Pad/internal/test/e2e/real_file_large_ime_test.go

package e2e_test

import (
	"fmt"
	"strings"
	"testing"
	"time"
	"unicode/utf8"

	"gioui.org/io/key"
	"pad/internal/editor"
	"pad/internal/ui"
)

// largeFileContent generates a deterministic ~1.6 MB text file: 40,000 lines
// of ~40 unique ASCII bytes each, with a multibyte rune ("é", "中") sprinkled
// through every 500th line so rune offsets and byte offsets genuinely differ.
// Generating in-test (t.TempDir, via realFileHarness) keeps the repository
// small while still exercising the >500 KB chunked/IME path the same way a
// real multi-hundred-kb note would.
func largeFileContent(lines int) string {
	var b strings.Builder
	for i := 0; i < lines; i++ {
		b.WriteString(fmt.Sprintf("large file line %05d pad", i))
		if i%500 == 0 {
			b.WriteString(" é 中")
		}
		b.WriteString(strings.Repeat(string(rune('a'+i%26)), 12))
		b.WriteString("\n")
	}
	return b.String()
}

// TestRealFile_LargeFile_IMECommits pins IME commit correctness on a large
// (~1.6 MB, well over the 500 KB threshold that the on-device "needs a
// gioui fork" claim referred to) file:
//
//   - the IME window invariants (window text, leading context, absolute
//     rune offset) hold for a window deep in the file;
//   - a replacement commit (a selection, as a swipe/autocorrect would send)
//     lands exactly in the scrolled window and nowhere else;
//   - an insert commit (empty range at the caret, a plain typed char) in a
//     second, deeper window lands exactly at the caret.
//
// The full-content comparison after each commit catches the classic
// corruption mode (the commit mapped through a stale/zero window start and
// clobbering text near the top of the file).
func TestRealFile_LargeFile_IMECommits(t *testing.T) {
	const (
		lines    = 40000
		scrollLn = 20000 // scroll deep into the file
		editLn   = 20008 // comfortably inside the ~18-line viewport window
	)
	model := largeFileContent(lines)
	h, path := realFileHarness(t, "large.txt", model)
	defer h.Cleanup()

	h.SendConfig(780, 400)
	if _, err := h.WaitForFrame(5 * time.Second); err != nil {
		t.Fatalf("wait for frame: %v", err)
	}

	// Scroll deep (5 dp into the line: avoid the float32/float64 line-pitch
	// boundary fragility, see TestRealFile_WindowSelectionAtScroll) and put a
	// selection on line editLn, 6 bytes into its filler.
	selStart := byteOffsetOfLine(t, model, editLn) + 31
	selEnd := selStart + 6
	if err := h.WithState(func(st *editor.State) {
		st.ScrollOffset = ui.Dp(16.8*scrollLn + 5)
		st.Editor.SelectionStart = selStart
		st.Editor.SelectionEnd = selEnd
		st.Editor.CursorPosition = selEnd
	}); err != nil {
		t.Fatalf("WithState: %v", err)
	}
	prev := h.FrameCount()
	h.SendConfig(780, 400)
	if _, err := h.WaitForFrameCount(prev+1, 5*time.Second); err != nil {
		t.Fatalf("wait for scrolled frame: %v", err)
	}

	// 1) IME window invariants for a window deep in a large file.
	var win struct {
		Start int
		Text  string
		Snip  string
	}
	v, err := h.Inspect(func(st *editor.State) any {
		e := &st.Editor
		return struct {
			Start int
			Text  string
			Snip  string
		}{e.IMEWindowStartByte, e.IMEWindowText, e.IMESnippetText}
	})
	if err != nil {
		t.Fatalf("Inspect: %v", err)
	}
	win = v.(struct {
		Start int
		Text  string
		Snip  string
	})
	if want := model[win.Start : win.Start+len(win.Text)]; win.Text != want {
		t.Fatalf("IME window text desync at start=%d", win.Start)
	}
	// The IME snippet is a hysteresis window around the caret (see
	// State.computeIMESnippetWindow); it must always be a fresh slice of
	// the model.
	if i := strings.Index(model, win.Snip); i < 0 {
		t.Fatalf("IME snippet not a slice of the model: %q", win.Snip[:min(40, len(win.Snip))])
	}

