Open pnkfelix opened 1 year ago
after having slept a bit... this requirement to have a significant set of tests in both places may be overkill.
A better middle ground will be to enable easy development of C-based tests; i.e., ensuring that all the Rust client requests have C equivalents, so that when oddities in a Rust-based test crop up, one of the available options for investigation should always be "Does the same problem arise in a C-based port of the test...?"
(spawned off of issue #18)
Until we have confidence that issues like #18 are not arising, we need to encode each basic test in C first and in Rust second.