Is there an existing issue for this?

• I have searched the existing issues

Problem statement

The standard output for ANTLR4 recognizers is just to print a basic message to standard err. It would be useful to collect these, and provide rudimentary additional information such as number of errors. This will enhance testing and also allow the builders and any other walkers to provide semantic errors or errors about inability to convert code (linting down the line, perhaps).

This will make it easier for test specifications to stop if parsing errors occur, amongst other things.

Proposed Solution

A quick and dirty collector will be written that collects errors without doing anything too fancy. Maybe go as far as printing the source line in error with a carat, but such things can always be added later.

Additional Context

No response