# Utf8 -\> latin2

**URL:** <https://rubytalk.org/t/utf8-latin2/8801>\
**Category:** ruby-talk\
**Created:** [14 November 2003 14:46 UTC](https://rubytalk.org/t/utf8-latin2/8801 "2003-11-14T14:46:07Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![Bedo\_Sandor1](https://avatars.discourse-cdn.com/v4/letter/b/8491ac/32.png) [@Bedo\_Sandor1](https://rubytalk.org/u/Bedo_Sandor1)\
**Post date:** [14 November 2003 14:46 UTC](https://rubytalk.org/t/utf8-latin2/8801/1 "2003-11-14T14:46:07Z")

</div>

Hi,

How can I convert utf-8 encoded strings to latin-2?  
I have tried it using libuconv with little success:

require ‘uconv’

class String  
def un\_utf8  
Uconv.u8tou16(self).gsub(/\000/, ‘’)  
end

```
    def to_utf8
            tmp = ""
            self.each_byte { |b|
                    tmp += b.chr + "\000"
            }
            Uconv.u16tou8(tmp)
    end

```

end

This program is ugly, and does not exactly what I want.  
u8tou16 generates a string with 16 bit long characters,  
for example “test”.un\_utf8 == “t\000e\000s\000t\000”.  
gsub clears the unnecessery “\000” characters from  
the string. But there are characters in Hungarian,  
that has non-zero second byte in the output of the  
u8tou16, so they fail to convert. Anyway this is an  
ugly hack.

How is it done nicely?

> **···**
>
> –  
> bSanyI

---

<div class="post-metadata">

**Author:** ![Wesley\_J\_Landaker](https://avatars.discourse-cdn.com/v4/letter/w/a698b9/32.png) [@Wesley\_J\_Landaker](https://rubytalk.org/u/Wesley_J_Landaker)\
**Post date:** [14 November 2003 15:31 UTC](https://rubytalk.org/t/utf8-latin2/8801/2 "2003-11-14T15:31:10Z")

</div>

I think the iconv module handles this nicely:

require ‘iconv’  
Iconv.conv(“utf-8”,“latin2”,“this is a test”)

* * *

Wesley J. Landaker - [wjl@icecavern.net](mailto:wjl@icecavern.net)  
OpenPGP FP: 4135 2A3B 4726 ACC5 9094 0097 F0A9 8A4C 4CD6 E3D2

> **···**
>
> On Friday 14 November 2003 7:46 am, Bedo Sandor wrote:
> 
> > Hi,
> > 
> > How can I convert utf-8 encoded strings to latin-2?  
> > I have tried it using libuconv with little success:
> > 
> > require ‘uconv’
> > 
> > class String  
> > def un\_utf8  
> > Uconv.u8tou16(self).gsub(/\000/, ‘’)  
> > end
> > 
> > ```
> > def to_utf8
> > tmp = ""
> > self.each_byte { |b|
> > tmp += b.chr + "\000"
> > }
> > Uconv.u16tou8(tmp)
> > end
> > 
> > ```
> > 
> > end
> > 
> > This program is ugly, and does not exactly what I want.  
> > u8tou16 generates a string with 16 bit long characters,  
> > for example “test”.un\_utf8 == “t\000e\000s\000t\000”.  
> > gsub clears the unnecessery “\000” characters from  
> > the string. But there are characters in Hungarian,  
> > that has non-zero second byte in the output of the  
> > u8tou16, so they fail to convert. Anyway this is an  
> > ugly hack.
> > 
> > How is it done nicely?