mirror/parsedown
1
0
Fork 0
mirror of https://github.com/erusev/parsedown.git synced 2023-08-10 21:13:06 +03:00
Code Releases Activity

Merge pull request #614 from aidantwoods/enhancement/performance-tweaks

Browse Source 
General and performance tweaks
This commit is contained in:
Aidan Woods 2018-05-07 11:18:21 +01:00 committed by GitHub
commit 8d172a2994
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 133 additions and 162 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

243
Parsedown.php
View File

@ -182,29 +182,17 @@ class Parsedown
continue; continue;
} }
if (strpos($line, "\t") !== false) while (($beforeTab = strstr($line, "\t", true)) !== false)
{ {
$parts = explode("\t", $line); $shortage = 4 - mb_strlen($beforeTab, 'utf-8') % 4;
$line = $parts[0]; $line = $beforeTab
. str_repeat(' ', $shortage)
unset($parts[0]); . substr($line, strlen($beforeTab) + 1)
;
foreach ($parts as $part)
{
$shortage = 4 - mb_strlen($line, 'utf-8') % 4;
$line .= str_repeat(' ', $shortage);
$line .= $part;
}
} }
$indent = 0; $indent = strspn($line, ' ');
while (isset($line[$indent]) and $line[$indent] === ' ')
{
$indent ++;
}
$text = $indent > 0 ? substr($line, $indent) : $line; $text = $indent > 0 ? substr($line, $indent) : $line;
@ -216,7 +204,8 @@ class Parsedown
if (isset($CurrentBlock['continuable'])) if (isset($CurrentBlock['continuable']))
{ {
$Block = $this->{'block'.$CurrentBlock['type'].'Continue'}($Line, $CurrentBlock); $methodName = 'block' . $CurrentBlock['type'] . 'Continue';
$Block = $this->$methodName($Line, $CurrentBlock);
if (isset($Block)) if (isset($Block))
{ {
@ -228,7 +217,8 @@ class Parsedown
{ {
if ($this->isBlockCompletable($CurrentBlock['type'])) if ($this->isBlockCompletable($CurrentBlock['type']))
{ {
$CurrentBlock = $this->{'block'.$CurrentBlock['type'].'Complete'}($CurrentBlock); $methodName = 'block' . $CurrentBlock['type'] . 'Complete';
$CurrentBlock = $this->$methodName($CurrentBlock);
} }
} }
} }
@ -254,7 +244,7 @@ class Parsedown
foreach ($blockTypes as $blockType) foreach ($blockTypes as $blockType)
{ {
$Block = $this->{'block'.$blockType}($Line, $CurrentBlock); $Block = $this->{"block$blockType"}($Line, $CurrentBlock);
if (isset($Block)) if (isset($Block))
{ {
@ -309,7 +299,8 @@ class Parsedown
if (isset($CurrentBlock['continuable']) and $this->isBlockCompletable($CurrentBlock['type'])) if (isset($CurrentBlock['continuable']) and $this->isBlockCompletable($CurrentBlock['type']))
{ {
$CurrentBlock = $this->{'block'.$CurrentBlock['type'].'Complete'}($CurrentBlock); $methodName = 'block' . $CurrentBlock['type'] . 'Complete';
$CurrentBlock = $this->$methodName($CurrentBlock);
} }
# ~ # ~
@ -393,15 +384,6 @@ class Parsedown
} }
} }
protected function blockCodeComplete($Block)
{
$text = $Block['element']['element']['text'];
$Block['element']['element']['text'] = $text;
return $Block;
}
# #
# Comment # Comment
@ -452,25 +434,35 @@ class Parsedown
protected function blockFencedCode($Line) protected function blockFencedCode($Line)
{ {
if (preg_match('/^(['.$Line['text'][0].']{3,})[ ]*([^`]+)?[ ]*$/', $Line['text'], $matches)) $marker = $Line['text'][0];
$openerLength = strspn($Line['text'], $marker);
if ($openerLength < 3)
{ {
return;
}
