lib/tests/htmlpurifier_test.php

   1 <?php
   2 // This file is part of Moodle - http://moodle.org/
   3 //
   4 // Moodle is free software: you can redistribute it and/or modify
   5 // it under the terms of the GNU General Public License as published by
   6 // the Free Software Foundation, either version 3 of the License, or
   7 // (at your option) any later version.
   8 //
   9 // Moodle is distributed in the hope that it will be useful,
  10 // but WITHOUT ANY WARRANTY; without even the implied warranty of
  11 // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  12 // GNU General Public License for more details.
  13 //
  14 // You should have received a copy of the GNU General Public License
  15 // along with Moodle.  If not, see <http://www.gnu.org/licenses/>.
  16
  17 /**
  18  * Unit tests for the HTMLPurifier integration
  19  *
  20  * @package    core
  21  * @category   phpunit
  22  * @copyright  2012 Petr Skoda {@link http://skodak.org}
  23  * @license    http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
  24  */
  25
  26 defined('MOODLE_INTERNAL') || die();
  27
  28
  29 /**
  30  * HTMLPurifier test case
  31  *
  32  * @package    core
  33  * @category   phpunit
  34  * @copyright  2012 Petr Skoda {@link http://skodak.org}
  35  * @license    http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
  36  */
  37 class core_htmlpurifier_testcase extends basic_testcase {
  38
  39     /**
  40      * Verify _blank target is allowed.
  41      */
  42     public function test_allow_blank_target() {
  43         $text = '<a href="http://moodle.org" target="_blank">Some link</a>';
  44         $result = format_text($text, FORMAT_HTML);
  45         $this->assertSame($text, $result);
  46
  47         $result = format_text('<a href="http://moodle.org" target="some">Some link</a>', FORMAT_HTML);
  48         $this->assertSame('<a href="http://moodle.org">Some link</a>', $result);
  49     }
  50
  51     /**
  52      * Verify our nolink tag accepted.
  53      */
  54     public function test_nolink() {
  55         // We can not use format text because nolink changes result.
  56         $text = '<nolink><div>no filters</div></nolink>';
  57         $result = purify_html($text, array());
  58         $this->assertSame($text, $result);
  59
  60         $text = '<nolink>xxx<em>xx</em><div>xxx</div></nolink>';
  61         $result = purify_html($text, array());
  62         $this->assertSame($text, $result);
  63     }
  64
  65     /**
  66      * Verify our tex tag accepted.
  67      */
  68     public function test_tex() {
  69         $text = '<tex>a+b=c</tex>';
  70         $result = purify_html($text, array());
  71         $this->assertSame($text, $result);
  72     }
  73
  74     /**
  75      * Verify our algebra tag accepted.
  76      */
  77     public function test_algebra() {
  78         $text = '<algebra>a+b=c</algebra>';
  79         $result = purify_html($text, array());
  80         $this->assertSame($text, $result);
  81     }
  82
  83     /**
  84      * Verify our hacky multilang works.
  85      */
  86     public function test_multilang() {
  87         $text = '<lang lang="en">hmmm</lang><lang lang="anything">hm</lang>';
  88         $result = purify_html($text, array());
  89         $this->assertSame($text, $result);
  90
  91         $text = '<span lang="en" class="multilang">hmmm</span><span lang="anything" class="multilang">hm</span>';
  92         $result = purify_html($text, array());
  93         $this->assertSame($text, $result);
  94
  95         $text = '<span lang="en">hmmm</span>';
  96         $result = purify_html($text, array());
  97         $this->assertNotSame($text, $result);
  98
  99         // Keep standard lang tags.
 100
 101         $text = '<span lang="de_DU" class="multilang">asas</span>';
 102         $result = purify_html($text, array());
 103         $this->assertSame($text, $result);
 104
 105         $text = '<lang lang="de_DU">xxxxxx</lang>';
 106         $result = purify_html($text, array());
 107         $this->assertSame($text, $result);
 108     }
 109
 110     /**
 111      * Tests the 'allowid' option for format_text.
 112      */
 113     public function test_format_text_allowid() {
 114         // Start off by not allowing ids (default).
 115         $options = array(
 116             'nocache' => true
 117         );
 118         $result = format_text('<div id="example">Frog</div>', FORMAT_HTML, $options);
 119         $this->assertSame('<div>Frog</div>', $result);
 120
 121         // Now allow ids.