	// 2) Replacement commit: absolute file runes (the coordinate space the
	//    real IME uses; the multibyte sprinkles make runes != bytes).
	absStart := utf8.RuneCountInString(model[:selStart])
	absEnd := utf8.RuneCountInString(model[:selEnd])
	wantModel := model[:selStart] + "ZZ" + model[selEnd:]
	h.SendInput([]ui.InputEvent{{
		Handler: editor.HandleKeyDown,
		Data: key.EditEvent{
			Range: key.Range{Start: absStart, End: absEnd},
			Text:  "ZZ",
		},
	}})
	time.Sleep(100 * time.Millisecond)

	got, err := h.FullContent()
	if err != nil {
		t.Fatalf("FullContent: %v", err)
	}
	if got != wantModel {
		t.Fatalf("after replacement commit: content differs (first diff at %d)", firstDiff(wantModel, got))
	}

	// 3) Insert commit in a deeper window: caret-only (empty range), the
	//    plain "typed a character" case.
	deepLn := 30000
	caret := byteOffsetOfLine(t, wantModel, deepLn) + 20
	if err := h.WithState(func(st *editor.State) {
		st.ScrollOffset = ui.Dp(16.8*float64(deepLn) + 5)
		st.Editor.SelectionStart = -1
		st.Editor.SelectionEnd = -1
		st.Editor.CursorPosition = caret
	}); err != nil {
		t.Fatalf("WithState: %v", err)
	}
	prev = h.FrameCount()
	h.SendConfig(780, 400)
	if _, err := h.WaitForFrameCount(prev+1, 5*time.Second); err != nil {
		t.Fatalf("wait for second scrolled frame: %v", err)
	}
	absCaret := utf8.RuneCountInString(wantModel[:caret])
	wantModel2 := wantModel[:caret] + "Q" + wantModel[caret:]
	h.SendInput([]ui.InputEvent{{
		Handler: editor.HandleKeyDown,
		Data: key.EditEvent{
			Range: key.Range{Start: absCaret, End: absCaret},
			Text:  "Q",
		},
	}})
	time.Sleep(100 * time.Millisecond)

	got, err = h.FullContent()
	if err != nil {
		t.Fatalf("FullContent: %v", err)
	}
	if got != wantModel2 {
		t.Fatalf("after insert commit: content differs (first diff at %d)", firstDiff(wantModel2, got))
	}
	if cp, err := h.Inspect(func(st *editor.State) any { return st.Editor.CursorPosition }); err != nil {
		t.Fatalf("Inspect: %v", err)
	} else if cp.(int) != caret+1 {
		t.Fatalf("cursor = %d, want %d", cp.(int), caret+1)
	}

	// 4) Persistence: the on-disk file matches the in-memory content.
	if err := h.Flush(); err != nil {
		t.Fatalf("flush: %v", err)
	}
	if disk := readDisk(t, path); disk != wantModel2 {
		t.Fatalf("disk content differs from in-memory (first diff at %d)", firstDiff(wantModel2, disk))
	}
}

// byteOffsetOfLine returns the absolute byte offset of the start of line n
// (0-indexed) in model. The generated lines are unique, so a plain index of
// the line prefix is unambiguous.
func byteOffsetOfLine(t *testing.T, model string, n int) int {
	t.Helper()
	prefix := fmt.Sprintf("large file line %05d", n)
	i := strings.Index(model, prefix)
	if i < 0 {
		t.Fatalf("line %d not found", n)
	}
	return i
}

