mirror/php-htmlpurifier
1
0
Fork 0
mirror of https://github.com/ezyang/htmlpurifier.git synced 2025-08-03 12:47:56 +02:00
Code Issues Releases Wiki Activity

Release 2.0.0, merged in 1026 to HEAD.

Browse Source 
git-svn-id: http://htmlpurifier.org/svnroot/htmlpurifier/branches/strict@1179 48356398-32a2-884e-a903-53898d9a118a
This commit is contained in:
Edward Z. Yang
2007-06-21 00:36:12 +00:00
parent c35eb3e95f
commit 0101311193
172 changed files with 7713 additions and 2520 deletions
Download Patch File Download Diff File Expand all files Collapse all files

28
library/HTMLPurifier/HTMLModule/Bdo.php
View File

@@ -11,30 +11,22 @@ class HTMLPurifier_HTMLModule_Bdo extends HTMLPurifier_HTMLModule
{
var $name = 'Bdo';
var $elements = array('bdo');
var $content_sets = array('Inline' => 'bdo');
var $attr_collections = array(
'I18N' => array('dir' => false)
);
function HTMLPurifier_HTMLModule_Bdo() {
$dir = new HTMLPurifier_AttrDef_Enum(array('ltr','rtl'), false);
$this->attr_collections['I18N']['dir'] = $dir;
$this->info['bdo'] = new HTMLPurifier_ElementDef();
$this->info['bdo']->attr = array(
0 => array('Core', 'Lang'),
'dir' => $dir, // required
// The Abstract Module specification has the attribute
// inclusions wrong for bdo: bdo allows
// xml:lang too (and we'll toss in lang for good measure,
// though it is not allowed for XHTML 1.1, this will
// be managed with a global attribute transform)
$bdo =& $this->addElement(
'bdo', true, 'Inline', 'Inline', array('Core', 'Lang'),
array(
'dir' => 'Enum#ltr,rtl', // required
// The Abstract Module specification has the attribute
// inclusions wrong for bdo: bdo allows Lang
)
);
$this->info['bdo']->content_model = '#PCDATA | Inline';
$this->info['bdo']->content_model_type = 'optional';
// provides fallback behavior if dir's missing (dir is required)
$this->info['bdo']->attr_transform_post['required-dir'] =
new HTMLPurifier_AttrTransform_BdoDir();
$bdo->attr_transform_post['required-dir'] = new HTMLPurifier_AttrTransform_BdoDir();
$this->attr_collections['I18N']['dir'] = 'Enum#ltr,rtl';
}
}

10
library/HTMLPurifier/HTMLModule/CommonAttributes.php
View File

@@ -1,5 +1,7 @@
<?php
require_once 'HTMLPurifier/HTMLModule.php';
class HTMLPurifier_HTMLModule_CommonAttributes extends HTMLPurifier_HTMLModule
{
var $name = 'CommonAttributes';
@@ -12,9 +14,7 @@ class HTMLPurifier_HTMLModule_CommonAttributes extends HTMLPurifier_HTMLModule
'id' => 'ID',
'title' => 'CDATA',
),
'Lang' => array(
'xml:lang' => false, // see constructor
),
'Lang' => array(),
'I18N' => array(
0 => array('Lang'), // proprietary, for xml:lang/lang
),
@@ -22,10 +22,6 @@ class HTMLPurifier_HTMLModule_CommonAttributes extends HTMLPurifier_HTMLModule
0 => array('Core', 'I18N')
)
);
function HTMLPurifier_HTMLModule_CommonAttributes() {
$this->attr_collections['Lang']['xml:lang'] = new HTMLPurifier_AttrDef_Lang();
}
}
?>

32
library/HTMLPurifier/HTMLModule/Edit.php
View File

@@ -11,28 +11,24 @@ class HTMLPurifier_HTMLModule_Edit extends HTMLPurifier_HTMLModule
{
var $name = 'Edit';
var $elements = array('del', 'ins');
var $content_sets = array('Inline' => 'del | ins');
function HTMLPurifier_HTMLModule_Edit() {
foreach ($this->elements as $element) {
$this->info[$element] = new HTMLPurifier_ElementDef();
$this->info[$element]->attr = array(
0 => array('Common'),
'cite' => 'URI',
// 'datetime' => 'Datetime' // Datetime not implemented
);
// Inline context ! Block context (exclamation mark is
// separator, see getChildDef for parsing)
$this->info[$element]->content_model =
'#PCDATA | Inline ! #PCDATA | Flow';
// HTML 4.01 specifies that ins/del must not contain block
// elements when used in an inline context, chameleon is
// a complicated workaround to acheive this effect
$this->info[$element]->content_model_type = 'chameleon';
}
$contents = 'Chameleon: #PCDATA | Inline ! #PCDATA | Flow';
$attr = array(
'cite' => 'URI',
// 'datetime' => 'Datetime', // not implemented
);
$this->addElement('del', true, 'Inline', $contents, 'Common', $attr);
$this->addElement('ins', true, 'Inline', $contents, 'Common', $attr);
}
// HTML 4.01 specifies that ins/del must not contain block
// elements when used in an inline context, chameleon is
// a complicated workaround to acheive this effect
// Inline context ! Block context (exclamation mark is
// separator, see getChildDef for parsing)
var $defines_child_def = true;
function getChildDef($def) {
if ($def->content_model_type != 'chameleon') return false;

29
library/HTMLPurifier/HTMLModule/Hypertext.php
View File

@@ -10,25 +10,22 @@ class HTMLPurifier_HTMLModule_Hypertext extends HTMLPurifier_HTMLModule
{
var $name = 'Hypertext';
var $elements = array('a');
var $content_sets = array('Inline' => 'a');
function HTMLPurifier_HTMLModule_Hypertext() {
$this->info['a'] = new HTMLPurifier_ElementDef();
$this->info['a']->attr = array(
0 => array('Common'),
// 'accesskey' => 'Character',
// 'charset' => 'Charset',
'href' => 'URI',
//'hreflang' => 'LanguageCode',
'rel' => new HTMLPurifier_AttrDef_HTML_LinkTypes('rel'),
'rev' => new HTMLPurifier_AttrDef_HTML_LinkTypes('rev'),
//'tabindex' => 'Number',
//'type' => 'ContentType',
$a =& $this->addElement(
'a', true, 'Inline', 'Inline', 'Common',
array(
// 'accesskey' => 'Character',
// 'charset' => 'Charset',
'href' => 'URI',
// 'hreflang' => 'LanguageCode',
'rel' => new HTMLPurifier_AttrDef_HTML_LinkTypes('rel'),
'rev' => new HTMLPurifier_AttrDef_HTML_LinkTypes('rev'),
// 'tabindex' => 'Number',
// 'type' => 'ContentType',
)
);
$this->info['a']->content_model = '#PCDATA | Inline';
$this->info['a']->content_model_type = 'optional';
$this->info['a']->excludes = array('a' => true);
$a->excludes = array('a' => true);
}
}

