fix(toml): parse() duplicates the character next to reserved escape sequences

[babiabeo]

Fix: #4187

[kt3k]

There are 2 cases that I expected:

• Ignores the reserved escape sequence:

parse('a = "a\\0b"'); // { a: "ab" }

• Throws an error as specified in the spec:

@babiabeo
Do you think ignoring is more useful than throwing?

[iuioiua]

I think we should stick to the spec and throw.

[babiabeo]

There are 2 cases that I expected:

• Ignores the reserved escape sequence:

parse('a = "a\\0b"'); // { a: "ab" }

• Throws an error as specified in the spec:

@babiabeo Do you think ignoring is more useful than throwing?

Currently, this PR solves the issue by ignoring reserved escape sequences.

However, I saw that most TOML parser packages throw an error in this case

• js-toml

import { load } from "npm:js-toml";

console.log(load('a = "a\\0b"'));

// error: Uncaught (in promise) SyntaxParseError: Syntax error
// unexpected character: ->"<- at offset: 4, skipped 3 characters.
// unexpected character: ->b<- at offset: 8, skipped 2 characters.

• tomllib

import tomllib

print(tomllib.loads('a = "a\\0b"'))

# File "D:\coding\programming\languages\python\Lib\tomllib\_parser.py", line 494, in parse_basic_str_escape
#     raise suffixed_err(src, pos, "Unescaped '\\' in a string") from None
# tomllib.TOMLDecodeError: Unescaped '\' in a string (at line 1, column 9)

• Go toml

package main

import (
	"fmt"

	"github.com/BurntSushi/toml"
)

type Data struct {
    A string `toml:"a"`
}

func main() {
    var data Data
    _, err := toml.Decode("a = \"a\\0b\"", &data)

    fmt.Println(err)

    // toml: line 1 (last key "a"): invalid escape in string '\0'
}

[iuioiua]

I think it best if we instead threw instead of ignoring, as to follow the spec. Are you able to rework this PR to do that?

[iuioiua]

@babiabeo, I've made some tweaks to improve the thrown error message and accuracy of the tests. Do these changes seem reasonable to you?

[babiabeo]

@babiabeo, I've made some tweaks to improve the thrown error message and accuracy of the tests. Do these changes seem reasonable to you?

Yes. I forgot to add \\. Thanks for adding it!

[iuioiua]

LGTM! Thank you, David.