$infostring = trim(substr($Line['text'], $openerLength), "\t ");
if (strpos($infostring, '`') !== false)
{
return;
}
$Element = array( $Element = array(
'name' => 'code', 'name' => 'code',
'text' => '', 'text' => '',
); );
if (isset($matches[2])) if ($infostring !== '')
{ {
$class = 'language-'.$matches[2]; $Element['attributes'] = array('class' => "language-$infostring");
$Element['attributes'] = array(
'class' => $class,
);
} }
$Block = array( $Block = array(
'char' => $Line['text'][0], 'char' => $marker,
'openerLength' => mb_strlen($matches[1]), 'openerLength' => $openerLength,
'element' => array( 'element' => array(
'name' => 'pre', 'name' => 'pre',
'element' => $Element, 'element' => $Element,
@ -479,7 +471,6 @@ class Parsedown
return $Block; return $Block;
} }
}
protected function blockFencedCodeContinue($Line, $Block) protected function blockFencedCodeContinue($Line, $Block)
{ {
@ -495,9 +486,8 @@ class Parsedown
unset($Block['interrupted']); unset($Block['interrupted']);
} }
if ( if (($len = strspn($Line['text'], $Block['char'])) >= $Block['openerLength']
preg_match('/^(['.preg_quote($Block['char']).']{3,})[ ]*$/', $Line['text'], $matches) and chop(substr($Line['text'], $len), ' ') === ''
and mb_strlen($matches[1]) >= $Block['openerLength']
) { ) {
$Block['element']['element']['text'] = substr($Block['element']['element']['text'], 1); $Block['element']['element']['text'] = substr($Block['element']['element']['text'], 1);
@ -511,26 +501,12 @@ class Parsedown
return $Block; return $Block;
} }
protected function blockFencedCodeComplete($Block)
{
$text = $Block['element']['element']['text'];
$Block['element']['element']['text'] = $text;
return $Block;
}
# #
# Header # Header
protected function blockHeader($Line) protected function blockHeader($Line)
{ {
$level = 1; $level = strspn($Line['text'], '#');
while (isset($Line['text'][$level]) and $Line['text'][$level] === '#')
{
$level ++;
}
if ($level > 6) if ($level > 6)
{ {
@ -565,9 +541,9 @@ class Parsedown
protected function blockList($Line, array $CurrentBlock = null) protected function blockList($Line, array $CurrentBlock = null)
{ {
list($name, $pattern) = $Line['text'][0] <= '-' ? array('ul', '[*+-]') : array('ol', '[0-9]{1,9}[.\)]'); list($name, $pattern) = $Line['text'][0] <= '-' ? array('ul', '[*+-]') : array('ol', '[0-9]{1,9}+[.\)]');
if (preg_match('/^('.$pattern.'([ ]+|$))(.*)/', $Line['text'], $matches)) if (preg_match('/^('.$pattern.'([ ]++|$))(.*+)/', $Line['text'], $matches))
{ {
$contentIndent = strlen($matches[2]); $contentIndent = strlen($matches[2]);
@ -582,19 +558,22 @@ class Parsedown
$matches[1] .= ' '; $matches[1] .= ' ';
} }
$markerWithoutWhitespace = strstr($matches[1], ' ', true);
$Block = array( $Block = array(
'indent' => $Line['indent'], 'indent' => $Line['indent'],
'pattern' => $pattern, 'pattern' => $pattern,
'data' => array( 'data' => array(
'type' => $name, 'type' => $name,
'marker' => $matches[1], 'marker' => $matches[1],
'markerType' => ($name === 'ul' ? strstr($matches[1], ' ', true) : substr(strstr($matches[1], ' ', true), -1)), 'markerType' => ($name === 'ul' ? $markerWithoutWhitespace : substr($markerWithoutWhitespace, -1)),
), ),
'element' => array( 'element' => array(
