RE: Bug 12543 and URL parsing in general

Thanks Silvia-

Travis and I had started looking into what changes would be required, and this commit includes those plus a lot more that we hadn't even gotten to yet. I will go ahead and assign bug 12543 to you to resolve with this merge.

From: Silvia Pfeiffer [mailto:silviapfeiffer1@gmail.com]
Sent: Sunday, February 17, 2013 12:02 AM
To: Julian Reschke
Cc: Travis Leithead; Erika Doyle Navara; public-html@w3.org
Subject: Re: Bug 12543 and URL parsing in general

Please note that I am about to cherry-pick the following commit from the WHATWG:
https://github.com/w3c/html/commit/5eba22bd9dfec1295f6108e002d3a3a157287767

It is making use of the new URL spec.

It would be good to analyse what open issues around URL remain after this and continue trying to resolve them.

Thanks,
Silvia.
On Thu, Feb 7, 2013 at 4:35 AM, Julian Reschke <julian.reschke@gmx.de<mailto:julian.reschke@gmx.de>> wrote:
On 2013-02-06 18:20, Travis Leithead wrote:
From: Julian Reschke [mailto:julian.reschke@gmx.de<mailto:julian.reschke@gmx.de>]
On 2013-02-05 23:36, Erika Doyle Navara wrote:
Bug 12543 [1] requests that step 12 (replacing "\" with "/") of the
"resolve a URL" algorithm be removed, as Gecko behavior has shown that
it isn't needed for compatibility. The discussion in a related bug 14693
[2] also highlights other problems with step 12 and the algorithm in
general. It has been suggested that the entire section of URL parsing be
removed and instead defer to the URL spec [3].

Are there any objections to this course of action? If not, I'll take a
stab at removing the relevant sections and stage them in a branch before
committing to HTML5.1 and resolving 12543 as fixed.
...

Deferring to another spec that has the same problem doesn't resolve the
issue; it just moves it elsewhere.

While conceptually correct, it does matter to the HTML5 editors :)

In any case, it seems as if there was a rough proposal to keep the "\" handling since it is implemented in all but Gecko browsers. To this end, the URL parsing spec does include it.

Julian, in my reading between the lines, are you proposing we should remove the "\" handling from both HTML5 and the URL spec?

Well, we shouldn't pretend that by referring to the "URL" spec any of URL-related problems gets solved :-)-

Best regards, Julian