SEO

I have a curses program which works fine on numerous other platforms,
including AIX Extended Curses on earlier AIXes. It uses foreground and
background colors. I'm now using curses rather than AIX Extended Curses
on AIX 5.3.

I took the normal vt100 and xterm terminfo entries and added some color
capabilities, to make vt100-color and xterm-color, I added
colors, pairs, setab, setaf, setb, setf, op, bce, ncv
I used "use" to pick up everything else from the base vt100 and xterm
definitions.

All is fine using vt100-color, but with xterm-color, I occassionally
get single character black spaces where I should get a space printed
with a different background colour.

Looking at the xterm-color and vt100-color entries, the only
significant difference is the fact that the xterm-color inherited a
couple of extra capabilities,
dch1, dch, dl, ich1, ich, il

There are some minor differences in padding, but I'm not really
expecting this to be a factor, given my choice of terminal emulator.

I note that when I see the problem, it looks as if the curses library
may have decided to use the ich1 or ich sequences. It certainly would
have been a reasonable thing to do at that point. For example, the line
on the screen might be changing from
word_word
to
wordy_wordy
where the entire line is yellow on red, and the _ is a space.
The (unexpected result) is that sometimes the space becomes (something
on) black.
My guess is that it is actually a white on black space.

I'm using PuTTY, and this emulates xterm (+extra stuff), and I have
confirmed
that it responds correctly to the \E[@ and \[P sequences used by ich
and dch, including the handling of the background.

Note that I do have bce in the terminfo definitions.


So I ask, are there any known issues with terminfo ich/dch and AIX
Curses on 5.3 ?

{{{ Andy

I have repeated the test using a real color xterm with the same result.
Therefore it looks more likely to be a AIX curses issue than a bug in
both PuTTY and real xterm.

{{{ Andy

ak@nyangau.fsnet.co.uk wrote:
> I have repeated the test using a real color xterm with the same result.
> Therefore it looks more likely to be a AIX curses issue than a bug in
> both PuTTY and real xterm.

Which color xterm was that?

PuTTY does differ from xterm - that's why there's a separate terminfo for it.
The ich/dch problem that you described would happen with rxvt (that's a
known bug in rxvt).

--
Thomas E. Dickey
http://invisible-island.net
ftp://invisible-island.net

It was xterm on Linux (not sure which, likely to have been RHAS 2.1 or
RHEL 3.0.

Incidentally, I compile the same program for Solaris (native curses)
or Linux (ncurses), and using xterm or PuTTY, things work fine.

PuTTY claims to emulate xterm, and seems to do a good job, but I'll
have to find the putty terminfo entry you refer to and see whether any
of the differences are of interest.

{{{ Andy

ak@nyangau.fsnet.co.uk wrote:
> It was xterm on Linux (not sure which, likely to have been RHAS 2.1 or
> RHEL 3.0.

> Incidentally, I compile the same program for Solaris (native curses)
> or Linux (ncurses), and using xterm or PuTTY, things work fine.

> PuTTY claims to emulate xterm, and seems to do a good job, but I'll
> have to find the putty terminfo entry you refer to and see whether any
> of the differences are of interest.

ftp://invisible-island.net/ncurses/terminfo.src.gz

(PuTTY doesn't handle wrapping the same as xterm).

--
Thomas E. Dickey
http://invisible-island.net
ftp://invisible-island.net