| 1 |
<?php declare(strict_types=1); |
| 2 |
|
| 3 |
namespace PhpParser; |
| 4 |
|
| 5 |
/* |
| 6 |
* This parser is based on a skeleton written by Moriyoshi Koizumi, which in |
| 7 |
* turn is based on work by Masato Bito. |
| 8 |
*/ |
| 9 |
|
| 10 |
use PhpParser\Node\Expr; |
| 11 |
use PhpParser\Node\Expr\Array_; |
| 12 |
use PhpParser\Node\Expr\Cast\Double; |
| 13 |
use PhpParser\Node\Identifier; |
| 14 |
use PhpParser\Node\InterpolatedStringPart; |
| 15 |
use PhpParser\Node\Name; |
| 16 |
use PhpParser\Node\Param; |
| 17 |
use PhpParser\Node\Scalar\InterpolatedString; |
| 18 |
use PhpParser\Node\Scalar\Int_; |
| 19 |
use PhpParser\Node\Scalar\String_; |
| 20 |
use PhpParser\Node\Stmt; |
| 21 |
use PhpParser\Node\Stmt\Class_; |
| 22 |
use PhpParser\Node\Stmt\ClassConst; |
| 23 |
use PhpParser\Node\Stmt\ClassMethod; |
| 24 |
use PhpParser\Node\Stmt\Else_; |
| 25 |
use PhpParser\Node\Stmt\ElseIf_; |
| 26 |
use PhpParser\Node\Stmt\Enum_; |
| 27 |
use PhpParser\Node\Stmt\Interface_; |
| 28 |
use PhpParser\Node\Stmt\Namespace_; |
| 29 |
use PhpParser\Node\Stmt\Nop; |
| 30 |
use PhpParser\Node\Stmt\Property; |
| 31 |
use PhpParser\Node\Stmt\TryCatch; |
| 32 |
use PhpParser\Node\UseItem; |
| 33 |
use PhpParser\NodeVisitor\CommentAnnotatingVisitor; |
| 34 |
|
| 35 |
abstract class ParserAbstract implements Parser { |
| 36 |
private const SYMBOL_NONE = -1; |
| 37 |
|
| 38 |
/** @var Lexer Lexer that is used when parsing */ |
| 39 |
protected Lexer $lexer; |
| 40 |
/** @var PhpVersion PHP version to target on a best-effort basis */ |
| 41 |
protected PhpVersion $phpVersion; |
| 42 |
|
| 43 |
/* |
| 44 |
* The following members will be filled with generated parsing data: |
| 45 |
*/ |
| 46 |
|
| 47 |
/** @var int Size of $tokenToSymbol map */ |
| 48 |
protected int $tokenToSymbolMapSize; |
| 49 |
/** @var int Size of $action table */ |
| 50 |
protected int $actionTableSize; |
| 51 |
/** @var int Size of $goto table */ |
| 52 |
protected int $gotoTableSize; |
| 53 |
|
| 54 |
/** @var int Symbol number signifying an invalid token */ |
| 55 |
protected int $invalidSymbol; |
| 56 |
/** @var int Symbol number of error recovery token */ |
| 57 |
protected int $errorSymbol; |
| 58 |
/** @var int Action number signifying default action */ |
| 59 |
protected int $defaultAction; |
| 60 |
/** @var int Rule number signifying that an unexpected token was encountered */ |
| 61 |
protected int $unexpectedTokenRule; |
| 62 |
|
| 63 |
protected int $YY2TBLSTATE; |
| 64 |
/** @var int Number of non-leaf states */ |
| 65 |
protected int $numNonLeafStates; |
| 66 |
|
| 67 |
/** @var int[] Map of PHP token IDs to internal symbols */ |
| 68 |
protected array $phpTokenToSymbol; |
| 69 |
/** @var array<int, bool> Map of PHP token IDs to drop */ |
| 70 |
protected array $dropTokens; |
| 71 |
/** @var int[] Map of external symbols (static::T_*) to internal symbols */ |
| 72 |
protected array $tokenToSymbol; |
| 73 |
/** @var string[] Map of symbols to their names */ |
| 74 |
protected array $symbolToName; |
| 75 |
/** @var array<int, string> Names of the production rules (only necessary for debugging) */ |
| 76 |
protected array $productions; |
| 77 |
|
| 78 |
/** @var int[] Map of states to a displacement into the $action table. The corresponding action for this |
| 79 |
* state/symbol pair is $action[$actionBase[$state] + $symbol]. If $actionBase[$state] is 0, the |
| 80 |
* action is defaulted, i.e. $actionDefault[$state] should be used instead. */ |
| 81 |
protected array $actionBase; |
| 82 |
/** @var int[] Table of actions. Indexed according to $actionBase comment. */ |
| 83 |
protected array $action; |
| 84 |
/** @var int[] Table indexed analogously to $action. If $actionCheck[$actionBase[$state] + $symbol] != $symbol |
| 85 |
* then the action is defaulted, i.e. $actionDefault[$state] should be used instead. */ |
| 86 |
protected array $actionCheck; |
| 87 |
/** @var int[] Map of states to their default action */ |
| 88 |
protected array $actionDefault; |
| 89 |
/** @var callable[] Semantic action callbacks */ |
| 90 |
protected array $reduceCallbacks; |
| 91 |
|
| 92 |
/** @var int[] Map of non-terminals to a displacement into the $goto table. The corresponding goto state for this |
| 93 |
* non-terminal/state pair is $goto[$gotoBase[$nonTerminal] + $state] (unless defaulted) */ |
| 94 |
protected array $gotoBase; |
| 95 |
/** @var int[] Table of states to goto after reduction. Indexed according to $gotoBase comment. */ |
| 96 |
protected array $goto; |
| 97 |
/** @var int[] Table indexed analogously to $goto. If $gotoCheck[$gotoBase[$nonTerminal] + $state] != $nonTerminal |
| 98 |
* then the goto state is defaulted, i.e. $gotoDefault[$nonTerminal] should be used. */ |
| 99 |
protected array $gotoCheck; |
| 100 |
/** @var int[] Map of non-terminals to the default state to goto after their reduction */ |
| 101 |
protected array $gotoDefault; |
| 102 |
|
| 103 |
/** @var int[] Map of rules to the non-terminal on their left-hand side, i.e. the non-terminal to use for |
| 104 |
* determining the state to goto after reduction. */ |
| 105 |
protected array $ruleToNonTerminal; |
| 106 |
/** @var int[] Map of rules to the length of their right-hand side, which is the number of elements that have to |
| 107 |
* be popped from the stack(s) on reduction. */ |
| 108 |
protected array $ruleToLength; |
| 109 |
|
| 110 |
/* |
| 111 |
* The following members are part of the parser state: |
| 112 |
*/ |
| 113 |
|
| 114 |
/** @var mixed Temporary value containing the result of last semantic action (reduction) */ |
| 115 |
protected $semValue; |
| 116 |
/** @var mixed[] Semantic value stack (contains values of tokens and semantic action results) */ |
| 117 |
protected array $semStack; |
| 118 |
/** @var int[] Token start position stack */ |
| 119 |
protected array $tokenStartStack; |
| 120 |
/** @var int[] Token end position stack */ |
| 121 |
protected array $tokenEndStack; |
| 122 |
|
| 123 |
/** @var ErrorHandler Error handler */ |
| 124 |
protected ErrorHandler $errorHandler; |
| 125 |
/** @var int Error state, used to avoid error floods */ |
| 126 |
protected int $errorState; |
| 127 |
|
| 128 |
