class BackdropHtmlToTextTestCase

class BackdropHtmlToTextTestCase extends BackdropWebTestCase {
  /**
   * Converts a string to its PHP source equivalent for display in test messages.
   *
   * @param $text
   *   The text string to convert.
   *
   * @return
   *   An HTML representation of the text string that, when displayed in a
   *   browser, represents the PHP source code equivalent of $text.
   */
  function stringToHtml($text) {
    return '"' .
      str_replace(
      array("\n", ' '), 
      array('\n', ' '), 
      check_plain($text)
      ) . '"';
  }

  /**
   * Helper function for testing backdrop_html_to_text().
   *
   * @param $html
   *   The source HTML string to be converted.
   * @param $text
   *   The expected result of converting $html to text.
   * @param $message
   *   A text message to display in the assertion message.
   * @param $allowed_tags
   *   (optional) An array of allowed tags, or NULL to default to the full
   *   set of tags supported by backdrop_html_to_text().
   */
  function assertHtmlToText($html, $text, $message, $allowed_tags = NULL) {
    preg_match_all('/<([a-z0-6]+)/', backdrop_strtolower($html), $matches);
    $tested_tags = implode(', ', array_unique($matches[1]));
    $message .= ' (' . $tested_tags . ')';
    $result = backdrop_html_to_text($html, $allowed_tags);
    $pass = $this->assertEqual($result, $text, check_plain($message));
    $verbose = 'html = <pre>' . $this->stringToHtml($html)
      . '</pre><br />result = <pre>' . $this->stringToHtml($result)
      . '</pre><br />expected = <pre>' . $this->stringToHtml($text)
      . '</pre>';
    $this->verbose($verbose);
    if (!$pass) {
      $this->pass("Previous test verbose info:<br />$verbose");
    }
  }

