cvs: pear /XML_XPath TODO

dallen Thu Jun 6 06:54:48 2002 EDT Modified files: /pear/XML_XPath TODO Log: updates Index: pear/XML_XPath/TODO diff -u pear/XML_XPath/TODO:1.3 pear/XML_XPath/TODO:1.4 --- pear/XML_XPath/TODO:1.3 Wed Jun 5 03:54:06 2002 +++ pear/XML_XPath/TODO Thu Jun 6 06:54:48 2002 @@ -27,13 +27,28 @@ usable in the class [*] implement getElementsByTagName and getElementById since there are domxml functions - now + now...return as getElementsByTagName as XML_XPath_result objects..getElementById + will be a move function [*] considering doing substringData for attribute nodes...just seems to make sense [*] in _quick_evaluate_init, make sure the pointer is not null, could happen coming off of a results object - [*] it would be better if getChildNodeByName() and getSiblingNodeByName() (not called that - now, but should be called that) are result object generation functions...this way - we can seperate the cache from getting mixed up in bad instances...word... + [@] functions for result object are now + next() + rewind() + end() + nextByNodeName() + nextByNodeType() + + [@] added getNodePath() to get the xpath location of the current node + + [@] when using next() on XML_XPath_result, it counts the first node the first time, so + it moves from off the stage to on the stage the first time, then it moves to the next + node after that + + [@] relative xpath queries + + [*] fix this stdClass hack we have to do to get the result object to work with non xpath + queries (in cases of using dom queries)