Standard library

nox.math

Floating-point mathematics backed by the C library's libm.

Text

import nox.math
from nox.math import sqrt, pi

Capability: libc_math — hosted programs only. For code that must run in a freestanding profile use nox.mathx, the same functions with no libc dependency.

Calling convention — read this first#

The libm functions are bound with extern def, which is never name-mangled, so they are called unqualified after the import:

Nox

import nox.math

print(sqrt(16.0), pow(2.0, 10.0), floor(2.7), ceil(2.1))

Output

4.0 1024.0 2.0 3.0

The pure-Nox helpers (ln, pi, e, min, max, abs) are ordinary functions and are called qualified: nox.math.pi(), nox.math.ln(x). from nox.math import sqrt, pi imports both kinds by name.

Functions#

Function	Description	Call as
`sqrt(x)`	square root	`sqrt(x)`
`pow(x, y)`	`x` raised to `y`	`pow(x, y)`
`floor(x)`, `ceil(x)`	round down / up to a whole number (still a `float`)	`floor(x)`
`sin(x)`, `cos(x)`, `tan(x)`	trigonometric functions (radians)	`sin(x)`
`atan2(y, x)`	the angle of the point `(x, y)`	`atan2(y, x)`
`exp(x)`	e to the power `x`	`exp(x)`
`log(x)`	natural logarithm (the libm name)	`log(x)`
`ln(x)`	natural logarithm (alias of `log`)	`nox.math.ln(x)`
`pi()`	π	`nox.math.pi()`
`e()`	Euler's number	`nox.math.e()`
`min(a, b)`, `max(a, b)`, `abs(x)`	`float` versions of the built-ins	`nox.math.min(a, b)`

All arguments and results are float. Domain errors follow IEEE-754 (sqrt(-1.0) is nan; log(0.0) is -inf) — nothing raises.

Nox

import nox.math

print(nox.math.pi() > 3.14159, nox.math.e() > 2.71828, nox.math.ln(1.0), nox.math.min(2.0, 3.0), nox.math.abs(-2.5))
print(sin(0.0), cos(0.0), atan2(1.0, 1.0) > 0.78, exp(0.0), log(1.0))

Output

True True 0.0 2.0 2.5
0.0 1.0 True 1.0 0.0

Choosing `nox.math` or `nox.mathx`#

	`nox.math`	`nox.mathx`
Backed by	the platform `libm`	Zig's own math (no libc)
Freestanding	no	yes
Call style	`sqrt(x)` (unqualified externs)	`nox.mathx.sqrt(x)` (always qualified)

Use nox.mathx for new code, and for anything that may run without a C library. Results may differ in the last bit between the two for some inputs.

Edit this page on GitHub

nox.math

Calling convention — read this first#

Functions#

Choosing nox.math or nox.mathx#

Choosing `nox.math` or `nox.mathx`#