Thijs/wiki.techinc.nl
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Browse source
This commit is contained in:
Leo Koppelkamm 2011-04-12 21:27:24 +00:00
parent d210706028
commit 109f7bcb9d
4 changed files with 362 additions and 261 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
includes/Sanitizer.php
View file

@ -369,7 +369,7 @@ class Sanitizer {
'strike', 'strong', 'tt', 'var', 'div', 'center',
'blockquote', 'ol', 'ul', 'dl', 'table', 'caption', 'pre',
'ruby', 'rt' , 'rb' , 'rp', 'p', 'span', 'abbr', 'dfn',
'kbd', 'samp'
'kbd', 'samp', 'thead', 'tbody', 'tfoot'
);
$htmlsingle = array(
'br', 'hr', 'li', 'dt', 'dd'

382
includes/parser/Parser.php
View file

@ -824,191 +824,285 @@ class Parser {
$lines = StringUtils::explode( "\n", $text );
$out = '';
$td_history = array(); # Is currently a td tag open?
$last_tag_history = array(); # Save history of last lag activated (td, th or caption)
$tr_history = array(); # Is currently a tr tag open?
$tr_attributes = array(); # history of tr attributes
$has_opened_tr = array(); # Did this table open a <tr> element?
$indent_level = 0; # indent level of the table
$output =& $out;
foreach ( $lines as $outLine ) {
$line = trim( $outLine );
if ( $line === '' ) { # empty line, go to next line
if ( $line == '') { //empty line, go to next line
$out .= $outLine."\n";
continue;
}
$first_character = $line[0];
$first_chars = $line[0];
if ( strlen($line) > 1) {
$first_chars .= in_array($line[1], array('}', '+', '-')) ? $line[1] : '';
}
$matches = array();
if ( preg_match( '/^(:*)\{\|(.*)$/', $line , $matches ) ) {
# First check if we are starting a new table
$indent_level = strlen( $matches[1] );
$tables[] = array();
$table =& $this->last($tables);
$table[0] = array(); //first row
$current_row =& $table[0];
$table['indent'] = strlen( $matches[1] );
$attributes = $this->mStripState->unstripBoth( $matches[2] );
$attributes = Sanitizer::fixTagAttributes( $attributes , 'table' );
$outLine = str_repeat( '<dl><dd>' , $indent_level ) . "<table{$attributes}>";
array_push( $td_history , false );
array_push( $last_tag_history , '' );
array_push( $tr_history , false );
array_push( $tr_attributes , '' );
array_push( $has_opened_tr , false );
} elseif ( count( $td_history ) == 0 ) {
# Don't do any of the following
if ( $attributes !== '' ) {
$table['attributes'] = $attributes;
}
} else if ( !isset($tables[0]) ) {
// we're outside the table
$out .= $outLine."\n";
continue;
} elseif ( substr( $line , 0 , 2 ) === '|}' ) {
# We are ending a table
$line = '</table>' . substr( $line , 2 );
$last_tag = array_pop( $last_tag_history );
} else if ( $first_chars === '|}' ) {
// trim the |} code from the line
$line = substr ( $line , 2 );
if ( !array_pop( $has_opened_tr ) ) {
$line = "<tr><td></td></tr>{$line}";
// Shorthand for last row
$last_row =& $this->last($table);
// a thead at the end becomes a tfoot, unless there is only one row
// Do this before deleting empty last lines to allow headers at the bottom of tables
