Fixes #1479. Added ElggAutoP. Removing [\n\r] from test strings before compare to deal with differing whitespace between tags among PHP versions.

author: Brett Profitt <brett.profitt@gmail.com> 2012-12-10 15:50:25 -0500
committer: Brett Profitt <brett.profitt@gmail.com> 2012-12-10 15:50:25 -0500
commit: 43a395ae735777bfb5474c4f6a37dc1cd0818a37 (patch)
tree: 6c156136680474a61cf162d8ef3aceef12cfc7dd
parent: 4c517db146cb1c59c8a54d9e87b9e5a4ae17987e (diff)
download: elgg-43a395ae735777bfb5474c4f6a37dc1cd0818a37.tar.gz
elgg-43a395ae735777bfb5474c4f6a37dc1cd0818a37.tar.bz2
3 files changed, 92 insertions, 87 deletions
diff --git a/engine/classes/ElggAutop.php b/engine/classes/ElggAutoP.php
index fa0c34225..89d77e583 100644
--- a/engine/classes/ElggAutop.php
+++ b/engine/classes/ElggAutoP.php
@@ -7,11 +7,8 @@
  *
  * In DIV elements, Ps are only added when there would be at
  * least two of them.
- *
- * @author Steve Clay <steve@mrclay.org>
- * @license http://www.opensource.org/licenses/mit-license.php  MIT License
  */
-class ElggAutop {
+class ElggAutoP {
 
 	public $encoding = 'UTF-8';
 
@@ -56,8 +53,7 @@ class ElggAutop {
 
 	protected $_unique = '';
 
-	public function __construct()
-	{
+	public function __construct() {
 		$this->_blocks = preg_split('@\\s+@', $this->_blocks);
 		$this->_descendList = preg_split('@\\s+@', $this->_descendList);
 		$this->_alterList = preg_split('@\\s+@', $this->_alterList);
@@ -67,13 +63,13 @@ class ElggAutop {
 
 	/**
 	 * Intance of class for singleton pattern.
-	 * @var ElggAutop
+	 * @var ElggAutoP
 	 */
 	private static $instance;
 	
 	/**
 	 * Singleton pattern.
-	 * @return ElggAutop
+	 * @return ElggAutoP
 	 */
 	public static function getInstance() {
 		$className = __CLASS__;
@@ -94,8 +90,7 @@ class ElggAutop {
 	 * @param string $html snippet
 	 * @return string|false output or false if parse error occurred
 	 */
-	public function process($html)
-	{
+	public function process($html) {
 		// normalize whitespace
 		$html = str_replace(array("\r\n", "\r"), "\n", $html);
 
@@ -107,7 +102,8 @@ class ElggAutop {
 		// parse to DOM, suppressing loadHTML warnings
 		// http://www.php.net/manual/en/domdocument.loadhtml.php#95463
 		libxml_use_internal_errors(true);
-		if (! @$this->_doc->loadHTML("<html><meta http-equiv='content-type' " 
+
+		if (!$this->_doc->loadHTML("<html><meta http-equiv='content-type' " 
 				. "content='text/html; charset={$this->encoding}'><body>{$html}</body>"
 				. "</html>")) {
 			return false;
@@ -130,7 +126,7 @@ class ElggAutop {
 
 		// re-parse so we can handle new AUTOP elements
 
-		if (! @$this->_doc->loadHTML($html)) {
+		if (!$this->_doc->loadHTML($html)) {
 			return false;
 		}
 		// must re-create XPath object after DOM load
@@ -149,14 +145,13 @@ class ElggAutop {
 					}
 				}
 			}
-			if (! $hasContent) {
+			if (!$hasContent) {
 				// strip w/ preg_replace later (faster than moving nodes out)
 				$autop->setAttribute("r", "1");
 			}
 		}
 
