Issue #8943 has been updated by luislavena (Luis Lavena).

Assignee set to cruby-windows

Hello,

Code in win32/registry predates the presence of encoding support built in Ruby (started in 1.9.1 and greater)

Feel free to send a patch that uses Wide characters functions to perform these operations, but don't bluntly assume people don't an idea on this topic.

If you have been affected by this for so many years, a better use of your time will be provide a patch instead of aggressively complaining.

The principle of Open Source is "scratch your own itch".
----------------------------------------
Bug #8943: Invalid encoding for Windows Registry
https://bugs.ruby-lang.org/issues/8943#change-41946

Author: davispuh (D??vis Mos??ns)
Status: Open
Priority: Normal
Assignee: cruby-windows
Category: 
Target version: 
ruby -v: ruby 2.0.0p247 (2013-06-27) [x64-mingw32]
Backport: 1.9.3: UNKNOWN, 2.0.0: UNKNOWN


=begin
It's not possible to correctly read non-ASCII registry values. It's also not possible to write them with correct encoding.

to test import ((<test.reg|URL:https://gist.github.com/davispuh/6676385#file-test-reg>))

then ((<test3.rb|URL:https://gist.github.com/davispuh/6676377#file-test3-rb>))

Output on Ruby 2.0 (it's same for 1.9.3 except encoding is US-ASCII not UTF-8, but both are wrong)

 N:\Projects>ruby test3.rb
 UTF-8
 ??????
 "\xE5??\xE4?\xE2"
 test

(it hangs here)

screenshot of registry after running it ((<URL:http://i.imm.io/1gFES.png>))


seems that whoever wrote Ruby code to interact with WinAPI had no idea about encodings/code pages Windows uses as there's loads of issues and they've been for years.
=end



-- 
http://bugs.ruby-lang.org/