Perl to PHP With s///;

Question

In perl, I can do: 1 while $var =~ s/a/b/;, and it will replace all a with b. In many cases, I would use it more like 1 while $var =~ s/^"(.*)"$/$1/; to remove all pairs of double quotes around a string.

Is there a way to do something similar to this in PHP, without having to do

while (preg_match('/^"(.*)"$/', $var)) {
    $var = preg_replace('/^"(.*)"$/', '$1', $var, 1);
}

Because apparently,

while ($var = preg_replace('/^"(.*)"$/', '$1', $var, 1)) { 1; }

doesn't work.

EDIT: The specific situation I'm working in involves replacing values in a string with values from an associative array:

$text = "This is [site_name], home of the [people_type]".
$array = ('site_name' => 'StackOverflow.com', 'people_type' => 'crazy coders');

where I would be doing:

while (preg_match('/\[.*?\]/', $text)) {
    $text = preg_replace('/\[(.*?)\]/', '$array[\'$1\']', $text, 1);
}

with the intended output being 'This is StackOverflow.com, home of the crazy coders'

Answer 1

What does your input data look like?

Because you're checking for double quotes only at the head and tail of the string. If that's accurate, then you don't need to capture a backreference at all. Also, that would make sending 1 as the 4th parameter completely superfluous.

$var = '"foo"';

// This works
echo preg_replace( '/^"(.*)"$/', '$1', $var );

// So does this
echo preg_replace( '/^"|"$/', '', $var );

But if your input data looks different, that would change my answer.

EDIT

Here's my take on your actual data

class VariableExpander
{
  protected $source;

  public function __construct( array $source )
  {
    $this->setSource( $source );
  }

  public function setSource( array $source )
  {
    $this->source = $source;
  }

  public function parse( $input )
  {
    return preg_replace_callback( '/\[([a-z_]+)\]/i', array( $this, 'expand' ), $input );
  }

  protected function expand( $matches )
  {
    return isset( $this->source[$matches[1]] )
      ? $this->source[$matches[1]]
      : '';
  }
}

$text = "This is [site_name], home of the [people_type]";
$data = array(
    'site_name'   => 'StackOverflow.com'
  , 'people_type' => 'crazy coders'
);

$ve = new VariableExpander( $data );
echo $ve->parse( $text );

The class is just for encapsulation - you could do this in a structured way if you wanted.

Answer 2

Could something like this work?

$var = preg_replace('/^("+)(.*)\1$', '$2', $var, 1);

Answer 3

Use do-while:

do {
  $var = preg_replace('/^"(.*)"$/', "$1", $var, 1, $count);
} while ($count == 1);

Requires at least php-5.1.0 due to its use of $count.

You could also write

do {
  $last = $var;
  $var = preg_replace('/^"(.*)"$/', "$1", $var);
} while ($last != $var);

Answer 4

preg_replace('#\[(.*?)\]#e', "\$array['$1']", $text);

Answer 5

In all of the cases, you can get rid of the loop by (e.g.) using the /g global replace option or rewriting the regexp:

$var =~ s/a/b/g;
$var =~ s/^("+)(.*)\1$/$2/;

The same patterns should work in PHP. You can also get rid of the $limit argument to preg_replace:

$text = preg_replace('/\[(.*?)\]/e', '$array[\'$1\']', $text);

Regular expressions can handle their own loops. Looping outside the RE is inefficient, since the RE has to process text it already processed in previous iterations.