Unclear how to set original base URL in Context Processing

[kasei]

(Related to #265, #356)

Context Processing Overview 4.1.1 says:

When an active context is initialized, the value of the original base URL is initialized from the original documentUrl of the document containing the initial context . This is necessary when resetting the active context by setting it to null to retain the original default base IRI .

What about when there is no documentUrl? Is it null, or is it the current base IRI? The documentUrl field is only relevant when the LoadDocumentCallback has been used.

This also impacts compact() step 3, and expand() step 4.

[gkellogg]

This is not dissimilar from Turtle serialization from a form, rather than a remote document. Also, the JSON-LD Playground. Implementations provide a reasonable default to use as a document base.

Perhaps if step 7 of the expand() API definition provided for the case where there is no remote document having a documentUrl. But, really, the use of a relative URL to locate a context from a form is non-sensical. Maybe step 5.2.1 should generate an error if context does not result in a valid IRI after resolution. Perhaps invalid IRI mapping could be used for this case too.

[gkellogg]

Actually, the LoadDocumentCallback will generate a "loading document failed" error if the URL can't be resolved, so no new error detection is required.

I did add changes to the expand() API definition to fallback to JsonLdOptions/base if there is no documentUrl, which is how it's typically handled in form-based parsers. Implementations may provide a default for this, but we don't need to describe that non-normative behavior.

[gkellogg]

@kasei see pending change in #377.

[kasei]

@gkellogg I think the changes to expand() in #377 are good, but the description in 4.1.1 is still misleading as in some cases there will simply not be a relevant documntUrl.

[gkellogg]

Updated in 50a50ca0356146974532507c5e01d3ce2798e324, and to give base a default of null.

[gkellogg]

Updated in #377.