Why does SPL token testing rely on spl-token-interface instead of local changes?

[victorchukwuemeka]

Hi SPL team,

I noticed that the tests in the program crate import instructions from spl-token-interface rather than the local implementation in interface. I have made some local modifications to experiment with new features.

I’m curious about the rationale:

• Is this for stability or backward compatibility?
• Are there design reasons for preferring the published interface over local changes during testing?
• How should contributors approach testing local extensions without affecting the main test suite?

Additionally, is it okay for contributors to build entire features, like a suspend authority, and submit a PR, or is there a preferred workflow for proposing larger changes?

Just trying to understand the design approach here. Thanks!