Thorsten Ruf has released a flow diagram that shows the Zend Framework dispatch process. This sheet could be useful for everyone has to develop controller plugins, custom routes, action helpers or just wants to learn more about Zend Framework internals.

Download

• Zend Framework dispatch overview - v. 1.01 (pdf)

Session support in PHP consists of a way to preserve certain data across subsequent accesses. Each visitor accessing the web site is assigned a unique id, the session id. This session id is either stored in a cookie or propagated in the url. The best solution is to store it in a cookie.

The cookie used to store the session id sets, by default, the cookie domain to the current domain. This means that if you are visiting www.domain.com, the session id will be stored in a cookie bounded to the domain www.domain.com. This behavior leads to a problem: if your site uses multiple subdomains, the session won’t be shared between the site’s subdomains.

To solve this problem you have to share the cookie between all subdomains, setting its domain it to “.domain.com”. The following code shows how to change the session cookie domain with Zend Framework:

Zend_Session::start(array('cookie_domain' => '.domain.com'));

Browers compatibility

The prefix dot in “.domain.com” is not always necessary, but it’s highly recommended to support all browsers.

References

• PHP session cookie multiple domains

There are a lot of web sites, online communities and social networks that apply a watermark over photos uploaded from users. This watermark, usually a logo or a text with the service name, should be applied in a region of the image where there are no important things, because no one likes to see a text over his face.

This simple feature requires an algorithm to automatically select the best corner where apply the watermark.

The algorithm

The watermark should be placed in the corner with the lowest number of graphic details, in order to be as less invasive as possible. To select the best corner, the proposed algorithm counts the number of different colours inside each corner and applies the watermark to the corner with the lowest number of different colours.

This algorithm is very simple but effective, and can be described with the following steps:

1. For each candidate image region (corners) calculate the number of different colours
2. Select the image region with the lowest number of different colours
3. Apply the watermark to the selected region

The code

Here is the code the PHP code of a function that selects the best corner of an image and then apply the watermark.

/**
 * Draw a watermark over an image (the watermark position is
 * selected automatically) and returns true. If the watermark
 * is bigger than the image, this method returns false.
 *
 * @param IMagick $image
 * @param IMagick $watermark
 * @param int $padding
 * @return bool
 */
private function drawWatermark($image, $watermark, $padding = 0)
{
	// Check if the watermark is bigger than the image
	$image_width 		= $image->getImageWidth();
	$image_height 		= $image->getImageHeight();
	$watermark_width 	= $watermark->getImageWidth();
	$watermark_height 	= $watermark->getImageHeight();

	if ($image_width < $watermark_width + $padding || $image_height < $watermark_height + $padding) {
		return false;
	}

	// Calculate each position
	$positions = array();
	$positions[] = array(0 + $padding, 0 + $padding);
	$positions[] = array($image_width - $watermark_width - $padding, 0 + $padding);
	$positions[] = array($image_width - $watermark_width - $padding, $image_height - $watermark_height - $padding);
	$positions[] = array(0 + $padding, $image_height - $watermark_height - $padding);

	// Initialization
	$min 		= null;
	$min_colors = 0;

	// Calculate the number of colors inside each region
	// and retrieve the minimum
	foreach($positions as $position)
	{
		$colors = $image->getImageRegion(
			$watermark_width,
			$watermark_height,
			$position[0],
			$position[1])->getImageColors();

		if ($min === null || $colors <= $min_colors)
		{
			$min 		= $position;
			$min_colors = $colors;
		}
	}

	// Draw the watermark
	$image->compositeImage(
		$watermark,
		Imagick::COMPOSITE_OVER,
		$min[0],
		$min[1]);

	return true;
}

(more…)

Ho trovato un bug nella classe Archive_Tar che si verifica quando si prova ad aggiungere ad un archivio compresso (gzip o bzip2) un file che, al suo interno, contiene il blocco utilizzato per delimitare la fine dell’archivio tar stesso (null padding di 1024 byte).

Questa patch, applicata alla versione 1.3.2 di Archive_Tar, dovrebbe risolvere il problema.

Riferimenti

• PHP - Archive_Tar
• Patch
• Formato tar