[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 #14 from Julian Reschke <julian.reschke at gmx.de>  2011-09-20 00:48:03 PST ---
(In reply to comment #10)
> Ok.  I've caught up on my reading.  I don't see a big reason to change behavior here.  The dream of having all URIs parse the same way is just a dream.  Some common URI schemes, such as javascript URIs, can't be done this way.  There are also obscure ones, like ed2k, that can't work this way either.
> 
> The main reason I see to change our parsing of data URLs is in connection with fragment references in SVG.  I'm not sure what the best solution there is, but using the RFC 3986 regexp on data URLs doesn't seem like the best approach, especially give that doing so has resulted in Firefox getting compat complaints.

Sounds like "(b) and (c) are weird as well, so we can't make (a) less weirder". 

ed2k, as you say, is obscure. Does it affect browsers?

javascript is indeed interesting. I still think these aren't proper identifiers at all, and only live in browsers, so browser people may want to reserve the scheme name, and just specify what they do with them as something being in use where URIs/IRIs appear, but not really belonging to this class of identifiers.

What makes data different from the other schemes you mentioned is:

- browser implementations differ (as far as I can tell)

- they are produced/consumed in other places (per spec)

- they actually denote a representation to which a fragment can be applied (I don't think this is the case for javascript or ed2k)

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