# # # Nim's Runtime Library # (c) Copyright 2015 Andreas Rumpf # # See the file "copying.txt", included in this # distribution, for details about the copyright. # ## Floating-point environment. Handling of floating-point rounding and ## exceptions (overflow, division by zero, etc.). ## The types, vars and procs are bindings for the C standard library ## [](https://en.cppreference.com/w/c/numeric/fenv) header. when defined(posix) and not defined(genode): {.passl: "-lm".} var FE_DIVBYZERO* {.importc, header: "".}: cint ## division by zero FE_INEXACT* {.importc, header: "".}: cint ## inexact result FE_INVALID* {.importc, header: "".}: cint ## invalid operation FE_OVERFLOW* {.importc, header: "".}: cint ## result not representable due to overflow FE_UNDERFLOW* {.importc, header: "".}: cint ## result not representable due to underflow FE_ALL_EXCEPT* {.importc, header: "".}: cint ## bitwise OR of all supported exceptions FE_DOWNWARD* {.importc, header: "".}: cint ## round toward -Inf FE_TONEAREST* {.importc, header: "".}: cint ## round to nearest FE_TOWARDZERO* {.importc, header: "".}: cint ## round toward 0 FE_UPWARD* {.importc, header: "".}: cint ## round toward +Inf FE_DFL_ENV* {.importc, header: "".}: cint ## macro of type pointer to `fenv_t` to be used as the argument ## to functions taking an argument of type `fenv_t`; in this ## case the default environment will be used type Tfenv* {.importc: "fenv_t", header: "", final, pure.} = object ## Represents the entire floating-point environment. The ## floating-point environment refers collectively to any ## floating-point status flags and control modes supported ## by the implementation. Tfexcept* {.importc: "fexcept_t", header: "", final, pure.} = object ## Represents the floating-point status flags collectively, ## including any status the implementation associates with the ## flags. A floating-point status flag is a system variable ## whose value is set (but never cleared) when a floating-point ## exception is raised, which occurs as a side effect of ## exceptional floating-point arithmetic to provide auxiliary ## information. A floating-point control mode is a system variable ## whose value may be set by the user to affect the subsequent ## behavior of floating-point arithmetic. proc feclearexcept*(excepts: cint): cint {.importc, header: "".} ## Clear the supported exceptions represented by `excepts`. proc fegetexceptflag*(flagp: ptr Tfexcept, excepts: cint): cint {. importc, header: "".} ## Store implementation-defined representation of the exception flags ## indicated by `excepts` in the object pointed to by `flagp`. proc feraiseexcept*(excepts: cint): cint {.importc, header: "".} ## Raise the supported exceptions represented by `excepts`. proc fesetexceptflag*(flagp: ptr Tfexcept, excepts: cint): cint {. importc, header: "".} ## Set complete status for exceptions indicated by `excepts` according to ## the representation in the object pointed to by `flagp`. proc fetestexcept*(excepts: cint): cint {.importc, header: "".} ## Determine which of subset of the exceptions specified by `excepts` are ## currently set. proc fegetround*(): cint {.importc, header: "".} ## Get current rounding direction. proc fesetround*(roundingDirection: cint): cint {.importc, header: "".} ## Establish the rounding direction represented by `roundingDirection`. proc fegetenv*(envp: ptr Tfenv): cint {.importc, header: "".} ## Store the current floating-point environment in the object pointed ## to by `envp`. proc feholdexcept*(envp: ptr Tfenv): cint {.importc, header: "".} ## Save the current environment in the object pointed to by `envp`, clear ## exception flags and install a non-stop mode (if available) for all ## exceptions. proc fesetenv*(a1: ptr Tfenv): cint {.importc, header: "".} ## Establish the floating-point environment represented by the object ## pointed to by `envp`. proc feupdateenv*(envp: ptr Tfenv): cint {.importc, header: "".