diff options
| -rw-r--r-- | _test/cases/inc/utf8_substr.test.php | 13 | ||||
| -rw-r--r-- | inc/fulltext.php | 5 |
2 files changed, 15 insertions, 3 deletions
diff --git a/_test/cases/inc/utf8_substr.test.php b/_test/cases/inc/utf8_substr.test.php index 92bc65fe3..5cb93034a 100644 --- a/_test/cases/inc/utf8_substr.test.php +++ b/_test/cases/inc/utf8_substr.test.php @@ -26,5 +26,18 @@ class utf8_substr_test extends UnitTestCase { } } + function test2_bug891() { + // we test multiple cases here - format: in, offset, length, out + $tests = array(); + + $str = str_pad('',strlen('в')*66000,'в').'@@'; + $tests[] = array($str, 65600, 1, 'в'); + $tests[] = array($str,0,66000,$str); + + foreach($tests as $test){ + $this->assertEqual(utf8_substr($test[0],$test[1],$test[2]),$test[3]); + } + } + } //Setup VIM: ex: et ts=4 enc=utf-8 : diff --git a/inc/fulltext.php b/inc/fulltext.php index fa3ec05d2..280ba0c89 100644 --- a/inc/fulltext.php +++ b/inc/fulltext.php @@ -333,7 +333,6 @@ switch ($algorithm) { // establish context, 100 bytes surrounding the match string // first look to see if we can go 100 either side, // then drop to 50 adding any excess if the other side can't go to 50, - // NOTE: these are byte adjustments and will have to be corrected for utf-8 $pre = min($utf8_idx-$utf8_offset,100); $post = min($len-$utf8_idx-$utf8_len,100); @@ -371,9 +370,9 @@ switch ($algorithm) { } else { // code for strings too large for utf8_substr // use a larger context number as its bytes not characters - $pre = 70; + // no need to check for short pre, $idx is nearly 64k $post = min(strlen($text)-$idx-strlen($str), 70); - if ($post < 70) { $pre = 70 - $post; } + $pre = ($post < 70) ? 140 - $post : 70; $start = utf8_correctIdx($text,$idx - $pre); $end = utf8_correctIdx($text, $idx + strlen($str) + $post); |