noinline

Noinline is a directive used in several programming languages to prevent the compiler from inlining a function or lambda. Inlining is the optimization that substitutes a function call with the body of the function, reducing call overhead and enabling further optimizations. When noinline is applied, the call remains a distinct call, preserving function boundaries, which can help with debugging, function pointers, and code layout, but may reduce certain performance benefits or increase code size.

In Kotlin, noinline is a modifier applied to a lambda parameter of an inline function to prevent that lambda from being inlined. This is necessary when the lambda is stored, passed as a value, or used in a way that must preserve its identity. Example usage is inline fun perform(noinline action: () -> Unit) { /* ... */ }.

In Rust, inlining is controlled at the function level with attributes. The attribute #[inline(never)] applied to a function disables inlining for that function. Example: #[inline(never)] fn compute(...) { ... } This can be useful to keep a stable call boundary for debugging or to avoid code size growth from aggressive inlining.

In C and C++, compilers such as GCC and Clang provide the noinline attribute to prevent inlining. It can be applied to a function declaration or definition, for example: void process(...) __attribute__((noinline)); or __attribute__((noinline)) void process(...); This helps manage performance characteristics, code size, and interactions with linker or profiling tools.

Noinline behavior is not portable across languages; syntax and semantics vary by compiler and language. It is commonly used to control trade-offs between speed, binary size, and debugging fidelity, or to preserve behavior that depends on distinct function boundaries.