lmc@denelcor.UUCP (Lyle McElhaney) (07/06/84)
I have been interested getting a definitive termcap set up for the vt100
terminal (ignore all variations for the moment; just 80 col vanilla, no
Advanced Video vt100). I've cadged a number of possibilities from the net
and have written a program to compare them. I'm anticipating a lot of
termcap/ terminfo work in the near future, and so I'm going to present my
vt100 termcap here, with comments, as a sort of sanity check and to get
some questions answered. Please reply by mail; I'll summarize to the net.
Generally, this is the longest vt100 termcap I've seen. This is so because
I've put everything in it. In particular, the status-line implementation
is included, along with variables for defining a set of line-drawing
characters (I've used these in a version of the Rand Editor).
Anyway, here it is. Criticisms invited.
dj|vt100|vt100 with status line on
size: 903 Long enough.
--------------------
Bj=v *
Bl=m *
Br=j * Line-drawing stuff proposed by
Cj=n * negaard@aerospace. Gs & Ge intro
Ge=1^O * and end sequences, others are (in
Gs=1^N * order): Bottom join, bottom left
Hl=q * corner,bottom right, cross, horiz
Lj=t * line, left join, right join, top
Rj=u * join,top left, top right, vertical
Tj=w * line, and other unique character.
Tl=l * This implementation assumes the
Tr=k * init setup below.
Tr=k *
Vl=x *
Xc=\140 *
ae=1^O
as=1^N start alternate char set. See is below
bl=1^G Shouldn't this default to ^G?
bs
cd=45\E[J
ce=2\E[K
cl=45\E[2J (non status-line)
cl=50\E7\E[23;80H\E[1J\E8 Erase lines 1-23.
cm=1\E[%i%d;%dH
co#80
cr=1^M
cs=1\E[%i%d;%dr
do=30^J I will use the shorter form whenever
possible. The long delay is used
whereever a scroll is possible.
ds= Can't disable status line.
es= I've seen this as enable status line
and as escapes safe in status line.
What's it to be?
fs=1\E8
ho=1\E[H Description says home cursor (if no
CUP). Is it not used otherwise?
hs
if=/usr/lib/tabset/vt100
is=\E>\E[?1;3;4l\E[?7;8h\E[1;24r\E( (overridden by:)
B\E)0
is=\E>\E[?1;3;4l\E[?7;8h\E[1;23r\E( init: set scroll region (set line 24
B\E)0 aside); select cursor key mode, jump
scrolling, 80 col, wraparound on, auto
repeat on, USASCII character set, and
line-drawing chars in shifted set.
Set scroll must follow option sets.
k1=\EOP
k2=\EOQ
k3=\EOR
k4=\EOS
ks=1\E= Not sure what "keypad transmit" is;
ke=1\E> used here to set keypad for function
keys, rather than numeric keypad.
kb=^H Is this default?
kd=\E[B A lot of people say kd=\E0B, but
kl=\E[D thats only in applications mode,
kr=\E[C which I reset and leave reset.
ku=\E[A
l1=PF1
l2=PF2
l3=PF3
l4=PF4
le=1^H
li#24 (overridden by:)
li#23 23 lines, 24th is status line.
ll=1\E[24H (overridden by:)
ll=1\E[23H
mb=1\E[5m
md=1\E[1m
me=1\E[m
mr=1\E[7m
ms
nd=1\E[C
nl=30^J
pt
rc=1\E8
rf=/usr/lib/tabset/vt100
rs=\E>\E[?1;3;4l\E[?7;8h\E[1;23r\E( Same reset as init.
B\E)0
rs=\E>\E[?1;3;4l\E[?7;8h\E[1;24r\E(
B\E)0
sc=1\E7
se=1\E[m
sf=30\ED
so=1\E[7m
sr=30\EM
ta=1^I is this defaulted?
te=1\E> Used by the Rand editor. Would ks/ke
ti=1\E= a more appropriate choice?
ts=4\E7\E[24;H\E[1K
ue=1\E[m
up=1\E[A
us=1\E[4m
vt#3 Where does this come from?
ws#80
Padding - I assumed 9600 baud, and used the advertized padding. Its quite
different from the padding I've seen in the other termcaps; Is there some
empirical evidence that it should be different?
Note lack of am and xn; a fellow from NADC wrote this about that:
The reason that VT100s and their clones do not work with automargins
enabled is that, regardless of whether :xn: is specified or not, the
cursor is still in column 80 when the software has presumed it has
wrapped. Thus, the software may decide that the best way to get to
column 1 a couple of lines down is to use line feeds rather than cursor
motion; this puts the VT100 cursor several lines down, but still in
column 80, and it has by then forgotten that it should be in column 1.
Here at NADC we religiously shun auto margins on all VT100 like
terminals. Instead, we let the :is=...: string set up the terminal with
wraparound enabled (for general use), but do not specify :am: in
termcap. This way, the software makes no incorrect assumptions.
However, I still have a problem. I can send the following sequence to my
vt100:
start with y in first and last cols, row 6; x in first and last
cols of other rows;
\E[6;80Hx\E[7;80Hy\E[6;1Hx (ignoring padding characters)
(put x in last col, row 6; put y in last col row 7; put x in first
col, row 6)
This results in all x's in col 80, and y's in col 7 and 8 of col 1. Do I
have a bad crt, or can others do this too?
