# Ferret w/ acts as ferret on windows

**URL:** <https://www.ruby-forum.com/t/ferret-w-acts-as-ferret-on-windows/86777>\
**Category:** Ferret\
**Created:** [February 25, 2007, 9:00pm UTC](https://www.ruby-forum.com/t/ferret-w-acts-as-ferret-on-windows/86777 "2007-02-25T21:00:05Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![neongrau](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/n/34f0e0/32.png) [@neongrau](https://www.ruby-forum.com/u/neongrau)\
**Post date:** [February 25, 2007, 9:00pm UTC](https://www.ruby-forum.com/t/ferret-w-acts-as-ferret-on-windows/86777/1 "2007-02-25T21:00:05Z")

</div>

hi all!

after hours of trying to find contents with german umlauts i stumbled  
upon a post where someone said ferret won’t work with utf-8 on  
windows???

is that really true?

do i really have to iconv everything to iso-8859-15 before indexing and  
do the same with the query to get it working?

i’m running ruby 1.8.5, ferret 0.10.9-mswin32, and rails 1.2.2 and just  
reinstalled aaf from svn yesterday (can’t find any version info, and  
forgot to remember the svn revision)

_banging_head_against_wall\*

---

<div class="post-metadata">

**Author:** ![neongrau](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/n/34f0e0/32.png) [@neongrau](https://www.ruby-forum.com/u/neongrau)\
**Post date:** [February 25, 2007, 10:18pm UTC](https://www.ruby-forum.com/t/ferret-w-acts-as-ferret-on-windows/86777/2 "2007-02-25T22:18:15Z")

</div>

On 2/26/07, neongrau \_\_ [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) wrote:

> hi all!
> 
> after hours of trying to find contents with german umlauts i stumbled  
> upon a post where someone said ferret won’t work with utf-8 on  
> windows???
> 
> is that really true?
> 
> do i really have to iconv everything to iso-8859-15 before indexing and  
> do the same with the query to get it working?

The StandardAnalyzer uses your current locale settings to determine  
what a letter is when tokenizing your data. As far as I was able to  
determine, Windows doesn’t have support for UTF-8 locales in C and the  
win32 libraries. (I’d love for someone to correct me on thise). What  
you can do is write a custom analyzer and UTF-8 should be fine. There  
has been plenty of discussion on creating your own analyzer in the  
past:

```
http://www.ruby-forum.com/search?query=ferret+analyzer&submit=Search

```

You can also look in the unit tests.