There were 2 small issues.
1. ValueObjectVTableChild was not fixing the addresses it was pulling
from signed pointers. This broke things like `SBValue::GetLoadAddress`
and identifying the function pointer type from debug info.
2. TestVTableValue.py made a lot of assumptions that did not hold on
arm64e. a. GetValueAsUnsigned will return a raw pointer value. Most of
the time, we needed GetValueAsAddress. b. The test was reading pointers
out of memory without fixing them up. c. The summary for a function
pointer on arm64e includes the load address. This isn't true on other
platforms.
Alex Langford
0e0127e5b2