@eskimo can you confirm that the bug in popen has been fixed? We're considering removing all popen calls right now.

Thanks for your help. It was an application crash during normal operation, not after wake. I wasn't able to extract any meaningful date from the timestamp. I've tried it with mach_timebase_info_data_t tb;   if (mach_timebase_info(&tb) == KERN_SUCCESS)     cout << 514306295821 * tb.numer / ( NSEC_PER_MSEC * tb.denom);

you don't have to leave it on Legacy. You can switch it back and it will still work. Only seems to matter that this setting changes at some point.

As always, thank you very much. I really don't know what Apple or rather all of us developers would do without you. I've read the pig-jato example, just wanted to make sure it transfers to a more complicated structure.

thanks for the link, that's very helpful! Changing the identifier also solved a problem we had with TCC. When the helper binary wasn't run from inside the bundle (e.g. when debugging in Xcode, it started the binary from outside the bundle (copying into bundle is a copy phase)) it constantly prompted us for allowance, although it was already added. That seems to be fixed now as well with the unique identifier.

Is this still the case or has anything changed?

Thanks Quinn, I've filed an enhancement request (FB9476062).

Thank you Quinn for your answer and for providing additional context, that's very helpful!