W3C home > Mailing lists > Public > ietf-http-wg@w3.org > April to June 2008

Re: ABNF switch: list rules

From: Frank Ellermann <nobody@xyzzy.claranet.de>
Date: Sat, 24 May 2008 18:02:46 +0200
To: ietf-http-wg@w3.org
Message-ID: <g19eqd$luu$1@ger.gmane.org>

Julian Reschke wrote:

> Is there consensus for this?

No.  #-LWS has to be killed or obsoleted.  ASCII art by commas
is only ridiculous, but "apparently empty lines" are dangerous.
Example, #x permits...

 ,,,   ,   ,   ,,,  <CRLF>
 ,  ,  ,   ,  ,     <CRLF>
 ,,,   ,   ,  , ,,, <CRLF>
 ,  ,  ,   ,  ,   , <CRLF>
 ,,,    ,,,    ,,,  <CRLF>

...and that's IMNSHO ugly.  But it also permits...

 ,,,,,,,,,<CRLF>
<SP>      <CRLF>
<HTAB>    <CRLF>
 ,,,,,,,,,<CRLF>
 
...and that is dangerous when applications don't get the subtle
difference between "really empty" <CRLF> and "apparently empty"
<SP><CRLF> or even <HTAB><CRLF> lines.

The ABNF has to be very obvious what is permitted, and what is
obsoleted, without misguided attempts to hide these horrors in
any #-constructs.  

The one and only point of the "ABNF-ication" is precisely this
#-LWS issue, otherwise nothing was wrong with the old 2068-BNF.

 Frank
Received on Saturday, 24 May 2008 16:13:55 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 06:50:48 GMT