ID: 51050 Updated by: degeb...@php.net Reported By: pecoes at gmail dot com Status: Bogus Bug Type: Filter related Operating System: WinXP PHP Version: 5.3.1 New Comment:
It is hardly within the scope of the manual to cover all possible attack vectors. http://php.net/security.variables already states a number of things you should always consider when dealing with user submitted data. Whether or not something is "safe" is entirely circumstantial and highly dependent on how trustworthy the user is. It's not the manual's job to figure that out, but your job as a programmer. Previous Comments: ------------------------------------------------------------------------ [2010-02-15 07:58:17] pecoes at gmail dot com Yes. You're perfectly right. Forget that last suggestion. But I still think, it would be helpful to clearly state in the documentation, that valid != safe. ------------------------------------------------------------------------ [2010-02-15 07:52:06] ras...@php.net But that wouldn't make it safe and as such would be worse than the current state because people would think it was. It takes more than just filtering out or encoding certain characters to make a user- supplied URL safe for direct display. Generally you need to apply some application-level logic to determine which domains and which primitives are valid. For example: javascript:alert(1) is a perfectly valid URL that has no special characters in it. Or file:/// can also cause problems. Even http://localhost/ can cause issues. There is a long list of things that can be problematic if you allow users to supply entire URLs. Usually all you want to accept from users are url fragments and you prepend the primitive and base domain and path. ------------------------------------------------------------------------ [2010-02-15 07:39:45] pecoes at gmail dot com Personally, I don't like sanitizing. I prefer to either accept or reject. I don't think modifying a user's input is a good idea. How about adding an optional FILTER_FLAG_ALLOW_SPECIAL_CHARS to FILTER_VALIDATE_URL? Or an optional FILTER_FLAG_DISALLOW_SPECIAL_CHARS, if that's what you prefer... Because, you know, using URLs in an HTML context is not such an exotic scenario. :0) ------------------------------------------------------------------------ [2010-02-15 07:26:07] ras...@php.net What you are after is a filter for the html-context. There is nothing wrong with your URL. You only have an issue with it if you use it in an HTML context. It is your target context you should be filtering for. The URL sanitizer is very explicitly documented as: Remove all characters except letters, digits and $- _.+!*'(),{}|\\^~[]`<>#%";/?:@&=. Have a look through: http://php.net/manual/en/filter.filters.sanitize.php What you are looking for is FILTER_SANITIZE_SPECIAL_CHARS ------------------------------------------------------------------------ [2010-02-15 07:20:03] pecoes at gmail dot com Okay, I accept that the URL I posted, is valid. But it deserves pointing out that valid does NOT mean safe. Btw FILTER_SANITIZE_URL has the same effect on this URL. ------------------------------------------------------------------------ The remainder of the comments for this report are too long. To view the rest of the comments, please view the bug report online at http://bugs.php.net/51050 -- Edit this bug report at http://bugs.php.net/?id=51050&edit=1