Danny Angus wrote:
What we're really interested in here is being able to comit designed
tests before we commit the code which passes the test.
IMO that *is* TDD
We who?
I think that before using TDD we should reach 100% test coverage.
What are the advantages of such complexity you want to introduce?
I see too much complexity for the provided advantages.
We have a coverage of 30% and we are talking of splitting tests in 2
separate test suites, using TDD and so on: this seems to me too much an
ideological, unrealistic, approach to things.
Stefano
What we're up against is the knowedge that some of the passes are a
low priority, and not critical enough to prevent others from working
on other aspects of James.
What we really want to do is to report against our designed tests, and
improve the pass rate, rather than naievly seek 100% pass rate at all
times.
I like the idea of having two test suites.
We could have mandatory-tests and optional-tests, the first halts on
failure, the second doesn't.
When an option test can be passed by James it is moved into mandatory,
so that any future changes continue to pass.
WDYT?
d
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
