Update TruncatedDivide Quantity unit semantics to match plain Divide

[dehall]

This PR updates the expected results for TruncatedDivide tests on Quantity operands, to match the unit conversion semantics from plain Divide. Specifically, (from https://cql.hl7.org/09-b-cqlreference.html#divide )

performing division of quantities with the same units, the result will have the default UCUM unit ('1')

The spec doesn't explicitly call out this same note for TruncatedDivide, it only says (from https://cql.hl7.org/09-b-cqlreference.html#truncated-divide )

For division operations involving quantities, the resulting quantity will have an appropriate unit as determined by application of the UCUM specification.

But, the spec also says the same "appropriate unit [...]" for plain Divide, so if instead the semantics are intentionally different for TruncatedDivide vs Divide, I think that would need to be made more explicit.

For quick reference, here is the corresponding test for plain Divide

cql-tests/tests/cql/CqlArithmeticFunctionsTest.xml

Lines 184 to 194 in 727219f

	<test name="Divide1Q1Q" version="1.0">
	<capability code="ucum-unit-conversion-support"/>
	<expression>1'g/cm3' / 1'g/cm3'</expression>
	<output>1.0'1'</output>
	<!-- See http://unitsofmeasure.org/ucum.html#section-Semantics
	particularly "the neutral element 1 (so called the unity)". -->
	<!-- See HL7 STU comment #1442. -->
	<!-- When dividing quantities with like units, the specification
	should state clearly that the expected UCUM unit would be the
	default unit (`1`). -->
	<!-- See https://github.com/cqframework/clinical_quality_language/issues/193 -->

[bryantaustin13]

Approved. Once merged, expect the reference engine to start failing, which is the point.

[cmoesel]

Approved (on behalf of MITRE's JS cql-execution team).