On Oct 22, 2006, at 1:02 PM, David Bovill wrote:
A little lost but very interested regarding this thread - is there a
stack here to test? My interest here would be in unit test type
frameworks for handlers and more complex chunks of code. I
implemeneted this by calling the handlers against a standard suite of
input data - testing these against output data - also timing them.
Changes to the handler could then be compared to see how it affected
the test suite.
I'd like to add this back again - is this what you are implementing?
I think we are exploring ideas on what direction something like this
might go. For example, should there be a card per question or a
stack per question or something else? Should pass/fail results (for
pass/fail questions) be returned or should a report be written to a
file? Should this include characteristics such as timing or distance
to the first base line in fields or just pass/fail?
Implementing is a good idea. Right now I'm just causing trouble.
I like the suite of data idea. This might be different for different
tests. Some of that might be general and automatic. I like to test
within 3 of limits and powers of two (or 256) for numbers and
sometimes powers of 10. So test makers might share such functions.
The same might apply to lengths of data. There might be other suites
of data for null testing or character set testing.
I like the idea of a general tool that can be used for both testing
Revolution features and our own handlers.
Dar
_______________________________________________
use-revolution mailing list
use-revolution@lists.runrev.com
Please visit this url to subscribe, unsubscribe and manage your subscription
preferences:
http://lists.runrev.com/mailman/listinfo/use-revolution
