Question

I have a strange bug where in the following code where the basename() is complete removing the $file:

$files = glob(DIR_APPLICATION . 'controller/module/*.php');

if ($files) {

    foreach ($files as $file) {

        $extension = basename($file, '.php');
        print($file).'<br />';

When debuging, if I add the print() code above the basename it works as expected. If I add it below it completely removing the $file.

What could be the reason for this?

if ($files) {

    foreach ($files as $file) {

        echo $file . '<br />';
        var_dump($file);
        echo '<br />';
        $extension = basename($file, '.php');
        var_dump($file);
        echo '<br /><br />';

PHP Version 5.3.10

Answer 1

var_dump gives different lengths for your strings than the number of human-readable characters it shows (specifically, each length is 6 greater than it should normally be).

This means that your pathnames contain multibyte characters, while the documentation for basename says

basename() is locale aware, so for it to see the correct basename with
multibyte character paths, the matching locale must be set using the
setlocale() function.

So, to solve the problem I suggest:

1. Use bin2hex or something similar to see the ordinals of all bytes that make up each path; this will tell you which character(s) in the pathnames are not single-byte. The 6 bytes difference means that it’s probably going to be three characters, so my psychic powers tell me that the www might not be exactly what it looks like.

2. If possible, rename the offending directory to make the paths be single-byte strings and the problem will automatically be solved.

3. If the above is not possible, then:

   a. Find out what encoding the returned paths are in

   b. Write your own implementation of basename that uses multibyte string functions, passing them the encoding.