View source
<?php
namespace Drupal\Tests\system\Functional\Mail;
use Drupal\Component\Utility\Html;
use Drupal\Core\Mail\MailFormatHelper;
use Drupal\Core\Site\Settings;
use Drupal\Tests\BrowserTestBase;
class HtmlToTextTest extends BrowserTestBase {
protected $defaultTheme = 'stark';
protected function stringToHtml($text) {
return '"' . str_replace([
"\n",
' ',
], [
'\\n',
' ',
], Html::escape($text)) . '"';
}
protected function assertHtmlToText($html, $text, $message, $allowed_tags = NULL) {
preg_match_all('/<([a-z0-6]+)/', mb_strtolower($html), $matches);
$tested_tags = implode(', ', array_unique($matches[1]));
$message .= ' (' . $tested_tags . ')';
$result = MailFormatHelper::htmlToText($html, $allowed_tags);
$this
->assertEquals($text, $result, Html::escape($message));
}
public function testTags() {
global $base_path, $base_url;
$tests = [
'<a href = "https://www.drupal.org">Drupal.org</a>' => "Drupal.org [1]\n\n[1] https://www.drupal.org\n",
"<a href = \"{$base_path}\">Homepage</a>" => "Homepage [1]\n\n[1] {$base_url}/\n",
'<address>Drupal</address>' => "Drupal\n",
'<address>Drupal</address><address>Drupal</address>' => "DrupalDrupal\n",
'<b>Drupal</b>' => "*Drupal*\n",
'<blockquote>Drupal</blockquote>' => ">Drupal\n",
'<blockquote>Drupal</blockquote><blockquote>Drupal</blockquote>' => ">Drupal\n>Drupal\n",
'<br />Drupal<br />Drupal<br /><br />Drupal' => "Drupal\nDrupal\nDrupal\n",
'<br/>Drupal<br/>Drupal<br/><br/>Drupal' => "Drupal\nDrupal\nDrupal\n",
'<br/>Drupal<br/>Drupal<br/><br/>Drupal<p>Drupal</p>' => "Drupal\nDrupal\nDrupal\nDrupal\n\n",
'<div>Drupal</div>' => "Drupal\n",
'<div>Drupal</div><div>Drupal</div>' => "DrupalDrupal\n",
'<em>Drupal</em>' => "/Drupal/\n",
'<h1>Drupal</h1>' => "======== Drupal ==============================================================\n\n",
'<h1>Drupal</h1><p>Drupal</p>' => "======== Drupal ==============================================================\n\nDrupal\n\n",
'<h2>Drupal</h2>' => "-------- Drupal --------------------------------------------------------------\n\n",
'<h2>Drupal</h2><p>Drupal</p>' => "-------- Drupal --------------------------------------------------------------\n\nDrupal\n\n",
'<h3>Drupal</h3>' => ".... Drupal\n\n",
'<h3>Drupal</h3><p>Drupal</p>' => ".... Drupal\n\nDrupal\n\n",
'<h4>Drupal</h4>' => ".. Drupal\n\n",
'<h4>Drupal</h4><p>Drupal</p>' => ".. Drupal\n\nDrupal\n\n",
'<h5>Drupal</h5>' => "Drupal\n\n",
'<h5>Drupal</h5><p>Drupal</p>' => "Drupal\n\nDrupal\n\n",
'<h6>Drupal</h6>' => "Drupal\n\n",
'<h6>Drupal</h6><p>Drupal</p>' => "Drupal\n\nDrupal\n\n",
'<hr />Drupal<hr />' => "------------------------------------------------------------------------------\nDrupal\n------------------------------------------------------------------------------\n",
'<hr/>Drupal<hr/>' => "------------------------------------------------------------------------------\nDrupal\n------------------------------------------------------------------------------\n",
'<hr/>Drupal<hr/><p>Drupal</p>' => "------------------------------------------------------------------------------\nDrupal\n------------------------------------------------------------------------------\nDrupal\n\n",
'<i>Drupal</i>' => "/Drupal/\n",
'<p>Drupal</p>' => "Drupal\n\n",
'<p>Drupal</p><p>Drupal</p>' => "Drupal\n\nDrupal\n\n",
'<strong>Drupal</strong>' => "*Drupal*\n",
'<table><tr><td>Drupal</td><td>Drupal</td></tr><tr><td>Drupal</td><td>Drupal</td></tr></table>' => "DrupalDrupalDrupalDrupal\n",
'<table><tr><td>Drupal</td></tr></table><p>Drupal</p>' => "Drupal\nDrupal\n\n",
'<u>Drupal</u>' => "Drupal\n",
'<ul><li>Drupal</li></ul>' => " * Drupal\n\n",