/** @var \SplObjectStorage<Array_, null>|null Array nodes created during parsing, for postprocessing of empty elements. */ |
| 129 |
protected ?\SplObjectStorage $createdArrays; |
| 130 |
|
| 131 |
/** @var Token[] Tokens for the current parse */ |
| 132 |
protected array $tokens; |
| 133 |
/** @var int Current position in token array */ |
| 134 |
protected int $tokenPos; |
| 135 |
|
| 136 |
/** |
| 137 |
* Initialize $reduceCallbacks map. |
| 138 |
*/ |
| 139 |
abstract protected function initReduceCallbacks(): void; |
| 140 |
|
| 141 |
/** |
| 142 |
* Creates a parser instance. |
| 143 |
* |
| 144 |
* Options: |
| 145 |
* * phpVersion: ?PhpVersion, |
| 146 |
* |
| 147 |
* @param Lexer $lexer A lexer |
| 148 |
* @param PhpVersion $phpVersion PHP version to target, defaults to latest supported. This |
| 149 |
* option is best-effort: Even if specified, parsing will generally assume the latest |
| 150 |
* supported version and only adjust behavior in minor ways, for example by omitting |
| 151 |
* errors in older versions and interpreting type hints as a name or identifier depending |
| 152 |
* on version. |
| 153 |
*/ |
| 154 |
public function __construct(Lexer $lexer, ?PhpVersion $phpVersion = null) { |
| 155 |
$this->lexer = $lexer; |
| 156 |
$this->phpVersion = $phpVersion ?? PhpVersion::getNewestSupported(); |
| 157 |
|
| 158 |
$this->initReduceCallbacks(); |
| 159 |
$this->phpTokenToSymbol = $this->createTokenMap(); |
| 160 |
$this->dropTokens = array_fill_keys( |
| 161 |
[\T_WHITESPACE, \T_OPEN_TAG, \T_COMMENT, \T_DOC_COMMENT, \T_BAD_CHARACTER], true |
| 162 |
); |
| 163 |
} |
| 164 |
|
| 165 |
/** |
| 166 |
* Parses PHP code into a node tree. |
| 167 |
* |
| 168 |
* If a non-throwing error handler is used, the parser will continue parsing after an error |
| 169 |
* occurred and attempt to build a partial AST. |
| 170 |
* |
| 171 |
* @param string $code The source code to parse |
| 172 |
* @param ErrorHandler|null $errorHandler Error handler to use for lexer/parser errors, defaults |
| 173 |
* to ErrorHandler\Throwing. |
| 174 |
* |
| 175 |
* @return Node\Stmt[]|null Array of statements (or null non-throwing error handler is used and |
| 176 |
* the parser was unable to recover from an error). |
| 177 |
*/ |
| 178 |
public function parse(string $code, ?ErrorHandler $errorHandler = null): ?array { |
| 179 |
$this->errorHandler = $errorHandler ?: new ErrorHandler\Throwing(); |
| 180 |
$this->createdArrays = new \SplObjectStorage(); |
| 181 |
|
| 182 |
$this->tokens = $this->lexer->tokenize($code, $this->errorHandler); |
| 183 |
$result = $this->doParse(); |
| 184 |
|
| 185 |
// Report errors for any empty elements used inside arrays. This is delayed until after the main parse, |
| 186 |
// because we don't know a priori whether a given array expression will be used in a destructuring context |
| 187 |
// or not. |
| 188 |
foreach ($this->createdArrays as $node) { |
| 189 |
foreach ($node->items as $item) { |
| 190 |
if ($item->value instanceof Expr\Error) { |
| 191 |
$this->errorHandler->handleError( |
| 192 |
new Error('Cannot use empty array elements in arrays', $item->getAttributes())); |
| 193 |
} |
| 194 |
} |
| 195 |
} |
| 196 |
|
| 197 |
// Clear out some of the interior state, so we don't hold onto unnecessary |
| 198 |
// memory between uses of the parser |
| 199 |
$this->tokenStartStack = []; |
| 200 |
$this->tokenEndStack = []; |
| 201 |
$this->semStack = []; |
| 202 |
$this->semValue = null; |
| 203 |
$this->createdArrays = null; |
| 204 |
|
| 205 |
if ($result !== null) { |
| 206 |
$traverser = new NodeTraverser(new CommentAnnotatingVisitor($this->tokens)); |
| 207 |
$traverser->traverse($result); |
| 208 |
} |
| 209 |
|
| 210 |
return $result; |
| 211 |
} |
| 212 |
|
| 213 |
public function getTokens(): array { |
| 214 |
return $this->tokens; |
| 215 |
} |
| 216 |
|
| 217 |
/** @return Stmt[]|null */ |
| 218 |
protected function doParse(): ?array { |
| 219 |
// We start off with no lookahead-token |
| 220 |
$symbol = self::SYMBOL_NONE; |
| 221 |
$tokenValue = null; |
| 222 |
$this->tokenPos = -1; |
| 223 |
|
| 224 |
// Keep stack of start and end attributes |
| 225 |
$this->tokenStartStack = []; |
| 226 |
$this->tokenEndStack = [0]; |
| 227 |
|
| 228 |
// Start off in the initial state and keep a stack of previous states |
| 229 |
$state = 0; |
| 230 |
$stateStack = [$state]; |
| 231 |
|
| 232 |
// Semantic value stack (contains values of tokens and semantic action results) |
| 233 |
$this->semStack = []; |
| 234 |
|
| 235 |
// Current position in the stack(s) |
| 236 |
$stackPos = 0; |
| 237 |
|
| 238 |
$this->errorState = 0; |
| 239 |
|
| 240 |
for (;;) { |
| 241 |
//$this->traceNewState($state, $symbol); |
| 242 |
|
| 243 |
if ($this->actionBase[$state] === 0) { |
| 244 |
$rule = $this->actionDefault[$state]; |
| 245 |
} else { |
| 246 |
if ($symbol === self::SYMBOL_NONE) { |
| 247 |
do { |
| 248 |
$token = $this->tokens[++$this->tokenPos]; |
| 249 |
$tokenId = $token->id; |
| 250 |
} while (isset($this->dropTokens[$tokenId])); |
| 251 |
|
| 252 |
// Map the lexer token id to the internally used symbols. |
| 253 |
$tokenValue = $token->text; |
| 254 |
if (!isset($this->phpTokenToSymbol[$tokenId])) { |
| 255 |
throw new \RangeException(sprintf( |
| 256 |
'The lexer returned an invalid token (id=%d, value=%s)', |
| 257 |
$tokenId, $tokenValue |
| 258 |
)); |
| 259 |
} |
| 260 |
$symbol = $this->phpTokenToSymbol[$tokenId]; |
| 261 |
|
| 262 |
//$this->traceRead($symbol); |
| 263 |
} |
| 264 |
|
| 265 |
$idx = $this->actionBase[$state] + $symbol; |
| 266 |
if ((($idx >= 0 && $idx < $this->actionTableSize && $this->actionCheck[$idx] === $symbol) |
| 267 |
|| ($state < $this->YY2TBLSTATE |
| 268 |
&& ($idx = $this->actionBase[$state + $this->numNonLeafStates] + $symbol) >= 0 |
| 269 |
&& $idx < $this->actionTableSize && $this->actionCheck[$idx] === $symbol)) |
| 270 |
&& ($action = $this->action[$idx]) !== $this->defaultAction) { |
| 271 |
/* |
| 272 |
* >= numNonLeafStates: shift and reduce |
| 273 |
* > 0: shift |
| 274 |
* = 0: accept |
| 275 |
* < 0: reduce |
| 276 |
* = -YYUNEXPECTED: error |
| 277 |
*/ |
| 278 |
if ($action > 0) { |
| 279 |
/* shift */ |
| 280 |
//$this->traceShift($symbol); |
| 281 |
|
| 282 |
++$stackPos; |
