| 1 |
<?php |
| 2 |
|
| 3 |
/* |
| 4 |
|
| 5 |
class.Diff.php |
| 6 |
|
| 7 |
A class containing a diff implementation |
| 8 |
|
| 9 |
Created by Stephen Morley - http://stephenmorley.org/ - and released under the |
| 10 |
terms of the CC0 1.0 Universal legal code: |
| 11 |
|
| 12 |
http://creativecommons.org/publicdomain/zero/1.0/legalcode |
| 13 |
|
| 14 |
*/ |
| 15 |
|
| 16 |
// A class containing functions for computing diffs and formatting the output. |
| 17 |
/** |
| 18 |
* Class Diff |
| 19 |
*/ |
| 20 |
class RSFirewall_Helper_Diff |
| 21 |
{ |
| 22 |
// define the constants |
| 23 |
const UNMODIFIED = 0; |
| 24 |
const DELETED = 1; |
| 25 |
const INSERTED = 2; |
| 26 |
|
| 27 |
/* Returns the diff for two strings. The return value is an array, each of |
| 28 |
* whose values is an array containing two values: a line (or character, if |
| 29 |
* $compareCharacters is true), and one of the constants DIFF::UNMODIFIED (the |
| 30 |
* line or character is in both strings), DIFF::DELETED (the line or character |
| 31 |
* is only in the first string), and DIFF::INSERTED (the line or character is |
| 32 |
* only in the second string). The parameters are: |
| 33 |
* |
| 34 |
* $string1 - the first string |
| 35 |
* $string2 - the second string |
| 36 |
* $compareCharacters - true to compare characters, and false to compare |
| 37 |
* lines; this optional parameter defaults to false |
| 38 |
*/ |
| 39 |
/** |
| 40 |
* @param $string1 |
| 41 |
* @param $string2 |
| 42 |
* @param bool $compareCharacters |
| 43 |
* |
| 44 |
* @return array |
| 45 |
*/ |
| 46 |
public static function compare( |
| 47 |
$string1, $string2, $compareCharacters = false){ |
| 48 |
|
| 49 |
// initialise the sequences and comparison start and end positions |
| 50 |
$start = 0; |
| 51 |
if ($compareCharacters){ |
| 52 |
$sequence1 = $string1; |
| 53 |
$sequence2 = $string2; |
| 54 |
$end1 = strlen($string1) - 1; |
| 55 |
$end2 = strlen($string2) - 1; |
| 56 |
}else{ |
| 57 |
$sequence1 = preg_split('/\R/', $string1); |
| 58 |
$sequence2 = preg_split('/\R/', $string2); |
| 59 |
$end1 = count($sequence1) - 1; |
| 60 |
$end2 = count($sequence2) - 1; |
| 61 |
} |
| 62 |
|
| 63 |
// skip any common prefix |
| 64 |
while ($start <= $end1 && $start <= $end2 |
| 65 |
&& $sequence1[$start] == $sequence2[$start]){ |
| 66 |
$start ++; |
| 67 |
} |
| 68 |
|
| 69 |
// skip any common suffix |
| 70 |
while ($end1 >= $start && $end2 >= $start |
| 71 |
&& $sequence1[$end1] == $sequence2[$end2]){ |
| 72 |
$end1 --; |
| 73 |
$end2 --; |
| 74 |
} |
| 75 |
|
| 76 |
// compute the table of longest common subsequence lengths |
| 77 |
$table = self::computeTable($sequence1, $sequence2, $start, $end1, $end2); |
| 78 |
|
| 79 |
// generate the partial diff |
| 80 |
$partialDiff = |
| 81 |
self::generatePartialDiff($table, $sequence1, $sequence2, $start); |
| 82 |
|
| 83 |
// generate the full diff |
| 84 |
$diff = array(); |
| 85 |
for ($index = 0; $index < $start; $index ++){ |
| 86 |
$diff[] = array($sequence1[$index], self::UNMODIFIED); |
| 87 |
} |
| 88 |
while (count($partialDiff) > 0) $diff[] = array_pop($partialDiff); |
| 89 |
for ($index = $end1 + 1; |
| 90 |
$index < ($compareCharacters ? strlen($sequence1) : count($sequence1)); |
| 91 |
$index ++){ |
| 92 |
$diff[] = array($sequence1[$index], self::UNMODIFIED); |
| 93 |
} |
| 94 |
|
| 95 |
// return the diff |
| 96 |
return $diff; |
| 97 |
|
| 98 |
} |
| 99 |
|
| 100 |
/* Returns the diff for two files. The parameters are: |
| 101 |
* |
| 102 |
* $file1 - the path to the first file |
| 103 |
* $file2 - the path to the second file |
| 104 |
* $compareCharacters - true to compare characters, and false to compare |
| 105 |
