[tap-l] What's with verbose rewriting the TAP?

[Michael Peters]

Andy Armstrong wrote:

> 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.

This is probably because no one writes TAP by hand they all just use Test::Builder and friends. But 
if TAP becomes more widespread, there will be more TAP producers in other languages and it would be 
nice if prove and friends don't change that TAP by default.

> Getting raw output should be easy - call raw() on the result objects  
> instead of as_string().

But it really should be the default, IMO. Normalizing data is an extra step that doesn't really give 
any benefit unless you're trying to track down why TAP::Harness is interpreting your TAP as 
something other than you intended, right?

-- 
Michael Peters
Plus Three, LP