| 283 |
$stateStack[$stackPos] = $state = $action; |
| 284 |
$this->semStack[$stackPos] = $tokenValue; |
| 285 |
$this->tokenStartStack[$stackPos] = $this->tokenPos; |
| 286 |
$this->tokenEndStack[$stackPos] = $this->tokenPos; |
| 287 |
$symbol = self::SYMBOL_NONE; |
| 288 |
|
| 289 |
if ($this->errorState) { |
| 290 |
--$this->errorState; |
| 291 |
} |
| 292 |
|
| 293 |
if ($action < $this->numNonLeafStates) { |
| 294 |
continue; |
| 295 |
} |
| 296 |
|
| 297 |
/* $yyn >= numNonLeafStates means shift-and-reduce */ |
| 298 |
$rule = $action - $this->numNonLeafStates; |
| 299 |
} else { |
| 300 |
$rule = -$action; |
| 301 |
} |
| 302 |
} else { |
| 303 |
$rule = $this->actionDefault[$state]; |
| 304 |
} |
| 305 |
} |
| 306 |
|
| 307 |
for (;;) { |
| 308 |
if ($rule === 0) { |
| 309 |
/* accept */ |
| 310 |
//$this->traceAccept(); |
| 311 |
return $this->semValue; |
| 312 |
} |
| 313 |
if ($rule !== $this->unexpectedTokenRule) { |
| 314 |
/* reduce */ |
| 315 |
//$this->traceReduce($rule); |
| 316 |
|
| 317 |
$ruleLength = $this->ruleToLength[$rule]; |
| 318 |
try { |
| 319 |
$callback = $this->reduceCallbacks[$rule]; |
| 320 |
if ($callback !== null) { |
| 321 |
$callback($this, $stackPos); |
| 322 |
} elseif ($ruleLength > 0) { |
| 323 |
$this->semValue = $this->semStack[$stackPos - $ruleLength + 1]; |
| 324 |
} |
| 325 |
} catch (Error $e) { |
| 326 |
if (-1 === $e->getStartLine()) { |
| 327 |
$e->setStartLine($this->tokens[$this->tokenPos]->line); |
| 328 |
} |
| 329 |
|
| 330 |
$this->emitError($e); |
| 331 |
// Can't recover from this type of error |
| 332 |
return null; |
| 333 |
} |
| 334 |
|
| 335 |
/* Goto - shift nonterminal */ |
| 336 |
$lastTokenEnd = $this->tokenEndStack[$stackPos]; |
| 337 |
$stackPos -= $ruleLength; |
| 338 |
$nonTerminal = $this->ruleToNonTerminal[$rule]; |
| 339 |
$idx = $this->gotoBase[$nonTerminal] + $stateStack[$stackPos]; |
| 340 |
if ($idx >= 0 && $idx < $this->gotoTableSize && $this->gotoCheck[$idx] === $nonTerminal) { |
| 341 |
$state = $this->goto[$idx]; |
| 342 |
} else { |
| 343 |
$state = $this->gotoDefault[$nonTerminal]; |
| 344 |
} |
| 345 |
|
| 346 |
++$stackPos; |
| 347 |
$stateStack[$stackPos] = $state; |
| 348 |
$this->semStack[$stackPos] = $this->semValue; |
| 349 |
$this->tokenEndStack[$stackPos] = $lastTokenEnd; |
| 350 |
if ($ruleLength === 0) { |
| 351 |
// Empty productions use the start attributes of the lookahead token. |
| 352 |
$this->tokenStartStack[$stackPos] = $this->tokenPos; |
| 353 |
} |
| 354 |
} else { |
| 355 |
/* error */ |
| 356 |
switch ($this->errorState) { |
| 357 |
case 0: |
| 358 |
$msg = $this->getErrorMessage($symbol, $state); |
| 359 |
$this->emitError(new Error($msg, $this->getAttributesForToken($this->tokenPos))); |
| 360 |
// Break missing intentionally |
| 361 |
// no break |
| 362 |
case 1: |
| 363 |
case 2: |
| 364 |
$this->errorState = 3; |
| 365 |
|
| 366 |
// Pop until error-expecting state uncovered |
| 367 |
while (!( |
| 368 |
(($idx = $this->actionBase[$state] + $this->errorSymbol) >= 0 |
| 369 |
&& $idx < $this->actionTableSize && $this->actionCheck[$idx] === $this->errorSymbol) |
| 370 |
|| ($state < $this->YY2TBLSTATE |
| 371 |
&& ($idx = $this->actionBase[$state + $this->numNonLeafStates] + $this->errorSymbol) >= 0 |
| 372 |
&& $idx < $this->actionTableSize && $this->actionCheck[$idx] === $this->errorSymbol) |
| 373 |
) || ($action = $this->action[$idx]) === $this->defaultAction) { // Not totally sure about this |
| 374 |
if ($stackPos <= 0) { |
| 375 |
// Could not recover from error |
| 376 |
return null; |
| 377 |
} |
| 378 |
$state = $stateStack[--$stackPos]; |
| 379 |
//$this->tracePop($state); |
| 380 |
} |
| 381 |
|
| 382 |
//$this->traceShift($this->errorSymbol); |
| 383 |
++$stackPos; |
| 384 |
$stateStack[$stackPos] = $state = $action; |
| 385 |
|
| 386 |
// We treat the error symbol as being empty, so we reset the end attributes |
| 387 |
// to the end attributes of the last non-error symbol |
| 388 |
$this->tokenStartStack[$stackPos] = $this->tokenPos; |
| 389 |
$this->tokenEndStack[$stackPos] = $this->tokenEndStack[$stackPos - 1]; |
| 390 |
break; |
| 391 |
|
| 392 |
case 3: |
| 393 |
if ($symbol === 0) { |
| 394 |
// Reached EOF without recovering from error |
| 395 |
return null; |
| 396 |
} |
| 397 |
|
| 398 |
//$this->traceDiscard($symbol); |
| 399 |
$symbol = self::SYMBOL_NONE; |
| 400 |
break 2; |
| 401 |
} |
| 402 |
} |
| 403 |
|
| 404 |
if ($state < $this->numNonLeafStates) { |
| 405 |
break; |
| 406 |
} |
| 407 |
|
| 408 |
/* >= numNonLeafStates means shift-and-reduce */ |
| 409 |
$rule = $state - $this->numNonLeafStates; |
| 410 |
} |
| 411 |
} |
| 412 |
|
| 413 |
throw new \RuntimeException('Reached end of parser loop'); |
| 414 |
} |
| 415 |
|
| 416 |
protected function emitError(Error $error): void { |
| 417 |
$this->errorHandler->handleError($error); |
| 418 |
} |
| 419 |
|
| 420 |
/** |
| 421 |
* Format error message including expected tokens. |
| 422 |
* |
| 423 |
* @param int $symbol Unexpected symbol |
| 424 |
* @param int $state State at time of error |
| 425 |
* |
| 426 |
* @return string Formatted error message |
| 427 |
*/ |
| 428 |
protected function getErrorMessage(int $symbol, int $state): string { |
| 429 |
$expectedString = ''; |
| 430 |
if ($expected = $this->getExpectedTokens($state)) { |
| 431 |
$expectedString = ', expecting ' . implode(' or ', $expected); |
| 432 |
} |
| 433 |
|
| 434 |
return 'Syntax error, unexpected ' . $this->symbolToName[$symbol] . $expectedString; |
| 435 |
} |
| 436 |
|
| 437 |
/** |
| 438 |
* Get limited number of expected tokens in given state. |
| 439 |
* |
| 440 |
* @param int $state State |
| 441 |
* |
| 442 |
* @return string[] Expected tokens. If too many, an empty array is returned. |
| 443 |
*/ |
| 444 |
protected function getExpectedTokens(int $state): array { |
| 445 |
$expected = []; |
| 446 |
|
| 447 |
$base = $this->actionBase[$state]; |
| 448 |
foreach ($this->symbolToName as $symbol => $name) { |
| 449 |
$idx = $base + $symbol; |
| 450 |
if ($idx >= 0 && $idx < $this->actionTableSize && $this->actionCheck[$idx] === $symbol |
| 451 |
|| $state < $this->YY2TBLSTATE |
| 452 |
&& ($idx = $this->actionBase[$state + $this->numNonLeafStates] + $symbol) >= 0 |
