Home > Articles

Error Handling

This chapter is from the book

This chapter is from the book

Exceptions

The methods covered to this point are all that was available before PHP5, and you can see that this poses some critical problems, especially when you are writing larger applications. The primary flaw is in returning errors to a user of a library. Consider the error checking that you just implemented in the passwd file reading function.

When you were building that example, you had two basic choices on how to handle a connection error:

  • Handle the error locally and return invalid data (such as false) back to the caller.

  • Propagate and preserve the error and return it to the caller instead of returning the result set.

In the passwd file reading function example, you did not select the first option because it would have been presumptuous for a library to know how the application wants it to handle the error. For example, if you are writing a database-testing suite, you might want to propagate the error in high granularity back to the top-level caller; on the other hand, in a Web application, you might want to return the user to an error page.

The preceding example uses the second method, but it is not much better than the first option. The problem with it is that it takes a significant amount of foresight and planning to make sure errors can always be correctly propagated through an application. If the result of a database query is a string, for example, how do you differentiate between that and an error string?

Further, propagation needs to be done manually: At every step, the error must be manually bubbled up to the caller, recognized as an error, and either passed along or handled. You saw in the last section just how difficult it is to handle this.

Exceptions are designed to handle this sort of situation. An exception is a flow-control structure that allows you to stop the current path of execution of a script and unwind the stack to a prescribed point. The error that you experienced is represented by an object that is set as the exception.

Exceptions are objects. To help with basic exceptions, PHP has a built-in Exception class that is designed specifically for exceptions. Although it is not necessary for exceptions to be instances of the Exception class, there are some benefits of having any class that you want to throw exceptions derive from Exception, which we'll discuss in a moment. To create a new exception, you instantiate an instance of the Exception class you want and you throw it.

When an exception is thrown, the Exception object is saved, and execution in the current block of code halts immediately. If there is an exception-handler block set in the current scope, the code jumps to that location and executes the handler. If there is no handler set in the current scope, the execution stack is popped, and the caller's scope is checked for an exception-handler block. This repeats until a handler is found or the main, or top, scope is reached.

Running this code:

<?php
  throw new Exception;
?>

returns the following:

> php uncaught-exception.php

Fatal error: Uncaught exception 'exception'! in Unknown on line 0

An uncaught exception is a fatal error. Thus, exceptions introduce their own maintenance requirements. If exceptions are used as warnings or possibly nonfatal errors in a script, every caller of that block of code must know that an exception may be thrown and must be prepared to handle it.

Exception handling consists of a block of statements you want to try and a second block that you want to enter if and when you trigger any errors there. Here is a simple example that shows an exception being thrown and caught:

try {
  throw new Exception;
  print "This code is unreached\n";
}
catch (Exception $e) {
  print "Exception caught\n";
}

In this case you throw an exception, but it is in a try block, so execution is halted and you jump ahead to the catch block. catch catches an Exception class (which is the class being thrown), so that block is entered. catch is normally used to perform any cleanup that might be necessary from the failure that occurred.

I mentioned earlier that it is not necessary to throw an instance of the Exception class. Here is an example that throws something other than an Exception class:

<?php

class AltException {}

try {
    throw new AltException;
}
catch (Exception $e) {
    print "Caught exception\n";
}
?>

Running this example returns the following:

> php failed_catch.php
Fatal error: Uncaught exception 'altexception'! in Unknown on line 0

This example failed to catch the exception because it threw an object of class AltException but was only looking to catch an object of class Exception.

Here is a less trivial example of how you might use a simple exception to facilitate error handling in your old favorite, the factorial function. The simple factorial function is valid only for natural numbers (integers > 0). You can incorporate this input checking into the application by throwing an exception if incorrect data is passed:

<?php
// factorial.inc
// A simple Factorial Function
function factorial($n) {
  if(!preg_match('/^\d+$/',$n) || $n < 0 ) {
    throw new Exception;
  } else if ($n == 0 || $n == 1) {
    return $n;
  }
  else {
    return $n * factorial($n – 1);
  }
}
?>

Incorporating sound input checking on functions is a key tenant of defensive programming.

Why the regex?

It might seem strange to choose to evaluate whether $n is an integer by using a regular expression instead of the is_int function. The is_int function, however, does not do what you want. It only evaluates whether $n has been typed as a string or as integer, not whether the value of $n is an integer. This is a nuance that will catch you if you use is_int to validate form data (among other things). We will explore dynamic typing in PHP in Chapter 20, "PHP and Zend Engine Internals."

