Drive capture selection through a CGEventTap

Panel event delivery is unreliable on macOS 26 (clicks pass through even
with the app active and the panel key). The selection flow is now driven
by a CGEventTap that sees and consumes mouse events system-wide: hover
highlights windows, clicks capture them (rounded), drags marquee-select,
Esc cancels. Requires Accessibility permission, prompted on first use;
falls back to panel events when unavailable.

💘 Generated with Crush

Assisted-by: Crush:deepseek-v4-flash-0731
This commit is contained in:
2026-09-09 12:08:57 +08:00
parent a7cad2fc2f
commit 7e9cc5b269
4 changed files with 265 additions and 189 deletions
+222 -12
View File
@@ -1,27 +1,47 @@
import AppKit
import CoreGraphics
import ScreenCaptureKit
import ApplicationServices
final class CaptureController: NSObject {
static let shared = CaptureController()
private var overlays: [NSWindow] = []
private var overlays: [OverlayWindow] = []
private(set) var isSelecting = false
private var previousApp: NSRunningApplication?
private var overlayIDs: [Int] = []
// Selection state (driven by the event tap; AppKit global coordinates)
private var marqueeStart: NSPoint?
private(set) var marqueeRect: NSRect?
private(set) var hoverFrame: NSRect?
private var trackedWindows: [(id: CGWindowID, frame: NSRect)] = []
private var lastWindowScan = Date.distantPast
private var eventTap: CFMachPort?
/// Session event log for debugging (dumped via Copy Diagnostics).
/// Also appended to /tmp/kapture-debug.log so it survives relaunches.
static var sessionLog: [String] = []
static func log(_ line: String) {
sessionLog.append(line)
if sessionLog.count > 100 { sessionLog.removeFirst() }
let stamped = String(format: "%@ %@", DateFormatter.localizedString(from: Date(), dateStyle: .none, timeStyle: .medium), line)
sessionLog.append(stamped)
if sessionLog.count > 200 { sessionLog.removeFirst() }
NSLog("[Kapture] \(line)")
let path = "/tmp/kapture-debug.log"
if let handle = FileHandle(forWritingAtPath: path) {
handle.seekToEndOfFile()
handle.write((stamped + "\n").data(using: .utf8)!)
handle.closeFile()
} else {
FileManager.default.createFile(atPath: path, contents: (stamped + "\n").data(using: .utf8))
}
}
/// Diagnostic snapshot for "Copy Diagnostics" menu item.
func diagnostics() -> String {
var lines: [String] = []
lines.append("Kapture diagnostics — version \(Bundle.main.infoDictionary?["CFBundleShortVersionString"] ?? "?")")
lines.append("accessibility=\(AXIsProcessTrusted()) eventTap=\(eventTap != nil)")
lines.append("--- session log (newest last) ---")
lines.append(contentsOf: Self.sessionLog)
lines.append("--- screens ---")
@@ -53,6 +73,8 @@ final class CaptureController: NSObject {
return lines.joined(separator: "\n")
}
// MARK: - Session lifecycle
func begin() {
guard !isSelecting else { return }
guard PermissionManager.isGranted else {
@@ -61,25 +83,22 @@ final class CaptureController: NSObject {
}
isSelecting = true
previousApp = NSWorkspace.shared.frontmostApplication
// Activation is ASYNC: isActive only flips on a later runloop tick
// (verified on macOS 26). Show overlays only once active, otherwise
// the first click can pass through to the windows below.
// Activation is async on macOS 26; wait a tick so the app is active
// before the overlays appear (otherwise clicks pass through).