| 453 |
&& $idx < $this->actionTableSize && $this->actionCheck[$idx] === $symbol |
| 454 |
) { |
| 455 |
if ($this->action[$idx] !== $this->unexpectedTokenRule |
| 456 |
&& $this->action[$idx] !== $this->defaultAction |
| 457 |
&& $symbol !== $this->errorSymbol |
| 458 |
) { |
| 459 |
if (count($expected) === 4) { |
| 460 |
/* Too many expected tokens */ |
| 461 |
return []; |
| 462 |
} |
| 463 |
|
| 464 |
$expected[] = $name; |
| 465 |
} |
| 466 |
} |
| 467 |
} |
| 468 |
|
| 469 |
return $expected; |
| 470 |
} |
| 471 |
|
| 472 |
/** |
| 473 |
* Get attributes for a node with the given start and end token positions. |
| 474 |
* |
| 475 |
* @param int $tokenStartPos Token position the node starts at |
| 476 |
* @param int $tokenEndPos Token position the node ends at |
| 477 |
* @return array<string, mixed> Attributes |
| 478 |
*/ |
| 479 |
protected function getAttributes(int $tokenStartPos, int $tokenEndPos): array { |
| 480 |
$startToken = $this->tokens[$tokenStartPos]; |
| 481 |
$afterEndToken = $this->tokens[$tokenEndPos + 1]; |
| 482 |
return [ |
| 483 |
'startLine' => $startToken->line, |
| 484 |
'startTokenPos' => $tokenStartPos, |
| 485 |
'startFilePos' => $startToken->pos, |
| 486 |
'endLine' => $afterEndToken->line, |
| 487 |
'endTokenPos' => $tokenEndPos, |
| 488 |
'endFilePos' => $afterEndToken->pos - 1, |
| 489 |
]; |
| 490 |
} |
| 491 |
|
| 492 |
/** |
| 493 |
* Get attributes for a single token at the given token position. |
| 494 |
* |
| 495 |
* @return array<string, mixed> Attributes |
| 496 |
*/ |
| 497 |
protected function getAttributesForToken(int $tokenPos): array { |
| 498 |
if ($tokenPos < \count($this->tokens) - 1) { |
| 499 |
return $this->getAttributes($tokenPos, $tokenPos); |
| 500 |
} |
| 501 |
|
| 502 |
// Get attributes for the sentinel token. |
| 503 |
$token = $this->tokens[$tokenPos]; |
| 504 |
return [ |
| 505 |
'startLine' => $token->line, |
| 506 |
'startTokenPos' => $tokenPos, |
| 507 |
'startFilePos' => $token->pos, |
| 508 |
'endLine' => $token->line, |
| 509 |
'endTokenPos' => $tokenPos, |
| 510 |
'endFilePos' => $token->pos, |
| 511 |
]; |
| 512 |
} |
| 513 |
|
| 514 |
/* |
| 515 |
* Tracing functions used for debugging the parser. |
| 516 |
*/ |
| 517 |
|
| 518 |
/* |
| 519 |
protected function traceNewState($state, $symbol): void { |
| 520 |
echo '% State ' . $state |
| 521 |
. ', Lookahead ' . ($symbol == self::SYMBOL_NONE ? '--none--' : $this->symbolToName[$symbol]) . "\n"; |
| 522 |
} |
| 523 |
|
| 524 |
protected function traceRead($symbol): void { |
| 525 |
echo '% Reading ' . $this->symbolToName[$symbol] . "\n"; |
| 526 |
} |
| 527 |
|
| 528 |
protected function traceShift($symbol): void { |
| 529 |
echo '% Shift ' . $this->symbolToName[$symbol] . "\n"; |
| 530 |
} |
| 531 |
|
| 532 |
protected function traceAccept(): void { |
| 533 |
echo "% Accepted.\n"; |
| 534 |
} |
| 535 |
|
| 536 |
protected function traceReduce($n): void { |
| 537 |
echo '% Reduce by (' . $n . ') ' . $this->productions[$n] . "\n"; |
| 538 |
} |
| 539 |
|
| 540 |
protected function tracePop($state): void { |
| 541 |
echo '% Recovering, uncovered state ' . $state . "\n"; |
| 542 |
} |
| 543 |
|
| 544 |
protected function traceDiscard($symbol): void { |
| 545 |
echo '% Discard ' . $this->symbolToName[$symbol] . "\n"; |
| 546 |
} |
| 547 |
*/ |
| 548 |
|
| 549 |
/* |
| 550 |
* Helper functions invoked by semantic actions |
| 551 |
*/ |
| 552 |
|
| 553 |
/** |
| 554 |
* Moves statements of semicolon-style namespaces into $ns->stmts and checks various error conditions. |
| 555 |
* |
| 556 |
* @param Node\Stmt[] $stmts |
| 557 |
* @return Node\Stmt[] |
| 558 |
*/ |
| 559 |
protected function handleNamespaces(array $stmts): array { |
| 560 |
$hasErrored = false; |
| 561 |
$style = $this->getNamespacingStyle($stmts); |
| 562 |
if (null === $style) { |
| 563 |
// not namespaced, nothing to do |
| 564 |
return $stmts; |
| 565 |
} |
| 566 |
if ('brace' === $style) { |
| 567 |
// For braced namespaces we only have to check that there are no invalid statements between the namespaces |
| 568 |
$afterFirstNamespace = false; |
| 569 |
foreach ($stmts as $stmt) { |
| 570 |
if ($stmt instanceof Node\Stmt\Namespace_) { |
| 571 |
$afterFirstNamespace = true; |
| 572 |
} elseif (!$stmt instanceof Node\Stmt\HaltCompiler |
| 573 |
&& !$stmt instanceof Node\Stmt\Nop |
| 574 |
&& $afterFirstNamespace && !$hasErrored) { |
| 575 |
$this->emitError(new Error( |
| 576 |
'No code may exist outside of namespace {}', $stmt->getAttributes())); |
| 577 |
$hasErrored = true; // Avoid one error for every statement |
| 578 |
} |
| 579 |
} |
| 580 |
return $stmts; |
| 581 |
} else { |
| 582 |
// For semicolon namespaces we have to move the statements after a namespace declaration into ->stmts |
| 583 |
$resultStmts = []; |
| 584 |
$targetStmts = &$resultStmts; |
| 585 |
$lastNs = null; |
| 586 |
foreach ($stmts as $stmt) { |
| 587 |
if ($stmt instanceof Node\Stmt\Namespace_) { |
| 588 |
if ($lastNs !== null) { |
| 589 |
$this->fixupNamespaceAttributes($lastNs); |
| 590 |
} |
| 591 |
if ($stmt->stmts === null) { |
| 592 |
$stmt->stmts = []; |
| 593 |
$targetStmts = &$stmt->stmts; |
| 594 |
$resultStmts[] = $stmt; |
| 595 |
} else { |
| 596 |
// This handles the invalid case of mixed style namespaces |
| 597 |
$resultStmts[] = $stmt; |
| 598 |
$targetStmts = &$resultStmts; |
| 599 |
} |
| 600 |
$lastNs = $stmt; |
| 601 |
} elseif ($stmt instanceof Node\Stmt\HaltCompiler) { |
| 602 |
// __halt_compiler() is not moved into the namespace |
| 603 |
$resultStmts[] = $stmt; |
| 604 |
} else { |
| 605 |
$targetStmts[] = $stmt; |
| 606 |
} |
| 607 |
} |
| 608 |
if ($lastNs !== null) { |
| 609 |
$this->fixupNamespaceAttributes($lastNs); |
| 610 |
} |
| 611 |
return $resultStmts; |
| 612 |
} |
| 613 |
} |
| 614 |
|
| 615 |
private function fixupNamespaceAttributes(Node\Stmt\Namespace_ $stmt): void { |
| 616 |
// We moved the statements into the namespace node, as such the end of the namespace node |
| 617 |
// needs to be extended to the end of the statements. |
| 618 |
if (empty($stmt->stmts)) { |
| 619 |
return; |
| 620 |
} |
| 621 |
|
| 622 |
// We only move the builtin end attributes here. This is the best we can do with the |