'name' => $name, 'name' => $name,
'elements' => array(), 'elements' => array(),
), ),
); );
$Block['data']['markerTypeRegex'] = preg_quote($Block['data']['markerType'], '/');
if ($name === 'ol') if ($name === 'ol')
{ {
@ -642,10 +621,10 @@ class Parsedown
and ( and (
( (
$Block['data']['type'] === 'ol' $Block['data']['type'] === 'ol'
and preg_match('/^[0-9]+'.preg_quote($Block['data']['markerType']).'(?:[ ]+(.*)|$)/', $Line['text'], $matches) and preg_match('/^[0-9]++'.$Block['data']['markerTypeRegex'].'(?:[ ]++(.*)|$)/', $Line['text'], $matches)
) or ( ) or (
$Block['data']['type'] === 'ul' $Block['data']['type'] === 'ul'
and preg_match('/^'.preg_quote($Block['data']['markerType']).'(?:[ ]+(.*)|$)/', $Line['text'], $matches) and preg_match('/^'.$Block['data']['markerTypeRegex'].'(?:[ ]++(.*)|$)/', $Line['text'], $matches)
) )
) )
) { ) {
@ -707,7 +686,7 @@ class Parsedown
if ( ! isset($Block['interrupted'])) if ( ! isset($Block['interrupted']))
{ {
$text = preg_replace('/^[ ]{0,'.$requiredIndent.'}/', '', $Line['body']); $text = preg_replace('/^[ ]{0,'.$requiredIndent.'}+/', '', $Line['body']);
$Block['li']['handler']['argument'] []= $text; $Block['li']['handler']['argument'] []= $text;
@ -736,7 +715,7 @@ class Parsedown
protected function blockQuote($Line) protected function blockQuote($Line)
{ {
if (preg_match('/^>[ ]?(.*)/', $Line['text'], $matches)) if (preg_match('/^>[ ]?+(.*+)/', $Line['text'], $matches))
{ {
$Block = array( $Block = array(
'element' => array( 'element' => array(
@ -760,7 +739,7 @@ class Parsedown
return; return;
} }
if ($Line['text'][0] === '>' and preg_match('/^>[ ]?(.*)/', $Line['text'], $matches)) if ($Line['text'][0] === '>' and preg_match('/^>[ ]?+(.*+)/', $Line['text'], $matches))
{ {
$Block['element']['handler']['argument'] []= $matches[1]; $Block['element']['handler']['argument'] []= $matches[1];
@ -780,7 +759,9 @@ class Parsedown
protected function blockRule($Line) protected function blockRule($Line)
{ {
if (preg_match('/^(['.$Line['text'][0].'])([ ]*\1){2,}[ ]*$/', $Line['text'])) $marker = $Line['text'][0];
if (substr_count($Line['text'], $marker) >= 3 and chop($Line['text'], " $marker") === '')
{ {
$Block = array( $Block = array(
'element' => array( 'element' => array(
@ -802,10 +783,8 @@ class Parsedown
return; return;
} }
if ( if ($Line['indent'] < 4 and chop(chop($Line['text'], ' '), $Line['text'][0]) === '')
chop(chop($Line['text'], ' '), $Line['text'][0]) === '' {
and $Line['indent'] < 4
) {
$Block['element']['name'] = $Line['text'][0] === '=' ? 'h1' : 'h2'; $Block['element']['name'] = $Line['text'][0] === '=' ? 'h1' : 'h2';
return $Block; return $Block;
@ -822,7 +801,7 @@ class Parsedown
return; return;
} }
if (preg_match('/^<[\/]?+(\w*)(?:[ ]*'.$this->regexHtmlAttribute.')*[ ]*(\/)?>/', $Line['text'], $matches)) if (preg_match('/^<[\/]?+(\w*)(?:[ ]*+'.$this->regexHtmlAttribute.')*+[ ]*+(\/)?>/', $Line['text'], $matches))
{ {
$element = strtolower($matches[1]); $element = strtolower($matches[1]);
@ -860,8 +839,9 @@ class Parsedown
protected function blockReference($Line) protected function blockReference($Line)
{ {
if (preg_match('/^\[(.+?)\]:[ ]*<?(\S+?)>?(?:[ ]+["\'(](.+)["\')])?[ ]*$/', $Line['text'], $matches)) if (strpos($Line['text'], ']') !== false
