W3C home > Mailing lists > Public > www-validator-css@w3.org > May 2007

Re: Lexical Error parsing prologue

From: Jukka K. Korpela <jkorpela@cs.tut.fi>
Date: Tue, 1 May 2007 21:27:04 +0300 (EEST)
To: Benjamin Tomos Lewis <ben.tc1415@gmail.com>
cc: www-validator-css@w3.org
Message-ID: <Pine.GSO.4.64.0705012116390.3843@hopeatilhi.cs.tut.fi>

On Tue, 1 May 2007, Benjamin Tomos Lewis wrote:

> I'm at a bit of a loss as to why this is occurring. When I validate the
> CSS of may site (http://benl.co.uk/) I always get the error:
>
> "Lexical error at line 1, column 5. Encountered: "I" (73), after : "$""

It looks strange indeed - as if the validator saw a page that has 
something like

/* $Id: node.css,v 1.2 2006/09/05 03:50:56 unconed Exp $ */

at the start, as in your CSS files.

The response headers from http://benl.co.uk/ contain the header
Vary: Accept-Encoding,User-Agent
which _suggests_ that the server _may_ send different responses to 
different user agents. This is a bit far-fetched, though. (I tested with a 
response containing an empty User-Agent value, and the response content is 
normal, i.e. the same as normal browsers get.)

> This is implying there is something wrong with the XML prologue - I
> believe - however the markup validator is not complaining, and a quick
> read of both the generated page and it's source in multiple editors
> reveals that the prologue does indeed read:
>
> "<?xml version="1.0" encoding="utf-8"?>" - no $s at all.

I wonder whether the prologue is useful in practice. It throws IE 6 into 
Quirks Mode, and it isn't really needed, at least in this case since
UTF-8 is explicitly specified in the Content-Type header.

-- 
Jukka "Yucca" Korpela, http://www.cs.tut.fi/~jkorpela/
Received on Tuesday, 1 May 2007 18:27:12 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 27 June 2012 00:14:19 GMT