| 623 |
// knowledge we have. |
| 624 |
$endAttributes = ['endLine', 'endFilePos', 'endTokenPos']; |
| 625 |
$lastStmt = $stmt->stmts[count($stmt->stmts) - 1]; |
| 626 |
foreach ($endAttributes as $endAttribute) { |
| 627 |
if ($lastStmt->hasAttribute($endAttribute)) { |
| 628 |
$stmt->setAttribute($endAttribute, $lastStmt->getAttribute($endAttribute)); |
| 629 |
} |
| 630 |
} |
| 631 |
} |
| 632 |
|
| 633 |
/** @return array<string, mixed> */ |
| 634 |
private function getNamespaceErrorAttributes(Namespace_ $node): array { |
| 635 |
$attrs = $node->getAttributes(); |
| 636 |
// Adjust end attributes to only cover the "namespace" keyword, not the whole namespace. |
| 637 |
if (isset($attrs['startLine'])) { |
| 638 |
$attrs['endLine'] = $attrs['startLine']; |
| 639 |
} |
| 640 |
if (isset($attrs['startTokenPos'])) { |
| 641 |
$attrs['endTokenPos'] = $attrs['startTokenPos']; |
| 642 |
} |
| 643 |
if (isset($attrs['startFilePos'])) { |
| 644 |
$attrs['endFilePos'] = $attrs['startFilePos'] + \strlen('namespace') - 1; |
| 645 |
} |
| 646 |
return $attrs; |
| 647 |
} |
| 648 |
|
| 649 |
/** |
| 650 |
* Determine namespacing style (semicolon or brace) |
| 651 |
* |
| 652 |
* @param Node[] $stmts Top-level statements. |
| 653 |
* |
| 654 |
* @return null|string One of "semicolon", "brace" or null (no namespaces) |
| 655 |
*/ |
| 656 |
private function getNamespacingStyle(array $stmts): ?string { |
| 657 |
$style = null; |
| 658 |
$hasNotAllowedStmts = false; |
| 659 |
foreach ($stmts as $i => $stmt) { |
| 660 |
if ($stmt instanceof Node\Stmt\Namespace_) { |
| 661 |
$currentStyle = null === $stmt->stmts ? 'semicolon' : 'brace'; |
| 662 |
if (null === $style) { |
| 663 |
$style = $currentStyle; |
| 664 |
if ($hasNotAllowedStmts) { |
| 665 |
$this->emitError(new Error( |
| 666 |
'Namespace declaration statement has to be the very first statement in the script', |
| 667 |
$this->getNamespaceErrorAttributes($stmt) |
| 668 |
)); |
| 669 |
} |
| 670 |
} elseif ($style !== $currentStyle) { |
| 671 |
$this->emitError(new Error( |
| 672 |
'Cannot mix bracketed namespace declarations with unbracketed namespace declarations', |
| 673 |
$this->getNamespaceErrorAttributes($stmt) |
| 674 |
)); |
| 675 |
// Treat like semicolon style for namespace normalization |
| 676 |
return 'semicolon'; |
| 677 |
} |
| 678 |
continue; |
| 679 |
} |
| 680 |
|
| 681 |
/* declare(), __halt_compiler() and nops can be used before a namespace declaration */ |
| 682 |
if ($stmt instanceof Node\Stmt\Declare_ |
| 683 |
|| $stmt instanceof Node\Stmt\HaltCompiler |
| 684 |
|| $stmt instanceof Node\Stmt\Nop) { |
| 685 |
continue; |
| 686 |
} |
| 687 |
|
| 688 |
/* There may be a hashbang line at the very start of the file */ |
| 689 |
if ($i === 0 && $stmt instanceof Node\Stmt\InlineHTML && preg_match('/\A#!.*\r?\n\z/', $stmt->value)) { |
| 690 |
continue; |
| 691 |
} |
| 692 |
|
| 693 |
/* Everything else if forbidden before namespace declarations */ |
| 694 |
$hasNotAllowedStmts = true; |
| 695 |
} |
| 696 |
return $style; |
| 697 |
} |
| 698 |
|
| 699 |
/** @return Name|Identifier */ |
| 700 |
protected function handleBuiltinTypes(Name $name) { |
| 701 |
if (!$name->isUnqualified()) { |
| 702 |
return $name; |
| 703 |
} |
| 704 |
|
| 705 |
$lowerName = $name->toLowerString(); |
| 706 |
if (!$this->phpVersion->supportsBuiltinType($lowerName)) { |
| 707 |
return $name; |
| 708 |
} |
| 709 |
|
| 710 |
return new Node\Identifier($lowerName, $name->getAttributes()); |
| 711 |
} |
| 712 |
|
| 713 |
/** |
| 714 |
* Get combined start and end attributes at a stack location |
| 715 |
* |
| 716 |
* @param int $stackPos Stack location |
| 717 |
* |
| 718 |
* @return array<string, mixed> Combined start and end attributes |
| 719 |
*/ |
| 720 |
protected function getAttributesAt(int $stackPos): array { |
| 721 |
return $this->getAttributes($this->tokenStartStack[$stackPos], $this->tokenEndStack[$stackPos]); |
| 722 |
} |
| 723 |
|
| 724 |
protected function getFloatCastKind(string $cast): int { |
| 725 |
$cast = strtolower($cast); |
| 726 |
if (strpos($cast, 'float') !== false) { |
| 727 |
return Double::KIND_FLOAT; |
| 728 |
} |
| 729 |
|
| 730 |
if (strpos($cast, 'real') !== false) { |
| 731 |
return Double::KIND_REAL; |
| 732 |
} |
| 733 |
|
| 734 |
return Double::KIND_DOUBLE; |
| 735 |
} |
| 736 |
|
| 737 |
/** @param array<string, mixed> $attributes */ |
| 738 |
protected function parseLNumber(string $str, array $attributes, bool $allowInvalidOctal = false): Int_ { |
| 739 |
try { |
| 740 |
return Int_::fromString($str, $attributes, $allowInvalidOctal); |
| 741 |
} catch (Error $error) { |
| 742 |
$this->emitError($error); |
| 743 |
// Use dummy value |
| 744 |
return new Int_(0, $attributes); |
| 745 |
} |
| 746 |
} |
| 747 |
|
| 748 |
/** |
| 749 |
* Parse a T_NUM_STRING token into either an integer or string node. |
| 750 |
* |
| 751 |
* @param string $str Number string |
| 752 |
* @param array<string, mixed> $attributes Attributes |
| 753 |
* |
| 754 |
* @return Int_|String_ Integer or string node. |
| 755 |
*/ |
| 756 |
protected function parseNumString(string $str, array $attributes) { |
| 757 |
if (!preg_match('/^(?:0|-?[1-9][0-9]*)$/', $str)) { |
| 758 |
return new String_($str, $attributes); |
| 759 |
} |
| 760 |
|
| 761 |
$num = +$str; |
| 762 |
if (!is_int($num)) { |
| 763 |
return new String_($str, $attributes); |
| 764 |
} |
| 765 |
|
| 766 |
return new Int_($num, $attributes); |
| 767 |
} |
| 768 |
|
| 769 |
/** @param array<string, mixed> $attributes */ |
| 770 |
protected function stripIndentation( |
| 771 |
string $string, int $indentLen, string $indentChar, |
| 772 |
bool $newlineAtStart, bool $newlineAtEnd, array $attributes |
| 773 |
): string { |
| 774 |
if ($indentLen === 0) { |
| 775 |
return $string; |
| 776 |
} |
| 777 |
|
| 778 |
$start = $newlineAtStart ? '(?:(?<=\n)|\A)' : '(?<=\n)'; |
| 779 |
$end = $newlineAtEnd ? '(?:(?=[\r\n])|\z)' : '(?=[\r\n])'; |
| 780 |
$regex = '/' . $start . '([ \t]*)(' . $end . ')?/'; |
| 781 |
return preg_replace_callback( |
| 782 |
$regex, |
| 783 |
function ($matches) use ($indentLen, $indentChar, $attributes) { |
| 784 |
$prefix = substr($matches[1], 0, $indentLen); |
| 785 |
