When a new DokuWiki version is released, some of the files that were needed previously may become obsolete.

Keeping the files may cause the following issues:

• Misleading error messages related to outdated plugins.
• They pose a security risk.
• They could break later installations.

So we recommend to delete them.

A list of all files that were removed in recent releases can be found in data/deleted.files. You should check if they still exist in your install. If they do, delete them.

You can also see the current list at https://github.com/dokuwiki/dokuwiki/raw/stable/data/deleted.files

Important: The list of files is case-sensitive. On case-insensitive file systems¹⁾ the scripts below may delete too much!

When upgrading your wiki using the upgrade Plugin, the plugin will take care of removing the deleted files automatically for you.

Alternatively the removeold Plugin can be used.

People with shell access to their server can use a script to automatically delete all old files.

This simple line should work on any Linux system

grep -Ev '^($|#)' data/deleted.files | xargs -n 1 rm -vf

Some systems may not support the “rm -d” option for directory removal. In that case, you have to use recursive removal (just be sure to double-check that the file list does not include any paths that will delete too much):

grep -Ev '^($|#)' data/deleted.files | xargs -n 1 rm -fr

Here's a Python script that will also print the files deleted

import os
for line in open("./data/deleted.files"):
    if line.isspace() or line[0] == '#':
        continue
    line = line.rstrip(os.linesep)
    try:
        if os.path.exists(line):
            print('File removed =>  ' + line)
            os.remove(line)
    except OSError:
        pass

Here's an alternative Python script that is case sensitive and will also delete directories included in the list

import os
import shutil
 
def exists_casesensitive(path):
    if not os.path.exists(path):
        return False
    directory, filename = os.path.split(path)
    return filename in os.listdir(directory)
 
with open("./data/deleted.files") as file:
    for line in file:
        line = line.strip()
        if line and not line.startswith('#'):
            path = line.rstrip(os.linesep)
            if exists_casesensitive(path):
                if os.path.isdir(path):
                    shutil.rmtree(path)
                    print('Directory removed =>  ' + path)
                else:
                    os.remove(path)
                    print('File removed =>  ' + path)
            else:
                #print('File not found => ' + path)
                pass

Here's a Ruby script doing the same

files = File.read("./data/deleted.files").split("\n").grep(/^[^$#]/)
 
files.each do |file|
  puts "Deleting #{file}."
  File.delete(file) rescue puts $!
end

The same for PHP:

<?php
/* Security function, comment this out to "activate" the script. */
exit('Check source');
 
$path = getcwd();
 
if(file_exists($path . '/data/deleted.files')) {
  $file = fopen($path . '/data/deleted.files', 'r');
  while(false !== ($line = fgets($file))) {
    $line = trim($line);
    if(!empty($line) && !preg_match('/^\#/', $line) && file_exists($path . '/' . $line)) {
      unlink($path . '/' . $line);
      echo '/' . $line . ' - deleted<br/>';
    }
  }
  echo 'Done!';
}

Here's a one-line powershell script doing the same

Get-Content .\data\deleted.files | Where-Object {$_ -notmatch '^($|#)' -and (Test-Path $_ )} | Foreach-Object {Remove-Item -path $_ -Force -Recurse}

Here's a one-line Windows's CMD script doing the same, although you may need to first reverse all the slashes in 'deleted.files' for the filenames to be accepted.

for /F %i in ('findstr /R /V /C:"^#" /C:"^$" data\deleted.files') do del "%i"