24
library/HTMLPurifier/HTMLModule/Image.php
View File

@@ -14,21 +14,21 @@ class HTMLPurifier_HTMLModule_Image extends HTMLPurifier_HTMLModule
{
var $name = 'Image';
var $elements = array('img');
var $content_sets = array('Inline' => 'img');
function HTMLPurifier_HTMLModule_Image() {
$this->info['img'] = new HTMLPurifier_ElementDef();
$this->info['img']->attr = array(
0 => array('Common'),
'alt' => 'Text',
'height' => 'Length',
'longdesc' => 'URI',
'src' => new HTMLPurifier_AttrDef_URI(true), // embedded
'width' => 'Length'
$img =& $this->addElement(
'img', true, 'Inline', 'Empty', 'Common',
array(
'alt*' => 'Text',
'height' => 'Length',
'longdesc' => 'URI',
'src*' => new HTMLPurifier_AttrDef_URI(true), // embedded
'width' => 'Length'
)
);
$this->info['img']->content_model_type = 'empty';
$this->info['img']->attr_transform_post[] =
// kind of strange, but splitting things up would be inefficient
$img->attr_transform_pre[] =
$img->attr_transform_post[] =
new HTMLPurifier_AttrTransform_ImgRequired();
}

125
library/HTMLPurifier/HTMLModule/Legacy.php
View File

@@ -1,5 +1,7 @@
<?php
require_once 'HTMLPurifier/AttrDef/HTML/Bool.php';
/**
* XHTML 1.1 Legacy module defines elements that were previously
* deprecated.
@@ -22,36 +24,115 @@ class HTMLPurifier_HTMLModule_Legacy extends HTMLPurifier_HTMLModule
// incomplete
var $name = 'Legacy';
var $elements = array('u', 's', 'strike');
var $non_standalone_elements = array('li', 'ol', 'address', 'blockquote');
function HTMLPurifier_HTMLModule_Legacy() {
// setup new elements
foreach ($this->elements as $name) {
$this->info[$name] = new HTMLPurifier_ElementDef();
// for u, s, strike, as more elements get added, add
// conditionals as necessary
$this->info[$name]->content_model = 'Inline | #PCDATA';
$this->info[$name]->content_model_type = 'optional';
$this->info[$name]->attr[0] = array('Common');
}
$this->addElement('basefont', true, 'Inline', 'Empty', false, array(
'color' => 'Color',
'face' => 'Text', // extremely broad, we should
'size' => 'Text', // tighten it
'id' => 'ID'
));
$this->addElement('center', true, 'Block', 'Flow', 'Common');
$this->addElement('dir', true, 'Block', 'Required: li', 'Common', array(
'compact' => 'Bool#compact'
));
$this->addElement('font', true, 'Inline', 'Inline', array('Core', 'I18N'), array(
'color' => 'Color',
'face' => 'Text', // extremely broad, we should
'size' => 'Text', // tighten it
));
$this->addElement('menu', true, 'Block', 'Required: li', 'Common', array(
'compact' => 'Bool#compact'
));
$this->addElement('s', true, 'Inline', 'Inline', 'Common');
$this->addElement('strike', true, 'Inline', 'Inline', 'Common');
$this->addElement('u', true, 'Inline', 'Inline', 'Common');
// setup modifications to old elements
foreach ($this->non_standalone_elements as $name) {
$this->info[$name] = new HTMLPurifier_ElementDef();
$this->info[$name]->standalone = false;
$align = 'Enum#left,right,center,justify';
$address =& $this->addBlankElement('address');
$address->content_model = 'Inline | #PCDATA | p';
$address->content_model_type = 'optional';
$address->child = false;
$blockquote =& $this->addBlankElement('blockquote');
$blockquote->content_model = 'Flow | #PCDATA';
$blockquote->content_model_type = 'optional';
$blockquote->child = false;
$br =& $this->addBlankElement('br');
$br->attr['clear'] = 'Enum#left,all,right,none';
$caption =& $this->addBlankElement('caption');
$caption->attr['align'] = 'Enum#top,bottom,left,right';
$div =& $this->addBlankElement('div');
$div->attr['align'] = $align;
$dl =& $this->addBlankElement('dl');
$dl->attr['compact'] = 'Bool#compact';
for ($i = 1; $i <= 6; $i++) {
$h =& $this->addBlankElement("h$i");
$h->attr['align'] = $align;
}
$this->info['li']->attr['value'] = new HTMLPurifier_AttrDef_Integer();
$this->info['ol']->attr['start'] = new HTMLPurifier_AttrDef_Integer();
$hr =& $this->addBlankElement('hr');
$hr->attr['align'] = $align;
$hr->attr['noshade'] = 'Bool#noshade';
$hr->attr['size'] = 'Pixels';
$hr->attr['width'] = 'Length';
$this->info['address']->content_model = 'Inline | #PCDATA | p';
$this->info['address']->content_model_type = 'optional';
$this->info['address']->child = false;
$img =& $this->addBlankElement('img');
$img->attr['align'] = 'Enum#top,middle,bottom,left,right';
$img->attr['border'] = 'Pixels';
$img->attr['hspace'] = 'Pixels';
$img->attr['vspace'] = 'Pixels';
$this->info['blockquote']->content_model = 'Flow | #PCDATA';
$this->info['blockquote']->content_model_type = 'optional';
$this->info['blockquote']->child = false;
// figure out this integer business
$li =& $this->addBlankElement('li');
$li->attr['value'] = new HTMLPurifier_AttrDef_Integer();
$li->attr['type'] = 'Enum#s:1,i,I,a,A,disc,square,circle';
$ol =& $this->addBlankElement('ol');
$ol->attr['compact'] = 'Bool#compact';
$ol->attr['start'] = new HTMLPurifier_AttrDef_Integer();
$ol->attr['type'] = 'Enum#s:1,i,I,a,A';
$p =& $this->addBlankElement('p');
$p->attr['align'] = $align;
$pre =& $this->addBlankElement('pre');
$pre->attr['width'] = 'Number';
// script omitted
$table =& $this->addBlankElement('table');
$table->attr['align'] = 'Enum#left,center,right';
$table->attr['bgcolor'] = 'Color';
$tr =& $this->addBlankElement('tr');
$tr->attr['bgcolor'] = 'Color';
$th =& $this->addBlankElement('th');
$th->attr['bgcolor'] = 'Color';
$th->attr['height'] = 'Length';
$th->attr['nowrap'] = 'Bool#nowrap';
$th->attr['width'] = 'Length';
$td =& $this->addBlankElement('td');
$td->attr['bgcolor'] = 'Color';
$td->attr['height'] = 'Length';
$td->attr['nowrap'] = 'Bool#nowrap';
$td->attr['width'] = 'Length';
$ul =& $this->addBlankElement('ul');
$ul->attr['compact'] = 'Bool#compact';
$ul->attr['type'] = 'Enum#square,disc,circle';
}