When you call factorial, you need to make sure that you execute it in a try block if you do not want to risk having the application die if bad data is passed in:

<html>
<form method="POST">
Compute the factorial of 
<input type="text" name="input" value="<?= $_POST['input'] ?>"><br>
<?php
include "factorial.inc";
if($_POST['input']) {
  try {
    $input = $_POST['input'];
    $output = factorial($input);
    echo "$_POST[input]! = $output";
  }
  catch (Exception $e) {
    echo "Only natural numbers can have their factorial computed.";
  }
}
?>
<br>
<input type=submit name=posted value="Submit">
</form>

Using Exception Hierarchies

You can have try use multiple catch blocks if you want to handle different errors differently. For example, we can modify the factorial example to also handle the case where $n is too large for PHP's math facilities:

class OverflowException {}
class NaNException {}
function factorial($n)
{
  if(!preg_match('/^\d+$/', $n) || $n < 0 ) {
    throw new NaNException;
  }
  else if ($n == 0 || $n == 1) {
    return $n;
  }
  else if ($n > 170 ) {
    throw new OverflowException;
  }
  else {
    return $n * factorial($n - 1);
  }
}

Now you handle each error case differently:

<?php
if($_POST['input']) {
  try {
    $input = $_POST['input'];
    $output = factorial($input);
    echo "$_POST[input]! = $output";
  }
  catch (OverflowException $e) {
    echo "The requested value is too large.";
  }
  catch (NaNException $e) {
    echo "Only natural numbers can have their factorial computed.";
  }
}
?>

As it stands, you now have to enumerate each of the possible cases separately. This is both cumbersome to write and potentially dangerous because, as the libraries grow, the set of possible exceptions will grow as well, making it ever easier to accidentally omit one.

To handle this, you can group the exceptions together in families and create an inheritance tree to associate them:

class MathException extends Exception {}
class NaNException extends MathException {}
class OverflowException extends MathException {}

You could now restructure the catch blocks as follows:

<?php
if($_POST['input']) {
  try {
    $input = $_POST['input'];
    $output = factorial($input);
    echo "$_POST[input]! = $output";
  }
  catch (OverflowException $e) {
    echo "The requested value is too large.";
  }
  catch (MathException $e) {
    echo "A generic math error occurred";
  }
  catch (Exception $e) {
    echo "An unknown error occurred";
  }
}
?>

In this case, if an OverflowException error is thrown, it will be caught by the first catch block. If any other descendant of MathException (for example, NaNException) is thrown, it will be caught by the second catch block. Finally, any descendant of Exception not covered by any of the previous cases will be caught.

This is the benefit of having all exceptions inherit from Exception: It is possible to write a generic catch block that will handle all exceptions without having to enumerate them individually. Catchall exception handlers are important because they allow you to recover from even the errors you didn't anticipate.

A Typed Exceptions Example

So far in this chapter, all the exceptions have been (to our knowledge, at least) attribute free. If you only need to identify the type of exception thrown and if you have been careful in setting up our hierarchy, this will satisfy most of your needs. Of course, if the only information you would ever be interested in passing up in an exception were strings, exceptions would have been implemented using strings instead of full objects. However, you would like to be able to include arbitrary information that might be useful to the caller that will catch the exception.

The base exception class itself is actually deeper than indicated thus far. It is a built-in class, meaning that it is implemented in C instead of PHP. It basically looks like this:

class Exception {
  Public function _ _construct($message=false, $code=false) {
    $this->file = _ _FILE_ _;
    $this->line = _ _LINE_ _;
    $this->message = $message; // the error message as a string
    $this->code = $code; // a place to stick a numeric error code
  }
  public function getFile() {
    return $this->file;
  }
  public function getLine() {
    return $this->line;
  }
  public function getMessage() {
    return $this->message;
  }
  public function getCode() {
    return $this->code;
  }
}

Tracking _ _FILE_ _ and _ _LINE_ _ for the last caller is often useless information. Imagine that you decide to throw an exception if you have a problem with a query in the DB_Mysql wrapper library:

class DB_Mysql {
  // ...
  public function execute($query) {
   if(!$this->dbh) {
    $this->connect();
   }
   $ret = mysql_query($query, $this->dbh);
   if(!is_resource($ret)) {
    throw new Exception;
   }
   return new MysqlStatement($ret);
  }
}

Now if you trigger this exception in the code by executing a syntactically invalid query, like this:

<?php
    require_once "DB.inc";
    try {
    $dbh = new DB_Mysql_Test;
    // ... execute a number of queries on our database connection
    $rows = $dbh->execute("SELECT * FROM")->fetchall_assoc();
    }
    catch (Exception $e) {
        print_r($e);
    }
?>

you get this:

exception Object
(
  [file] => /Users/george/Advanced PHP/examples/chapter-3/DB.inc
  [line] => 42
)

Line 42 of DB.inc is the execute() statement itself! If you executed a number of queries within the try block, you would have no insight yet into which one of them caused the error. It gets worse, though: If you use your own exception class and manually set $file and $line (or call parent::_ _construct to run Exception's constructor), you would actually end up with the first callers _ _FILE_ _ and _ _LINE_ _ being the constructor itself! What you want instead is a full backtrace from the moment the problem occurred.

