juju-php/msgfmt.class.php

<?php
	namespace Juju;

	class MsgFmt {
		public static function from(string $popath){
			if(!file_exists($popath)){
				throw new \Exception("PO file {$popath} does not exist.");
			}
			$hash = [];
			$temp = [];
			$state = null;
			$fuzzy = false;
			// iterate over lines
			foreach(explode(
				"\n",
				str_replace(
					["\r\n", "\r"],
					["\n", "\n"],
					file_get_contents($popath)
				)
			) as $line){
				$line = trim($line);
				if ($line !== ''){
					$exp = explode(' ', $line, 2);
					$key = $exp[0];
					if(isset($exp[1])){
						$data = $exp[1];
					}
					switch($key){
						case '#,': // flag...
							$fuzzy = in_array('fuzzy', preg_split('/,\s*/', $data));
						case '#': // translator-comments
						case '#.': // extracted-comments
						case '#:': // reference...
						case '#|': // msgid previous-untranslated-string
							// start a new entry
							if(sizeof($temp) && array_key_exists('msgid', $temp) && array_key_exists('msgstr', $temp)){
								if (!$fuzzy){
									$hash[] = $temp;
								}
								$temp = [];
								$state = null;
								$fuzzy = false;
							}
						break;
						case 'msgctxt': // context
						case 'msgid': // untranslated-string
						case 'msgid_plural': // untranslated-string-plural
							$state = $key;
							$temp[$state] = $data;
						break;
						case 'msgstr': // translated-string
							$state= 'msgstr';
							$temp[$state][] = $data;
						break;
						default:
							if(strpos($key, 'msgstr[') !== false){ // translated-string-case-n
								$state= 'msgstr';
								$temp[$state][] = $data;
							}else{ // continued lines
								switch($state){
									case 'msgctxt':
									case 'msgid':
									case 'msgid_plural':
										$temp[$state] .= "\n" . $line;
									break;
									case 'msgstr':
										$temp[$state][sizeof($temp[$state]) - 1] .= "\n" . $line;
									break;
									default: // parse error
										return false;
								}
							}
						break;
					}
				}
			}
			// add final entry
			if ($state === 'msgstr'){
				$hash[] = $temp;
			}
			// Cleanup data, merge multiline entries, reindex hash for ksort
			$temp= $hash;
			$hash= [];
			foreach($temp as $entry){
				foreach($entry as &$value){
					$value = self::clean($value);
					if($value === false){ // parse error
						return false;
					}
				}
				$hash[$entry['msgid']] = $entry;
			}
			return $hash;
		}
		public static function to(string $mopath, array $hash){
			// sort by msgid
			ksort($hash, SORT_STRING);
			// our mo file data
			$mo = '';
			// header data
			$offsets = [];
			$ids = '';
			$strings = '';
			foreach($hash as $entry){
				$id = $entry['msgid'];
				if(isset($entry['msgid_plural'])){
					$id .= "\x00" . $entry['msgid_plural'];
				}
				// context is merged into id, separated by EOT (\x04)
				if(array_key_exists('msgctxt', $entry)){
					$id = $entry['msgctxt'] . "\x04" . $id;
				}
				// plural msgstrs are NUL-separated
				$str = implode("\x00", $entry['msgstr']);
				// keep track of offsets
				$offsets[]= [
					strlen($ids),
					strlen($id),
					strlen($strings),
					strlen($str)
				];
				// plural msgids are not stored (?)
				$ids .= $id . "\x00";
				$strings .= $str . "\x00";
			}
			// keys start after the header (7 words) + index tables ($#hash * 4 words)
			$key_start = 7 * 4 + sizeof($hash) * 4 * 4;
			// values start right after the keys
			$value_start = $key_start + strlen($ids);
			// first all key offsets, then all value offsets
			$key_offsets = [];
			$value_offsets = [];
			// calculate
			foreach($offsets as $value){
				list($o1, $l1, $o2, $l2) = $value;
				$key_offsets[] = $l1;
				$key_offsets[] = $o1 + $key_start;
				$value_offsets[] = $l2;
				$value_offsets[] = $o2 + $value_start;
			}
			$offsets= array_merge($key_offsets, $value_offsets);
			// write header
			$mo .= pack('Iiiiiii', 0x950412de, // magic number
				0, // version
				sizeof($hash), // number of entries in the catalog
				7 * 4, // key index offset
				7 * 4 + sizeof($hash) * 8, // value index offset,
				0, // hashtable size (unused, thus 0)
				$key_start // hashtable offset
			);
			// offsets
			foreach($offsets as $offset){
				$mo .= pack('i', $offset);
			}
			// ids
			$mo .= $ids;
			// strings
			$mo .= $strings;
			file_put_contents($mopath, $mo);
		}
		public static function convert(string $popath, string $mopath = null){
			if(is_null($mopath)){
				$mopath = dirname($popath).'/'.basename($popath, '.po').'.mo';
			}
			self::to($mopath, self::from($popath));
		}
		private static function clean($value){
			if(is_array($value)){
				foreach($value as $k => $v){
					$value[$k] = self::clean($v);
				}
			}else{
				if ($value[0] == '"'){
					$value = substr($value, 1, -1);
				}
				$value = (string)str_replace(
					'$',
					'\\$',
					str_replace("\"\n\"", '', $value)
				);
			}
			return $value;
		}
	}
?>