mirror of
https://github.com/ezyang/htmlpurifier.git
synced 2025-01-03 05:11:52 +00:00
[2.0.1] Add preliminary auto-paragraph implementation. It needs to be aggressively refactored and generalized.
git-svn-id: http://htmlpurifier.org/svnroot/htmlpurifier/trunk@1202 48356398-32a2-884e-a903-53898d9a118a
This commit is contained in:
Edward Z. Yang
parent
03657ad51a
commit
eee45fed37
@ -4,6 +4,16 @@ require_once 'HTMLPurifier/Strategy.php';
|
||||
require_once 'HTMLPurifier/HTMLDefinition.php';
|
||||
require_once 'HTMLPurifier/Generator.php';
|
||||
|
||||
HTMLPurifier_ConfigSchema::define(
|
||||
'Core', 'AutoParagraph', false, 'bool', '
|
||||
<p>
|
||||
This directive will cause HTML Purifier to automatically paragraph text
|
||||
in the document fragment root based on two newlines and block tags.
|
||||
This directive has been available since 2.0.1.
|
||||
</p>
|
||||
'
|
||||
);
|
||||
|
||||
/**
|
||||
* Takes tokens makes them well-formed (balance end tags, etc.)
|
||||
*/
|
||||
@ -15,10 +25,17 @@ class HTMLPurifier_Strategy_MakeWellFormed extends HTMLPurifier_Strategy
|
||||
$generator = new HTMLPurifier_Generator();
|
||||
$result = array();
|
||||
$current_nesting = array();
|
||||
|
||||
$escape_invalid_tags = $config->get('Core', 'EscapeInvalidTags');
|
||||
foreach ($tokens as $token) {
|
||||
$auto_paragraph = $config->get('Core', 'AutoParagraph');
|
||||
|
||||
for ($k = 0, $tokens_count = count($tokens); $k < $tokens_count; $k++) {
|
||||
$token = $tokens[$k];
|
||||
if (empty( $token->is_tag )) {
|
||||
$result[] = $token;
|
||||
if ($auto_paragraph && $token->type === 'text') {
|
||||
$this->autoParagraphText($result, $current_nesting, $tokens, $k, $token, $context, $config);
|
||||
}
|
||||
if ($token) $result[] = $token;
|
||||
continue;
|
||||
}
|
||||
|
||||
@ -74,6 +91,8 @@ class HTMLPurifier_Strategy_MakeWellFormed extends HTMLPurifier_Strategy
|
||||
$current_nesting[] = $parent; // undo the pop
|
||||
}
|
||||
|
||||
if ($auto_paragraph) $this->autoParagraphStart($result, $current_nesting, $tokens, $k, $token, $context, $config);
|
||||
|
||||
$result[] = $token;
|
||||
$current_nesting[] = $token;
|
||||
continue;
|
||||
@ -155,6 +174,117 @@ class HTMLPurifier_Strategy_MakeWellFormed extends HTMLPurifier_Strategy
|
||||
return $result;
|
||||
}
|
||||
|
||||
/**
|
||||
* Sub-function call for auto-paragraphing for any old text node.
|
||||
* This will eventually
|
||||
* be factored out into a generic Formatter class
|
||||
* @note This function does not care at all about ending paragraph
|
||||
* tags: the rest of MakeWellFormed handles that!
|
||||
*/
|
||||
function autoParagraphText(&$result, &$current_nesting, $tokens, $k, &$token, &$context, $config) {
|
||||
$dnl = PHP_EOL . PHP_EOL; // double-newline
|
||||
// paragraphing is on
|
||||
if (empty($current_nesting)) {
|
||||
// we're in root node, great time to start a paragraph
|
||||
// since we're also dealing with a text node
|
||||
$result[] = new HTMLPurifier_Token_Start('p');
|
||||
$current_nesting[] = new HTMLPurifier_Token_Start('p');
|
||||
$this->autoParagraphSplitText($result, $current_nesting, $tokens, $k, $token, $context, $config);
|
||||
} else {
|
||||
// we're not in root node, so let's see whether or not
|
||||
// we're in a paragraph
|
||||
|
||||
// losslessly access the parent element
|
||||
$parent = array_pop($current_nesting);
|
||||
$current_nesting[] = $parent;
|
||||
|
||||
if ($parent->name === 'p') {
|
||||
$this->autoParagraphSplitText($result, $current_nesting, $tokens, $k, $token, $context, $config);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Sub-function for auto-paragraphing that takes a token and splits it
|
||||
* up into paragraphs unconditionally. Requires that a paragraph was
|
||||
* already started
|
||||
*/
|
||||
function autoParagraphSplitText(&$result, &$current_nesting, $tokens, $k, &$token, &$context, $config) {
|
||||
$dnl = PHP_EOL . PHP_EOL; // double-newline
|
||||
$definition = $config->getHTMLDefinition();
|
||||
|
||||
$raw_paragraphs = explode($dnl, $token->data);
|
||||
|