if ( isset($last_row['type'] ) && $last_row['type'] == 'thead' && isset($table[1])) {
$last_row['type'] = 'tfoot';
for($i = 0; isset($last_row[$i]); $i++ ) {
$last_row[$i]['type'] = 'td';
}
}
if ( array_pop( $tr_history ) ) {
$line = "</tr>{$line}";
// Delete empty last lines
if ( empty($last_row) ) {
$last_row = NULL;
}
$o = $this->printTableHtml( array_pop($tables) ) . $line;
if ( count($tables) > 0 ) {
$table =& $this->last($tables);
$current_row =& $this->last($table);
$current_element =& $this->last($current_row);
$output =& $current_element['content'];
} else {
$output =& $out;
}
if ( array_pop( $td_history ) ) {
$line = "</{$last_tag}>{$line}";
}
array_pop( $tr_attributes );
$outLine = $line . str_repeat( '</dd></dl>' , $indent_level );
} elseif ( substr( $line , 0 , 2 ) === '|-' ) {
# Now we have a table row
$line = preg_replace( '#^\|-+#', '', $line );
$output .= $o;
# Whats after the tag is now only attributes
} else if ( $first_chars === '|-' ) {
// start a new row element
// but only when we haven't started one already
if( count($current_row) != 0 ) {
$table[] = array();
$current_row =& $this->last($table);
}
// Get the attributes, there's nothing else useful in $line now
$line = substr ( $line , 2 );
$attributes = $this->mStripState->unstripBoth( $line );
$attributes = Sanitizer::fixTagAttributes( $attributes, 'tr' );
array_pop( $tr_attributes );
array_push( $tr_attributes, $attributes );
if( $attributes !== '') {
$current_row['attributes'] = $attributes;
}
} else if ( $first_chars === '|+' ) {
// a table caption
$line = substr ( $line , 2 );
$c = $this->getCellAttr($line , 'caption');
$table['caption'] = array();
$table['caption']['content'] = $c[0];
if(isset($c[1])) $table['caption']['attributes'] = $c[1];
unset($c);
$line = '';
$last_tag = array_pop( $last_tag_history );
array_pop( $has_opened_tr );
array_push( $has_opened_tr , true );
$output =& $table['caption'];
} else if ( $first_chars === '|' || $first_chars === '!' || $first_chars === '!+' ) {
// Which kind of cells are we dealing with
$this_tag = 'td';
$line = substr ( $line , 1 );
if ( array_pop( $tr_history ) ) {
$line = '</tr>';
if ( $first_chars === '!' || $first_chars === '!+' ) {
$line = str_replace ( '!!' , '||' , $line );
$this_tag = 'th';
}
if ( array_pop( $td_history ) ) {
$line = "</{$last_tag}>{$line}";
}
$outLine = $line;
array_push( $tr_history , false );
array_push( $td_history , false );
array_push( $last_tag_history , '' );
} elseif ( $first_character === '|' || $first_character === '!' || substr( $line , 0 , 2 ) === '|+' ) {
# This might be cell elements, td, th or captions
if ( substr( $line , 0 , 2 ) === '|+' ) {
$first_character = '+';
$line = substr( $line , 1 );
}
$line = substr( $line , 1 );
if ( $first_character === '!' ) {
$line = str_replace( '!!' , '||' , $line );
}
# Split up multiple cells on the same line.
# FIXME : This can result in improper nesting of tags processed
# by earlier parser steps, but should avoid splitting up eg
# attribute values containing literal "||".
// Split up multiple cells on the same line.
