[BUG] consecutive_digit_separator has no effect

[dubrowgn]

Description

First, thanks for the great library!

Perhaps I am doing this wrong, or misunderstanding what the flag is supposed to do, but it appears that consecutive_digit_separator has no effect on the results of parsing. Setting it to false still allows multiple consecutive digit separators in the source string.

Prerequisites

• Rust version: rustc 1.68.0 (2c8cc3432 2023-03-06)
• lexical version:
  lexical-util v0.8.5
  lexical-parse-integer v0.8.6
  lexical-parse-float v0.8.5
  lexical-core v0.8.5
  lexical v6.1.1
• Cargo.toml

[dependencies.lexical]
default-features = false
features = [ "parse-floats", "parse-integers" ]
version = "6.x"

[dependencies.lexical-parse-integer]
default-features = false
features = [ "format", "power-of-two" ]
version = "x"

[dependencies.lexical-parse-float]
default-features = false
features = [ "format", "power-of-two" ]
version = "x"

Test case

let opts = lexical_parse_float::Options::new();
const FMT: u128 = NumberFormatBuilder::new()
	.digit_separator(num::NonZeroU8::new(b'_'))
	.leading_digit_separator(true)
	.internal_digit_separator(true)
	.trailing_digit_separator(true)
	.consecutive_digit_separator(false)
	.build();
println!("{:?}", lexical::parse_with_options::<f64, _, FMT>("__1__234__", &opts));

// Ok(1234.0)

[Alexhuszagh]

Confirmed this affects both integer and floating point parsing, it seems like this corner case was never tested in our API tests, only the positive case. It seems there's something wrong with the peek_1! macro.