View source
<?php
namespace Symfony\Component\Yaml;
use Symfony\Component\Yaml\Exception\ParseException;
class Parser {
const BLOCK_SCALAR_HEADER_PATTERN = '(?P<separator>\\||>)(?P<modifiers>\\+|\\-|\\d+|\\+\\d+|\\-\\d+|\\d+\\+|\\d+\\-)?(?P<comments> +#.*)?';
const FOLDED_SCALAR_PATTERN = self::BLOCK_SCALAR_HEADER_PATTERN;
private $offset = 0;
private $totalNumberOfLines;
private $lines = array();
private $currentLineNb = -1;
private $currentLine = '';
private $refs = array();
public function __construct($offset = 0, $totalNumberOfLines = null) {
$this->offset = $offset;
$this->totalNumberOfLines = $totalNumberOfLines;
}
public function parse($value, $exceptionOnInvalidType = false, $objectSupport = false) {
if (!preg_match('//u', $value)) {
throw new ParseException('The YAML value does not appear to be valid UTF-8.');
}
$this->currentLineNb = -1;
$this->currentLine = '';
$value = $this
->cleanup($value);
$this->lines = explode("\n", $value);
if (null === $this->totalNumberOfLines) {
$this->totalNumberOfLines = count($this->lines);
}
if (function_exists('mb_internal_encoding') && (int) ini_get('mbstring.func_overload') & 2) {
$mbEncoding = mb_internal_encoding();
mb_internal_encoding('UTF-8');
}
$data = array();
$context = null;
while ($this
->moveToNextLine()) {
if ($this
->isCurrentLineEmpty()) {
continue;
}
if ("\t" === $this->currentLine[0]) {
throw new ParseException('A YAML file cannot contain tabs as indentation.', $this
->getRealCurrentLineNb() + 1, $this->currentLine);
}
$isRef = $isInPlace = $isProcessed = false;
if (preg_match('#^\\-((?P<leadspaces>\\s+)(?P<value>.+?))?\\s*$#u', $this->currentLine, $values)) {
if ($context && 'mapping' == $context) {
throw new ParseException('You cannot define a sequence item when in a mapping', $this
->getRealCurrentLineNb() + 1, $this->currentLine);
}
$context = 'sequence';
if (isset($values['value']) && preg_match('#^&(?P<ref>[^ ]+) *(?P<value>.*)#u', $values['value'], $matches)) {
$isRef = $matches['ref'];
$values['value'] = $matches['value'];
}
if (!isset($values['value']) || '' == trim($values['value'], ' ') || 0 === strpos(ltrim($values['value'], ' '), '#')) {
$c = $this
->getRealCurrentLineNb() + 1;
$parser = new self($c, $this->totalNumberOfLines);
$parser->refs =& $this->refs;
$data[] = $parser
->parse($this
->getNextEmbedBlock(null, true), $exceptionOnInvalidType, $objectSupport);
}
else {
if (isset($values['leadspaces']) && preg_match('#^(?P<key>' . Inline::REGEX_QUOTED_STRING . '|[^ \'"\\{\\[].*?) *\\:(\\s+(?P<value>.+?))?\\s*$#u', $values['value'], $matches)) {
$c = $this
->getRealCurrentLineNb();
$parser = new self($c, $this->totalNumberOfLines);
$parser->refs =& $this->refs;
$block = $values['value'];
if ($this
->isNextLineIndented()) {
$block .= "\n" . $this
->getNextEmbedBlock($this
->getCurrentLineIndentation() + strlen($values['leadspaces']) + 1);
}
$data[] = $parser
->parse($block, $exceptionOnInvalidType, $objectSupport);
}
else {
$data[] = $this
->parseValue($values['value'], $exceptionOnInvalidType, $objectSupport);
}
}
if ($isRef) {
$this->refs[$isRef] = end($data);
}
}
elseif (preg_match('#^(?P<key>' . Inline::REGEX_QUOTED_STRING . '|[^ \'"\\[\\{].*?) *\\:(\\s+(?P<value>.+?))?\\s*$#u', $this->currentLine, $values) && (false === strpos($values['key'], ' #') || in_array($values['key'][0], array(
'"',
"'",
)))) {
if ($context && 'sequence' == $context) {
throw new ParseException('You cannot define a mapping item when in a sequence', $this->currentLineNb + 1, $this->currentLine);
}
$context = 'mapping';
Inline::parse(null, $exceptionOnInvalidType, $objectSupport, $this->refs);
try {
$key = Inline::parseScalar($values['key']);
} catch (ParseException $e) {
$e
->setParsedLine($this
->getRealCurrentLineNb() + 1);
$e
->setSnippet($this->currentLine);
throw $e;
}
if (is_float($key)) {
$key = (string) $key;
}
if ('<<' === $key) {
if (isset($values['value']) && 0 === strpos($values['value'], '*')) {
