cool, could we add a fail-on-error option which is set to true by
default on all the steps, but could be overridden if different
behavior is desired.
also any thoughts on ?
also wondering 'out loud' if whilst we are here to investigate any
extension attributes to p:data / p:document (or p:load) to make
encoding explicit between binary/unicode/other encoding ... I know we
have content-type for p:data but I think we default to whats inside
the document to inform with p:document.