Yeah!!! Managed to get a small test working on Firefox 4 beta 6 using Selenium RC v1.0.1
Here are the steps that I followed.
- Rename the selenium-server.jar to selenium-server.zip (as it is actually a zip file).
- Unpack the selenium-server.xip into a folder named selff4.
- Search inside selff4 folder for all filesÂ named install.rdf
- Open all these install.rdf files,Â five in all, one by one.
- Update the maxVersion line with theÂ 4.0. so it looked like
- Zip up the contents of the folderÂ selff4 back into the selenium-server.zip
- Rename the selenium-server.zip backÂ to selenium-server.jar
I used the command similar to the following to run the test.
java -jar "selenium-server.jar" -htmlSuite "*firefox C:\Program Files (x86)\Mozilla Firefox 4.0 Beta 3\firefox.exe" "http://yourservernameorURLhere/" "C:\fullpath\to\test-Suite.htm" "C:\fullpath\to\logfile.txt"
And it worked!!!
I did seeÂ the following error message in the console, but the test ran just fine.
22:52:57.474 ERROR - Couldn't determine version number
The test suiteÂ contained very limited number of comands, perhaps one of you can give it a more thorough workout. Also notice how the path to the firefox executable can be specified.
Just an tiny observation: My Firefox beta 6 is still in the folder saying beta 3.
I have one question. Is there a way when using -htmlSuite option of RC to get a result even if all tests doesn’t finish. Because now only if all results are done we are getting result file. Otherwise it is empty.
The tests must complete to get the results file. They can pass or fail, but must complete.
So this mean that if RC timeout happens no results can be collected. This problem is that from testing perspective if give 2 hours timeout for 1.5 hours set execution it is bad if the only result is timeout. Doesn’t help much. I would prefer to see results of all finished tests to this particular moment of the timeout.
Also I really want to be able to see runtime results but not from the runner but from the html file. Runner has problem with clicking around already executed tests
Just sharing my opinion
My understanding is that in the current implementation the results are sent from the browser to the Selenium RC server on completion of the tests. If Selenium RC determines that the tests are stuck somewhere (using the timeout), it will kill the browser and hence there is no log. I do agree that there is a need for improvement. I was planning to give that area a quick look sometime soon to see if my understanding is correct and perhaps improve it a bit. In the meanwhile, feel free to create an issue on the official Selenium website.