sin and tan

[nunull]

sin and tan don't seem to work correctly. sin(pi) gives 1.2246467991473515E-16 and tan(pi) -1.2246467991473515E-16, while cos works as expected.

[alfredxing]

I've noticed this too, but it's likely a problem with Go itself, not with calc: https://play.golang.org/p/c2-ZAgTXmt

[rpopplewell]

Float64 has about 15 significant digits, so why not truncate?: http://play.golang.org/p/rmdSC3197k

[Rufflewind]

The utility uses floating-point numbers, so this is working as intended.

… sin(pi) gives 1.2246467991473515E-16 …

sin(pi) in floating-point arithmetic is not the same as sin(π) in math. What sin(pi) calculates is the approximate value of sine at the approximate value pi. Since pi is not equal to the true π (in fact, it's off by about 1.22E-16 … does this number look familiar?), that imprecision leads to an imprecision in the value of sin(pi).

…, so why not truncate?: http://play.golang.org/p/rmdSC3197k

Float64 has 15-18 significant digits, not decimal places. 1.22464679914735E-16 truncated to 15 significant digits is still 1.22464679914735E-16.