You can now start to convert the DB wrapper libraries to use exceptions. In addition to populating the backtrace data, you can also make a best-effort attempt to set the message and code attributes with the MySQL error information:

class MysqlException extends Exception {
 public $backtrace;
 public function _ _construct($message=false, $code=false) {
  if(!$message) {
   $this->message = mysql_error();
  }
  if(!$code) {
   $this->code = mysql_errno();
  }
  $this->backtrace = debug_backtrace();
 }
}

If you now change the library to use this exception type:

class DB_Mysql {
 public function execute($query) {
  if(!$this->dbh) {
   $this->connect();
  }
  $ret = mysql_query($query, $this->dbh);
  if(!is_resource($ret)) {
   throw new MysqlException;
  }
  return new MysqlStatement($ret);
 }
}

and repeat the test:

<?php
    require_once "DB.inc";
    try {
    $dbh = new DB_Mysql_Test;
    // ... execute a number of queries on our database connection
    $rows = $dbh->execute("SELECT * FROM")->fetchall_assoc();
    }
    catch (Exception $e) {
        print_r($e);
    }
?>

you get this:

mysqlexception Object
(
 [backtrace] => Array
  (
   [0] => Array
    (
     [file] => /Users/george/Advanced PHP/examples/chapter-3/DB.inc
     [line] => 45
     [function] => _ _construct
     [class] => mysqlexception
     [type] => ->
     [args] => Array
      (
      )
     )
   [1] => Array
    (
     [file] => /Users/george/Advanced PHP/examples/chapter-3/test.php
     [line] => 5
     [function] => execute
     [class] => mysql_test
     [type] => ->
     [args] => Array
      (
       [0] => SELECT * FROM
      )
    )
  )

 [message] => You have an error in your SQL syntax near '' at line 1
 [code] => 1064
)

Compared with the previous exception, this one contains a cornucopia of information:

  • Where the error occurred

  • How the application got to that point

  • The MySQL details for the error

You can now convert the entire library to use this new exception:

class MysqlException extends Exception { 
 public $backtrace;
 public function _ _construct($message=false, $code=false) {
  if(!$message) {
   $this->message = mysql_error();
  }
  if(!$code) {
   $this->code = mysql_errno();
  }
  $this->backtrace = debug_backtrace();
 }
}
class DB_Mysql {
 protected $user;
 protected $pass;
 protected $dbhost;
 protected $dbname;
 protected $dbh;

 public function _ _construct($user, $pass, $dbhost, $dbname) {
  $this->user = $user;
  $this->pass = $pass;
  $this->dbhost = $dbhost;
  $this->dbname = $dbname;
 }
 protected function connect() {
  $this->dbh = mysql_pconnect($this->dbhost, $this->user, $this->pass);
  if(!is_resource($this->dbh)) {
   throw new MysqlException;
  }
  if(!mysql_select_db($this->dbname, $this->dbh)) {
   throw new MysqlException;
  }
 }
 public function execute($query) {
  if(!$this->dbh) {
   $this->connect();
  }
  $ret = mysql_query($query, $this->dbh); 
  if(!$ret) {
   throw new MysqlException;
  }
  else if(!is_resource($ret)) {
   return TRUE;
  } else {
   return new DB_MysqlStatement($ret);
  }
 }
 public function prepare($query) {
  if(!$this->dbh) {
   $this->connect();
  }
  return new DB_MysqlStatement($this->dbh, $query);
 }
}
class DB_MysqlStatement {
 protected $result;
 protected $binds;
 public $query;
 protected $dbh;
 public function _ _construct($dbh, $query) {
  $this->query = $query;
  $this->dbh = $dbh;
  if(!is_resource($dbh)) {
   throw new MysqlException("Not a valid database connection");
  }
 }
 public function bind_param($ph, $pv) {
  $this->binds[$ph] = $pv;
 }
 public function execute() {
  $binds = func_get_args();
  foreach($binds as $index => $name) {
   $this->binds[$index + 1] = $name;
  }
  $cnt = count($binds);
  $query = $this->query;
  foreach ($this->binds as $ph => $pv) {
   $query = str_replace(":$ph", "'".mysql_escape_string($pv)."'", $query);
  }
  $this->result = mysql_query($query, $this->dbh);
  if(!$this->result) {
   throw new MysqlException;
  }
 }
 public function fetch_row() {
  if(!$this->result) {
   throw new MysqlException("Query not executed");
  }
  return mysql_fetch_row($this->result);
 }
 public function fetch_assoc() {
  return mysql_fetch_assoc($this->result);
 }
 public function fetchall_assoc() {
  $retval = array();
  while($row = $this->fetch_assoc()) {
   $retval[] = $row;
  }
  return $retval;
 }
}