30
library/HTMLPurifier/HTMLModule/List.php
View File

@@ -9,7 +9,6 @@ class HTMLPurifier_HTMLModule_List extends HTMLPurifier_HTMLModule
{
var $name = 'List';
var $elements = array('dl', 'dt', 'dd', 'ol', 'ul', 'li');
// According to the abstract schema, the List content set is a fully formed
// one or more expr, but it invariably occurs in an optional declaration
@@ -19,26 +18,19 @@ class HTMLPurifier_HTMLModule_List extends HTMLPurifier_HTMLModule
// Furthermore, the actual XML Schema may disagree. Regardless,
// we don't have support for such nested expressions without using
// the incredibly inefficient and draconic Custom ChildDef.
var $content_sets = array('List' => 'dl | ol | ul', 'Flow' => 'List');
var $content_sets = array('Flow' => 'List');
function HTMLPurifier_HTMLModule_List() {
foreach ($this->elements as $element) {
$this->info[$element] = new HTMLPurifier_ElementDef();
$this->info[$element]->attr = array(0 => array('Common'));
if ($element == 'li' || $element == 'dd') {
$this->info[$element]->content_model = '#PCDATA | Flow';
$this->info[$element]->content_model_type = 'optional';
} elseif ($element == 'ol' || $element == 'ul') {
$this->info[$element]->content_model = 'li';
$this->info[$element]->content_model_type = 'required';
}
}
$this->info['dt']->content_model = '#PCDATA | Inline';
$this->info['dt']->content_model_type = 'optional';
$this->info['dl']->content_model = 'dt | dd';
$this->info['dl']->content_model_type = 'required';
// this could be a LOT more robust
$this->info['li']->auto_close = array('li' => true);
$this->addElement('ol', true, 'List', 'Required: li', 'Common');
$this->addElement('ul', true, 'List', 'Required: li', 'Common');
$this->addElement('dl', true, 'List', 'Required: dt | dd', 'Common');
$li =& $this->addElement('li', true, false, 'Flow', 'Common');
$li->auto_close = array('li' => true);
$this->addElement('dd', true, false, 'Flow', 'Common');
$this->addElement('dt', true, false, 'Inline', 'Common');
}
}

16
library/HTMLPurifier/HTMLModule/NonXMLCommonAttributes.php Normal file
View File

@@ -0,0 +1,16 @@
<?php
require_once 'HTMLPurifier/HTMLModule.php';
class HTMLPurifier_HTMLModule_NonXMLCommonAttributes extends HTMLPurifier_HTMLModule
{
var $name = 'NonXMLCommonAttributes';
var $attr_collections = array(
'Lang' => array(
'lang' => 'LanguageCode',
)
);
}
?>

23
library/HTMLPurifier/HTMLModule/Presentation.php
View File

@@ -16,23 +16,16 @@ class HTMLPurifier_HTMLModule_Presentation extends HTMLPurifier_HTMLModule
{
var $name = 'Presentation';
var $elements = array('b', 'big', 'hr', 'i', 'small', 'sub', 'sup', 'tt');
var $content_sets = array(
'Block' => 'hr',
'Inline' => 'b | big | i | small | sub | sup | tt'
);
function HTMLPurifier_HTMLModule_Presentation() {
foreach ($this->elements as $element) {
$this->info[$element] = new HTMLPurifier_ElementDef();
$this->info[$element]->attr = array(0 => array('Common'));
if ($element == 'hr') {
$this->info[$element]->content_model_type = 'empty';
} else {
$this->info[$element]->content_model = '#PCDATA | Inline';
$this->info[$element]->content_model_type = 'optional';
}
}
$this->addElement('b', true, 'Inline', 'Inline', 'Common');
$this->addElement('big', true, 'Inline', 'Inline', 'Common');
$this->addElement('hr', true, 'Block', 'Empty', 'Common');
$this->addElement('i', true, 'Inline', 'Inline', 'Common');
$this->addElement('small', true, 'Inline', 'Inline', 'Common');
$this->addElement('sub', true, 'Inline', 'Inline', 'Common');
$this->addElement('sup', true, 'Inline', 'Inline', 'Common');
$this->addElement('tt', true, 'Inline', 'Inline', 'Common');
}
}

4
library/HTMLPurifier/HTMLModule/Scripting.php
View File

@@ -46,8 +46,12 @@ class HTMLPurifier_HTMLModule_Scripting extends HTMLPurifier_HTMLModule
// blockquote's custom definition (we would use it but
// blockquote's contents are optional while noscript's contents
// are required)
// TODO: convert this to new syntax, main problem is getting
// both content sets working
foreach ($this->elements as $element) {
$this->info[$element] = new HTMLPurifier_ElementDef();
$this->info[$element]->safe = false;
}
$this->info['noscript']->attr = array( 0 => array('Common') );
$this->info['noscript']->content_model = 'Heading | List | Block';