if (false !== strpos($prefix, $indentChar === " " ? "\t" : " ")) { |
| 786 |
$this->emitError(new Error( |
| 787 |
'Invalid indentation - tabs and spaces cannot be mixed', $attributes |
| 788 |
)); |
| 789 |
} elseif (strlen($prefix) < $indentLen && !isset($matches[2])) { |
| 790 |
$this->emitError(new Error( |
| 791 |
'Invalid body indentation level ' . |
| 792 |
'(expecting an indentation level of at least ' . $indentLen . ')', |
| 793 |
$attributes |
| 794 |
)); |
| 795 |
} |
| 796 |
return substr($matches[0], strlen($prefix)); |
| 797 |
}, |
| 798 |
$string |
| 799 |
); |
| 800 |
} |
| 801 |
|
| 802 |
/** |
| 803 |
* @param string|(Expr|InterpolatedStringPart)[] $contents |
| 804 |
* @param array<string, mixed> $attributes |
| 805 |
* @param array<string, mixed> $endTokenAttributes |
| 806 |
*/ |
| 807 |
protected function parseDocString( |
| 808 |
string $startToken, $contents, string $endToken, |
| 809 |
array $attributes, array $endTokenAttributes, bool $parseUnicodeEscape |
| 810 |
): Expr { |
| 811 |
$kind = strpos($startToken, "'") === false |
| 812 |
? String_::KIND_HEREDOC : String_::KIND_NOWDOC; |
| 813 |
|
| 814 |
$regex = '/\A[bB]?<<<[ \t]*[\'"]?([a-zA-Z_\x7f-\xff][a-zA-Z0-9_\x7f-\xff]*)[\'"]?(?:\r\n|\n|\r)\z/'; |
| 815 |
$result = preg_match($regex, $startToken, $matches); |
| 816 |
assert($result === 1); |
| 817 |
$label = $matches[1]; |
| 818 |
|
| 819 |
$result = preg_match('/\A[ \t]*/', $endToken, $matches); |
| 820 |
assert($result === 1); |
| 821 |
$indentation = $matches[0]; |
| 822 |
|
| 823 |
$attributes['kind'] = $kind; |
| 824 |
$attributes['docLabel'] = $label; |
| 825 |
$attributes['docIndentation'] = $indentation; |
| 826 |
|
| 827 |
$indentHasSpaces = false !== strpos($indentation, " "); |
| 828 |
$indentHasTabs = false !== strpos($indentation, "\t"); |
| 829 |
if ($indentHasSpaces && $indentHasTabs) { |
| 830 |
$this->emitError(new Error( |
| 831 |
'Invalid indentation - tabs and spaces cannot be mixed', |
| 832 |
$endTokenAttributes |
| 833 |
)); |
| 834 |
|
| 835 |
// Proceed processing as if this doc string is not indented |
| 836 |
$indentation = ''; |
| 837 |
} |
| 838 |
|
| 839 |
$indentLen = \strlen($indentation); |
| 840 |
$indentChar = $indentHasSpaces ? " " : "\t"; |
| 841 |
|
| 842 |
if (\is_string($contents)) { |
| 843 |
if ($contents === '') { |
| 844 |
$attributes['rawValue'] = $contents; |
| 845 |
return new String_('', $attributes); |
| 846 |
} |
| 847 |
|
| 848 |
$contents = $this->stripIndentation( |
| 849 |
$contents, $indentLen, $indentChar, true, true, $attributes |
| 850 |
); |
| 851 |
$contents = preg_replace('~(\r\n|\n|\r)\z~', '', $contents); |
| 852 |
$attributes['rawValue'] = $contents; |
| 853 |
|
| 854 |
if ($kind === String_::KIND_HEREDOC) { |
| 855 |
$contents = String_::parseEscapeSequences($contents, null, $parseUnicodeEscape); |
| 856 |
} |
| 857 |
|
| 858 |
return new String_($contents, $attributes); |
| 859 |
} else { |
| 860 |
assert(count($contents) > 0); |
| 861 |
if (!$contents[0] instanceof Node\InterpolatedStringPart) { |
| 862 |
// If there is no leading encapsed string part, pretend there is an empty one |
| 863 |
$this->stripIndentation( |
| 864 |
'', $indentLen, $indentChar, true, false, $contents[0]->getAttributes() |
| 865 |
); |
| 866 |
} |
| 867 |
|
| 868 |
$newContents = []; |
| 869 |
foreach ($contents as $i => $part) { |
| 870 |
if ($part instanceof Node\InterpolatedStringPart) { |
| 871 |
$isLast = $i === \count($contents) - 1; |
| 872 |
$part->value = $this->stripIndentation( |
| 873 |
$part->value, $indentLen, $indentChar, |
| 874 |
$i === 0, $isLast, $part->getAttributes() |
| 875 |
); |
| 876 |
if ($isLast) { |
| 877 |
$part->value = preg_replace('~(\r\n|\n|\r)\z~', '', $part->value); |
| 878 |
} |
| 879 |
$part->setAttribute('rawValue', $part->value); |
| 880 |
$part->value = String_::parseEscapeSequences($part->value, null, $parseUnicodeEscape); |
| 881 |
if ('' === $part->value) { |
| 882 |
continue; |
| 883 |
} |
| 884 |
} |
| 885 |
$newContents[] = $part; |
| 886 |
} |
| 887 |
return new InterpolatedString($newContents, $attributes); |
| 888 |
} |
| 889 |
} |
| 890 |
|
| 891 |
protected function createCommentFromToken(Token $token, int $tokenPos): Comment { |
| 892 |
assert($token->id === \T_COMMENT || $token->id == \T_DOC_COMMENT); |
| 893 |
return \T_DOC_COMMENT === $token->id |
| 894 |
? new Comment\Doc($token->text, $token->line, $token->pos, $tokenPos, |
| 895 |
$token->getEndLine(), $token->getEndPos() - 1, $tokenPos) |
| 896 |
: new Comment($token->text, $token->line, $token->pos, $tokenPos, |
| 897 |
$token->getEndLine(), $token->getEndPos() - 1, $tokenPos); |
| 898 |
} |
| 899 |
|
| 900 |
/** |
| 901 |
* Get last comment before the given token position, if any |
| 902 |
*/ |
| 903 |
protected function getCommentBeforeToken(int $tokenPos): ?Comment { |
| 904 |
while (--$tokenPos >= 0) { |
| 905 |
$token = $this->tokens[$tokenPos]; |
| 906 |
if (!isset($this->dropTokens[$token->id])) { |
| 907 |
break; |
| 908 |
} |
| 909 |
|
| 910 |
if ($token->id === \T_COMMENT || $token->id === \T_DOC_COMMENT) { |
| 911 |
return $this->createCommentFromToken($token, $tokenPos); |
| 912 |
} |
| 913 |
} |
| 914 |
return null; |
| 915 |
} |
| 916 |
|
| 917 |
/** |
| 918 |
* Create a zero-length nop to capture preceding comments, if any. |
| 919 |
*/ |
| 920 |
protected function maybeCreateZeroLengthNop(int $tokenPos): ?Nop { |
| 921 |
$comment = $this->getCommentBeforeToken($tokenPos); |
| 922 |
if ($comment === null) { |
| 923 |
return null; |
| 924 |
} |
| 925 |
|
| 926 |
$commentEndLine = $comment->getEndLine(); |
| 927 |
$commentEndFilePos = $comment->getEndFilePos(); |
| 928 |
$commentEndTokenPos = $comment->getEndTokenPos(); |
| 929 |
$attributes = [ |
| 930 |
'startLine' => $commentEndLine, |
| 931 |
'endLine' => $commentEndLine, |
| 932 |
'startFilePos' => $commentEndFilePos + 1, |
| 933 |
'endFilePos' => $commentEndFilePos, |
| 934 |
'startTokenPos' => $commentEndTokenPos + 1, |
| 935 |
'endTokenPos' => $commentEndTokenPos, |
| 936 |
]; |
| 937 |
return new Nop($attributes); |
| 938 |
} |
| 939 |
|
| 940 |
protected function maybeCreateNop(int $tokenStartPos, int $tokenEndPos): ?Nop { |
| 941 |
