py/modmath: Add domain error checking to sqrt, log, log2, log10.

These functions will raise 'ValueError: math domain error' on invalid
input.
pull/1712/head
Michael Buesch 2015-12-10 13:28:37 +01:00 zatwierdzone przez Damien George
rodzic f7c4f9a640
commit 17298af61e
2 zmienionych plików z 41 dodań i 14 usunięć

Wyświetl plik

@ -25,6 +25,7 @@
*/
#include "py/builtin.h"
#include "py/nlr.h"
#if MICROPY_PY_BUILTINS_FLOAT && MICROPY_PY_MATH
@ -35,7 +36,10 @@
/// The `math` module provides some basic mathematical funtions for
/// working with floating-point numbers.
//TODO: Change macros to check for overflow and raise OverflowError or RangeError
STATIC NORETURN void math_error(void) {
nlr_raise(mp_obj_new_exception_msg_varg(&mp_type_ValueError, "math domain error"));
}
#define MATH_FUN_1(py_name, c_name) \
STATIC mp_obj_t mp_math_ ## py_name(mp_obj_t x_obj) { return mp_obj_new_float(MICROPY_FLOAT_C_FUN(c_name)(mp_obj_get_float(x_obj))); } \
STATIC MP_DEFINE_CONST_FUN_OBJ_1(mp_math_## py_name ## _obj, mp_math_ ## py_name);
@ -52,6 +56,16 @@
STATIC mp_obj_t mp_math_ ## py_name(mp_obj_t x_obj) { mp_int_t x = MICROPY_FLOAT_C_FUN(c_name)(mp_obj_get_float(x_obj)); return mp_obj_new_int(x); } \
STATIC MP_DEFINE_CONST_FUN_OBJ_1(mp_math_## py_name ## _obj, mp_math_ ## py_name);
#define MATH_FUN_1_ERRCOND(py_name, c_name, error_condition) \
STATIC mp_obj_t mp_math_ ## py_name(mp_obj_t x_obj) { \
mp_float_t x = mp_obj_get_float(x_obj); \
if (error_condition) { \
math_error(); \
} \
return mp_obj_new_float(MICROPY_FLOAT_C_FUN(c_name)(x)); \
} \
STATIC MP_DEFINE_CONST_FUN_OBJ_1(mp_math_## py_name ## _obj, mp_math_ ## py_name);
#if MP_NEED_LOG2
// 1.442695040888963407354163704 is 1/_M_LN2
#define log2(x) (log(x) * 1.442695040888963407354163704)
@ -59,7 +73,7 @@
/// \function sqrt(x)
/// Returns the square root of `x`.
MATH_FUN_1(sqrt, sqrt)
MATH_FUN_1_ERRCOND(sqrt, sqrt, (x < (mp_float_t)0.0))
/// \function pow(x, y)
/// Returns `x` to the power of `y`.
MATH_FUN_2(pow, pow)
@ -69,9 +83,9 @@ MATH_FUN_1(exp, exp)
/// \function expm1(x)
MATH_FUN_1(expm1, expm1)
/// \function log2(x)
MATH_FUN_1(log2, log2)
MATH_FUN_1_ERRCOND(log2, log2, (x <= (mp_float_t)0.0))
/// \function log10(x)
MATH_FUN_1(log10, log10)
MATH_FUN_1_ERRCOND(log10, log10, (x <= (mp_float_t)0.0))
/// \function cosh(x)
MATH_FUN_1(cosh, cosh)
/// \function sinh(x)
@ -139,11 +153,19 @@ MATH_FUN_1(lgamma, lgamma)
// log(x[, base])
STATIC mp_obj_t mp_math_log(mp_uint_t n_args, const mp_obj_t *args) {
mp_float_t l = MICROPY_FLOAT_C_FUN(log)(mp_obj_get_float(args[0]));
mp_float_t x = mp_obj_get_float(args[0]);
if (x <= (mp_float_t)0.0) {
math_error();
}
mp_float_t l = MICROPY_FLOAT_C_FUN(log)(x);
if (n_args == 1) {
return mp_obj_new_float(l);
} else {
return mp_obj_new_float(l / MICROPY_FLOAT_C_FUN(log)(mp_obj_get_float(args[1])));
mp_float_t base = mp_obj_get_float(args[1]);
if (base <= (mp_float_t)0.0) {
math_error();
}
return mp_obj_new_float(l / MICROPY_FLOAT_C_FUN(log)(base));
}
}
STATIC MP_DEFINE_CONST_FUN_OBJ_VAR_BETWEEN(mp_math_log_obj, 1, 2, mp_math_log);

Wyświetl plik

@ -9,15 +9,14 @@ except ImportError:
test_values = [-100., -1.23456, -1, -0.5, 0.0, 0.5, 1.23456, 100.]
test_values_small = [-10., -1.23456, -1, -0.5, 0.0, 0.5, 1.23456, 10.] # so we don't overflow 32-bit precision
p_test_values = [0.1, 0.5, 1.23456]
unit_range_test_values = [-1., -0.75, -0.5, -0.25, 0., 0.25, 0.5, 0.75, 1.]
functions = [('sqrt', sqrt, p_test_values),
functions = [('sqrt', sqrt, test_values),
('exp', exp, test_values_small),
('expm1', expm1, test_values_small),
('log', log, p_test_values),
('log2', log2, p_test_values),
('log10', log10, p_test_values),
('log', log, test_values),
('log2', log2, test_values),
('log10', log10, test_values),
('cosh', cosh, test_values_small),
('sinh', sinh, test_values_small),
('tanh', tanh, test_values_small),
@ -41,7 +40,10 @@ functions = [('sqrt', sqrt, p_test_values),
for function_name, function, test_vals in functions:
print(function_name)
for value in test_vals:
print("{:.5g}".format(function(value)))
try:
print("{:.5g}".format(function(value)))
except ValueError as e:
print(str(e))
tuple_functions = [('frexp', frexp, test_values),
('modf', modf, test_values),
@ -59,10 +61,13 @@ binary_functions = [('copysign', copysign, [(23., 42.), (-23., 42.), (23., -42.)
('atan2', atan2, ((1., 0.), (0., 1.), (2., 0.5), (-3., 5.), (-3., -4.),)),
('fmod', fmod, ((1., 1.), (0., 1.), (2., 0.5), (-3., 5.), (-3., -4.),)),
('ldexp', ldexp, ((1., 0), (0., 1), (2., 2), (3., -2), (-3., -4),)),
('log', log, ((2., 2.), (3., 2.), (4., 5.))),
('log', log, ((2., 2.), (3., 2.), (4., 5.), (0., 1.), (1., 0.), (-1., 1.), (1., -1.))),
]
for function_name, function, test_vals in binary_functions:
print(function_name)
for value1, value2 in test_vals:
print("{:.5g}".format(function(value1, value2)))
try:
print("{:.5g}".format(function(value1, value2)))
except ValueError as e:
print(str(e))