<?php
/**
 * File walk by batches: stack of folders relative to ABSPATH plus the position
 * inside the current folder, both kept in ahcl_state between two requests.
 *
 * Symbolic links are not followed: a linked folder is listed as not traversed,
 * a linked file is counted as unreadable (both in the "Not looked at" section).
 *
 * @package Timaub\AfterHackChecklist
 */

namespace Timaub\AfterHackChecklist\WordPress;

use Timaub\AfterHackChecklist\Domain\Assumptions;
use Timaub\AfterHackChecklist\Domain\Context;
use Timaub\AfterHackChecklist\Domain\FileEntry;
use Timaub\AfterHackChecklist\Domain\FileInspector;
use Timaub\AfterHackChecklist\Domain\PassAccumulator;

defined( 'ABSPATH' ) || exit;

/**
 * File walk in batches.
 */
final class Walker {

	/** Initial walk state: the whole WordPress folder. */
	public static function initial() {
		return array(
			'stack'  => array( '' ),
			'dir'    => null,
			'offset' => 0,
			'files'  => 0,
		);
	}

	/**
	 * Walks until the deadline (at least one file) or the end.
	 *
	 * @param array           $walk     Walk state, updated.
	 * @param Context         $ctx      Context.
	 * @param PassAccumulator $acc      Accumulator.
	 * @param float           $deadline microtime(true) limit.
	 * @param array           $batch    Collected items for ahcl_batch_collected, appended.
	 * @return bool True when the walk is over.
	 */
	public static function run( array &$walk, Context $ctx, PassAccumulator $acc, $deadline, array &$batch ) {
		$root      = untrailingslashit( Site::abspath() );
		$processed = 0;
		$listeners = has_action( 'ahcl_batch_collected' );
		while ( true ) {
			if ( null === $walk['dir'] ) {
				if ( ! $walk['stack'] ) {
					return true;
				}
				$walk['dir']    = array_pop( $walk['stack'] );
				$walk['offset'] = 0;
			}
			$dir     = $walk['dir'];
			$abs     = '' === $dir ? $root : $root . '/' . $dir;
			$entries = is_readable( $abs ) ? scandir( $abs ) : false;
			if ( false === $entries ) {
				$acc->addNotTraversed( '' === $dir ? '/' : $dir );
				$walk['dir'] = null;
				continue;
			}
			$entries = array_values( array_diff( $entries, array( '.', '..' ) ) );
			$count   = count( $entries );
			for ( $i = $walk['offset']; $i < $count; $i++ ) {
				if ( $processed > 0 && microtime( true ) >= $deadline ) {
					$walk['offset'] = $i;
					return false;
				}
				$name = $entries[ $i ];
				$rel  = '' === $dir ? $name : $dir . '/' . $name;
				$full = $abs . '/' . $name;
				if ( is_link( $full ) ) {
					if ( is_dir( $full ) ) {
						$acc->addNotTraversed( $rel );
					} else {
						$acc->addFile(
							FileEntry::fromArray(
								array(
									'path'     => $rel,
									'readable' => false,
								)
							),
							$ctx
						);
					}
					continue;
				}
				if ( is_dir( $full ) ) {
					$walk['stack'][] = $rel;
					continue;
				}
				if ( ! is_file( $full ) ) {
					continue;
				}
				$entry = self::read( $full, $rel, $ctx );
				$acc->addFile( $entry, $ctx );
				++$processed;
				++$walk['files'];
				if ( $listeners ) {
					$batch[] = self::batchItem( $entry, $full );
				}
			}
			$walk['dir'] = null;
		}
	}

	/**
	 * Item of ahcl_batch_collected for one file: the md5 already read, else computed here,
	 * except above Assumptions::HASH_MAX_BYTES, where md5 stays null (D-05 holds with a listener).
	 *
	 * @param FileEntry $entry File as read.
	 * @param string    $full  Absolute path.
	 * @return array{path:string, md5:?string, size:int}
	 */
	private static function batchItem( FileEntry $entry, $full ) {
		$md5 = $entry->md5;
		if ( null === $md5 && $entry->readable && $entry->size <= Assumptions::HASH_MAX_BYTES ) {
			$md5 = md5_file( $full );
		}
		return array(
			'path' => $entry->path,
			'md5'  => false === $md5 ? null : $md5,
			'size' => $entry->size,
		);
	}

	/**
	 * Reads what the domain asks for this file (FileInspector::needs).
	 *
	 * @param string  $full Absolute path.
	 * @param string  $rel  Path relative to ABSPATH.
	 * @param Context $ctx  Context.
	 * @return FileEntry
	 */
	private static function read( $full, $rel, Context $ctx ) {
		$size  = (int) filesize( $full );
		$mtime = filemtime( $full );
		$a     = array(
			'path'     => $rel,
			'size'     => $size,
			'readable' => is_readable( $full ),
			'mtime'    => false === $mtime ? null : $mtime,
		);
		if ( ! $a['readable'] ) {
			return FileEntry::fromArray( $a );
		}
		$needs = FileInspector::needs( $rel, $size, $ctx );
		if ( $needs['md5'] ) {
			$md5 = md5_file( $full );
			if ( false === $md5 ) {
				$a['readable'] = false;
				return FileEntry::fromArray( $a );
			}
			$a['md5'] = $md5;
		}
		if ( $needs['sha256'] ) {
			$sha = hash_file( 'sha256', $full );
			if ( false !== $sha ) {
				$a['sha256'] = $sha;
			}
		}
		if ( $needs['content'] && $size <= Assumptions::CONTENT_MAX_BYTES ) {
			$content = file_get_contents( $full ); // phpcs:ignore WordPress.WP.AlternativeFunctions.file_get_contents_file_get_contents -- local file of the site, read only.
			if ( false === $content ) {
				$a['readable'] = false;
				return FileEntry::fromArray( $a );
			}
			$a['content'] = $content;
		} elseif ( $needs['head'] ) {
			$head = file_get_contents( $full, false, null, 0, Assumptions::HEAD_BYTES ); // phpcs:ignore WordPress.WP.AlternativeFunctions.file_get_contents_file_get_contents -- local file of the site, read only.
			if ( false === $head ) {
				$a['readable'] = false;
				return FileEntry::fromArray( $a );
			}
			$a['head'] = $head;
		}
		return FileEntry::fromArray( $a );
	}
}