On Wed, 2 Dec 1998, Dave Huang wrote:

> On Tue, 1 Dec 1998, Frederick Bruckman wrote:
> > Actually, I do have a -current system, and X isn't wrapping. I took a
> > closer look at "dt", and it isn't exactly wrapping, either. What
> > happens is, white space in the rightmost column is replaced with a
> > random character; any printable character displays correctly. I'll
> > fool around with it this week, but if I don't get anywhere, I'll just
> > submit the package as is.
> 
> I think the problem is the #define VT_MAXCOLS 132 in vt.h. With an
> 800x600 screen and the small (6x10) font, you get 133 columns. The
> problem gets worse with higher resolutions. Perhaps the values of
> VT_MAXROWS and VT_MAXCOLS should be calculated instead of hardcoded...

I see what you mean. The worst part is that drawscreen and friends all
write v->numtcols characters. This causes the last character of a long
line to be put outside the boundaries of the static row arrays, which
are only VT_MAXCOLS long! Most of the time, it simply wraps to the
next line, but there are no guarantees. Needless to say, this is
REALLY BAD.

I'm leaning toward malloc'ing the row arrays in vt_init. I've got to
leave for work now, more later...