||||
$token = false; // token has been completely dismantled
|
||||
|
||||
// remove empty paragraphs
|
||||
$paragraphs = array();
|
||||
foreach ($raw_paragraphs as $par) {
|
||||
if (trim($par) !== '') $paragraphs[] = $par;
|
||||
}
|
||||
|
||||
if (empty($paragraphs) && count($raw_paragraphs) > 1) {
|
||||
$result[] = new HTMLPurifier_Token_End('p');
|
||||
array_pop($current_nesting);
|
||||
return;
|
||||
}
|
||||
|
||||
foreach ($paragraphs as $data) {
|
||||
$result[] = new HTMLPurifier_Token_Text($data);
|
||||
$result[] = new HTMLPurifier_Token_End('p');
|
||||
$result[] = new HTMLPurifier_Token_Start('p');
|
||||
}
|
||||
array_pop($result); // remove trailing start token
|
||||
|
||||
// check the outside to determine whether or not
|
||||
// another start tag is needed
|
||||
$end_paragraph = $this->autoParagraphEndParagraph($tokens, $k, $definition);
|
||||
if (!$end_paragraph) {
|
||||
array_pop($result);
|
||||
} else {
|
||||
array_pop($current_nesting);
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
/**
|
||||
* Determines if up-coming code requires an end-paragraph tag,
|
||||
* otherwise, keep the paragraph open (don't make another one)
|
||||
* @protected
|
||||
*/
|
||||
function autoParagraphEndParagraph($tokens, $k, $definition) {
|
||||
$end_paragraph = false;
|
||||
for ($j = $k + 1; isset($tokens[$j]); $j++) {
|
||||
if ($tokens[$j]->type == 'start' || $tokens[$j]->type == 'empty') {
|
||||
if ($tokens[$j]->name == 'p') $end_paragraph = true;
|
||||
else $end_paragraph = isset($definition->info['p']->auto_close[$tokens[$j]->name]);
|
||||
break;
|
||||
} elseif ($tokens[$j]->type == 'text') {
|
||||
if (!$tokens[$j]->is_whitespace) {
|
||||
$end_paragraph = false;
|
||||
break;
|
||||
}
|
||||
} elseif ($tokens[$j]->type == 'end') {
|
||||
// nonsensical case
|
||||
$end_paragraph = false;
|
||||
break;
|
||||
}
|
||||
}
|
||||
return $end_paragraph;
|
||||
}
|
||||
|
||||
/**
|
||||
* Sub-function for auto-paragraphing that processes element starts
|
||||
*/
|
||||
function autoParagraphStart(&$result, &$current_nesting, $tokens, $k, &$token, &$context, $config) {
|
||||
if (!empty($current_nesting)) return;
|
||||
$definition = $config->getHTMLDefinition();
|
||||
// a better check would be to use the projected new algorithm
|
||||
// for auto_close
|
||||
if (isset($definition->info['p']->auto_close[$token->name])) return;
|
||||
$result[] = $current_nesting[] = new HTMLPurifier_Token_Start('p');
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
?>
|
||||
@ -74,6 +74,83 @@ class HTMLPurifier_Strategy_MakeWellFormedTest extends HTMLPurifier_StrategyHarn
|
||||
|
||||
}
|
||||
|
||||
function testAutoParagraph() {
|
||||
$this->config = array('Core.AutoParagraph' => true);
|
||||
|
||||
$this->assertResult(
|
||||
'Foobar',
|
||||
'<p>Foobar</p>'
|
||||
);
|
||||
|
||||
$this->assertResult(
|
||||
'Par 1
|
||||
Par 1 still',
|
||||
'<p>Par 1
|
||||
Par 1 still</p>'
|
||||
);
|
||||
|
||||
$this->assertResult(
|
||||
'Par1
|
||||
|
||||
Par2',
|
||||
'<p>Par1</p><p>Par2</p>'
|
||||
);
|
||||
|
||||
$this->assertResult(
|
||||
'<b>Par1</b>
|
||||
|
||||
<i>Par2</i>',
|
||||
'<p><b>Par1</b></p><p><i>Par2</i></p>'
|
||||
);
|
||||
|
||||
|
||||
$this->assertResult(
|
||||
'<b>Par1
|
||||
|
||||
Par2</b>',
|
||||
'<p><b>Par1
|
||||
|
||||
Par2</b></p>'
|
||||
);
|
||||
|
||||
$this->assertResult(
|
||||
'Par1<p>Par2</p>',
|
||||
'<p>Par1</p><p>Par2</p>'
|
||||
);
|
||||
|
||||
$this->assertResult(
|
||||
'<b>Par1',
|
||||
'<p><b>Par1</b></p>'
|
||||
);
|
||||
|
||||
$this->assertResult(
|
||||
'<pre>Par1
|
||||
|
||||
Par1</pre>'
|
||||
);
|
||||
|
||||
$this->assertResult(
|
||||
'Par1
|
||||
|
||||
',
|
||||
'<p>Par1</p>'
|
||||
);
|
||||
$this->assertResult(
|
||||
'Par1
|
||||
|
||||
<div>Par2</div>
|
||||
|
||||
Par3',
|
||||
'<p>Par1</p><div>Par2</div><p>Par3</p>'
|
||||
);
|
||||
|
||||
$this->assertResult(
|
||||
'Par<b>1</b>',
|
||||
'<p>Par<b>1</b></p>'
|
||||
);
|
||||
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
?>
|
||||
Loading…
Reference in New Issue
Block a user