  /**
   * Test all supported tags of backdrop_html_to_text().
   */
  function testTags() {
    global $base_path, $base_url;
    $tests = array(
      // @todo Trailing linefeeds should be trimmed.
      '<a href = "https://backdropcms.org">BackdropCMS.org</a>' => "BackdropCMS.org [1]\n\n[1] https://backdropcms.org\n",
      // @todo Footer URLs should be absolute.
      "<a href = \"$base_path\">Homepage</a>" => "Homepage [1]\n\n[1] $base_url/\n",
      '<address>Backdrop</address>' => "Backdrop\n",
      // @todo The <address> tag is currently not supported.
      '<address>Backdrop</address><address>Backdrop</address>' => "BackdropBackdrop\n",
      '<b>Backdrop</b>' => "*Backdrop*\n",
      // @todo There should be a space between the '>' and the text.
      '<blockquote>Backdrop</blockquote>' => ">Backdrop\n",
      '<blockquote>Backdrop</blockquote><blockquote>Backdrop</blockquote>' => ">Backdrop\n>Backdrop\n",
      '<br />Backdrop<br />Backdrop<br /><br />Backdrop' => "Backdrop\nBackdrop\nBackdrop\n",
      '<br/>Backdrop<br/>Backdrop<br/><br/>Backdrop' => "Backdrop\nBackdrop\nBackdrop\n",
      // @todo There should be two line breaks before the paragraph.
      '<br/>Backdrop<br/>Backdrop<br/><br/>Backdrop<p>Backdrop</p>' => "Backdrop\nBackdrop\nBackdrop\nBackdrop\n\n",
      '<div>Backdrop</div>' => "Backdrop\n",
      // @todo The <div> tag is currently not supported.
      '<div>Backdrop</div><div>Backdrop</div>' => "BackdropBackdrop\n",
      '<em>Backdrop</em>' => "/Backdrop/\n",
      '<h1>Backdrop</h1>' => "======== BACKDROP ============================================================\n\n",
      '<h1>Backdrop</h1><p>Backdrop</p>' => "======== BACKDROP ============================================================\n\nBackdrop\n\n",
      '<h2>Backdrop</h2>' => "-------- BACKDROP ------------------------------------------------------------\n\n",
      '<h2>Backdrop</h2><p>Backdrop</p>' => "-------- BACKDROP ------------------------------------------------------------\n\nBackdrop\n\n",
      '<h3>Backdrop</h3>' => ".... Backdrop\n\n",
      '<h3>Backdrop</h3><p>Backdrop</p>' => ".... Backdrop\n\nBackdrop\n\n",
      '<h4>Backdrop</h4>' => ".. Backdrop\n\n",
      '<h4>Backdrop</h4><p>Backdrop</p>' => ".. Backdrop\n\nBackdrop\n\n",
      '<h5>Backdrop</h5>' => "Backdrop\n\n",
      '<h5>Backdrop</h5><p>Backdrop</p>' => "Backdrop\n\nBackdrop\n\n",
      '<h6>Backdrop</h6>' => "Backdrop\n\n",
      '<h6>Backdrop</h6><p>Backdrop</p>' => "Backdrop\n\nBackdrop\n\n",
      '<hr />Backdrop<hr />' => "------------------------------------------------------------------------------\nBackdrop\n------------------------------------------------------------------------------\n",
      '<hr/>Backdrop<hr/>' => "------------------------------------------------------------------------------\nBackdrop\n------------------------------------------------------------------------------\n",
      '<hr/>Backdrop<hr/><p>Backdrop</p>' => "------------------------------------------------------------------------------\nBackdrop\n------------------------------------------------------------------------------\nBackdrop\n\n",
      '<i>Backdrop</i>' => "/Backdrop/\n",
      '<p>Backdrop</p>' => "Backdrop\n\n",
      '<p>Backdrop</p><p>Backdrop</p>' => "Backdrop\n\nBackdrop\n\n",
      '<strong>Backdrop</strong>' => "*Backdrop*\n",
      // @todo Tables are currently not supported.
      '<table><tr><td>Backdrop</td><td>Backdrop</td></tr><tr><td>Backdrop</td><td>Backdrop</td></tr></table>' => "BackdropBackdropBackdropBackdrop\n",
