RE: Testcase attrcreatetextnode2.xml in level1/core does not exer cise the test

[ca] I'm pretty sure that attrcreatetextnode3 does what the author of
attrcreatetextnode was wanting to do but didn't quite accomplish.
 
[mb] I disagree -- it was simply a mistake in invoking the entity.  These
tests were originally authored here at NIST, so I'm privy to what the 
author intended.  This one was written early on and has simply been
translated into java and then DOM TSML with the error.  The intent of 
the test was to set it to "Y&ent1;" and it was simply an error.  


[ca]
Maybe it is the mail lag, but you say that you disagree and then say exactly
what I thought I was saying.   Maybe it is the extra level of entity
processing involved with DOMTSML, for me to get "Y&ent1;" in the code, the
text of the DOMTSML has to read "Y&ent1;".

The even numbered tests (2) and (4) do the same test but using
setNodeValue() instead of setValue() since there was no other test that
checked calling setNodeValue() for an attribute and reusing the existing
test was the most expedient way of covering that function since setNodeValue
and setValue are equivalent.

With the two corrected tests, the value of the two flawed tests is minimal.
They don't particularly hurt, but they don't help either.

I was taking it as a general principle that we don't make substantive
changes to a test after release so that the URL to a released test always
means the same thing.  However, since we haven't really established any
convention for addressing tests, the disruption of making this particular
change in place in inconsequential.

So, I assume that your preference would be to delete attrcreatetextnode3 and
4 and to modify attrcreatetextnode and attrcreatetextnode2 so they do the
same thing that 3 and 4 do now.  If not, what would you suggest?

Received on Wednesday, 24 April 2002 15:35:22 UTC