* lines; this optional parameter defaults to false |
| 106 |
*/ |
| 107 |
/** |
| 108 |
* @param $file1 |
| 109 |
* @param $file2 |
| 110 |
* @param bool $compareCharacters |
| 111 |
* |
| 112 |
* @return array |
| 113 |
*/ |
| 114 |
public static function compareFiles( |
| 115 |
$file1, $file2, $compareCharacters = false){ |
| 116 |
|
| 117 |
// return the diff of the files |
| 118 |
return self::compare( |
| 119 |
file_get_contents($file1), |
| 120 |
file_get_contents($file2), |
| 121 |
$compareCharacters); |
| 122 |
|
| 123 |
} |
| 124 |
|
| 125 |
/* Returns the table of longest common subsequence lengths for the specified |
| 126 |
* sequences. The parameters are: |
| 127 |
* |
| 128 |
* $sequence1 - the first sequence |
| 129 |
* $sequence2 - the second sequence |
| 130 |
* $start - the starting index |
| 131 |
* $end1 - the ending index for the first sequence |
| 132 |
* $end2 - the ending index for the second sequence |
| 133 |
*/ |
| 134 |
/** |
| 135 |
* @param $sequence1 |
| 136 |
* @param $sequence2 |
| 137 |
* @param $start |
| 138 |
* @param $end1 |
| 139 |
* @param $end2 |
| 140 |
* |
| 141 |
* @return array |
| 142 |
*/ |
| 143 |
private static function computeTable( |
| 144 |
$sequence1, $sequence2, $start, $end1, $end2){ |
| 145 |
|
| 146 |
// determine the lengths to be compared |
| 147 |
$length1 = $end1 - $start + 1; |
| 148 |
$length2 = $end2 - $start + 1; |
| 149 |
|
| 150 |
// initialise the table |
| 151 |
$table = array(array_fill(0, $length2 + 1, 0)); |
| 152 |
|
| 153 |
// loop over the rows |
| 154 |
for ($index1 = 1; $index1 <= $length1; $index1 ++){ |
| 155 |
|
| 156 |
// create the new row |
| 157 |
$table[$index1] = array(0); |
| 158 |
|
| 159 |
// loop over the columns |
| 160 |
for ($index2 = 1; $index2 <= $length2; $index2 ++){ |
| 161 |
|
| 162 |
// store the longest common subsequence length |
| 163 |
if ($sequence1[$index1 + $start - 1] |
| 164 |
== $sequence2[$index2 + $start - 1]){ |
| 165 |
$table[$index1][$index2] = $table[$index1 - 1][$index2 - 1] + 1; |
| 166 |
}else{ |
| 167 |
$table[$index1][$index2] = |
| 168 |
max($table[$index1 - 1][$index2], $table[$index1][$index2 - 1]); |
| 169 |
} |
| 170 |
|
| 171 |
} |
| 172 |
} |
| 173 |
|
| 174 |
// return the table |
| 175 |
return $table; |
| 176 |
|
| 177 |
} |
| 178 |
|
| 179 |
/* Returns the partial diff for the specificed sequences, in reverse order. |
| 180 |
* The parameters are: |
| 181 |
* |
| 182 |
* $table - the table returned by the computeTable function |
| 183 |
* $sequence1 - the first sequence |
| 184 |
* $sequence2 - the second sequence |
| 185 |
* $start - the starting index |
| 186 |
*/ |
| 187 |
/** |
| 188 |
* @param $table |
| 189 |
* @param $sequence1 |
| 190 |
* @param $sequence2 |
| 191 |
* @param $start |
| 192 |
* |
| 193 |
* @return array |
| 194 |
*/ |
| 195 |
private static function generatePartialDiff( |
| 196 |
$table, $sequence1, $sequence2, $start){ |
| 197 |
|
| 198 |
// initialise the diff |
| 199 |
$diff = array(); |
| 200 |
|
| 201 |
// initialise the indices |
| 202 |
$index1 = count($table) - 1; |
| 203 |
$index2 = count($table[0]) - 1; |
| 204 |
|
| 205 |
// loop until there are no items remaining in either sequence |
| 206 |
while ($index1 > 0 || $index2 > 0){ |
| 207 |
|
| 208 |
// check what has happened to the items at these indices |
| 209 |
if ($index1 > 0 && $index2 > 0 |
| 210 |
&& $sequence1[$index1 + $start - 1] |
| 211 |
== $sequence2[$index2 + $start - 1]){ |
| 212 |
|
| 213 |
// update the diff and the indices |
| 214 |
$diff[] = array($sequence1[$index1 + $start - 1], self::UNMODIFIED); |
| 215 |
$index1 --; |
| 216 |
$index2 --; |
| 217 |
|
| 218 |
}elseif ($index2 > 0 |
| 219 |
