[wtp-dev] Status of RC3: explaination of JUnit result changes

[David M Williams <david_williams@xxxxxxxxxx>]

Just so everyone knows the same thing,


if any of you watch wtp-releng closely,
you would have noticed that the 
Sun, 11 Dec 2005 -- 05:22 (UTC) 
version of RC3 had some JUnit failures,
twice, and then did not. 

Here's the explaination. When I first
noticed the JUnit failures, I immediately re-ran the 
tests again, and they failed again.
Then, the next morning, I re-ran the tests again, and they 
succeeded. 

I think what's happening here, especially
given some recent posts on newsgroup, is that sometimes
a server from which we try to get DTDs/Schemas
 can sometimes be having "special" problems 
which we do not handle well. I have
opened following two bugs to improve our testing and handling 
of these cases in future .. but, I don't
think there's a fundamental problem or regression in our code. 

120275
maj wst.xml We should never ever catch Exception and do nothing with it

120277
nor wst.inte We should test the many cases of server responses when getting
remote resources 

Just FYI, after all the above "re-tests", I
re-ran complete build "from scratch", just to make sure there

was not a fluke in our "re-test" procedure.

This was the 
Sun, 11 Dec 2005 -- 17:34 (UTC) 
version of RC3. And it succeeded. Its actual code content
should be identical with the 05:22 version, 
but if/when I declare an RC3 today, it will be with the
17:34 version, since it was done "from scratch".