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 /**
  18  * Tests our html2text hacks
  19  *
  20  * Note: includes original tests from testweblib.php
  21  *
  22  * @package    core
  23  * @category   phpunit
  24  * @copyright  2012 Petr Skoda {@link http://skodak.org}
  25  * @license    http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
  26  */
  27
  28
  29 defined('MOODLE_INTERNAL') || die();
  30
  31
  32 class core_html2text_testcase extends basic_testcase {
  33
  34     /**
  35      * ALT as image replacements.
  36      */
  37     public function test_images() {
  38         $this->assertSame('[edit]', html_to_text('<img src="edit.png" alt="edit" />'));
  39
  40         $text = 'xx<img src="gif.gif" alt="some gif" />xx';
  41         $result = html_to_text($text, null, false, false);
  42         $this->assertSame($result, 'xx[some gif]xx');
  43     }
  44
  45     /**
  46      * No magic quotes messing.
  47      */
  48     public function test_no_strip_slashes() {
  49         $this->assertSame('[\edit]', html_to_text('<img src="edit.png" alt="\edit" />'));
  50
  51         $text = '\\magic\\quotes\\are\\\\horrible';
  52         $result = html_to_text($text, null, false, false);
  53         $this->assertSame($result, $text);
  54     }
  55
  56     /**
  57      * core_text integration.
  58      */
  59     public function test_core_text() {
  60         $text = '<strong>Žluťoučký koníček</strong>';
  61         $result = html_to_text($text, null, false, false);
  62         $this->assertSame($result, 'ŽLUŤOUČKÝ KONÍČEK');
  63     }
  64
  65     /**
  66      * Protect 0.
  67      */
  68     public function test_zero() {
  69         $text = '0';
  70         $result = html_to_text($text, null, false, false);
  71         $this->assertSame($result, $text);
  72
  73         $this->assertSame('0', html_to_text('0'));
  74     }
  75
  76     /**
  77      * Test the links list enumeration.
  78      */
  79     public function test_build_link_list() {
  80
  81         // Note the trailing whitespace left intentionally in the text after first link.
  82         $text = 'Total of <a title="List of integrated issues"
  83             href="http://tr.mdl.org/sh.jspa?r=1&j=p+%3D+%22I+d%22+%3D">     ' . '
  84             <strong>27 issues</strong></a> and <a href="http://another.url/?f=a&amp;b=2">some</a> other
  85 have been fixed <strong><a href="http://third.url/view.php">last week</a></strong>';
  86
  87         // Do not collect links.
  88         $result = html_to_text($text, 5000, false);
  89         $this->assertSame('Total of 27 ISSUES and some other have been fixed LAST WEEK', $result);
  90
  91         // Collect and enumerate links.
  92         $result = html_to_text($text, 5000, true);
  93         $this->assertSame(0, strpos($result, 'Total of 27 ISSUES [1] and some [2] other have been fixed LAST WEEK [3]'));
  94         $this->assertSame(false, strpos($result, '[0]'));
  95         $this->assertSame(1, preg_match('|^'.preg_quote('[1] http://tr.mdl.org/sh.jspa?r=1&j=p+%3D+%22I+d%22+%3D').'$|m', $result));
  96         $this->assertSame(1, preg_match('|^'.preg_quote('[2] http://another.url/?f=a&amp;b=2').'$|m', $result));
  97         $this->assertSame(1, preg_match('|^'.preg_quote('[3] http://third.url/view.php').'$|m', $result));
  98         $this->assertSame(false, strpos($result, '[4]'));
  99
 100         // Test multiple occurrences of the same URL.
 101         $text = '<p>See <a href="http://moodle.org">moodle.org</a>,
 102             <a href="http://www.google.fr">google</a>, <a href="http://www.univ-lemans.fr">univ-lemans</a>
 103             and <a href="http://www.google.fr">google</a>.
