Question

The routine below does two scans over an input stream of hypertext. The first pass is a spin replacement on user defined phrase options. The second pass is a find replace on the tags collection in the doReplace function below.

I’m just looking for suggestions on how it might be optimized. I’m having no performance issues as is. But I want to build for scalability.

/* FIND REPLACE SPIN
--------------------------------------------------------------------*/
function doReplace($content)
{
// content is a precompiled text document formatted with html and 
// special using replacement tags matching the $tags array collection below
    $tags = array('[blog-name]', '[blog-url]', '[blog-email]');
    $replacements = array('value1', 'value2', 'value3');
    $content = str_replace($tags, $replacements, $content);
    return $content;
}

function doSpin($content) {
// the content also has phrase option tags denoted by [%phrase1|phrase2_|phrase3%]
// delimiters throughout the text.
  return preg_replace_callback('!\[%(.*?)%\]!', 'pick_one', $content);
}

function pick_one($matches) {
  $choices = explode('|', $matches[1]);
  return $choices[rand(0, count($choices)-1)];
}

$my_source_page = file_get_contents('path/to/source';}
$my_source1_spin = doSpin($my_source_page);
$my_source1_replace = doReplace($my_source1_spin);
$my_source1_final = addslashes($my_source1_replace);

//Now do something with $my_source1_final

Answer 1

To be honest, I don’t see anything wrong with the code you’ve posted. The main bottleneck in the code is likely going to be the file_get_contents call.

The only thing I can see myself is that you’re allocating the string to different variables (four variables beginning $my_source) which will use more memory than if you just used 1 or 2 variables.

But unless you’re reading a large amount of text into memory very frequently on a busy site, then I don’t think you need to worry about the code you’ve posted. And you said yourself, you’re not having any performance issues at the moment 😉