Currently, COBOL parser generated code produced by ANTLR is exactly the same for all ANTLR versions supported by various Spark version starting from Spark 2.4.
But if fat JARs are not carefully generated by a user of Cobrix, ANTLR version included in the COBOL parser can conflict with Spark.
This has not yet been observed, but it is good thing to protect against future potential issues.
Feature
Shade ANTLR runtime in the parser to avoid ANTLR potential incompatibility issues.
Example [Optional]
--
Proposed Solution [Optional]
Investigate how sharing can be done in published artifact in sbt.
Background
Currently, COBOL parser generated code produced by ANTLR is exactly the same for all ANTLR versions supported by various Spark version starting from Spark 2.4. But if fat JARs are not carefully generated by a user of Cobrix, ANTLR version included in the COBOL parser can conflict with Spark.
This has not yet been observed, but it is good thing to protect against future potential issues.
Feature
Shade ANTLR runtime in the parser to avoid ANTLR potential incompatibility issues.
Example [Optional]
--
Proposed Solution [Optional]
Investigate how sharing can be done in published artifact in sbt.