Post not yet marked as solved
Can NSXPCInterface work with non-void return type?
The official document says, "All messages must be 'void' return type." (Reference)
However, in the file provider sample code of wwdc21, it uses it with non-void return type.
(Reference)
I wondered whether it can or can not be used with non-void return type?
Is the answer changed after swift support for async/await?
Post not yet marked as solved
how to obtain information about transmission errors such as overrun, framing, parity, and
break signals？
Post not yet marked as solved
How can I exchange information easily and securely between 2 apps on macOS?
1 of the app will infrequently request a short amount of data from the other one.
Here are the options that I can see:
DistributedNotificationCenter : very easy to implement. However, the notifications are broadcast to any apps that wants to listen to it, and apple's documentation clearly states that it's not secure. I would ideally like a mechanism that is as simple as this, but with a secure communication between 2 aps
Apple Events. I am not sure how to make an app respond to apple event. And I think it would be the same problem : any other app could talk to these 2 apps and get information from them. I ideally want this to be more secure.
XPC, this seems overly complex for what I want to do, as my understanding is that this essentially involves creating a third process that will regulate communication.
Implementing IPC via sockets. It seems reasonable I think, as I can probably secure the communication better. However it feels over-enigneered to set up sockets that always listen for incoming connections, etc
Am I missing a simple mechanism on macOS that could help me in that use case?
Or am I looking at things incorrectly for one of these options?
Thanks!
Post not yet marked as solved
Hi,
We are building an sandbox enabled app which contains
1. One launch daemon
2. One launch agent
3. System extension which is contained in launch agent
The launch daemon is outside of the sandbox. The launch agent and system extension is inside the sandbox.
The launch agent is in good communicating with system daemon already. But recently the use case I am meeting is to comunnication between launch agent, launch daemon and systen extension daemon.
1. Launch agent sends request to launch daemon
2. launch daemon sends response to launch agent
And
1. System extension sends request to launch daemon
2. Launch daemon sends response to system extension
I have read some articles on the forum and understand that we can use machServices to make the XPC connection fulfilled in launch daemon.
And we can use com.apple.security.temporary-exception.mach-lookup.global-name to eliminate the sandbox limitation between the daemon and agents.
But when we do experiment, it always return
Error Domain=NSCocoaErrorDomain Code=4097 "connection to service named com.*****.******" UserInfo={NSDebugDescription=connection to service named com.****.*****}
I understand that the suggested debug method is making anonymous listener in the same process. But that looks like more for XPC service. I am not sure how to debug
in the launchd. Is there any suggestion?
BTW, the daemon is mainly implemented in C++ and the agent is in swift. So I use NSXPCConnection on both sides. I am wondering if it is the best fit for our purpose.
Is there any good example that I can follow?
Post not yet marked as solved
I discovered that an app I downloaded for a desktop clock had been starting at login automatically. When I changed that selection so it wasn't starting at login, I started getting this error. When I changed it back to "start at login" I stopped getting this error. I hope this is helpful.
Post not yet marked as solved
XPC is the preferred inter-process communication (IPC) mechanism on Apple platforms. XPC has two APIs:
The high-level NSXPCConnection API, for Objective-C and Swift
The low-level C API, which, while callable from all languages, works best with C-based languages
General:
DevForums tag: XPC
NSXPCConnection class documentation
XPC C API documentation
XPC has extensive man pages — For the C API, start with the xpc man page; this is the original source for the XPC C API documentation and still contains tidbits that you can’t find elsewhere. Also read the xpcservice.plist man page, which documents the property list format used by XPC services.
Daemons and Services Programming Guide archived documentation
Technote 2083 Daemons and Agents — It hasn’t been updated in… well… decades, but it’s still remarkably relevant.
TN3113 Testing and Debugging XPC Code With an Anonymous Listener
Related tags include:
Inter-process communication, for other IPC mechanisms
Service Management, for installing and uninstalling Service Management login items, launchd agents, and launchd daemons
Share and Enjoy
—
Quinn “The Eskimo!” @ Developer Technical Support @ Apple
let myEmail = "eskimo" + "1" + "@" + "apple.com"
Is it better to
Create an NSXPCConnection, keep it around, and create proxies using that as needed, or
Create an NSXPCConnection, create a proxy off of it, and then close the connection when done?
