Add support for executeScript/evaluateScript args

[mvorisek]

The args support is necessary as webdriver objects must be passed natively to the underlaying drivers & cannot be serialized directly into the 1st script argument.

Example executeScript method prototype of one of the most popular driver - https://github.com/php-webdriver/php-webdriver/blob/8ffa927b270e932449e8015abf4d38bb0eff24b7/lib/Remote/RemoteWebDriver.php#L324

[codecov]

Codecov Report

❌ Patch coverage is 75.00000% with 3 lines in your changes missing coverage. Please review.
✅ Project coverage is 98.17%. Comparing base (19e5890) to head (b37c535).
⚠️ Report is 60 commits behind head on master.

Files with missing lines	Patch %	Lines
src/Driver/DriverInterface.php	0.00%	3 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff              @@
##             master     #826      +/-   ##
============================================
- Coverage     98.47%   98.17%   -0.30%     
  Complexity      345      345              
============================================
  Files            23       24       +1     
  Lines           983      986       +3     
============================================
  Hits            968      968              
- Misses           15       18       +3     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[aik099]

Adding more parameters to the DriverInterface interface might be a BC break (not sure if an optional argument causes BC breaks). I guess you'll need to create PR for every driver to pass through these arguments. I guess the Selenium2 is the only one capable of it.

[mvorisek]

Yes, it is fatal error BC break - https://3v4l.org/kO2hW.

I will create PRs to these drivers - https://github.com/minkphp/MinkSelenium2Driver and https://github.com/silverstripe/MinkFacebookWebDriver shortly. What are the other popular drivers?

[stof]

We simply cannot do such BC break in a minor version. That would mean we stop applying semver.

[mvorisek]

Are there any objections to release a v2?

[stof]

even for doing a v2, we would want to provide a migration path