Sorry, you're right - I should have tested the example I just used first.
What I mean to point out, though, is: shouldn't the parser be able to differentiate based on indentation in my original example?
item details:
- term: "term0"
id: 7890
type: typeA
- term:
- "term1"
- "term2"
id: 1234
type: typeB
should create ... {"term": ["term1", "term2"]} ...
and
item details:
- term: "term0"
id: 7890
type: typeA
- term:
- "term1"
- "term2"
id: 1234
type: typeB
should create ... {"term": ""}, {(""): ["term1", "term2"]} ...
Right now this parser interprets both as ... {"term": ""}, {(""): ["term1", "term2"]} ... unless I rearrange the key entries like so:
item details:
- term: "term0"
id: 7890
type: typeA
- id: 1234
type: typeB
term:
- "term1"
- "term2"
in which case, it interprets it as I expected: ... {"term": ["term1", "term2"]} ... (but this also seems like inconsistent behavior).