[Lldb-commits] [lldb] [lldb] Change ValueObject::AddressOf() to return Expected (NFC) (PR #106831)

[via lldb-commits]

jimingham wrote:

This is more an architectural issue.  Right now, we're always handing out ValueObjectSP's which means there are two ways of indicating failure, having an empty ValueObjectSP, and having a ValueObjectSP that has a ValueObject in it, but it's Error state is set.  That's pretty awkward at times, and we should come up with a better scheme than that, but I'm not sure what that is without thinking more about it.  

That doesn't seem so much like a new contributor task.  

However, there are lots of other API's in lldb that fit the pattern Adrian referred to that don't relate to ValueObjects.  Those are likely to be much less problematic.

https://github.com/llvm/llvm-project/pull/106831