mirror/php-htmlpurifier
1
0
Fork 0
mirror of https://github.com/ezyang/htmlpurifier.git synced 2025-08-04 05:07:55 +02:00
Code Issues Releases Wiki Activity

Migrate strategies to separate classes complete.

Browse Source 
git-svn-id: http://htmlpurifier.org/svnroot/htmlpurifier/trunk@119 48356398-32a2-884e-a903-53898d9a118a
This commit is contained in:
Edward Z. Yang
2006-07-24 02:49:37 +00:00
parent f75e8dab8c
commit 619d5d9bc1
10 changed files with 33 additions and 558 deletions
Download Patch File Download Diff File Expand all files Collapse all files

282
library/HTMLPurifier/Definition.php
View File

@@ -33,8 +33,17 @@ class HTMLPurifier_Definition
'ul' => true
);
function instance() {
static $instance = null;
if (!$instance) {
$instance = new HTMLPurifier_Definition();
}
return $instance;
}
function HTMLPurifier_Definition() {
$this->generator = new HTMLPurifier_Generator();
$this->loadData();
}
function loadData() {
@@ -154,279 +163,6 @@ class HTMLPurifier_Definition
}
function purifyTokens($tokens) {
if (empty($this->info)) $this->loadData();
$tokens = $this->removeForeignElements($tokens);
$tokens = $this->makeWellFormed($tokens);
$tokens = $this->fixNesting($tokens);
$tokens = $this->validateAttributes($tokens);
return $tokens;
}
function removeForeignElements($tokens) {
if (empty($this->info)) $this->loadData();
$result = array();
foreach($tokens as $token) {
if (!empty( $token->is_tag )) {
if (!isset($this->info[$token->name])) {
// invalid tag, generate HTML and insert in
$token = new HTMLPurifier_Token_Text(
$this->generator->generateFromToken($token)
);
}
} elseif ($token->type == 'comment') {
// strip comments
continue;
} elseif ($token->type == 'text') {
} else {
continue;
}
$result[] = $token;
}
return $result;
}
function makeWellFormed($tokens) {
if (empty($this->info)) $this->loadData();
$result = array();
$current_nesting = array();
foreach ($tokens as $token) {
if (empty( $token->is_tag )) {
$result[] = $token;
continue;
}
$info = $this->info[$token->name]; // assumption but valid
// test if it claims to be a start tag but is empty
if ($info->child_def->type == 'empty' &&
$token->type == 'start' ) {
$result[] = new HTMLPurifier_Token_Empty($token->name,
$token->attributes);
continue;
}
// test if it claims to be empty but really is a start tag
if ($info->child_def->type != 'empty' &&
$token->type == 'empty' ) {
$result[] = new HTMLPurifier_Token_Start($token->name,
$token->attributes);
$result[] = new HTMLPurifier_Token_End($token->name);
continue;
}
// automatically insert empty tags
if ($token->type == 'empty') {
$result[] = $token;
continue;
}
// we give start tags precedence, so automatically accept unless...
// it's one of those special cases
if ($token->type == 'start') {
// if there's a parent, check for special case
if (!empty($current_nesting)) {
$current_parent = array_pop($current_nesting);
// check if we're closing a P tag
if ($current_parent->name == 'p' &&
isset($this->info_closes_p[$token->name])
) {
$result[] = new HTMLPurifier_Token_End('p');
$result[] = $token;
$current_nesting[] = $token;
continue;
}
// check if we're closing a LI tag
if ($current_parent->name == 'li' &&
$token->name == 'li'
) {
$result[] = new HTMLPurifier_Token_End('li');
$result[] = $token;
$current_nesting[] = $token;
continue;
}
// this is more TIDY stuff
// we should also get some TABLE related code
// mismatched h#
$current_nesting[] = $current_parent; // undo the pop
}
$result[] = $token;
$current_nesting[] = $token;
continue;
}
// sanity check
if ($token->type != 'end') continue;
// okay, we're dealing with a closing tag
// make sure that we have something open
if (empty($current_nesting)) {
$result[] = new HTMLPurifier_Token_Text(
$this->generator->generateFromToken($token)
);
continue;
}
// first, check for the simplest case: everything closes neatly
// current_nesting is modified
$current_parent = array_pop($current_nesting);
