https://github.com/DavidSpickett commented:
> This also prevents a new process from aliasing stale entries when an > allocator reuses a RegisterType address. I do not understand how this could happen. I think there is a RegisterTypeBuilder per Target object, and the RegisterType addresses are created in ProcessGDBRemote. So you'd have to swap out the ProcessGDBRemote while keeping the same Target object with the same RegisterTypeBuilder. Though the fact that I had to think about it this much does suggest that this is all far too complex with too many assumptions. Tell me if I'm misunderstanding that or you were thinking of another scenario. But either way, it may be that a object UID is just overall simpler to reason about. https://github.com/llvm/llvm-project/pull/215874 _______________________________________________ lldb-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
