lib/tests/html2text_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 namespace core;
  18
  19 /**
  20  * Tests our html2text hacks
  21  *
  22  * Note: includes original tests from testweblib.php
  23  *
  24  * @package    core
  25  * @category   test
  26  * @copyright  2012 Petr Skoda {@link http://skodak.org}
  27  * @license    http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
  28  */
  29 class html2text_test extends \basic_testcase {
  30
  31     /**
  32      * ALT as image replacements.
  33      */
  34     public function test_images() {
  35         $this->assertSame('[edit]', html_to_text('<img src="edit.png" alt="edit" />'));
  36
  37         $text = 'xx<img src="gif.gif" alt="some gif" />xx';
  38         $result = html_to_text($text, null, false, false);
  39         $this->assertSame($result, 'xx[some gif]xx');
  40     }
  41
  42     /**
  43      * No magic quotes messing.
  44      */
  45     public function test_no_strip_slashes() {
  46         $this->assertSame('[\edit]', html_to_text('<img src="edit.png" alt="\edit" />'));
  47
  48         $text = '\\magic\\quotes\\are\\\\horrible';
  49         $result = html_to_text($text, null, false, false);
  50         $this->assertSame($result, $text);
  51     }
  52
  53     /**
  54      * core_text integration.
  55      */
  56     public function test_core_text() {
  57         $text = '<strong>Žluťoučký koníček</strong>';
  58         $result = html_to_text($text, null, false, false);
  59         $this->assertSame($result, 'ŽLUŤOUČKÝ KONÍČEK');
  60     }
  61
  62     /**
  63      * Protect 0.
  64      */
  65     public function test_zero() {
  66         $text = '0';
  67         $result = html_to_text($text, null, false, false);
  68         $this->assertSame($result, $text);
  69
  70         $this->assertSame('0', html_to_text('0'));
  71     }
  72
  73     /**
  74      * Test the links list enumeration.
  75      */
  76     public function test_build_link_list() {
  77
  78         // Note the trailing whitespace left intentionally in the text after first link.
  79         $text = 'Total of <a title="List of integrated issues"
  80             href="http://tr.mdl.org/sh.jspa?r=1&j=p+%3D+%22I+d%22+%3D">     ' . '
  81             <strong>27 issues</strong></a> and <a href="http://another.url/?f=a&amp;b=2">some</a> other
  82 have been fixed <strong><a href="http://third.url/view.php">last week</a></strong>';
  83
  84         // Do not collect links.
  85         $result = html_to_text($text, 5000, false);
  86         $this->assertSame('Total of 27 ISSUES and some other have been fixed LAST WEEK', $result);
  87
  88         // Collect and enumerate links.
  89         $result = html_to_text($text, 5000, true);
  90         $this->assertSame(0, strpos($result, 'Total of 27 ISSUES [1] and some [2] other have been fixed LAST WEEK [3]'));
  91         $this->assertSame(false, strpos($result, '[0]'));
  92         $this->assertSame(1, preg_match('|^'.preg_quote('[1] http://tr.mdl.org/sh.jspa?r=1&j=p+%3D+%22I+d%22+%3D').'$|m', $result));
  93         $this->assertSame(1, preg_match('|^'.preg_quote('[2] http://another.url/?f=a&amp;b=2').'$|m', $result));
  94         $this->assertSame(1, preg_match('|^'.preg_quote('[3] http://third.url/view.php').'$|m', $result));
  95         $this->assertSame(false, strpos($result, '[4]'));
  96
  97         // Test multiple occurrences of the same URL.
  98         $text = '<p>See <a href="http://moodle.org">moodle.org</a>,
  99             <a href="http://www.google.fr">google</a>, <a href="http://www.univ-lemans.fr">univ-lemans</a>
 100             and <a href="http://www.google.fr">google</a>.
