Unlike fields, tags, methods, and method parameters, type names don't play any role in transportation. Types have names, and "_type" hint field in serialization format, but they're only for human programmers. These "_type" hints are ignored when a program interprets a serialized value.
So the parser should disallow to add a behind name to type declarations.
Currently it even raises an error if two types have the same behind name (but two different facial names) for no reason:
record foo/dup (text x);
record bar/dup (text y);
examples/type-behind-name.nrm:4:1:
the behind type name `dup` is duplicated
^
Unlike fields, tags, methods, and method parameters, type names don't play any role in transportation. Types have names, and
"_type"
hint field in serialization format, but they're only for human programmers. These"_type"
hints are ignored when a program interprets a serialized value.So the parser should disallow to add a behind name to type declarations.
Currently it even raises an error if two types have the same behind name (but two different facial names) for no reason: