[Webkit-unassigned] [Bug 68089] WebKit doesn't parse "#" as delimiter for fragment identifier in data URIs

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=68089





--- Comment #23 from Robert O'Callahan <roc at ocallahan.org>  2011-09-20 13:20:06 PST ---
(In reply to comment #22)
> > I can't find that in the thread.
> 
> It was mentioned in Comment #10 and Comment #13 in this bug.  In the thread,
> it was mentioned in Boris Zbarsky's message on Sat, Sep 10, 2011 at 5:32 PM.

Thanks.

> If we'd had this discussion back in April, I would have been much more excited
> about digging in here and finishing the URL spec.  At the moment, I'd rather
> give the IRIbis working group some breathing room and a chance to succeed.
> Once they've either succeeded or failed, we should talk more about writing a
> high-quality specification for URL handling.

Fair enough, although the status quo is quite unsatisfactory.

Of the three options:
1) Current behavior of Firefox and Opera: allow data: URIs to have fragment references, parse everything after the first # as a fragment reference, and allow relative URIs consisting of fragment references to resolve relative to a data: URI.
2) Your suggestion in comment #18 applied consistently to SVG images: disallow fragment references in data: URIs, resolve all URIs relative to a data: URI to an invalid URI.
3) Your suggestion in comment #18, except with something magical happening for SVG (possibly only in an image context) (e.g., have SVG special-case fragment-only relative URIs to bypass normal relative URI resolution)
#2 would cripple data: URIs for SVG images. #3 seems unnecessarily complex, and limiting it to SVG would be difficult as we increase the integration of SVG with CSS and HTML. #1 seems best to me.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.