if ($this->getCommentBeforeToken($tokenStartPos) === null) { |
| 942 |
return null; |
| 943 |
} |
| 944 |
return new Nop($this->getAttributes($tokenStartPos, $tokenEndPos)); |
| 945 |
} |
| 946 |
|
| 947 |
protected function handleHaltCompiler(): string { |
| 948 |
// Prevent the lexer from returning any further tokens. |
| 949 |
$nextToken = $this->tokens[$this->tokenPos + 1]; |
| 950 |
$this->tokenPos = \count($this->tokens) - 2; |
| 951 |
|
| 952 |
// Return text after __halt_compiler. |
| 953 |
return $nextToken->id === \T_INLINE_HTML ? $nextToken->text : ''; |
| 954 |
} |
| 955 |
|
| 956 |
protected function inlineHtmlHasLeadingNewline(int $stackPos): bool { |
| 957 |
$tokenPos = $this->tokenStartStack[$stackPos]; |
| 958 |
$token = $this->tokens[$tokenPos]; |
| 959 |
assert($token->id == \T_INLINE_HTML); |
| 960 |
if ($tokenPos > 0) { |
| 961 |
$prevToken = $this->tokens[$tokenPos - 1]; |
| 962 |
assert($prevToken->id == \T_CLOSE_TAG); |
| 963 |
return false !== strpos($prevToken->text, "\n") |
| 964 |
|| false !== strpos($prevToken->text, "\r"); |
| 965 |
} |
| 966 |
return true; |
| 967 |
} |
| 968 |
|
| 969 |
/** |
| 970 |
* @return array<string, mixed> |
| 971 |
*/ |
| 972 |
protected function createEmptyElemAttributes(int $tokenPos): array { |
| 973 |
return $this->getAttributesForToken($tokenPos); |
| 974 |
} |
| 975 |
|
| 976 |
protected function fixupArrayDestructuring(Array_ $node): Expr\List_ { |
| 977 |
$this->createdArrays->detach($node); |
| 978 |
return new Expr\List_(array_map(function (Node\ArrayItem $item) { |
| 979 |
if ($item->value instanceof Expr\Error) { |
| 980 |
// We used Error as a placeholder for empty elements, which are legal for destructuring. |
| 981 |
return null; |
| 982 |
} |
| 983 |
if ($item->value instanceof Array_) { |
| 984 |
return new Node\ArrayItem( |
| 985 |
$this->fixupArrayDestructuring($item->value), |
| 986 |
$item->key, $item->byRef, $item->getAttributes()); |
| 987 |
} |
| 988 |
return $item; |
| 989 |
}, $node->items), ['kind' => Expr\List_::KIND_ARRAY] + $node->getAttributes()); |
| 990 |
} |
| 991 |
|
| 992 |
protected function postprocessList(Expr\List_ $node): void { |
| 993 |
foreach ($node->items as $i => $item) { |
| 994 |
if ($item->value instanceof Expr\Error) { |
| 995 |
// We used Error as a placeholder for empty elements, which are legal for destructuring. |
| 996 |
$node->items[$i] = null; |
| 997 |
} |
| 998 |
} |
| 999 |
} |
| 1000 |
|
| 1001 |
/** @param ElseIf_|Else_ $node */ |
| 1002 |
protected function fixupAlternativeElse($node): void { |
| 1003 |
// Make sure a trailing nop statement carrying comments is part of the node. |
| 1004 |
$numStmts = \count($node->stmts); |
| 1005 |
if ($numStmts !== 0 && $node->stmts[$numStmts - 1] instanceof Nop) { |
| 1006 |
$nopAttrs = $node->stmts[$numStmts - 1]->getAttributes(); |
| 1007 |
if (isset($nopAttrs['endLine'])) { |
| 1008 |
$node->setAttribute('endLine', $nopAttrs['endLine']); |
| 1009 |
} |
| 1010 |
if (isset($nopAttrs['endFilePos'])) { |
| 1011 |
$node->setAttribute('endFilePos', $nopAttrs['endFilePos']); |
| 1012 |
} |
| 1013 |
if (isset($nopAttrs['endTokenPos'])) { |
| 1014 |
$node->setAttribute('endTokenPos', $nopAttrs['endTokenPos']); |
| 1015 |
} |
| 1016 |
} |
| 1017 |
} |
| 1018 |
|
| 1019 |
protected function checkClassModifier(int $a, int $b, int $modifierPos): void { |
| 1020 |
try { |
| 1021 |
Modifiers::verifyClassModifier($a, $b); |
| 1022 |
} catch (Error $error) { |
| 1023 |
$error->setAttributes($this->getAttributesAt($modifierPos)); |
| 1024 |
$this->emitError($error); |
| 1025 |
} |
| 1026 |
} |
| 1027 |
|
| 1028 |
protected function checkModifier(int $a, int $b, int $modifierPos): void { |
| 1029 |
// Jumping through some hoops here because verifyModifier() is also used elsewhere |
| 1030 |
try { |
| 1031 |
Modifiers::verifyModifier($a, $b); |
| 1032 |
} catch (Error $error) { |
| 1033 |
$error->setAttributes($this->getAttributesAt($modifierPos)); |
| 1034 |
$this->emitError($error); |
| 1035 |
} |
| 1036 |
} |
| 1037 |
|
| 1038 |
protected function checkParam(Param $node): void { |
| 1039 |
if ($node->variadic && null !== $node->default) { |
| 1040 |
$this->emitError(new Error( |
| 1041 |
'Variadic parameter cannot have a default value', |
| 1042 |
$node->default->getAttributes() |
| 1043 |
)); |
| 1044 |
} |
| 1045 |
} |
| 1046 |
|
| 1047 |
protected function checkTryCatch(TryCatch $node): void { |
| 1048 |
if (empty($node->catches) && null === $node->finally) { |
| 1049 |
$this->emitError(new Error( |
| 1050 |
'Cannot use try without catch or finally', $node->getAttributes() |
| 1051 |
)); |
| 1052 |
} |
| 1053 |
} |
| 1054 |
|
| 1055 |
protected function checkNamespace(Namespace_ $node): void { |
| 1056 |
if (null !== $node->stmts) { |
| 1057 |
foreach ($node->stmts as $stmt) { |
| 1058 |
if ($stmt instanceof Namespace_) { |
| 1059 |
$this->emitError(new Error( |
| 1060 |
'Namespace declarations cannot be nested', $stmt->getAttributes() |
| 1061 |
)); |
| 1062 |
} |
| 1063 |
} |
| 1064 |
} |
| 1065 |
} |
| 1066 |
|
| 1067 |
private function checkClassName(?Identifier $name, int $namePos): void { |
| 1068 |
if (null !== $name && $name->isSpecialClassName()) { |
| 1069 |
$this->emitError(new Error( |
| 1070 |
sprintf('Cannot use \'%s\' as class name as it is reserved', $name), |
| 1071 |
$this->getAttributesAt($namePos) |
| 1072 |
)); |
| 1073 |
} |
| 1074 |
} |
| 1075 |
|
| 1076 |
/** @param Name[] $interfaces */ |
| 1077 |
private function checkImplementedInterfaces(array $interfaces): void { |
| 1078 |
foreach ($interfaces as $interface) { |
| 1079 |
if ($interface->isSpecialClassName()) { |
| 1080 |
$this->emitError(new Error( |
| 1081 |
sprintf('Cannot use \'%s\' as interface name as it is reserved', $interface), |
| 1082 |
$interface->getAttributes() |
| 1083 |
)); |
| 1084 |
} |
| 1085 |
} |
| 1086 |
} |
| 1087 |
|
| 1088 |
protected function checkClass(Class_ $node, int $namePos): void { |
| 1089 |
$this->checkClassName($node->name, $namePos); |
| 1090 |
|
| 1091 |
if ($node->extends && $node->extends->isSpecialClassName()) { |
| 1092 |
$this->emitError(new Error( |
| 1093 |
sprintf('Cannot use \'%s\' as class name as it is reserved', $node->extends), |
| 1094 |
