W3C home > Mailing lists > Public > public-html@w3.org > May 2008

Re: Parsing: Equals sign in unquoted attributes

From: Henri Sivonen <hsivonen@iki.fi>
Date: Fri, 30 May 2008 21:23:56 +0300
Cc: "public-html@w3.org" <public-html@w3.org>
Message-Id: <A6F34F6D-125E-4E74-9515-01E45180E7FD@iki.fi>
To: Simon Pieters <simonp@opera.com>

On May 30, 2008, at 13:26, Simon Pieters wrote:

> This seems to result in false positives for:
>
>   <a href=foo?bar=baz>
>
> ...which is pretty common.


It seems to me (from Philip's data) that making = always silently  
allowed in unquoted attribute values is a bad idea as it would conceal  
the kind of mistakes that would be helpful to reveal to the author  
(attributes running together mostly with the previous one taking loose  
enough strings for the result not to violate datatype constraints). If  
we wanted to allow = in unquoted URIs, we should make the tokenizer  
check if the current attribute is src or href. Do we really want to go  
there?

If we want to reduce author hoop jumping by making the conformance  
definition closer to practice, I suggest we start by allowing <img  
border=0> and <script language=javascript>.

-- 
Henri Sivonen
hsivonen@iki.fi
http://hsivonen.iki.fi/
Received on Friday, 30 May 2008 18:24:40 UTC

This archive was generated by hypermail 2.3.1 : Monday, 29 September 2014 09:38:55 UTC