Closed j0ni closed 7 years ago
Otherwise looks great.
I implemented the CRC check in the crudest way, and because the MarshalJSON
code makes calls to libspector
to resolve package and package version, this results in duplicate calls occurring per run.
To continue using the CRC check, I think I should either pre-load all that data, or instead of capturing the data structure, capture the JSON blob of bytes.
LOOKS GOOD TO PHILLIP
This introduces libspector, and supports two use cases:
agent.conf
file, and track processes which match.There's a companion coalmine PR appcanary/coalmine#196 which adds datamodel for the second use case, as well as an endpoint. Not sure there's much point in keeping the first part of the implementation here, as configuring it will simply duplicate a subset of the data generated by the second use case.