Thank you for taking the time to write to us, but this is not
a bug. Please double-check the documentation available at
http://www.php.net/manual/ and the instructions on how to report
a bug at http://bugs.php.net/how-to-report.php

When used as a string we need the fully qualified name as we don't know where the parameter is coming from.

Hi Johannes,

I double-checked the documentation and found no mention that the string passed to class_exists() must be fully qualified. Perhaps this is a documentation bug.

It is certainly seems inconsistent/counter-intuitive:-

class_exists('Things\Car'); // FALSE, class does not exist
$car = new Things\Car(); // HUH? Class does exist after all

So PHP is able to figure out there is a "use Foo/Things" namespace in effect on one line, but not on the other? Lame.

I agree that this behavior is counter-intuitive. I seem to have just run into the same problem, myself, where my autoload works just fine where class_exists() returns false.

In short... class_exists() should really infer the given namespace of the current file, just like it's inferred in all other uses.

I have a very similar problem, not with class exists, but with dynamic instantiation: 

  return ($dynamic? new $className : new SelectMysql);

I can fix it concatenating __NAMESPACE__, but it seems weird that PHP cannot determine this.

Here's a working example:

Select.php
<?php

namespace sql;

abstract class Select {
   abstract function limit($limit, $offset);
   static function factory($dynamic = true) {
       $className = "SelectMysql";
       return ($dynamic? new $className : new SelectMysql);
   }
}
?>

SelectMysql.php
<?php

namespace sql;

class SelectMysql {
   function limit($limit, $offset) {
       return;
   }
}
?>

test.php
<?php

include('Select.php');
include('SelectMysql.php');

// this works
$select = sql\Select::factory(true);

// this does not
$select = sql\Select::factory(true);

?>