[13347] in Athena Bugs
curses library: yet another followup
daemon@ATHENA.MIT.EDU (bcb@MIT.EDU)
Tue Mar 21 02:39:22 1995
From: bcb@MIT.EDU
Date: Tue, 21 Mar 95 02:39:17 -0500
To: bugs@MIT.EDU
Okay, I've found part of the source of my problem. My description of it may
have been misleading.
The problem is that the combination of "newterm," "initscr," and manual
(but correct) setting of the LINES and COLS global variables do not, when
combined, actually tell curses to use the dimensions of the terminal it's
talking *to*. Somewhere "under the covers" of curses, it maintains a deep
belief that if the program was started in a 22x81 window, for instnace,
then the window it's talking to must be 22x81 as well (as opposed to, say,
25x80). This results in all sorts of weird, untraceable bugs; they are
untraceable for exactly the reason that as far as the *interface* to curses
is concerned, the screen in 25x80 and it just wrote on line 24. However,
deep down inside, it *knows* that the screen is only 22x81, and when it
thought it was being told to write on line 24, it really had to scroll
twice and write on line 22. It then reports back to its upper levels that
it did what it was supposed to.
I haven't found a fix for this yet. If I do, I'll let you know. If you know
one already, please tell me.
Brian