$cells = StringUtils::explodeMarkup( '||' , $line );
$line = ''; // save memory
$outLine = '';
# Loop through each table cell
foreach ( $cells as $cell ) {
$previous = '';
if ( $first_character !== '+' ) {
$tr_after = array_pop( $tr_attributes );
if ( !array_pop( $tr_history ) ) {
$previous = "<tr{$tr_after}>\n";
}
array_push( $tr_history , true );
array_push( $tr_attributes , '' );
array_pop( $has_opened_tr );
array_push( $has_opened_tr , true );
}
$last_tag = array_pop( $last_tag_history );
if ( array_pop( $td_history ) ) {
$previous = "</{$last_tag}>\n{$previous}";
}
if ( $first_character === '|' ) {
$last_tag = 'td';
} elseif ( $first_character === '!' ) {
$last_tag = 'th';
} elseif ( $first_character === '+' ) {
$last_tag = 'caption';
} else {
$last_tag = '';
}
array_push( $last_tag_history , $last_tag );
# A cell could contain both parameters and data
$cell_data = explode( '|' , $cell , 2 );
# Bug 553: Note that a '|' inside an invalid link should not
# be mistaken as delimiting cell parameters
if ( strpos( $cell_data[0], '[[' ) !== false ) {
$cell = "{$previous}<{$last_tag}>{$cell}";
} elseif ( count( $cell_data ) == 1 ) {
$cell = "{$previous}<{$last_tag}>{$cell_data[0]}";
} else {
$attributes = $this->mStripState->unstripBoth( $cell_data[0] );
$attributes = Sanitizer::fixTagAttributes( $attributes , $last_tag );
$cell = "{$previous}<{$last_tag}{$attributes}>{$cell_data[1]}";
}
$outLine .= $cell;
array_push( $td_history , true );
// decide whether thead to tbody
if ( !array_key_exists('type', $current_row) ) {
$current_row['type'] = ( $first_chars === '!' ) ? 'thead' : 'tbody' ;
} else if( $first_chars === '|' ) {
$current_row['type'] = 'tbody';
}
// Loop through each table cell
foreach ( $cells as $cell ) {
// a new cell
$current_row[] = array();
$current_element =& $this->last($current_row);
$current_element['type'] = $this_tag;
$c = $this->getCellAttr($cell , $this_tag);
$current_element['content'] = $c[0];
if(isset($c[1])) $current_element['attributes'] = $c[1];
unset($c);
}
$output =& $current_element['content'];
} else {
$output .= $outLine."\n";
}
$out .= $outLine . "\n";
}
# Closing open td, tr && table
while ( count( $td_history ) > 0 ) {
if ( array_pop( $td_history ) ) {
$out .= "</td>\n";
}
if ( array_pop( $tr_history ) ) {
$out .= "</tr>\n";
}
if ( !array_pop( $has_opened_tr ) ) {
$out .= "<tr><td></td></tr>\n" ;
}
$out .= "</table>\n";
}
# Remove trailing line-ending (b/c)
if ( substr( $out, -1 ) === "\n" ) {
$out = substr( $out, 0, -1 );
}
# special case: don't return empty table
if ( $out === "<table>\n<tr><td></td></tr>\n</table>" ) {
$out = '';
#Close any unclosed tables
if (isset($tables) && count($tables) > 0 ) {
for ($i = 0; $i < count($tables); $i++) {
$out .= $this->printTableHtml( array_pop($tables) );
}
}
wfProfileOut( __METHOD__ );
return $out;
}
/**
* Helper function for doTableStuff() separating the contents of cells from
* attributes. Particularly useful as there's a possible bug and this action
* is repeated twice.
*
* @private
*/
function getCellAttr ($cell , $tag_name) {
$content = null;
$attributes = null;
$cell = trim ( $cell );
// A cell could contain both parameters and data
$cell_data = explode ( '|' , $cell , 2 );
// Bug 553: Note that a '|' inside an invalid link should not
// be mistaken as delimiting cell parameters
if ( strpos( $cell_data[0], '[[' ) !== false ) {
$content = trim ( $cell );
}
else if ( count ( $cell_data ) == 1 ) {
$content = trim ( $cell_data[0] );
}
else {
$attributes = $this->mStripState->unstripBoth( $cell_data[0] );
$attributes = Sanitizer::fixTagAttributes( $attributes , $tag_name );
$content = trim ( $cell_data[1] );
}
return array($content, $attributes);
}
/**
* Helper function for doTableStuff(). This converts the structured array into html.
