Wait for rest of escape sequence before parsing (#4901)

Read loop dropped its escDelay retry budget after every successful byte,
so a sequence split across reads reached the parser as a fragment, parsed
as ALT-[ with the remainder left behind as query text.

- fzf queries DECRQM at startup since dab626b, so a terminal answering
  late leaked "?2004;2$y" into the query
- Same split leaked modified keys and mouse sequences: CTRL-UP left "5A",
  SGR mouse left "0;1;1M"
- Bound unchanged, a stall longer than escDelay still falls back to ALT

Fix #4899

src/tui/light_escape_test.go

@@ -0,0 +1,53 @@
+package tui
+
+import (
+	"strings"
+	"testing"
+)
+
+func TestIncompleteEscape(t *testing.T) {
+	for _, c := range []struct {
+		buffer string
+		want   bool
+	}{
+		// Complete sequences: nothing to wait for
+		{"\x1b[A", false},
+		{"\x1bOA", false},
+		{"\x1b[1;5A", false},
+		{"\x1b[200~", false},
+		{"\x1b[<0;1;1M", false},
+		{"\x1b[12;34R", false},
+		{"\x1b[?2004;2$y", false},
+		{"\x1b[?1;2c", false},
+
+		// Fragments: keep waiting
+		{"\x1b[", true},
+		{"\x1b[?", true},
+		{"\x1b[1;", true},
+		{"\x1b[?2004;2$", true},
+		{"\x1bO", true},
+		{"\x1b[<0;1;", true},
+
+		// Only the trailing sequence matters
+		{"ab\x1b[?2004;2$", true},
+		{"\x1b[A\x1b[", true},
+		{"\x1b[A\x1b[B", false},
+
+		// Long buffers: only the tail is scanned, so an introducer further
+		// back than escapeLookback is not waited for
+		{strings.Repeat("a", 100000), false},
+		{"\x1b[" + strings.Repeat("a", 100000), false},
+		{strings.Repeat("a", 100000) + "\x1b[1;", true},
+
+		// Not a sequence fzf waits on
+		{"", false},
+		{"abc", false},
+		{"\x1b", false},      // lone ESC, handled by the existing escDelay branch
+		{"\x1ba", false},     // ALT-a
+		{"\x1b[\x01", false}, // malformed, do not stall on it
+	} {
+		if got := incompleteEscape([]byte(c.buffer)); got != c.want {
+			t.Errorf("incompleteEscape(%q) = %v, want %v", c.buffer, got, c.want)
+		}
+	}
+}