107
library/HTMLPurifier/HTMLModule/Tables.php
View File

@@ -10,75 +10,60 @@ class HTMLPurifier_HTMLModule_Tables extends HTMLPurifier_HTMLModule
{
var $name = 'Tables';
var $elements = array('caption', 'table', 'td', 'th', 'tr', 'col',
'colgroup', 'tbody', 'thead', 'tfoot');
var $content_sets = array('Block' => 'table');
function HTMLPurifier_HTMLModule_Tables() {
foreach ($this->elements as $e) {
$this->info[$e] = new HTMLPurifier_ElementDef();
$this->info[$e]->attr = array(0 => array('Common'));
$attr =& $this->info[$e]->attr;
if ($e == 'caption') continue;
if ($e == 'table'){
$attr['border'] = 'Pixels';
$attr['cellpadding'] = 'Length';
$attr['cellspacing'] = 'Length';
$attr['frame'] = new HTMLPurifier_AttrDef_Enum(array(
'void', 'above', 'below', 'hsides', 'lhs', 'rhs',
'vsides', 'box', 'border'
), false);
$attr['rules'] = new HTMLPurifier_AttrDef_Enum(array(
'none', 'groups', 'rows', 'cols', 'all'
), false);
$attr['summary'] = 'Text';
$attr['width'] = 'Length';
continue;
}
if ($e == 'col' || $e == 'colgroup') {
$attr['span'] = 'Number';
$attr['width'] = 'MultiLength';
}
if ($e == 'td' || $e == 'th') {
$attr['abbr'] = 'Text';
$attr['colspan'] = 'Number';
$attr['rowspan'] = 'Number';
}
$attr['align'] = new HTMLPurifier_AttrDef_Enum(array(
'left', 'center', 'right', 'justify', 'char'
), false);
$attr['valign'] = new HTMLPurifier_AttrDef_Enum(array(
'top', 'middle', 'bottom', 'baseline'
), false);
$attr['charoff'] = 'Length';
}
$this->info['caption']->content_model = '#PCDATA | Inline';
$this->info['caption']->content_model_type = 'optional';
// Is done directly because it doesn't leverage substitution
// mechanisms. True model is:
// 'caption?, ( col* | colgroup* ), (( thead?, tfoot?, tbody+ ) | ( tr+ ))'
$this->info['table']->child = new HTMLPurifier_ChildDef_Table();
$this->addElement('caption', true, false, 'Inline', 'Common');
$this->info['td']->content_model =
$this->info['th']->content_model = '#PCDATA | Flow';
$this->info['td']->content_model_type =
$this->info['th']->content_model_type = 'optional';
$this->addElement('table', true, 'Block',
new HTMLPurifier_ChildDef_Table(), 'Common',
array(
'border' => 'Pixels',
'cellpadding' => 'Length',
'cellspacing' => 'Length',
'frame' => 'Enum#void,above,below,hsides,lhs,rhs,vsides,box,border',
'rules' => 'Enum#none,groups,rows,cols,all',
'summary' => 'Text',
'width' => 'Length'
)
);
$this->info['tr']->content_model = 'td | th';
$this->info['tr']->content_model_type = 'required';
// common attributes
$cell_align = array(
'align' => 'Enum#left,center,right,justify,char',
'charoff' => 'Length',
'valign' => 'Enum#top,middle,bottom,baseline',
);
$this->info['col']->content_model_type = 'empty';
$cell_t = array_merge(
array(
'abbr' => 'Text',
'colspan' => 'Number',
'rowspan' => 'Number',
),
$cell_align
);
$this->addElement('td', true, false, 'Flow', 'Common', $cell_t);
$this->addElement('th', true, false, 'Flow', 'Common', $cell_t);
$this->info['colgroup']->content_model = 'col';
$this->info['colgroup']->content_model_type = 'optional';
$this->addElement('tr', true, false, 'Required: td | th', 'Common', $cell_align);
$this->info['tbody']->content_model =
$this->info['thead']->content_model =
$this->info['tfoot']->content_model = 'tr';
$this->info['tbody']->content_model_type =
$this->info['thead']->content_model_type =
$this->info['tfoot']->content_model_type = 'required';
$cell_col = array_merge(
array(
'span' => 'Number',
'width' => 'MultiLength',
),
$cell_align
);
$this->addElement('col', true, false, 'Empty', 'Common', $cell_col);
$colgroup =& $this->addElement('colgroup', true, false, 'Optional: col', 'Common', $cell_col);
$colgroup->auto_close = $this->makeLookup(
'thead', 'tbody', 'tfoot', 'tr'
);
$this->addElement('tbody', true, false, 'Required: tr', 'Common', $cell_align);
$this->addElement('thead', true, false, 'Required: tr', 'Common', $cell_align);
$this->addElement('tfoot', true, false, 'Required: tr', 'Common', $cell_align);
}

9
library/HTMLPurifier/HTMLModule/Target.php
View File

@@ -9,13 +9,12 @@ class HTMLPurifier_HTMLModule_Target extends HTMLPurifier_HTMLModule
{
var $name = 'Target';
var $elements = array('a');
function HTMLPurifier_HTMLModule_Target() {
foreach ($this->elements as $e) {
$this->info[$e] = new HTMLPurifier_ElementDef();
$this->info[$e]->standalone = false;
$this->info[$e]->attr = array(
$elements = array('a');
foreach ($elements as $name) {
$e =& $this->addBlankElement($name);
$e->attr = array(
'target' => new HTMLPurifier_AttrDef_HTML_FrameTarget()
);
}

89
library/HTMLPurifier/HTMLModule/Text.php
View File

@@ -10,65 +10,60 @@ require_once 'HTMLPurifier/HTMLModule.php';
* - Block Structural (div, p)
* - Inline Phrasal (abbr, acronym, cite, code, dfn, em, kbd, q, samp, strong, var)
* - Inline Structural (br, span)
* We have elected not to follow suite, but this may change.
* This module, functionally, does not distinguish between these
* sub-modules, but the code is internally structured to reflect
* these distinctions.
*/
class HTMLPurifier_HTMLModule_Text extends HTMLPurifier_HTMLModule
{
var $name = 'Text';
var $elements = array('abbr', 'acronym', 'address', 'blockquote',
'br', 'cite', 'code', 'dfn', 'div', 'em', 'h1', 'h2', 'h3',
'h4', 'h5', 'h6', 'kbd', 'p', 'pre', 'q', 'samp', 'span', 'strong',
'var');
var $content_sets = array(
'Heading' => 'h1 | h2 | h3 | h4 | h5 | h6',
'Block' => 'address | blockquote | div | p | pre',
'Inline' => 'abbr | acronym | br | cite | code | dfn | em | kbd | q | samp | span | strong | var',
'Flow' => 'Heading | Block | Inline'
);
function HTMLPurifier_HTMLModule_Text() {
foreach ($this->elements as $element) {
$this->info[$element] = new HTMLPurifier_ElementDef();
// attributes
if ($element == 'br') {
$this->info[$element]->attr = array(0 => array('Core'));
} elseif ($element == 'blockquote' || $element == 'q') {
$this->info[$element]->attr = array(0 => array('Common'), 'cite' => 'URI');
} else {
$this->info[$element]->attr = array(0 => array('Common'));
}
// content models
if ($element == 'br') {
$this->info[$element]->content_model_type = 'empty';
} elseif ($element == 'blockquote') {
$this->info[$element]->content_model = 'Heading | Block | List';
$this->info[$element]->content_model_type = 'optional';
} elseif ($element == 'div') {
$this->info[$element]->content_model = '#PCDATA | Flow';
$this->info[$element]->content_model_type = 'optional';
} else {
$this->info[$element]->content_model = '#PCDATA | Inline';
$this->info[$element]->content_model_type = 'optional';
}
}
// SGML permits exclusions for all descendants, but this is
// not possible with DTDs or XML Schemas. W3C has elected to
// use complicated compositions of content_models to simulate
// exclusion for children, but we go the simpler, SGML-style
// route of flat-out exclusions. Note that the Abstract Module
// is blithely unaware of such distinctions.
$this->info['pre']->excludes = array_flip(array(
'img', 'big', 'small',
'object', 'applet', 'font', 'basefont' // generally not allowed
));
$this->info['p']->auto_close = array_flip(array(
// Inline Phrasal -------------------------------------------------
$this->addElement('abbr', true, 'Inline', 'Inline', 'Common');
$this->addElement('acronym', true, 'Inline', 'Inline', 'Common');
$this->addElement('cite', true, 'Inline', 'Inline', 'Common');
$this->addElement('code', true, 'Inline', 'Inline', 'Common');
$this->addElement('dfn', true, 'Inline', 'Inline', 'Common');
$this->addElement('em', true, 'Inline', 'Inline', 'Common');
$this->addElement('kbd', true, 'Inline', 'Inline', 'Common');
$this->addElement('q', true, 'Inline', 'Inline', 'Common', array('cite' => 'URI'));
$this->addElement('samp', true, 'Inline', 'Inline', 'Common');
$this->addElement('strong', true, 'Inline', 'Inline', 'Common');
$this->addElement('var', true, 'Inline', 'Inline', 'Common');
// Inline Structural ----------------------------------------------
$this->addElement('span', true, 'Inline', 'Inline', 'Common');
$this->addElement('br', true, 'Inline', 'Empty', 'Core');
// Block Phrasal --------------------------------------------------
$this->addElement('address', true, 'Block', 'Inline', 'Common');
$this->addElement('blockquote', true, 'Block', 'Optional: Heading | Block | List', 'Common', array('cite' => 'URI') );
$pre =& $this->addElement('pre', true, 'Block', 'Inline', 'Common');
$pre->excludes = $this->makeLookup(
'img', 'big', 'small', 'object', 'applet', 'font', 'basefont' );
$this->addElement('h1', true, 'Heading', 'Inline', 'Common');
$this->addElement('h2', true, 'Heading', 'Inline', 'Common');
$this->addElement('h3', true, 'Heading', 'Inline', 'Common');
$this->addElement('h4', true, 'Heading', 'Inline', 'Common');
$this->addElement('h5', true, 'Heading', 'Inline', 'Common');
$this->addElement('h6', true, 'Heading', 'Inline', 'Common');
// Block Structural -----------------------------------------------
$p =& $this->addElement('p', true, 'Block', 'Inline', 'Common');
// this seems really ad hoc: implementing some general
// heuristics would probably be better
$p->auto_close = $this->makeLookup(
'address', 'blockquote', 'dd', 'dir', 'div', 'dl', 'dt',
'h1', 'h2', 'h3', 'h4', 'h5', 'h6', 'hr', 'ol', 'p', 'pre',
'table', 'ul'
));
'table', 'ul' );
$this->addElement('div', true, 'Block', 'Flow', 'Common');
}
}

