How to minify php page html output?

Question

CSS and Javascript

Consider the following link to minify Javascript/CSS files: https://github.com/mrclay/minify

HTML

Tell Apache to deliver HTML with GZip - this generally reduces the response size by about 70%. (If you use Apache, the module configuring gzip depends on your version: Apache 1.3 uses mod_gzip while Apache 2.x uses mod_deflate.)

Accept-Encoding: gzip, deflate

Content-Encoding: gzip

Use the following snippet to remove white-spaces from the HTML with the help ob_start's buffer:

<?php

function sanitize_output($buffer) {

    $search = array(
        '/\>[^\S ]+/s',     // strip whitespaces after tags, except space
        '/[^\S ]+\</s',     // strip whitespaces before tags, except space
        '/(\s)+/s',         // shorten multiple whitespace sequences
        '/<!--(.|\s)*?-->/' // Remove HTML comments
    );

    $replace = array(
        '>',
        '<',
        '\\1',
        ''
    );

    $buffer = preg_replace($search, $replace, $buffer);

    return $buffer;
}

ob_start("sanitize_output");

?>

Turn on gzip if you want to do it properly. You can also just do something like this:

$this->output = preg_replace(
    array(
        '/ {2,}/',
        '/<!--.*?-->|\t|(?:\r?\n[ \t]*)+/s'
    ),
    array(
        ' ',
        ''
    ),
    $this->output
);

This removes about 30% of the page size by turning your html into one line, no tabs, no new lines, no comments. Mileage may vary

I've tried several minifiers and they either remove too little or too much.

This code removes redundant empty spaces and optional HTML (ending) tags. Also it plays it safe and does not remove anything that could potentially break HTML, JS or CSS.

Also the code shows how to do that in Zend Framework:

class Application_Plugin_Minify extends Zend_Controller_Plugin_Abstract {

  public function dispatchLoopShutdown() {
    $response = $this->getResponse();
    $body = $response->getBody(); //actually returns both HEAD and BODY

    //remove redundant (white-space) characters
    $replace = array(
        //remove tabs before and after HTML tags
        '/\>[^\S ]+/s'   => '>',
        '/[^\S ]+\</s'   => '<',
        //shorten multiple whitespace sequences; keep new-line characters because they matter in JS!!!
        '/([\t ])+/s'  => ' ',
        //remove leading and trailing spaces
        '/^([\t ])+/m' => '',
        '/([\t ])+$/m' => '',
        // remove JS line comments (simple only); do NOT remove lines containing URL (e.g. 'src="http://server.com/"')!!!
        '~//[a-zA-Z0-9 ]+$~m' => '',
        //remove empty lines (sequence of line-end and white-space characters)
        '/[\r\n]+([\t ]?[\r\n]+)+/s'  => "\n",
        //remove empty lines (between HTML tags); cannot remove just any line-end characters because in inline JS they can matter!
        '/\>[\r\n\t ]+\</s'    => '><',
        //remove "empty" lines containing only JS's block end character; join with next line (e.g. "}\n}\n</script>" --> "}}</script>"
        '/}[\r\n\t ]+/s'  => '}',
        '/}[\r\n\t ]+,[\r\n\t ]+/s'  => '},',
        //remove new-line after JS's function or condition start; join with next line
        '/\)[\r\n\t ]?{[\r\n\t ]+/s'  => '){',
        '/,[\r\n\t ]?{[\r\n\t ]+/s'  => ',{',
        //remove new-line after JS's line end (only most obvious and safe cases)
        '/\),[\r\n\t ]+/s'  => '),',
        //remove quotes from HTML attributes that does not contain spaces; keep quotes around URLs!
        '~([\r\n\t ])?([a-zA-Z0-9]+)="([a-zA-Z0-9_/\\-]+)"([\r\n\t ])?~s' => '$1$2=$3$4', //$1 and $4 insert first white-space character found before/after attribute
    );
    $body = preg_replace(array_keys($replace), array_values($replace), $body);

    //remove optional ending tags (see http://www.w3.org/TR/html5/syntax.html#syntax-tag-omission )
    $remove = array(
        '</option>', '</li>', '</dt>', '</dd>', '</tr>', '</th>', '</td>'
    );
    $body = str_ireplace($remove, '', $body);

    $response->setBody($body);
  }
}

But note that when using gZip compression your code gets compressed a lot more that any minification can do so combining minification and gZip is pointless, because time saved by downloading is lost by minification and also saves minimum.

Here are my results (download via 3G network):

 Original HTML:        150kB       180ms download
 gZipped HTML:          24kB        40ms
 minified HTML:        120kB       150ms download + 150ms minification
 min+gzip HTML:         22kB        30ms download + 150ms minification

All of the preg_replace() solutions above have issues of single line comments, conditional comments and other pitfalls. I'd recommend taking advantage of the well-tested Minify project rather than creating your own regex from scratch.

In my case I place the following code at the top of a PHP page to minify it:

function sanitize_output($buffer) {
    require_once('min/lib/Minify/HTML.php');
    require_once('min/lib/Minify/CSS.php');
    require_once('min/lib/JSMin.php');
    $buffer = Minify_HTML::minify($buffer, array(
        'cssMinifier' => array('Minify_CSS', 'minify'),
        'jsMinifier' => array('JSMin', 'minify')
    ));
    return $buffer;
}
ob_start('sanitize_output');