Unreliable simulate-location with iOS 17

[filmackay]

Is anyone else having set work fine, but clear not do anything?

[doronz88]

I just tested this again and this worked fine. Also, when closing the context-manager from the CLI it also reset my location to the original one.

[filmackay]

OK that's weird. This is was entirely reproducible or me.

I can (sometimes) get it to work - but when I "clear" (oslogs look like the correct action is being taken at least) my location is very inaccurate for a substantial period of time. Do you see any of these symptoms as well?

[doronz88]

No, I haven't encountered them at least in my setup scenario. I also tried to write a small snippet that sets and then resets the location immediately and it also worked just as expected

[filmackay]

OK - I just tried again (admittedly on head this time) and it works. I am using the USB-C cable that came with a iPhone 15 Pro - as opposed to another USB-C "charge cable" (still Apple, I think).

Is this a factor (charge cable v data cable?). Should I expect if I can see oslogs over the cable that everything should work; or is my different experiences explained with the different cables perhaps?

I also get some oslog errors being reported (stack trace) - should I put a PR up to resolve this, it crashes resonably frequently for me. Checking if there is some background here at all :)

[doronz88]

The cable isn't supposed to effect anything. If you discovered a bug and you know of a fix then please by any means submit a PR

[filmackay]

I'll put up a PR for the oslog crash (crashes trying to parse out a formatted message), but I've managed to eliminate the cable (and other factors). Essentially, the set/clear cycle seems to work perfectly on occasion - but others (maybe 60% of the time) the set works but the clear does not (by clear I mean the 'reset' after the set completes).

I am on macOS, presuming you're testing on that as well..

I'll prepare a contrast of oslog for the fail/success case - anything else I should look at in particular? One other symptom I get: sometimes on sudo pymobiledevice3 remote start-quic-tunnel I get ERROR Device is not connected. This persists for a while (minutes) but then starts working again.. 🤷🏻

I see three cases on clearing the location (back to original):

1. nothing happens (stays at simulated location, even though logs indicate it is not simulated any more)
2. returns to non-simulated location, with precise location
3. returns to non-simulated location, but very imprecise location (large circle of KM's) - and takes many minutes for the precision to return to normal.

[doronz88]

Yes, I'm testing on macOS also. Just making sure you know the oslog issue is completely unrelated, though I would appreciate any PRs to make this project better.

What macOS version are you using specifically? I know the monterey USB ethernet driver is somewhat unpredicatable.

Also, I'm not sure all the scenarios you are encountering aren't bugs on the device-end. Do you have other references you can compare it to? I'll be able to sniff the diff in the protocol messages.

[filmackay]

Yes, I'm testing on macOS also. Just making sure you know the oslog issue is completely unrelated, though I would appreciate any PRs to make this project better.

Yes definitely unrelated - it's just a formatting issue where it fails the pattern match and there is no 'message' to communicate.

What macOS version are you using specifically? I know the monterey USB ethernet driver is somewhat unpredicatable.

I was on Monterey, just switched to Sonoma. I've only attempted once so far, and it did work - but the 'clear' returned my location to low-precision (suburb level) physical locale. But it did come good after while..

Also, I'm not sure all the scenarios you are encountering aren't bugs on the device-end. Do you have other references you can compare it to? I'll be able to sniff the diff in the protocol messages.

Other devices? It is a iPhone 15 Pro on my side, I could try something else..

[doronz88]

I cannot find reasons why this happens, nor am I sure this is not Apple's own bug. If you can point at another scenario where the issue is fixed I can sniff it and compare.

[c00ni]

I presume you meant clear actually puts the phone in an unable to find current location mode (grey circle in google maps) centred around the simulated location.

iOS can have "trouble" finding your real location again if you've been simulating a location far away from your real location.
To speed it up significantly, set the location to somewhere close to your real location first and then reset.

This pre-dates iOS 17 (and many previous versions as well).

[Fidetro]

I found out that clear doesn't work properly in this case.

1. sudo python3 -m pymobiledevice3 remote start-quic-tunnel
2. pymobiledevice3 developer dvt simulate-location set --rsd host port -- 60 60
3. Disconnect iPhone by unplugging the cable;
4. sudo python3 -m pymobiledevice3 remote start-quic-tunnel
5. pymobiledevice3 developer dvt simulate-location clear --rsd host port,it doesn't work.

but you still can set new location.

[JoeUchihaa]

In my case, any command without dvt doesnt work, for example:

I can simulate location using:

pymobiledevice3 developer dvt simulate-location set --rsd host port -- x y

but not : pymobiledevice3 developer simulate-location set --rsd host port -- x y

similarly, I cannot simulate motion, speed etc. because there is no dvt option for the .gpx file:

sudo pymobiledevice3 developer simulate-location play "path/to/gpx/file" --rsd host port

any fix for this or a workaround?? I'm on iOS 17 btw.

[c00ni]

In my case, any command without dvt doesnt work
any fix for this or a workaround?? I'm on iOS 17 btw.

Why are you seeking a fix or workaround? You need dvt for things to work on iOS 17 due to the changes made.

[JoeUchihaa]

In my case, any command without dvt doesnt work
any fix for this or a workaround?? I'm on iOS 17 btw.

Why are you seeking a fix or workaround? You need dvt for things to work on iOS 17 due to the changes made.

the thing is play gpx doesnt work either, so i was just pointing that out

[omggodlike123]

any plans for support fake location for ios 17 on windows? @doronz88

[doronz88]

@lindaamanda123 see #569

[Dat-Bois]

the thing is play gpx doesnt work either, so i was just pointing that out

I am also interested if this feature can be implemented sometime soon for IOS 17. Since simulate-location has been implemented already, I hope this shouldn't be too large of an undertaking. No worries if you are busy.

[doronz88]

Oh the issue here is the GPX file specifically? It's harder to keep track of different issues in the same one. Can you please create another one for that specifically so I can also tag it in a PR?

Also, most of pymobile subcommands provide an abstraction to the way the protocol messages are consrtructed. In iOS17 the simulate location works very differently over DVT instead of its own relay in the DDI.

[Dat-Bois]

Also, most of pymobile subcommands provide an abstraction to the way the protocol messages are consrtructed. In iOS17 the simulate location works very differently over DVT instead of its own relay in the DDI.

I would love to learn more about how this project works and the DVT and DDI protocols you just mentioned, any resources you can suggest?