Write exact results past the range of a double (closes #16)

An exact result outside the normal range of a double, such as 2^1200,
10^400 or 2^-1200, was refused because format wrote every result
through float64. Such a result is now written from its exact value,
rounded to 17 significant digits with trailing zeros dropped, through a
big.Float precise enough to round as the exact value would. Results
inside the range are written as before, and a power computed in float64
must still be a normal double. The tests check the written text of these
results, the refusals just past the 4096-bit limit, and the time taken
to write the largest and smallest numbers under it.

Model: opus-5-5

internal/bot/bot_test.go

@@ -18,6 +18,7 @@ func TestReply(t *testing.T) {
 		"5 * 5/2": "12.5",
 		"2^10":    "1024",
 		"7 % 3":   "1",
+		"2^1200":  "1.7218479456385751e+361",
 	} {
 		if got := bot.Reply(in); got != want {
 			t.Errorf("Reply(%q) = %q, want %q", in, got, want)
@@ -27,7 +28,7 @@ func TestReply(t *testing.T) {
 	for in, want := range map[string]string{
 		"hello":    "I only understand arithmetic",
 		"1 / 0":    "I cannot divide by zero.",
-		"1e400":    "That needs a number too large or too small for me.",
+		"1e1300":   "That needs a number too large or too small for me.",
 		"1e-1300":  "That needs a number too large or too small for me.",
 		"(-8)^0.5": "A negative number to a fractional power has no real",
 		strings.Repeat("1+", calc.MaxInputLength) + "1": "That is too long for me",