On 10/14/2014 05:49 AM, Anne van Kesteren wrote:
On Tue, Oct 14, 2014 at 11:38 AM, Sam Ruby <ru...@intertwingly.net> wrote:
At the present time, all I can say is that the https://url.spec.whatwg.org/,
https://github.com/w3c/web-platform-tests/blob/master/url/, and
https://github.com/annevk/url are inconsistent.
I recommend not looking at annevk/url.
To illustrate, try pasting http://f:b/c into:
http://www.lookout.net/test/url/url-liveview.html
Relevant excerpt from that page:
var url = new URL(input, base);
urlHref.textContent = url.href;
And the results for http://f:b/c after applying urltestparser.js against
urltestdata.js is as follows:
{"input":"http://f:b/c","base":"http://example.org/foo/bar","scheme
":"","username":"","password":null,"host":"","port":"","path":"","query":"","fra
gment":"","href":"http://f:b/c","protocol":":","search":"","hash":""}
That seems correct. You hit "b" in the port state and that will return
failure (from memory, did not check).
How does this not match the specification?
Here's my original statement:
"The expected results are an object that returns the original href, but
empty values for all other properties. I don't see this behavior in the
spec: https://url.spec.whatwg.org/#url-parsing"
http://lists.w3.org/Archives/Public/public-whatwg-archive/2014Oct/0159.html
If you could be so kind as to point out what I am missing, I would
appreciate it.
I'll look further into why the results provided by Opera and
https://rubygems.org/gems/addressable don't appear to match RFC 3491.
Note that RFC 3491 is not a normative dependency for any of the algorithms.
RFC 3491 is a normative dependency for RFC 3490, Internationalizing
Domain Names in Applications (IDNA).
You said, "per IDNA those are ignored".
http://lists.w3.org/Archives/Public/public-whatwg-archive/2014Oct/0166.html
- Sam Ruby