Problem with UTF-8 file/directory names

[zaak]

I have following directory structure:

.
└── files
    ├── 繁體中文字
    │   └── test.txt
    └── 繁體中文字.txt

When I call $filesystem->listContents('files') I'm getting following output:

array (size=2)
  0 => 
    array (size=6)
      'dirname' => string 'files' (length=5)
      'basename' => string 'files' (length=5)
      'filename' => string 'files' (length=5)
      'path' => string 'files/繁體中文字' (length=21)
      'type' => string 'dir' (length=3)
      'timestamp' => int 1409223899
  1 => 
    array (size=8)
      'dirname' => string 'files' (length=5)
      'basename' => string '.txt' (length=4)
      'extension' => string 'txt' (length=3)
      'filename' => string '' (length=0)
      'path' => string 'files/繁體中文字.txt' (length=25)
      'type' => string 'file' (length=4)
      'timestamp' => int 1409223047
      'size' => int 0

As you can see filename and basename are invalid. This is caused by lack of multibyte support in pathinfo used in Util::pathinfo(). Expected output is:

array (size=2)
  0 => 
    array (size=7)
      'dirname' => string 'files' (length=5)
      'basename' => string '繁體中文字' (length=15)
      'extension' => string '' (length=0)
      'filename' => string '繁體中文字' (length=15)
      'path' => string 'files/繁體中文字' (length=21)
      'type' => string 'dir' (length=3)
      'timestamp' => int 1409223899
  1 => 
    array (size=8)
      'dirname' => string 'files' (length=5)
      'basename' => string '繁體中文字.txt' (length=19)
      'extension' => string 'txt' (length=3)
      'filename' => string '繁體中文字' (length=15)
      'path' => string 'files/繁體中文字.txt' (length=25)
      'type' => string 'file' (length=4)
      'timestamp' => int 1409223047
      'size' => int 0

[zaak]

Gah, it was server locale settings messing around. Closing.

[zaak]

Anyone who encountered similar issue - make sure you set locale in your code with setlocale, like:

setlocale(LC_ALL, "en_US.utf8");

[ThaDafinser]

@zaak did you had this issue under windows or unix?

[zaak]

@ThaDafinser Linux. On Windows it fails completely if it's about UTF-8 names, but it's PHP itself at fault, not the Flysystem.

[ThaDafinser]

@zaak thank you.

Just tried on PHP7 and still not solved.

I found this for windows (in case someone comes around here):
https://github.com/kenjiuno/php-wfio

Alternative with com extension:

<?php
$path = '..YOUR_PATH...';

/*
 * COM
 */
$fso = new \COM('Scripting.FileSystemObject', null, CP_UTF8);

$folder = $fso->GetFolder($path);

foreach ($folder->Files as $obj) {
    var_dump($obj->Name);
}

[adrianopedro]

I also found this problem...
The problem is that at the origin the files are url encoded for safety, but then are not reverted.
In the pathinfo, just treat the filename with urldecode.

public static function pathinfo($path)
    {

       $pathinfo = pathinfo($path) + compact('path');

        //Fix problem with filenames having spacing or special chars 
        //that are url encoded for safety, but then are not reverted.
        $pathinfo['filename'] = urldecode($pathinfo['filename']);

        $pathinfo['dirname'] = array_key_exists('dirname', $pathinfo)
            ? static::normalizeDirname($pathinfo['dirname'])
            : '';

        return $pathinfo;
    }

[ThaDafinser]

Most or all of the issues here should be resolved with php 7.1
https://github.com/php/php-src/blob/PHP-7.1.0beta2/UPGRADING#L321