'<ul><li>Drupal <em>Drupal</em> Drupal</li></ul>' => " * Drupal /Drupal/ Drupal\n\n",
'<ul><li>Drupal</li><li><ol><li>Drupal</li><li>Drupal</li></ol></li></ul>' => " * Drupal\n * 1) Drupal\n 2) Drupal\n \n\n",
'<ul><li>Drupal</li><li><ol><li>Drupal</li></ol></li><li>Drupal</li></ul>' => " * Drupal\n * 1) Drupal\n \n * Drupal\n\n",
'<ul><li>Drupal</li><li>Drupal</li></ul>' => " * Drupal\n * Drupal\n\n",
'<ul><li>Drupal</li></ul><p>Drupal</p>' => " * Drupal\n\nDrupal\n\n",
'<ol><li>Drupal</li></ol>' => " 1) Drupal\n\n",
'<ol><li>Drupal</li><li><ul><li>Drupal</li><li>Drupal</li></ul></li></ol>' => " 1) Drupal\n 2) * Drupal\n * Drupal\n \n\n",
'<ol><li>Drupal</li><li>Drupal</li></ol>' => " 1) Drupal\n 2) Drupal\n\n",
'<ol>Drupal</ol>' => "Drupal\n\n",
'<ol><li>Drupal</li></ol><p>Drupal</p>' => " 1) Drupal\n\nDrupal\n\n",
'<dl><dt>Drupal</dt></dl>' => "Drupal\n\n",
'<dl><dt>Drupal</dt><dd>Drupal</dd></dl>' => "Drupal\n Drupal\n\n",
'<dl><dt>Drupal</dt><dd>Drupal</dd><dt>Drupal</dt><dd>Drupal</dd></dl>' => "Drupal\n Drupal\nDrupal\n Drupal\n\n",
'<dl><dt>Drupal</dt><dd>Drupal</dd></dl><p>Drupal</p>' => "Drupal\n Drupal\n\nDrupal\n\n",
'<dl><dt>Drupal<dd>Drupal</dl>' => "Drupal\n Drupal\n\n",
'<dl><dt>Drupal</dt></dl><p>Drupal</p>' => "Drupal\n\nDrupal\n\n",
'<ul><li>Drupal</li><li><dl><dt>Drupal</dt><dd>Drupal</dd><dt>Drupal</dt><dd>Drupal</dd></dl></li><li>Drupal</li></ul>' => " * Drupal\n * Drupal\n Drupal\n Drupal\n Drupal\n \n * Drupal\n\n",
'<br>Drupal<br>Drupal' => "Drupal\nDrupal\n",
'<hr>Drupal<hr>Drupal' => "------------------------------------------------------------------------------\nDrupal\n------------------------------------------------------------------------------\nDrupal\n",
'<ol><li>Drupal<li>Drupal</ol>' => " 1) Drupal\n 2) Drupal\n\n",
'<ul><li>Drupal <em>Drupal</em> Drupal</ul></ul>' => " * Drupal /Drupal/ Drupal\n\n",
'<ul><li>Drupal<li>Drupal</ol>' => " * Drupal\n * Drupal\n\n",
'<ul><li>Drupal<li>Drupal</ul>' => " * Drupal\n * Drupal\n\n",
'<ul>Drupal</ul>' => "Drupal\n\n",
'Drupal</ul></ol></dl><li>Drupal' => "Drupal\n * Drupal\n",
'<dl>Drupal</dl>' => "Drupal\n\n",
'<dl>Drupal</dl><p>Drupal</p>' => "Drupal\n\nDrupal\n\n",
'<dt>Drupal</dt>' => "Drupal\n",
'<html>Drupal</html>' => "Drupal\n",
'<script type="text/javascript">Drupal</script>' => "Drupal\n",
'<q>I <em>will</em> be back…</q>' => "I /will/ be back…\n",
'FrançAIS is ÜBER-åwesome' => "FrançAIS is ÜBER-åwesome\n",
];
foreach ($tests as $html => $text) {
$this
->assertHtmlToText($html, $text, 'Supported tags');
}
}
public function testDrupalHtmlToTextArgs() {
$this
->assertHtmlToText('Drupal <b>Drupal</b> Drupal', "Drupal *Drupal* Drupal\n", 'Allowed <b> tag found', [
'b',
]);
$this
->assertHtmlToText('Drupal <h1>Drupal</h1> Drupal', "Drupal Drupal Drupal\n", 'Disallowed <h1> tag not found', [
'b',
]);
$this
->assertHtmlToText('Drupal <p><em><b>Drupal</b></em><p> Drupal', "Drupal Drupal Drupal\n", 'Disallowed <p>, <em>, and <b> tags not found', [
'a',
'br',
'h1',
]);
$this
->assertHtmlToText('<html><body>Drupal</body></html>', "Drupal\n", 'Unsupported <html> and <body> tags not found', [
'html',
'body',
]);
}
public function testDrupalHtmltoTextCollapsesWhitespace() {
$input = "<p>Drupal Drupal\n\nDrupal<pre>Drupal Drupal\n\nDrupal</pre>Drupal Drupal\n\nDrupal</p>";
$collapsed = "Drupal Drupal\n\nDrupalDrupal Drupal\n\nDrupalDrupal Drupal\n\nDrupal\n\n";
$this
->assertHtmlToText($input, $collapsed, 'Whitespace is collapsed', [
'p',
]);
}
public function testDrupalHtmlToTextBlockTagToNewline() {
$input = <<<'EOT'
[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]
EOT;
$input = str_replace([
"\r",
"\n",
], '', $input);