&& $table[$index1][$index2] == $table[$index1][$index2 - 1]){ |
| 220 |
|
| 221 |
// update the diff and the indices |
| 222 |
$diff[] = array($sequence2[$index2 + $start - 1], self::INSERTED); |
| 223 |
$index2 --; |
| 224 |
|
| 225 |
}else{ |
| 226 |
|
| 227 |
// update the diff and the indices |
| 228 |
$diff[] = array($sequence1[$index1 + $start - 1], self::DELETED); |
| 229 |
$index1 --; |
| 230 |
|
| 231 |
} |
| 232 |
|
| 233 |
} |
| 234 |
|
| 235 |
// return the diff |
| 236 |
return $diff; |
| 237 |
|
| 238 |
} |
| 239 |
|
| 240 |
/* Returns a diff as a string, where unmodified lines are prefixed by ' ', |
| 241 |
* deletions are prefixed by '- ', and insertions are prefixed by '+ '. The |
| 242 |
* parameters are: |
| 243 |
* |
| 244 |
* $diff - the diff array |
| 245 |
* $separator - the separator between lines; this optional parameter defaults |
| 246 |
* to "\n" |
| 247 |
*/ |
| 248 |
/** |
| 249 |
* @param $diff |
| 250 |
* @param string $separator |
| 251 |
* |
| 252 |
* @return string |
| 253 |
*/ |
| 254 |
public static function toString($diff, $separator = "\n"){ |
| 255 |
|
| 256 |
// initialise the string |
| 257 |
$string = ''; |
| 258 |
|
| 259 |
// loop over the lines in the diff |
| 260 |
foreach ($diff as $line){ |
| 261 |
|
| 262 |
// extend the string with the line |
| 263 |
switch ($line[1]){ |
| 264 |
case self::UNMODIFIED : $string .= ' ' . $line[0];break; |
| 265 |
case self::DELETED : $string .= '- ' . $line[0];break; |
| 266 |
case self::INSERTED : $string .= '+ ' . $line[0];break; |
| 267 |
} |
| 268 |
|
| 269 |
// extend the string with the separator |
| 270 |
$string .= $separator; |
| 271 |
|
| 272 |
} |
| 273 |
|
| 274 |
// return the string |
| 275 |
return $string; |
| 276 |
|
| 277 |
} |
| 278 |
|
| 279 |
/* Returns a diff as an HTML string, where unmodified lines are contained |
| 280 |
* within 'span' elements, deletions are contained within 'del' elements, and |
| 281 |
* insertions are contained within 'ins' elements. The parameters are: |
| 282 |
* |
| 283 |
* $diff - the diff array |
| 284 |
* $separator - the separator between lines; this optional parameter defaults |
| 285 |
* to '<br>' |
| 286 |
*/ |
| 287 |
/** |
| 288 |
* @param $diff |
| 289 |
* @param string $separator |
| 290 |
* |
| 291 |
* @return string |
| 292 |
*/ |
| 293 |
public static function toHTML($diff, $separator = '<br>'){ |
| 294 |
|
| 295 |
// initialise the HTML |
| 296 |
$html = ''; |
| 297 |
|
| 298 |
// loop over the lines in the diff |
| 299 |
foreach ($diff as $line){ |
| 300 |
|
| 301 |
// extend the HTML with the line |
| 302 |
switch ($line[1]){ |
| 303 |
case self::UNMODIFIED : $element = 'span'; break; |
| 304 |
case self::DELETED : $element = 'del'; break; |
| 305 |
case self::INSERTED : $element = 'ins'; break; |
| 306 |
} |
| 307 |
$html .= |
| 308 |
'<' . $element . '>' |
| 309 |
. htmlspecialchars($line[0]) |
| 310 |
. '</' . $element . '>'; |
| 311 |
|
| 312 |
// extend the HTML with the separator |
| 313 |
$html .= $separator; |
| 314 |
|
| 315 |
} |
| 316 |
|
| 317 |
// return the HTML |
| 318 |
return $html; |
| 319 |
|
| 320 |
} |
| 321 |
|
| 322 |
/* Returns a diff as an HTML table. The parameters are: |
| 323 |
* |
| 324 |
* $diff - the diff array |
| 325 |
* $indentation - indentation to add to every line of the generated HTML; this |
| 326 |
* optional parameter defaults to '' |
| 327 |
* $separator - the separator between lines; this optional parameter |
| 328 |
* defaults to '<br>' |
| 329 |
*/ |
| 330 |
/** |
| 331 |
* @param $diff |
| 332 |
* @param string $indentation |
| 333 |
* @param string $separator |
| 334 |
* @param array $headers |
| 335 |
* |
| 336 |
* @return string |
| 337 |
*/ |
| 338 |
