Received-SPF: pass (google.com: domain of abnf-discuss-bounces@ietf.org designates 2001:1890:126c::1:1e as permitted sender) client-ip=2001:1890:126c::1:1e;
Message-ID: <50E5B9DB.4030500@alum.mit.edu>
Date: Thu, 03 Jan 2013 12:03:23 -0500
From: Paul Kyzivat <pkyzivat@alum.mit.edu>
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.7;
	rv:17.0) Gecko/17.0 Thunderbird/17.0
MIME-Version: 1.0
To: abnf-discuss@ietf.org
References: <50E4A900.1070408@alum.mit.edu>
	<alpine.LSU.2.00.1301030943590.15409@hermes-1.csi.cam.ac.uk>
	<50E5B175.1030308@gmail.com>
In-Reply-To: <50E5B175.1030308@gmail.com>
Subject: Re: [abnf-discuss] How to indent ABNF figures in draft?
Precedence: list
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset="us-ascii"; Format="flowed"
Sender: abnf-discuss-bounces@ietf.org
Errors-To: abnf-discuss-bounces@ietf.org

On 1/3/13 11:27 AM, Dave Crocker wrote:
>
>
> On 1/3/2013 2:09 AM, Tony Finch wrote:
>> I suggest the following, which still requires rules to start on new lines
>> but removes sensitivity to horizontal white space.
>>
>>           rulelist       =  1*( *c-wsp rule ) *c-wsp
>>
>>           rule           =  rulename defined-as elements c-nl
>>
>>           c-wsp          =  WSP / c-nl
>
>
> If intent is to require indentation for a rule that wraps around to an
> additional line, I think you mean:
>
>              c-wsp          =  WSP / (c-nl WSP)
>
> (I'm completely ignoring the issue of starting a new rule indented.  We
> seem to be handling that adequately, already...)

I'm not sure what you mean by "handling that adequately"?

If you mean "we have some tools that handle indented ABNF", then maybe.

If you mean "RFC5234 unambiguously specifies how ABNF may be indented" 
then I disagree. We have an ABNF for ABNF that says there may be no 
whitespace at the beginning of a rule. And then there is some text, that 
may or may not be normative, that informally describes how ABNF may be 
indented.

And even if that text is considered normative, it isn't entirely clear 
what is and isn't valid. I *think* it says that the beginning of all 
rules must be indented the same amount, and continuation lines must be 
indented more. So that the file could be preprocessed by finding the 
first non-whitespace line, removing leading whitespace from that line, 
and then removing identical leading whitespace from all subsequent 
lines. Then process the result according to the ABNF for ABNF.

But that becomes messy when both tabs and spaces are used. In doing 
leading whitespace removal, what if some lines start with spaces and 
some start with a tab? E.g. what if I have:

<sp><sp><sp><sp>foo-stmt = "foo:" bar
<tab>"," baz

Is that valid? Should the <tab> be removed from the 2nd line? If so, 
does that mean the line starts without whitespace and so isn't part of 
the rule from the first line? Or should the <tab> be considered 
equivalent to some number of spaces? If it is considered equivalent to 8 
spaces, then the 2nd line is a continuation. If it is considered 
equivalent to four spaces then the 2nd line isn't a continuation.

IMO this isn't "adequate". :-)

	Thanks,
	Paul
_______________________________________________
abnf-discuss mailing list
abnf-discuss@ietf.org
https://www.ietf.org/mailman/listinfo/abnf-discuss