utmapp / UTM

Virtual machines for iOS and macOS

Home Page:https://getutm.app

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

UTM run in App Translocation downloaded from App Store

simbados opened this issue · comments

BEFORE SUBMITTING YOUR ISSUE, PLEASE LOOK AT THE PINNED ISSUES AND USE THE SEARCH FUNCTION TO MAKE SURE IT IS NOT ALREADY REPORTED. ALWAYS COMMENT ON AN EXISTING ISSUE INSTEAD OF MAKING A NEW ONE.

Describe the issue
Version 4.5.2 of UTM from the App Store will run in App Translocation. Why is that so?
To see that it is run in App Translocation you can run 'ps xw | grep UTM'

Configuration

  • UTM Version: 4.5.2
  • macOS Version: 14.5
  • Mac Chip (Intel, M1, ...): M3

Tried reinstalling it, but still in App Translocation.
Issue started with upgrade from 4.4.5 -> 4.5.2

Have you tried dragging UTM.app to the desktop and back to the Applications folder?

Seems odd that Translocation would be enabled for an App Store app though. Somehow the OS thinks it's a new app with questionable notarization instead of a properly signed and notarized app update.

commented

Yes I tried that and the app is no longer opened in Translocation. However if I uninstall and reinstall it, then it will start again in Translocation (tried this multiple times). Maybe I broke some config files, because I copied some VMs outside their location. But normally Translocation should not happen on apps downloaded from the App Store and I never installed UTM outside of the App Store.

commented

If anyone is interested why this happens, here is a good article explaining it: https://eclecticlight.co/2024/05/27/app-store-apps-can-be-translocated/

This is probably an unintended side effect from apples own translocation rules

In particular, see the comment thread there.

Short version: it seems like the package submitted to the App Store for v4.5.2 might have had a com.apple.quarantine EA on the app bundle. Why the backend App Store processing did not strip that (if it was really there) is an open question. But the default for productbuild(1) is to preserve any EAs that are present in the package payload, so if Apple didn't strip it, it probably would get installed. Which is what appeared to be happening with that version (maybe?) but not with v4.5.3.