241
library/HTMLPurifier/HTMLModule/Tidy.php Normal file
View File

@@ -0,0 +1,241 @@
<?php
require_once 'HTMLPurifier/HTMLModule.php';
HTMLPurifier_ConfigSchema::define(
'HTML', 'TidyLevel', 'medium', 'string', '
<p>General level of cleanliness the Tidy module should enforce.
There are four allowed values:</p>
<dl>
<dt>none</dt>
<dd>No extra tidying should be done</dd>
<dt>light</dt>
<dd>Only fix elements that would be discarded otherwise due to
lack of support in doctype</dd>
<dt>medium</dt>
<dd>Enforce best practices</dd>
<dt>heavy</dt>
<dd>Transform all deprecated elements and attributes to standards
compliant equivalents</dd>
</dl>
<p>This directive has been available since 2.0.0</p>
' );
HTMLPurifier_ConfigSchema::defineAllowedValues(
'HTML', 'TidyLevel', array('none', 'light', 'medium', 'heavy')
);
HTMLPurifier_ConfigSchema::define(
'HTML', 'TidyAdd', array(), 'lookup', '
Fixes to add to the default set of Tidy fixes as per your level. This
directive has been available since 2.0.0.
' );
HTMLPurifier_ConfigSchema::define(
'HTML', 'TidyRemove', array(), 'lookup', '
Fixes to remove from the default set of Tidy fixes as per your level. This
directive has been available since 2.0.0.
' );
/**
* Abstract class for a set of proprietary modules that clean up (tidy)
* poorly written HTML.
*/
class HTMLPurifier_HTMLModule_Tidy extends HTMLPurifier_HTMLModule
{
/**
* List of supported levels. Index zero is a special case "no fixes"
* level.
*/
var $levels = array(0 => 'none', 'light', 'medium', 'heavy');
/**
* Default level to place all fixes in. Disabled by default
*/
var $defaultLevel = null;
/**
* Lists of fixes used by getFixesForLevel(). Format is:
* HTMLModule_Tidy->fixesForLevel[$level] = array('fix-1', 'fix-2');
*/
var $fixesForLevel = array(
'light' => array(),
'medium' => array(),
'heavy' => array()
);
/**
* Lazy load constructs the module by determining the necessary
* fixes to create and then delegating to the populate() function.
* @todo Wildcard matching and error reporting when an added or
* subtracted fix has no effect.
*/
function construct($config) {
// create fixes, initialize fixesForLevel
$fixes = $this->makeFixes();
$this->makeFixesForLevel($fixes);
// figure out which fixes to use
$level = $config->get('HTML', 'TidyLevel');
$fixes_lookup = $this->getFixesForLevel($level);
// get custom fix declarations: these need namespace processing
$add_fixes = $config->get('HTML', 'TidyAdd');
$remove_fixes = $config->get('HTML', 'TidyRemove');
foreach ($fixes as $name => $fix) {
// needs to be refactored a little to implement globbing
if (
isset($remove_fixes[$name]) ||
(!isset($add_fixes[$name]) && !isset($fixes_lookup[$name]))
) {
unset($fixes[$name]);
}
}
// populate this module with necessary fixes
$this->populate($fixes);
}
/**
* Retrieves all fixes per a level, returning fixes for that specific
* level as well as all levels below it.
* @param $level String level identifier, see $levels for valid values
* @return Lookup up table of fixes
*/
function getFixesForLevel($level) {
if ($level == $this->levels[0]) {
return array();
}
$activated_levels = array();
for ($i = 1, $c = count($this->levels); $i < $c; $i++) {
$activated_levels[] = $this->levels[$i];
if ($this->levels[$i] == $level) break;
}
if ($i == $c) {
trigger_error(
'Tidy level ' . htmlspecialchars($level) . ' not recognized',
E_USER_WARNING
);
return array();
}
$ret = array();
foreach ($activated_levels as $level) {
foreach ($this->fixesForLevel[$level] as $fix) {
$ret[$fix] = true;
}
}
return $ret;
}
/**
* Dynamically populates the $fixesForLevel member variable using
* the fixes array. It may be custom overloaded, used in conjunction
* with $defaultLevel, or not used at all.
*/
function makeFixesForLevel($fixes) {
if (!isset($this->defaultLevel)) return;
if (!isset($this->fixesForLevel[$this->defaultLevel])) {
trigger_error(
'Default level ' . $this->defaultLevel . ' does not exist',
E_USER_ERROR
);
return;
}
$this->fixesForLevel[$this->defaultLevel] = array_keys($fixes);
}
/**
* Populates the module with transforms and other special-case code
* based on a list of fixes passed to it
* @param $lookup Lookup table of fixes to activate
*/
function populate($fixes) {
foreach ($fixes as $name => $fix) {
// determine what the fix is for
list($type, $params) = $this->getFixType($name);
switch ($type) {
case 'attr_transform_pre':
case 'attr_transform_post':
$attr = $params['attr'];
if (isset($params['element'])) {
$element = $params['element'];
if (empty($this->info[$element])) {
$e =& $this->addBlankElement($element);
} else {
$e =& $this->info[$element];
}
} else {
$type = "info_$type";
$e =& $this;
}
$f =& $e->$type;
$f[$attr] = $fix;
break;
case 'tag_transform':
$this->info_tag_transform[$params['element']] = $fix;
break;
case 'child':
case 'content_model_type':
$element = $params['element'];
if (empty($this->info[$element])) {
$e =& $this->addBlankElement($element);
} else {
$e =& $this->info[$element];
}
$e->$type = $fix;
break;
default:
trigger_error("Fix type $type not supported", E_USER_ERROR);
break;
}
}
}
/**
* Parses a fix name and determines what kind of fix it is, as well
* as other information defined by the fix
* @param $name String name of fix
* @return array(string $fix_type, array $fix_parameters)
* @note $fix_parameters is type dependant, see populate() for usage
* of these parameters
*/
function getFixType($name) {
// parse it
$property = $attr = null;
if (strpos($name, '#') !== false) list($name, $property) = explode('#', $name);
if (strpos($name, '@') !== false) list($name, $attr) = explode('@', $name);
// figure out the parameters
$params = array();
if ($name !== '') $params['element'] = $name;
if (!is_null($attr)) $params['attr'] = $attr;
// special case: attribute transform
if (!is_null($attr)) {
if (is_null($property)) $property = 'pre';
$type = 'attr_transform_' . $property;
return array($type, $params);
}
// special case: tag transform
if (is_null($property)) {
return array('tag_transform', $params);
}
return array($property, $params);
}
/**
* Defines all fixes the module will perform in a compact
* associative array of fix name to fix implementation.
* @abstract
*/
function makeFixes() {}
}
?>