 		// remove a single AUTOP inside certain elements
-		
 		foreach ($this->_xpath->query('//div') as $el) {
 			$autops = $this->_xpath->query('./autop', $el);
 			if ($autops->length === 1) {
@@ -164,9 +159,9 @@ class ElggAutop {
 				$autops->item(0)->setAttribute("r", "1");
 			}
 		}
-		
+
 		$html = $this->_doc->saveHTML();
-		
+
 		// trim to the contents of BODY
 		$bodyStart = strpos($html, '<body>');
 		$bodyEnd = strpos($html, '</body>', $bodyStart + 6);
@@ -189,8 +184,7 @@ class ElggAutop {
 	 *
 	 * @param DOMElement $el
 	 */
-	protected function _addParagraphs(DOMElement $el)
-	{
+	protected function _addParagraphs(DOMElement $el) {
 		// no need to recurse, just queue up
 		$elsToProcess = array($el);
 		$inlinesToProcess = array();
diff --git a/engine/lib/output.php b/engine/lib/output.php
index cce1c7cba..bff0bf6e9 100644
--- a/engine/lib/output.php
+++ b/engine/lib/output.php
@@ -16,7 +16,7 @@
  **/
 function parse_urls($text) {
 	// @todo this causes problems with <attr = "val">
-	// must be ing <attr="val"> format (no space).
+	// must be in <attr="val"> format (no space).
 	// By default htmlawed rewrites tags to this format.
 	// if PHP supported conditional negative lookbehinds we could use this:
 	// $r = preg_replace_callback('/(?<!=)(?<![ ])?(?<!["\'])((ht|f)tps?:\/\/[^\s\r\n\t<>"\'\!\(\),]+)/i',
@@ -46,6 +46,7 @@ function parse_urls($text) {
  *
  * @param string $pee The string
  * @deprecated Use elgg_autop instead
+ * @todo Add deprecation warning in 1.9
  *
  * @return string
  **/
@@ -56,12 +57,12 @@ function autop($pee) {
 /**
  * Create paragraphs from text with line spacing
  *
- * @param string $pee The string
+ * @param string $string The string
  *
  * @return string
  **/
-function elgg_autop($pee) {
-	return ElggAutop::getInstance()->process($pee);
+function elgg_autop($string) {
+	return ElggAutoP::getInstance()->process($string);
 }
 
