Re: [patch 1/4] do not force suid color

[Jan Engelhardt]

On Saturday 2008-06-21 19:37, Jim Meyering wrote:
>Jan Engelhardt <address@hidden> wrote:
>> By default, ls highlights setuid/setgid/etc. files with a color, but
>> there is no way to restore the old (coreutils 5.x?) behavior, i.e.
>> that the setuid file gets the same color as it would when not having
>> suid.[...]
>
>Thanks for explaining the problem and taking the time to prepare those
>patches.  However, upstream ls has never relied on /etc/DIR_COLORS,
>and I'm confident that many people would complain if --color stopped
>colorizing the output of ls.[...]
>And dircolors cannot yet emit such settings:
>
>    $ echo SETGID|dircolors -
>    dircolors: -:1: invalid line;  missing second token
>
>If you (or anyone else) are interested in fixing that (maybe as easy
>as removing the test and ensuring that downstream code can handle NULL
>pointers or empty strings) and adding tests and documentation, that
>would accelerate the process.  Otherwise, I'll do it eventually.
>
>>From 9dde28f1ee2ccd8bbfff1f2ec893e446abc02175 Mon Sep 17 00:00:00 2001
>From: Jim Meyering <address@hidden>
>Date: Sat, 21 Jun 2008 19:05:02 +0200
>Subject: [PATCH] ls: ignore a normally-colorizable attribute if its string has 
>length 0
>
>* ls.c (print_color_indicator): Skip each test for an attribute
>like C_SETUID, C_SETGID, C_STICKY_OTHER_WRITABLE, etc. if the
>corresponding string has 0 length.
>Prompted by a report from Jan Engelhardt.

This patch has not yet made it in, it seems.