$isInPlace = substr($values['value'], 1);
if (!array_key_exists($isInPlace, $this->refs)) {
throw new ParseException(sprintf('Reference "%s" does not exist.', $isInPlace), $this
->getRealCurrentLineNb() + 1, $this->currentLine);
}
}
else {
if (isset($values['value']) && $values['value'] !== '') {
$value = $values['value'];
}
else {
$value = $this
->getNextEmbedBlock();
}
$c = $this
->getRealCurrentLineNb() + 1;
$parser = new self($c, $this->totalNumberOfLines);
$parser->refs =& $this->refs;
$parsed = $parser
->parse($value, $exceptionOnInvalidType, $objectSupport);
$merged = array();
if (!is_array($parsed)) {
throw new ParseException('YAML merge keys used with a scalar value instead of an array.', $this
->getRealCurrentLineNb() + 1, $this->currentLine);
}
elseif (isset($parsed[0])) {
foreach (array_reverse($parsed) as $parsedItem) {
if (!is_array($parsedItem)) {
throw new ParseException('Merge items must be arrays.', $this
->getRealCurrentLineNb() + 1, $parsedItem);
}
$merged = array_merge($parsedItem, $merged);
}
}
else {
$merged = array_merge($merged, $parsed);
}
$isProcessed = $merged;
}
}
elseif (isset($values['value']) && preg_match('#^&(?P<ref>[^ ]+) *(?P<value>.*)#u', $values['value'], $matches)) {
$isRef = $matches['ref'];
$values['value'] = $matches['value'];
}
if ($isProcessed) {
$data = $isProcessed;
}
elseif (!isset($values['value']) || '' == trim($values['value'], ' ') || 0 === strpos(ltrim($values['value'], ' '), '#')) {
if (!$this
->isNextLineIndented() && !$this
->isNextLineUnIndentedCollection()) {
$data[$key] = null;
}
else {
$c = $this
->getRealCurrentLineNb() + 1;
$parser = new self($c, $this->totalNumberOfLines);
$parser->refs =& $this->refs;
$data[$key] = $parser
->parse($this
->getNextEmbedBlock(), $exceptionOnInvalidType, $objectSupport);
}
}
else {
if ($isInPlace) {
$data = $this->refs[$isInPlace];
}
else {
$data[$key] = $this
->parseValue($values['value'], $exceptionOnInvalidType, $objectSupport);
}
}
if ($isRef) {
$this->refs[$isRef] = $data[$key];
}
}
else {
if ('---' === $this->currentLine) {
throw new ParseException('Multiple documents are not supported.', $this->currentLineNb + 1, $this->currentLine);
}
if ($this->lines[0] === trim($value)) {
try {
$value = Inline::parse($this->lines[0], $exceptionOnInvalidType, $objectSupport, $this->refs);
} catch (ParseException $e) {
$e
->setParsedLine($this
->getRealCurrentLineNb() + 1);
$e
->setSnippet($this->currentLine);
throw $e;
}
if (is_array($value)) {
$first = reset($value);
if (is_string($first) && 0 === strpos($first, '*')) {
$data = array();
foreach ($value as $alias) {
$data[] = $this->refs[substr($alias, 1)];
}
$value = $data;
}
}
if (isset($mbEncoding)) {
mb_internal_encoding($mbEncoding);
}
return $value;
}
switch (preg_last_error()) {
case PREG_INTERNAL_ERROR:
$error = 'Internal PCRE error.';
break;
case PREG_BACKTRACK_LIMIT_ERROR:
$error = 'pcre.backtrack_limit reached.';
break;
case PREG_RECURSION_LIMIT_ERROR:
$error = 'pcre.recursion_limit reached.';
break;
case PREG_BAD_UTF8_ERROR:
$error = 'Malformed UTF-8 data.';
break;
case PREG_BAD_UTF8_OFFSET_ERROR:
$error = 'Offset doesn\'t correspond to the begin of a valid UTF-8 code point.';
break;
default:
$error = 'Unable to parse.';
}
throw new ParseException($error, $this
->getRealCurrentLineNb() + 1, $this->currentLine);
}
}
if (isset($mbEncoding)) {
mb_internal_encoding($mbEncoding);
}
return empty($data) ? null : $data;
}
private function getRealCurrentLineNb() {
return $this->currentLineNb + $this->offset;
}
private function getCurrentLineIndentation() {
return strlen($this->currentLine) - strlen(ltrim($this->currentLine, ' '));
}
private function getNextEmbedBlock($indentation = null, $inSequence = false) {
$oldLineIndentation = $this
->getCurrentLineIndentation();
$blockScalarIndentations = array();
if ($this
->isBlockScalarHeader()) {
$blockScalarIndentations[] = $this
->getCurrentLineIndentation();
}
if (!$this
