On 28 Apr 2011, at 12:19, Rich Tibbett wrote: > Scott Wilson wrote: >> Right, I've done some basic work on the tests in Apache Wookie - most of >> them seem to work OK so far; I need to do interactive testing next - I've >> tested processing the update element in config.xml and acquiring and >> validating the UDD, but not actually processing the updates. >> >> However there are some issues I've found so far: >> >> Test ta-preparation1-8 >> This fails to parse as a valid widget as the XML isn't well-formed >> >> Test ta-processing2-3 >> I get a content type of text/html when getting the UDD >> >> Test ta-processing2-7 >> I get an error of the UDD XML not being well-formed >> >> Test ta-processing2-9 >> I get a content type of text/html when getting the UDD > > All the above tests have been fixed.
Great, thanks! > >> >> Test ta-processing2-16 >> We throw a UDD parse exception when a UDD has a relative URL for the src >> attribute, but the test seems to assume you just silently fail to update the >> widget? > > The processing in this case was underspecified in the spec. I've updated the > specification to allow for relative URLs to be included in the 'src' > attribute. The resolution system for relative URL src attributes to absolute > valid IRIs has also been included in the spec (i.e. resolve relative URLs > relative to the UDDs IRI). > > Relative URLs are not allowed in the update-description element of a widget's > config.xml. > > We've therefore replaced ta-processing2-16 test with a test to check that the > resolution of src attribute relative URLs is working correctly. Makes sense. > >> >> Test ta-processing2-19 >> I get a content type of text/html when getting the UDD >> >> Test ta-processing2-20 >> I get a content type of text/html when getting the UDD >> > > These tests have been fixed. > > HTH, Rich Its a big help, thanks. One more thing, the widget at: http://people.opera.com/harig/wupdres/resources/pass.wgt ... isn't valid as its not well formed - there should be ?> at end of line 1. S
