On 05/11/2014 00:06, Martin Buchholz wrote:
:
So ... there is the question whether we are tightening the spec for
Attributes and its subclasses or only for Attributes.
Subclasses of Attribute that would change the order are likely to be
very rare (is there a use case?) - I'm OK with those becoming
non-spec-compliant subclasses, especially given that they will
continue to work just fine in practice. So in my proposal the
tightening applies to subclasses as well.
What you have is okay but I think if I were doing this then I would keep
the existing wording in the constructor (the update to the class
description is good). The reason is that the new wording might distract
the reader into thinking about the iteration order of the given
Attributes when it's just not interesting here.
-Alan.