18
library/HTMLPurifier/HTMLModule/Tidy/Proprietary.php Normal file
View File

@@ -0,0 +1,18 @@
<?php
require_once 'HTMLPurifier/HTMLModule/Tidy.php';
class HTMLPurifier_HTMLModule_Tidy_Proprietary extends
HTMLPurifier_HTMLModule_Tidy
{
var $name = 'Tidy_Proprietary';
var $defaultLevel = 'light';
function makeFixes() {
return array();
}
}
?>

21
library/HTMLPurifier/HTMLModule/Tidy/XHTML.php Normal file
View File

@@ -0,0 +1,21 @@
<?php
require_once 'HTMLPurifier/HTMLModule/Tidy.php';
require_once 'HTMLPurifier/AttrTransform/Lang.php';
class HTMLPurifier_HTMLModule_Tidy_XHTML extends
HTMLPurifier_HTMLModule_Tidy
{
var $name = 'Tidy_XHTML';
var $defaultLevel = 'medium';
function makeFixes() {
$r = array();
$r['@lang'] = new HTMLPurifier_AttrTransform_Lang();
return $r;
}
}
?>

193
library/HTMLPurifier/HTMLModule/Tidy/XHTMLAndHTML4.php Normal file
View File

@@ -0,0 +1,193 @@
<?php
require_once 'HTMLPurifier/HTMLModule/Tidy.php';
require_once 'HTMLPurifier/TagTransform/Simple.php';
require_once 'HTMLPurifier/TagTransform/Font.php';
require_once 'HTMLPurifier/AttrTransform/BgColor.php';
require_once 'HTMLPurifier/AttrTransform/BoolToCSS.php';
require_once 'HTMLPurifier/AttrTransform/Border.php';
require_once 'HTMLPurifier/AttrTransform/Name.php';
require_once 'HTMLPurifier/AttrTransform/Length.php';
require_once 'HTMLPurifier/AttrTransform/ImgSpace.php';
require_once 'HTMLPurifier/AttrTransform/EnumToCSS.php';
class HTMLPurifier_HTMLModule_Tidy_XHTMLAndHTML4 extends
HTMLPurifier_HTMLModule_Tidy
{
function makeFixes() {
$r = array();
// == deprecated tag transforms ===================================
$r['font'] = new HTMLPurifier_TagTransform_Font();
$r['menu'] = new HTMLPurifier_TagTransform_Simple('ul');
$r['dir'] = new HTMLPurifier_TagTransform_Simple('ul');
$r['center'] = new HTMLPurifier_TagTransform_Simple('div', 'text-align:center;');
$r['u'] = new HTMLPurifier_TagTransform_Simple('span', 'text-decoration:underline;');
$r['s'] = new HTMLPurifier_TagTransform_Simple('span', 'text-decoration:line-through;');
$r['strike'] = new HTMLPurifier_TagTransform_Simple('span', 'text-decoration:line-through;');
// == deprecated attribute transforms =============================
$r['caption@align'] =
new HTMLPurifier_AttrTransform_EnumToCSS('align', array(
// we're following IE's behavior, not Firefox's, due
// to the fact that no one supports caption-side:right,
// W3C included (with CSS 2.1). This is a slightly
// unreasonable attribute!
'left' => 'text-align:left;',
'right' => 'text-align:right;',
'top' => 'caption-side:top;',
'bottom' => 'caption-side:bottom;' // not supported by IE
));
// @align for img -------------------------------------------------
$r['img@align'] =
new HTMLPurifier_AttrTransform_EnumToCSS('align', array(
'left' => 'float:left;',
'right' => 'float:right;',
'top' => 'vertical-align:top;',
'middle' => 'vertical-align:middle;',
'bottom' => 'vertical-align:baseline;',
));
// @align for table -----------------------------------------------
$r['table@align'] =
new HTMLPurifier_AttrTransform_EnumToCSS('align', array(
'left' => 'float:left;',
'center' => 'margin-left:auto;margin-right:auto;',
'right' => 'float:right;'
));
// @align for hr -----------------------------------------------
$r['hr@align'] =
new HTMLPurifier_AttrTransform_EnumToCSS('align', array(
// we use both text-align and margin because these work
// for different browsers (IE and Firefox, respectively)
// and the melange makes for a pretty cross-compatible
// solution
'left' => 'margin-left:0;margin-right:auto;text-align:left;',
'center' => 'margin-left:auto;margin-right:auto;text-align:center;',
'right' => 'margin-left:auto;margin-right:0;text-align:right;'
));
// @align for h1, h2, h3, h4, h5, h6, p, div ----------------------
// {{{
$align_lookup = array();
$align_values = array('left', 'right', 'center', 'justify');
foreach ($align_values as $v) $align_lookup[$v] = "text-align:$v;";
// }}}
$r['h1@align'] =
$r['h2@align'] =
$r['h3@align'] =
$r['h4@align'] =
$r['h5@align'] =
$r['h6@align'] =
$r['p@align'] =
$r['div@align'] =
new HTMLPurifier_AttrTransform_EnumToCSS('align', $align_lookup);
// @bgcolor for table, tr, td, th ---------------------------------
$r['table@bgcolor'] =
$r['td@bgcolor'] =
$r['th@bgcolor'] =
new HTMLPurifier_AttrTransform_BgColor();
// @border for img ------------------------------------------------
$r['img@border'] = new HTMLPurifier_AttrTransform_Border();
// @clear for br --------------------------------------------------
$r['br@clear'] =
new HTMLPurifier_AttrTransform_EnumToCSS('clear', array(
'left' => 'clear:left;',
'right' => 'clear:right;',
'all' => 'clear:both;',
'none' => 'clear:none;',
));
// @height for td, th ---------------------------------------------
$r['td@height'] =
$r['th@height'] =
new HTMLPurifier_AttrTransform_Length('height');
// @hspace for img ------------------------------------------------
$r['img@hspace'] = new HTMLPurifier_AttrTransform_ImgSpace('hspace');
// @name for img, a -----------------------------------------------
$r['img@name'] =
$r['a@name'] = new HTMLPurifier_AttrTransform_Name();
// @noshade for hr ------------------------------------------------
// this transformation is not precise but often good enough.
// different browsers use different styles to designate noshade
$r['hr@noshade'] =
new HTMLPurifier_AttrTransform_BoolToCSS(
'noshade',
'color:#808080;background-color:#808080;border:0;'
);
// @nowrap for td, th ---------------------------------------------
$r['td@nowrap'] =
$r['th@nowrap'] =
new HTMLPurifier_AttrTransform_BoolToCSS(
'nowrap',
'white-space:nowrap;'
);
// @size for hr --------------------------------------------------
$r['hr@size'] = new HTMLPurifier_AttrTransform_Length('size', 'height');
// @type for li, ol, ul -------------------------------------------
// {{{
$ul_types = array(
'disc' => 'list-style-type:disc;',
'square' => 'list-style-type:square;',
'circle' => 'list-style-type:circle;'
);
$ol_types = array(
'1' => 'list-style-type:decimal;',
'i' => 'list-style-type:lower-roman;',
'I' => 'list-style-type:upper-roman;',
'a' => 'list-style-type:lower-alpha;',
'A' => 'list-style-type:upper-alpha;'
);
$li_types = $ul_types + $ol_types;
// }}}
$r['ul@type'] = new HTMLPurifier_AttrTransform_EnumToCSS('type', $ul_types);
$r['ol@type'] = new HTMLPurifier_AttrTransform_EnumToCSS('type', $ol_types, true);
$r['li@type'] = new HTMLPurifier_AttrTransform_EnumToCSS('type', $li_types, true);
// @vspace for img ------------------------------------------------
$r['img@vspace'] = new HTMLPurifier_AttrTransform_ImgSpace('vspace');
// @width for hr, td, th ------------------------------------------
$r['td@width'] =
$r['th@width'] =
$r['hr@width'] = new HTMLPurifier_AttrTransform_Length('width');
return $r;
}
}
class HTMLPurifier_HTMLModule_Tidy_Transitional extends
HTMLPurifier_HTMLModule_Tidy_XHTMLAndHTML4
{
var $name = 'Tidy_Transitional';
var $defaultLevel = 'heavy';
}
class HTMLPurifier_HTMLModule_Tidy_Strict extends
HTMLPurifier_HTMLModule_Tidy_XHTMLAndHTML4
{
var $name = 'Tidy_Strict';
var $defaultLevel = 'light';
}
?>

