builtintrap

Builtintrap is a programming construct used by several languages and their compilers to intentionally trigger a trap, an abrupt interruption of program execution. It represents a non-returning operation that transfers control to the runtime or environment in a way that is typically not catchable as a normal error. The exact effect varies by architecture and implementation, but it is generally used to indicate unreachable code, critical failures, or a deliberate abort in low-level code.

In practice, a builtin trap is usually implemented as a compiler intrinsic. In C and C++, compilers such as GCC and Clang provide a function like __builtin_trap(), which expands to an architecture-specific trap instruction. The result is an immediate termination of the program, often without performing cleanup that would occur with a standard abort. The location in code where a trap is emitted is treated as a non-returning path by the optimizer, helping it reason about control flow.

Other ecosystems describe similar capabilities under different names, often mapped to an underlying LLVM or machine instruction (for example, an intrinsic that corresponds to a trap or breakpoint on the target architecture). Some languages offer higher-level equivalents, such as panic in safe languages or explicit abort paths, but builtintrap specifically emphasizes a low-level, non-returning trap.

Use cases include asserting invariants in performance-critical code, providing a compact way to implement unreachable branches, and enabling debuggers or tools to recognize intentional termination points. However, its behavior can be platform-dependent and may produce core dumps or differ in how cleanup and signal handling are performed. Therefore, builtintrap is generally reserved for low-level, architecture-aware scenarios and not a substitute for proper error handling.

See also: trap instruction, __builtin_unreachable, abort, panic, unreachable.