Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!linus!philabs!cmcl2!harvard!seismo!umcp-cs!chris
From: chris@umcp-cs.UUCP (Chris Torek)
Newsgroups: net.unix-wizards,net.unix,net.info-terms
Subject: Re: terminfo, termcap, etc
Message-ID: <1624@umcp-cs.UUCP>
Date: Wed, 21-May-86 02:13:27 EDT
Article-I.D.: umcp-cs.1624
Posted: Wed May 21 02:13:27 1986
Date-Received: Sat, 24-May-86 02:14:58 EDT
References: <1135@bunker.UUCP> <155@molihp.UUCP> <2774@pegasus.UUCP> <299@enmasse.UUCP> <162@njitcccc.UUCP>
Reply-To: chris@maryland.UUCP (Chris Torek)
Distribution: net
Organization: University of Maryland, Dept. of Computer Sci.
Lines: 83
Xref: linus net.unix-wizards:15103 net.unix:7244 net.info-terms:513

In article <162@njitcccc.UUCP> ken@njitcccc.UUCP (Kenneth Ng) writes:
>In article <299@enmasse.UUCP>, keith@enmasse.UUCP (Keith Crews) writes:
>>While we are on the subject of terminfo, does anyone know of a
>>documented list of terminfo entries that are sufficient to allow
>>curses to work properly, if not optimally, in all cases?  [KC]

>I have found that the minimal requirements are (I think): col
>and row numbers, clear screen, cursor addressing, and indexing.
>By indexing Un*x means the control sequence to scroll the screen
>up a line.  [KN]

Actually, at least under 4BSD, vi requires:

	1) cursor motion, including EITHER:
		a) up and left and absolute, OR
		b) up, down, left, and right
	2) clear screen
	3) screen size (`co#' and `li#')

4BSD vi assumes a line feed from the last line will force a scroll.
I once worked with a terminal emulator for which this was not true;
it was rather painful.  (Adding `ns', `terminal is a CRT but cannot
scroll', makes vi use open mode: workable but unpleasant.)

4BSD curses may have a different set of minimums; at any rate I
remember *something* needed a clear-to-end-of-line as well.

Also, some attributes may force others to be required.  For example,
`db' implies a need for `cd=', though vi will iterate `ce=' if it
must.  (It may even use blanks if necessary; I have never tried it.)
Moreover, some attributes (e.g., `am') will cause what look like
random trouble if they are specified incorrectly.  If you want
instead to know which capabilites are *examined*, well, again,
that depends on the program.

My own termcap Emacs screen driver has similar (though not quite
identical) requirements as vi; but it currently *uses* all of these:

Strings (`cl='):
	al bc cd ce cl ch cm cr cv dc dl dm ds ed ei ho ic im ip
	pc ll nd nl se so ta te ti ue up us vb ve vs

Numerics (`co#'):
	co li sg tw ug

Flags (`am'):
	MT am bs db hz in km mi ms nc nn rn ul xn xs xt

(Alas, it is missing the new AL, DL, etc. entries; nor will it use
`cs='.  `rn', `nn', `tw#', and `ds=' are locals.)

I just found vi's strings and flags too:

	al bc bt cd ce cl cm cr cs dc dl dm do ed ei k0 k1 k2 k3 k4
	k5 k6 k7 k8 k9 ho ic im ip kd ke kh kl kr ks ku ll nd nl pc
	rc sc se sf so sr ta te ti up vb vs ve AL DL UP DO LE RI 

and

	am bs da db eo hc hz in mi nc ns os ul xb xn xt xx 

respectively.

>But of course the manual does not say what the minimal requirements
>are, I guess they just assume you can read minds.

I seem to recall seeing a list of requirements somewhere.  Perhaps
it was in a bit of source code.  The real problem, though, is that
the `minimal' set depends upon the program, and could change between
system releases; and writing a complete description is always
sufficient, if time-consuming.

[And a bit off the beaten track---in fact in .signature territory:]

>Vulcan jealousy: "I fail to see the logic in prefering Stan over me"

Actually, it was `Stonn'; I am not sure about the rest of the quote.
But I do recall this (probably inexactly): `You may find that *having*
something is not quite so pleasing a thing as *wanting* it.'
-- 
In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 1516)
UUCP:	seismo!umcp-cs!chris
CSNet:	chris@umcp-cs		ARPA:	chris@mimsy.umd.edu