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.

game/sticks.go

@@ -31,14 +31,30 @@ func (g *RogueGame) doZap() {
 		return
 	}
 
-	// A malformed Which yields no handler, which is exactly what C did in a
-	// non-MASTER build: its do_zap switch matched no case, fell out, and
-	// still ran o_charges--. (The MASTER-only "what a bizarre schtick!"
-	// message for that arm is issue #13, not this bounds fix.)
-	if h := g.data.zapHandler(obj); h != nil {
+	// C's switch has a case for every one of the 14 WS_ kinds, so its
+	// closing "otherwise: msg(...)" arm is reachable only for an o_which
+	// outside the table — the malformed objects hasValidWhich screens
+	// for, which is why no handler and a legal Which can only mean WS_NOP.
+	//
+	// The message is under #ifdef MASTER, not under a wizard test: C
+	// printed it for every player of a MASTER build, which is the build
+	// this port is (see the '+' command, issue #11). Do not gate it on
+	// g.Wizard.
+	//
+	// WS_NOP is a case of its own ("when WS_NOP: break;"): the wand that
+	// deliberately does nothing says nothing either. All three arms fall
+	// out of the switch into o_charges--, so even the bizarre schtick
+	// costs a charge.
+	h := g.data.zapHandler(obj)
+
+	switch {
+	case h != nil:
 		if !h(g, obj) {
 			return // the zap aborted; no charge is used
 		}
+	case obj.hasValidWhich(): // WS_NOP
+	default:
+		g.msg("what a bizarre schtick!")
 	}
 
 	obj.Charges--