// firstDiff returns the byte offset of the first difference between a and b
// (min(len) when they are a common prefix), for useful failure messages.
func firstDiff(a, b string) int {
	n := len(a)
	if len(b) < n {
		n = len(b)
	}
	for i := 0; i < n; i++ {
		if a[i] != b[i] {
			return i
		}
	}
	return n
}

// TestRealFile_IMEForceResync pins the desync self-heal: after several
// consecutive anomalous IME commits (drift-snapped or empty — the pattern
// Gboard produces when its local text has desynchronized from the app and it
// keeps re-sending the same fix), the next frame arms a forced snippet
// re-push (a restartInput that makes the IME re-fetch the real text). A
// normal commit resets the streak, so isolated anomalies never resync.
func TestRealFile_IMEForceResync(t *testing.T) {
	model := largeFileContent(4000)
	h, _ := realFileHarness(t, "resync.txt", model)
	defer h.Cleanup()

	h.SendConfig(780, 400)
	if _, err := h.WaitForFrame(5 * time.Second); err != nil {
		t.Fatalf("wait for frame: %v", err)
	}

	// A caret deep in the file, in absolute file runes.
	const caretByte = 10000
	caretRune := utf8.RuneCountInString(model[:caretByte])
	if err := h.WithState(func(st *editor.State) {
		st.Editor.CursorPosition = caretByte
	}); err != nil {
		t.Fatalf("WithState: %v", err)
	}

	commit := func(start, end int, text string) {
		h.SendInput([]ui.InputEvent{{
			Handler: editor.HandleIMECommit,
			Data:    editor.IMECommit{StartRune: start, EndRune: end, Text: text},
		}})
		// The logic goroutine applies commits asynchronously; wait for the
		// resulting frame before asserting on the derived state.
		prev := h.FrameCount()
		if _, err := h.WaitForFrameCount(prev+1, 5*time.Second); err != nil {
			t.Fatalf("wait for commit frame: %v", err)
		}
	}
	// forceInLatestFrame reports whether the latest delivered frame's editor
	// element carries the forced-resync flag (the state flag is consumed by
	// the very frame that follows the arming commit, so the frame is the
	// observable artifact).
	forceInLatestFrame := func() bool {
		frames := h.GetFrames()
		if len(frames) == 0 {
			t.Fatal("no frames")
		}
		for _, el := range frames[len(frames)-1] {
			if tf, ok := el.(ui.TextField); ok && tf.ID() == "editor_text" {
				return tf.IMEForceResync
			}
		}
		t.Fatal("no editor_text element in latest frame")
		return false
	}

	// Two anomalous commits (an empty fix-up and a drifted insertion):
	// below the threshold, no resync.
	commit(caretRune+5, caretRune+5, "")
	commit(caretRune+6, caretRune+6, "z")
	if forceInLatestFrame() {
		t.Fatalf("resync armed after only 2 anomalous commits")
	}

	// A normal commit resets the streak: two more anomalies must not arm.
	commit(caretRune+10, caretRune+10, "q")
	commit(caretRune+11, caretRune+11, "")
	commit(caretRune+12, caretRune+12, "")
	if forceInLatestFrame() {
		t.Fatalf("resync armed after a reset streak")
	}

	// One more anomaly makes three in a row: the frame emitted by that
	// commit carries the flag on its editor element (the state flag is
	// consumed by that same frame).
	commit(caretRune+13, caretRune+13, "")
	if !forceInLatestFrame() {
		t.Fatalf("resync not armed after 3 consecutive anomalous commits")
	}

	// A later frame (no new anomalies) carries no flag: the resync fires
	// exactly once.
	prev := h.FrameCount()
	h.SendConfig(780, 400)
	if _, err := h.WaitForFrameCount(prev+1, 5*time.Second); err != nil {
		t.Fatalf("wait for frame: %v", err)
	}
	if forceInLatestFrame() {
		t.Fatalf("resync flag not one-shot")
	}
}