Full MiniMessage support

[HaHaWTH]

This PR brings full support of MiniMessage to AuthMe.

Current TODOs:

• Gradient working
• HoverEvents working
• Legacy format char compatibility
• Version compatibility (from 1.8.8 to latest) (Tested versions: 1.12.2, 1.21)
• Support for kick message
• Support for deny login message
• Support for chat message

[TheFloodDragon]

Here offers a more effective mothod (Written in Kotlin):

import net.kyori.adventure.text.serializer.legacy.LegacyComponentSerializer.AMPERSAND_CHAR
import net.kyori.adventure.text.serializer.legacy.LegacyComponentSerializer.SECTION_CHAR

fun convertLegacyToMiniMessage(source: String, vararg tagResolver: TagResolver): Component =
    LegacyComponentSerializer.deserialize(translateAmpersandColor(mark(source)))
            .let { MiniMessageSerializer.serialize(it) }
            .let { MiniMessageSerializer.deserialize(deMark(it), *tagResolver) }

private fun mark(source: String) =
   source.replaceNonEscaped("<", "{marked:start}").replaceNonEscaped(">", "{marked:end}")

private fun deMark(source: String) =
    source.replaceNonEscaped("{marked:start}", "<").replaceNonEscaped("{marked:end}", ">")

/**
 * 将 '&' 转换成 '§'
 */
fun translateAmpersandColor(target: String) = target.replace(AMPERSAND_CHAR, SECTION_CHAR)

/**
 * 替换非转义字符
 * @see String.replace
 */
@JvmOverloads
fun String.replaceNonEscaped(
    oldValue: String,
    newValue: String,
    ignoreCase: Boolean = false,
    startIndex: Int = 0,
    escapeChar: String = ESCAPE_CHAR,
): String = buildString {
    // 索引记录
    var lastIndex = 0
    // 匹配字符串
    allIndexOf(oldValue, startIndex, ignoreCase) { index ->
        // 从上次找到的位置到当前找到的位置之前的字符串
        val segment = this@replaceNonEscaped.substring(lastIndex, index)
        // 检查转义字符串
        if (this@replaceNonEscaped.startsWith(escapeChar, index - escapeChar.length))
            append(segment.dropLast(escapeChar.length)).append(oldValue)
        else append(segment).append(newValue)
        // 更新索引
        lastIndex = index + oldValue.length
    }
    append(this@replaceNonEscaped.substring(lastIndex)) // 尾处理
}

[TheFloodDragon]

具体逻辑：

• 提前替换 "<" ">"，因为直接用LegacyComponentSerializer的话 "<" ">" 会被转义，为避免转义对其预先处理，替换成过渡字符
• 此时已经得到了已经通过旧版解析的Componet，用MiniMessageSerializer再将其序列化成字符串，相当于说，把旧版的形式转换成了MiniMessage的形式
• 然后将过渡字符还原成 "<" ">"，再交给MiniMessageSerializer，就得到最终成品了

ps: 宽松点的话，replaceNonEscaped 可以不要