Support encode/decode of RDF Values in the REST API for the "access path" methods, providing correct encode/decode of the URIs and Literals. For example, a Literal with an embedded quotation mark or a URI with an embedded GT or LT symbol. The most general solution for this is to handle the SPARQL code point sequences , which is a bit of work (in fact, Sesame only handles one of the two code point sequence forms
- it seems that JavaCC does not recognize the 8 character code point sequences. For the moment, the implementation ignores this issue. The relevant code is in EncodeDecodeValue? and can be enhanced once the API change is proven useful. The most critical problem that someone is likely to encounter is a quotation mark embedded within a Literal. At present, that will not come across the API correctly. (It will be reported as a bad request exception at the HTTP layer.)
See EncodeDecodeValue and TestEncodeDecodeValue and the link above to the SPARQL code point escape sequence specification. We only really need to handle the <code>\u hh hh hh hh</code> sequence for now as the Sesame SPARQL parser does not handle the 8 hex digit decode.
The changes need to be made against both the TERMS_REFACTOR_BRANCH and the 1.0.x maintenance branch.