Don't treat class attributes as sub attributes in perly.y

[leonerd]

We need to use a dedicated grammar type; or at the very least, not use subattrlist for doing this. The previous use of subattrlist had the side-effect of passing every attribute list for a class via the apply_builtin_cv_attributes() function. That was pointlessly redundant, though currently harmless. It may stop being harmless in a future improvement to the way attributes are handled, so best not to do that any more.

• This set of changes does not require a perldelta entry.

[happy-barney]

Just one low priority comment (mostly coming from my ongoing adaptation of grammar for GLR parser):

It will be nice to have dedicated non-terminals for each usage of attributes, for example:

class_attributes
    :   optattrlist
    ;
field_attributes
    :   optattrlist
    ;

Reasoning:
in future they can be upgraded to something like this (meaning, not implementation):

class_attributes
    :   { push_attribute_context (CLASS); }
        optional_attributes
        { pop_attribute_context (); $$ = $optional_attributes; }
    ;

[leonerd]

It will be nice to have dedicated non-terminals for each usage of attributes, for example:

I think if that sort of thing were to be added, it would be easy enough to split the grammar rules at that point. I'd prefer currently to keep the rules agnostic about what their target is, as it keeps them simpler. There's no need to have different kinds just at the moment.

[happy-barney]

Comment was low priority, just to make it little bit nicer. I guess you are right, related change once it will introduced may be even better to be part of that changeset.