This is just a suggestion, but I thought it might be useful for catching regressions.
We could assemble a collection of wild Unison code, starting with whatever's out there at the moment like Runar's repo. We could have tests that send it through the edit + slurp cycle, and use them to detect regressions, whether in the parser, pretty printer, or typechecker.
This is just a suggestion, but I thought it might be useful for catching regressions.
We could assemble a collection of wild Unison code, starting with whatever's out there at the moment like Runar's repo. We could have tests that send it through the edit + slurp cycle, and use them to detect regressions, whether in the parser, pretty printer, or typechecker.
Thoughts?