This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
L3 Core test domconfigurationcansetparameter06 tests the functionality of isElementContentWhitespace, however "if there is no declaration for the containing element, or there are multiple declarations, this property has no value for white space characters" [1] In hc_staff.xml, there is no declaration for the element strong. Suggest modifying the test so that barfoo.xml is loaded instead. Similarly, for LS test elementcontentwhitespace01, whitespace cannot be expected to be eliminated if there is no declaration for the containing element p. Suggest modifying the test so that test3.xml is used instead. [1]http://www.w3.org/TR/2004/REC-xml-infoset-20040204/#infoitem.character
Created attachment 354 [details] Proposed patch with reworked element content whitespace tests
The domconfigurationcansetparameter* tests seem seriously confused. From the title, you'd assume that they'd basically check the behavior of the canSetParameter. However, almost all of the test is checking that asking if you can set a value to a specific value effectively changes the value by looking at the changes after normalizing a document. If there was such a flaw, there is a very high probability that it would actually change the property and you won't need to try to detect it indirectly. So you could do: <canSetProperty var="canSet" obj="domConfig" name='"element-content-whitespace"' value="false"/> <getProperty var="paramVal" obj="domConfig" name='"element-content-whitespace"'/> <assertTrue actual="paramVal" id="shouldNotHaveActuallyChangedValue"/> The proposed domconfigurationcansetparameter06 also assumes that whitespace exists in the document, but it may have been eliminated on load. In addition, several of the tests specify willBeModified="false", but then call normalizeDocument. I've walked through the tests and tried to correct the problems I saw. elementcontentwhitespace looked unnecessarily complex, but valid to me. I've stripped it down, but should be equivalent to the old test.
Created attachment 361 [details] Alternative fix for raised issues