#38000 [Opn->Bgs]: preg_match fails if strlen >= 1630

ID: 38000 Updated by: nlopess@php.net Reported By: dave at smartboy dot com -Status: Open +Status: Bogus Bug Type: PCRE related Operating System: Windows XP PHP Version: 5.1.4 New Comment: Well what I can reproduce is a pcre backtracking and recursion limit. This is because your regex is particularly difficult to run (you may note that it will need to backtrack a lot!). The other problem is that the bundled pcre library was upgraded in PHP 5.1.3. And that version has some 'S'tudy features disabled, and you are beeing hited by that. PCRE was previously optimizing your complex regex but now it isn't. We'll have to wait for the next release of PCRE. In the mean time, you may want to upgrade to php 5.2, where you can tweak the backtrack and recursion limits on run-time (anyway you can always tweak them on compile time on any version). Previous Comments: ------------------------------------------------------------------------ [2006-07-04 21:33:55] dave at smartboy dot com $str = str_repeat('a', 1000000); //works in PHP 5.1.2 ------------------------------------------------------------------------ [2006-07-04 21:23:49] dave at smartboy dot com It turns out that no test file is needed $str = str_repeat('a', 1575); //works $str = str_repeat('a', 1576); //fails There are really two issues: - PCRE not working for "long" strings (although 1576 bytes is not really a "long" string). This greatly limits the usefulness of regexp pattern matching. - When PCRE fails the error message is VERY misleading. Surely an E_NOTICE should be issued by preg_match() if match has failed due to out of memory, etc.? "Could not open input file" is just plain wrong. ------------------------------------------------------------------------ [2006-07-04 14:40:01] nlopess@php.net please provide the zzz file. (either post a link to it or send it to my e-mail). ------------------------------------------------------------------------ [2006-07-04 00:46:02] dave at smartboy dot com Description: ------------ preg_match() with regexp to test for valid UTF-8 sequence - fails and causes the error message: 'Could not open input file: <name of script>.php' IF the subject string passed to preg_match() is longer than 1629 characters. (Or in this case the size of the file 'zzz' which contains ASCII) There was no such limitation in preg_match() in the previous version of PHP (5.1.2) Reproduce code: --------------- $str = file_get_contents('zzz'); echo "Loaded file...\n"; $result = preg_match('/^([\x00-\x7f]|[\xc2-\xdf][\x80-\xbf]|\xe0[\xa0-\xbf][\x80-\xbf]|' . '[\xe1-\xec][\x80-\xbf]{2}|\xed[\x80-\x9f][\x80-\xbf]|[\xee-\xef][\x80-\xbf]{2}|' . '\xf0[\x90-\xbf][\x80-\xbf]{2}|[\xf1-\xf3][\x80-\xbf]{3}|\xf4[\x80-\x8f][\x80-\xbf]{2})*$/S', $str) === 1; echo "Back from preg_match()\n"; var_export($result); echo "\n"; Expected result: ---------------- Loaded file... Back from preg_match() true Actual result: -------------- Loaded file... Could not open input file: u8.php ------------------------------------------------------------------------ -- Edit this bug report at http://bugs.php.net/?id=38000&edit=1