Fetching data islands in RDFa too zealous

[doriantaylor]

Hi,

I have RDFa markup that has <script> tags that have a type that is Parsable but they have src attributes so there's nothing in the bodies. The following code thus blows up:

rdflib.js/reference/fetcher-classes.js

Line 117 in 85a72c8

rdfParse(scripts[i].textContent, kb, xhr.original.uri, contentType)

Would it be possible to add an option to Fetcher.load to turn this behaviour off? Or otherwise skip over it if there's an src attribute, or otherwise have it no-op or wrap in a try block or something?

[csarven]

That makes sense. The code should do something like textContent.trim().length == 0 before committing to parse.

Dorian, if your script src is referring to JavaScript, you can omit the type attribute ( https://html.spec.whatwg.org/multipage/scripting.html#attr-script-type ) - perhaps that's of some help for the time being.

[csarven]

#693

[doriantaylor]

nah it's my lazy bastard transclusion mechanism. The thing that's spitting that out though (or rather its neighbours; warning: SLOW) has content negotiation with a JSON-LD variant so just added another variant to the one that was tripping up.