[Lldb-commits] [lldb] [lldb][NFCI] Minor refactor to CommandObjectProcessHandle::VerifyCommandOptionValue (PR #79901)

[via lldb-commits]

jimingham wrote:

This version is still a little awkward because there's an ambiguity about what the incoming string being empty means.  In the VerifyCommandOptionValue, it means "No Value" and the return is not distinguishable from "error" - but in a bunch of the places where you use VerifyCommandOptionValue an empty string just means a value wasn't provided,  so you end up with constructs where you set the optional (checking the incoming string for `empty`) then check the input string again so you can tell empty apart from error, THEN make up the error on the outside w/o knowing what actually was wrong with the string.

I wonder if it would be better to add a Status to the call.  Then empty input -> "Empty Optional but no error"; error in parsing -> "Empty Optional but error" and result means success.  Then you wouldn't have to do two checks on the input string in different places, or make up the error string.

https://github.com/llvm/llvm-project/pull/79901