Every serializer/deserializer for nested, dynamically sized blocks of binary data contains a buggy, ad hoc variant of this. I'd already be happy with something

Every serializer/deserializer for nested, dynamically sized blocks of binary data contains a buggy, ad hoc variant of this. I'd already be happy with something like Erlang's standard binary matching library[0] as part of each language's stdlib. Block deduplication, block reference counting (and the requirement to in-file-GC unused blocks), n:m references are things that arise in some formats that don't seem to be part of this language? [0] http://erlang.org/doc/efficiency_guide/binaryhandling.html http://erlang.org/doc/efficiency_guide/binaryhandling.html