{ and preg_match('/^\[(.+?)\]:[ ]*+<?(\S+?)>?(?:[ ]+["\'(](.+)["\')])?[ ]*+$/', $Line['text'], $matches)
) {
$id = strtolower($matches[1]); $id = strtolower($matches[1]);
$Data = array( $Data = array(
@ -970,7 +950,7 @@ class Parsedown
$alignment = $alignments[$index]; $alignment = $alignments[$index];
$HeaderElement['attributes'] = array( $HeaderElement['attributes'] = array(
'style' => 'text-align: '.$alignment.';', 'style' => "text-align: $alignment;",
); );
} }
@ -1021,7 +1001,7 @@ class Parsedown
$row = trim($row); $row = trim($row);
$row = trim($row, '|'); $row = trim($row, '|');
preg_match_all('/(?:(\\\\[|])|[^|`]|`[^`]+`|`)+/', $row, $matches); preg_match_all('/(?:(\\\\[|])|[^|`]|`[^`]++`|`)++/', $row, $matches);
$cells = array_slice($matches[0], 0, count($Block['alignments'])); $cells = array_slice($matches[0], 0, count($Block['alignments']));
@ -1124,13 +1104,18 @@ class Parsedown
{ {
$Elements = array(); $Elements = array();
$nonNestables = (empty($nonNestables)
? array()
: array_combine($nonNestables, $nonNestables)
);
# $excerpt is based on the first occurrence of a marker # $excerpt is based on the first occurrence of a marker
while ($excerpt = strpbrk($text, $this->inlineMarkerList)) while ($excerpt = strpbrk($text, $this->inlineMarkerList))
{ {
$marker = $excerpt[0]; $marker = $excerpt[0];
$markerPosition = strpos($text, $marker); $markerPosition = strlen($text) - strlen($excerpt);
$Excerpt = array('text' => $excerpt, 'context' => $text); $Excerpt = array('text' => $excerpt, 'context' => $text);
@ -1138,12 +1123,12 @@ class Parsedown
{ {
# check to see if the current inline type is nestable in the current context # check to see if the current inline type is nestable in the current context
if ( ! empty($nonNestables) and in_array($inlineType, $nonNestables)) if (isset($nonNestables[$inlineType]))
{ {
continue; continue;
} }
$Inline = $this->{'inline'.$inlineType}($Excerpt); $Inline = $this->{"inline$inlineType"}($Excerpt);
if ( ! isset($Inline)) if ( ! isset($Inline))
{ {
@ -1166,10 +1151,11 @@ class Parsedown
# cause the new element to 'inherit' our non nestables # cause the new element to 'inherit' our non nestables
foreach ($nonNestables as $non_nestable)
{ $Inline['element']['nonNestables'] = isset($Inline['element']['nonNestables'])
$Inline['element']['nonNestables'][] = $non_nestable; ? array_merge($Inline['element']['nonNestables'], $nonNestables)
} : $nonNestables
;
# the text that comes before the inline # the text that comes before the inline
$unmarkedText = substr($text, 0, $Inline['position']); $unmarkedText = substr($text, 0, $Inline['position']);
@ -1200,14 +1186,13 @@ class Parsedown
$InlineText = $this->inlineText($text); $InlineText = $this->inlineText($text);
$Elements[] = $InlineText['element']; $Elements[] = $InlineText['element'];
$Elements = array_map( foreach ($Elements as &$Element)
function ($Element) { {
$Element['autobreak'] = isset($Element['autobreak']) if ( ! isset($Element['autobreak']))
? $Element['autobreak'] : false; {
return $Element; $Element['autobreak'] = false;
}, }
$Elements }
);
return $Elements; return $Elements;
} }
@ -1220,33 +1205,17 @@ class Parsedown
{ {
