30316 2018-11-07 19:36:54 +0000 [FO31] percent encoding rules in F&O for fn:resolve-uri are opposite to the same rules in XP31 2018-11-07 19:36:54 +0000 1 1 1 Unclassified XPath / XQuery / XSLT Functions and Operators 3.1 Recommendation PC Windows NT NEW P2 major --- 1 abel.online mike public-qt-comments oldest_to_newest 129382 0 abel.online 2018-11-07 19:36:54 +0000 In F&O 3.1 we say "No percent encoding takes place" under the section of fn:resolve-uri. In the XPath 3.1 spec, we say under "2.4.6 Resolving a Relative URI Reference": "Any process that attempts to resolve URI against a base URI, or to dereference the URI, may apply percent-encoding or decoding as defined in the relevant RFCs. " Apart from the type ("resolve *a* URI"), this rule appears to be opposite to the one under fn:resolve-uri. Even more so, because sections that involve resolving URIs often point to either location. Furthermore: test-case 'fn-resolve-uri-31' explicitly tests that no percent decoding should take place. Both rules, same wording, were also present in XP30 and FO30. The question now is: which rule has prevalence? May processors apply percent-de-/encoding? From relevant discussions I think the answer should be a resounding "NO", but since both specs have REC status, I don't know what the policy is here. Disallowing it could break existing processors. Allowing it wouldn't break anything, because of the "MAY". Since resolving URIs is essentially everywhere, I marked the bug "major".