php - Regex: How to validate that a path does not contain- Stack Overflow

I need help to create a regex (for JavaScript .match and PHP preg_match) that validates a unix type absolute path to a file (with international characters such as åäöøæð and so on) so that:

1. /path/to/someWhere is valid
2. /path/tø/sömewhere is valid
3. /path/to//somewhere is invalid
4. path/to/somewhere is invalid
5. /path/to/somewhere/ is invalid

The regex needs to handle paths regardless of their depth (/path/to or /path/to/somewhere or /path/to/somewhere/else)

I have a regexp that marks 1 to 3 as valid /^\/.+[^\/]$/ , the problem is to make this regex not to mark 3 as valid as it contains // without any other character in between.

I need help to create a regex (for JavaScript .match and PHP preg_match) that validates a unix type absolute path to a file (with international characters such as åäöøæð and so on) so that:

1. /path/to/someWhere is valid
2. /path/tø/sömewhere is valid
3. /path/to//somewhere is invalid
4. path/to/somewhere is invalid
5. /path/to/somewhere/ is invalid

The regex needs to handle paths regardless of their depth (/path/to or /path/to/somewhere or /path/to/somewhere/else)

I have a regexp that marks 1 to 3 as valid /^\/.+[^\/]$/ , the problem is to make this regex not to mark 3 as valid as it contains // without any other character in between.

• php
• javascript
• regex
• path

Share Improve this question Follow asked Sep 28, 2010 at 13:45 tirithentirithen 3,5271111 gold badges4444 silver badges6767 bronze badges 2

• I love regex posts like this, to an outsider we all look like spambots or something. – Incognito Commented Sep 28, 2010 at 13:52
• Unix/Linux pathnames actually work just fine if they end in a slash or have double-slashes. – Pointy Commented Sep 28, 2010 at 14:30

Add a ment  | 

6 Answers 6

Sorted by: Reset to default 4

Regex isn't really needed here. As far as I can see, there are three things you want to ensure:

1. The string starts with /
2. The string doesn't end with /, unless the whole string is /
3. The string doesn't contain any instances of //

All three of the above can be done with string functions.

In PHP:

if ($string != '/' && ($string[0] != '/' || $string[strlen($string)-1] == '/' || strpos($string, '//') > -1))
{
  // string is invalid
}

In Javascript:

if (string != '/' && (string.charAt(0) != '/' || string.charAt(string.length - 1) == '/' || string.indexOf('//') > -1))
{
  // string is invalid
}

Resources:

• PHP's strpos function
• Javascript's String.charAt and String.indexOf functions

A Solution for PHP:

    $lines =  array(
        "/path/to/someWhere",
        "/path/tø/sömewhere",
        "/path/to//somewhere",
        "path/to/somewhere",
        "/path/to/somewhere/",
    );

    foreach($lines as $line){
        var_dump(preg_match('#^(/[^/]+)+$#',$line)); // dumps int(1) int(1) int(0) int(0) int(0) 
    }

I think this will do it:

^(:?\/$|(:?\/[^/]+)+$)

That says to accept any string that's either just a /, or any string formed from a sequence of one or more repetitions of a / followed by one or more non-/ characters.

This uses all greedy quantifiers so it should be fast; also, for performance, the ^ anchor is factored out.

That's a Javascript regex. I'm not a PHP programmer so the main thing I don't know is whether the non-capturing group syntax works in PHP. Also, I'm not sure how you'd handle "quoting" the slash characters.

This should work:

^/[^/]?$|^/[^/]([^/]|/[^/])*?[^/]$

It allows any character except /, or a / followed by any character except /. It also makes sure that the last character isn’t a /, and that the second character isn’t one either.

Finally, this uses / without escaping. To use it in PHP, don’t use / as the regex delimiter – this just makes the regular expression hard to read. Use any other character, e.g. ; to delimit the expression instead:

;^/[^/]?$|^/[^/]([^/]|/[^/])*?[^/]$;

EDIT: Added special handing for the root path, "/", and paths that consist of a single letter directory.

If the path matches ^[^\/]|\/\/|.\/$, it is invalid. Otherwise it is valid.

it's not regex, but works just as well.

str_replace('//', '/', $file)