<?php

declare( strict_types=1 );

namespace Rowbot\Idna;

use Normalizer;
use Rowbot\Punycode\Exception\PunycodeException;
use Rowbot\Punycode\Punycode;

use function array_merge;
use function count;
use function explode;
use function implode;
use function preg_match;
use function strlen;
use function strncmp;
use function substr;

/**
 * @see https://www.unicode.org/reports/tr46/
 */
final class Idna {
	public const UNICODE_VERSION = '14.0.0';

	private const DEFAULT_UNICODE_OPTIONS = [
		'CheckHyphens'            => true,
		'CheckBidi'               => true,
		'CheckJoiners'            => true,
		'UseSTD3ASCIIRules'       => true,
		'Transitional_Processing' => false,
	];
	private const DEFAULT_ASCII_OPTIONS = self::DEFAULT_UNICODE_OPTIONS + [
		'VerifyDnsLength' => true,
	];

	public const ERROR_EMPTY_LABEL = 1;
	public const ERROR_LABEL_TOO_LONG = 2;
	public const ERROR_DOMAIN_NAME_TOO_LONG = 4;
	public const ERROR_LEADING_HYPHEN = 8;
	public const ERROR_TRAILING_HYPHEN = 0x10;
	public const ERROR_HYPHEN_3_4 = 0x20;
	public const ERROR_LEADING_COMBINING_MARK = 0x40;
	public const ERROR_DISALLOWED = 0x80;
	public const ERROR_PUNYCODE = 0x100;
	public const ERROR_LABEL_HAS_DOT = 0x200;
	public const ERROR_INVALID_ACE_LABEL = 0x400;
	public const ERROR_BIDI = 0x800;
	public const ERROR_CONTEXTJ = 0x1000;
	public const ERROR_CONTEXTO_PUNCTUATION = 0x2000;
	public const ERROR_CONTEXTO_DIGITS = 0x4000;

	private const MAX_DOMAIN_SIZE = 253;
	private const MAX_LABEL_SIZE = 63;

	/**
	 * @codeCoverageIgnore
	 */
	private function __construct() {
	}

	/**
	 * @see https://www.unicode.org/reports/tr46/#ProcessingStepMap
	 *
	 * @param  array<string, bool>  $options
	 */
	private static function mapCodePoints( string $domain, array $options, DomainInfo $info ): string {
		$str                   = '';
		$useSTD3ASCIIRules     = $options['UseSTD3ASCIIRules'];
		$transitional          = $options['Transitional_Processing'];
		$errors                = 0;
		$transitionalDifferent = false;

		foreach ( CodePoint::utf8Decode( $domain ) as $codePoint ) {
			$data = CodePointStatus::lookup( $codePoint, $useSTD3ASCIIRules );

			switch ( $data['status'] ) {
				case 'disallowed':
					$errors |= Idna::ERROR_DISALLOWED;

				// no break.

				case 'valid':
					$str .= CodePoint::encode( $codePoint );

					break;

				case 'ignored':
					// Do nothing.
					break;

				case 'mapped':
					$str .= $data['mapping'];

					break;

				case 'deviation':
					$transitionalDifferent = true;
					$str                   .= ( $transitional ? $data['mapping'] : CodePoint::encode( $codePoint ) );

					break;
			}
		}

		$info->addError( $errors );

		if ( $transitionalDifferent ) {
			$info->setTransitionalDifferent();
		}

		return $str;
	}

	/**
	 * @see https://www.unicode.org/reports/tr46/#Processing
	 *
	 * @param  array<string, bool>  $options
	 *
	 * @return list<string>
	 */
	private static function process( string $domain, array $options, DomainInfo $info ): array {
		// If VerifyDnsLength is not set, we are doing ToUnicode otherwise we are doing ToASCII and
		// we need to respect the VerifyDnsLength option.
		$checkForEmptyLabels = ! isset( $options['VerifyDnsLength'] ) || $options['VerifyDnsLength'];

		if ( $checkForEmptyLabels && $domain === '' ) {
			$info->addError( self::ERROR_EMPTY_LABEL );

			return [ '' ];
		}

		// Step 1. Map each code point in the domain name string
		$domain = self::mapCodePoints( $domain, $options, $info );

		// Step 2. Normalize the domain name string to Unicode Normalization Form C.
		if ( ! Normalizer::isNormalized( $domain, Normalizer::FORM_C ) ) {
			$originalDomain = $domain;
			$domain         = Normalizer::normalize( $domain, Normalizer::FORM_C );