? >

Cascading Exceptions

Sometimes you might want to handle an error but still pass it along to further error handlers. You can do this by throwing a new exception in the catch block:

<?php
try {
    throw new Exception;
}
catch (Exception $e) {
    print "Exception caught, and rethrown\n";
    throw new Exception;
}
?>

The catch block catches the exception, prints its message, and then throws a new exception. In the preceding example, there is no catch block to handle this new exception, so it goes uncaught. Observe what happens as you run the code:

> php re-throw.php
Exception caught, and rethrown

Fatal error: Uncaught exception 'exception'! in Unknown on line 0

In fact, creating a new exception is not necessary. If you want, you can rethrow the current Exception object, with identical results:

<?php
try {
    throw new Exception;
}
catch (Exception $e) {
    print "Exception caught, and rethrown\n";
    throw $e;
}
?>

Being able to rethrow an exception is important because you might not be certain that you want to handle an exception when you catch it. For example, say you want to track referrals on your Web site. To do this, you have a table:

CREATE TABLE track_referrers (
  url varchar2(128) not null primary key,
  counter int
);

The first time a URL is referred from, you need to execute this:

INSERT INTO track_referrers VALUES('http://some.url/', 1)

On subsequent requests, you need to execute this:

UPDATE track_referrers SET counter=counter+1 where url = 'http://some.url/'

You could first select from the table to determine whether the URL's row exists and choose the appropriate query based on that. This logic contains a race condition though: If two referrals from the same URL are processed by two different processes simultaneously, it is possible for one of the inserts to fail.

A cleaner solution is to blindly perform the insert and call update if the insert failed and produced a unique key violation. You can then catch all MysqlException errors and perform the update where indicated:

<?php
include "DB.inc";

function track_referrer($url) { 
    $insertq = "INSERT INTO referrers (url, count) VALUES(:1, :2)";
    $updateq = "UPDATE referrers SET count=count+1 WHERE url = :1";
    $dbh = new DB_Mysql_Test;
    try {
        $sth = $dbh->prepare($insertq);
        $sth->execute($url, 1);
    }
    catch (MysqlException $e) {
        if($e->getCode == 1062) {
            $dbh->prepare($updateq)->execute($url);
        }
        else {
            throw $e;
        }
    }
}
?> 

Alternatively, you can use a purely typed exception solution where execute itself throws different exceptions based on the errors it incurs:

class Mysql_Dup_Val_On_Index extends MysqlException {}
//...
class DB_Mysql {
 // ...
 public function execute($query) {
  if(!$this->dbh) {
   $this->connect();
  }
  $ret = mysql_query($query, $this->dbh); 
  if(!$ret) {
   if(mysql_errno() == 1062) {
    throw new Mysql_Dup_Val_On_Index;
   else {
    throw new MysqlException;
   }
  }
  else if(!is_resource($ret)) {
    return TRUE;
  } else {
   return new MysqlStatement($ret);
  }
 }
}

Then you can perform your checking, as follows:

function track_referrer($url) { 
 $insertq = "INSERT INTO referrers (url, count) VALUES('$url', 1)";
 $updateq = "UPDATE referrers SET count=count+1 WHERE url = '$url'";
 $dbh = new DB_Mysql_Test;
 try {
  $sth = $dbh->execute($insertq);
 }
 catch (Mysql_Dup_Val_On_Index $e) {
  $dbh->execute($updateq);
 }
}

Both methods are valid; it's largely a matter of taste and style. If you go the path of typed exceptions, you can gain some flexibility by using a factory pattern to generate your errors, as in this example:

class MysqlException { 
 // ...
 static function createError($message=false, $code=false) {
   if(!$code) {
    $code = mysql_errno();
   }
   if(!$message) {
    $message = mysql_error();
   }
   switch($code) {
    case 1062:
     return new Mysql_Dup_Val_On_Index($message, $code);
     break;
    default:
     return new MysqlException($message, $code);
     break;
   }
  }
}

There is the additional benefit of increased readability. Instead of a cryptic constant being thrown, you get a suggestive class name. The value of readability aids should not be underestimated.

Now instead of throwing specific errors in your code, you just call this:

throw MysqlException::createError();

Handling Constructor Failure

Handling constructor failure in an object is a difficult business. A class constructor in PHP must return an instance of that class, so the options are limited:

  • You can use an initialized attribute in the object to mark it as correctly initialized.

  • You can perform no initialization in the constructor.

  • You can throw an exception in the constructor.

The first option is very inelegant, and we won't even consider it seriously. The second option is a pretty common way of handling constructors that might fail. In fact, in PHP4, it is the preferable way of handling this.

