makary/phpbb
1
0
Fork 0
mirror of https://github.com/phpbb/phpbb.git synced 2025-06-07 20:08:53 +00:00
Code Issues Projects Releases Packages Wiki Activity

[ticket/17387] Fix handling unicode strings

Browse source 
PHPBB-17387
This commit is contained in:
rxu 2024-09-02 22:26:58 +07:00
parent 472b36877c
commit c26ded6025
No known key found for this signature in database
GPG key ID: 955F0567380E586A
2 changed files with 7 additions and 8 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

13
phpBB/includes/functions_content.php
View file

@ -336,7 +336,7 @@ function get_context(string $text, array $words, int $length = 400): string
$text = htmlspecialchars_decode($text);
// Replace all spaces/invisible characters with single spaces
$text = preg_replace("/\s+/u", ' ', $text);
$text = preg_replace("/[\p{Z}\h\v]+/u", ' ', $text);
$text_length = utf8_strlen($text);
@ -351,7 +351,6 @@ function get_context(string $text, array $words, int $length = 400): string
$word_indexes[$pos] = $word;
}
}
if (!empty($word_indexes))
{
ksort($word_indexes);
@ -400,21 +399,21 @@ function get_context(string $text, array $words, int $length = 400): string
$fragment_end = $end - $start + 1;
// Find the first valid alphanumeric character in the fragment to don't cut words
if ($start > 0 && preg_match('/[^\p{L}\p{N}][\p{L}\p{N}]/ui', $fragment, $matches, PREG_OFFSET_CAPTURE))
if ($start > 0 && preg_match('/[^\p{L}\p{N}][\p{L}\p{N}]/u', $fragment, $matches, PREG_OFFSET_CAPTURE))
{
$fragment_start = (int) $matches[0][1] + 1; // first valid alphanumeric character
$fragment_start = mb_strlen(substr($fragment, 0, (int) $matches[0][1])) + 1;
}
// Find the last valid alphanumeric character in the fragment to don't cut words
if ($end < $text_length - 1 && preg_match_all('/[\p{L}\p{N}][^\p{L}\p{N}]/ui', $fragment, $matches, PREG_OFFSET_CAPTURE))
if ($end < $text_length - 1 && preg_match_all('/[\p{L}\p{N}][^\p{L}\p{N}]/u', $fragment, $matches, PREG_OFFSET_CAPTURE))
{
$fragment_end = end($matches[0])[1]; // last valid alphanumeric character
$fragment_end = mb_strlen(substr($fragment, 0, end($matches[0])[1]));
}
$output[] = utf8_substr($fragment, $fragment_start, $fragment_end - $fragment_start + 1);
}
return ($fragments[0][0] !== 0 ? '... ' : '') . htmlspecialchars(implode(' ... ', $output)) . ($end < $text_length - 1 ? ' ...' : '');
return ($fragments[0][0] !== 0 ? '... ' : '') . utf8_htmlspecialchars(implode(' ... ', $output)) . ($end < $text_length - 1 ? ' ...' : '');
}
/**

2
tests/functions_content/get_context_test.php
View file

@ -117,7 +117,7 @@ class phpbb_functions_content_get_context_test extends TestCase
'text' => 'Manzana,plátano,naranja,fresa,mango,uva,piña,pera,kiwi,cereza,sandía,melón,papaya,arándano,durazno',
'words' => ['piña'],
'length' => 20,
'expected' => '... uva,piña,pera, ...',
'expected' => '... uva,piña,pera ...',
]
];
}