Post not yet marked as solved
I've configured an item and placeholder in NSFileProviderExtension iOS extension, and an XPC service in the provider (based on the template for XPC service for Mac, but as part of the NSFileProviderExtension
When connecting to the service from the app using code example in getFileProviderServicesForItem I am getting an error straight in the completion handler of that function:
Error Domain=NSCocoaErrorDomain Code=4097 "Error while sending identifierForItemAtURL:completionHandler:" UserInfo={NSDebugDescription=Error while sending identifierForItemAtURL:completionHandler:, NSUnderlyingError=0x2833640c0 {Error Domain=NSCocoaErrorDomain Code=4097 "connection from pid 30324 on anonymousListener or serviceListener" UserInfo={NSDebugDescription=connection from pid 30324 on anonymousListener or serviceListener}}}
Could you maybe suggest what I am missing? Or, is there an example somewhere of the FileProvider extension with the service being called from the app?
Code:
File Provider service definition: https://github.com/simplex-chat/simplex-chat/blob/af3dcc4a9a9b24751bf9d74af67cf8e7d119597a/apps/ios/SimpleX%20Service/SimpleXFPService.swift
Application code that calls the service: https://github.com/simplex-chat/simplex-chat/blob/af3dcc4a9a9b24751bf9d74af67cf8e7d119597a/apps/ios/Shared/FPService.swift
Thank you!
Post not yet marked as solved
I thought Swift wasn't supposed to get them, which is part of the reason why I chose to use it for my network extension. But we're getting crashes occasionally, that look like:
Thread 4 Crashed:: Dispatch queue: com.apple.NSXPCConnection.user.endpoint
0 com.kithrup.MyApp.NExt 0x102c4ffe2 MyExt.sendData(_:data:completion:) + 610
1 com.kithrup.MyApp.NExt 0x102c5091f @objc MyExt.sendData(_:data:completion:) + 255
2 Foundation 0x7ff81ef97490 __NSXPCCONNECTION_IS_CALLING_OUT_TO_EXPORTED_OBJECT_S3__ + 10
3 Foundation 0x7ff81ef3fa1f -[NSXPCConnection _decodeAndInvokeMessageWithEvent:flags:] + 2322
4 Foundation 0x7ff81eef641e message_handler + 206
5 libxpc.dylib 0x7ff81de24b6c _xpc_connection_call_event_handler + 56
6 libxpc.dylib 0x7ff81de23947 _xpc_connection_mach_event + 1382
7 libdispatch.dylib 0x7ff81df2e3b1 _dispatch_client_callout4 + 9
8 libdispatch.dylib 0x7ff81df47041 _dispatch_mach_msg_invoke + 445
9 libdispatch.dylib 0x7ff81df341cd _dispatch_lane_serial_drain + 342
10 libdispatch.dylib 0x7ff81df47b77 _dispatch_mach_invoke + 484
11 libdispatch.dylib 0x7ff81df341cd _dispatch_lane_serial_drain + 342
12 libdispatch.dylib 0x7ff81df34e30 _dispatch_lane_invoke + 417
13 libdispatch.dylib 0x7ff81df3eeee _dispatch_workloop_worker_thread + 753
14 libsystem_pthread.dylib 0x7ff81e0e1fd0 _pthread_wqthread + 326
The XPC method is func sendData(_: UUID, data: Data?, completion: @escaping (_: Error?) -> Void)
It's crashing on address 0x10, so pretty clearly a NULL-dereference.
Since this is happening in my extension, it's in Swift (as I said above), so I have no idea what could be NULL without the compiler yelling at me first.
Post not yet marked as solved
Hello
We are developing our own iOS Network Extensions-based VPN and it has an HTTP proxy in the VPN.
In addition, we also use PAC (Proxy auto-configuration) script to configure what kind of HTTP/HTTPS traffic should route to our proxy in the VPN.
However, we get this kind of message "Received XPC error Connection invalid for message type 3 kCFNetworkAgentXPCMessageTypePACQuery" randomly on iOS 15.5.
We have not been aware of any weird behavior of iOS based on the error message. We are afraid of this error message is caused by our VPN solution.
Is there any suggestion that should consider or follow to fix this error?
