Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Commit c9a1c09

Browse files
RichardBradleyfabpot
RichardBradley
authored and
fabpot
committed
#20411 fix Yaml parsing for very long quoted strings
1 parent 01a0250 commit c9a1c09

File tree

3 files changed

+77
-39
lines changed

3 files changed

+77
-39
lines changed

‎src/Symfony/Component/Yaml/Inline.php

Lines changed: 8 additions & 8 deletions
Original file line numberDiff line numberDiff line change
@@ -149,8 +149,8 @@ public static function dump($value, $exceptionOnInvalidType = false, $objectSupp
149149
case Escaper::requiresDoubleQuoting($value):
150150
return Escaper::escapeWithDoubleQuotes($value);
151151
case Escaper::requiresSingleQuoting($value):
152-
case preg_match(self::getHexRegex(), $value):
153-
case preg_match(self::getTimestampRegex(), $value):
152+
case Parser::preg_match(self::getHexRegex(), $value):
153+
case Parser::preg_match(self::getTimestampRegex(), $value):
154154
return Escaper::escapeWithSingleQuotes($value);
155155
default:
156156
return $value;
@@ -242,10 +242,10 @@ public static function parseScalar($scalar, $delimiters = null, $stringDelimiter
242242
$i += strlen($output);
243243

244244
// remove comments
245-
if (preg_match('/[ \t]+#/', $output, $match, PREG_OFFSET_CAPTURE)) {
245+
if (Parser::preg_match('/[ \t]+#/', $output, $match, PREG_OFFSET_CAPTURE)) {
246246
$output = substr($output, 0, $match[0][1]);
247247
}
248-
} elseif (preg_match('/^(.+?)('.implode('|', $delimiters).')/', substr($scalar, $i), $match)) {
248+
} elseif (Parser::preg_match('/^(.+?)('.implode('|', $delimiters).')/', substr($scalar, $i), $match)) {
249249
$output = $match[1];
250250
$i += strlen($output);
251251
} else {
@@ -272,7 +272,7 @@ public static function parseScalar($scalar, $delimiters = null, $stringDelimiter
272272
*/
273273
private static function parseQuotedScalar($scalar, &$i)
274274
{
275-
if (!preg_match('/'.self::REGEX_QUOTED_STRING.'/Au', substr($scalar, $i), $match)) {
275+
if (!Parser::preg_match('/'.self::REGEX_QUOTED_STRING.'/Au', substr($scalar, $i), $match)) {
276276
throw new ParseException(sprintf('Malformed inline YAML string: %s.', substr($scalar, $i)));
277277
}
278278

@@ -520,16 +520,16 @@ private static function evaluateScalar($scalar, $references = array())
520520

521521
return '0' == $scalar[1] ? octdec($scalar) : (((string) $raw === (string) $cast) ? $cast : $raw);
522522
case is_numeric($scalar):
523-
case preg_match(self::getHexRegex(), $scalar):
523+
case Parser::preg_match(self::getHexRegex(), $scalar):
524524
return '0x' === $scalar[0].$scalar[1] ? hexdec($scalar) : (float) $scalar;
525525
case '.inf' === $scalarLower:
526526
case '.nan' === $scalarLower:
527527
return -log(0);
528528
case '-.inf' === $scalarLower:
529529
return log(0);
530-
case preg_match('/^(-|\+)?[0-9,]+(\.[0-9]+)?$/', $scalar):
530+
case Parser::preg_match('/^(-|\+)?[0-9,]+(\.[0-9]+)?$/', $scalar):
531531
return (float) str_replace(',', '', $scalar);
532-
case preg_match(self::getTimestampRegex(), $scalar):
532+
case Parser::preg_match(self::getTimestampRegex(), $scalar):
533533
$timeZone = date_default_timezone_get();
534534
date_default_timezone_set('UTC');
535535
$time = strtotime($scalar);

‎src/Symfony/Component/Yaml/Parser.php

Lines changed: 57 additions & 31 deletions
Original file line numberDiff line numberDiff line change
@@ -61,7 +61,7 @@ public function __construct($offset = 0, $totalNumberOfLines = null, array $skip
6161
*/
6262
public function parse($value, $exceptionOnInvalidType = false, $objectSupport = false, $objectForMap = false)
6363
{
64-
if (!preg_match('//u', $value)) {
64+
if (false === preg_match('//u', $value)) {
6565
throw new ParseException('The YAML value does not appear to be valid UTF-8.');
6666
}
6767
$this->currentLineNb = -1;
@@ -92,13 +92,13 @@ public function parse($value, $exceptionOnInvalidType = false, $objectSupport =
9292
}
9393

