Fix for #1824 s,http://rdlib.net,http://rdflib.net,g

[ghost]

Summary of changes

Fix for #1824 reported typo of http://rdlib.net/, change http://rdlib.net to http://rdflib.net in three files.

• Fixes Typo "http://rdlib.net/" #1824

Checklist

• Checked that there aren't other open pull requests for the same change.
• Considered granting push permissions to the PR branch,
  so maintainers can fix minor issues and keep your PR up to date.

[coveralls]

Coverage increased (+0.0006%) to 88.423% when pulling 5247aae on gjhiggins:fix-for-issue1824 into ccb9c4a on RDFLib:master.

[edmondchuc]

Is this considered a breaking change?

[ghost]

Is this considered a breaking change?

Now that you mention it, probably yes, given that the skolemization oftimbl-card.n3pre- and post-change yields graphs that are not ==.

So, do I just close the PR and keep it around for when we're ready?

[edmondchuc]

Not sure what to do. I just thought I'd mention it.

[niklasl]

Given that neither domain is under the control of RDFLib, oughtn't we remake it further anyway? E.g. using rdflib.dev (if we bet on that being maintained during the lifetime of this project).

Not that skolemized blank ID:s "are to be" dereferenced, but as it is designed so (with a .well-known rather than e.g. a tag: URI), I believe the prudent and proper, qualitative way is to avoid "abanonable" domains...

[aucampia]

Given that neither domain is under the control of RDFLib, oughtn't we remake it further anyway? E.g. using rdflib.dev (if we bet on that being maintained during the lifetime of this project).

I'm in favor of this, we should not use domains we don't own so rdflib.dev will be better another option may be rdflib.github.io.

[ghost]

Given that neither domain is under the control of RDFLib, oughtn't we remake it further anyway? E.g. using rdflib.dev (if we bet on that being maintained during the lifetime of this project).

I'm in favor of this, we should not use domains we don't own so rdflib.dev will be better another option may be rdflib.github.io.

Changed.

Note to reviewers - although there were only a couple of corrected typos in the original PR, the wholesale change away from rdflib.net has resulted in the number of changed files increasing from 2 to 7.

[sonarcloud]

Kudos, SonarCloud Quality Gate passed!   

0 Bugs
0 Vulnerabilities
0 Security Hotspots
0 Code Smells

No Coverage information
0.0% Duplication

[aucampia]

This PR includes the following changes with runtime impact:

• Change the binding for genid that gets set under some circumstances:

  rdflib/rdflib/graph.py

  Lines 2051 to 2053 in c906f64

  	self.bind(
  	"genid", "http://rdflib.net" + rdflib_skolem_genid, override=False
  	)

  I would consider this a breaking change as under some strange circumstances this could result in user code doing something different than before, but technically it was wrong before as it should match the prefixed used by BNode().skolemize() to have the desired effect, and it won't - so the current value is a bugged value to me anyway. I'm also not sure setting this in this place makes a lot of sense, I would say this should either always be set, or never. The current behavior for setting this seems somewhat arbitrary.

  My determination: Treat it as a non breaking change because the situations in which it would effect end user behavior is extremely arbitrary. In future we should re-consider what we are doing here though as I think it may be best to not do that, and either make genid part of our default namespace prefix set or not.

• Change the user-agent comment used in some contexts:

  rdflib/rdflib/parser.py

  Lines 185 to 187 in c906f64

  	headers = {
  	"User-agent": "rdflib-%s (http://rdflib.net/; eikeon@eikeon.com)" % __version__
  	}

  The user agent comment has no specific semantics associated with it in IETF RFC 7231 [ref] - however many systems do effectively assign some semantics to it. The comment in our user-agent however seems to not really follow the semantics that is assigned to it [ref].

  My determination: Treat it as a non breaking change as nothing mandates that semantics be assigned to the user-agent comment and in this case it is unclear why someone would assign semantics to it. Our user-agent comment anyway seems entirely arbitrary. In future I would recommend removing the user-agent comment entirely as it is unclear what benefit it serves.

• Change the default skolemization authority:

  rdflib/rdflib/term.py

  Lines 484 to 485 in 8cf7494

  	if authority is None:
  	authority = "https://rdflib.github.io/"

  I don't see this as breaking, as there is nothing anywhere that I can find that assigns any semantic meaning to the authority used for generating Skolem IRIs. I also see our current value as a bug as we are not in control of the authority we currently use, which seems to go against what is said in the RDF 1.1 abstract syntax spec.

  My determination: Treat it as a non breaking change.

I would like to know if anyone differs on any of these points: @RDFLib/core

[ghost]

I would like to know if anyone differs on any of these points: @RDFLib/core

I'm in agreement.

[aucampia]

LGTM, as discussed earlier I don't think we should treat this as a breaking change, but open to more input on the matter.

[aucampia]

@gjhiggins I also added an explicit test for this behavior:

rdflib/rdflib/graph.py

Lines 2047 to 2056 in 7c51027

	def graph(self, identifier=None, base=None):
	if identifier is None:
	from rdflib.term import _SKOLEM_DEFAULT_AUTHORITY, rdflib_skolem_genid
	self.bind(
	"genid",
	_SKOLEM_DEFAULT_AUTHORITY + rdflib_skolem_genid,
	override=False,
	)
	identifier = BNode().skolemize()

Test is here: https://github.com/gjhiggins/rdflib/blob/fix-for-issue1824/test/test_dataset/test_dataset.py#L239-L268

[edmondchuc]

I agree with your points @aucampia and I think using https://rdflib.github.io/ is a good idea.

[niklasl]

Very sound reasoning; I also agree!