[systemd-devel] vt220 default for serial console still relevant?

[Lennart Poettering]

On Mo, 20.07.20 15:03, smcv at collabora.com (smcv at collabora.com) wrote:

> On Sat, 11 Jul 2020 at 21:04:18 +0200, Lennart Poettering wrote:
> > widely-supported TERM value
>
> For a value of TERM to work (at all), it must be something that is reliably
> present in the terminfo/termcap data visible to TUI programs.
>
> Debian divides terminfo into two categories. It wouldn't surprise me
> if other distributions have a similar split, with the line drawn in a
> different place.
>
> ncurses-base is marked as Essential (so every Debian system is required
> to have it) and supports a relatively small number of common terminals:
> https://packages.debian.org/sid/all/ncurses-base/filelist
> (In particular this includes linux, vt220, xterm, xterm-256color, ansi,
> screen and tmux.)

So what I find interesting: it lists vt300 too, but vt300 is a series
of terminals, and the first model in the series is apparently vt320,
followed by vt330, and vt340, where the latter added color. Now, if
the whole series is grouped under one moniker "vt300" (and there's no
model named like that), it sounds like one could convince people to
enable color for that...

I think what's interesting to ask is not whether some terminal type
such as vt100 or vt220 actually do color, but whether they gracefully
handle color sequences. i.e. if I send color sequences to the original
vt100, what happens? does it spew them as weird characters onto the
screen, or does it simply ignore it? What about the vt220? Because, if
they all just take it and ignore it, there's a good case to be made to
convince "ls" and such tools to just enable color for them, since it
has no effect if connected to an actual vt100/vt220 terminal, but
works wonders for everyone else...

Lennart

--
Lennart Poettering, Berlin