			// This shouldn't be possible since the input string is run through the UTF-8 decoder
			// when mapping the code points above, but lets account for it anyway.
			if ( $domain === false ) {
				$info->addError( self::ERROR_INVALID_ACE_LABEL );
				$domain = $originalDomain;
			}

			unset( $originalDomain );
		}

		// Step 3. Break the string into labels at U+002E (.) FULL STOP.
		$labels         = explode( '.', $domain );
		$lastLabelIndex = count( $labels ) - 1;
		$validator      = new LabelValidator( $info );

		// Step 4. Convert and validate each label in the domain name string.
		foreach ( $labels as $i => $label ) {
			$validationOptions = $options;

			if ( strncmp( $label, 'xn--', 4 ) === 0 ) {
				try {
					$label = Punycode::decode( substr( $label, 4 ) );
				} catch ( PunycodeException $e ) {
					$info->addError( self::ERROR_PUNYCODE );

					continue;
				}

				$validationOptions['Transitional_Processing'] = false;
				$labels[ $i ]                                 = $label;
			}

			$validator->validate( $label, $validationOptions, $i > 0 && $i === $lastLabelIndex );
		}

		if ( $info->isBidiDomain() && ! $info->isValidBidiDomain() ) {
			$info->addError( self::ERROR_BIDI );
		}

		// Any input domain name string that does not record an error has been successfully
		// processed according to this specification. Conversely, if an input domain_name string
		// causes an error, then the processing of the input domain_name string fails. Determining
		// what to do with error input is up to the caller, and not in the scope of this document.
		return $labels;
	}

	/**
	 * @see https://www.unicode.org/reports/tr46/#ToASCII
	 *
	 * @param  array<string, bool>  $options
	 */
	public static function toAscii( string $domain, array $options = [] ): IdnaResult {
		$options = array_merge( self::DEFAULT_ASCII_OPTIONS, $options );
		$info    = new DomainInfo();
		$labels  = self::process( $domain, $options, $info );

		foreach ( $labels as $i => $label ) {
			// Only convert labels to punycode that contain non-ASCII code points
			if ( preg_match( '/[^\x00-\x7F]/', $label ) === 1 ) {
				try {
					$label = 'xn--' . Punycode::encode( $label );
				} catch ( PunycodeException $e ) {
					$info->addError( self::ERROR_PUNYCODE );
				}

				$labels[ $i ] = $label;
			}
		}

		if ( $options['VerifyDnsLength'] ) {
			self::validateDomainAndLabelLength( $labels, $info );
		}

		return new IdnaResult( implode( '.', $labels ), $info );
	}

	/**
	 * @see https://www.unicode.org/reports/tr46/#ToUnicode
	 *
	 * @param  array<string, bool>  $options
	 */
	public static function toUnicode( string $domain, array $options = [] ): IdnaResult {
		// VerifyDnsLength is not a valid option for toUnicode, so remove it if it exists.
		unset( $options['VerifyDnsLength'] );
		$options = array_merge( self::DEFAULT_UNICODE_OPTIONS, $options );
		$info    = new DomainInfo();
		$labels  = self::process( $domain, $options, $info );

		return new IdnaResult( implode( '.', $labels ), $info );
	}

	/**
	 * @param  list<string>  $labels
	 */
	private static function validateDomainAndLabelLength( array $labels, DomainInfo $info ): void {
		$maxDomainSize = self::MAX_DOMAIN_SIZE;
		$length        = count( $labels );
		$totalLength   = $length - 1;

		// If the last label is empty and it is not the first label, then it is the root label.
		// Increase the max size by 1, making it 254, to account for the root label's "."
		// delimiter. This also means we don't need to check the last label's length for being too
		// long.
		if ( $length > 1 && $labels[ $length - 1 ] === '' ) {
			++ $maxDomainSize;
			-- $length;
		}

		for ( $i = 0; $i < $length; ++ $i ) {
			$bytes       = strlen( $labels[ $i ] );
			$totalLength += $bytes;

			if ( $bytes > self::MAX_LABEL_SIZE ) {
				$info->addError( self::ERROR_LABEL_TOO_LONG );
			}
		}

		if ( $totalLength > $maxDomainSize ) {
			$info->addError( self::ERROR_DOMAIN_NAME_TOO_LONG );
		}
	}
}