27
library/HTMLPurifier/HTMLModule/Tidy/XHTMLStrict.php Normal file
View File

@@ -0,0 +1,27 @@
<?php
require_once 'HTMLPurifier/HTMLModule/Tidy.php';
require_once 'HTMLPurifier/ChildDef/StrictBlockquote.php';
class HTMLPurifier_HTMLModule_Tidy_XHTMLStrict extends
HTMLPurifier_HTMLModule_Tidy
{
var $name = 'Tidy_XHTMLStrict';
var $defaultLevel = 'light';
function makeFixes() {
$r = array();
$r['blockquote#content_model_type'] = 'strictblockquote';
return $r;
}
var $defines_child_def = true;
function getChildDef($def) {
if ($def->content_model_type != 'strictblockquote') return false;
return new HTMLPurifier_ChildDef_StrictBlockquote($def->content_model);
}
}
?>

200
library/HTMLPurifier/HTMLModule/TransformToStrict.php
View File

@@ -1,200 +0,0 @@
<?php
require_once 'HTMLPurifier/ChildDef/StrictBlockquote.php';
require_once 'HTMLPurifier/TagTransform/Simple.php';
require_once 'HTMLPurifier/TagTransform/Center.php';
require_once 'HTMLPurifier/TagTransform/Font.php';
require_once 'HTMLPurifier/AttrTransform/Lang.php';
require_once 'HTMLPurifier/AttrTransform/BgColor.php';
require_once 'HTMLPurifier/AttrTransform/BoolToCSS.php';
require_once 'HTMLPurifier/AttrTransform/Border.php';
require_once 'HTMLPurifier/AttrTransform/Name.php';
require_once 'HTMLPurifier/AttrTransform/Length.php';
require_once 'HTMLPurifier/AttrTransform/ImgSpace.php';
require_once 'HTMLPurifier/AttrTransform/EnumToCSS.php';
/**
* Proprietary module that transforms deprecated elements into Strict
* HTML (see HTML 4.01 and XHTML 1.0) when possible.
*/
class HTMLPurifier_HTMLModule_TransformToStrict extends HTMLPurifier_HTMLModule
{
var $name = 'TransformToStrict';
// we're actually modifying these elements, not defining them
var $elements = array('h1', 'h2', 'h3', 'h4', 'h5', 'h6', 'p',
'blockquote', 'table', 'td', 'th', 'tr', 'img', 'a', 'hr', 'br',
'caption', 'ul', 'ol', 'li');
var $info_tag_transform = array(
// placeholders, see constructor for definitions
'font' => false,
'menu' => false,
'dir' => false,
'center'=> false
);
var $attr_collections = array(
'Lang' => array(
'lang' => false // placeholder
)
);
var $info_attr_transform_post = array(
'lang' => false // placeholder
);
function HTMLPurifier_HTMLModule_TransformToStrict() {
// behavior with transformations when there's another CSS property
// working on it is interesting: the CSS will *always* override
// the deprecated attribute, whereas an inline CSS declaration will
// override the corresponding declaration in, say, an external
// stylesheet. This behavior won't affect most people, but it
// does represent an operational difference we CANNOT fix.
// deprecated tag transforms
$this->info_tag_transform['font'] = new HTMLPurifier_TagTransform_Font();
$this->info_tag_transform['menu'] = new HTMLPurifier_TagTransform_Simple('ul');
$this->info_tag_transform['dir'] = new HTMLPurifier_TagTransform_Simple('ul');
$this->info_tag_transform['center'] = new HTMLPurifier_TagTransform_Center();
foreach ($this->elements as $name) {
$this->info[$name] = new HTMLPurifier_ElementDef();
$this->info[$name]->standalone = false;
}
// deprecated attribute transforms
// align battery
$align_lookup = array();
$align_values = array('left', 'right', 'center', 'justify');
foreach ($align_values as $v) $align_lookup[$v] = "text-align:$v;";
$this->info['h1']->attr_transform_pre['align'] =
$this->info['h2']->attr_transform_pre['align'] =
$this->info['h3']->attr_transform_pre['align'] =
$this->info['h4']->attr_transform_pre['align'] =
$this->info['h5']->attr_transform_pre['align'] =
$this->info['h6']->attr_transform_pre['align'] =
$this->info['p'] ->attr_transform_pre['align'] =
new HTMLPurifier_AttrTransform_EnumToCSS('align', $align_lookup);
// xml:lang <=> lang mirroring, implement in TransformToStrict,
// this is overridden in TransformToXHTML11
$this->info_attr_transform_post['lang'] = new HTMLPurifier_AttrTransform_Lang();
$this->attr_collections['Lang']['lang'] = new HTMLPurifier_AttrDef_Lang();
// this should not be applied to XHTML 1.0 Transitional, ONLY
// XHTML 1.0 Strict. We may need three classes
$this->info['blockquote']->content_model_type = 'strictblockquote';
$this->info['blockquote']->child = false; // recalculate please!
$this->info['table']->attr_transform_pre['bgcolor'] =
$this->info['tr']->attr_transform_pre['bgcolor'] =
$this->info['td']->attr_transform_pre['bgcolor'] =
$this->info['th']->attr_transform_pre['bgcolor'] = new HTMLPurifier_AttrTransform_BgColor();
$this->info['img']->attr_transform_pre['border'] = new HTMLPurifier_AttrTransform_Border();
$this->info['img']->attr_transform_pre['name'] =
$this->info['a']->attr_transform_pre['name'] = new HTMLPurifier_AttrTransform_Name();
$this->info['td']->attr_transform_pre['width'] =
$this->info['th']->attr_transform_pre['width'] =
