[Lldb-commits] [lldb] [lldb] Fix reading 32-bit signed integers (PR #169150)

[David Spickett via lldb-commits]

https://github.com/DavidSpickett commented:

Overall looks good, the fixes are quite obvious in hindsight.

I know this seems redundant, but do we have coverage for 64-bit signed too?

Now would be a good time to add that. To show that reading 64-bit signed into int64_t gives -1, and reading 32-bit signed into int64_t also gives -1, *because* of the correctly applied sign extension.

Unlikely we'd ever break the 64-bit ones but it's cheap to check it here and it's a hint toward what the tests expect to see for 32-bit.

Your example is on Arm 32-bit but I think this applies to any target as long as the requested signed int is 32-bit. Did I understand correctly?

https://github.com/llvm/llvm-project/pull/169150