<?php

	// parsecvs.php - Platon SDG CVS commit reports prettyfier
	// Developed by Martin Karas <wolcano@pobox.sk>
	// Tag support by Ondrej Jombik <nepto@php.net> [11/1/2003]
	// Copyright (c) 2002-2003 Platon SDG, http://www.platon.sk/
	// Licensed under terms of GNU General Public License

	// $Platon: scripts/php/parsecvs/parsecvs.php,v 1.26 2005-02-07 22:38:46 rajo Exp $

	#
	# CONSTANTS
	#

	// Specify input/output streams (ie. files, sockets or std's)
	$input_file  = 'php://stdin';
	$output_file = 'php://stdout';

	// Made more liberal (20 compulsory dashes, others are arbitrary)
	// old: $begin_line = '----------------------------------------------';
	$begin_line = '--------------------[-]*';
	$msg_log_begin = '^Log Message:';
	$date_format = '^([A-Z][a-z]{2} *)?([A-Z][a-z]{2} *[0-9]+ *' .
			'[0-9]{2}:[0-9]{2}:[0-9]{2}) *CES?T *([0-9]{4})';
	$user_format = '^CVSUSER=([0-9a-zA-Z]*)';
	$directory_format = '^In directory';
	$project_name_format = '^Update of /home/cvs/(.*)';
	$cvsweblink_format = '^(http:[^ 	]*)';
	$tag_oper_begin  = '^Tag operation: (.*)$';
	$tag_name_begin  = '^Tag name: (.*)$';
	$add_files_begin = '^Added Files:';
	$mod_files_begin = '^Modified Files:';
	$del_files_begin = '^Removed Files:';
	$in_file_list = "^(\t|        )";
	$project_head = '--[ Project %s ]--';
	$project_head_folding_char = '-';
	$sub_projects_separator = "----\n";
	$log_message_folding = '  ';
	
	#
	# FUNCTIONS
	#

	function my_array_walk($x)
	{
		// This will do array_unique() on third-level values of $x
		$ik = array_keys($x);
		for ($i = 0; $i < sizeof($ik); $i++) {
			$jk = array_keys($x[$ik[$i]]);
			for ($j = 0; $j < sizeof($jk); $j++)
				if (is_array($x[$ik[$i]][$jk[$j]])) {

					/* To prevent merging [0] => array(), [1] => array(), ...
					   into one [0] => array().  Note that merging algorithm
					   in array_unique() is string-based, thus it saw
					   [0] => 'Array', [1] => 'Array', ... and it logically
					   merge it into one [0] => 'Array'. However, this is not
					   desirable for us.

					   -- Nepto [11/1/2003] */

					$all_values_are_arrays = true;
					foreach ($x[$ik[$i]][$jk[$j]] as $val) {
						if (! is_array($val)) {
							$all_values_are_arrays = false;
							break;
						}
					}
					// Skip merging process if all values are arrays
					if (! $all_values_are_arrays) {
						$x[$ik[$i]][$jk[$j]] = array_unique($x[$ik[$i]][$jk[$j]]);
					}
				}
		}
		return $x;
	} // my_array_walk()

	function write_out_file_set($op = '*', $file_ar, $tags)
	{
		if (! is_array($file_ar) || count($file_ar) <= 0)
			return;

		$additional_file_infos = array();
		if (is_array($tags)) {
			foreach ($tags as $tag_ar) {
				foreach ($tag_ar['files'] as $file) {
					$additional_file_infos[$file][] =
						$tag_ar['operation'].' '.$tag_ar['name'];
				}
			}
		}

		global $out, $indent_str;
		$prefix = "$indent_str$op ";
		foreach ($file_ar as $filename) {
			fwrite($out, $prefix);
			if (is_array($additional_file_infos[$filename])) {
				for ($i = 0; $i < count($additional_file_infos[$filename]); $i++) {
					if ($i == 0) {
						$filename2 = str_pad($filename, 35, ' ', STR_PAD_RIGHT);
					} else if ($i == 1) {
						$filename2 = str_repeat(' ', 35 + strlen($prefix));
					}
					$add_info = $additional_file_infos[$filename][$i];
					fwrite($out, "$filename2 $add_info\n");
				}
			} else {
				fwrite($out, "$filename\n");
			}
		}
	}