if ($current_parent->name == $token->name) {
$result[] = $token;
continue;
}
// undo the array_pop
$current_nesting[] = $current_parent;
// okay, so we're trying to close the wrong tag
// scroll back the entire nest, trying to find our tag
// feature could be to specify how far you'd like to go
$size = count($current_nesting);
// -2 because -1 is the last element, but we already checked that
$skipped_tags = false;
for ($i = $size - 2; $i >= 0; $i--) {
if ($current_nesting[$i]->name == $token->name) {
// current nesting is modified
$skipped_tags = array_splice($current_nesting, $i);
break;
}
}
// we still didn't find the tag, so translate to text
if ($skipped_tags === false) {
$result[] = new HTMLPurifier_Token_Text(
$this->generator->generateFromToken($token)
);
continue;
}
// okay, we found it, close all the skipped tags
// note that skipped tags contains the element we need closed
$size = count($skipped_tags);
for ($i = $size - 1; $i >= 0; $i--) {
$result[] = new HTMLPurifier_Token_End($skipped_tags[$i]->name);
}
// done!
}
// we're at the end now, fix all still unclosed tags
if (!empty($current_nesting)) {
$size = count($current_nesting);
for ($i = $size - 1; $i >= 0; $i--) {
$result[] =
new HTMLPurifier_Token_End($current_nesting[$i]->name);
}
}
return $result;
}
function fixNesting($tokens) {
if (empty($this->info)) $this->loadData();
// insert implicit "parent" node, will be removed at end
array_unshift($tokens, new HTMLPurifier_Token_Start('div'));
$tokens[] = new HTMLPurifier_Token_End('div');
for ($i = 0, $size = count($tokens) ; $i < $size; ) {
$child_tokens = array();
// scroll to the end of this node, and report number
for ($j = $i, $depth = 0; ; $j++) {
if ($tokens[$j]->type == 'start') {
$depth++;
// skip token assignment on first iteration
if ($depth == 1) continue;
} elseif ($tokens[$j]->type == 'end') {
$depth--;
// skip token assignment on last iteration
if ($depth == 0) break;
}
$child_tokens[] = $tokens[$j];
}
// $i is index of start token
// $j is index of end token
// have DTD child def validate children
$element_def = $this->info[$tokens[$i]->name];
$result = $element_def->child_def->validateChildren($child_tokens);
// process result
if ($result === true) {
// leave the nodes as is
} elseif($result === false) {
// WARNING WARNING WARNING!!!
// While for the original DTD, there will never be
// cascading removal, more complex ones may have such
// a problem.
// If you modify the info array such that an element
// that requires children may contain a child that requires
// children, you need to also scroll back and re-check that
// elements parent node
$length = $j - $i + 1;
// remove entire node
array_splice($tokens, $i, $length);
// change size
$size -= $length;
// ensure that we scroll to the next node
$i--;
} else {
$length = $j - $i - 1;
// replace node with $result
array_splice($tokens, $i + 1, $length, $result);
// change size
$size -= $length;
$size += count($result);
}
// scroll to next node
$i++;
while ($i < $size and $tokens[$i]->type != 'start') $i++;
}
// remove implicit divs
array_shift($tokens);
array_pop($tokens);
return $tokens;
}
function validateAttributes($tokens) {
if (empty($this->info)) $this->loadData();
}
}
class HTMLPurifier_ElementDef

2
library/HTMLPurifier/Strategy/FixNesting.php
View File

@@ -8,7 +8,7 @@ class HTMLPurifier_Strategy_FixNesting extends HTMLPurifier_Strategy
var $definition;
function HTMLPurifier_Definition() {
function HTMLPurifier_Strategy_FixNesting() {
$this->definition = HTMLPurifier_Definition::instance();
}

2
library/HTMLPurifier/Strategy/RemoveForeignElements.php
View File

@@ -10,7 +10,7 @@ class HTMLPurifier_Strategy_RemoveForeignElements extends HTMLPurifier_Strategy
var $generator;
var $definition;
function HTMLPurifier_Definition() {
function HTMLPurifier_Strategy_RemoveForeignElements() {
$this->generator = new HTMLPurifier_Generator();
$this->definition = HTMLPurifier_Definition::instance();
}