I agree with you on both counts. The second check for node type seem
superfluous, and hasNSDecls is never reset, so we'll always skip the
addNSDeclsToAttrs call. Do you think that may be the cause of John's
problems?
Myriam
Gary L Peskin
<garyp@firstec To: Xalan Development
h.com> <[EMAIL PROTECTED]>
cc: (bcc: Myriam Midy/CAM/Lotus)
07/19/01 04:28 Subject: TreeWalker2Result.startNode()
PM questions
Please respond
to xalan-dev
Joe or Myriam or Scott upon return --
Could you please look at
org.apache.xalan.transformer.TreeWalker2Result.startNode()? I am either
going sane (an unlikely possibility) or there are some serious problems
in startNode().
At line 150, we have
if (DTM.ELEMENT_NODE == m_dtm.getNodeType(node))
this begins an if block. Inside there at line 167, we have
if (DTM.ELEMENT_NODE == m_dtm.getNodeType(node))
I don't see how the type of node could have changed and node itself
didn't change so shouldn't this inner if test just be eliminated and the
code moved to the block beginning at line 150 (actually the "else"
clause of that).
Second, on line 169 we declare
boolean hasNSDecls = false;
This is later tested but never changed. Shouldn't the body of the loop
at line 174 maybe set this to true?
Please advise.
Gary