To implement that, you would do something like this:

class ResourceClass {
 protected $resource;
 public function _ _construct() {
  // set username, password, etc
 }
 public function init() {
  if(($this->resource = resource_connect()) == false) {
   return false;
  }
  return true;
 }
}

When the user creates a new ResourceClass object, there are no actions taken, which can mean the code fails. To actually initialize any sort of potentially faulty code, you call the init() method. This can fail without any issues.

The third option is usually the best available, and it is reinforced by the fact that it is the standard method of handling constructor failure in more traditional object-oriented languages such as C++. In C++ the cleanup done in a catch block around a constructor call is a little more important than in PHP because memory management might need to be performed. Fortunately, in PHP memory management is handled for you, as in this example:

class Stillborn {
 public function _ _construct() {
  throw new Exception;
 }
 public function _ _destruct() {
  print "destructing\n";
 }
}
try {
 $sb = new Stillborn;
}
catch(Stillborn $e) {}

Running this generates no output at all:

>php stillborn.php
>

The Stillborn class demonstrates that the object's destructors are not called if an exception is thrown inside the constructor. This is because the object does not really exist until the constructor is returned from.

Installing a Top-Level Exception Handler

An interesting feature in PHP is the ability to install a default exception handler that will be called if an exception reaches the top scope and still has not been caught. This handler is different from a normal catch block in that it is a single function that will handle any uncaught exception, regardless of type (including exceptions that do not inherit from Exception).

The default exception handler is particularly useful in Web applications, where you want to prevent a user from being returned an error or a partial page in the event of an uncaught exception. If you use PHP's output buffering to delay sending content until the page is fully generated, you gracefully back out of any error and return the user to an appropriate page.

To set a default exception handler, you define a function that takes a single parameter:

function default_exception_handler($exception) {}

You set this function like so:

$old_handler = set_exception_handler('default_exception_handler');

The previously defined default exception handler (if one exists) is returned.

User-defined exception handlers are held in a stack, so you can restore the old handler either by pushing another copy of the old handler onto the stack, like this:

set_exception_handler($old_handler);

or by popping the stack with this:

restore_exception_handler();

An example of the flexibility this gives you has to do with setting up error redirects for errors incurred for generation during a page. Instead of wrapping every questionable statement in an individual try block, you can set up a default handler that handles the redirection. Because an error can occur after partial output has been generated, you need to make sure to set output buffering on in the script, either by calling this at the top of each script:

ob_start();

or by setting the php.ini directive:

output_buffering = On

The advantage of the former is that it allows you to more easily toggle the behavior on and off in individual scripts, and it allows for more portable code (in that the behavior is dictated by the content of the script and does not require any nondefault .ini settings). The advantage of the latter is that it allows for output buffering to be enabled in every script via a single setting, and it does not require adding output buffering code to every script. In general, if I am writing code that I know will be executed only in my local environment, I prefer to go with .ini settings that make my life easier. If I am authoring a software product that people will be running on their own servers, I try to go with a maximally portable solution. Usually it is pretty clear at the beginning of a project which direction the project is destined to take.

The following is an example of a default exception handler that will automatically generate an error page on any uncaught exception:

<?php
function redirect_on_error($e) {
 ob_end_clean();
 include("error.html");
}
set_exception_handler("redirect_on_error");
ob_start();
// ... arbitrary page code goes here
?>

This handler relies on output buffering being on so that when an uncaught exception is bubbled to the top calling scope, the handler can discard all content that has been generated up to this point and return an HTML error page instead.

You can further enhance this handler by adding the ability to handle certain error conditions differently. For example, if you raise an AuthException exception, you can redirect the person to the login page instead of displaying the error page:

<?php
function redirect_on_error($e) {
 ob_end_clean();
 if(is_a($e, "AuthException")) {
  header("Location: /login/php");
 }
 else {
  include("error.html");
 }
}
set_exception_handler("redirect_on_error");
ob_start();
// ... arbitrary page code goes here
? >

Data Validation

A major source of bugs in Web programming is a lack of validation for client-provided data. Data validation involves verification that the data you receive from a client is in fact in the form you planned on receiving. Unvalidated data causes two major problems in code:

  • Trash data

  • Maliciously altered data

Trash data is information that simply does not match the specification of what it should be. Consider a user registration form where users can enter their geographic information. If a user can enter his or her state free form, then you have exposed yourself to getting states like

  • New Yrok (typo)

  • Lalalala (intentionally obscured)

A common tactic used to address this is to use drop-down option boxes to provide users a choice of state. This only solves half the problem, though: You've prevented people from accidentally entering an incorrect state, but it offers no protection from someone maliciously altering their POST data to pass in a non-existent option.