public static function toTable($diff, $indentation = '', $separator = '<br>', $headers = array()){ |
| 339 |
// initialise the HTML |
| 340 |
$html = $indentation . "<table class=\"diff table table-bordered\">\n"; |
| 341 |
|
| 342 |
// Add headers if requested |
| 343 |
if ($headers) { |
| 344 |
$html .= |
| 345 |
"<thead>\n". |
| 346 |
"<tr>\n". |
| 347 |
"<th>".htmlspecialchars($headers[0], ENT_COMPAT, 'utf-8')."</th>\n". |
| 348 |
"<th>".htmlspecialchars($headers[1], ENT_COMPAT, 'utf-8')."</th>\n". |
| 349 |
"</tr>\n". |
| 350 |
"</thead>\n"; |
| 351 |
} |
| 352 |
|
| 353 |
// loop over the lines in the diff |
| 354 |
$index = 0; |
| 355 |
while ($index < count($diff)){ |
| 356 |
|
| 357 |
// determine the line type |
| 358 |
switch ($diff[$index][1]){ |
| 359 |
|
| 360 |
// display the content on the left and right |
| 361 |
case self::UNMODIFIED: |
| 362 |
$leftCell = |
| 363 |
self::getCellContent( |
| 364 |
$diff, $indentation, $separator, $index, self::UNMODIFIED); |
| 365 |
$rightCell = $leftCell; |
| 366 |
break; |
| 367 |
|
| 368 |
// display the deleted on the left and inserted content on the right |
| 369 |
case self::DELETED: |
| 370 |
$leftCell = |
| 371 |
self::getCellContent( |
| 372 |
$diff, $indentation, $separator, $index, self::DELETED); |
| 373 |
$rightCell = |
| 374 |
self::getCellContent( |
| 375 |
$diff, $indentation, $separator, $index, self::INSERTED); |
| 376 |
break; |
| 377 |
|
| 378 |
// display the inserted content on the right |
| 379 |
case self::INSERTED: |
| 380 |
$leftCell = ''; |
| 381 |
$rightCell = |
| 382 |
self::getCellContent( |
| 383 |
$diff, $indentation, $separator, $index, self::INSERTED); |
| 384 |
break; |
| 385 |
|
| 386 |
} |
| 387 |
|
| 388 |
// extend the HTML with the new row |
| 389 |
$html .= |
| 390 |
$indentation |
| 391 |
. " <tr>\n" |
| 392 |
. $indentation |
| 393 |
. ' <td class="diff' |
| 394 |
. ($leftCell == $rightCell |
| 395 |
? 'Unmodified' |
| 396 |
: ($leftCell == '' ? 'Blank' : 'Deleted')) |
| 397 |
. '">' |
| 398 |
. $leftCell |
| 399 |
. "</td>\n" |
| 400 |
. $indentation |
| 401 |
. ' <td class="diff' |
| 402 |
. ($leftCell == $rightCell |
| 403 |
? 'Unmodified' |
| 404 |
: ($rightCell == '' ? 'Blank' : 'Inserted')) |
| 405 |
. '">' |
| 406 |
. $rightCell |
| 407 |
. "</td>\n" |
| 408 |
. $indentation |
| 409 |
. " </tr>\n"; |
| 410 |
|
| 411 |
} |
| 412 |
|
| 413 |
// return the HTML |
| 414 |
return $html . $indentation . "</table>\n"; |
| 415 |
|
| 416 |
} |
| 417 |
|
| 418 |
/* Returns the content of the cell, for use in the toTable function. The |
| 419 |
* parameters are: |
| 420 |
* |
| 421 |
* $diff - the diff array |
| 422 |
* $indentation - indentation to add to every line of the generated HTML |
| 423 |
* $separator - the separator between lines |
| 424 |
* $index - the current index, passes by reference |
| 425 |
* $type - the type of line |
| 426 |
*/ |
| 427 |
/** |
| 428 |
* @param $diff |
| 429 |
* @param $indentation |
| 430 |
* @param $separator |
| 431 |
* @param $index |
| 432 |
* @param $type |
| 433 |
* |
| 434 |
* @return string |
| 435 |
*/ |
| 436 |
private static function getCellContent( |
| 437 |
$diff, $indentation, $separator, &$index, $type){ |
| 438 |
|
| 439 |
// initialise the HTML |
| 440 |
$html = ''; |
| 441 |
|
| 442 |
// loop over the matching lines, adding them to the HTML |
| 443 |
while ($index < count($diff) && $diff[$index][1] == $type){ |
| 444 |
$html .= |
| 445 |
'<span>' |
| 446 |
. htmlspecialchars($diff[$index][0]) |
| 447 |
. '</span>' |
| 448 |
. $separator; |
| 449 |
$index ++; |
| 450 |
} |
| 451 |
|
| 452 |
// return the HTML |
| 453 |
return $html; |
| 454 |
|
| 455 |
} |
| 456 |
|
| 457 |
} |
| 458 |
|