Post not yet marked as solved
In my project I have a host application and a husk application.
What I want to do is, every time the user launches a new view in the host application, I want to create one more dock icon by launching another instance of husk application. Then the husk application behaves like it is the view itself by monitoring the event of click/quit and send them to the host application through XPC.
The XPC tutorial tells me an XPC service embedded in an application is invisible to the processes outside the bundle. To communicate between two foreground applications it seems that I need to create a third helperTool/agent/daemon which venders a Mach/XPC service.
But I wonder if I can put husk application inside the bundle of the hose application. So they can directly connect to the XPC service which is also embedded in the same bundle.
If the answer is no, maybe NSDistributedNotificationCenter is much better and simpler in my scenario?
Post not yet marked as solved
I built an app which hosts a CMIOExtension. The app works, and it can activate the extension. The extension loads in e.g. Photo Booth and shows the expected video (a white horizontal line which moves down the picture).
I have a couple of questions about this though.
The sample Camera Extension is built with a CMIOExtension dictionary with just one entry, CMIOExtensionMachServiceName which is $(TeamIdentifierPrefix)$(PRODUCT_BUNDLE_IDENTIFIER)
This Mach service name won't work though. When attempting to activate the extension, sysextd says that the extensions has an invalid mach service name or is not signed, the value must be prefixed with one of the App Groups in the entitlement.
So in order to get the sample extension to activate from my app, I have to change its CMIOExtensionMachServiceName to
<my team ID>.com.mycompany.my-app-group.<myextensionname>
Is this to be expected?
The template CMIOExtension generates its own video using a timer. My app is intended to capture video from a source, filter that video, then feed it to the CMIOExtension, somehow. The template creates an app group called "$(TeamIdentifierPrefix)com.example.app-group", which suggests that it might be possible to use XPC to send frames from the app to the extension.
However, I've been unable to do so. I've used
NSXPCConnection * connection = [[NSXPCConnection alloc] initWithMachServiceName:, using the CMIOExtensionMachServiceName with no options and with the NSXPCConnectionPrivileged option. I've tried NSXPCConnection * connection = [[NSXPCConnection alloc] initWithServiceName: using the extension's bundle identifier. In all cases when I send the first message I get an error in the remote object proxy's handler:
Error Domain=NSCocoaErrorDomain Code=4099 "The connection to service named <whatever name I try> was invalidated: failed at lookup with error 3 - No such process."
According to the "Daemons and Services Programming Guide" an XPC service should have a CFBundlePackageType of XPC!, but a CMIOExtension is of type SYSX. It can't be both.
Does the CMIOExtension loading apparatus cook up a synthetic name for the XPC service, and if so, what is it? If none, how is one expected to get pixel buffers into the camera extension?
Post not yet marked as solved
I have an application which is doing screen recording, now I move the screen recording feature to a standalone native XPC module for better performance due to some reason that the app is tied an old lib which cannot generate native code for M1 (Intel only).
My question is that, this new xpc module is belong to the App (demanded by the app), if I give the screen recording permission to the app, will the xpc screen scraping module be granted to the permission?
Right now looks like it is not after I granted the application with the screen recording permission since display stream won't produce the frame data.
Post not yet marked as solved
We'd prefer our security application not be worked around by the complex task of typing sudo launchctl unload /Library/LaunchDaemons/foo.plist 😄. Is there a way to prevent that? (We're not using ServiceManagement because we need ot control some of the plist entries, sadly.)
Post not yet marked as solved
XPC connection keeps getting interrupted.
I'm creating an xpc endpoint in FxPlug plugin for FCP X using xpc_endpoint_create.
This endpoint is then passed to a helper mach service running in the background and stored there.
Next, our main application is launched and retrieves the stored endpoint from the helper service.
It creates the communication channel using xpc_connection_create_from_endpoint
The main application communicates with FxPlug plugin using that endpoint.
It all works well when I am debugging either our application or FxPlug.
The moment I use the release build on both, the connection works fine for a while but is very quickly interrupted (usually 2-10 seconds), FxPlug plugin gets flagged as non-responsive and is unloaded by FCP X. This behavior is erratic and may cease after some time on some machines.
We've been working on this and some other issues with FxPlug team for months and some changes have been made, but we're stuck with that one last bit.