To protect against this, you should always validate user data in the script as well. You can do this by manually validating user input before doing anything with it:

<?php
$STATES = array('al' => 'Alabama',
         /* ... */,
        'wy' => 'Wyoming');
function is_valid_state($state) {
 global $STATES;
 return array_key_exists($STATES, $state);
}
?>

I often like to add a validation method to classes to help encapsulate my efforts and ensure that I don't miss validating any attributes. Here's an example of this:

<?php

class User {
 public id;
 public name;
 public city;
 public state;
 public zipcode;
 public function _ _construct($attr = false) {
  if($attr) {
   $this->name = $attr['name'];
   $this->email = $attr['email'];
   $this->city = $attr['city'];
   $this->state = $attr['state'];
   $this->zipcode = $attr['zipcode'];
  }
 }
 public function validate() {
  if(strlen($this->name) > 100) {
   throw new DataException;
  }
  if(strlen($this->city) > 100) {
   throw new DataException;
  }
  if(!is_valid_state($this->state)) {
   throw new DataException;
  }
  if(!is_valid_zipcode($this->zipcode)) {
   throw new DataException;
  }
 }
}

?>

The validate() method fully validates all the attributes of the User object, including the following:

  • Compliance with the lengths of database fields

  • Handling foreign key data constraints (for example, the user's U.S. state being valid)

  • Handling data form constraints (for example, the zip code being valid)

To use the validate() method, you could simply instantiate a new User object with untrusted user data:

$user = new User($_POST);

and then call validate on it

try {
  $user->validate();
}
catch (DataException $e) {
  /* Do whatever we should do if the users data is invalid */
}

Again, the benefit of using an exception here instead of simply having validate() return true or false is that you might not want to have a try block here at all; you might prefer to allow the exception to percolate up a few callers before you decide to handle it.

Malicious data goes well beyond passing in nonexistent state names, of course. The most famous category of bad data validation attacks are referred to as cross-site scripting attacks. Cross-site scripting attacks involve putting malicious HTML (usually client-side scripting tags such as JavaScript tags) in user-submitted forms.

The following case is a simple example. If you allow users of a site to list a link to their home page on the site and display it as follows:

<a href="<?= $url ?>">Click on my home page</a>

where url is arbitrary data that a user can submit, they could submit something like this:

$url ='http://example.foo/" onClick=bad_javascript_func foo="';

When the page is rendered, this results in the following being displayed to the user:

<a href="'http://example.foo/" onClick=bad_javascript_func foo="">
 Click on my home page
</a>

This will cause the user to execute bad_javascript_func when he or she clicks the link. What's more, because it is being served from your Web page, the JavaScript has full access to the user's cookies for your domain. This is, of course, really bad because it allows malicious users to manipulate, steal, or otherwise exploit other users' data.

Needless to say, proper data validation for any user data that is to be rendered on a Web page is essential to your site's security. The tags that you should filter are of course regulated by your business rules. I prefer to take a pretty draconian approach to this filtering, declining any text that even appears to be JavaScript. Here's an example:

<?php
$UNSAFE_HTML[] = "!javascript\s*:!is";
$UNSAFE_HTML[] = "!vbscri?pt\s*:!is";
$UNSAFE_HTML[] = "!<\s*embed.*swf!is";
$UNSAFE_HTML[] = "!<[^>]*[^a-z]onabort\s*=!is";
$UNSAFE_HTML[] = "!<[^>]*[^a-z]onblur\s*=!is";
$UNSAFE_HTML[] = "!<[^>]*[^a-z]onchange\s*=!is";
$UNSAFE_HTML[] = "!<[^>]*[^a-z]onfocus\s*=!is";
$UNSAFE_HTML[] = "!<[^>]*[^a-z]onmouseout\s*=!is";
$UNSAFE_HTML[] = "!<[^>]*[^a-z]onmouseover\s*=!is";
$UNSAFE_HTML[] = "!<[^>]*[^a-z]onload\s*=!is";
$UNSAFE_HTML[] = "!<[^>]*[^a-z]onreset\s*=!is";
$UNSAFE_HTML[] = "!<[^>]*[^a-z]onselect\s*=!is";
$UNSAFE_HTML[] = "!<[^>]*[^a-z]onsubmit\s*=!is";
$UNSAFE_HTML[] = "!<[^>]*[^a-z]onunload\s*=!is";
$UNSAFE_HTML[] = "!<[^>]*[^a-z]onerror\s*=!is";
$UNSAFE_HTML[] = "!<[^>]*[^a-z]onclick\s*=!is";

function unsafe_html($html) {
 global $UNSAFE_HTML;
 $html = html_entities($html, ENT_COMPAT, ISO-8859-1_
 foreach ( $UNSAFE_HTML as $match ) {
  if( preg_match($match, $html, $matches) ) {
   return $match;
  }
 }
 return false;
}
?>

If you plan on allowing text to be directly integrated into tags (as in the preceding example), you might want to go so far as to ban any text that looks at all like client-side scripting tags, as in this example:

$UNSAFE_HTML[] = "!onabort\s*=!is";
$UNSAFE_HTML[] = "!onblur\s*=!is";
$UNSAFE_HTML[] = "!onchange\s*=!is";
$UNSAFE_HTML[] = "!onfocus\s*=!is";
$UNSAFE_HTML[] = "!onmouseout\s*=!is";
$UNSAFE_HTML[] = "!onmouseover\s*=!is";
$UNSAFE_HTML[] = "!onload\s*=!is";
$UNSAFE_HTML[] = "!onreset\s*=!is";
$UNSAFE_HTML[] = "!onselect\s*=!is";
$UNSAFE_HTML[] = "!onsubmit\s*=!is";
$UNSAFE_HTML[] = "!onunload\s*=!is";
$UNSAFE_HTML[] = "!onerror\s*=!is";
$UNSAFE_HTML[] = "!onclick\s*=!is";

It is often tempting to turn on magic_quotes_gpc in you php.ini file. magic_quotes automatically adds quotes to any incoming data. I do not care for magic_quotes. For one, it can be a crutch that makes you feel safe, although it is simple to craft examples such as the preceding ones that are exploitable even with magic_quotes on.

With data validation (especially with data used for display purposes), there is often the option of performing filtering and conversion inbound (when the data is submitted) or outbound (when the data is displayed). In general, filtering data when it comes in is more efficient and safer. Inbound filtering needs to be performed only once, and you minimize the risk of forgetting to do it somewhere if the data is displayed in multiple places. The following are two reasons you might want to perform outbound filtering:

  • You need highly customizable filters (for example, multilingual profanity filters).

  • Your content filters change rapidly.

In the latter case, it is probably best to filter known malicious content on the way in and add a second filtering step on the way out.

Further Data Validation

Web page display is not the only place that unvalidated data can be exploited. Any and all data that is received from a user should be checked and cleaned before usage. In database queries, for instance, proper quoting of all data for insert should be performed. There are convenience functions to help perform these conversion operations.

A high-profile example of this are the so-called SQL injection attacks. A SQL injection attack works something like this: Suppose you have a query like this:

$query = "SELECT * FROM users where userid = $userid";

If $userid is passed in, unvalidated, from the end user, a malicious user could pass in this:

$userid = "10; DELETE FROM users;";

Because MySQL (like many other RDBMS systems) supports multiple queries inline, if this value is passed in unchecked, you will have lost your user's table. This is just one of a number of variations on this sort of attack. The moral of the story is that you should always validate any data in queries.

InformIT Promotional Mailings & Special Offers

I would like to receive exclusive offers and hear about products from InformIT and its family of brands. I can unsubscribe at any time.

Overview


Pearson Education, Inc., 221 River Street, Hoboken, New Jersey 07030, (Pearson) presents this site to provide information about products and services that can be purchased through this site.

This privacy notice provides an overview of our commitment to privacy and describes how we collect, protect, use and share personal information collected through this site. Please note that other Pearson websites and online products and services have their own separate privacy policies.

Collection and Use of Information


To conduct business and deliver products and services, Pearson collects and uses personal information in several ways in connection with this site, including:

Questions and Inquiries

For inquiries and questions, we collect the inquiry or question, together with name, contact details (email address, phone number and mailing address) and any other additional information voluntarily submitted to us through a Contact Us form or an email. We use this information to address the inquiry and respond to the question.

Online Store

For orders and purchases placed through our online store on this site, we collect order details, name, institution name and address (if applicable), email address, phone number, shipping and billing addresses, credit/debit card information, shipping options and any instructions. We use this information to complete transactions, fulfill orders, communicate with individuals placing orders or visiting the online store, and for related purposes.

Surveys

Pearson may offer opportunities to provide feedback or participate in surveys, including surveys evaluating Pearson products, services or sites. Participation is voluntary. Pearson collects information requested in the survey questions and uses the information to evaluate, support, maintain and improve products, services or sites, develop new products and services, conduct educational research and for other purposes specified in the survey.

Contests and Drawings

Occasionally, we may sponsor a contest or drawing. Participation is optional. Pearson collects name, contact information and other information specified on the entry form for the contest or drawing to conduct the contest or drawing. Pearson may collect additional personal information from the winners of a contest or drawing in order to award the prize and for tax reporting purposes, as required by law.

Newsletters

If you have elected to receive email newsletters or promotional mailings and special offers but want to unsubscribe, simply email information@informit.com.

Service Announcements

On rare occasions it is necessary to send out a strictly service related announcement. For instance, if our service is temporarily suspended for maintenance we might send users an email. Generally, users may not opt-out of these communications, though they can deactivate their account information. However, these communications are not promotional in nature.

Customer Service

We communicate with users on a regular basis to provide requested services and in regard to issues relating to their account we reply via email or phone in accordance with the users' wishes when a user submits their information through our Contact Us form.

