Possible bug: f128 constant only has 16 decimals of precision when printed

Using 0.5.0 the following code:

var z : f128 = 1.738954882048422195749048390299;
warn("{:.34}\n",z);

Shows this:

1.7389548820484222000000000000000000e+00

It's like the compile time constant is 64 bit rather than 128 bit. The documentation implies that floating constants are 128 bit.

Is this a bug? Or am I doing something wrong?

This is because the current float formatter internally works on 64bit floats. #1299 should fix this.

dupe of #1181?

std.fmt: add ryu floating-point formatting

Navigation Menu