$this->info['hr']->attr_transform_pre['width'] = new HTMLPurifier_AttrTransform_Length('width');
$this->info['td']->attr_transform_pre['nowrap'] =
$this->info['th']->attr_transform_pre['nowrap'] = new HTMLPurifier_AttrTransform_BoolToCSS('nowrap', 'white-space:nowrap;');
$this->info['td']->attr_transform_pre['height'] =
$this->info['th']->attr_transform_pre['height'] = new HTMLPurifier_AttrTransform_Length('height');
$this->info['img']->attr_transform_pre['hspace'] = new HTMLPurifier_AttrTransform_ImgSpace('hspace');
$this->info['img']->attr_transform_pre['vspace'] = new HTMLPurifier_AttrTransform_ImgSpace('vspace');
$this->info['hr']->attr_transform_pre['size'] = new HTMLPurifier_AttrTransform_Length('size', 'height');
// this transformation is not precise but often good enough.
// different browsers use different styles to designate noshade
$this->info['hr']->attr_transform_pre['noshade'] = new HTMLPurifier_AttrTransform_BoolToCSS('noshade', 'color:#808080;background-color:#808080;border: 0;');
$this->info['br']->attr_transform_pre['clear'] =
new HTMLPurifier_AttrTransform_EnumToCSS('clear', array(
'left' => 'clear:left;',
'right' => 'clear:right;',
'all' => 'clear:both;',
'none' => 'clear:none;',
));
// this is a slightly unreasonable attribute
$this->info['caption']->attr_transform_pre['align'] =
new HTMLPurifier_AttrTransform_EnumToCSS('align', array(
// we're following IE's behavior, not Firefox's, due
// to the fact that no one supports caption-side:right,
// W3C included (with CSS 2.1)
'left' => 'text-align:left;',
'right' => 'text-align:right;',
'top' => 'caption-side:top;',
'bottom' => 'caption-side:bottom;' // not supported by IE
));
$this->info['table']->attr_transform_pre['align'] =
new HTMLPurifier_AttrTransform_EnumToCSS('align', array(
'left' => 'float:left;',
'center' => 'margin-left:auto;margin-right:auto;',
'right' => 'float:right;'
));
$this->info['img']->attr_transform_pre['align'] =
new HTMLPurifier_AttrTransform_EnumToCSS('align', array(
'left' => 'float:left;',
'right' => 'float:right;',
'top' => 'vertical-align:top;',
'middle' => 'vertical-align:middle;',
'bottom' => 'vertical-align:baseline;',
));
$this->info['hr']->attr_transform_pre['align'] =
new HTMLPurifier_AttrTransform_EnumToCSS('align', array(
'left' => 'margin-left:0;margin-right:auto;text-align:left;',
'center' => 'margin-left:auto;margin-right:auto;text-align:center;',
'right' => 'margin-left:auto;margin-right:0;text-align:right;'
));
$ul_types = array(
'disc' => 'list-style-type:disc;',
'square' => 'list-style-type:square;',
'circle' => 'list-style-type:circle;'
);
$ol_types = array(
'1' => 'list-style-type:decimal;',
'i' => 'list-style-type:lower-roman;',
'I' => 'list-style-type:upper-roman;',
'a' => 'list-style-type:lower-alpha;',
'A' => 'list-style-type:upper-alpha;'
);
$li_types = $ul_types + $ol_types;
$this->info['ul']->attr_transform_pre['type'] =
new HTMLPurifier_AttrTransform_EnumToCSS('type', $ul_types);
$this->info['ol']->attr_transform_pre['type'] =
new HTMLPurifier_AttrTransform_EnumToCSS('type', $ol_types, true);
$this->info['li']->attr_transform_pre['type'] =
new HTMLPurifier_AttrTransform_EnumToCSS('type', $li_types, true);
}
var $defines_child_def = true;
function getChildDef($def) {
if ($def->content_model_type != 'strictblockquote') return false;
return new HTMLPurifier_ChildDef_StrictBlockquote($def->content_model);
}
}
?>

36
library/HTMLPurifier/HTMLModule/TransformToXHTML11.php
View File

@@ -1,36 +0,0 @@
<?php
require_once 'HTMLPurifier/AttrTransform/Lang.php';
/**
* Proprietary module that transforms XHTML 1.0 deprecated aspects into
* XHTML 1.1 compliant ones, when possible. For maximum effectiveness,
* HTMLPurifier_HTMLModule_TransformToStrict must also be loaded
* (otherwise, elements that were deprecated from Transitional to Strict
* will not be transformed).
*
* XHTML 1.1 compliant document are automatically XHTML 1.0 compliant too,
* although they may not be as friendly to legacy browsers.
*/
class HTMLPurifier_HTMLModule_TransformToXHTML11 extends HTMLPurifier_HTMLModule
{
var $name = 'TransformToXHTML11';
var $attr_collections = array(
'Lang' => array(
'lang' => false // remove it
)
);
var $info_attr_transform_post = array(
'lang' => false // remove it
);
function HTMLPurifier_HTMLModule_TransformToXHTML11() {
$this->info_attr_transform_pre['lang'] = new HTMLPurifier_AttrTransform_Lang();
}
}
?>

16
library/HTMLPurifier/HTMLModule/XMLCommonAttributes.php Normal file
View File

@@ -0,0 +1,16 @@
<?php
require_once 'HTMLPurifier/HTMLModule.php';
class HTMLPurifier_HTMLModule_XMLCommonAttributes extends HTMLPurifier_HTMLModule
{
var $name = 'XMLCommonAttributes';
var $attr_collections = array(
'Lang' => array(
'xml:lang' => 'LanguageCode',
)
);
}
?>