 101             Also try <a href="https://www.google.fr">google via HTTPS</a>.';
 102         $result = html_to_text($text, 5000, true);
 103         $this->assertSame(0, strpos($result, 'See moodle.org [1], google [2], univ-lemans [3] and google [2]. Also try google via HTTPS [4].'));
 104         $this->assertSame(false, strpos($result, '[0]'));
 105         $this->assertSame(1, preg_match('|^'.preg_quote('[1] http://moodle.org').'$|m', $result));
 106         $this->assertSame(1, preg_match('|^'.preg_quote('[2] http://www.google.fr').'$|m', $result));
 107         $this->assertSame(1, preg_match('|^'.preg_quote('[3] http://www.univ-lemans.fr').'$|m', $result));
 108         $this->assertSame(1, preg_match('|^'.preg_quote('[4] https://www.google.fr').'$|m', $result));
 109         $this->assertSame(false, strpos($result, '[5]'));
 110     }
 111
 112     /**
 113      * Various invalid HTML typed by users that ignore html strict.
 114      **/
 115     public function test_invalid_html() {
 116         $text = 'Gin & Tonic';
 117         $result = html_to_text($text, null, false, false);
 118         $this->assertSame($result, $text);
 119
 120         $text = 'Gin > Tonic';
 121         $result = html_to_text($text, null, false, false);
 122         $this->assertSame($result, $text);
 123
 124         $text = 'Gin < Tonic';
 125         $result = html_to_text($text, null, false, false);
 126         $this->assertSame($result, $text);
 127     }
 128
 129     /**
 130      * Basic text formatting.
 131      */
 132     public function test_simple() {
 133         $this->assertSame("_Hello_ WORLD!\n", html_to_text('<p><i>Hello</i> <b>world</b>!</p>'));
 134         $this->assertSame("All the WORLD’S a stage.\n\n-- William Shakespeare\n", html_to_text('<p>All the <strong>world’s</strong> a stage.</p><p>-- William Shakespeare</p>'));
 135         $this->assertSame("HELLO WORLD!\n\n", html_to_text('<h1>Hello world!</h1>'));
 136         $this->assertSame("Hello\nworld!", html_to_text('Hello<br />world!'));
 137     }
 138
 139     /**
 140      * Test line wrapping.
 141      */
 142     public function test_text_nowrap() {
 143         $long = "Here is a long string, more than 75 characters long, since by default html_to_text wraps text at 75 chars.";
 144         $wrapped = "Here is a long string, more than 75 characters long, since by default\nhtml_to_text wraps text at 75 chars.";
 145         $this->assertSame($long, html_to_text($long, 0));
 146         $this->assertSame($wrapped, html_to_text($long));
 147     }
 148
 149     /**
 150      * Whitespace removal.
 151      */
 152     public function test_trailing_whitespace() {
 153         $this->assertSame('With trailing whitespace and some more text', html_to_text("With trailing whitespace   \nand some   more text", 0));
 154     }
 155
 156     /**
 157      * PRE parsing.
 158      */
 159     public function test_html_to_text_pre_parsing_problem() {
 160         $strorig = 'Consider the following function:<br /><pre><span style="color: rgb(153, 51, 102);">void FillMeUp(char* in_string) {'.
 161             '<br />  int i = 0;<br />  while (in_string[i] != \'\0\') {<br />    in_string[i] = \'X\';<br />    i++;<br />  }<br />'.
 162             '}</span></pre>What would happen if a non-terminated string were input to this function?<br /><br />';
 163
 164         // Note, the spaces in the <pre> section are Unicode NBSPs - they may not be displayed in your editor.
 165         $strconv = 'Consider the following function:
 166
 167 void FillMeUp(char* in_string) {
 168   int i = 0;
 169   while (in_string[i] != \'\0\') {
 170     in_string[i] = \'X\';
 171     i++;
 172   }
 173 }
 174 What would happen if a non-terminated string were input to this function?
 175
 176 ';
 177
 178         $this->assertSame($strconv, html_to_text($strorig));
 179     }
 180
 181     /**
 182      * Scripts should be stripped.
 183      */
 184     public function test_strip_scripts() {
 185         $this->assertSame('Interesting text',
 186                 html_to_text('Interesting <script type="text/javascript">var what_a_mess = "Yuck!";</script> text', 0));
 187     }
 188 }