->moveToNextLine()) {
return;
}
if (null === $indentation) {
$newIndent = $this
->getCurrentLineIndentation();
$unindentedEmbedBlock = $this
->isStringUnIndentedCollectionItem();
if (!$this
->isCurrentLineEmpty() && 0 === $newIndent && !$unindentedEmbedBlock) {
throw new ParseException('Indentation problem.', $this
->getRealCurrentLineNb() + 1, $this->currentLine);
}
}
else {
$newIndent = $indentation;
}
$data = array();
if ($this
->getCurrentLineIndentation() >= $newIndent) {
$data[] = substr($this->currentLine, $newIndent);
}
else {
$this
->moveToPreviousLine();
return;
}
if ($inSequence && $oldLineIndentation === $newIndent && isset($data[0][0]) && '-' === $data[0][0]) {
$this
->moveToPreviousLine();
return;
}
$isItUnindentedCollection = $this
->isStringUnIndentedCollectionItem();
if (empty($blockScalarIndentations) && $this
->isBlockScalarHeader()) {
$blockScalarIndentations[] = $this
->getCurrentLineIndentation();
}
$previousLineIndentation = $this
->getCurrentLineIndentation();
while ($this
->moveToNextLine()) {
$indent = $this
->getCurrentLineIndentation();
if (!empty($blockScalarIndentations) && $indent < $previousLineIndentation && trim($this->currentLine) !== '') {
foreach ($blockScalarIndentations as $key => $blockScalarIndentation) {
if ($blockScalarIndentation >= $this
->getCurrentLineIndentation()) {
unset($blockScalarIndentations[$key]);
}
}
}
if (empty($blockScalarIndentations) && !$this
->isCurrentLineComment() && $this
->isBlockScalarHeader()) {
$blockScalarIndentations[] = $this
->getCurrentLineIndentation();
}
$previousLineIndentation = $indent;
if ($isItUnindentedCollection && !$this
->isStringUnIndentedCollectionItem() && $newIndent === $indent) {
$this
->moveToPreviousLine();
break;
}
if ($this
->isCurrentLineBlank()) {
$data[] = substr($this->currentLine, $newIndent);
continue;
}
if (empty($blockScalarIndentations) && $this
->isCurrentLineComment()) {
continue;
}
if ($indent >= $newIndent) {
$data[] = substr($this->currentLine, $newIndent);
}
elseif (0 == $indent) {
$this
->moveToPreviousLine();
break;
}
else {
throw new ParseException('Indentation problem.', $this
->getRealCurrentLineNb() + 1, $this->currentLine);
}
}
return implode("\n", $data);
}
private function moveToNextLine() {
if ($this->currentLineNb >= count($this->lines) - 1) {
return false;
}
$this->currentLine = $this->lines[++$this->currentLineNb];
return true;
}
private function moveToPreviousLine() {
$this->currentLine = $this->lines[--$this->currentLineNb];
}
private function parseValue($value, $exceptionOnInvalidType, $objectSupport) {
if (0 === strpos($value, '*')) {
if (false !== ($pos = strpos($value, '#'))) {
$value = substr($value, 1, $pos - 2);
}
else {
$value = substr($value, 1);
}
if (!array_key_exists($value, $this->refs)) {
throw new ParseException(sprintf('Reference "%s" does not exist.', $value), $this->currentLineNb + 1, $this->currentLine);
}
return $this->refs[$value];
}
if (preg_match('/^' . self::BLOCK_SCALAR_HEADER_PATTERN . '$/', $value, $matches)) {
$modifiers = isset($matches['modifiers']) ? $matches['modifiers'] : '';
return $this
->parseBlockScalar($matches['separator'], preg_replace('#\\d+#', '', $modifiers), (int) abs($modifiers));
}
try {
return Inline::parse($value, $exceptionOnInvalidType, $objectSupport, $this->refs);
} catch (ParseException $e) {
$e
->setParsedLine($this
->getRealCurrentLineNb() + 1);
$e
->setSnippet($this->currentLine);
throw $e;
}
}
private function parseBlockScalar($style, $chomping = '', $indentation = 0) {
$notEOF = $this
->moveToNextLine();
if (!$notEOF) {
return '';
}
$isCurrentLineBlank = $this
->isCurrentLineBlank();
$blockLines = array();
while ($notEOF && $isCurrentLineBlank) {
if ($notEOF = $this
->moveToNextLine()) {
$blockLines[] = '';
$isCurrentLineBlank = $this
->isCurrentLineBlank();
}
}
if (0 === $indentation) {
if (preg_match('/^ +/', $this->currentLine, $matches)) {
$indentation = strlen($matches[0]);