} ## Save current exceptions in temporary storage, install environment ## represented by object pointed to by `envp` and raise exceptions ## according to saved exceptions. const FLT_RADIX = 2 ## the radix of the exponent representation FLT_MANT_DIG = 24 ## the number of base FLT_RADIX digits in the mantissa part of a float FLT_DIG = 6 ## the number of digits of precision of a float FLT_MIN_EXP = -125 ## the minimum value of base FLT_RADIX in the exponent part of a float FLT_MAX_EXP = 128 ## the maximum value of base FLT_RADIX in the exponent part of a float FLT_MIN_10_EXP = -37 ## the minimum value in base 10 of the exponent part of a float FLT_MAX_10_EXP = 38 ## the maximum value in base 10 of the exponent part of a float FLT_MIN = 1.17549435e-38'f32 ## the minimum value of a float FLT_MAX = 3.40282347e+38'f32 ## the maximum value of a float FLT_EPSILON = 1.19209290e-07'f32 ## the difference between 1 and the least value greater than 1 of a float DBL_MANT_DIG = 53 ## the number of base FLT_RADIX digits in the mantissa part of a double DBL_DIG = 15 ## the number of digits of precision of a double DBL_MIN_EXP = -1021 ## the minimum value of base FLT_RADIX in the exponent part of a double DBL_MAX_EXP = 1024 ## the maximum value of base FLT_RADIX in the exponent part of a double DBL_MIN_10_EXP = -307 ## the minimum value in base 10 of the exponent part of a double DBL_MAX_10_EXP = 308 ## the maximum value in base 10 of the exponent part of a double DBL_MIN = 2.2250738585072014E-308 ## the minimal value of a double DBL_MAX = 1.7976931348623157E+308 ## the minimal value of a double DBL_EPSILON = 2.2204460492503131E-16 ## the difference between 1 and the least value greater than 1 of a double template fpRadix*: int = FLT_RADIX ## The (integer) value of the radix used to represent any floating ## point type on the architecture used to build the program. template mantissaDigits*(T: typedesc[float32]): int = FLT_MANT_DIG ## Number of digits (in base `floatingPointRadix`) in the mantissa ## of 32-bit floating-point numbers. template digits*(T: typedesc[float32]): int = FLT_DIG ## Number of decimal digits that can be represented in a ## 32-bit floating-point type without losing precision. template minExponent*(T: typedesc[float32]): int = FLT_MIN_EXP ## Minimum (negative) exponent for 32-bit floating-point numbers. template maxExponent*(T: typedesc[float32]): int = FLT_MAX_EXP ## Maximum (positive) exponent for 32-bit floating-point numbers. template min10Exponent*(T: typedesc[float32]): int = FLT_MIN_10_EXP ## Minimum (negative) exponent in base 10 for 32-bit floating-point ## numbers. template max10Exponent*(T: typedesc[float32]): int = FLT_MAX_10_EXP ## Maximum (positive) exponent in base 10 for 32-bit floating-point ## numbers. template minimumPositiveValue*(T: typedesc[float32]): float32 = FLT_MIN ## The smallest positive (nonzero) number that can be represented in a ## 32-bit floating-point type. template maximumPositiveValue*(T: typedesc[float32]): float32 = FLT_MAX ## The largest positive number that can be represented in a 32-bit ## floating-point type. template epsilon*(T: typedesc[float32]): float32 = FLT_EPSILON ## The difference between 1.0 and the smallest number greater than ## 1.0 that can be represented in a 32-bit floating-point type. template mantissaDigits*(T: typedesc[float64]): int = DBL_MANT_DIG ## Number of digits (in base `floatingPointRadix`) in the mantissa ## of 64-bit floating-point numbers. template digits*(T: typedesc[float64]): int = DBL_DIG ## Number of decimal digits that can be represented in a ## 64-bit floating-point type without losing precision. template minExponent*(T: typedesc[float64]): int = DBL_MIN_EXP ## Minimum (negative) exponent for 64-bit floating-point numbers. template maxExponent*(T: typedesc[float64]): int = DBL_MAX_EXP ## Maximum (positive) exponent for 64-bit floating-point numbers. template min10Exponent*(T: typedesc[float64]): int = DBL_MIN_10_EXP ## Minimum (negative) exponent in base 10 for 64-bit floating-point ## numbers. template max10Exponent*(T: typedesc[float64]): int = DBL_MAX_10_EXP ## Maximum (positive) exponent in base 10 for 64-bit floating-point ## numbers. template minimumPositiveValue*(T: typedesc[float64]): float64 = DBL_MIN ## The smallest positive (nonzero) number that can be represented in a ## 64-bit floating-point type. template maximumPositiveValue*(T: typedesc[float64]): float64 = DBL_MAX ## The largest positive number that can be represented in a 64-bit ## floating-point type. template epsilon*(T: typedesc[float64]): float64 = DBL_EPSILON ## The difference between 1.0 and the smallest number greater than ## 1.0 that can be represented in a 64-bit floating-point type.