I want to stress the following: when I use a debug version of either plugin or our app, everything works fine, fxplug is never unloaded or marked as unresponsive, the connection is stable. When both components are using release builds, it all comes apart for no apparent reason.
Both plugin and application can normally recover and reconnect after being unloaded and restored.
Any thoughts on why an xpc connection would be interrupted in this way?
Post not yet marked as solved
Hi,
I’d like to perform client-side certificate authentication from https based connection in macOS.
I’m using the method didReceiveChallenge from URLSession. However, I cannot read the keychain directly since my process is running as Daemon, and my client certificate reside in login keychain.
So I've followed the guidance from this question https://developer.apple.com/forums/thread/106851, and sent this authentication request to a user-based process which is running in the current user so it has access to the keychain.
After I acquire the NSURLCredential object, I’d like to return it back to the Daemon, so it may run the completionHandler with that credential.
However, After I successfully create the NSURLCredential in the user process, and send it back using some reply callback. It looks like the object didn’t serialized properly and I get the following error :
Exception: decodeObjectForKey: Object of class "NSURLCredential" returned nil from -initWithCoder: while being decoded for key <no key>
Here’s my client side code ( I made sure that the server side create a valid NSURLCredential object).
and the problem occur after I send the XPC request, right when i’m about to get the callback response (reply)
- (void)URLSession:(NSURLSession *)session
didReceiveChallenge:(NSURLAuthenticationChallenge *)challenge
completionHandler:(void (^)(NSURLSessionAuthChallengeDisposition disposition, NSURLCredential *credential))completionHandler {
if (challenge.protectionSpace.authenticationMethod == NSURLAuthenticationMethodClientCertificate) {
[myXpcService getCertIdentityWithAcceptedIssuers:challenge.protectionSpace.distinguishedNames
withReply:^(NSURLCredential *cred, NSError *error) {
if (error != nil) {
completionHandler(NSURLSessionAuthChallengeCancelAuthenticationChallenge, nil);
} else {
completionHandler(NSURLSessionAuthChallengeUseCredential, cred);
}
}];
}
Perhaps anybody can tell me what did I do wrong here ? Does XPC is capable to pass complex objects like NSURLCredentials ?
thanks !
Post not yet marked as solved
I am trying to pass an array of C-structs to an XPC Service, but the service receives only the first element. Following is the C struct
struct MyStruct
{
char *name;
unsigned char v1;
unsigned char v2;
Status status; // a c-style enum
};
and I am using it like this
struct MyStruct structs[3] = {{"name1", 0, 0, success}, {"name2", 1,1, success}, {0}};
[[_connectionToService remoteObjectProxy] doSomething:structs];
and doSomething is declared as
- (void)doSomething: (struct MyStruct[]) structs;
The document Creating XPC Services mentions that C structures and arrays containing only the types listed above are supported, but I am unable to get it to work even for an array of c-strings or ints.
Also, there's an API for making synchronous RPC calls but there is no documentation available for it.
- (id)synchronousRemoteObjectProxyWithErrorHandler:(void (^)(NSError *error))handler
It does seem to block but only if the remote method has a reply block. Is this the expected behaviour? And is it safe to cache the proxy object returned by this method?
I have an app with the following simple architecture:
Main App: A regular macOS app bundle with UI that allows users to customize settings of the app
Helper: Another macOS app bundle with no UI (LSUIElement=1 in Info.plist) that is packaged inside the main app in the LoginItems directory doing the core tasks of the app in the background
My requirements are:
Distribution via the MAS (=sandbox enabled for both targets)
Both apps should be able to communicate via XPC
The main app should be closable by the user at any time, should not keep running after being closed, whereas the helper app should as it performs actions for which it needs to be kept running in the background
Launch-on-login of the helper app should not (and according to 2.4.5 (iii) of the ASRG must not) happen automatically w/o user consent and therefore I assume should always be a checkbox optional to the user
For sharing settings changed by the user in the main app with the helper too, I've added the Application Group capability to both targets to allow usage of a common user defaults suite.
While that works fine, there's the requirement that from within the main app I'd also need to request information and call a method from the background process (bidirectional communication) which is where I'm currently stuck.
