Improve Grid Testing
The goal of the improved testing framework is to thoroughly define which grid combinations are fully tested, as well as to provide a mechanism to better track errors back to the tests. The combined parameter space of Grid methods is huge, so it is import to specify exactly what configurations of grids, coordinates, distributions and remappings are fully functional, and when errors arrise, to track them back to the exact configuration that failed. This will be an expanded version of the current testing code in place for redistribution and regridding. The testing framework intends to be more general and comprehensive than the previous testing code and would report the results as human readable descriptions of the test configuration, using the proposed taxonomy for field representation, redistribution and regridding.
Task breakdown
- Configurable test suite (separate by major function using proposed field syntax)
- Parser to translate field syntax to action
- Automated generation of reports specifying tested grid configurations
