Algorithm updates and tests to allow @container: @set on @type
#41
Conversation
…n alias in 1.1. Remains an error to redefine `@type` in 1.0. For w3c/json-ld-syntax#34.
| "name": "Keywords may not be redefined", | ||
| "purpose": "Verifies that an exception is raised on expansion when processing an invalid context attempting to define @container on a keyword", | ||
| "input": "expand/e042-in.jsonld", | ||
| "expect": "keyword redefinition" |
There was a problem hiding this comment.
Does this need to be in 1.0 mode?
There was a problem hiding this comment.
I added a test to be sure that a 1.0 processor would fail if @type were so defined. We didn’t have good relevant tests before.
There was a problem hiding this comment.
If processingMode is not set in the test definition, it defaults to 1.0. expand/e042-in.jsonld is the following:
{
"@context": {
"@type": {"@container": "@set"}
},
"@type": "http://example.org/type"
}which a 1.0 processor should fail on.
There was a problem hiding this comment.
Yeah, I was forgetting it defaulted to 1.0.
| "expect": "compact/0105-out.jsonld", | ||
| "option": {"processingMode": "json-ld-1.1", "specVersion": "json-ld-1.1"} | ||
| }, { | ||
| "@id": "#t0106", |
There was a problem hiding this comment.
Won't this test be a negative test in 1.0 mode?
There was a problem hiding this comment.
I think the tests are reasonable to check both 1.0 and 1.1 modes. The negative test is in expansion, to make sure a 1.0 processor fails. A 1.0 processor compacting with an alias of @type will allow a @container definition but ignore it when actually compacting.
There was a problem hiding this comment.
Ah, I thought it failed in 1.0.
There was a problem hiding this comment.
We actually never checked that keyword aliases should only be for simple term definitions; we just ignore everything else. We could consider tightening these up in 1.1, and require that keyword aliases only be simple term definitions.
or an alias in 1.1. Remains an error to redefine
@typein 1.0.For w3c/json-ld-syntax#34.