fix: upgrade puppeteer to v21 now using "Chrome for Testing"

BREAKING CHANGE:

Puppeteer now uses "Chrome for Testing" instead of Chromium. This should make the installation of puppeteer's browser more stable especially on Mac computers with Apple silicon (M1/M2/M3).

We now use the new headless mode of Chrome.

A new README section related to Puppeteer is now available: https://github.com/amtrack/sfdx-browserforce-plugin/blob/master/README.md#puppeteer

README.md

@@ -11,7 +11,7 @@ Further benefits:
 
 - comfortable configuration using JSON Schema (similar to the [Scratch Org Definition Configuration](https://developer.salesforce.com/docs/atlas.en-us.sfdx_dev.meta/sfdx_dev/sfdx_dev_scratch_orgs_def_file.htm))
 - idempotency of the `apply` command only applies what's necessary and allows re-execution (concept similar to [terraform](https://www.terraform.io/docs/commands/apply.html))
-- browser automation powered by [Puppeteer](https://github.com/GoogleChrome/puppeteer) (Chromium)
+- browser automation powered by Puppeteer and "Chrome for Testing", [learn more about Puppeteer and Browserforce](#puppeteer)
 - implement your own custom preferences (a.k.a. plugins; to be improved)
 
 # Installation
@@ -162,6 +162,31 @@ Here is a full blown example showing most of the supported settings in action:
 - `BROWSERFORCE_RETRY_MAX_RETRIES`: number of retries on failures opening a page (default: `4`)
 - `BROWSERFORCE_RETRY_TIMEOUT_MS`: initial time between retries in exponential mode (default: `4000`)
 
+# Puppeteer
+
+We use [Puppeteer](https://github.com/puppeteer/puppeteer) for browser automation which comes with its own "Chrome for Testing" browser.
+
+The puppeteer [installation doc](https://github.com/puppeteer/puppeteer#installation) describes how this works:
+
+> When you install Puppeteer, it automatically downloads a recent version of
+> [Chrome for Testing](https://goo.gle/chrome-for-testing) (~170MB macOS, ~282MB Linux, ~280MB Windows) that is [guaranteed to
+> work](https://pptr.dev/faq#q-why-doesnt-puppeteer-vxxx-work-with-chromium-vyyy)
+> with Puppeteer. The browser is downloaded to the `$HOME/.cache/puppeteer` folder
+> by default (starting with Puppeteer v19.0.0).
+
+In most of the cases this just works! If you still want to skip the download and use another browser installation, you can do this as follows:
+
+```console
+export PUPPETEER_SKIP_CHROMIUM_DOWNLOAD=true
+sf plugins install sfdx-browserforce-plugin
+export PUPPETEER_EXECUTABLE_PATH=/usr/bin/chromium-browser
+sf browserforce:apply ...
+```
+
+Troubleshooting:
+
+- The installation is triggered via the `postinstall` hook of npm/yarn. If you've disabled running scripts with npm (`--ignore-scripts` or via config file), it will not download the browser.
+
 # Contributing
 
 Please see [CONTRIBUTING.md](CONTRIBUTING.md) for getting started.

package.json

@@ -10,7 +10,7 @@
     "@mdapi-issues/listmetadata-standardvalueset": "2.0.3",
     "@salesforce/sf-plugins-core": "4.1.1",
     "p-retry": "4.6.2",
-    "puppeteer": "19.2.0"
+    "puppeteer": "21.5.1"
   },
   "devDependencies": {
     "@salesforce/dev-config": "4.1.0",

src/browserforce.ts

@@ -29,7 +29,7 @@ export class Browserforce {
         // workaround for navigating frames https://github.com/puppeteer/puppeteer/issues/5123
         '--disable-features=site-per-process'
       ],
-      headless: !(process.env.BROWSER_DEBUG === 'true')
+      headless: process.env.BROWSER_DEBUG === 'true' ? false : 'new'
     });
     await this.openPage(
       `secur/frontdoor.jsp?sid=${this.org.getConnection().accessToken}&retURL=${encodeURIComponent(POST_LOGIN_PATH)}`