}
}
if ($indentation > 0) {
$pattern = sprintf('/^ {%d}(.*)$/', $indentation);
while ($notEOF && ($isCurrentLineBlank || preg_match($pattern, $this->currentLine, $matches))) {
if ($isCurrentLineBlank && strlen($this->currentLine) > $indentation) {
$blockLines[] = substr($this->currentLine, $indentation);
}
elseif ($isCurrentLineBlank) {
$blockLines[] = '';
}
else {
$blockLines[] = $matches[1];
}
if ($notEOF = $this
->moveToNextLine()) {
$isCurrentLineBlank = $this
->isCurrentLineBlank();
}
}
}
elseif ($notEOF) {
$blockLines[] = '';
}
if ($notEOF) {
$blockLines[] = '';
$this
->moveToPreviousLine();
}
elseif (!$notEOF && !$this
->isCurrentLineLastLineInDocument()) {
$blockLines[] = '';
}
if ('>' === $style) {
$text = '';
$previousLineIndented = false;
$previousLineBlank = false;
for ($i = 0; $i < count($blockLines); ++$i) {
if ('' === $blockLines[$i]) {
$text .= "\n";
$previousLineIndented = false;
$previousLineBlank = true;
}
elseif (' ' === $blockLines[$i][0]) {
$text .= "\n" . $blockLines[$i];
$previousLineIndented = true;
$previousLineBlank = false;
}
elseif ($previousLineIndented) {
$text .= "\n" . $blockLines[$i];
$previousLineIndented = false;
$previousLineBlank = false;
}
elseif ($previousLineBlank || 0 === $i) {
$text .= $blockLines[$i];
$previousLineIndented = false;
$previousLineBlank = false;
}
else {
$text .= ' ' . $blockLines[$i];
$previousLineIndented = false;
$previousLineBlank = false;
}
}
}
else {
$text = implode("\n", $blockLines);
}
if ('' === $chomping) {
$text = preg_replace('/\\n+$/', "\n", $text);
}
elseif ('-' === $chomping) {
$text = preg_replace('/\\n+$/', '', $text);
}
return $text;
}
private function isNextLineIndented() {
$currentIndentation = $this
->getCurrentLineIndentation();
$EOF = !$this
->moveToNextLine();
while (!$EOF && $this
->isCurrentLineEmpty()) {
$EOF = !$this
->moveToNextLine();
}
if ($EOF) {
return false;
}
$ret = false;
if ($this
->getCurrentLineIndentation() > $currentIndentation) {
$ret = true;
}
$this
->moveToPreviousLine();
return $ret;
}
private function isCurrentLineEmpty() {
return $this
->isCurrentLineBlank() || $this
->isCurrentLineComment();
}
private function isCurrentLineBlank() {
return '' == trim($this->currentLine, ' ');
}
private function isCurrentLineComment() {
$ltrimmedLine = ltrim($this->currentLine, ' ');
return '' !== $ltrimmedLine && $ltrimmedLine[0] === '#';
}
private function isCurrentLineLastLineInDocument() {
return $this->offset + $this->currentLineNb >= $this->totalNumberOfLines - 1;
}
private function cleanup($value) {
$value = str_replace(array(
"\r\n",
"\r",
), "\n", $value);
$count = 0;
$value = preg_replace('#^\\%YAML[: ][\\d\\.]+.*\\n#u', '', $value, -1, $count);
$this->offset += $count;
$trimmedValue = preg_replace('#^(\\#.*?\\n)+#s', '', $value, -1, $count);
if ($count == 1) {
$this->offset += substr_count($value, "\n") - substr_count($trimmedValue, "\n");
$value = $trimmedValue;
}
$trimmedValue = preg_replace('#^\\-\\-\\-.*?\\n#s', '', $value, -1, $count);
if ($count == 1) {
$this->offset += substr_count($value, "\n") - substr_count($trimmedValue, "\n");
$value = $trimmedValue;
$value = preg_replace('#\\.\\.\\.\\s*$#', '', $value);
}
return $value;
}
private function isNextLineUnIndentedCollection() {
$currentIndentation = $this
->getCurrentLineIndentation();
$notEOF = $this
->moveToNextLine();
while ($notEOF && $this
->isCurrentLineEmpty()) {
$notEOF = $this
->moveToNextLine();
}
if (false === $notEOF) {
return false;
}
$ret = false;
if ($this
->getCurrentLineIndentation() == $currentIndentation && $this
->isStringUnIndentedCollectionItem()) {
$ret = true;
}
$this
->moveToPreviousLine();
return $ret;
}
private function isStringUnIndentedCollectionItem() {
return '-' === rtrim($this->currentLine) || 0 === strpos($this->currentLine, '- ');
}
private function isBlockScalarHeader() {
return (bool) preg_match('~' . self::BLOCK_SCALAR_HEADER_PATTERN . '$~', $this->currentLine);
}
}