What

high_performance_async_writer::write() in src/impl/async/high_performance_async_writer.cpp:79-86 calls to_string() three times per log entry — for message, file, and function — creating three heap allocations on every log call.

Why

• Logging is on the hot path of every system in the kcenon ecosystem
• Three std::string allocations per message adds measurable overhead under high throughput
• The downstream batch_entry constructor could accept std::string_view for zero-copy

Where

• src/impl/async/high_performance_async_writer.cpp:79-86

How

Acceptance Criteria

• batch_entry accepts std::string_view parameters
• write() path avoids unnecessary string copies
• Benchmark shows reduced allocation count per log message