lexer now produces a vector of heap allocated tokens

This removes the problem of possibly expensive copies occurring due to
working with tokens produced from the lexer (that C++ just... does):
now we hold pointers where the copy operator is a lot easier to use.

I want expensive stuff to be done by me and for a reason: I want to
be holding the shotgun.

asm/lexer.hpp

@@ -91,6 +91,6 @@ enum class lerr_t
 };
 const char *lerr_as_cstr(lerr_t);
 
-lerr_t tokenise_buffer(std::string_view, std::vector<token_t> &);
+lerr_t tokenise_buffer(std::string_view, std::vector<token_t *> &);
 
 #endif