Re: rxvt-unicode-9.15nb1 and irssi-0.8.15nb4 vs netbsd-5 curses

["John D. Baker" <jdbaker%mylinuxisp.com@localhost>]

On Wed, 11 Apr 2012, Daniel Horecki wrote:

> "John D. Baker" <jdbaker%mylinuxisp.com@localhost> writes:

> > Is there any way to get 'irssi' to recognize the color capability of
> > the new 'urxvt' on netbsd-5 without changing the reported TERM type?
>
> This is weird, because other products like mc, emacs, ninja or even bash
> works with this TERM (after linking to ~/.termcap).
> In irssi alone you can do
>
> /set term_force_colors on
>
> and you'll get colors.

I was trying to avoid that since I occasionally do use a terminal
that is not color-capable but see below.

> You can include more files in termcap.db:
>
> cap_mkdb -f ~/.termcap ~/.termcap /usr/share/misc/termcap
>
> which will eliminate problem of missing terminals defs.

Yes, and I tried that.  It made no difference.  Also, I don't like
per-user duplication of system-wide data.  If it's not found in the
user's custom termcap database, consult the system-provided database.
In between, it'd be even better to consult a site-local database.  That
way updating is less of a pain.

At least with a text-only ~/.termcap file, the system-wide termcap file
and the user's ~/.termcap file are used together.

> Or you could do upgrade to netbsd-6. ;-)

I'm working towards that, but I've got a few systems that have quirks
under netbsd-6 and -current, so they're stuck at netbsd-5 for a little
longer.

What I ultimately discovered was the following:

The example termcap file defines an entry for "rxvt-unicode" which
lacks the colors (Co#) and color-pairs (pa#) capabilities.  It then
defines "rxvt-unicode-256color" with these capabilites and continuation
with the previous "rxvt-unicode" entry.

The system termcap file also defines "rxvt-unicode" lacking color and
color-pair capabilities, but rxvt-unicode-9.06 was seen as color-capable
anyway when named "rxvt-unicode".  The system "rxvt" entry defines only
8 colors and 64 color pairs.

With v9.15, ":tc=rxvt-unicode:" causes the color capabilities to be
ignored, regardless of which "rxvt-unicode" entry is used for the con-
tinuation even when the color and color-pair capabilities were set to
match those of the system "rxvt" entry .

If I make the "rxvt-unicode-256color" entry ("rxvt-256color" on OpenBSD)
use just the "rxvt" entry as its continuation, v9.15 is seen as color-
capable.  I haven't tested whether ":Co#256:pa#32767:" is properly over-
riding "rxvt"'s ":Co#8:pa#64:" or vice versa.

--
|/"\ John D. Baker, KN5UKS               NetBSD     Darwin/MacOS X
|\ / jdbaker[snail]mylinuxisp[flyspeck]com    OpenBSD            FreeBSD
| X  No HTML/proprietary data in email.   BSD just sits there and works!
|/ \ GPGkeyID:  D703 4A7E 479F 63F8 D3F4  BD99 9572 8F23 E4AD 1645