Custom capabilities are not seen in browser.capabilities within beforeScenario hook (Chrome sessions ONLY)

[VladSelikh]

I have a problem with capabilities for Chrome tests local run. Here is my config file:

export const config: WebdriverIO.Config = {
 runner: 'local',
  autoCompileOpts: {
    autoCompile: true,
    tsNodeOpts: {
      project: './tsconfig.json',
      transpileOnly: true
    }
  },
  specs: [
    `../features/tests/${brand}/historyTests.feature`,
    // `../features/tests/Mein Konto/**/*.feature`,
    // '../features/tests/*.feature'
  ],
  exclude: [
    // 'path/to/excluded/files'
  ],
  maxInstances: 50,
  capabilities: [{
       browserName: 'chrome',
       'goog:chromeOptions': {
       binary: 'C:\\Program Files\\Google\\Chrome\\Application\\chrome.exe',
       excludeSwitches: [ 'enable-logging' ],
       args: [
           'no-sandbox',
           'disable-gpu',
           'dns-prefetch-disable',
           'disable-extensions',
           'allow-insecure-localhost',
           'disable-notifications',
           'remote-allow-origins=*',
           'disable-features=WebShare',
           'window-size=1920,1080',
           'headless=new'
           ],
       prefs: { 'intl.accept_languages': 'de,DE' }
   },
   acceptInsecureCerts: true,
     'custom:platform': 'CHROME',
     'custom:resolution': '1920x1080',
     'custom:language': 'DE'
   }],
  logLevel: 'silent',
  bail: 0,
  baseUrl: getSpecificUrl(environment.toUpperCase() as types.Environment, brand.toUpperCase() as types.Brand, language.toUpperCase() as types.Language),
  waitforTimeout: 50000,
  connectionRetryTimeout: 120000,
  connectionRetryCount: 3,
  services: [],
  maxInstancesPerCapability: 1,
  framework: 'cucumber',
  reporters: [
    ['spec', { showPreface: false }],
    ['allure', {
      outputDir: './allure-results',
      disableWebdriverStepsReporting: true,
      disableWebdriverScreenshotsReporting: false,
      useCucumberStepReporter: true,
      addConsoleLogs: true
    }]
  ],
  cucumberOpts: {
    require: ['./features/step-definitions/**/*.ts'],
    backtrace: false,
    requireModule: [],
    dryRun: false,
    failFast: false,
    snippets: true,
    source: true,
    strict: false,
    tags: filter + 'and (not @need-to-be-refactored and not @need-to-be-improved)',
    timeout: 60000,
    ignoreUndefinedDefinitions: false
  },
  before: function () {
    const commandNames = Object.keys(commands) as Array<keyof typeof commands>;
    commandNames.forEach((key) => {
      browser.addCommand(key, commands[key]);
    })
  },
  beforeScenario: function() {
    console.log(browser.requestedCapabilities);
    console.log(browser.capabilities)
  },
  afterStep: async function (step, scenario, { error, duration, passed }, context) {
    await browser.takeScreenshot();
  },
}

For the fist console.log() the output is:

{
[0-0]   browserName: 'chrome',
[0-0]   'goog:chromeOptions': {
[0-0]     binary: 'C:\\Program Files\\Google\\Chrome\\Application\\chrome.exe',
[0-0]     excludeSwitches: [ 'enable-logging' ],
[0-0]     args: [
[0-0]       'no-sandbox',
[0-0]       'disable-gpu',
[0-0]       'dns-prefetch-disable',
[0-0]       'disable-extensions',
[0-0]       'allow-insecure-localhost',
[0-0]       'disable-notifications',
[0-0]       'remote-allow-origins=*',
[0-0]       'disable-features=WebShare',
[0-0]       'window-size=1920,1080',
[0-0]       'headless=new'
[0-0]     ],
[0-0]     prefs: { 'intl.accept_languages': 'de,DE' }
[0-0]   },
[0-0]   acceptInsecureCerts: true,
[0-0]   'custom:platform': 'CHROME',
[0-0]   'custom:resolution': '1920x1080',
[0-0]   'custom:language': 'DE'
[0-0] }

For the second one is:

{
[0-0]   acceptInsecureCerts: true,
[0-0]   browserName: 'chrome',
[0-0]   browserVersion: '123.0.6312.124',
[0-0]   chrome: {
[0-0]     chromedriverVersion: '123.0.6312.122 (31f8248cdd90acbac59f700b603fed0b5967ca50-refs/branch-heads/6312@{#824})',
[0-0]     userDataDir: 'C:\\Users\\vladi\\AppData\\Local\\Temp\\scoped_dir35480_1760099228'
[0-0]   },
[0-0]   'fedcm:accounts': true,
[0-0]   'goog:chromeOptions': { debuggerAddress: 'localhost:60326' },
[0-0]   networkConnectionEnabled: false,
[0-0]   pageLoadStrategy: 'normal',
[0-0]   platformName: 'windows',
[0-0]   proxy: {},
[0-0]   setWindowRect: true,
[0-0]   timeouts: { implicit: 0, pageLoad: 300000, script: 30000 },
[0-0]   unhandledPromptBehavior: 'dismiss and notify',
[0-0]   'webauthn:extension:credBlob': true,
[0-0]   'webauthn:extension:largeBlob': true,
[0-0]   'webauthn:extension:minPinLength': true,
[0-0]   'webauthn:extension:prf': true,
[0-0]   'webauthn:virtualAuthenticators': true,
[0-0]   'wdio:driverPID': 35480
[0-0] }

I can not understand why desired capabilities are not the same as session capabilities, AND ONLY for Chrome, AND ONLY for local run. Does anyone have a guess?

[christian-bromann]

I can not understand why desired capabilities are not the same as session capabilities

That is expected behavior, browser.capabilities returns the capabilities that were accepted by the driver initiating the session. Chromedriver does not know about your custom capabilities, hence will ignore it. In order to access the requested capabilities, use browser.requestedCapabilities, see more in https://webdriver.io/docs/api/browser/#properties

[VladSelikh]

But why then does Geckodriver know about them? This situation is only true for tests run on Chrome.

[christian-bromann]

@VladSelikh this is not defined in the spec, browser may behave differently here.