Format decimal 0.0 with no precision as just 0

[frmdstryr]

Currently it prints 00 when no precision is given, This makes it just 0. See https://godbolt.org/z/4jDgNv

const std = @import("std");
pub fn main() void {
    const zero: f32 = 0.0;
    const one: f32 = 1.0;
    std.debug.warn("One is: {d}\n", .{one});
    std.debug.warn("Zero is: {d}\n", .{zero}); // Should be a single 0
}

Outputs

One is: 1
Zero is: 00

[adontz]

@andrewrk Did you consider "David M. Gay, Correctly Rounded Binary-Decimaland Decimal-Binary Conversions"?
paper: https://ampl.com/REFS/rounding.pdf
implementation: http://www.netlib.org/fp/ dtoa

[LemonBoy]

Did you consider "David M. Gay, Correctly Rounded Binary-Decimaland Decimal-Binary Conversions"?

The dtoa routine uses the Errol3 algorithm that's an improvement over the original one devised by Gay. The state of the art at the moment is the Ryū algorithm and it's supposed to be adopted by the stdlib (#1299) sometimes soon.