In the code below, A key is remapped to B key, and vice versa. The remapping is activated via a SwiftUI toggle switch.
In example presented here the same block of code is used in three different functions.
Additionally, the loop that iterates through the function call is also used in all three of these functions.
I've been struggling to simplify this code and make it less redundant for more than a day. Any help would be greatly appreciated.
let aKey: UInt64 = 0x700000004
let bKey: UInt64 = 0x700000005
func isKeyboardServiceClientForUsagePage(_ serviceClient: IOHIDServiceClient, _ usagePage: UInt32, _ usage: UInt32) -> Bool {
return IOHIDServiceClientConformsTo(serviceClient, usagePage, usage) == 1
}
func updateKeyboardKeyMapping(_ keyMap: [[String: UInt64]]) {
let eventSystemClient = IOHIDEventSystemClientCreateSimpleClient(kCFAllocatorDefault)
guard let serviceClients = IOHIDEventSystemClientCopyServices(eventSystemClient) as? [IOHIDServiceClient] else {
return
}
for serviceClient in serviceClients {
let usagePage = UInt32(kHIDPage_GenericDesktop)
let usage = UInt32(kHIDUsage_GD_Keyboard)
if isKeyboardServiceClientForUsagePage(serviceClient, usagePage, usage) {
IOHIDServiceClientSetProperty(serviceClient, kIOHIDUserKeyUsageMapKey as CFString, keyMap as CFArray)
}
}
}
func areKeysMappedOnAnyServiceClient() -> Bool {
let eventSystemClient = IOHIDEventSystemClientCreateSimpleClient(kCFAllocatorDefault)
guard let serviceClients = IOHIDEventSystemClientCopyServices(eventSystemClient) as? [IOHIDServiceClient] else {
return false
}
for serviceClient in serviceClients {
let usagePage = UInt32(kHIDPage_GenericDesktop)
let usage = UInt32(kHIDUsage_GD_Keyboard)
if isKeyboardServiceClientForUsagePage(serviceClient, usagePage, usage) {
guard let keyMapping = IOHIDServiceClientCopyProperty(serviceClient, kIOHIDUserKeyUsageMapKey as CFString) as? [[String: UInt64]] else {
return false
}
if keyMapping.contains(where: { $0[kIOHIDKeyboardModifierMappingSrcKey] == aKey && $0[kIOHIDKeyboardModifierMappingDstKey] == bKey }) &&
keyMapping.contains(where: { $0[kIOHIDKeyboardModifierMappingSrcKey] == bKey && $0[kIOHIDKeyboardModifierMappingDstKey] == aKey })
{
return true
}
}
}
return false
}
func remapABBA() {
let keyMap: [[String: UInt64]] = [
[
kIOHIDKeyboardModifierMappingSrcKey: aKey,
kIOHIDKeyboardModifierMappingDstKey: bKey,
],
[
kIOHIDKeyboardModifierMappingSrcKey: bKey,
kIOHIDKeyboardModifierMappingDstKey: aKey,
],
]
updateKeyboardKeyMapping(keyMap)
}
func resetKeyMapping() {
updateKeyboardKeyMapping([])
}
And here’s the SwiftUI part if you would like to try the app:
import SwiftUI
struct ContentView: View {
@State private var remapKeys = areKeysMappedOnAnyServiceClient()
var body: some View {
HStack {
Spacer()
Toggle(isOn: $remapKeys, label: { Text("Remap A → B and B → A.") })
.toggleStyle(SwitchToggleStyle())
.onChange(of: remapKeys, perform: toggleKeyboardRemapping)
Spacer()
}
}
}
private func toggleKeyboardRemapping(_ remapKeys: Bool) {
if remapKeys {
remapABBA()
} else {
resetKeyMapping()
}
}
OK... this is going to take some time to answer.
It seems like you're lacking in a place to store things. That's why you have to use the same block of code over and over. We can solve that with a view model...
In here I'm going to hide away the logic of what is happening from the view and only expose what the view needs access to in order to display itself.
I'm going to make extensions of
IOHIDServiceClientandIOHIDEventSystemClientRefto encapsulate your logic...Doing all of this means that your view can look something like this...
This contains all your same logic and TBH wasn't too bad already.
My main changes were to take the free functions and vars and add them to their respective types.
So, the
updateandareKeysMapped...functions now belong to theIOHIDEventSystemClientReftype.The
isForGDKeyboardandkeyMappingvars now belong to theIOHIDServiceClienttype.Doing this removed a lot of the repeated code you had as you no longer had to continuously call free functions. It also meant we unlocked some very Swifty
keyPathusage which helped make some of the logic more concise.Then we made a view model. This allowed us to keep all the moving parts of the view in one place. It had a place to easily get hold of the client. It also meant we could hide a lot of the stuff inside the view model by making it private.
This meant that the view only had one thing it could do. Which is to use the binding to the
toggleState. Everything else was behind closed doors to the view.