Drupal

tag not found', array('b') ); $this->assertHtmlToText( 'Drupal

Drupal

Drupal', "Drupal Drupal Drupal\n", 'Disallowed

, , and tags not found', array('a', 'br', 'h1') ); $this->assertHtmlToText( 'Drupal', "Drupal\n", 'Unsupported and tags not found', array('html', 'body') ); } /** * Test that whitespace is collapsed. */ function testDrupalHtmltoTextCollapsesWhitespace() { $input = "

Drupal Drupal\n\nDrupal

Drupal  Drupal\n\nDrupal

Drupal Drupal\n\nDrupal

"; // @todo The whitespace should be collapsed. $collapsed = "Drupal Drupal\n\nDrupalDrupal Drupal\n\nDrupalDrupal Drupal\n\nDrupal\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 testDrupalHtmlToTextBlockTagToNewline() { $input = '[text]' . '

[blockquote]

' . '
[br]' . '

[dl-dt]

' . '

[dt]

' . '

[dd]

' . '

[dd-dl]

' . '

[h1]

' . '

[h2]

' . '

[h3]

' . '

[h4]

' . '

[h5]

' . '

[h6]

' . '

---

[hr]' . '

1. [ol-li]
' . '
2. [li]
' . '
3. [li-ol]

' . '

[p]

' . '

• [ul-li]
' . '
• [li-ul]

' . '[text]'; $output = drupal_html_to_text($input); $pass = $this->assertFalse( preg_match('/\][^\n]*\[/s', $output), 'Block-level HTML tags should force newlines' ); if (!$pass) { $this->verbose($this->stringToHtml($output)); } $output_upper = drupal_strtoupper($output); $upper_input = drupal_strtoupper($input); $upper_output = drupal_html_to_text($upper_input); $pass = $this->assertEqual( $upper_output, $output_upper, 'Tag recognition should be case-insensitive' ); if (!$pass) { $this->verbose( $upper_output . '
should be equal to
' . $output_upper ); } } /** * Test that headers are properly separated from surrounding text. */ function testHeaderSeparation() { $html = 'Drupal

Drupal

Drupal'; // @todo There should be more space above the header than below it. $text = "Drupal\n======== DRUPAL ==============================================================\n\nDrupal\n"; $this->assertHtmlToText($html, $text, 'Text before and after

tag'); $html = '

Drupal

Drupal

Drupal'; // @todo There should be more space above the header than below it. $text = "Drupal\n\n======== DRUPAL ==============================================================\n\nDrupal\n"; $this->assertHtmlToText($html, $text, 'Paragraph before and text after

tag'); $html = 'Drupal

Drupal

Drupal

'; // @todo There should be more space above the header than below it. $text = "Drupal\n======== DRUPAL ==============================================================\n\nDrupal\n\n"; $this->assertHtmlToText($html, $text, 'Text before and paragraph after

tag'); $html = '

Drupal

Drupal

Drupal

'; $text = "Drupal\n\n======== DRUPAL ==============================================================\n\nDrupal\n\n"; $this->assertHtmlToText($html, $text, 'Paragraph before and after

tag'); } /** * Test that footnote references are properly generated. */ function testFootnoteReferences() { global $base_path, $base_url; $source = 'Host and path' . '
Host, no path' . '
Path, no host' . '
Relative path'; // @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 testDrupalHtmlToTextParagraphs() { $tests = array(); $tests[] = array( 'html' => "

line 1
\nline 2
line 3\n
line 4

paragraph

", // @todo Trailing line breaks should be trimmed. 'text' => "line 1\nline 2\nline 3\nline 4\n\nparagraph\n\n", ); $tests[] = array( 'html' => "

line 1
line 2

line 4
line 5

0

", // @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 drupal_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 * is 1000 characters." */ function testVeryLongLineWrap() { $input = 'Drupal

' . str_repeat('x', 2100) . '
Drupal'; $output = drupal_html_to_text($input); // This awkward construct comes from includes/mail.inc lines 8-13. $eol = variable_get('mail_line_endings', MAIL_LINE_ENDINGS); // We must use strlen() rather than drupal_strlen() in order to count // octets rather than characters. $line_length_limit = 1000 - drupal_strlen($eol); $maximum_line_length = 0; foreach (explode($eol, $output) as $line) { // We must use strlen() rather than drupal_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.'; // @todo This should assert that $maximum_line_length <= 1000. $this->pass($verbose); } }