*
* @private
*/
function printTableHtml (&$t) {
$r = "\n";
$r .= str_repeat( '<dl><dd>' , $t['indent'] );
$r .= '<table';
$r .= isset($t['attributes']) ? $t['attributes'] : '';
$r .= '>';
unset($t['attributes']);
if ( isset($t['caption']) ) {
$r .= "\n<caption";
$r .= isset($t['caption']['attributes']) ? $t['caption']['attributes'] : '';
$r .= '>';
$r .= $t['caption']['content'];
$r .= '</caption>';
}
$last_section = '';
$empty = true;
$simple = true;
//If we only have tbodies, mark table as simple
for($i = 0; isset($t[$i]); $i++ ) {
if ( !count( $t[$i]) ) continue;
if ( !$last_section ) {
$last_section = $t[$i]['type'];
} else if ($last_section != $t[$i]['type']) {
$simple = false;
break;
}
}
$last_section = '';
for($i = 0; isset($t[$i]); $i++ ) {
// Check for empty tables
if ( count( $t[$i]) ) {
$empty = false;
} else {
continue;
}
if( $t[$i]['type'] != $last_section && !$simple ) {
$r .= "\n<" . $t[$i]['type'] . '>';
}
$r .= "\n<tr";
$r .= isset($t[$i]['attributes']) ? $t[$i]['attributes'] : '';
$r .= '>';
for($j = 0; isset($t[$i][$j]); $j++ ) {
$r .= "\n<" . $t[$i][$j]['type'];
$r .= isset($t[$i][$j]['attributes']) ? $t[$i][$j]['attributes'] : '';
$r .= '>';
$r .= $t[$i][$j]['content'];
$r .= '</' . $t[$i][$j]['type'] . '>';
unset($t[$i][$j]);
}
$r .= "\n</tr>";
if( ( !isset($t[$i+1]) && !$simple )|| ( isset($t[$i+1]) && ($t[$i]['type'] != $t[$i+1]['type'])) ) {
$r .= '</' . $t[$i]['type'] . '>';
}
$last_section = $t[$i]['type'];
unset($t[$i]);
}
if ( $empty ) {
if ( isset($t['caption']) ) {
$r .= "\n<tr><td></td></tr>";
} else {
return '';
}
}
$r .= "\n</table>";
$r .= str_repeat( '</dd></dl>' , $t['indent'] );
return $r;
}
/**
* like end() but only works on the numeric array index and php's internal pointers
* returns a reference to the last element of an array much like "\$arr[-1]" in perl
* ignores associative elements and will create a 0 key will a NULL value if there were
* no numric elements and an array itself if not previously defined.
*
* @private
*/
function &last (&$arr) {
for($i = count($arr); (!isset($arr[$i]) && $i > 0); $i--) { }
return $arr[$i];
}
/**
* Helper function for parse() that transforms wiki markup into
* HTML. Only called for $mOutputType == self::OT_HTML.
@ -2239,7 +2333,7 @@ class Parser {
'<td|<th|<\\/?div|<hr|<\\/pre|<\\/p|'.$this->mUniqPrefix.'-pre|<\\/li|<\\/ul|<\\/ol|<\\/?center)/iS', $t );
if ( $openmatch or $closematch ) {
$paragraphStack = false;
# TODO bug 5718: paragraph closed
# TODO bug 5718: paragraph closed
$output .= $this->closeParagraph();
if ( $preOpenMatch and !$preCloseMatch ) {
$this->mInPre = true;

9
skins/common/wikibits.js
View file

@ -551,14 +551,7 @@ window.sortables_init = function() {
};
window.ts_makeSortable = function( table ) {
var firstRow;
if ( table.rows && table.rows.length > 0 ) {
