AFAICT the purpose of a problem matcher is to capture errors in the output, parse them and provide links to location where the problem originated. Which sounds useful for BATS - we could e.g. link to failing test definition.
However, I'm not sure if BATS provides enough information to do that, it seems you may need at least a file and line number to do that, and bats only provides test name and (in --tap mode) sequence number which I'm not sure if you can correlate back to location.
Also it's not quite clear what the schema of the JSON file for defining problem matchers is, the best thing I could find so far is actions/toolkit#56 which points to this vscode doc.
AFAICT the purpose of a problem matcher is to capture errors in the output, parse them and provide links to location where the problem originated. Which sounds useful for BATS - we could e.g. link to failing test definition.
However, I'm not sure if BATS provides enough information to do that, it seems you may need at least a file and line number to do that, and bats only provides test name and (in
--tapmode) sequence number which I'm not sure if you can correlate back to location.Also it's not quite clear what the schema of the JSON file for defining problem matchers is, the best thing I could find so far is actions/toolkit#56 which points to this vscode doc.