Summary

Now that we can run unittest ( #1587 ) I think we should start running the CPython tests (Lib/test) to test RustPython.

Detailed Explanation

My suggestion is to do the following:

1. Copy all tests that can compile and run from CPython Lib/test to RustPython Lib/test.
2. Mark all the failed test with @unittest.skip("TODO: fix RustPython").
3. Add a job for both linux and windows CI that will run the tests.

Drawbacks, Rationale, and Alternatives

• There are many tests in CPython so we might want to consider choosing only a subsets of tests to work on at first.
• We should see how slow it would take to run all the tests. We might want to run the tests in parallel.

Unresolved Questions

• Should we mark the tests that currently don't pass with @unittest.expectedFailure so we will know when something is fixed?
• What should happen to the snippets tests? Should we keep a duplicate set of tests?
• Would we add tests to CPython set of tests?