[tap-l] What's with verbose rewriting the TAP?
andy at hexten.net
Thu Feb 19 11:29:00 GMT 2009
On 18 Feb 2009, at 20:54, Ovid wrote:
>> From: Michael G Schwern <schwern at pobox.com>
>> I've just noticed another instance of "prove -v" rewriting the TAP
>> so that
>> you're not really seeing what the test output.
> Mea culpa. I did that when I first wrote TAP::Parser to ensure a
> nice, consistent output showing the end-user what TAP::Parser is
> interpreting each TAP line as.
> So yeah, I could see this being an issue. Should we have --raw to
> show the original TAP? Or --normalize to show what TAP::Parser is
> intending? I think Schwern is right that we should show the actual
> output by default. (/me wonders if this will break any tools
> relying on the new parser)
Yeah, I wonder that too. Since nobody else has ever complained about
the munging we do I'd lean towards adding a --raw option rather than
changing the behaviour of -v.
Getting raw output should be easy - call raw() on the result objects
instead of as_string().
Andy Armstrong, Hexten
More information about the tap-l