NSApp.activate(ignoringOtherApps: true)
CaptureController.log("begin: activate issued, active=\(NSApp.isActive)")
DispatchQueue.main.asyncAfter(deadline: .now() + 0.05) { [weak self] in
guard let self, self.isSelecting else { return }
CaptureController.log("begin: active=\(NSApp.isActive) — showing overlays")
self.overlays = NSScreen.screens.map { screen in
OverlayWindow(
screen: screen,
onComplete: { [weak self] rect in self?.finishSelection(rect) },
onCancel: { [weak self] in self?.cancelSelection() },
onWindowPick: { [weak self] rect in self?.finishWindowPick(rect) }
)
let panel = OverlayWindow(screen: screen, controller: self)
(panel.contentView as? OverlayView)?.controller = self
return panel
}
self.overlays.forEach { $0.makeKeyAndOrderFront(nil) }
self.overlayIDs = self.overlays.map { $0.windowNumber }
CaptureController.log("begin: \(self.overlays.count) overlays, ids=\(self.overlayIDs), active=\(NSApp.isActive)")
self.installEventTap()
}
}
@@ -113,8 +132,12 @@ final class CaptureController: NSObject {
}
private func hideOverlays() {
removeEventTap()
overlays.forEach { $0.orderOut(nil) }
overlays = []
marqueeRect = nil
hoverFrame = nil
marqueeStart = nil
}
private func restorePreviousApp() {
@@ -122,6 +145,193 @@ final class CaptureController: NSObject {
previousApp = nil
}
private func renderOverlays() {
for overlay in overlays {
(overlay.contentView as? OverlayView)?.needsDisplay = true
}
}
// MARK: - Event tap
private func installEventTap() {
guard eventTap == nil else { return }
guard AXIsProcessTrusted() else {
CaptureController.log("tap: accessibility not granted — falling back to panel events")
PermissionManager.requestAccessibility()
return
}
let mask: CGEventMask = (1 << CGEventType.mouseMoved.rawValue)
| (1 << CGEventType.leftMouseDown.rawValue)
| (1 << CGEventType.leftMouseDragged.rawValue)
| (1 << CGEventType.leftMouseUp.rawValue)
| (1 << CGEventType.keyDown.rawValue)
guard let tap = CGEvent.tapCreate(
tap: .cghidEventTap,
place: .headInsertEventTap,
options: .defaultTap,
eventsOfInterest: CGEventMask(mask),
callback: Self.tapCallback,
userInfo: Unmanaged.passUnretained(self).toOpaque()
) else {
CaptureController.log("tap: creation FAILED — falling back to panel events")
return
}
eventTap = tap
if let source = CFMachPortCreateRunLoopSource(nil, tap, 0) {
CFRunLoopAddSource(CFRunLoopGetMain(), source, .commonModes)
}
CGEvent.tapEnable(tap: tap, enable: true)
CaptureController.log("tap: installed")
}
private func removeEventTap() {
guard let tap = eventTap else { return }
CGEvent.tapEnable(tap: tap, enable: false)
if let source = CFMachPortCreateRunLoopSource(nil, tap, 0) {
CFRunLoopRemoveSource(CFRunLoopGetMain(), source, .commonModes)
}
CFMachPortInvalidate(tap)
eventTap = nil
CaptureController.log("tap: removed")
}
private static let tapCallback: CGEventTapCallBack = { _, type, event, refcon in
guard let refcon else { return Unmanaged.passUnretained(event) }
let controller = Unmanaged<CaptureController>.fromOpaque(refcon).takeUnretainedValue()
if let result = controller.handleTap(type, event) {
return Unmanaged.passUnretained(result)
}
return nil
}
/// Convert a CG global point (top-left origin) to AppKit global.
private func cgPointToAppKit(_ p: NSPoint) -> NSPoint {
guard let mainScreen = NSScreen.screens.first else { return p }
return NSPoint(x: p.x, y: mainScreen.frame.height - p.y)
}
/// Scan on-screen windows (AppKit global frames), front to back.
