Child pages
  • ERSelenium Framework

Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.
Comment: StandaloneRunner information added

...

  • Play around with test cases from ERSeleniumExample/Resources/Selenium/ using Selenium IDE.

...

  • Selenium IDE Firefox plugin and XPath Checker can be very handy for creating and editing Selenium tests. Selenium IDE Selenese source plugin with proper comments support resides in ERSelenium/Resources/selenese-ide-plugin.js.

Standalone runner

ERSelenium offers tests' developers several nice features - like automatic test suite generation, metacommands and URL independence. Unfortunately this leads to some troubles when trying to execute your tests with Selenium-RC. This is where StandaloneRunner can be very helpful.

Selenium-RC is essentially a selenium-server and a set of client libraries that you can use. You can write any kind of client application that sends particular commands to the server and receives back status codes. Among these commands are typical selenium commands ("click", "type" and so on) and several specific ("open specified browser"). Selenium-RC does a great job of preparing the browser profile (it turns off confirmation dialog boxes, clears cookies and so on) and gives a lot of other nice features. One of the most useful is the ability to execute tests from the command line. It's a typical task that is usually executed on build server on regular or per-commit basis.

The problem is that if you have your tests written with ERSelenium flavor, you won't be able to run them directly with Selenium-RC - as you won't even have a test suite file. But you can do this with StandaloneRunner. To execute the tests you must have Selenium-RC server running in the background. Note that it should be started in windowed mode (not via SSH) - i.e. not in the headless mode. Assuming that you have your application built, you should use the following commands:

Code Block

YourApp.woa/YourApp <tests root folder> <application root url> <selenium-rc server host> <selenium-rc server port> <browser type> \
     -DWOApplicationClass=er.selenium.rc.StandaloneRunner -Dfile.encoding=utf-8

This will execute the tests using the Selenium-RC server on the specified host and port with the specified browser. The <application root url> will be used as a root url for all urls that are used in tests. Tests from <tests root folder> will be executed (they will be searched for recursively).

Note, that two last arguments are essential: -DWOApplicationClass substitutes your application's application class with er.selenium.rc.StandaloneRunner, which will do all the testing and then exit without entering WOApplication's requests handling loop. The last argument ensures proper encoding.

Here is the real-world example commands:

Code Block

cd /Library/WebObjects/Applications
./YourApp.woa/YourApp ./YourApp.woa/Contents/Resources/Selenium http://localhost/cgi-bin/WebObjects/YourApp.woa localhost 4444 '*firefox' \
     -DWOApplicationClass=er.selenium.rc.StandaloneRunner -Dfile.encoding=utf-8

This will execute tests from /Library/WebObjects/Applications/YourApp.woa/Contents/Resources/Selenium using the http://localhost/cgi-bin/WebObjects/YourApp.woa as the root url in firefox browser using Selenium-RC server on localhost on its default port 4444.

The sample of successful output is:

Code Block

- test '/usr/local/wondercap2/dep/Songsterr/14662/dep/dist/Songsterr.woa/./Contents/Resources/Selenium/Commenting/LoginViaCommenting.sel' PASSED
- test '/usr/local/wondercap2/dep/Songsterr/14662/dep/dist/Songsterr.woa/./Contents/Resources/Selenium/Commenting/NotifyAboutReplies.sel' PASSED
- test '/usr/local/wondercap2/dep/Songsterr/14662/dep/dist/Songsterr.woa/./Contents/Resources/Selenium/Commenting/SignUpViaCommenting.sel' PASSED
- test '/usr/local/wondercap2/dep/Songsterr/14662/dep/dist/Songsterr.woa/./Contents/Resources/Selenium/Commenting/SuccessfulComment.sel' PASSED

All tests in tests root folder are always executed (even if there are failures). The executed command will fail if one of the tests fails.