note 57824 added to function.eregi

There seems to be a lot of confusion about escaping special characters within a bracket expression enclosed in [] The reference is the regex man page at http://www.tin.org/bin/man.cgi?section=7&topic=regex , Linked from the "Regular Expression Functions (POSIX Extended)" page in the PHP manual at http://php.planetmirror.com/manual/en/ref.regex.php The simple rule is there should be no backslash `\' in a bracket expression unless you are matching it literally. ALL special characters (the few exceptions are listed below), lose their special significance in a bracket expression. This includes .[$()|*+?{\ all of which have special meaning outside of a bracket expression, inside they will match LITERALLY! The backslash character can never escape anything in a bracket expression including itself! It is always interpreted literally. Here are the exceptions ^ at the start of the bracket expression will match on all but the set of characters within the bracket expression - (hyphen) Indicates range, so [a-z] will match a single character in the range a-z (case insensitive for eregi). To match a literal `-' it must be at the start or end of the bracket expression or the end of a range, or enclosed in [. .] to be the start of a range. Examples: [0-9-] [-0-9] Both the above will match a single character from the set comprising the digits and hyphen `-' [!--] is the same as [!"#$%&'()*+,-] i.e. it matches a single character from the range of characters from !' to hyphen -' [[.-.]-9] is the same as [-./0123456789] or [./0123456789-] and will match a single digit from the range of characters from hyphen `-' to digit 9 ] The closing bracket indicates the end of a bracket expression but not if it immediately follows the opening bracket [' or [^' in that case the closing bracket is matched literally. Examples: [][0-9] is the same as []0-9[] and matches a single character from the set of digits and a literal closing bracket ]' and opening bracket ]' [^][0-9] is the same as [^]0-9[] and matches a single character from the group of all characters except the set of digits and literal ]' or [' Look at the regex man page for more on the above and the following special cases [.characters.] A collating element [=characters=] An equivalence class [:alnum:] A character class That's it for special cases! ALL other characters match literally within a bracket expression and can never be escaped with `\' Here is a list of some of the postings here that have mistakenly tried to use special characters within a bracket expression. 28-Jan-2002 12:56, 09-Feb-2002 06:28, 24-Sep-2002 01:21, 08-Mar-2003 02:21, 11-Apr-2003 03:22 (uses |' as if it was or' it is not, it is a literal `|'), 27-Jun-2003 04:03, 06-Sep-2003 07:17, 20-Apr-2004 04:55, 24-Nov-2004 09:12, 25-Nov-2004 02:11, 06-Jan-2005 08:54, 24-Feb-2005 06:05 It seems people are copying each others mistakes so this needs to be cleared up. Most of the mistakes attempt to escape special characters within a bracket expression with backslash. That never works! A simple fix is to simply remove any backslashes within the bracket expression that are not meant to be matched literally. The attempt to use | as 'or' within a bracketed expression should be put in parentheses instead. I used a simple script attached to a form to test these results. Here's the script. <?php if(isset($_POST['pattern']) && isset($_POST['teststring'])) { $pattern = stripslashes($_POST['pattern']); $teststring = stripslashes($_POST['teststring']); $match = ereg($pattern, $teststring, $regs); if(!$match) $match = 0; echo('<p>The regular expression is: '.$pattern.'</p>'); echo('<p>The string for testing is: '.$teststring.'</p>'); echo('<p>The number of matching characters is: '.$match.'</p>'); } ?> Just use a form with input fields named "pattern" and "teststring" Hope this helps and clears up some misconceptions Jason Smart Monash University Student ---- Manual Page -- http://www.php.net/manual/en/function.eregi.php Edit -- http://master.php.net/manage/user-notes.php?action=edit+57824 Delete: added to the manual -- http://master.php.net/manage/user-notes.php?action=delete+57824&report=yes&reason=added+to+the+manual Delete: bad code -- http://master.php.net/manage/user-notes.php?action=delete+57824&report=yes&reason=bad+code Delete: spam -- http://master.php.net/manage/user-notes.php?action=delete+57824&report=yes&reason=spam Delete: useless -- http://master.php.net/manage/user-notes.php?action=delete+57824&report=yes&reason=useless Delete: non-english -- http://master.php.net/manage/user-notes.php?action=delete+57824&report=yes&reason=non-english Delete: already in docs -- http://master.php.net/manage/user-notes.php?action=delete+57824&report=yes&reason=already+in+docs Delete: other reasons -- http://master.php.net/manage/user-notes.php?action=delete+57824&report=yes Reject -- http://master.php.net/manage/user-notes.php?action=reject+57824&report=yes Search -- http://master.php.net/manage/user-notes.php