[rfc-i] tabs [was: sourcecode indentation]

tony.l.hansen at gmail.com (Tony Hansen) Tue, 16 February 2016 14:30 UTC

From: tony.l.hansen at gmail.com (Tony Hansen)
Date: Tue, 16 Feb 2016 09:30:02 -0500
Subject: [rfc-i] tabs [was: sourcecode indentation]
In-Reply-To: <F5B55E50-D7AA-427A-932C-37D4872F9BD4@cisco.com>
References: <B7E2CCFD-F56B-4749-B8D2-C3F4CFF3EF5F@cisco.com> <56C2B0E8.7070900@it.aoyama.ac.jp> <F5B55E50-D7AA-427A-932C-37D4872F9BD4@cisco.com>
Message-ID: <4998B3A5-750C-4095-8FDD-3168FE6CABCA@gmail.com>

> On Feb 16, 2016, at 1:14 AM, Joe Hildebrand (jhildebr) <jhildebr at cisco.com> wrote:
> 
> I think the current situation is pretty close to 2119's "SHOULD NOT".  You SHOULD NOT put a tab character in <sourcecode>, unless it is absolutely required in the target language.  To deal with this, we could add a tabs="true" attribute to <sourcecode>, and give an error if there's an unintended tab character.

A light tweak would be tabs="8" if tabs are present and to be allowed, and error if tabs is not defined.

> Alternately, we could just say that tab-encrusted Makefile's aren't supported in an RFC, and always throw an error on tabs.  Tabs can be worked around in GNU make (https://www.gnu.org/software/make/manual/html_node/Special-Variables.html):
> 
> .RECIPEPREFIX = >
> all:
>> @echo Hello, world

Interesting.

- Tony