Review #2304 for code improvements: guards for a missing .query in set operations #2646

This is not a bug report. It is an invitation to look at #2304 again, because it may guard against a symptom rather than fix the cause. The goal is to understand the change and decide whether anything should be improved. The notes below come from a quick read of the diff, not from a full analysis.

The concern:

• Guards instead of a cause. In queryfn3 (src/38query.js) eight places now check query.unionallfn.query && ... (and the same for unionfn, exceptfn, intersectfn) before setting the right operand's modifier. The question is why the compiled function of a set operation sometimes has no .query. If that case is legitimate, setting the modifier where the operand is compiled, in one place, would remove the eight checks.
• Evaluate UNION / EXCEPT / INTERSECT chains left to right to close #202 #2582 (set operation chains, 5.0) replaces these branches with one applySetOp; the guards may be worth folding into that.

Related: #635, #2582.