On 2011-11-02 20:02, Anne van Kesteren wrote:
On Wed, 02 Nov 2011 11:59:14 -0700, Julian Reschke
<[email protected]> wrote:
On 2011-11-02 19:46, Anne van Kesteren wrote:
If you are confused about the terms, they are defined in HTML.
But they aren't linked, as far as I can tell. It would be good if they
were, and then we can review that text again.
They are in
http://www.whatwg.org/specs/web-apps/current-work/multipage/network.html#parsing-websocket-urls
so you can use that for review.
...
OK, I have tried to understand what's going on here, but I largely failed.
0) The algorithm "Resolve the url string" doesn't exist in the spec, nor
does it reference anything. If nobody has complained about this yet then
my conclusion is that nobody has read that section. Not surprisingly, as
it was added past last-call.
1) Anne explains that it's supposed to reference
<http://www.whatwg.org/specs/web-apps/current-work/multipage/urls.html#resolve-a-url>,
so that change should be made.
2) Assuming that change *will* be made...: the difference to what the
protocol spec says seems to be...
2a) it deals with IRIs (converting them to URIs),
2b) it specifies certain fixup operations for broken input, such as
converting \ to /, which, last time I checked, were controversial even
among browser implementers.
Reminder: this was a past-LC change. I think I'm not asking too much
when I'm asking for a precise explanation of what the rational for this
change was.
Best regards, Julian
