Updated addLiteral to properly support XMLLiteral

[jamsden]

Parsing an XMLLiteral should store a normalized (sorted attributes) XML
string as the value for the literal, not the parsed XML dom.

Addresses issue #75.

[dmitrizagidulin]

@jamsden @timbl - is this PR ready to merge?

[jamsden]

It doesn’t do the attribution normalization (sorting the attributes) so === works on XMLLiterals that are the same except for ordering, but it does at least capture the correct contents.

Not sure === on XMLLiterals is a high priority, so this can possibly be deferred.

On Jul 19, 2016, at 4:16 PM, Dmitri Zagidulin notifications@github.com wrote:

@jamsden https://github.com/jamsden @timbl https://github.com/timbl - is this PR ready to merge?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub #133 (comment), or mute the thread https://github.com/notifications/unsubscribe-auth/ABECqmFsmnFo0esSvscMZSud51m-2NcMks5qXTCVgaJpZM4JPIrn.

[timbl]

Sounds like a definite improvement - suggest merge

[timbl]

(I had earlier left the DOM pointer as th value, as that is what is useful to software. - if we have the DOM then maybe it would be be usefulk to leave a pointer to it, so that the caller doesn't have to waste time parsing it again?

[jamsden]

Good point, but it doesn't follow the RDF/XLM spec and creates inconsistencies for consumers. For example, it breaks OSLC resources.

Sent from my iPad

On Aug 22, 2016, at 2:02 PM, Tim Berners-Lee notifications@github.com wrote:

(I had earlier left the DOM pointer as th value, as that is what is useful to software. - if we have the DOM then maybe it would be be usefulk to leave a pointer to it, so that the caller doesn't have to waste time parsing it again?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub, or mute the thread.

[timbl]

This now needs to be patched in
rdfxmlparser.js
at line:
if (this.parent.datatype) {
Line 156 in rdfxmlparser.js