25130 2014-03-24 04:39:03 +0000 spec disallows fragments in both "absolute URL" and "relative URL" 2014-05-22 10:27:14 +0000 1 1 1 Unclassified WHATWG URL unspecified PC All RESOLVED NEEDSINFO http://url.spec.whatwg.org/#concept-absolute-url P2 normal Unsorted 1 mike annevk mike sideshowbarker+urlspec oldest_to_newest 102806 0 mike 2014-03-24 04:39:03 +0000 I suggest rewriting the definition of "absolute URL" in the spec to explicitly state that it can optionally contain a fragment: — An absolute URL must be a scheme, followed by ":", followed by either a scheme-relative URL, if scheme is a relative scheme, or scheme data otherwise, optionally followed by "?" and a query, *optionally followed by "#" and a fragment*. And similarly for "relative URL". Then change the definition of "URL" to just: — A URL must be written as either a relative URL or an absolute URL. That is, remove from there the part that says 'optionally followed by "#" and a fragment' (because it's moved instead to the definitions of "absolute URL" and "relative URL". Rationale: I want to directly reference just the definition of "absolute URL" from the spec, but if I do that now I end up with something that can't contain a fragment. (So right now I'd have to instead reference the spec by saying something like, "A URL that is an absolute URL.") 102817 1 annevk 2014-03-24 11:34:04 +0000 That would no longer match https://tools.ietf.org/html/rfc3986#section-4.3 Though the current definition for relative URL does not match relative-ref either so maybe it doesn't matter. Depends a bit on the dependencies... 102822 2 mike 2014-03-24 14:22:28 +0000 (In reply to Anne from comment #1) > That would no longer match https://tools.ietf.org/html/rfc3986#section-4.3 Ah, OK. I guess there much be some reason for it having been defined that way previously. > Though the current definition for relative URL does not match relative-ref > either so maybe it doesn't matter. Depends a bit on the dependencies... I guess for "absolute URL" we shouldn't violate the previous definition unless there's a good reason to. I can't say that I personally know of any really compelling reasons to change it. 106630 3 annevk 2014-05-22 10:27:14 +0000 I guess once there's a dependency that shows this is needed we can revisit this.