Pierre-Antoine Champin
Pierre-Antoine Champin
I'm not sure about that... Being a valid IRI (resp. language tag) is part of the contract of the datatypes you expose. So letting the parser produce such "impure" data...
The limitations in PN_LOCALS are not all related to IRI syntax. @drobilla pointed out to me some time ago that, for example, slashes are forbidden in Turtle pnames *only* because...
> I love this name. It was intended as a working title, but why not ;) The English version is "care bear", but I'm not sure it has the same...
> although we'd need to agree on what the grammar is I don't think that we *need* that. The rationale of the non-strict mode is not to define an alternate...
@drobilla Granted, the analogy with HTML has its limits. Humans are more tolerant than machines, and so lenient HTML is less of a problem than lenient RDF. In any case,...
Just gave it a quick try, but unfortunately, this change is much more disruptive than I anticipated... * if we go that way, `Subject` and `Term` can not implement `Copy`...
I can't really think of a better word :-/ Actually, if another (useful) way of generalizing RDF triples/quads was proposed, I would consider adding support for it in gtrig. If...
Great. I just integrated it in sophia https://github.com/pchampin/sophia_rs/commit/f3ed2f8d010f89540d5bd23cf3e857d802e31621 . I'll have a look at your IRI parser.
Renaming this issue, as it is not specific to XML (the same is true of Turtle or TriG.
I see that https://github.com/oxigraph/rio/releases/tag/v0.7.1 addresses it for the Turtle-family of parsers. Nice! :heart: