Use screen crop for owned popup windows

desktopIndependentWindow composites owned popups together with their
parent window and scales the group to the popup's size. Detect the case
at pick time — another window from the same app overlapping the picked
one — and capture the screen crop with a rounded mask instead, which
contains only the popup.

💘 Generated with Crush

Assisted-by: Crush:deepseek-v4-flash-0731

Sources/Kapture/CaptureController.swift

@@ -14,8 +14,9 @@ final class CaptureController: NSObject {
     // 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(set) var hoverWindow: (id: CGWindowID, frame: NSRect, pid: Int)?
+    var hoverFrame: NSRect? { hoverWindow?.frame }
+    private var trackedWindows: [(id: CGWindowID, frame: NSRect, pid: Int)] = []
     private var lastWindowScan = Date.distantPast
     private var eventTap: CFMachPort?
 
@@ -102,14 +103,19 @@ final class CaptureController: NSObject {
         }
     }
 
-    /// A window was picked from hover — capture it with rounded corners.
-    private func finishWindowPick(_ rect: NSRect) {
+    /// A window was picked from hover — capture it directly (real rounded
+    /// corners come from the window's own alpha).
+    private func finishWindowPick(_ windowID: CGWindowID, _ rect: NSRect, ownedRisk: Bool) {
         let ids = overlayIDs
         hideOverlays()
         isSelecting = false
         restorePreviousApp()
         DispatchQueue.main.async { [self] in
-            capture(rect, excludingWindowIDs: ids, roundedCorners: true)
+            if ownedRisk {
+                capture(rect, excludingWindowIDs: ids, roundedCorners: true)
+            } else {
+                capture(rect, excludingWindowIDs: ids, windowID: windowID)
+            }
         }
     }
 
@@ -136,7 +142,7 @@ final class CaptureController: NSObject {
         overlays.forEach { $0.orderOut(nil) }
         overlays = []
         marqueeRect = nil
-        hoverFrame = nil
+        hoverWindow = nil
         marqueeStart = nil
     }
 
@@ -217,11 +223,12 @@ final class CaptureController: NSObject {
         ) as? [[String: Any]] else { trackedWindows = []; return }
 
         let ownIDs = Set(overlayIDs)
-        var found: [(id: CGWindowID, frame: NSRect)] = []
+        var found: [(id: CGWindowID, frame: NSRect, pid: Int)] = []
         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,
+                  let pid = info[kCGWindowOwnerPID as String] as? Int,
                   layer == 0
             else { continue }
             if ownIDs.contains(Int(cid)) { continue }
@@ -233,12 +240,7 @@ final class CaptureController: NSObject {
             )
             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))
+            found.append((id: cid, frame: rect, pid: pid))
         }
         trackedWindows = found
     }
@@ -251,10 +253,10 @@ final class CaptureController: NSObject {
         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
+    private func topmostWindowContaining(_ appKitPoint: NSPoint) -> (id: CGWindowID, frame: NSRect, pid: Int)? {
+        for entry in trackedWindows {
+            if entry.frame.contains(appKitPoint) {
+                return entry
             }
         }
         return nil
@@ -265,11 +267,12 @@ final class CaptureController: NSObject {
         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)
+        case .tapDisabledByUserInput, .tapDisabledByTimeout:
+            // A timed-out tap must be re-enabled; guard against a torn-down tap.
+            if let tap = eventTap {
+                CGEvent.tapEnable(tap: tap, enable: true)
+                CaptureController.log("tap: re-enabled after \(type)")
+            }
             return event
         case .keyDown:
             if event.getIntegerValueField(.keyboardEventKeycode) == 53 {
@@ -286,9 +289,14 @@ final class CaptureController: NSObject {
             }
             let appKitPoint = cgPointToAppKit(NSPoint(x: event.location.x, y: event.location.y))
             if type == .mouseMoved {
-                let frame = topmostWindowContaining(appKitPoint)
-                if frame != hoverFrame {
-                    hoverFrame = frame
+                let window = topmostWindowContaining(appKitPoint)
+                if window?.frame != hoverFrame {
+                    hoverWindow = window
+                    if let w = window {
+                        CaptureController.log("hover: picked id=\(w.id) frame=\(w.frame) (tracked=\(trackedWindows.count))")
+                    } else {
+                        CaptureController.log("hover: none (tracked=\(trackedWindows.count))")
+                    }
                     renderOverlays()
                 }
                 return event
@@ -302,7 +310,7 @@ final class CaptureController: NSObject {
                 height: abs(appKitPoint.y - start.y)
             )
             marqueeRect = rect
-            hoverFrame = nil
+            hoverWindow = nil
             renderOverlays()
             return nil
         case .leftMouseDown:
@@ -317,9 +325,15 @@ final class CaptureController: NSObject {
             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 if let hover = hoverWindow {
+                // An overlapping window from the same app means this is an
+                // owned popup — desktopIndependentWindow would composite the
+                // parent too, so capture the screen crop instead.
+                let ownedRisk = trackedWindows.contains {
+                    $0.id != hover.id && $0.pid == hover.pid && $0.frame.intersects(hover.frame)
+                }
+                CaptureController.log("tap: window pick id=\(hover.id) pid=\(hover.pid) ownedRisk=\(ownedRisk) \(hover.frame)")
+                finishWindowPick(hover.id, hover.frame, ownedRisk: ownedRisk)
             } else {
                 CaptureController.log("tap: click with no hover/marquee — cancel")
                 cancelSelection()
@@ -332,8 +346,8 @@ final class CaptureController: NSObject {
 
