<?php
/**
 * php.mo 0.1 by Joss Crowcroft (http://www.josscrowcroft.com)
 * 
 * Converts gettext translation '.po' files to binary '.mo' files in PHP.
 * 
 * Usage: 
 * <?php require('php-mo.php'); phpmo_convert( 'input.po', [ 'output.mo' ] ); ?>
 * 
 * NB:
 * - If no $output_file specified, output filename is same as $input_file (but .mo)
 * - Returns true/false for success/failure
 * - No warranty, but if it breaks, please let me know
 * 
 * More info:
 * https://github.com/josscrowcroft/php.mo
 * 
 * Based on php-msgfmt by Matthias Bauer (Copyright © 2007), a command-line PHP tool
 * for converting .po files to .mo.
 * (http://wordpress-soc-2007.googlecode.com/svn/trunk/moeffju/php-msgfmt/msgfmt.php)
 * 
 * License: GPL v3 http://www.opensource.org/licenses/gpl-3.0.html
 */

/**
 * The main .po to .mo function
 */
function phpmo_convert($input, $output = false) {
	if ( !$output )
		$output = str_replace( '.po', '.mo', $input );

	$hash = phpmo_parse_po_file( $input );
	if ( $hash === false ) {
		return false;
	} else {
		phpmo_write_mo_file( $hash, $output );
		return true;
	}
}

function phpmo_clean_helper($x) {
	if (is_array($x)) {
		foreach ($x as $k => $v) {
			$x[$k] = phpmo_clean_helper($v);
		}
	} else {
		if ($x[0] == '"')
			$x = substr($x, 1, -1);
		$x = str_replace("\"\n\"", '', $x);
		$x = str_replace('$', '\\$', $x);
		$x = @ eval ("return \"$x\";");
	}
	return $x;
}

/* Parse gettext .po files. */
/* @link http://www.gnu.org/software/gettext/manual/gettext.html#PO-Files */
function phpmo_parse_po_file($in) {
	// read .po file
	$fc = file_get_contents($in);
	// normalize newlines
	$fc = str_replace(array (
		"\r\n",
		"\r"
	), array (
		"\n",
		"\n"
	), $fc);

	// results array
	$hash = array ();
	// temporary array
	$temp = array ();
	// state
	$state = null;
	$fuzzy = false;

	// iterate over lines
	foreach (explode("\n", $fc) as $line) {
		$line = trim($line);
		if ($line === '')
			continue;

		list ($key, $data) = explode(' ', $line, 2);

		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 = array ();
					$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;
	$hash = array ();
	foreach ($temp as $entry) {
		foreach ($entry as & $v) {
			$v = phpmo_clean_helper($v);
			if ($v === FALSE) {
				// parse error
				return FALSE;
			}
		}
		$hash[$entry['msgid']] = $entry;
	}

	return $hash;
}

/* Write a GNU gettext style machine object. */
/* @link http://www.gnu.org/software/gettext/manual/gettext.html#MO-Files */
function phpmo_write_mo_file($hash, $out) {
	// sort by msgid
	ksort($hash, SORT_STRING);
	// our mo file data
	$mo = '';
	// header data
	$offsets = array ();
	$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
		$offsets[] = array (
			strlen($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 = array ();
	$value_offsets = array ();
	// calculate
	foreach ($offsets as $v) {
		list ($o1, $l1, $o2, $l2) = $v;
		$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($out, $mo);
}
?>