Question about "ul" in termcap

[ajs at hpfcla.UUCP]

Thanks for the quick and succinct responses.  Here is the question again,
followed by excerpts from the answers.

Alan Silverstein

Q> From termcap(5):
Q>	If your terminal correctly generates	underlined  characters
Q>	(with	 no  special  codes  needed)  even  though it does not
Q>	overstrike, then you should give the capability ul.
Q> It's not clear HOW the terminal  should  generate  underline  characters
Q> without special codes.  Are there terminals that can backspace and add a
Q> non-destructive underscore (only)?  Is that what this is intended for?

A1> There are terminals which cannot normally overstrike but which, if sent
A1> the sequence "x<BACKSPACE>_" will display an underscored x.  These
A1> terminals are described by "ul".  Terminals described by "os" obviously
A1> can do this.  Thus, it is correct for more(1) to use this scheme to
A1> underline on terminals described by either of these attributes.

A2> When I found that the 4.1c BSD termcap entry for Ann Arbor Ambassadors
A2> had 'ul' and that they didn't work with more(1) or ul(1), I suspected
A2> that the definition was funny.
A2> I interpret the definition "terminal underlines though it doesn't
A2> overstrike" as meaning "some terminals can take the sequence character-
A2> backspace-underscore and translate it into underscore".

A3> Perhaps your termcap lacks the us= and ue= entries which tell how to
A3> active and terminate the underlining.  (In fact my terminal as I write
A3> this does not have the ul property but does underline (because this
A3> terminal does have magic cookies around the underlining).

	No, the termcap has "us" and "ue" (and even "so" and "se"); the
	question was when it also needed "ul", and why.

A4> Some terminals (I think there's a Concept that does this) transform
A4> <character><backspace>_ (or actually any _ that is printed on top of
A4> a visible character) into an underlined character.  So foo\b\b\b___
A4> makes "foo" become underlined also.