search: query printer preserves field aliases where possible
Created by: rvantonder
We typically normalize things like content:xyz
so when this goes through a pretty printer you only get xyz
out. We store whether a term was aliased, so we can (for some fields) recover content:xyz
or r:foo
(which matches the user input). Turns out the information on content:
is significant for bug fix I'll need in a follow up PR.
Test plan
Updates tests.