PS: maybe better using function_declarator_p??? I think I regression tested that variant too, at some point.

Paolo.

Reply via email to