if ( table.tHead && table.tHead.rows.length > 0 ) {
firstRow = table.tHead.rows[table.tHead.rows.length-1];
} else {
firstRow = table.rows[0];
}
}
var firstRow = table.rows[0];
if ( !firstRow ) {
return;
}

230
tests/parser/parserTests.txt
View file

@ -1237,8 +1237,9 @@ A table with nothing but a caption
|}
!! result
<table>
<caption> caption
</caption><tr><td></td></tr></table>
<caption>caption</caption>
<tr><td></td></tr>
</table>
!! end
@ -1253,13 +1254,14 @@ Simple table
!! result
<table>
<tr>
<td> 1 </td>
<td> 2
</td></tr>
<td>1</td>
<td>2</td>
</tr>
<tr>
<td> 3 </td>
<td> 4
</td></tr></table>
<td>3</td>
<td>4</td>
</tr>
</table>
!! end
@ -1288,49 +1290,46 @@ Multiplication table
|}
!! result
<table border="1" cellpadding="2">
<caption>Multiplication table
</caption>
<caption>Multiplication table</caption>
<thead>
<tr>
<th> &#215; </th>
<th> 1 </th>
<th> 2 </th>
<th> 3
</th></tr>
<th>&#215;</th>
<th>1</th>
<th>2</th>
<th>3</th>
</tr></thead>
<tbody>
<tr>
<th> 1
</th>
<td> 1 </td>
<td> 2 </td>
<td> 3
</td></tr>
<th>1</th>
<td>1</td>
<td>2</td>
<td>3</td>
</tr>
<tr>
<th> 2
</th>
<td> 2 </td>
<td> 4 </td>
<td> 6
</td></tr>
<th>2</th>
<td>2</td>
<td>4</td>
<td>6</td>
</tr>
<tr>
<th> 3
</th>
<td> 3 </td>
<td> 6 </td>
<td> 9
</td></tr>
<th>3</th>
<td>3</td>
<td>6</td>
<td>9</td>
</tr>
<tr>
<th> 4
</th>
<td> 4 </td>
<td> 8 </td>
<td> 12
</td></tr>
<th>4</th>
<td>4</td>
<td>8</td>
<td>12</td>
</tr>
<tr>
<th> 5
</th>
<td> 5 </td>
<td> 10 </td>
<td> 15
</td></tr></table>
<th>5</th>
<td>5</td>
<td>10</td>
<td>15</td>
</tr></tbody>
</table>
!! end
@ -1348,17 +1347,15 @@ Table rowspan
!! result
<table align="right" border="1">
<tr>
<td> Cell 1, row 1
</td>
<td rowspan="2"> Cell 2, row 1 (and 2)
</td>
<td> Cell 3, row 1
</td></tr>
<td>Cell 1, row 1</td>
<td rowspan="2">Cell 2, row 1 (and 2)</td>
<td>Cell 3, row 1</td>
</tr>
<tr>
<td> Cell 1, row 2
</td>
<td> Cell 3, row 2
</td></tr></table>
<td>Cell 1, row 2</td>
<td>Cell 3, row 2</td>
</tr>
</table>
!! end
@ -1378,19 +1375,19 @@ Nested table
!! result
<table border="1">
<tr>
<td> &#945;
</td>
<td>&#945;</td>
<td>
<table bgcolor="#ABCDEF" border="2">
<tr>
<td>nested
</td></tr>
<td>nested</td>
</tr>
<tr>
<td>table
</td></tr></table>
</td>
<td>the original table again
</td></tr></table>
<td>table</td>
</tr>
</table></td>
<td>the original table again</td>
</tr>
</table>
!! end
@ -1403,8 +1400,9 @@ Invalid attributes in table cell (bug 1830)
!! result
<table>
<tr>
<td>broken
</td></tr></table>
<td>broken</td>
</tr>
</table>
!! end
@ -1418,8 +1416,7 @@ Table security: embedded pipes (http://lists.wikimedia.org/mailman/htdig/wikitec
<table>
<tr>
<td>[<a rel="nofollow" class="external free" href="ftp://%7Cx">ftp://%7Cx</a></td>
<td>]" onmouseover="alert(document.cookie)"&gt;test
</td>
<td>]" onmouseover="alert(document.cookie)"&gt;test</td>
</tr>
</table>
