Adam Spiers <[EMAIL PROTECTED]> writes:
> Piers Cawley ([EMAIL PROTECTED]) wrote:
> > Adam Spiers <[EMAIL PROTECTED]> writes:
> >
> > > In TestRunner, start_test() prints "." for each test, add_error()
> > > prints "E", add_error() prints "F", and add_pass() does nothing.
> > >
> > > Wouldn't it be less confusing to change start_test() to do nothing,
> > > and add_pass() to print "." ? That way, the number of characters
> > > output by a given set of tests always remains constant, no matter
> > > which fail or encounter errors.
> >
> > What's JUnit's behaviour in this context? We should do that.
>
> It looks like add_pass is something we (Brian?) added. I like it, and
> I think we should take advantage of it -- I'd prefer us to improve on
> JUnit and then document unexpected differences rather than aim for a
> 100% clone. Does that conflict with other people's ideas of the goals
> of PerlUnit?
Given that this is purely behaviour of a Listener I suppose we could
subclass TestRunner to give 'TestRunner::StrictXUnit',
'TestRunner::PerlUnit' or some such pair. Then users get to choose. If
we then document how this works well, they can even roll their own to
suit their own environment (cf Test::Unit::HarnessUnit).
I'd still like to see something that looks just like the default JUnit
behaviour, if only so that folks can have consistent test output in
environments where they're working in both java and perl (say).
--
Piers
_______________________________________________
Perlunit-devel mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/perlunit-devel
