Dear community, Recently, I’ve been trying to debug an arbitrary precision algorithms/formulae. I stumbled upon to the fact that it is by default users are allowed to have mixed precision input without any warnings, e.g. big(1.3) 1.3000000000000000444089209850062616169452667236328125 gives you …

You can just redefine the constructor: function BigFloat(x::Float64) throw(error("Imprecise.")) end

Disable promotion from Float64 to BigFloat

GunnarFarneback June 2, 2025, 5:28pm 21

While all of this is correct, for debugging purposes it’s perfectly fine to make a dirty redefinition of a constructor if that helps you pin down mistakes in your code.

5 Likes

Topic		Replies	Views
Promotion to BigFloat -- no zero padding General Usage bigfloat	8	786	August 9, 2021
A question regarding BigFloat General Usage question	11	1937	November 5, 2018
BigFloat converted to string returns different number? Numerics numbers , bigfloat	8	619	August 28, 2022
Handling large values correctly General Usage precision , bigfloat	4	306	July 27, 2024
BigFloat number conversion problem New to Julia	15	1338	January 21, 2019

Disable promotion from Float64 to BigFloat

Related topics