Add indexed-repeat tests

[lognaturel]

TODO:

• Modify first test to use a relative reference for index and add a group around the repeat so the contextualization is clear
• Add nested repeat tests
  getodk/web-forms@541cd12#diff-d7f5e2a9fa88fc71566d393928de170be5700174a890bc68e5195a4f8c6e1657R335
• Demonstrate that the order of the repeat/index pairs doesn't matter
• Multiple repeat/index pairs where the repeat refs are relative
• Include the equivalent XPath path expression and also evaluate that

What has been done to verify that this works as intended?

These are test-only additions intended to capture the intended indexed-repeat behavior. It also updates the function's comment to better reflect the language that we use these days to talk about repeats.

I wasn't there when the function was added by my understanding is that it was introduced so end users wouldn't have to write XPath predicates in the common case where they want to use a numeric index to access a specific repeat instance's value. XPath predicates are harder to reason about because the predicate expression is evaluated in the context of the path that's being filtered. It's particularly baffling when using parallel repeats because you end up having to write something like position(.) = position(current()/..).

The indexed-repeat function makes it so the index expression is evaluated using the bound node's context which is intuitive to users. The first parameter is the full reference to a field that needs to get indexed, and then there are some number of pairs of repeat references and indexes.

A simple algorithm to think about the intended behavior (not exactly what JR does):

• contextualize all the parameters based on the single node binding
• evaluate all the index parameters (param 3, 5, 7, etc), these should now be literal integers
• for each repeat reference parameter (param 2, 4, 6, etc), add a positional predicate at that reference level in the base reference (param 1)

Why is this the best possible solution? Were any other approaches considered?

How does this change affect users? Describe intentional changes to behavior and behavior that could have accidentally been affected by code changes. In other words, what are the regression risks?

Do we need any specific form for testing your changes? If so, please attach one.

Does this change require updates to documentation? If so, please file an issue here and include the link below.