@ -2685,8 +2682,9 @@ BUG 553: link with two variables in a piped link
!! result
<table>
<tr>
<td>[[{{{1}}}|{{{2}}}]]
</td></tr></table>
<td>[[{{{1}}}|{{{2}}}]]</td>
</tr>
</table>
!! end
@ -2795,13 +2793,14 @@ foo {{table}}
</p>
<table>
<tr>
<td> 1 </td>
<td> 2
</td></tr>
<td>1</td>
<td>2</td>
</tr>
<tr>
<td> 3 </td>
<td> 4
</td></tr></table>
<td>3</td>
<td>4</td>
</tr>
</table>
!! end
@ -2815,13 +2814,14 @@ foo
</p>
<table>
<tr>
<td> 1 </td>
<td> 2
</td></tr>
<td>1</td>
<td>2</td>
</tr>
<tr>
<td> 3 </td>
<td> 4
</td></tr></table>
<td>3</td>
<td>4</td>
</tr>
</table>
!! end
@ -4370,8 +4370,9 @@ Table multiple attributes correction
!! result
<table>
<tr>
<th class="awesome"> status
</th></tr></table>
<th class="awesome">status</th>
</tr>
</table>
!!end
@ -4815,8 +4816,9 @@ Table attribute legitimate extension
!! result
<table>
<tr>
<th style="color:blue"> status
</th></tr></table>
<th style="color:blue">status</th>
</tr>
</table>
!!end
@ -4829,8 +4831,9 @@ Table attribute safety
!! result
<table>
<tr>
<th style="/* insecure input */"> status
</th></tr></table>
<th style="/* insecure input */">status</th>
</tr>
</table>
!! end
@ -5452,8 +5455,7 @@ Fuzz testing: Parser13
!! result
<table>
<tr>
<td>
</td>
<td></td>
</tr>
</table>
@ -5501,8 +5503,7 @@ noxml
<th>https://</th>
<th></th>
<th></th>
<th>
</td>
<th></th>
</tr>
</table>
@ -5517,10 +5518,8 @@ Fuzz testing: Parser21
!! result
<table>
<tr>
<th> <a rel="nofollow" class="external free" href="irc://{{ftp://a">irc://{{ftp://a</a>" onmouseover="alert('hello world');"
</th>
<td>
</td>
<th><a rel="nofollow" class="external free" href="irc://{{ftp://a">irc://{{ftp://a</a>" onmouseover="alert('hello world');"</th>
<td></td>
</tr>
</table>
@ -5542,6 +5541,22 @@ http://===r:::https://b
!! end
# Known to produce bad XML for now
# Note: the current result listed for this is not what the original one was,
# but the original bug was JavaScript injection, which is fixed in any case.
# It's not clear that the original result listed was any more correct than the
# current one. Original result:
# <table>
# {{{|
# <u class="&#124;">}}}} &gt;
# <br style="onmouseover=&#39;alert(document.cookie);&#39;" />
#
# MOVE YOUR MOUSE CURSOR OVER THIS TEXT
# <tr>
# <td></u>
# </td>
# </tr>
# </table>
!! test
Fuzz testing: Parser24
!! options
@ -5556,15 +5571,14 @@ noxml
MOVE YOUR MOUSE CURSOR OVER THIS TEXT
|
!! result
<table>
{{{|
<p>{{{|
<u class="&#124;">}}}} &gt;
<br style="onmouseover=&#39;alert(document.cookie);&#39;" />
MOVE YOUR MOUSE CURSOR OVER THIS TEXT
</p><p>MOVE YOUR MOUSE CURSOR OVER THIS TEXT
</p>
<table>
<tr>
<td></u>
</td>
<td></u></td>
</tr>
</table>
@ -7735,13 +7749,13 @@ y
</p>
<table>
<tr>
<td> 1 </td>
<td> 2
</td></tr>
<td>1</td>
<td>2</td>
</tr>
<tr>
<td> 3 </td>
<td> 4
</td></tr></table>
<td>3</td>
<td>4</td>
</tr></table>
<p>y
</p>
!! end