History.txt

# Version 0.3.5

Release date: 2010-02-26

This is a mostly backwards compatible release, it does break
the API in some minor places, which should hopefully not affect
too many users, please read the release notes carefully!

### Breaking

* Relative searching in a node (e.g. find('//p').all('//a')) will now follow XPath standard
  this means that if you want to find descendant nodes only, you'll need to prefix a dot!
* `visit` now accepts fully qualified URLs for drivers that support it.
* Capybara will always try to run a rack server, unless you set Capybara.run_sever = false

### Changed

* thin is preferred over mongrel and webrick, since it is Ruby 1.9 compatible
* click_button and click will find <input type="button">, clicking them does nothing in RackTest

### Added

* Much improved error messages in a multitude of places
* More semantic page querying with has_link?, has_button?, etc...
* Option to ignore hidden elements when querying and interacting with the page
* Support for multiple selects

### Fixed

* find_by_id is no longer broken
* clicking links where the image's alt attribute contains the text is now possible
* within_fieldset and within_table work when the default selector is CSS
* boolean attributes work the same across drivers (return true/false)