- From: Lars Marius Garshol <larsga@garshol.priv.no>
- Date: Sat, 20 May 2000 07:33:22 -0400 (EDT)
- To: www-dom@w3.org
The last couple of days I have been putting together a test suite for
javadom.py, which is a simple wrapper interface that allows Python
scripts running in JPython to use Java DOM implementations with a
Python DOM interface.
When writing this test suite I found a number of bugs in the
underlying DOM implementations (and, of course, my own interface),
which I have reported. I also found a couple of weak spots, I believe,
in the specifications, where implementations differ.
1. What is the result of
document.createAttribute("a")._get_nodeValue()?
Is it None or ""? As far as I can see neither of the level 1 and
level 2 specs answer this question. None would be the most
reasonable answer, methinks, but some implementations think
otherwise.
Xerces: ""
Sun: None
Brownell: ""
4DOM: ""
Indelv: None
SXP: None
OpenXML: ""
2. What is the result of calling
element.removeAttribute("non-existent")?
The specs (1 and 2) seem to say that nothing should happen, and that
is my own interpretation.
Xerces: nothing
Sun: raise DOMException
Brownell: raise DOMException
4DOM: nothing
Indelv: nothing
SXP: nothing
OpenXML: nothing
--Lars M.
Received on Sunday, 21 May 2000 11:02:11 UTC