error) {
-				// ignored
-			}
-			return false;
+	async function testExecutionContext(contextId) {
+		try {
+			const { result } = await Runtime.evaluate({
+				expression: "singlefile !== undefined",
+				contextId
+			});
+			return result.value === true;
+		} catch (error) {
+			// ignored
 		}
-	});
+		return false;
+	}
 }
 
 function waitForLoadTimeout(abortSignal, maxDelay) {
@@ -262,11 +264,9 @@ async function waitForPageReady({ Page }, topFrameId, options, navigationSignal)
 
 			function onLifecycleEvent({ params }) {
 				const { name, frameId } = params;
-				if (frameId === topFrameId) {
-					if (name === options.browserWaitUntil || name === INTERACTIVE_TIME_STATE || navigationSignal.aborted) {
-						Page.removeEventListener(LIFE_CYCLE_EVENT, onLifecycleEvent);
-						resolve();
-					}
+				if (frameId === topFrameId && (name === options.browserWaitUntil || name === DEFAULT_NETWORK_STATE || navigationSignal.aborted)) {
+					Page.removeEventListener(LIFE_CYCLE_EVENT, onLifecycleEvent);
+					resolve();
 				}
 			}
 		});

package/lib/deno-polyfill.js

@@ -114,7 +114,6 @@ const DenoAPI = {
 };
 
 const pathAPI = {
-	toFileUrl,
 	dirname
 };
 
@@ -208,15 +207,6 @@ function exit(code) {
 	}
 }
 
-async function toFileUrl(filePath) {
-	if (DENO_RUNTIME_DETECTED) {
-		return path.toFileUrl(filePath);
-	} else {
-		const url = await import(NODE_MODULES["url"]);
-		return new URL(url.pathToFileURL(filePath));
-	}
-}
-
 async function dirname(filePath) {
 	return path.dirname(filePath);
 }

package/lib/version.js

@@ -1 +1 @@
-export const version = "2.0.21";
+export const version = "2.0.23";

package/options.js

@@ -60,7 +60,7 @@ const OPTIONS_INFO = {
 	"browser-cookie": { description: "Ordered list of cookie parameters separated by a comma (name,value,domain,path,expires,httpOnly,secure,sameSite,url)", type: "string[]" },
 	"browser-cookies-file": { description: "Path of the cookies file formatted as a JSON file or a Netscape text file", type: "string" },
 	"browser-ignore-insecure-certs": { description: "Ignore HTTPs errors", type: "boolean" },
-	"browser-remote-debugging-url": { description: "Remote debugging URL", type: "string" },
+	"browser-remote-debugging-URL": { description: "Remote debugging URL", type: "string" },
 	"compress-content": { description: "Compress the output file into a ZIP file", type: "boolean" },
 	"compress-CSS": { description: "Compress CSS stylesheets", type: "boolean" },
 	"compress-HTML": { description: "Compress HTML content", type: "boolean", defaultValue: true },

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "single-file-cli",
-	"version": "2.0.21",
+	"version": "2.0.23",
 	"description": "SingleFile CLI",
 	"author": "Gildas Lormeau",
 	"engines": {
@@ -13,7 +13,7 @@
 		"single-file": "./single-file-node.js"
 	},
 	"dependencies": {
-		"simple-cdp": "^1.8.4",
+		"simple-cdp": "^1.8.5",
 		"ws": "^8.16.0"
 	}
 }

package/single-file-cli-api.js

@@ -79,7 +79,7 @@ const STATE_PROCESSED = "processed";
 const { readTextFile, writeTextFile, writeFile, stdout, mkdir, stat, errors } = Deno;
 let tasks = [], maxParallelWorkers, sessionFilename;
 
-export { DEFAULT_OPTIONS, VALID_URL_TEST, initialize };
+export { initialize };
 
 async function initialize(options) {
 	options = Object.assign({}, DEFAULT_OPTIONS, options);
@@ -203,21 +203,28 @@ function testMaxDepth(task) {
 
 function createTask(url, options, parentTask, rootTask) {
 	url = parentTask ? rewriteURL(url, options.crawlRemoveURLFragment, options.crawlRewriteRules) : url;
-	if (VALID_URL_TEST.test(url)) {
-		const isInnerLink = rootTask && url.startsWith(getHostURL(rootTask.url));
-		const rootBaseURIMatch = rootTask && rootTask.url.match(/(.*?)[^/]*$/);
-		const isChild = isInnerLink && rootBaseURIMatch && rootBaseURIMatch[1] && url.startsWith(rootBaseURIMatch[1]);
-		return {
-			url,
-			isInnerLink,
-			isChild,
-			originalUrl: url,
-			rootBaseURI: rootBaseURIMatch && rootBaseURIMatch[1],
-			depth: parentTask ? parentTask.depth + 1 : 0,
-			externalLinkDepth: isInnerLink ? -1 : parentTask ? parentTask.externalLinkDepth + 1 : -1,
-			options
-		};
+	if (!VALID_URL_TEST.test(url)) {
+		try {
+			url = url.replace(/\\/g, "/");
+			url = url.replace(/#/g, "%23"); 
+			url = new URL(url, import.meta.url).href;
+		} catch (error) {
+			throw new Error("Invalid URL or file path: " + url);
+		}
 	}
+	const isInnerLink = rootTask && url.startsWith(getHostURL(rootTask.url));
+	const rootBaseURIMatch = rootTask && rootTask.url.match(/(.*?)[^/]*$/);
+	const isChild = isInnerLink && rootBaseURIMatch && rootBaseURIMatch[1] && url.startsWith(rootBaseURIMatch[1]);
+	return {
+		url,
+		isInnerLink,
+		isChild,
+		originalUrl: url,
+		rootBaseURI: rootBaseURIMatch && rootBaseURIMatch[1],
+		depth: parentTask ? parentTask.depth + 1 : 0,
+		externalLinkDepth: isInnerLink ? -1 : parentTask ? parentTask.externalLinkDepth + 1 : -1,
+		options
+	};
 }
 
 async function saveTasks() {

package/single-file-launcher.js

@@ -23,21 +23,17 @@
 
 /* global URL */
 
-import { VALID_URL_TEST, initialize } from "./single-file-cli-api.js";
-import { Deno, path } from "./lib/deno-polyfill.js";
+import { initialize } from "./single-file-cli-api.js";
+import { Deno } from "./lib/deno-polyfill.js";
 import options from "./options.js";
 
 const { readTextFile, exit } = Deno;
-const { toFileUrl } = path;
 
 export { run };
 
 async function run() {
 	try {
 		let urls;
-		if (options.url && !VALID_URL_TEST.test(options.url)) {
-			options.url = (await toFileUrl(new URL(options.url, import.meta.url).pathname)).href;
-		}
 		if (options.urlsFile) {
 			urls = (await readTextFile(options.urlsFile)).split("\n");
 		} else {

package/single-file.bat

@@ -1,2 +1,2 @@
 @echo off
-node "%~dp0\single-file" %*
+deno run --allow-read --allow-write --allow-net --allow-env --allow-run "%~dp0\single-file" %*