Other Collection and Use of Information


Application and System Logs

Pearson automatically collects log data to help ensure the delivery, availability and security of this site. Log data may include technical information about how a user or visitor connected to this site, such as browser type, type of computer/device, operating system, internet service provider and IP address. We use this information for support purposes and to monitor the health of the site, identify problems, improve service, detect unauthorized access and fraudulent activity, prevent and respond to security incidents and appropriately scale computing resources.

Web Analytics

Pearson may use third party web trend analytical services, including Google Analytics, to collect visitor information, such as IP addresses, browser types, referring pages, pages visited and time spent on a particular site. While these analytical services collect and report information on an anonymous basis, they may use cookies to gather web trend information. The information gathered may enable Pearson (but not the third party web trend services) to link information with application and system log data. Pearson uses this information for system administration and to identify problems, improve service, detect unauthorized access and fraudulent activity, prevent and respond to security incidents, appropriately scale computing resources and otherwise support and deliver this site and its services.

Cookies and Related Technologies

This site uses cookies and similar technologies to personalize content, measure traffic patterns, control security, track use and access of information on this site, and provide interest-based messages and advertising. Users can manage and block the use of cookies through their browser. Disabling or blocking certain cookies may limit the functionality of this site.

Do Not Track

This site currently does not respond to Do Not Track signals.

Security


Pearson uses appropriate physical, administrative and technical security measures to protect personal information from unauthorized access, use and disclosure.

Children


This site is not directed to children under the age of 13.

Marketing


Pearson may send or direct marketing communications to users, provided that

  • Pearson will not use personal information collected or processed as a K-12 school service provider for the purpose of directed or targeted advertising.
  • Such marketing is consistent with applicable law and Pearson's legal obligations.
  • Pearson will not knowingly direct or send marketing communications to an individual who has expressed a preference not to receive marketing.
  • Where required by applicable law, express or implied consent to marketing exists and has not been withdrawn.

Pearson may provide personal information to a third party service provider on a restricted basis to provide marketing solely on behalf of Pearson or an affiliate or customer for whom Pearson is a service provider. Marketing preferences may be changed at any time.

Correcting/Updating Personal Information


If a user's personally identifiable information changes (such as your postal address or email address), we provide a way to correct or update that user's personal data provided to us. This can be done on the Account page. If a user no longer desires our service and desires to delete his or her account, please contact us at customer-service@informit.com and we will process the deletion of a user's account.

Choice/Opt-out


Users can always make an informed choice as to whether they should proceed with certain services offered by InformIT. If you choose to remove yourself from our mailing list(s) simply visit the following page and uncheck any communication you no longer want to receive: www.informit.com/u.aspx.

Sale of Personal Information


Pearson does not rent or sell personal information in exchange for any payment of money.

While Pearson does not sell personal information, as defined in Nevada law, Nevada residents may email a request for no sale of their personal information to NevadaDesignatedRequest@pearson.com.

Supplemental Privacy Statement for California Residents


California residents should read our Supplemental privacy statement for California residents in conjunction with this Privacy Notice. The Supplemental privacy statement for California residents explains Pearson's commitment to comply with California law and applies to personal information of California residents collected in connection with this site and the Services.

Sharing and Disclosure


Pearson may disclose personal information, as follows:

  • As required by law.
  • With the consent of the individual (or their parent, if the individual is a minor)
  • In response to a subpoena, court order or legal process, to the extent permitted or required by law
  • To protect the security and safety of individuals, data, assets and systems, consistent with applicable law
  • In connection the sale, joint venture or other transfer of some or all of its company or assets, subject to the provisions of this Privacy Notice
  • To investigate or address actual or suspected fraud or other illegal activities
  • To exercise its legal rights, including enforcement of the Terms of Use for this site or another contract
  • To affiliated Pearson companies and other companies and organizations who perform work for Pearson and are obligated to protect the privacy of personal information consistent with this Privacy Notice
  • To a school, organization, company or government agency, where Pearson collects or processes the personal information in a school setting or on behalf of such organization, company or government agency.

Links


This web site contains links to other sites. Please be aware that we are not responsible for the privacy practices of such other sites. We encourage our users to be aware when they leave our site and to read the privacy statements of each and every web site that collects Personal Information. This privacy statement applies solely to information collected by this web site.

Requests and Contact


Please contact us about this Privacy Notice or if you have any requests or questions relating to the privacy of your personal information.

Changes to this Privacy Notice


We may revise this Privacy Notice through an updated posting. We will identify the effective date of the revision in the posting. Often, updates are made to provide greater clarity or to comply with changes in regulatory requirements. If the updates involve material changes to the collection, protection, use or disclosure of Personal Information, Pearson will provide notice of the change through a conspicuous notice on this site or other appropriate way. Continued use of the site after the effective date of a posted revision evidences acceptance. Please contact us if you have questions or concerns about the Privacy Notice or any objection to any revisions.

Last Update: November 17, 2020