Chris@0: <?php
Chris@0: 
Chris@0: namespace Drupal\Tests\search\Functional;
Chris@0: 
Chris@0: use Drupal\Component\Utility\Unicode;
Chris@0: 
Chris@0: /**
Chris@0:  * Tests that the search_simply() function works as intended.
Chris@0:  *
Chris@0:  * @group search
Chris@0:  */
Chris@0: class SearchSimplifyTest extends SearchTestBase {
Chris@0:   /**
Chris@0:    * Tests that all Unicode characters simplify correctly.
Chris@0:    */
Chris@0:   public function testSearchSimplifyUnicode() {
Chris@0:     // This test uses a file that was constructed so that the even lines are
Chris@0:     // boundary characters, and the odd lines are valid word characters. (It
Chris@0:     // was generated as a sequence of all the Unicode characters, and then the
Chris@0:     // boundary characters (punctuation, spaces, etc.) were split off into
Chris@0:     // their own lines).  So the even-numbered lines should simplify to nothing,
Chris@0:     // and the odd-numbered lines we need to split into shorter chunks and
Chris@0:     // verify that simplification doesn't lose any characters.
Chris@0:     $input = file_get_contents(\Drupal::root() . '/core/modules/search/tests/UnicodeTest.txt');
Chris@0:     $basestrings = explode(chr(10), $input);
Chris@0:     $strings = [];
Chris@0:     foreach ($basestrings as $key => $string) {
Chris@0:       if ($key % 2) {
Chris@0:         // Even line - should simplify down to a space.
Chris@0:         $simplified = search_simplify($string);
Chris@0:         $this->assertIdentical($simplified, ' ', "Line $key is excluded from the index");
Chris@0:       }
Chris@0:       else {
Chris@0:         // Odd line, should be word characters.
Chris@0:         // Split this into 30-character chunks, so we don't run into limits
Chris@0:         // of truncation in search_simplify().
Chris@0:         $start = 0;
Chris@0:         while ($start < Unicode::strlen($string)) {
Chris@0:           $newstr = Unicode::substr($string, $start, 30);
Chris@0:           // Special case: leading zeros are removed from numeric strings,
Chris@0:           // and there's one string in this file that is numbers starting with
Chris@0:           // zero, so prepend a 1 on that string.
Chris@0:           if (preg_match('/^[0-9]+$/', $newstr)) {
Chris@0:             $newstr = '1' . $newstr;
Chris@0:           }
Chris@0:           $strings[] = $newstr;
Chris@0:           $start += 30;
Chris@0:         }
Chris@0:       }
Chris@0:     }
Chris@0:     foreach ($strings as $key => $string) {
Chris@0:       $simplified = search_simplify($string);
Chris@0:       $this->assertTrue(Unicode::strlen($simplified) >= Unicode::strlen($string), "Nothing is removed from string $key.");
Chris@0:     }
Chris@0: 
Chris@0:     // Test the low-numbered ASCII control characters separately. They are not
Chris@0:     // in the text file because they are problematic for diff, especially \0.
Chris@0:     $string = '';
Chris@0:     for ($i = 0; $i < 32; $i++) {
Chris@0:       $string .= chr($i);
Chris@0:     }
Chris@0:     $this->assertIdentical(' ', search_simplify($string), 'Search simplify works for ASCII control characters.');
Chris@0:   }
Chris@0: 
Chris@0:   /**
Chris@0:    * Tests that search_simplify() does the right thing with punctuation.
Chris@0:    */
Chris@0:   public function testSearchSimplifyPunctuation() {
Chris@0:     $cases = [
Chris@0:       ['20.03/94-28,876', '20039428876', 'Punctuation removed from numbers'],
Chris@0:       ['great...drupal--module', 'great drupal module', 'Multiple dot and dashes are word boundaries'],
Chris@0:       ['very_great-drupal.module', 'verygreatdrupalmodule', 'Single dot, dash, underscore are removed'],
Chris@0:       ['regular,punctuation;word', 'regular punctuation word', 'Punctuation is a word boundary'],
Chris@0:     ];
Chris@0: 
Chris@0:     foreach ($cases as $case) {
Chris@0:       $out = trim(search_simplify($case[0]));
Chris@0:       $this->assertEqual($out, $case[1], $case[2]);
Chris@0:     }
Chris@0:   }
Chris@0: 
Chris@0: }