Best IT Recipes

Natural ghost growth

I recently switched from FirefoxDriver to GhostDriver.

My test suite (approximately 150 to the end of the tests) used to run 25 minutes. Now with Ghostdriver it takes 23 minutes. So in total I got 2 minutes of runtime.

This is an improvement, but not as huge as I expected. Is such a low gain normal? Or should I expect a much higher downtime from switching to a headless test?

I am using the .NET version of webdriver / ghostdriver.

I would be very interested to compare this "test" with someone who recently switched to GhostDriver

+2
selenium webdriver functional-testing headless
Jan 07 '13 at 10:50
source share
2 answers

GhostDriver will not offer any performance benefits on ChromeDriver, since the only thing it does not do is render graphics on the screen, which Chrome uses the graphics driver, not the processor.

The only thing I see is not the appearance of the browser window and, possibly, when working on the CI server.

Some statistics of my test run are on my blog: http://watirmelon.com/2013/02/05/watir-webdriver-with-ghostdriver-on-osx-headless-browser-testing/

+2
Feb 05 '13 at 12:23
source share

I feel this is a slightly vague question. For example, if your test contains many calls to Thread.sleep() , then the tests will not be improved.

My tests use ChromeDriver, and I did not think about switching. But I feel that this simple switch will not bring you much, perhaps also consider refactoring the code

+2
Jan 07 '13 at 11:00
source share


More articles:


All Articles