	#
	# MAIN JOB
	#

	$in = file($input_file);  // input split into array by lines
	$i  = 0;                  // starting line to parse
	$j  = -1;                 // current parsed report

	// Rewind to the beginning of the first commit. This will allow us
	// to parse commits with garbage texts on the top (ie. e-mails).
	for (; ! ereg($begin_line, $in[$i]); $i++);

	while ($i < sizeof($in)) {
		// BEGIN NEW COMMIT
		if (ereg($begin_line, $in[$i])) {
			// increasing report index
			$j++;
			// reseting all flags
			$in_msg_log   = false;
			$in_add_files = false;
			$in_mod_files = false;
			$in_del_files = false;
			// moving to the next line
			$i++;
			continue;
		}

		// LOG MESSAGE
		if (ereg($msg_log_begin, $in[$i])) {
			$in_msg_log = true;
			$i++;
			continue;
		}

		if ($in_msg_log) {
			if (ereg($begin_line, $in[$i])) {
				$in_msg_log = false;
				$i++;
				continue;
			}
			else {
				$reports[$j]['log_msg'][sizeof($reports[$j]['log_msg']) - 1] .=
					rtrim($in[$i]);
				//$reports[$j]['log_msg'][0] .= rtrim($in[$i]) . "\n";
			}
		}

		// ADDED FILES
		if (ereg($add_files_begin, $in[$i])) {
			$in_add_files = true;
			$i++;
			continue;
		}
		
		if ($in_add_files) {
			if (!ereg($in_file_list, $in[$i])) {
				$in_add_files = false;
				continue;
			}

			$reports[$j]['added'] = 
				array_merge($reports[$j]['added'],
					explode(' ', trim($in[$i])));
		}
		
		// MODIFIED FILES
		if (ereg($mod_files_begin, $in[$i])) {
			$in_mod_files = true;
			$i++;
			continue;
		}
		
		if ($in_mod_files) {
			if (!ereg($in_file_list, $in[$i])) {
				$in_mod_files = false;
				continue;
			}

			$reports[$j]['modified'] = 
				@array_merge($reports[$j]['modified'],
				explode(' ', trim($in[$i])));
		}

		// DELETED FILES
		if (ereg($del_files_begin, $in[$i])) {
			$in_del_files = true;
			$i++;
			continue;
		}
		
		if ($in_del_files) {
			if (!ereg($in_file_list, $in[$i])) {
				$in_del_files = false;
				continue;
			}

			$reports[$j]['deleted'] = 
				array_merge($reports[$j]['deleted'],
					explode(' ', trim($in[$i])));
		}

		// DATE		//Accepted format: Mon May 20 18:06:09 CEST 2002
		if (ereg($date_format, $in[$i], $regs)) {
			$reports[$j]['date'][] = $regs[2] . ' ' . $regs[3];
		}

		// USERS
		if (ereg($user_format, $in[$i], $regs)) {
			$reports[$j]['users'][] = $regs[1];
		}

		// DIRECTORY
		if (ereg($directory_format, $in[$i])) {
			// This is redundant, simply ignoring it
		}

		// NAME
		if (ereg($project_name_format, $in[$i], $regs)) {
			// only project name
			$reports[$j]['name'] = trim(strtok($regs[1], '/'));
			$reports[$j]['directory'] = trim($regs[1]);
		}

		// CVSWEB link
		if (ereg($cvsweblink_format, $in[$i], $regs)) {
			// only URL
			$reports[$j]['cvsweblink'] = 
				@array_merge($reports[$j]['cvsweblink'], trim($regs[1]));
		}

		// TAG OPERATION
		if (ereg($tag_oper_begin, $in[$i], $regs)) {
			//$reports[$j]['tag_operation'][0] = trim($regs[1]);
			$reports[$j]['tag'][0]['operation'] = trim($regs[1]);
		}
		
		// TAG NAME
		if (ereg($tag_name_begin, $in[$i], $regs)) {
			//$reports[$j]['tag_name'][0] = trim($regs[1]);
			$reports[$j]['tag'][0]['name'] = trim($regs[1]);
		}

		$i++;
	}

	$merged = array();

