Tag Archives: Computer programming

Default Error Handling in Zend Framework

Zend_Controller_Action_Exception

There’s a cool feature in Zend Framework when a controller or action doesn’t exists. That’s really useful, because you simply write an ErrorController.php in you default module and every action’s exception sends the user to this controller/action and the job is done for you. What is important here is to show the error message at least on the development stage. Actually the request stores it as a parameter. Here’s some source:

<?php
 
class ErrorController extends Zend_Controller_Action
{
    public function errorAction()
    {
        $errorHandler = $this->getRequest()->getParam('error_handler');
        /* @var $error Zend_Controller_Action_Exception */
        $error = $errorHandler->exception;
        echo $error->getMessage();
    }
}

Zend Framework: Simple Acl Front Controller Plugin

access with Zend_Acl

Almost every web site need some abstraction over the access control list (ACL) to grant access of its users. As usual Zend Framework has quite good mechanism to deal with this – Zend_Acl.

Out in the web there are a lot of resources about Zend_Acl’s usage, so I ain’t going to cover it one more time, but simply copy/paste a very small front controller plugin implementing the basic usage of Zend_Acl.

Note that instead of defining the __construct() here is called preDispatch where the request is passed as a parameter. However only by copy pasting not every answer will be given. That’s why I’m going to write more about Zend_Acl in my future posts, for now only the source code:

<?php
 
class AclInit extends Zend_Controller_Plugin_Abstract
{
    public function preDispatch(Zend_Controller_Request_Abstract $request)
    {
        // create new acl object
        $acl = new Zend_Acl();
 
        // define resources. typically there are
        // only four resources from the CRUD functionality
        // but there can be added more resources
        $acl->add(new Zend_Acl_Resource('index'))
            ->add(new Zend_Acl_Resource('create'))
            ->add(new Zend_Acl_Resource('read'))
            ->add(new Zend_Acl_Resource('update'))
            ->add(new Zend_Acl_Resource('delete'));
 
        // define roles
        $acl->addRole(new Zend_Acl_Role('guest'))
            ->addRole(new Zend_Acl_Role('admin'));
 
        // define privileges
        $acl->allow('guest', array('index', 'read'))
            ->allow('admin');
 
        // setup acl in the registry for more
        Zend_Registry::set('acl', $acl);
 
        // check permissions
        if (!$acl->isAllowed('guest', $request->getActionName())) {
            $request->setControllerName('error');
            $request->setActionName('error');
        }
    }
}

PHP: preg_match Give Names to the Matches

So Called – Subpatterns

Patterns - preg_match

In PHP 5.2.2+ you can name the sub patterns returned from preg_match with a specific syntax.

Named subpatterns now accept the syntax (?<name>) and (?’name’) as well as (?P<name>). Previous versions accepted only (?P<name>)

This is extremely helpful, when dealing with long patterns. As you may know you can simply use the “old school” way and to call the matches by their number based index:

    $haystack = '01 Jan 1970';
    $pattern = '/(\d{1,2})\ (Jan|Feb)\ (19\d\d)/';
 
    preg_match($pattern, $haystack, $matches);
 
    print_r($matches);

Although it may look difficult to maintain, now you can simply name the sub patterns of preg_match and to call them with their associative array keys. This is more clear when writing code and it’s definitely more maintainable.

    $haystack = '01 Jan 1970';
    $pattern = '/(?<day>\d{1,2})\ (?<month>Jan|Feb)\ (?<year>19\d\d)/';
 
    preg_match($pattern, $haystack, $matches);
 
    print_r($matches);
 
    // now there's $matches['day'], $matches['month'] ...

Zend_View_Helper_InlineScript Appends Scripts Twice

Double Trouble

duplicate
This is a well known problem, which is well described as a bug in Zend Framework’s issue tracker. Once you add some scripts with the inlineScript helper you’ll receive them twice in the code. This makes the page slower and leads to some crashes.

$scripts->appendFile('my-file1.js')
        ->appendFile('my-file2.js')
        ->appendScript('alert("test")');

This will print the alert twice.

<script src="my-file1.js"></script>
<script src="my-file2.js"></script>
<script>
//<!--
alert("test");
//-->
</script>
<script>
//<!--
alert("test");
//-->
</script>

Quick Solution

Although this is the worst solution I’ve ever made – it works! There is a check for duplicates in the HeadScript helper around line 240:

if (!$this->_isDuplicate($content)) {
...

and then a check in the _isDuplicate method(), ~ line 270:

if (($item->source === null)
    && array_key_exists('src', $item->attributes)
    && ($file == $item->attributes['src']))
{
    return true;
}

The thing you’ve to do is to add such a check for the appendScript() and not only for appendFile() method – again ~ line 230:

case 'script':
	if (!$this->_isDuplicate($content)) {
	    $item = $this->createData($type, $attrs, $content);
	    if ('offsetSet' == $action) {
	        $this->offsetSet($index, $item);
	    } else {
	        $this->$action($item);
	    }
	}
	break;

And then slightly modify the _isDuplicate method:

foreach ($this->getContainer() as $item) {
    if (($item->source === null)
        && array_key_exists('src', $item->attributes)
        && ($file == $item->attributes['src']))
    {
        return true;
    }
    if ($item->source == $file) {
        return true;
    }
}
return false;