$Inline = array( $Inline = array(
'extent' => strlen($text), 'extent' => strlen($text),
'element' => array( 'element' => array(),
'elements' => array(),
),
); );
if ($this->breaksEnabled) $safeText = self::escape($text, true);
{
$Inline['element']['elements'] = self::pregReplaceElements( $Inline['element']['rawHtml'] = preg_replace(
'/[ ]*\n/', $this->breaksEnabled ? '/[ ]*+\n/' : '/(?:[ ]*+\\\\|[ ]{2,}+)\n/',
array( "<br />\n",
array('name' => 'br'), $safeText
array('text' => "\n"),
),
$text
); );
} $Inline['element']['allowRawHtmlInSafeMode'] = true;
else
{
$Inline['element']['elements'] = self::pregReplaceElements(
'/(?:[ ][ ]+|[ ]*\\\\)\n/',
array(
array('name' => 'br'),
array('text' => "\n"),
),
$text
);
}
return $Inline; return $Inline;
} }
@ -1255,10 +1224,10 @@ class Parsedown
{ {
$marker = $Excerpt['text'][0]; $marker = $Excerpt['text'][0];
if (preg_match('/^('.$marker.'+)[ ]*(.+?)[ ]*(?<!'.$marker.')\1(?!'.$marker.')/s', $Excerpt['text'], $matches)) if (preg_match('/^(['.$marker.']++)[ ]*+(.+?)[ ]*+(?<!['.$marker.'])\1(?!'.$marker.')/s', $Excerpt['text'], $matches))
{ {
$text = $matches[2]; $text = $matches[2];
$text = preg_replace("/[ ]*\n/", ' ', $text); $text = preg_replace('/[ ]*+\n/', ' ', $text);
return array( return array(
'extent' => strlen($matches[0]), 'extent' => strlen($matches[0]),
@ -1284,7 +1253,7 @@ class Parsedown
if ( ! isset($matches[2])) if ( ! isset($matches[2]))
{ {
$url = 'mailto:' . $url; $url = "mailto:$url";
} }
return array( return array(
@ -1414,7 +1383,7 @@ class Parsedown
return; return;
} }
if (preg_match('/^[(]\s*+((?:[^ ()]++|[(][^ )]+[)])++)(?:[ ]+("[^"]*"|\'[^\']*\'))?\s*[)]/', $remainder, $matches)) if (preg_match('/^[(]\s*+((?:[^ ()]++|[(][^ )]+[)])++)(?:[ ]+("[^"]*+"|\'[^\']*+\'))?\s*+[)]/', $remainder, $matches))
{ {
$Element['attributes']['href'] = $matches[1]; $Element['attributes']['href'] = $matches[1];
@ -1463,7 +1432,7 @@ class Parsedown
return; return;
} }
if ($Excerpt['text'][1] === '/' and preg_match('/^<\/\w[\w-]*[ ]*>/s', $Excerpt['text'], $matches)) if ($Excerpt['text'][1] === '/' and preg_match('/^<\/\w[\w-]*+[ ]*+>/s', $Excerpt['text'], $matches))
{ {
return array( return array(
'element' => array('rawHtml' => $matches[0]), 'element' => array('rawHtml' => $matches[0]),
@ -1471,7 +1440,7 @@ class Parsedown
); );
} }
if ($Excerpt['text'][1] === '!' and preg_match('/^<!---?[^>-](?:-?[^-])*-->/s', $Excerpt['text'], $matches)) if ($Excerpt['text'][1] === '!' and preg_match('/^<!---?[^>-](?:-?+[^-])*-->/s', $Excerpt['text'], $matches))
{ {
return array( return array(
'element' => array('rawHtml' => $matches[0]), 'element' => array('rawHtml' => $matches[0]),
@ -1479,7 +1448,7 @@ class Parsedown
); );
} }
if ($Excerpt['text'][1] !== ' ' and preg_match('/^<\w[\w-]*(?:[ ]*'.$this->regexHtmlAttribute.')*[ ]*\/?>/s', $Excerpt['text'], $matches)) if ($Excerpt['text'][1] !== ' ' and preg_match('/^<\w[\w-]*+(?:[ ]*+'.$this->regexHtmlAttribute.')*+[ ]*+\/?>/s', $Excerpt['text'], $matches))
{ {
return array( return array(
'element' => array('rawHtml' => $matches[0]), 'element' => array('rawHtml' => $matches[0]),
@ -1490,8 +1459,9 @@ class Parsedown
protected function inlineSpecialCharacter($Excerpt) protected function inlineSpecialCharacter($Excerpt)
