[PATCH weston] terminal: Fix segmentation fault when processing DCH ANSI escape code

[aaron at correspondwith.me]

All good points, and your previous review pointed out that
terminal_shift_line is also used for ICH, though in that case I think
the parameter is always positive, and so this would not apply.

I think in the case of ICH it would wrap, whereas the VT docs seem to
say that DCH has no effect outside the margins. Then again, the docs
could be wrong, and I see something said to that effect in the Xterm
changelog (http://invisible-island.net/xterm/xterm.log.html#xterm_316).

I will look into both the desired behaviour of DCH, and possibly moving
this logic to the calculation of d, if it makes sense to.

Thanks