WIP: restore lost C behaviors (schtick message, forced redraw, greeting)

Incomplete checkpoint, preserved after the implementing session hit a
capacity limit partway through mutation verification. NOT reviewed and NOT
gate-clean: make fmt has not been run, so fmt-check currently fails on
ARCHITECTURE.md and TODO.md.

Work still outstanding before this is fit for review:
  - run make fmt and fold the result in
  - finish mutation-proving each of the three behaviors
  - verify every message string byte-for-byte against the C sources
  - confirm TestSeedCompatItemTables passes with its golden untouched

Refs #13.

ARCHITECTURE.md

@@ -1488,6 +1488,8 @@ type Terminal interface {
     ReadChar() (byte, bool)   // event loop → C char codes (arrows→hjkl, ^C→quit);
                               // ok is false when Interrupt woke the read
     Interrupt()               // wake a blocked ReadChar (signal goroutine only)
+    Repaint()                 // forced full redraw for CTRL-R
+                              // (curses clearok(curscr,TRUE)+wrefresh(curscr))
     Fini()                    // restore the device (curses endwin) on exit
 }
 
@@ -1507,13 +1509,22 @@ func (w *Window) Printwf(format string, a ...any); func (w *Window) Inch(y, x in
 func (w *Window) Clear(); func (w *Window) Clrtoeol(); func (w *Window) Standout(on bool)
 func (s *Screen) Refresh()                  // blit stdscr to the device
 func (s *Screen) RefreshWin(w *Window)      // blit any window
+func (s *Screen) Repaint()                  // force a full redraw of the device
 func (s *Screen) Fini()                     // tear the device down
 ```
 
 Ported drawing code keeps its structure: `mvaddch(y, x, ch)` →
 `g.scr.Std.MvAddCh(y, x, ch)`. Curses `mvinch` reads come from the Window
 buffer, preserving the "screen is a data structure" idiom without touching the
-real terminal. `md_readchar`'s escape decoding is deleted; tcell's `EventKey`
+real terminal. `Repaint` is the one drawing call that is not a blit: it is the
+`CTRL('R')` command's `clearok(curscr, TRUE)` plus `wrefresh(curscr)`,
+implemented as tcell's `Screen.Sync`, and it exists because `Render` cannot do
+its job. Both diff a new frame against the device's record of the old one and
+send nothing where they agree, which is the wrong answer precisely when the
+screen has been corrupted by something else's output — the only reason a player
+types `CTRL('R')`. Like C's, it repaints what was last drawn rather than
+re-blitting `stdscr`, so it takes no window. `md_readchar`'s escape decoding is
+deleted; tcell's `EventKey`
 provides decoded keys and we translate to the byte codes `command()` already
 handles (KeyUp → 'k', etc.). Resize is handled by tcell, which registers only
 SIGWINCH — SIGTSTP is not among the signals it takes, and is dropped (§9).