{ {
if (preg_match('/^&(#?+[0-9a-zA-Z]++);/', $Excerpt['text'], $matches)) if ($Excerpt['text'][1] !== ' ' and strpos($Excerpt['text'], ';') !== false
{ and preg_match('/^&(#?+[0-9a-zA-Z]++);/', $Excerpt['text'], $matches)
) {
return array( return array(
'element' => array('rawHtml' => '&' . $matches[1] . ';'), 'element' => array('rawHtml' => '&' . $matches[1] . ';'),
'extent' => strlen($matches[0]), 'extent' => strlen($matches[0]),
@ -1531,8 +1501,9 @@ class Parsedown
return; return;
} }
if (preg_match('/\bhttps?:[\/]{2}[^\s<]+\b\/*/ui', $Excerpt['context'], $matches, PREG_OFFSET_CAPTURE)) if (strpos($Excerpt['context'], 'http') !== false
{ and preg_match('/\bhttps?+:[\/]{2}[^\s<]+\b\/*+/ui', $Excerpt['context'], $matches, PREG_OFFSET_CAPTURE)
) {
$url = $matches[0][0]; $url = $matches[0][0];
$Inline = array( $Inline = array(
@ -1553,7 +1524,7 @@ class Parsedown
protected function inlineUrlTag($Excerpt) protected function inlineUrlTag($Excerpt)
{ {
if (strpos($Excerpt['text'], '>') !== false and preg_match('/^<(\w+:\/{2}[^ >]+)>/i', $Excerpt['text'], $matches)) if (strpos($Excerpt['text'], '>') !== false and preg_match('/^<(\w++:\/{2}[^ >]++)>/i', $Excerpt['text'], $matches))
{ {
$url = $matches[1]; $url = $matches[1];
@ -1611,9 +1582,9 @@ class Parsedown
{ {
$Element = $this->handle($Element); $Element = $this->handle($Element);
} }
}
unset($Element['handler']); unset($Element['handler']);
}
return $Element; return $Element;
} }
@ -1707,7 +1678,7 @@ class Parsedown
continue; continue;
} }
$markup .= ' '.$name.'="'.self::escape($value).'"'; $markup .= " $name=\"".self::escape($value).'"';
} }
} }
} }
@ -1777,8 +1748,8 @@ class Parsedown
continue; continue;
} }
$autoBreakNext = (isset($Element['autobreak']) && $Element['autobreak'] $autoBreakNext = (isset($Element['autobreak'])
|| ! isset($Element['autobreak']) && isset($Element['name']) ? $Element['autobreak'] : isset($Element['name'])
); );
// (autobreak === false) covers both sides of an element // (autobreak === false) covers both sides of an element
$autoBreak = !$autoBreak ? $autoBreak : $autoBreakNext; $autoBreak = !$autoBreak ? $autoBreak : $autoBreakNext;
@ -1959,8 +1930,8 @@ class Parsedown
); );
protected $StrongRegex = array( protected $StrongRegex = array(
'*' => '/^[*]{2}((?:\\\\\*|[^*]|[*][^*]*[*])+?)[*]{2}(?![*])/s', '*' => '/^[*]{2}((?:\\\\\*|[^*]|[*][^*]*+[*])+?)[*]{2}(?![*])/s',
'_' => '/^__((?:\\\\_|[^_]|_[^_]*_)+?)__(?!_)/us', '_' => '/^__((?:\\\\_|[^_]|_[^_]*+_)+?)__(?!_)/us',
); );
protected $EmRegex = array( protected $EmRegex = array(
@ -1968,7 +1939,7 @@ class Parsedown
'_' => '/^_((?:\\\\_|[^_]|__[^_]*__)+?)_(?!_)\b/us', '_' => '/^_((?:\\\\_|[^_]|__[^_]*__)+?)_(?!_)\b/us',
); );
protected $regexHtmlAttribute = '[a-zA-Z_:][\w:.-]*(?:\s*=\s*(?:[^"\'=<>`\s]+|"[^"]*"|\'[^\']*\'))?'; protected $regexHtmlAttribute = '[a-zA-Z_:][\w:.-]*+(?:\s*+=\s*+(?:[^"\'=<>`\s]+|"[^"]*+"|\'[^\']*+\'))?+';
protected $voidElements = array( protected $voidElements = array(
'area', 'base', 'br', 'col', 'command', 'embed', 'hr', 'img', 'input', 'link', 'meta', 'param', 'source', 'area', 'base', 'br', 'col', 'command', 'embed', 'hr', 'img', 'input', 'link', 'meta', 'param', 'source',