      '<table><tr><td>Backdrop</td></tr></table><p>Backdrop</p>' => "Backdrop\nBackdrop\n\n",
      // @todo The <u> tag is currently not supported.
      '<u>Backdrop</u>' => "Backdrop\n",
      '<ul><li>Backdrop</li></ul>' => " * Backdrop\n\n",
      '<ul><li>Backdrop <em>Backdrop</em> Backdrop</li></ul>' => " * Backdrop /Backdrop/ Backdrop\n\n",
      // @todo Lines containing nothing but spaces should be trimmed.
      '<ul><li>Backdrop</li><li><ol><li>Backdrop</li><li>Backdrop</li></ol></li></ul>' => " * Backdrop\n *  1) Backdrop\n    2) Backdrop\n   \n\n",
      '<ul><li>Backdrop</li><li><ol><li>Backdrop</li></ol></li><li>Backdrop</li></ul>' => " * Backdrop\n *  1) Backdrop\n   \n * Backdrop\n\n",
      '<ul><li>Backdrop</li><li>Backdrop</li></ul>' => " * Backdrop\n * Backdrop\n\n",
      '<ul><li>Backdrop</li></ul><p>Backdrop</p>' => " * Backdrop\n\nBackdrop\n\n",
      '<ol><li>Backdrop</li></ol>' => " 1) Backdrop\n\n",
      '<ol><li>Backdrop</li><li><ul><li>Backdrop</li><li>Backdrop</li></ul></li></ol>' => " 1) Backdrop\n 2)  * Backdrop\n     * Backdrop\n    \n\n",
      '<ol><li>Backdrop</li><li>Backdrop</li></ol>' => " 1) Backdrop\n 2) Backdrop\n\n",
      '<ol>Backdrop</ol>' => "Backdrop\n\n",
      '<ol><li>Backdrop</li></ol><p>Backdrop</p>' => " 1) Backdrop\n\nBackdrop\n\n",
      '<dl><dt>Backdrop</dt></dl>' => "Backdrop\n\n",
      '<dl><dt>Backdrop</dt><dd>Backdrop</dd></dl>' => "Backdrop\n    Backdrop\n\n",
      '<dl><dt>Backdrop</dt><dd>Backdrop</dd><dt>Backdrop</dt><dd>Backdrop</dd></dl>' => "Backdrop\n    Backdrop\nBackdrop\n    Backdrop\n\n",
      '<dl><dt>Backdrop</dt><dd>Backdrop</dd></dl><p>Backdrop</p>' => "Backdrop\n    Backdrop\n\nBackdrop\n\n",
      '<dl><dt>Backdrop<dd>Backdrop</dl>' => "Backdrop\n    Backdrop\n\n",
      '<dl><dt>Backdrop</dt></dl><p>Backdrop</p>' => "Backdrop\n\nBackdrop\n\n",
      // @todo Again, lines containing only spaces should be trimmed.
      '<ul><li>Backdrop</li><li><dl><dt>Backdrop</dt><dd>Backdrop</dd><dt>Backdrop</dt><dd>Backdrop</dd></dl></li><li>Backdrop</li></ul>' => " * Backdrop\n * Backdrop\n       Backdrop\n   Backdrop\n       Backdrop\n   \n * Backdrop\n\n",
      // Tests malformed HTML tags.
      '<br>Backdrop<br>Backdrop' => "Backdrop\nBackdrop\n",
      '<hr>Backdrop<hr>Backdrop' => "------------------------------------------------------------------------------\nBackdrop\n------------------------------------------------------------------------------\nBackdrop\n",
      '<ol><li>Backdrop<li>Backdrop</ol>' => " 1) Backdrop\n 2) Backdrop\n\n",
      '<ul><li>Backdrop <em>Backdrop</em> Backdrop</ul></ul>' => " * Backdrop /Backdrop/ Backdrop\n\n",
      '<ul><li>Backdrop<li>Backdrop</ol>' => " * Backdrop\n * Backdrop\n\n",
      '<ul><li>Backdrop<li>Backdrop</ul>' => " * Backdrop\n * Backdrop\n\n",
      '<ul>Backdrop</ul>' => "Backdrop\n\n",
      'Backdrop</ul></ol></dl><li>Backdrop' => "Backdrop\n * Backdrop\n",
      '<dl>Backdrop</dl>' => "Backdrop\n\n",
      '<dl>Backdrop</dl><p>Backdrop</p>' => "Backdrop\n\nBackdrop\n\n",
      '<dt>Backdrop</dt>' => "Backdrop\n",
      // Tests some unsupported HTML tags.
      '<html>Backdrop</html>' => "Backdrop\n",
      // @todo Perhaps the contents of <script> tags should be dropped.
      '<script type="text/javascript">Backdrop</script>' => "Backdrop\n",
      // A couple of tests for Unicode characters.
      '<q>I <em>will</em> be back…</q>' => "I /will/ be back…\n",
      // cspell:disable-next-line
      'FrançAIS is ÜBER-åwesome' => "FrançAIS is ÜBER-åwesome\n",
    );