Oh, well, enough allready. Send me comments.
--
Lyle McElhaney
(hao,brl-bmd,nbires,csu-cs,scgvaxd)!denelcor!lmcmark@cbhydra.uucp (07/08/84)
I'm enclosing my current terminfo description for the vt100.
I notice two significant problems with the one previously posted.
(1) Your method of dealing with graphics characters is different
from what AT&T is going with, (2) you put the status line on the
bottom rather than on the top. (Most people, given a choice, prefer
it on the top, also it doesn't get clobbered when you clear the screen.)
I have no quarrel with your observations about automatic margins, but
neither method seems to handle the vt100 right (outputting a char in
col 80 sets a flag in the terminal, outputting a char elsewhere seems to
be the only way to clear that flag.)
Explanation of graphics characters. The vt100 is the definition, but only
for those chars that are line drawing in nature. The acsc (alternate char
set chars) lists pairs of chars, for aAbBcC, to get standard a you output A,
to get b output B, etc. enacs must be output before sending anything out
in the acs (once) to get the terminal in the mode. smacs/rmacs turn on
and off the alternate char set (the one with graphics in it.)
# Vanilla vt100 definitions.
vt100|vt100-am|dec vt100,
cr=^M, cud1=^J, ind=^J, bel=^G, cols#80, lines#24, it#8,
clear=\E[H\E[2J$<50>, cub1=^H, am, cup=\E[%i%p1%d;%p2%dH$<5>,
cuf1=\E[C$<2>, cuu1=\E[A$<2>, el=\E[K$<3>, ed=\E[J$<50>,
cud=\E[%p1%dB, cuu=\E[%p1%dA, cub=\E[%p1%dD, cuf=\E[%p1%dC,
smso=\E[7m$<2>, rmso=\E[m$<2>, smul=\E[4m$<2>, rmul=\E[m$<2>,
bold=\E[1m$<2>, rev=\E[7m$<2>, blink=\E[5m$<2>, sgr0=\E[m$<2>,
sgr=\E[%?%p1%t;7%;%?%p2%t;4%;%?%p3%t;7%;%?%p4%t;5%;%?%p6%t;1%;m,
rs2=\E>\E[?3l\E[?4l\E[?5l\E[?7h\E[?8h, smkx=\E[?1h\E=, rmkx=\E[?1l\E>,
tbc=\E[3g, hts=\EH, home=\E[H,
enacs=\E(0, smacs=^O, rmacs=^N,
acsc=aaffggjjkkllmmnnooppqqrrssttuuvvwwxxyyzz||}},
kcuu1=\EOA, kcud1=\EOB, kcuf1=\EOC, kcub1=\EOD, kbs=^H,
kf1=\EOP, kf2=\EOQ, kf3=\EOR, kf4=\EOS, ht=^I, ri=\EM$<5>,
vt#3, xenl, xon, sc=\E7, rc=\E8, csr=\E[%i%p1%d;%p2%dr,
vt100-nam|vt100 w/no am,
am@, xenl@, use=vt100-am,
# Ordinary vt100 in 132 column ("wide") mode.
vt100-w|vt100-w-am|dec vt100 132 cols (w/advanced video),
cols#132, lines#24, rs2=\E>\E[?3h\E[?4l\E[?5l\E[?8h, use=vt100-am,
vt100-w-nam|vt100-nam-w|dec vt100 132 cols (w/advanced video),
cols#132, lines#24, rs2=\E>\E[?3h\E[?4l\E[?5l\E[?8h, vt@, use=vt100-nam,
# vt100's with no advanced video.
vt100-nav|vt100 without advanced video option,
sgr@, smso=\E[7m , rmso=\E[m , xmc#1,
smul@, rmul@, bold@, rev@, blink@, sgr0@, use=vt100,
vt100-nav-w|vt100-w-nav|dec vt100 132 cols 14 lines (no advanced video option),
lines#14, cols#132, use=vt100-nav,
# vt100's with one of the 24 lines used as a status line.
# We put the status line on the top.
vt100-s|vt100-s-top|vt100-top-s|vt100 for use with sysline,
lines#23, is2=\E7\E[2;24r\E8, clear=\E[2;1H\E[J$<50>, home=\E[2;1H,
dsl=\E7\E[1;24r\E8,
cup=\E[%i%p1%{1}%+%d;%p2%dH$<5>, csr=\E[%i%i%p1%d;%p2%dr,
hs, eslok, tsl=\E7\E[1;%p1%dH\E[1K, fsl=\E8, use=vt100-am,
# Status line at bottom. Clearing the screen will clobber status line.
vt100-s-bot|vt100-bot-s|vt100 for use with sysline,
lines#23, is2=\E[1;23r\E[23;1H, dsl=\E7\E[1;24r\E8,
hs, eslok, tsl=\E7\E[24;%p1%dH\E[1K, fsl=\E8, use=vt100-am,gwyn@brl-smoke.ARPA (Doug Gwyn ) (08/17/86)
Elsie!ado has correctly pointed out to me that my VT100 termcap entry should have the "ch" capability removed, since it doesn't work right for a motion to the first column. I hadn't noticed this before since our software would always use \r to do that. ch=\r\E[%dC WRONG; remove this capability