private func refreshTrackedWindows() {
guard let list = CGWindowListCopyWindowInfo(
[ .optionOnScreenOnly, .excludeDesktopElements ], kCGNullWindowID
) as? [[String: Any]] else { trackedWindows = []; return }
let ownIDs = Set(overlayIDs)
var found: [(id: CGWindowID, frame: NSRect)] = []
for info in list {
guard let cid = info[kCGWindowNumber as String] as? CGWindowID,
let bounds = info[kCGWindowBounds as String] as? [String: CGFloat],
let layer = info[kCGWindowLayer as String] as? Int,
layer == 0
else { continue }
if ownIDs.contains(Int(cid)) { continue }
let cgRect = NSRect(
x: bounds["X"] ?? 0,
y: bounds["Y"] ?? 0,
width: bounds["Width"] ?? 0,
height: bounds["Height"] ?? 0
)
guard cgRect.width >= 80, cgRect.height >= 60 else { continue }
let rect = Self.cgToAppKit(rect: cgRect)
// Skip windows filling an entire screen (fullscreen apps)
let coversWholeScreen = NSScreen.screens.contains { screen in
abs(screen.frame.width - rect.width) < 2 && abs(screen.frame.height - rect.height) < 2
}
guard !coversWholeScreen else { continue }
found.append((cid, rect))
}
trackedWindows = found
}
/// Convert a CGWindowList (top-left origin) frame into AppKit screen
/// coordinates (bottom-left origin).
private static func cgToAppKit(rect: NSRect) -> NSRect {
guard let mainScreen = NSScreen.screens.first else { return rect }
let y = mainScreen.frame.height - rect.origin.y - rect.height
return NSRect(x: rect.origin.x, y: y, width: rect.width, height: rect.height)
}
private func topmostWindowContaining(_ appKitPoint: NSPoint) -> NSRect? {
for (_, frame) in trackedWindows {
if frame.contains(appKitPoint) {
return frame
}
}
return nil
}
/// Returns the event to pass through, or nil to consume it.
func handleTap(_ type: CGEventType, _ event: CGEvent) -> CGEvent? {
guard isSelecting else { return event }
switch type {
case .tapDisabledByUserInput:
CGEvent.tapEnable(tap: eventTap!, enable: true)
return event
case .tapDisabledByTimeout:
CGEvent.tapEnable(tap: eventTap!, enable: true)
return event
case .keyDown:
if event.getIntegerValueField(.keyboardEventKeycode) == 53 {
CaptureController.log("tap: Esc — cancel")
cancelSelection()
return nil
}
return event
case .mouseMoved, .leftMouseDragged:
// Throttle window scans while moving
if Date().timeIntervalSince(lastWindowScan) > 0.15 {
lastWindowScan = Date()
refreshTrackedWindows()
}
let appKitPoint = cgPointToAppKit(NSPoint(x: event.location.x, y: event.location.y))
if type == .mouseMoved {
let frame = topmostWindowContaining(appKitPoint)
if frame != hoverFrame {
hoverFrame = frame
renderOverlays()
}
return event
}
// Drag: grow the marquee
guard let start = marqueeStart else { return nil }
let rect = NSRect(
x: min(start.x, appKitPoint.x),
y: min(start.y, appKitPoint.y),
width: abs(appKitPoint.x - start.x),
height: abs(appKitPoint.y - start.y)
)
marqueeRect = rect
hoverFrame = nil
renderOverlays()
return nil
case .leftMouseDown:
marqueeStart = cgPointToAppKit(NSPoint(x: event.location.x, y: event.location.y))
marqueeRect = nil
CaptureController.log("tap: mouseDown at \(marqueeStart.map { "\($0)" } ?? "?")")
return nil
case .leftMouseUp:
defer { marqueeStart = nil; marqueeRect = nil }
let appKitPoint = cgPointToAppKit(NSPoint(x: event.location.x, y: event.location.y))
let dragDistance = marqueeStart.map { hypot($0.x - appKitPoint.x, $0.y - appKitPoint.y) } ?? 0
if let rect = marqueeRect, dragDistance >= 4, rect.width >= 2, rect.height >= 2 {
CaptureController.log("tap: marquee complete \(rect)")
finishSelection(rect)
} else if let hover = hoverFrame {
CaptureController.log("tap: window pick \(hover)")
finishWindowPick(hover)
} else {
CaptureController.log("tap: click with no hover/marquee — cancel")
cancelSelection()
}
return nil
default:
return event
}
}
// MARK: - Capture
private func capture(_ rect: NSRect, excludingWindowIDs ids: [Int], roundedCorners: Bool = false) {
ScreenCapture.shared.capture(rect, excludingWindowIDs: ids, roundedCorners: roundedCorners) { [weak self] image in
DispatchQueue.main.async {