     // 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
+    private func capture(_ rect: NSRect, excludingWindowIDs ids: [Int], roundedCorners: Bool = false, windowID: CGWindowID? = nil) {
+        ScreenCapture.shared.capture(rect, excludingWindowIDs: ids, roundedCorners: roundedCorners, windowID: windowID) { [weak self] image in
             DispatchQueue.main.async {
                 self?.isSelecting = false
                 self?.restorePreviousApp()
@@ -342,6 +356,7 @@ final class CaptureController: NSObject {
                         ResultWindow.shared.show(image: ResultWindow.failureImage("Capture failed. Check Screen Recording permission."))
                         return
                     }
+                    CaptureController.log("capture done: \(image.width)x\(image.height) — showing result window")
                     ResultWindow.shared.show(image: image)
                 }
             }
@@ -417,7 +432,7 @@ final class ScreenCapture {
         return content.displays.first(where: { $0.displayID == cgID }) ?? content.displays.first
     }
 
-    func capture(_ rect: NSRect, excludingWindowIDs excludedIDs: [Int], roundedCorners: Bool = false, completion: @escaping (CGImage?) -> Void) {
+    func capture(_ rect: NSRect, excludingWindowIDs excludedIDs: [Int], roundedCorners: Bool = false, windowID: CGWindowID? = nil, completion: @escaping (CGImage?) -> Void) {
         if #available(macOS 14.0, *) {
             Task {
                 guard let content = try? await SCShareableContent.current,
@@ -440,7 +455,7 @@ final class ScreenCapture {
                 let backingScale = screen.backingScaleFactor
                 config.width = Int(CGFloat(display.width) * backingScale)
                 config.height = Int(CGFloat(display.height) * backingScale)
-                config.showsCursor = true
+                config.showsCursor = false
 
                 // ScreenCaptureKit uses a different origin than AppKit for
                 // screens stacked around the main display: per-display,
@@ -452,6 +467,37 @@ final class ScreenCapture {
                     height: rect.height
                 )
 
+                // Window pick: capture the window DIRECTLY — the image comes
+                // with the app's real rounded-corner alpha (no mask guessing).
+                if let windowID,
+                   let scWindow = content.windows.first(where: { $0.windowID == windowID }) {
+                    var windowConfig = SCStreamConfiguration()
+                    windowConfig.width = Int(scWindow.frame.width * backingScale)
+                    windowConfig.height = Int(scWindow.frame.height * backingScale)
+                            windowConfig.showsCursor = false
+                    if let windowImage = try? await SCScreenshotManager.captureImage(
+                        contentFilter: SCContentFilter(desktopIndependentWindow: scWindow),
+                        configuration: windowConfig
+                    ) {
+                        let expectedW = CGFloat(scWindow.frame.width) * backingScale
+                        let expectedH = CGFloat(scWindow.frame.height) * backingScale
+                        let overflow = max(
+                            CGFloat(windowImage.width) / max(expectedW, 1),
+                            CGFloat(windowImage.height) / max(expectedH, 1)
+                        )
+                        if overflow > 1.35 {
+                            // SCKit composited extra windows (owned popups etc.)
+                            // — reject and use the screen crop instead.
+                            CaptureController.log("capture: WINDOW id=\(windowID) overflow=\(overflow) (\(windowImage.width)x\(windowImage.height) vs \(Int(expectedW))x\(Int(expectedH))) — rejecting, using crop+mask")
+                        } else {
+                            CaptureController.log("capture: WINDOW id=\(windowID) image=\(windowImage.width)x\(windowImage.height) (native alpha corners)")
+                            completion(windowImage)
+                            return
+                        }
+                    }
+                    CaptureController.log("capture: window filter failed for id=\(windowID) — falling back to crop+mask")
+                }
+
                 guard let image = try? await SCScreenshotManager.captureImage(contentFilter: filter, configuration: config) else {
                     completion(nil)
                     return

build-app.sh

@@ -108,9 +108,9 @@ cat > "$APP/Contents/Info.plist" <<'PLIST'
     <key>CFBundleIconFile</key>
     <string>Icon</string>
     <key>CFBundleShortVersionString</key>
-    <string>1.7.1</string>
+    <string>1.8.2</string>
     <key>CFBundleVersion</key>
-    <string>71</string>
+    <string>82</string>
     <key>LSUIElement</key>
     <true/>
     <key>NSHighResolutionCapable</key>