OpenSolaris_b135/cmd/terminfo/README
#
# CDDL HEADER START
#
# The contents of this file are subject to the terms of the
# Common Development and Distribution License, Version 1.0 only
# (the "License"). You may not use this file except in compliance
# with the License.
#
# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
# or http://www.opensolaris.org/os/licensing.
# See the License for the specific language governing permissions
# and limitations under the License.
#
# When distributing Covered Code, include this CDDL HEADER in each
# file and include the License file at usr/src/OPENSOLARIS.LICENSE.
# If applicable, add the following below this CDDL HEADER, with the
# fields enclosed by brackets "[]" replaced with your own identifying
# information: Portions Copyright [yyyy] [name of copyright owner]
#
# CDDL HEADER END
#
#ident "%Z%%M% %I% %E% SMI" SVr4.0 1.6
# From: SVr4.0 terminfo:README 1.6
1 Within the curses component, exists other conversion tools which are
much more robust than those described below. They are called infocmp,
and captoinfo. The cvt files are provided here only for those possible
cases where a user has the terminfo component without the libcurses
component.
The captoinfo and infocmp utilities cannot be included here, as they
require the user to have libcurses. Although we know of no instance
when a user would have one and not the other, we have provided the
cvt files (described below) for those limited cases.
2 The files in this directory with the .ti suffix are terminfo sources.
They should be compiled (separately or by catting them together into
terminfo.src) with tic, placing the results in /usr/lib/terminfo.
Please send any updates to AT&T Bell Laboratories UNIX support,
via UNIX mail to attunix!terminfo.
3 The cvt files are useful tools for converting termcap to terminfo.
They are not 100% accurate, but do most of the conversion for you.
cvt.ex is an ex script to convert a termcap description into a
terminfo description. Note that it will not convert padding
specifications, so they must be done by hand. Note also that typical
termcap entries do not give as much information as terminfo, so the
resulting terminfo entry is often incomplete (e.g. won't tell you the
terminal uses xon/xoff handshaking, or what extra function keys send).
You are urged to read the list of terminfo capabilities and augment your
terminfo descriptions accordingly.
The cvt.h file is useful for a quick hack at converting termcap programs
which use uppercase 2 letter names for capabilities to use terminfo.
Since tget* are provided anyway, this is of questionable value unless
your program barely fits on a pdp-11.
The cvt.sed script is useful for actually editing the source of the same
class of programs. It requires a sed that understands \< and \>, the
mod is trivial to make if you look at the corresponding code in ex or
grep.
3 There are other incompatibilities at the user level between termcap and
terminfo. A program which creates a termcap description and then
passes it to tgetent (e.g. vi used to do this if no TERM was set) or
which puts such a description in the environment for a child cannot
possibly work, since terminfo puts the parser into the compiler, not
the user program. If you want to give a child a smaller window, set
up the LINES and COLUMNS environment variables or implement the JWINSIZE
ioctl.
4 If you want to edit your own personal terminfo descriptions (and are not
a super user on your system) the method is different. Set
TERMINFO=$HOME/term (or wherever you put the compiled tree) in your
environment, then compile your source with tic. Tic and user programs
will check in $TERMINFO before looking in /usr/lib/terminfo/*/*
5 Philosophy in adding new terminfo capabilities:
Capabilities were cheap in termcap, since no code supported them
and they need only be documented. In terminfo, they add size to
the structure and the binaries, so don't add them in mass quantities.
Add a capability only if there is an application that wants to use it.
Lots of terminals have a half duplex forms editing mode, but no UNIX
applications use it, so we don't include it.
Before you add a capability, try to hold off until there are at least
2 or 3 different terminals implementing similar features. That way,
you can get a better idea of the general model that the capability
should have, rather than coming up with something that only works
on one kind of terminal. For example, the status line capabilities
were added after we had seen the h19, the tvi950, and the vt100 run
sysline. The original program, called h19sys, only worked on an h19
and addressed the cursor to line 25. This model doesn't fit other
terminals with a status line.
Note that capabilities must be added at the end of ../screen/caps.
Furthermore, if you add a private capability, you should check with
someone to make sure your capability goes into the master file,
otherwise someone else will add a different capability and
compatibility between two systems is destroyed. There must be one
master set of capabilities. This list is maintained at AT&T UNIX
Development. Comments should be sent to attunix!terminfo.
6 Current murky areas include:
Color - there is demand for colors but it isn't clear what to do yet.
Some terminals support only 2 or 4 or 8 or 16 colors, others have a
palette of some huge selection. What are the standard colors? How
does graphics fit into this (terminfo is alphanumeric oriented?)
Curses can have another 16 bits added, or some routine set to decide
which 9 attribute bits have meaning in any given program. An
alternative is that if you just want color alphanumerics for a simple
application, e.g. highlighting certain fields, decide how you would
want your application to behave on a B/W terminal (e.g. a vt100),
using reverse for one thing, blinking for another, bold for another,
invisible for another, etc.
(Invis may be useful for colored fields with no information in them.)
Then make a terminfo entry with blink=xxx, bold=yyy, etc, where xxx
and yyy are sequences to go into the colors you really want. This way
your application also works on B/W terminals.
Graphics: Giles Billingsley at Berkeley did something called MFBCAP
once, it was like termcap but 3 times as big and handled graphics.
I don't think it was ever finished. I don't know how to do graphics
in curses, one might add it to terminfo at very high cost.
Input: things that send escape sequences to your program to be decoded
are a hard issue. You have to somehow deal with typeahead and with
terminals that can't do it. This includes "request cursor position",
for which a better solution is to immediately address the cursor to
a known position. (Curses also has filter mode that won't assume
the line but will assume the column.) Mice also fall into this
category. Scanf style strings (tparm is printf style) might be able
to decode these sequences, but I have no experience with them.
Alternate character set: the vt100 set seems to be becoming a defacto
standard, although it doesn't do much. I almost standardized on the
Teletype 5410, which was a nice superset of the vt100, but then Teletype
updated the 5410 to make it a vt100 duplicate, so now all I've put in
are the vt100 line drawing characters. HP has a more complete set,
but it has some really weird things in it and the mappings are
nonstandard.
Any extension should be able to handle both kinds of terminals, and
handle common programs without assuming an HP (or even a vt100).
------------------------------------
7 Additional modules:
ckout shell script, analyzes file errs for diagnostics
and displays number of entries built
Doc.sed sed script to be run on ti files.
prints documentation of ti files.