Re: What is the best way to handle LS_COLORS ?

[Thomas Dickey]

On Fri, Mar 27, 2020 at 02:48:20PM -0500, Bryan Christ wrote:
> I need advice as to handle LS_COLORS.  IMO, it's an absolute nightmare.
> Instead of using
> SGR 38:2 and 48:2 variants (which allow you know what the desired color is)
> it use SGR 38:5 and 48:5 which indicate a specific color number.  The

That depends on the particular system (the upstream source in coreutils
doesn't have any of that -- though it _does_ say that "vt100" does ANSI
colors).

On the other hand, coreutils has some documentation.

The packagers as a general rule don't bother to document anything
past checkin-comments.

> problem is, there's no real way to know what the color value is supposed to
> be.  It might be possible to query the color values with color_content()
> but as was discussed here...
> 
> https://lists.gnu.org/archive/html/bug-ncurses/2020-02/msg00001.html
> 
> ...it's not guaranteed to work.  It seems the fundamental problem is that
> LS_COLORS takes advantage of information known only to the terminal and
> whoever specified the color values.  Am I missing something here?

I wouldn't give it that much credit.  It's a hard-coded set of color
values, which happens to "work" for a default configuration.

Since it's hard-coded (and essentially copy/paste scripting),
one can just assume that the color palette is xterm's default palette.

-- 
Thomas E. Dickey <address@hidden>
https://invisible-island.net
ftp://ftp.invisible-island.net

signature.asc
Description: PGP signature