	// Merge recursively all reports into one array
	// eg. join reports in same directory
	for ($i = 0; $i < sizeof($reports); $i++) {

		if (isset($reports[$i]['tag'][0]['name'])) {			
			$reports[$i]['tag'][0]['files'] = 
				(array) $reports[$i]['added'] +
				(array) $reports[$i]['modified'] +
				(array) $reports[$i]['deleted'];

			switch (@$reports[$i]['tag'][0]['operation']) {
				case 'add': $op = '+'; break;
				case 'del': $op = '-'; break;
				default:    $op = '?'; break;
			}
			$reports[$i]['tag'][0]['operation'] = $op;
		}
		unset($a);
		$a[$reports[$i]['directory']] = $reports[$i];
		$merged = array_merge_recursive($merged, $a);
	}
	unset($i);
	unset($a);

	$merged_unique = my_array_walk($merged);
	ksort($merged_unique);

	// DEBUG
	//var_dump($reports);
	//var_dump($merged);
	//var_dump($merged_unique);

	unset($reports);
	unset($merged);

	$out  = fopen($output_file, 'w');
	$keys = array_keys($merged_unique);
	$i    = 0;
	$name = '';

	// OUTPUT:
	// Print out the merged/unique to $out

	while ($i < sizeof($merged_unique)) {
		$k = $keys[$i];
		$m = $merged_unique[$k];
		$new_name = is_array($m['name']) ? join(', ', $m['name']) : $m['name'];
		
		if (strcmp($new_name, $name)) {
			// Print new project name
			fwrite($out, sprintf($project_head, $new_name));
			;
			fwrite($out, str_repeat($project_head_folding_char, 70 - strlen($new_name) - 16));
			fwrite($out, "\n");
			$name = $new_name;
		}
		else {
			// Separate subdirectories of the same project
			fwrite($out, $sub_projects_separator);
		}

		// DIRECTORY
		fwrite($out, 'Path:  ');
		fwrite($out, is_array($m['directory'])
			? join(', ', $m['directory'])
			: $m['directory']);
		fwrite($out, "/\n");

		// DATE
		if (is_array($m['date'])) {
			fwrite($out, 'Date:  ' . $m['date'][sizeof($m['date']) - 1]);
			fwrite($out, "\n");
		}

		// EDITORS
		if (is_array($m['users'])) {
			/* there is no fprintf() in PHP,
			   so we must use fwrite()/sprintf() combo */
			fwrite($out, sprintf('User%s ',
						count($m['users']) == 1 ? ': ' : 's:'));
			fwrite($out, implode(', ', $m['users']) . "\n");
		}

		// MODIFIED FILES
		if (is_array($m['added']) || is_array($m['modified']) ||
			is_array($m['deleted'])) {
			if ((count($m['added']) + count($m['modified']) +
				count($m['deleted'])) > 1)
				fwrite($out, "Files:\n");
			else
				fwrite($out, "File:\n");
		}
			
		// mod = mod - add
		if (is_array($m['added']) && is_array($m['modified'])) {
			$m['modified'] = array_diff($m['modified'], $m['added']);
		}
		
		// $indent_str is used as global in write_out_file_set()
		$indent_str = str_repeat(' ', strlen('Files') + 2);

		write_out_file_set('+', $m['added'],    $m['tag']);
		write_out_file_set('*', $m['modified'], $m['tag']);
		write_out_file_set('-', $m['deleted'],  $m['tag']);

		// LOG MESSAGES
		if (is_array($m['log_msg'])) {
			fwrite($out, "Message:\n");
			for ($l = 0; $l < count($m['log_msg']); $l++) {
				if (trim($m['log_msg'][$l]) == '')
					unset($m['log_msg'][$l]);
			}

			fwrite($out, $log_message_folding .
				implode("\n$log_message_folding",
				//str_replace("\n", "\n$log_message_folding", $m['log_msg'])
				$m['log_msg']
			) . "\n");
		}

		// LOG MESSAGES
		if (is_array($m['cvsweblink'])) {
			fwrite($out, "CVSweb URL:\n");

			fwrite($out, $log_message_folding .
				implode("\n$log_message_folding",
				$m['cvsweblink']
			) . "\n");
		}

		$i++;
	}
	
	fflush($out);
?>