Follow-up on make vs. Ninja refresh_cache behavior

[thomas-bc]

F´ Version
Affected Component

---

Problem Description

Follow up to nasa/fprime-tools#276

There's a discrepancy between how Ninja vs. Make refereshes cache (i) by themselves and (ii) how we trigger them through fprime-tools.
See the PR comments. Need to investigate and make sure we haven't broken the Make pattern too much