Java target should not allow any token or rule to match the grammar name

Currently, the code generation process does not report an error when an imported token (via the `tokenVocab` option) defines a token which matches the name of the parser grammar. This should be reported as an error for either the Java target specifically, or for the tool in general.

Start by tracing Java-target code generation and how imported tokenVocab definitions are validated against the parser grammar name. Reproduce the case with a minimal grammar and imported token definition, then verify that generation reports an error when either a token or rule matches the grammar name.

Written by the indexing model from the issue text.