Warning C26475

  • Article

Do not use function style C-casts.

C++ Core Guidelines: ES.49: If you must use a cast, use a named cast

Function-style casts (for example, int(1.1)) are another form of C-style casts (like (int)1.1), which have questionable safety. Specifically, the compiler doesn't try to check if any data loss can occur either in C-casts or in function casts. In both cases, it's better either to avoid casting or to use a braced initializer if possible. If neither works, static casts may be suitable, but it's still better to use utilities from the Guidelines Support Library:

  • gsl::narrow ensures lossless conversion and throws gsl::narrowing_error if it's not possible.
  • gsl::narrow_cast clearly states that conversion can lose data and it's acceptable.

Remarks

  • This rule fires only for constants of primitive types. The compiler can clearly detect data loss in these cases and emits an error if a braced initializer is used. The cases that would require run-time execution are flagged by C26493 NO_CSTYLE_CAST.

  • Default initializers aren't flagged (for example int()).

Example

Dangerous conversion example:

constexpr auto planck_constant = float( 6.62607004082e-34 ); // C26475

Compiler error for dangerous conversion, detecting potential data loss:

constexpr auto planck_constant = float{ 6.62607004082e-34 }; // Error C2397

To correct the dangerous conversion, use an appropriately sized primitive type:

constexpr auto planck_constant = double{ 6.62607004082e-34 };