SeleniumLibrary is a web testing library for Robot Framework that utilizes the Selenium tool internally. SeleniumLibrary 3.0.0b3 is a new release with Python 3 support and rewritten architecture. The new architecture should be mostly invisible from the Robot Framework test date point of view. But there are many changes in internal methods and classes, which may affect users who have written libraries extending SeleniumLibrary private methods and classes. There are also many other smaller bigger enhancements.

All issues targeted for SeleniumLibrary v3.0.0 can be found from the issue tracker.

If you have pip installed, just run

pip install --pre --upgrade robotframework-seleniumlibrary

to install the latest available release or use

pip install robotframework-seleniumlibrary==3.0.0b3

to install exactly this version. Alternatively you can download the source distribution from PyPI and install it manually.

SeleniumLibrary 3.0.0b3 was released on Thursday September 28, 2017.

The SeleniunLibrary was was enhanced to support Python 3.3 or newer. More precisely, the currently supported Python versions are 2.7, 3.3 and newer. Installation on Python 3 works exactly as it does for Python 2, and the recommended installation method is with pip.

To ease support for Python 3.3, the support for Python 2.6 was dropped.

Because Selenium moved from version 2 to version 3, we felt that the old name, Selenium2Library, did not anymore describe what the library supports and helps user to achieve. Therefore we renamed the original SeleniumLibrary to OldSeleniumLibrary. Then the Selenium2Libary was renamed to SeleniumLibrary and new Selenium2Library was created. The new Selenium2Library is just a wrapper for the SeleniumLibrary and is meant only to ease migration from Selenium2Library to SeleniumLibrary.

The old architecture did contained many practices which are against good coding practices and made the development of the library slower or more difficult. The old architecture made also hard to define clear API interface for the user who are building their own libraries on top of the SeleniumLibrary.

The new architecture relies on the Robot Framework Dynamic library API, PythonLibCore and using context object to share common methods between different classes and methods. Because of the new architecture, many

private methods or classes have been changed or removed totally.

The project documentation, expect keyword documentation, has been rewritten to better serve user of the SeleniumLibrary.

The current release was automatically tested with Python 2.7, Python 3.3 and Python 3.6. The Jython, PyPy and IronPython compatibility was tested manually. We found that SeleniumLibrary supports Jython and PyPy, but IronPython is not supported.

Next steps are add the Jython and PyPy testing in CI.

New strategy:value syntax to specify locator strategy in addition to current strategy=value. The current locator strategy causes problems with Robot Framework keyword argument syntax, because then the equal sing must be escaped, example xpath\=//div | &{kw_args} |. The locator syntax is aimed to ease the transition when the keyword arguments are in future taken in use.

The previous releases, the table keywords locator strategy did differ greatly from the rest of keywords which did interact with elements in the browsers. This is now changed and tables can be located with same locator strategies which can be used for the rest of the library.

Library documentation, excluding individual keyword documentation, has been rewritten.

The new architecture should not change how the keywords are used in Robot Framework test data. But causes changes how the SeleniumLibrary can be used to build new libraries. We have deprecated many private methods and have created many new public methods or attributes which should make the extending more easier in the future.

In this release the keyword returned by the Register keyword to run on failure keyword can be always used to restore the original state. User do not anymore need to have special logic in Robot Framework test data to restore the Run On Failure keyword.

The Capture Page Screenshot keyword now verifies from the file system that screenshot file does not exist in the file system. If the file exist, it will create new index in the file name until it find a file name which does not exist.

If the filename does not contain index, the filename is always overwritten.

The minimum requires Robot Framework version is now updated to 2.8.7

The minimum required Selenium version is now updated to 2.53.6, which is the latest Selenium 2 release.

The boolean arguments are handled in similar way as in Robot Framework. More details in library Boolean arguments documentation.

The Firefox profile was removed from the SeleniumLibrary and Selenium default profile is used instead. Although this should not cause any problems, there is low risk for some backwards incompatible change.

• New strategy:value syntax to specify locator strategy in addition to current strategy=value (#908, beta 3)
• Drop Python 2.6 support (#620, alpha 1)
• Remove unused expiry argument from Add Cookie keyword (#847, alpha 1)
• Modify Get Webelements not to raise exception when no elements are found (#805, beta 3)
• Drop Python 2.6 support (#620, alpha 1)

• Change library name from Selenium2Library to SeleniumLibrary (#777, alpha 1)
• Externally useful attributes should be declared public and not start with an underscore (#882, beta 3)

Many thanks to "thaffenden" to add Get Locations keyword. "davidshepherd7" to fixing issue when getting window information and switching windows on browsers that do not support javascript. Many thanks to many other contributors who have helped to make this release.

Altogether 33 issues. View on the issue tracker.