 122         $options['allowid'] = true;
 123         $result = format_text('<div id="example">Frog</div>', FORMAT_HTML, $options);
 124         $this->assertSame('<div id="example">Frog</div>', $result);
 125     }
 126
 127     public function test_allowobjectembed() {
 128         global $CFG;
 129
 130         $this->assertSame('0', $CFG->allowobjectembed);
 131
 132         $text = '<object width="425" height="350">
 133 <param name="movie" value="http://www.youtube.com/v/AyPzM5WK8ys" />
 134 <param name="wmode" value="transparent" />
 135 <embed src="http://www.youtube.com/v/AyPzM5WK8ys" type="application/x-shockwave-flash" wmode="transparent" width="425" height="350" />
 136 </object>hmmm';
 137         $result = purify_html($text, array());
 138         $this->assertSame('hmmm', trim($result));
 139
 140         $CFG->allowobjectembed = '1';
 141
 142         $expected = '<object width="425" height="350" data="http://www.youtube.com/v/AyPzM5WK8ys" type="application/x-shockwave-flash">
 143 <param name="allowScriptAccess" value="never" />
 144 <param name="allowNetworking" value="internal" />
 145 <param name="movie" value="http://www.youtube.com/v/AyPzM5WK8ys" />
 146 <param name="wmode" value="transparent" />
 147 <embed src="http://www.youtube.com/v/AyPzM5WK8ys" type="application/x-shockwave-flash" wmode="transparent" width="425" height="350" allowscriptaccess="never" allownetworking="internal" />
 148 </object>hmmm';
 149         $result = purify_html($text, array());
 150         $this->assertSame(str_replace("\n", '', $expected), str_replace("\n", '', $result));
 151
 152         $CFG->allowobjectembed = '0';
 153
 154         $result = purify_html($text, array());
 155         $this->assertSame('hmmm', trim($result));
 156     }
 157
 158     /**
 159      * Test if linebreaks kept unchanged.
 160      */
 161     public function test_line_breaking() {
 162         $text = "\n\raa\rsss\nsss\r";
 163         $this->assertSame($text, purify_html($text));
 164     }
 165
 166     /**
 167      * Test fixing of strict problems.
 168      */
 169     public function test_tidy() {
 170         $text = "<p>xx";
 171         $this->assertSame('<p>xx</p>', purify_html($text));
 172
 173         $text = "<P>xx</P>";
 174         $this->assertSame('<p>xx</p>', purify_html($text));
 175
 176         $text = "xx<br>";
 177         $this->assertSame('xx<br />', purify_html($text));
 178     }
 179
 180     /**
 181      * Test nesting - this used to cause problems in earlier versions.
 182      */
 183     public function test_nested_lists() {
 184         $text = "<ul><li>One<ul><li>Two</li></ul></li><li>Three</li></ul>";
 185         $this->assertSame($text, purify_html($text));
 186     }
 187
 188     /**
 189      * Test that XSS protection works, complete smoke tests are in htmlpurifier itself.
 190      */
 191     public function test_cleaning_nastiness() {
 192         $text = "x<SCRIPT>alert('XSS')</SCRIPT>x";
 193         $this->assertSame('xx', purify_html($text));
 194
 195         $text = '<DIV STYLE="background-image:url(javascript:alert(\'XSS\'))">xx</DIV>';
 196         $this->assertSame('<div>xx</div>', purify_html($text));
 197
 198         $text = '<DIV STYLE="width:expression(alert(\'XSS\'));">xx</DIV>';
 199         $this->assertSame('<div>xx</div>', purify_html($text));
 200
 201         $text = 'x<IFRAME SRC="javascript:alert(\'XSS\');"></IFRAME>x';
 202         $this->assertSame('xx', purify_html($text));
 203
 204         $text = 'x<OBJECT TYPE="text/x-scriptlet" DATA="http://ha.ckers.org/scriptlet.html"></OBJECT>x';
 205         $this->assertSame('xx', purify_html($text));
 206
 207         $text = 'x<EMBED SRC="http://ha.ckers.org/xss.swf" AllowScriptAccess="always"></EMBED>x';
 208         $this->assertSame('xx', purify_html($text));
 209
 210         $text = 'x<form></form>x';
 211         $this->assertSame('xx', purify_html($text));
 212     }
 213
 214     /**
 215      * Test internal function used for clean_text() speedup.
