Fix a malformed cache set

[cameronterry]

This PR resolves issue: https://github.com/cameronterry/dark-matter/issues/98

After changing the logic for Primary domains database calls and cache setting, moving that into DarkMatter_Domains, it appears that one small bug and a leftover call in DarkMatter_Primary is causing the excessive database calls.

• The small bug is a malformed call to wp_cache_set(), which forgot to include the value - the domain (mappeddomain1.test) - and was instead setting it to dark-matter which is the cache group.
• The get() method DarkMatter_Primary was still calling its own set() method, which was causing even cache miss to essentially issue an UPDATE SQL query. This has been removed as it is handled by DarkMatter_Domains now.

[cameronterry]

The fix during this was to flush cache, however that might not be feasible for some installs. Checking if that is actually a requirement.

[cameronterry]

The fix during this was to flush cache, however that might not be feasible for some installs. Checking if that is actually a requirement.

Simulating the deploy of 2.3.1 by switching between this branch and develop (which is 2.3.0), this change would have required a cache flush. However by moving the check and primary domain cache update to a different place, the need for installations using Dark Matter needing to run wp cache flush can be mitigated: https://github.com/cameronterry/dark-matter/pull/99/commits/5aa734aab61236dbd491544cf57a8ec850d121e4.