direct mode ought be able to place and discover cursor

[dankamongmen]

While thinking about some things, it became clear that we ought support cursor movement in direct mode. One also ought be able to get the cursor position. For an example of why this would be useful, let me quote HTPwNotcurses:

it is possible to, for instance, periodically acquire the current cursor position, move elsewhere on
the screen, update a HUD, and return to the departure position. Scrolling could be addressed by retaining a copy of any obliterated output. This would suffer a startup period of one screen, during which the area scrolled above the HUD would be cleared. This could be avoided by aligning the HUD with the top of the terminal.

if we actually hacked up such a tool (and see #381), we should definitely run it in the demo somehow.

[dankamongmen]

DECSTBM (set scrolling region) is kinda perfect for this, though it only lets us specify rows.

[dankamongmen]

• detect terminal geometry in direct mode
• set cursor position in direct mode
• discover cursor position in direct mode

[dankamongmen]

Detecting cursor position is tough, since we need send the Device Status Inquiry, and then read its output. We don't have any active context with which to read said output, though I intend to add one. Going ahead and marking this closed.

[dankamongmen]

Updated python, readme, man page, C++