- From: Shane McCarron <shane@spec-ops.io>
- Date: Sat, 11 Jun 2016 12:18:12 -0500
- To: Discussions about Test Development <testdev@lists.spec-ops.io>, W3C Public Annotation List <public-annotation@w3.org>
- Message-ID: <CAJdbnOC78Og6zH36xU2Qr+=KHDs2=F0VTCA4eTdy=S6Ab3ia8w@mail.gmail.com>
In the current declarative structure for JSON tests, we have a concept that you can nest lists of assertions... assertionList : [ thing.json, otherThing.json, thirdThing.json, [ anotherThing.json, yetAnother.json] ] This sort of structure, while nested, does not impose semantics on the item that is nested. My proposal is that we actually do not support a nested list. Instead, we rely upon my new ConditionObject. A ConditionObject looks like this: { "title": "Condition Object", "description": "A pseudo-test that will get a result from the aggregate of its children", "assertionType": "must", "expectedResult": "valid", "errorMessage": "Error: None of the various options were present", "compareWith": "or", "assertions": [ "common/has_context.json", "common/has_id.json" ] } In this way the nested list itself represents a sort of a test - it has a result and its subtests have results too. Frankly, I don't know if any of this is necessary. But having this structure means that when something gets nested we can have clear explanatory text associated with it, an explicit result from the aggregate of what is nested, etc. So, the questions: 1. Has anyone used the nested assertion lists? 2. If so, would it be relatively easy to change? 3. If not, do we need any sort of nesting support at all? 4. If so, is this change acceptable? -- Shane McCarron Projects Manager, Spec-Ops
Received on Saturday, 11 June 2016 17:19:06 UTC