    foreach ($tests as $html => $text) {
      $this->assertHtmlToText($html, $text, 'Supported tags');
    }
  }

  /**
   * Test $allowed_tags argument of backdrop_html_to_text().
   */
  function testBackdropHtmlToTextArgs() {
    // The second parameter of backdrop_html_to_text() overrules the allowed tags.
    $this->assertHtmlToText(
    'Backdrop <b>Backdrop</b> Backdrop', 
    "Backdrop *Backdrop* Backdrop\n", 
    'Allowed <b> tag found', 
    array('b')
    );
    $this->assertHtmlToText(
    'Backdrop <h1>Backdrop</h1> Backdrop', 
    "Backdrop Backdrop Backdrop\n", 
    'Disallowed <h1> tag not found', 
    array('b')
    );

    $this->assertHtmlToText(
    'Backdrop <p><em><b>Backdrop</b></em><p> Backdrop', 
    "Backdrop Backdrop Backdrop\n", 
    'Disallowed <p>, <em>, and <b> tags not found', 
    array('a', 'br', 'h1')
    );

    $this->assertHtmlToText(
    '<html><body>Backdrop</body></html>', 
    "Backdrop\n", 
    'Unsupported <html> and <body> tags not found', 
    array('html', 'body')
    );
  }

  /**
   * Tests that backdrop_wrap_mail() removes trailing whitespace before newlines.
   */
  function testBackdropHtmltoTextRemoveTrailingWhitespace() {
    $text = "Hi there! \nFoo Bar";
    $mail_lines = explode("\n", backdrop_wrap_mail($text));
    $this->assertNotEqual(" ", substr($mail_lines[0], -1), 'Trailing whitespace removed.');
  }

  /**
   * Tests backdrop_wrap_mail() retains whitespace from Usenet style signatures.
   *
   * RFC 3676 says, "This is a special case; an (optionally quoted or quoted and
   * stuffed) line consisting of DASH DASH SP is neither fixed nor flowed."
   */
  function testBackdropHtmltoTextUsenetSignature() {
    $text = "Hi there!\n-- \nFoo Bar";
    $mail_lines = explode("\n", backdrop_wrap_mail($text));
    $this->assertEqual("-- ", $mail_lines[1], 'Trailing whitespace not removed for dash-dash-space signatures.');

    $text = "Hi there!\n--  \nFoo Bar";
    $mail_lines = explode("\n", backdrop_wrap_mail($text));
    $this->assertEqual("--", $mail_lines[1], 'Trailing whitespace removed for incorrect dash-dash-space signatures.');
  }

  /**
   * Test that whitespace is collapsed.
   */
  function testBackdropHtmltoTextCollapsesWhitespace() {
    $input = "<p>Backdrop  Backdrop\n\nBackdrop<pre>Backdrop  Backdrop\n\nBackdrop</pre>Backdrop  Backdrop\n\nBackdrop</p>";
    // @todo The whitespace should be collapsed.
    $collapsed = "Backdrop  Backdrop\n\nBackdropBackdrop  Backdrop\n\nBackdropBackdrop  Backdrop\n\nBackdrop\n\n";
    $this->assertHtmlToText(
    $input, 
    $collapsed, 
    'Whitespace is collapsed', 
    array('p')
    );
  }

  /**
   * Test that text separated by block-level tags in HTML get separated by
   * (at least) a newline in the plaintext version.
   */
  function testBackdropHtmlToTextBlockTagToNewline() {
    $input = '[text]'
      . '<blockquote>[blockquote]</blockquote>'
      . '<br />[br]'
      . '<dl><dt>[dl-dt]</dt>'
      . '<dt>[dt]</dt>'
      . '<dd>[dd]</dd>'
      . '<dd>[dd-dl]</dd></dl>'
      . '<h1>[h1]</h1>'
      . '<h2>[h2]</h2>'
      . '<h3>[h3]</h3>'
      . '<h4>[h4]</h4>'
      . '<h5>[h5]</h5>'
      . '<h6>[h6]</h6>'
      . '<hr />[hr]'
      . '<ol><li>[ol-li]</li>'
      . '<li>[li]</li>'
      . '<li>[li-ol]</li></ol>'
      . '<p>[p]</p>'
      . '<ul><li>[ul-li]</li>'
      . '<li>[li-ul]</li></ul>'
      . '[text]';
    $output = backdrop_html_to_text($input);
    $pass = $this->assertFalse(
    preg_match('/\][^\n]*\[/s', $output), 
    'Block-level HTML tags should force newlines'
    );
    if (!$pass) {
      $this->verbose($this->stringToHtml($output));
    }
    $output_upper = backdrop_strtoupper($output);
    $upper_input = backdrop_strtoupper($input);
    $upper_output = backdrop_html_to_text($upper_input);
    $pass = $this->assertEqual(
    $upper_output, 
    $output_upper, 
    'Tag recognition should be case-insensitive'
    );
    if (!$pass) {
      $this->verbose(
      $upper_output
        . '<br />should  be equal to <br />'
        . $output_upper
        );
    }
  }

