Re: Testcase attrcreatetextnode2.xml in level1/core does not exercise the test

From the description of the test, the original author were trying to find a
processor that would set an attribute to "Yes" by mistakening expanding an
reference to the ent1 entity on a call to setNodeValue or equivalent.
However, they used "Y%ent1" instead of "Y&ent1;".

The tests were not bad tests, they were just weak tests.  However, after a
test is released, there should be not substantive change to the body of the
test.  In this case, since the existing tests weren't bad, they did not need
to be removed from the suite.  However, to actually test what they had
intended to test required adding new tests.

p.s. I did check the log and it appears that attrcreatetextnode2 was a
variation of attrcreatetextnode that I made to provide coverage for
Attr.setNodeValue about 6 months ago (the original test uses Attr.setValue),
but I did not change the contributor and date from the original.  I've
changed the metadata on the CVS version.  The original bug reports addressed
weaknesses that existed in both attrcreatetextnode and attrcreatetextnode2.

Received on Tuesday, 23 April 2002 11:37:04 UTC