Prevent keyword alias to be declared as prefix #154
Conversation
There was a problem hiding this comment.
From the discussion, we should also have a test without @prefix: true that would not flag an error, but also not expand.
Resolution #4: When encountering a term in the form of a compact IRI for which the prefix would expand to a keyword, do not expand the term.
| <li>If the <a>prefix flag</a> of <var>definition</var>is set to <code>true</code>, | ||
| and its <a>IRI mapping</a> is a <a>keyword</a>, | ||
| an <a data-link-for="JsonLdErrorCode">invalid term definition</a> | ||
| has been detected and processing is aborted.</li> | ||
| </ol> |
There was a problem hiding this comment.
Might want to say "has the form of a keyword", so that @the-next-keyword would be rejected too. This is related to another discussion from the F2F, for which we don't yet have an open issue: we want to discourage the use of things that may be keywords in the future (should => must).
There was a problem hiding this comment.
Per w3c/json-ld-syntax#16, keyword-like strings MUST be ignored and SHOULD result in a processor warning. I'll add text and tests to ensure they're ignored (with and without @vocab), but we can't test for warnings.
|
Also, after changing the test manifest, please run |
|
I have more tests coming on variations of using keyword patterns, to satisfy w3c/json-ld-syntax#16 as well. |
NB: should we have a dedicated error for that?
…egin with `@`, or the term maps to something of this form.
…o add toRdf verion of pr33.
gkellogg
force-pushed
the
prevent-keyword-prefix
branch
from
84cf81e to
c9e9d90
Compare
addresses #153
Preview | Diff