  /**
   * Test that headers are properly separated from surrounding text.
   */
  function testHeaderSeparation() {
    $html = 'Backdrop<h1>Backdrop</h1>Backdrop';
    // @todo There should be more space above the header than below it.
    $text = "Backdrop\n======== BACKDROP ============================================================\n\nBackdrop\n";
    $this->assertHtmlToText($html, $text, 
    'Text before and after <h1> tag');
    $html = '<p>Backdrop</p><h1>Backdrop</h1>Backdrop';
    // @todo There should be more space above the header than below it.
    $text = "Backdrop\n\n======== BACKDROP ============================================================\n\nBackdrop\n";
    $this->assertHtmlToText($html, $text, 
    'Paragraph before and text after <h1> tag');
    $html = 'Backdrop<h1>Backdrop</h1><p>Backdrop</p>';
    // @todo There should be more space above the header than below it.
    $text = "Backdrop\n======== BACKDROP ============================================================\n\nBackdrop\n\n";
    $this->assertHtmlToText($html, $text, 
    'Text before and paragraph after <h1> tag');
    $html = '<p>Backdrop</p><h1>Backdrop</h1><p>Backdrop</p>';
    $text = "Backdrop\n\n======== BACKDROP ============================================================\n\nBackdrop\n\n";
    $this->assertHtmlToText($html, $text, 
    'Paragraph before and after <h1> tag');
  }

  /**
   * Test that footnote references are properly generated.
   */
  function testFootnoteReferences() {
    global $base_path, $base_url;
    $source = '<a href="http://www.example.com/node/1">Host and path</a>'
      . '<br /><a href="http://www.example.com">Host, no path</a>'
      . '<br /><a href="' . $base_path . 'node/1">Path, no host</a>'
      . '<br /><a href="node/1">Relative path</a>';
    // @todo Footnote URLs should be absolute.
    $tt = "Host and path [1]"
      . "\nHost, no path [2]"
      // @todo The following two references should be combined.
      . "\nPath, no host [3]"
      . "\nRelative path [4]"
      . "\n"
      . "\n[1] http://www.example.com/node/1"
      . "\n[2] http://www.example.com"
      // @todo The following two references should be combined.
      . "\n[3] $base_url/node/1"
      . "\n[4] node/1\n";
    $this->assertHtmlToText($source, $tt, 'Footnotes');
  }

  /**
   * Test that combinations of paragraph breaks, line breaks, linefeeds,
   * and spaces are properly handled.
   */
  function testBackdropHtmlToTextParagraphs() {
    $tests = array();
    $tests[] = array(
      'html' => "<p>line 1<br />\nline 2<br />line 3\n<br />line 4</p><p>paragraph</p>",
      // @todo Trailing line breaks should be trimmed.
      'text' => "line 1\nline 2\nline 3\nline 4\n\nparagraph\n\n",
    );
    $tests[] = array(
      'html' => "<p>line 1<br /> line 2</p> <p>line 4<br /> line 5</p> <p>0</p>",
      // @todo Trailing line breaks should be trimmed.
      'text' => "line 1\nline 2\n\nline 4\nline 5\n\n0\n\n",
    );
    foreach ($tests as $test) {
      $this->assertHtmlToText($test['html'], $test['text'], 'Paragraph breaks');
    }
  }

  /**
   * Tests that backdrop_html_to_text() wraps before 1000 characters.
   *
   * RFC 3676 says, "The Text/Plain media type is the lowest common
   * denominator of Internet email, with lines of no more than 998 characters."
   *
   * RFC 2046 says, "SMTP [RFC-821] allows a maximum of 998 octets before the
   * next CRLF sequence."
   *
   * RFC 821 says, "The maximum total length of a text line including the
   * <CRLF> is 1000 characters."
   */
  function testVeryLongLineWrap() {
    $input = 'Backdrop<br /><p>' . str_repeat('x', 2100) . '</p><br />Backdrop';
    $output = backdrop_html_to_text($input);
    // This awkward construct comes from includes/mail.inc lines 8-13.
    $eol = settings_get('mail_line_endings', MAIL_LINE_ENDINGS);
    // We must use strlen() rather than backdrop_strlen() in order to count
    // octets rather than characters.
    $line_length_limit = 1000 - backdrop_strlen($eol);
    $maximum_line_length = 0;
    foreach (explode($eol, $output) as $line) {
      // We must use strlen() rather than backdrop_strlen() in order to count
      // octets rather than characters.
      $maximum_line_length = max($maximum_line_length, strlen($line . $eol));
    }
    $verbose = 'Maximum line length found was ' . $maximum_line_length . ' octets.';
    $this->pass($verbose);
    $this->assertTrue($maximum_line_length <= 1000, $verbose);
  }
}