In #26 (comment) it was noted that the current approach to permissive parsing is limited by the fact that the OpenPulseNodeVisitor (and some routines it relies on, like the @span decorator) do not correctly handle ANTLR errors. It may be worth identifying what useful error recovery would look like, and how it could be implemented. This strikes me as most relevant for developers who are experimenting with grammar changes, although perhaps a case can be made for a broader base of users.
In #26 (comment) it was noted that the current approach to permissive parsing is limited by the fact that the
OpenPulseNodeVisitor(and some routines it relies on, like the@spandecorator) do not correctly handle ANTLR errors. It may be worth identifying what useful error recovery would look like, and how it could be implemented. This strikes me as most relevant for developers who are experimenting with grammar changes, although perhaps a case can be made for a broader base of users.