$output = MailFormatHelper::htmlToText($input);
$this
->assertDoesNotMatchRegularExpression('/\\][^\\n]*\\[/s', $output, 'Block-level HTML tags should force newlines');
$output_upper = mb_strtoupper($output);
$upper_input = mb_strtoupper($input);
$upper_output = MailFormatHelper::htmlToText($upper_input);
$this
->assertEquals($output_upper, $upper_output, 'Tag recognition should be case-insensitive');
}
public function testHeaderSeparation() {
$html = 'Drupal<h1>Drupal</h1>Drupal';
$text = "Drupal\n======== Drupal ==============================================================\n\nDrupal\n";
$this
->assertHtmlToText($html, $text, 'Text before and after <h1> tag');
$html = '<p>Drupal</p><h1>Drupal</h1>Drupal';
$text = "Drupal\n\n======== Drupal ==============================================================\n\nDrupal\n";
$this
->assertHtmlToText($html, $text, 'Paragraph before and text after <h1> tag');
$html = 'Drupal<h1>Drupal</h1><p>Drupal</p>';
$text = "Drupal\n======== Drupal ==============================================================\n\nDrupal\n\n";
$this
->assertHtmlToText($html, $text, 'Text before and paragraph after <h1> tag');
$html = '<p>Drupal</p><h1>Drupal</h1><p>Drupal</p>';
$text = "Drupal\n\n======== Drupal ==============================================================\n\nDrupal\n\n";
$this
->assertHtmlToText($html, $text, 'Paragraph before and after <h1> tag');
}
public function testFootnoteReferences() {
global $base_path, $base_url;
$source = <<<EOT
<a href="http://www.example.com/node/1">Host and path</a>
<br /><a href="http://www.example.com">Host, no path</a>
<br /><a href="{<span class="php-variable">$base_path</span>}node/1">Path, no host</a>
<br /><a href="node/1">Relative path</a>
EOT;
$source = str_replace([
"\r",
"\n",
], '', $source);
$text = <<<EOT
Host and path [1]
Host, no path [2]
Path, no host [3]
Relative path [4]
[1] http://www.example.com/node/1
[2] http://www.example.com
[3] {<span class="php-variable">$base_url</span>}/node/1
[4] node/1
EOT;
$this
->assertHtmlToText($source, $text, 'Footnotes');
}
public function testDrupalHtmlToTextParagraphs() {
$tests = [];
$tests[] = [
'html' => "<p>line 1<br />\nline 2<br />line 3\n<br />line 4</p><p>paragraph</p>",
'text' => "line 1\nline 2\nline 3\nline 4\n\nparagraph\n\n",
];
$tests[] = [
'html' => "<p>line 1<br /> line 2</p> <p>line 4<br /> line 5</p> <p>0</p>",
'text' => "line 1\nline 2\n\nline 4\nline 5\n\n0\n\n",
];
foreach ($tests as $test) {
$this
->assertHtmlToText($test['html'], $test['text'], 'Paragraph breaks');
}
}
public function testVeryLongLineWrap() {
$input = 'Drupal<br /><p>' . str_repeat('x', 2100) . '</p><br />Drupal';
$output = MailFormatHelper::htmlToText($input);
$eol = Settings::get('mail_line_endings', PHP_EOL);
$maximum_line_length = 0;
foreach (explode($eol, $output) as $line) {
$maximum_line_length = max($maximum_line_length, strlen($line . $eol));
}
$this
->assertLessThanOrEqual(1000, $maximum_line_length);
}
public function testRemoveTrailingWhitespace() {
$text = "Hi there! \nHerp Derp";
$mail_lines = explode("\n", MailFormatHelper::wrapMail($text));
$this
->assertNotEquals(" ", substr($mail_lines[0], -1), 'Trailing whitespace removed.');
}
public function testUsenetSignature() {
$text = "Hi there!\n-- \nHerp Derp";
$mail_lines = explode("\n", MailFormatHelper::wrapMail($text));
$this
->assertEquals("-- ", $mail_lines[1], 'Trailing whitespace not removed for dash-dash-space signatures.');
$text = "Hi there!\n-- \nHerp Derp";
$mail_lines = explode("\n", MailFormatHelper::wrapMail($text));
$this
->assertEquals("--", $mail_lines[1], 'Trailing whitespace removed for incorrect dash-dash-space signatures.');
}
}