Re: [code-review] WebService::Validator::HTML::W3C

* at 14/11 17:01 +0000 Struan Donald said:
> Hi,
> 
> Any thoughts/improvements/whatever on the attached?

ok, so slightly belated thanks[0] for the feedback.

Am I supposed to give feedback on the feedback? Here goes anyway.

Class::Accessort makes sense, especially now there's a second module
in the distribution (the error one Tony suggested).

Darren's idea of returning the response as an object kind of makes
sense but I prefer the original interface. And passing an array of
URIs to the validate method seems too much like encouraging people to
throw a whole load of links at the validator at once. (which is a bit
of a straw man I'll admit)

I've refactored the validate method a little to make it easier to
subclass. I imagine more could be done but I need to think a little
how to make it work best.

It's not possible to tell what doctype the validator has attempted to
validate a page as without parsing the XML it returns and as I don't
want to force the XML stuff on people it's not doable within the
confines I've set. Whether they're sensible confines is another matter
but there you go.

Errors are now objects instead of hash references.

oh, and I've made some other code tweaks along with what seem to be
generally held beliefs about readability.

thanks for what's been useful feedback

Struan

[0] errant mailfilters to blame. sigh.