 /**
@@ -358,7 +359,7 @@ function elgg_get_friendly_time($time) {
 /**
  * Strip tags and offer plugins the chance.
  * Plugins register for output:strip_tags plugin hook.
- * 	Original string included in $params['original_string']
+ * Original string included in $params['original_string']
  *
  * @param string $string Formatted string
  *
diff --git a/engine/tests/api/output.php b/engine/tests/api/output.php
index eb1a66b29..c3d5aa8c6 100644
--- a/engine/tests/api/output.php
+++ b/engine/tests/api/output.php
@@ -1,64 +1,74 @@
-<?php
-/**
- * Test case for ElggAutop functionality.
- * @author Steve Clay <steve@mrclay.org>
- */
-class ElggCoreOutputAutoPTest extends ElggCoreUnitTest {
-
-	/**
-	 * @var ElggAutop
-	 */
-	protected $_autop;
-
-	public function setUp() {
-		$this->_autop = new ElggAutop();
-	}
-	
-	public function testDomRoundtrip()
-	{
-		$d = dir(dirname(__DIR__) . '/test_files/output/autop');
-		$in = file_get_contents($d->path . "/domdoc_in.html");
-		$exp = file_get_contents($d->path . "/domdoc_exp.html");
-
-		$doc = new DOMDocument();
-		libxml_use_internal_errors(true);
-		$doc->loadHTML("<html><meta http-equiv='content-type' content='text/html; charset=utf-8'><body>"
-				. $in . '</body></html>');
-		$serialized = $doc->saveHTML();
-		list(,$out) = explode('<body>', $serialized, 2);
-		list($out) = explode('</body>', $out, 2);
-
-		$this->assertEqual($exp, $out, "DOMDocument's parsing/serialization roundtrip");
-	}
-
-	public function testProcess()
-	{
-		$data = $this->provider();
-		foreach ($data as $row) {
-			list($test, $in, $exp) = $row;
-			$out = $this->_autop->process($in);
-			$this->assertEqual($exp, $out, "Equality case {$test}");
-		}
-	}
-
-	public function provider()
-	{
-		$d = dir(dirname(__DIR__) . '/test_files/output/autop');
-		$tests = array();
-		while (false !== ($entry = $d->read())) {
-			if (preg_match('/^([a-z\\-]+)\.in\.html$/i', $entry, $m)) {
-				$tests[] = $m[1];
-			}
-		}
-
-		$data = array();
-		foreach ($tests as $test) {
-			$data[] = array(
-				$test,
-				file_get_contents($d->path . '/' . "{$test}.in.html"),
-				file_get_contents($d->path . '/' . "{$test}.exp.html"),
-			);
-		}
-		return $data;
-	}
-}
+<?php
+/**
+ * Test case for ElggAutoP functionality.
+ */
+class ElggCoreOutputAutoPTest extends ElggCoreUnitTest {
+
+	/**
+	 * @var ElggAutoP
+	 */
+	protected $_autop;
+
+	public function setUp() {
+		$this->_autop = new ElggAutoP();
+	}
+	
+	public function testDomRoundtrip() {
+		$d = dir(dirname(dirname(__FILE__)) . '/test_files/output/autop');
+		$in = file_get_contents($d->path . "/domdoc_in.html");
+		$exp = file_get_contents($d->path . "/domdoc_exp.html");
+		$exp = $this->flattenString($exp);
+
+		$doc = new DOMDocument();
+		libxml_use_internal_errors(true);
+		$doc->loadHTML("<html><meta http-equiv='content-type' content='text/html; charset=utf-8'><body>"
+				. $in . '</body></html>');
+		$serialized = $doc->saveHTML();
+		list(,$out) = explode('<body>', $serialized, 2);
+		list($out) = explode('</body>', $out, 2);
+		$out = $this->flattenString($out);
+
+		$this->assertEqual($exp, $out, "DOMDocument's parsing/serialization roundtrip");
+	}
+
+	public function testProcess() {
+		$data = $this->provider();
+		foreach ($data as $row) {
+			list($test, $in, $exp) = $row;
+			$exp = $this->flattenString($exp);
+			$out = $this->_autop->process($in);
+			$out = $this->flattenString($out);
+			
+			$this->assertEqual($exp, $out, "Equality case {$test}");
+		}
+	}
+
+	public function provider() {
+		$d = dir(dirname(dirname(__FILE__)) . '/test_files/output/autop');
+		$tests = array();
+		while (false !== ($entry = $d->read())) {
+			if (preg_match('/^([a-z\\-]+)\.in\.html$/i', $entry, $m)) {
+				$tests[] = $m[1];
+			}
+		}
+
+		$data = array();
+		foreach ($tests as $test) {
+			$data[] = array(
+				$test,
+				file_get_contents($d->path . '/' . "{$test}.in.html"),
+				file_get_contents($d->path . '/' . "{$test}.exp.html"),
+			);
+		}
+		return $data;
+	}
+
+	/**
+	 * Different versions of PHP return different whitespace between tags.
+	 * Removing all line breaks normalizes that.
+	 */
+	public function flattenString($string) {
+		$r = preg_replace('/[\n\r]+/', '', $string);
+		return $r;
+	}
+}
+\ No newline at end of file
author	Brett Profitt <brett.profitt@gmail.com>	2012-12-10 15:50:25 -0500
committer	Brett Profitt <brett.profitt@gmail.com>	2012-12-10 15:50:25 -0500
commit	43a395ae735777bfb5474c4f6a37dc1cd0818a37 (patch)
tree	6c156136680474a61cf162d8ef3aceef12cfc7dd
parent	4c517db146cb1c59c8a54d9e87b9e5a4ae17987e (diff)
download	elgg-43a395ae735777bfb5474c4f6a37dc1cd0818a37.tar.gz elgg-43a395ae735777bfb5474c4f6a37dc1cd0818a37.tar.bz2