 104             Also try <a href="https://www.google.fr">google via HTTPS</a>.';
 105         $result = html_to_text($text, 5000, true);
 106         $this->assertSame(0, strpos($result, 'See moodle.org [1], google [2], univ-lemans [3] and google [2]. Also try google via HTTPS [4].'));
 107         $this->assertSame(false, strpos($result, '[0]'));
 108         $this->assertSame(1, preg_match('|^'.preg_quote('[1] http://moodle.org').'$|m', $result));
 109         $this->assertSame(1, preg_match('|^'.preg_quote('[2] http://www.google.fr').'$|m', $result));
 110         $this->assertSame(1, preg_match('|^'.preg_quote('[3] http://www.univ-lemans.fr').'$|m', $result));
 111         $this->assertSame(1, preg_match('|^'.preg_quote('[4] https://www.google.fr').'$|m', $result));
 112         $this->assertSame(false, strpos($result, '[5]'));
 113     }
 114
 115     /**
 116      * Various invalid HTML typed by users that ignore html strict.
 117      **/
 118     public function test_invalid_html() {
 119         $text = 'Gin & Tonic';
 120         $result = html_to_text($text, null, false, false);
 121         $this->assertSame($result, $text);
 122
 123         $text = 'Gin > Tonic';
 124         $result = html_to_text($text, null, false, false);
 125         $this->assertSame($result, $text);
 126
 127         $text = 'Gin < Tonic';
 128         $result = html_to_text($text, null, false, false);
 129         $this->assertSame($result, $text);
 130     }
 131
 132     /**
 133      * Basic text formatting.
 134      */
 135     public function test_simple() {
 136         $this->assertSame("_Hello_ WORLD!\n", html_to_text('<p><i>Hello</i> <b>world</b>!</p>'));
 137         $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>'));
 138         $this->assertSame("HELLO WORLD!\n\n", html_to_text('<h1>Hello world!</h1>'));
 139         $this->assertSame("Hello\nworld!", html_to_text('Hello<br />world!'));
 140     }
 141
 142     /**
 143      * Test line wrapping.
 144      */
 145     public function test_text_nowrap() {
 146         $long = "Here is a long string, more than 75 characters long, since by default html_to_text wraps text at 75 chars.";
 147         $wrapped = "Here is a long string, more than 75 characters long, since by default\nhtml_to_text wraps text at 75 chars.";
 148         $this->assertSame($long, html_to_text($long, 0));
 149         $this->assertSame($wrapped, html_to_text($long));
 150     }
 151
 152     /**
 153      * Whitespace removal.
 154      */
 155     public function test_trailing_whitespace() {
 156         $this->assertSame('With trailing whitespace and some more text', html_to_text("With trailing whitespace   \nand some   more text", 0));
 157     }
 158
 159     /**
 160      * PRE parsing.
 161      */
 162     public function test_html_to_text_pre_parsing_problem() {
 163         $strorig = 'Consider the following function:<br /><pre><span style="color: rgb(153, 51, 102);">void FillMeUp(char* in_string) {'.
 164             '<br />  int i = 0;<br />  while (in_string[i] != \'\0\') {<br />    in_string[i] = \'X\';<br />    i++;<br />  }<br />'.
 165             '}</span></pre>What would happen if a non-terminated string were input to this function?<br /><br />';
 166
 167         // Note, the spaces in the <pre> section are Unicode NBSPs - they may not be displayed in your editor.
 168         $strconv = 'Consider the following function:
 169
 170 void FillMeUp(char* in_string) {
 171   int i = 0;
 172   while (in_string[i] != \'\0\') {
 173     in_string[i] = \'X\';
 174     i++;
 175   }
 176 }
 177 What would happen if a non-terminated string were input to this function?
 178
 179 ';
 180
 181         $this->assertSame($strconv, html_to_text($strorig));
 182     }
 183
 184     /**
 185      * Scripts should be stripped.
 186      */
 187     public function test_strip_scripts() {
 188         $this->assertSame('Interesting text',
 189                 html_to_text('Interesting <script type="text/javascript">var what_a_mess = "Yuck!";</script> text', 0));
 190     }
 191 }