Fix error handling in NPM/Puppeteer

src/drivers/npm/browsers/puppeteer.js

@@ -68,100 +68,100 @@ class PuppeteerBrowser extends Browser {
   }
 
   async visit(url) {
-    let done = false;
-    let browser;
+    return new Promise(async (resolve, reject) => {
+      let done = false;
+      let browser;
+
+      try {
+        browser = await this.browser();
+      } catch (error) {
+        reject(new Error(error.message || error.toString()));
+      }
 
-    try {
-      browser = await this.browser();
-    } catch (error) {
-      throw new Error(error.message || error.toString());
-    }
+      browser.on('disconnected', () => {
+        if (!done) {
+          reject(new Error('Disconnected'));
+        }
+      });
 
-    browser.on('disconnected', () => {
-      if (!done) {
-        throw new Error('Disconnected');
-      }
-    });
+      try {
+        const page = await browser.newPage();
 
-    try {
-      const page = await browser.newPage();
+        page.setDefaultTimeout(this.options.maxWait);
 
-      page.setDefaultTimeout(this.options.maxWait);
+        await page.setRequestInterception(true);
 
-      await page.setRequestInterception(true);
+        page.on('error', reject);
 
-      page.on('error', (error) => {
-        throw new Error(error.message || error.toString());
-      });
+        page.on('request', request => request.continue());
 
-      page.on('request', request => request.continue());
+        page.on('response', (response) => {
+          if (response.status() === 301 || response.status() === 302) {
+            return;
+          }
 
-      page.on('response', (response) => {
-        if (response.status() === 301 || response.status() === 302) {
-          return;
-        }
+          if (!this.statusCode) {
+            this.statusCode = response.status();
 
-        if (!this.statusCode) {
-          this.statusCode = response.status();
+            this.headers = {};
 
-          this.headers = {};
+            const headers = response.headers();
 
-          const headers = response.headers();
+            Object.keys(headers).forEach((key) => {
+              this.headers[key] = Array.isArray(headers[key]) ? headers[key] : [headers[key]];
+            });
 
-          Object.keys(headers).forEach((key) => {
-            this.headers[key] = Array.isArray(headers[key]) ? headers[key] : [headers[key]];
-          });
+            this.contentType = headers['content-type'] || null;
+          }
+        });
 
-          this.contentType = headers['content-type'] || null;
-        }
-      });
+        page.on('console', ({ _type, _text, _location }) => this.log(`${_text} (${_location.url}: ${_location.lineNumber})`, _type));
 
-      page.on('console', ({ _type, _text, _location }) => this.log(`${_text} (${_location.url}: ${_location.lineNumber})`, _type));
+        await page.setUserAgent(this.options.userAgent);
 
-      await page.setUserAgent(this.options.userAgent);
+        await Promise.race([
+          page.goto(url, { waitUntil: 'networkidle2' }),
+          new Promise(_resolve => setTimeout(_resolve, this.options.maxWait)),
+        ]);
 
-      await Promise.race([
-        page.goto(url, { waitUntil: 'networkidle2' }),
-        new Promise(resolve => setTimeout(resolve, this.options.maxWait)),
-      ]);
+        // eslint-disable-next-line no-undef
+        const links = await page.evaluateHandle(() => Array.from(document.getElementsByTagName('a')).map(({
+          hash, hostname, href, pathname, protocol, rel,
+        }) => ({
+          hash,
+          hostname,
+          href,
+          pathname,
+          protocol,
+          rel,
+        })));
 
-      // eslint-disable-next-line no-undef
-      const links = await page.evaluateHandle(() => Array.from(document.getElementsByTagName('a')).map(({
-        hash, hostname, href, pathname, protocol, rel,
-      }) => ({
-        hash,
-        hostname,
-        href,
-        pathname,
-        protocol,
-        rel,
-      })));
-
-      this.links = await links.jsonValue();
+        this.links = await links.jsonValue();
 
-      // eslint-disable-next-line no-undef
-      const scripts = await page.evaluateHandle(() => Array.from(document.getElementsByTagName('script')).map(({
-        src,
-      }) => src));
+        // eslint-disable-next-line no-undef
+        const scripts = await page.evaluateHandle(() => Array.from(document.getElementsByTagName('script')).map(({
+          src,
+        }) => src));
 
-      this.scripts = (await scripts.jsonValue()).filter(script => script);
+        this.scripts = (await scripts.jsonValue()).filter(script => script);
 
-      this.js = await page.evaluate(getJs);
+        this.js = await page.evaluate(getJs);
 
-      this.cookies = (await page.cookies()).map(({
-        name, value, domain, path,
-      }) => ({
-        name, value, domain, path,
-      }));
+        this.cookies = (await page.cookies()).map(({
+          name, value, domain, path,
+        }) => ({
+          name, value, domain, path,
+        }));
 
-      this.html = await page.content();
-    } catch (error) {
-      throw new Error(error.message || error.toString());
-    } finally {
-      done = true;
+        this.html = await page.content();
+      } catch (error) {
+        reject(error.message || error.toString());
+      } finally {
+        done = true;
 
-      await browser.close();
-    }
+        await browser.close();
+      }
+    });
   }
 }
 

src/drivers/npm/package.json

@@ -2,7 +2,7 @@
   "name": "wappalyzer",
   "description": "Uncovers the technologies used on websites",
   "homepage": "https://github.com/AliasIO/Wappalyzer",
-  "version": "5.9.12",
+  "version": "5.9.13",
   "author": "Elbert Alias",
   "license": "GPL-3.0",
   "repository": {

src/wappalyzer.js

@@ -322,7 +322,7 @@ class Wappalyzer {
    *
    */
   ping() {
-    if (Object.keys(this.hostnameCache).length > 100) {
+    if (Object.keys(this.hostnameCache).length > 25) {
       this.driver.ping(this.hostnameCache);
 
       this.hostnameCache = {};