I understand that an XPC Service (.xpc) would not be suitable for the helper here because it is automatically terminated when the parent app dies and may also not be suitable for my use cases as the helper needs to be able to request Screen Capture permissions from the user and I doubt this is possible for XPC bundles. I also understand that an XPC service which utilizes a mach-service XPC listener will only work in a sandboxed environment through the use of Service Management's SMLoginItemSetEnabled() API.
My main issue here is that the mandatory requirement to leave the option to launch the helper on login open to the user conflicts with the requirement of being able to communicate with the helper via XPC any time the main app is open, regardless of user choices.
If there wasn't the requirement to sandbox both apps, I would solve this issue with a launchd user agent that is kept alive but only runs at load if the user checked the launch-on-login box in the Settings of the main app. With sandbox enabled though, I'm currently launching the helper app manually if launch-on-login is disabled and let the Service Management API handle the lifecycle if it is enabled. For the first case, I haven't been able to establish an XPC connection w/o calling SMLoginItemSetEnabled() and I assume that is by design.
Is there something obvious I've missed here as I kinda feel like this is a typical app setup many other 3rd party devs are having as well?
Post not yet marked as solved
Hi,
Greetings for the day!
We would like to update you that we have created Content Filter NetworkExtension and this extension is working fine till Big Sur M1 however we are facing some strange problem in M1 Monterey. Intermittently, When we try to browse websites, it does not respond and after 3-5 minutes its opened the websites correctly.
We would like to update you that our subclass overrides handleNewFlow, handleInboundDataFromFlow, handleOutboundDataFromFlow, handleInboundDataCompleteForFlow and handleOutboundDataCompleteForFlow. In all these methods we first check whether NEFilterFlow is nil or not and then pauseVerdict and once asynchronous methods completes execution then we call resumeFlow with verdict (allowVerdict/dropVerdict).
When above mentioned issue generated we collected console streaming log and found these lines in the logs (Not from our application):
Ignoring resume command for flow 3c8faf3c4a9f7 which does not exist
Ignoring resume command for flow 3c90795d4d6f9 which does not exist
Ignoring resume command for flow 3c9086d1ede69 which does not exist
Ignoring resume command for flow 3c909b251d53b which does not exist
We are not sure how above line get printed because we don’t have this logs in our source code so we would need your help to understand this problem and resolution so that we can solve this issue.
We have couple of extra queries:
What is flow mentioned in above logs in bold text? Is it NEFilterFlow's identifier or something else?
How we can validate whether NEFilterFlow is valid or not before calling resumeFlow
Why above line is getting printed in log which says flow does not exist.
Is there any timeout maintained by NetworkExtension?
We are using XPC for interprocess communication so our question is that, Is NetworkExtension/XPC maintain the queue size and if it overflow the size then above line is getting printed. If this is the case then how we can handle that?
Is it known issue in NetworkExtension framework itself on M1 Monterey?
Thanks & Regards,
Mohmad Vasim
Post not yet marked as solved
I am building a somewhat-VPN-like system. I have a daemon that handles the networking and a couple of agents that interact with the logged-in user. The daemon and agent communicate via XPC.
I am trying to get this to work "smoothly" with fast-user-switching. Empirically, I find that the daemon can correlate XPC connection from different agents in the different login sessions via xpc_connection_get_asid(), which appears to be equivalent to the security session id, though it's not clear this this equivalence is always they case, nor will always be the case in the future. If I had a way to get the security session id for a pid, I would use xpc_connection_get_pid()
When the daemon gets a network connection I want to find the pid the connection is coming from (it can only be from the local machine) and figure out which security session that pid belongs to so that I can direct any necessary user interaction and permission checking to the agent in the security session from which the network request is coming. Finding the pid from the TCP port is arduous and inefficient, but doable. However, once I've found the pid, I don't know how to determine the security session id (or audit session id) that that pid is part of.
GetSessionInfo appears to permit me to get information about my own session id including my session id, or additional info about another session, if I already have that session id.
For my purposes, a viable alternative to being able to get the security session for a pid might be to be able to ask the question "Is this pid part of this security session?" since I expect to have agents in a small number of security sessions.
Getting the username of the pid is obviously doable, however this would mean that I would be unable to determine where a sudo'ed process was coming from (at least not without walking up the parent heirarchy). This feels dicey.
Am I missing something?