9494
$isRef = $mergeNode = false;
95-
if (preg_match('#^\-((?P<leadspaces>\s+)(?P<value>.+?))?\s*$#u', $this->currentLine, $values)) {
95+
if (self::preg_match('#^\-((?P<leadspaces>\s+)(?P<value>.+))?$#u', rtrim($this->currentLine), $values)) {
9696
if ($context && 'mapping' == $context) {
9797
throw new ParseException('You cannot define a sequence item when in a mapping', $this->getRealCurrentLineNb() + 1, $this->currentLine);
9898
}
9999
$context = 'sequence';
100100

101-
if (isset($values['value']) && preg_match('#^&(?P<ref>[^ ]+) *(?P<value>.*)#u', $values['value'], $matches)) {
101+
if (isset($values['value']) && self::preg_match('#^&(?P<ref>[^ ]+) *(?P<value>.*)#u', $values['value'], $matches)) {
102102
$isRef = $matches['ref'];
103103
$values['value'] = $matches['value'];
104104
}
@@ -108,7 +108,7 @@ public function parse($value, $exceptionOnInvalidType = false, $objectSupport =
108108
$data[] = $this->parseBlock($this->getRealCurrentLineNb() + 1, $this->getNextEmbedBlock(null, true), $exceptionOnInvalidType, $objectSupport, $objectForMap);
109109
} else {
110110
if (isset($values['leadspaces'])
111-
&& preg_match('#^(?P<key>'.Inline::REGEX_QUOTED_STRING.'|[^ \'"\{\[].*?) *\:(\s+(?P<value>.+?))?\s*$#u', $values['value'], $matches)
111+
&& self::preg_match('#^(?P<key>'.Inline::REGEX_QUOTED_STRING.'|[^ \'"\{\[].*?) *\:(\s+(?P<value>.+))?$#u', rtrim($values['value']), $matches)
112112
) {
113113
// this is a compact notation element, add to next block and parse
114114
$block = $values['value'];
@@ -124,7 +124,10 @@ public function parse($value, $exceptionOnInvalidType = false, $objectSupport =
124124
if ($isRef) {
125125
$this->refs[$isRef] = end($data);
126126
}
127-
} 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('"', "'")))) {
127+
} elseif (
128+
self::preg_match('#^(?P<key>'.Inline::REGEX_QUOTED_STRING.'|[^ \'"\[\{].*?) *\:(\s+(?P<value>.+))?$#u', rtrim($this->currentLine), $values)
129+
&& (false === strpos($values['key'], ' #') || in_array($values['key'][0], array('"', "'")))
130+
) {
128131
if ($context && 'sequence' == $context) {
129132
throw new ParseException('You cannot define a mapping item when in a sequence', $this->currentLineNb + 1, $this->currentLine);
130133
}
@@ -203,7 +206,7 @@ public function parse($value, $exceptionOnInvalidType = false, $objectSupport =
203206
}
204207
}
205208
}
206-
} elseif (isset($values['value']) && preg_match('#^&(?P<ref>[^ ]+) *(?P<value>.*)#u', $values['value'], $matches)) {
209+
} elseif (isset($values['value']) && self::preg_match('#^&(?P<ref>[^ ]+) *(?P<value>.*)#u', $values['value'], $matches)) {
207210
$isRef = $matches['ref'];
208211
$values['value'] = $matches['value'];
209212
}
@@ -266,27 +269,7 @@ public function parse($value, $exceptionOnInvalidType = false, $objectSupport =
266269
return $value;
267270
}
268271

269-
switch (preg_last_error()) {
270-
case PREG_INTERNAL_ERROR:
271-
$error = 'Internal PCRE error.';
272-
break;
273-
case PREG_BACKTRACK_LIMIT_ERROR:
274-
$error = 'pcre.backtrack_limit reached.';
275-
break;
276-
case PREG_RECURSION_LIMIT_ERROR:
277-
$error = 'pcre.recursion_limit reached.';
278-
break;
279-
case PREG_BAD_UTF8_ERROR:
280-
$error = 'Malformed UTF-8 data.';
281-
break;
282-
case PREG_BAD_UTF8_OFFSET_ERROR:
283-
$error = 'Offset doesn\'t correspond to the begin of a valid UTF-8 code point.';
284-
break;
285-
default:
286-
$error = 'Unable to parse.';
287-
}
288-
289-
throw new ParseException($error, $this->getRealCurrentLineNb() + 1, $this->currentLine);
272+
throw new ParseException('Unable to parse', $this->getRealCurrentLineNb() + 1, $this->currentLine);
290273
}
291274
}
292275

