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

@@ -14,9 +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 hoverWindow: (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)] = []
+    private var trackedWindows: [(id: CGWindowID, frame: NSRect, pid: Int)] = []
     private var lastWindowScan = Date.distantPast
     private var eventTap: CFMachPort?
 
@@ -105,13 +105,17 @@ final class CaptureController: NSObject {
 
     /// 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) {
+    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, windowID: windowID)
+            if ownedRisk {
+                capture(rect, excludingWindowIDs: ids, roundedCorners: true)
+            } else {
+                capture(rect, excludingWindowIDs: ids, windowID: windowID)
+            }
         }
     }
 
@@ -219,11 +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 }
@@ -235,7 +240,7 @@ final class CaptureController: NSObject {
             )
             guard cgRect.width >= 80, cgRect.height >= 60 else { continue }
             let rect = Self.cgToAppKit(rect: cgRect)
-            found.append((cid, rect))
+            found.append((id: cid, frame: rect, pid: pid))
         }
         trackedWindows = found
     }
@@ -248,7 +253,7 @@ final class CaptureController: NSObject {
         return NSRect(x: rect.origin.x, y: y, width: rect.width, height: rect.height)
     }
 
-    private func topmostWindowContaining(_ appKitPoint: NSPoint) -> (id: CGWindowID, frame: NSRect)? {
+    private func topmostWindowContaining(_ appKitPoint: NSPoint) -> (id: CGWindowID, frame: NSRect, pid: Int)? {
         for entry in trackedWindows {
             if entry.frame.contains(appKitPoint) {
                 return entry
@@ -287,6 +292,11 @@ final class CaptureController: NSObject {
                 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
@@ -316,8 +326,14 @@ final class CaptureController: NSObject {
                 CaptureController.log("tap: marquee complete \(rect)")
                 finishSelection(rect)
             } else if let hover = hoverWindow {
-                CaptureController.log("tap: window pick id=\(hover.id) \(hover.frame)")
-                finishWindowPick(hover.id, hover.frame)
+                // 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()
@@ -463,9 +479,21 @@ final class ScreenCapture {
                         contentFilter: SCContentFilter(desktopIndependentWindow: scWindow),
                         configuration: windowConfig
                     ) {
-                        CaptureController.log("capture: WINDOW id=\(windowID) image=\(windowImage.width)x\(windowImage.height) (native alpha corners)")
-                        completion(windowImage)
-                        return
+                        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")
                 }

build-app.sh

@@ -108,9 +108,9 @@ cat > "$APP/Contents/Info.plist" <<'PLIST'
     <key>CFBundleIconFile</key>
     <string>Icon</string>
     <key>CFBundleShortVersionString</key>
-    <string>1.8.0</string>
+    <string>1.8.2</string>
     <key>CFBundleVersion</key>
-    <string>80</string>
+    <string>82</string>
     <key>LSUIElement</key>
     <true/>
     <key>NSHighResolutionCapable</key>