Escaping variables

问题:

I've read that it's enough and even recommended to escape characters on the output, not on the input.

It could be easily applied to all get variables as they are not injected to the database from the form level.

However, I'm not sure what to do with all post variables. If it doesn't come from the database, so if it's a raw input data, escaping is fully needed. But I'm using PDO prepare/execute to escape all variables. Questions now:

1. Is it OK to use PDO's prepare/execute both in select and insert statements? Isn't it escaping variables twice ?
2. Let's say I get some variable through PDO exeute statement - is it ok to display this variable simply with $_POST['variable'] without escaping it (if it was already done in PDO function) ?
3. Is htmlspecialchars() enough to escape things such as GET variables which don't come from the database?

And the most important - is all of this, PDO prepare/execute, and htmlspecialchars(), enough to prevent all XSS attacks? Or should I also do more? If so, what should this be? Removing all html tags from the input? Using BB-Code instead?

回答1:

I've read that it's enough and even recommended to escape characters on the output, not on the input.

Typically, you want to:

• Validate input and store it using prepared statements. Prepared statements will protect your database against SQL injections. Typically, you don't want to strip out HTML tags on input because doing so could lead to a loss of data integrity.
• When displaying user-generated data (output), you can guard against XSS by using a combination of htmlentities and mb_convert_encoding.

Note on the htmlspecialchars function from another question:

Even if you use htmlspecialchars($string) outside of HTML tags, you are still vulnerable to multi-byte charset attack vectors.

The most effective you can be is to use the a combination of mb_convert_encoding and htmlentities as follows.

$str = mb_convert_encoding($str, ‘UTF-8′, ‘UTF-8′); $str = htmlentities($str, ENT_QUOTES, ‘UTF-8′); 

回答2:

escape characters on the output, not on the input

Yes.

easily applied to all get variables

But $_GET is by definition input

Isn't it escaping variables twice ?

No - by escaping the content you're just insulating it from mis-interpretation by the processing agent. The database doesn't store the escaped data, it stores the original data.

Hence if start with

O'Reilly 

Then escape to splice it into a SQL string....

O\'Reilly 

Then the value stored in the database, and retrieved by a SELECT statement is

O'Reilly 

And when you want to output it your HTML, then you pass it though htmlspecialchars() to get

O"Reilly 

You use an appropriate method for escaping the data depending on where it's going - hence you use mysql_real_escape() or paramter binding or similar when putting stuff INTO your database, and htmlspecialchars() when putting stuff INTO html

回答3:

1. whenever data is coming from user, sanitize it(take special attention if its storing in database.). So PDO with prepared statement is a must. What else you do is added bonus.

2. Yes (opinions will differ here from person to person) for preventing sql injection (assuming you are using prepared statement). though I prefer storing the raw data in database even if it means sacrificing for some malicious XSS code may contain it. While outputting, take utmost care.

3. No. use htmlpurifier (with a view that you are outputting from a database.)

回答4:

I use mysqli_real_escape_string and preg_replace

$email = mysqli_real_escape_string($dbc, trim($_POST['email'])); $password = mysqli_real_escape_string($dbc, trim($_POST['password'])); $domain = preg_replace('/^[a-zA-Z0-9][a-zA-Z0-9\._\-&!?=#]*@/', '', $email); 

Also, here is a link to a similar post regarding PDO escaping Escape arguments for PDO statements?

文章来源: Escaping variables