@@ -520,7 +503,7 @@ private function parseValue($value, $exceptionOnInvalidType, $objectSupport, $ob
520503
return $this->refs[$value];
521504
}
522505

523-
if (preg_match('/^'.self::BLOCK_SCALAR_HEADER_PATTERN.'$/', $value, $matches)) {
506+
if (self::preg_match('/^'.self::BLOCK_SCALAR_HEADER_PATTERN.'$/', $value, $matches)) {
524507
$modifiers = isset($matches['modifiers']) ? $matches['modifiers'] : '';
525508

526509
return $this->parseBlockScalar($matches['separator'], preg_replace('#\d+#', '', $modifiers), (int) abs($modifiers));
@@ -566,7 +549,7 @@ private function parseBlockScalar($style, $chomping = '', $indentation = 0)
566549

567550
// determine indentation if not specified
568551
if (0 === $indentation) {
569-
if (preg_match('/^ +/', $this->currentLine, $matches)) {
552+
if (self::preg_match('/^ +/', $this->currentLine, $matches)) {
570553
$indentation = strlen($matches[0]);
571554
}
572555
}
@@ -577,7 +560,7 @@ private function parseBlockScalar($style, $chomping = '', $indentation = 0)
577560
while (
578561
$notEOF && (
579562
$isCurrentLineBlank ||
580-
preg_match($pattern, $this->currentLine, $matches)
563+
self::preg_match($pattern, $this->currentLine, $matches)
581564
)
582565
) {
583566
if ($isCurrentLineBlank && strlen($this->currentLine) > $indentation) {
@@ -800,6 +783,49 @@ private function isStringUnIndentedCollectionItem()
800783
*/
801784
private function isBlockScalarHeader()
802785
{
803-
return (bool) preg_match('~'.self::BLOCK_SCALAR_HEADER_PATTERN.'$~', $this->currentLine);
786+
return (bool) self::preg_match('~'.self::BLOCK_SCALAR_HEADER_PATTERN.'$~', $this->currentLine);
787+
}
788+
789+
/**
790+
* A local wrapper for `preg_match` which will throw a ParseException if there
791+
* is an internal error in the PCRE engine.
792+
*
793+
* This avoids us needing to check for "false" every time PCRE is used
794+
* in the YAML engine
795+
*
796+
* @throws ParseException on a PCRE internal error
797+
*
798+
* @see preg_last_error()
799+
*
800+
* @internal
801+
*/
802+
public static function preg_match($pattern, $subject, &$matches = null, $flags = 0, $offset = 0)
803+
{
804+
$ret = preg_match($pattern, $subject, $matches, $flags, $offset);
805+
if ($ret === false) {
806+
switch (preg_last_error()) {
807+
case PREG_INTERNAL_ERROR:
808+
$error = 'Internal PCRE error.';
809+
break;
810+
case PREG_BACKTRACK_LIMIT_ERROR:
811+
$error = 'pcre.backtrack_limit reached.';
812+
break;
813+
case PREG_RECURSION_LIMIT_ERROR:
814+
$error = 'pcre.recursion_limit reached.';
815+
break;
816+
case PREG_BAD_UTF8_ERROR:
817+
$error = 'Malformed UTF-8 data.';
818+
break;
819+
case PREG_BAD_UTF8_OFFSET_ERROR:
820+
$error = 'Offset doesn\'t correspond to the begin of a valid UTF-8 code point.';
821+
break;
822+
default:
823+
$error = 'Error.';
824+
}
825+
826+
throw new ParseException($error);
827+
}
828+
829+
return $ret;
804830
}
805831
}

‎src/Symfony/Component/Yaml/Tests/ParserTest.php

Lines changed: 12 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -16,6 +16,7 @@
1616

1717
class ParserTest extends \PHPUnit_Framework_TestCase
1818
{
19+
/** @var Parser */
1920
protected $parser;
2021

2122
protected function setUp()
@@ -1143,6 +1144,17 @@ public function parserThrowsExceptionWithCorrectLineNumberProvider()
11431144
),
11441145
);
11451146
}
1147+
1148+
public function testCanParseVeryLongValue()
1149+
{
1150+
$longStringWithSpaces = str_repeat('xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx ', 20000);
1151+
$trickyVal = array('x' => $longStringWithSpaces);
1152+
1153+
$yamlString = Yaml::dump($trickyVal);
1154+
$arrayFromYaml = $this->parser->parse($yamlString);
1155+
1156+
$this->assertEquals($trickyVal, $arrayFromYaml);
1157+
}
11461158
}
11471159

11481160
class B

0 commit comments

Comments
 (0)