$node->extends->getAttributes() |
| 1095 |
)); |
| 1096 |
} |
| 1097 |
|
| 1098 |
$this->checkImplementedInterfaces($node->implements); |
| 1099 |
} |
| 1100 |
|
| 1101 |
protected function checkInterface(Interface_ $node, int $namePos): void { |
| 1102 |
$this->checkClassName($node->name, $namePos); |
| 1103 |
$this->checkImplementedInterfaces($node->extends); |
| 1104 |
} |
| 1105 |
|
| 1106 |
protected function checkEnum(Enum_ $node, int $namePos): void { |
| 1107 |
$this->checkClassName($node->name, $namePos); |
| 1108 |
$this->checkImplementedInterfaces($node->implements); |
| 1109 |
} |
| 1110 |
|
| 1111 |
protected function checkClassMethod(ClassMethod $node, int $modifierPos): void { |
| 1112 |
if ($node->flags & Modifiers::STATIC) { |
| 1113 |
switch ($node->name->toLowerString()) { |
| 1114 |
case '__construct': |
| 1115 |
$this->emitError(new Error( |
| 1116 |
sprintf('Constructor %s() cannot be static', $node->name), |
| 1117 |
$this->getAttributesAt($modifierPos))); |
| 1118 |
break; |
| 1119 |
case '__destruct': |
| 1120 |
$this->emitError(new Error( |
| 1121 |
sprintf('Destructor %s() cannot be static', $node->name), |
| 1122 |
$this->getAttributesAt($modifierPos))); |
| 1123 |
break; |
| 1124 |
case '__clone': |
| 1125 |
$this->emitError(new Error( |
| 1126 |
sprintf('Clone method %s() cannot be static', $node->name), |
| 1127 |
$this->getAttributesAt($modifierPos))); |
| 1128 |
break; |
| 1129 |
} |
| 1130 |
} |
| 1131 |
|
| 1132 |
if ($node->flags & Modifiers::READONLY) { |
| 1133 |
$this->emitError(new Error( |
| 1134 |
sprintf('Method %s() cannot be readonly', $node->name), |
| 1135 |
$this->getAttributesAt($modifierPos))); |
| 1136 |
} |
| 1137 |
} |
| 1138 |
|
| 1139 |
protected function checkClassConst(ClassConst $node, int $modifierPos): void { |
| 1140 |
if ($node->flags & Modifiers::STATIC) { |
| 1141 |
$this->emitError(new Error( |
| 1142 |
"Cannot use 'static' as constant modifier", |
| 1143 |
$this->getAttributesAt($modifierPos))); |
| 1144 |
} |
| 1145 |
if ($node->flags & Modifiers::ABSTRACT) { |
| 1146 |
$this->emitError(new Error( |
| 1147 |
"Cannot use 'abstract' as constant modifier", |
| 1148 |
$this->getAttributesAt($modifierPos))); |
| 1149 |
} |
| 1150 |
if ($node->flags & Modifiers::READONLY) { |
| 1151 |
$this->emitError(new Error( |
| 1152 |
"Cannot use 'readonly' as constant modifier", |
| 1153 |
$this->getAttributesAt($modifierPos))); |
| 1154 |
} |
| 1155 |
} |
| 1156 |
|
| 1157 |
protected function checkProperty(Property $node, int $modifierPos): void { |
| 1158 |
if ($node->flags & Modifiers::ABSTRACT) { |
| 1159 |
$this->emitError(new Error('Properties cannot be declared abstract', |
| 1160 |
$this->getAttributesAt($modifierPos))); |
| 1161 |
} |
| 1162 |
|
| 1163 |
if ($node->flags & Modifiers::FINAL) { |
| 1164 |
$this->emitError(new Error('Properties cannot be declared final', |
| 1165 |
$this->getAttributesAt($modifierPos))); |
| 1166 |
} |
| 1167 |
} |
| 1168 |
|
| 1169 |
protected function checkUseUse(UseItem $node, int $namePos): void { |
| 1170 |
if ($node->alias && $node->alias->isSpecialClassName()) { |
| 1171 |
$this->emitError(new Error( |
| 1172 |
sprintf( |
| 1173 |
'Cannot use %s as %s because \'%2$s\' is a special class name', |
| 1174 |
$node->name, $node->alias |
| 1175 |
), |
| 1176 |
$this->getAttributesAt($namePos) |
| 1177 |
)); |
| 1178 |
} |
| 1179 |
} |
| 1180 |
|
| 1181 |
/** |
| 1182 |
* Creates the token map. |
| 1183 |
* |
| 1184 |
* The token map maps the PHP internal token identifiers |
| 1185 |
* to the identifiers used by the Parser. Additionally it |
| 1186 |
* maps T_OPEN_TAG_WITH_ECHO to T_ECHO and T_CLOSE_TAG to ';'. |
| 1187 |
* |
| 1188 |
* @return array<int, int> The token map |
| 1189 |
*/ |
| 1190 |
protected function createTokenMap(): array { |
| 1191 |
$tokenMap = []; |
| 1192 |
|
| 1193 |
for ($i = 0; $i < 1000; ++$i) { |
| 1194 |
if ($i < 256) { |
| 1195 |
// Single-char tokens use an identity mapping. |
| 1196 |
$tokenMap[$i] = $i; |
| 1197 |
} elseif (\T_DOUBLE_COLON === $i) { |
| 1198 |
// T_DOUBLE_COLON is equivalent to T_PAAMAYIM_NEKUDOTAYIM |
| 1199 |
$tokenMap[$i] = static::T_PAAMAYIM_NEKUDOTAYIM; |
| 1200 |
} elseif (\T_OPEN_TAG_WITH_ECHO === $i) { |
| 1201 |
// T_OPEN_TAG_WITH_ECHO with dropped T_OPEN_TAG results in T_ECHO |
| 1202 |
$tokenMap[$i] = static::T_ECHO; |
| 1203 |
} elseif (\T_CLOSE_TAG === $i) { |
| 1204 |
// T_CLOSE_TAG is equivalent to ';' |
| 1205 |
$tokenMap[$i] = ord(';'); |
| 1206 |
} elseif ('UNKNOWN' !== $name = token_name($i)) { |
| 1207 |
if (defined($name = static::class . '::' . $name)) { |
| 1208 |
// Other tokens can be mapped directly |
| 1209 |
$tokenMap[$i] = constant($name); |
| 1210 |
} |
| 1211 |
} |
| 1212 |
} |
| 1213 |
|
| 1214 |
// Assign tokens for which we define compatibility constants, as token_name() does not know them. |
| 1215 |
$tokenMap[\T_FN] = static::T_FN; |
| 1216 |
$tokenMap[\T_COALESCE_EQUAL] = static::T_COALESCE_EQUAL; |
| 1217 |
$tokenMap[\T_NAME_QUALIFIED] = static::T_NAME_QUALIFIED; |
| 1218 |
$tokenMap[\T_NAME_FULLY_QUALIFIED] = static::T_NAME_FULLY_QUALIFIED; |
| 1219 |
$tokenMap[\T_NAME_RELATIVE] = static::T_NAME_RELATIVE; |
| 1220 |
$tokenMap[\T_MATCH] = static::T_MATCH; |
| 1221 |
$tokenMap[\T_NULLSAFE_OBJECT_OPERATOR] = static::T_NULLSAFE_OBJECT_OPERATOR; |
| 1222 |
$tokenMap[\T_ATTRIBUTE] = static::T_ATTRIBUTE; |
| 1223 |
$tokenMap[\T_AMPERSAND_NOT_FOLLOWED_BY_VAR_OR_VARARG] = static::T_AMPERSAND_NOT_FOLLOWED_BY_VAR_OR_VARARG; |
| 1224 |
$tokenMap[\T_AMPERSAND_FOLLOWED_BY_VAR_OR_VARARG] = static::T_AMPERSAND_FOLLOWED_BY_VAR_OR_VARARG; |
| 1225 |
$tokenMap[\T_ENUM] = static::T_ENUM; |
| 1226 |
$tokenMap[\T_READONLY] = static::T_READONLY; |
| 1227 |
|
| 1228 |
// We have create a map from PHP token IDs to external symbol IDs. |
| 1229 |
// Now map them to the internal symbol ID. |
| 1230 |
$fullTokenMap = []; |
| 1231 |
foreach ($tokenMap as $phpToken => $extSymbol) { |
| 1232 |
$intSymbol = $this->tokenToSymbol[$extSymbol]; |
| 1233 |
if ($intSymbol === $this->invalidSymbol) { |
| 1234 |
continue; |
| 1235 |
} |
| 1236 |
$fullTokenMap[$phpToken] = $intSymbol; |
| 1237 |
} |
| 1238 |
|
| 1239 |
return $fullTokenMap; |
| 1240 |
} |
| 1241 |
} |
| 1242 |
|