 216      */
 217     public function test_is_purify_html_necessary() {
 218         // First our shortcuts.
 219         $text = "";
 220         $this->assertFalse(is_purify_html_necessary($text));
 221         $this->assertSame($text, purify_html($text));
 222
 223         $text = "666";
 224         $this->assertFalse(is_purify_html_necessary($text));
 225         $this->assertSame($text, purify_html($text));
 226
 227         $text = "abc\ndef \" ' ";
 228         $this->assertFalse(is_purify_html_necessary($text));
 229         $this->assertSame($text, purify_html($text));
 230
 231         $text = "abc\n<p>def</p>efg<p>hij</p>";
 232         $this->assertFalse(is_purify_html_necessary($text));
 233         $this->assertSame($text, purify_html($text));
 234
 235         $text = "<br />abc\n<p>def<em>efg</em><strong>hi<br />j</strong></p>";
 236         $this->assertFalse(is_purify_html_necessary($text));
 237         $this->assertSame($text, purify_html($text));
 238
 239         // Now failures.
 240         $text = "&nbsp;";
 241         $this->assertTrue(is_purify_html_necessary($text));
 242
 243         $text = "Gin & Tonic";
 244         $this->assertTrue(is_purify_html_necessary($text));
 245
 246         $text = "Gin > Tonic";
 247         $this->assertTrue(is_purify_html_necessary($text));
 248
 249         $text = "Gin < Tonic";
 250         $this->assertTrue(is_purify_html_necessary($text));
 251
 252         $text = "<div>abc</div>";
 253         $this->assertTrue(is_purify_html_necessary($text));
 254
 255         $text = "<span>abc</span>";
 256         $this->assertTrue(is_purify_html_necessary($text));
 257
 258         $text = "<br>abc";
 259         $this->assertTrue(is_purify_html_necessary($text));
 260
 261         $text = "<p class='xxx'>abc</p>";
 262         $this->assertTrue(is_purify_html_necessary($text));
 263
 264         $text = "<p>abc<em></p></em>";
 265         $this->assertTrue(is_purify_html_necessary($text));
 266
 267         $text = "<p>abc";
 268         $this->assertTrue(is_purify_html_necessary($text));
 269     }
 270
 271     public function test_allowed_schemes() {
 272         // First standard schemas.
 273         $text = '<a href="http://www.example.com/course/view.php?id=5">link</a>';
 274         $this->assertSame($text, purify_html($text));
 275
 276         $text = '<a href="https://www.example.com/course/view.php?id=5">link</a>';
 277         $this->assertSame($text, purify_html($text));
 278
 279         $text = '<a href="ftp://user@ftp.example.com/some/file.txt">link</a>';
 280         $this->assertSame($text, purify_html($text));
 281
 282         $text = '<a href="nntp://example.com/group/123">link</a>';
 283         $this->assertSame($text, purify_html($text));
 284
 285         $text = '<a href="news:groupname">link</a>';
 286         $this->assertSame($text, purify_html($text));
 287
 288         $text = '<a href="mailto:user@example.com">link</a>';
 289         $this->assertSame($text, purify_html($text));
 290
 291         // Extra schemes allowed in moodle.
 292         $text = '<a href="irc://irc.example.com/3213?pass">link</a>';
 293         $this->assertSame($text, purify_html($text));
 294
 295         $text = '<a href="rtsp://www.example.com/movie.mov">link</a>';
 296         $this->assertSame($text, purify_html($text));
 297
 298         $text = '<a href="teamspeak://speak.example.com/?par=val?par2=val2">link</a>';
 299         $this->assertSame($text, purify_html($text));
 300
 301         $text = '<a href="gopher://gopher.example.com/resource">link</a>';
 302         $this->assertSame($text, purify_html($text));
 303
 304         $text = '<a href="mms://www.example.com/movie.mms">link</a>';
 305         $this->assertSame($text, purify_html($text));
 306
 307         // Now some borked or dangerous schemes.
 308         $text = '<a href="javascript://www.example.com">link</a>';
 309         $this->assertSame('<a>link</a>', purify_html($text));
 310
 311         $text = '<a href="hmmm://www.example.com">link</a>';
 312         $this->assertSame('<a>link</a>', purify_html($text));
 313     }
 314 }