Tim Visher wrote:
Wow... let me feel sheepish for a bit as I blithely mentioned the
acquisition of Ant in Action to its author (sometimes, reading
signatures is helpful)...
I took it at a complement. If you'd said it hadn't helped, then I'd be
worried
I suppose that means that help from that
book may be a little bit of a long shot, or was the junit task covered
somewhat in depth in your book, Steve?
oh yes, there is lots of coverage of it, and the stuff on top, such as
Cactus, even AntUnit
but
- nothing on Junit4. That was not just because Ant1.7 can't handle it,
but because cactus didnt (does it now?) like it, and I think cactus is
wonderful.
-nothing on TestNG. They have their own task and when I do sit down to
redo the report format, I want to do something they work with too.
-The HtmlUnit people have been sending me emails telling me off for
talking about HttpUnit instead :)
Ant1.7 does work with Junit4, but it doesnt report everything as well as
it should. There are some sublties. Similarly, TestNG has a good notion
of 'skipped' tests; these are tests that you decide you dont want to run
(or they decide themselves to be skipped). this is incredibly useful
when testing stuff that is machine/network specific, but you do want to
keep track of how many tests are being skipped and what they are.
I'm trying to test complex systems (imagine: a hadoop cluster spread
over a set of machines, with the logs from all the boxes merged in to
the results so that you can show/hide the different log levels. ) What
we have for reporting doesnt work for that, nor does it handle the test
scenario of "same tests, Java 1.5 and 1.6 on windows/linux/OSX' , where
you want the results merged under each test, rather than sorted by machine.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
