TTD API cleanup chores (follow-up from PR #1125 review)

[xusheng6]

Follow-up chores from @plafosse's review of #1125 (approved contingent only on the std::pair<bool,…> → std::optional fix, which has been done). These are the remaining "holistic / light mechanical refactor" items. They aren't specific to the register-write code — most apply to the surrounding TTD query APIs generally — so tracking them here rather than blocking that PR.

C++

• std::pair<bool, T> → std::optional<T> everywhere. Done for the register-write methods in Add TTD go to previous/next register write (#1123) #1125; the pre-existing TTD memory-access methods (GetTTDNext/PrevMemoryAccess) still return std::pair<bool, TTDMemoryEvent> and should be migrated for consistency.
• Register-name parameters: const std::string& → std::string_view. Applies to the TTD query methods that take a register name (api/debuggerapi.h, controller, adapter).
• ParseSingleTTDRegisterWriteObject: take expression by value (std::string) and std::move it into wExpression rather than const std::string& + copy (core/adapters/dbgengttdadapter.cpp).
• Logging: avoid Log* without a logger, and prefer the Log*F variants (faster, more readable, more featureful). The new TTD code uses bare LogInfo/LogError/LogWarn with printf-style args — this pattern is repo-wide in the debugger, so worth a broader sweep.

Python (api/python/debuggercontroller.py)

• Make the event classes frozen @dataclasses instead of hand-rolled classes with manual __eq__/__hash__/__setattr__ (TTDRegisterWriteEvent, and the sibling TTDMemoryEvent / related event classes).
• Use a typedef'd register-name type instead of native str. Mirror architecture.py's RegisterName = NewType('RegisterName', str) for register-name parameters.

Review comments: #1125 (review)