Ideally we don't have to reinvent the wheel on this. Sounds as if JDK6 solves 
this in a clean way by
using a reader. While Click itself cannot use this mechanism yet, users could 
plug in a custom
MessageMap that uses the UTF-8 reader.

Bob

On 3/11/2010 03:47, Lorenzo Simionato wrote:
> If in the next release Click will have a pluggable messages Map this could 
> resolve most 
> of the problems, in fact as of now one must modify the click jar!
> If click targets java 1.5 it seems that the only solution is the self made 
> resource bundle (given
> that the hack 
> is not a very clean solution).
> 
> How the situation is handled in a  couple of existing framework:
> -Tapestry:
> properties file encoded in UTF-8 by default.
> They are read using an hack, see http://tinyurl.com/2g5m8zu
> 
> -Wicket:
> properties file read using ResourceBundle (so ISO8859)
> Ability to use read XML based properties in UTF-8, see 
> http://tinyurl.com/37frn4m
> 
> -Grails:
> properties file encoded in UTF-8 by default
> It seems they are read using an hack, see 
> http://www.grails.org/Internationalization
> 
> --
> Lorenzo
> 
> On Nov 2, 2010, at 0:58 , Bob Schellink wrote:
> 
>> Hi Lorenzo,
>>
>> his has been brought up before but I can't recall if a proper solution was 
>> suggested.
>>
>> I know Java 5 introduced XML based properties, but I think key/value 
>> properties are preferred as
>> language packs. Java6 introduced more flexibility around loading bundles, 
>> but I'm not sure how that
>> works, and Click targets Java5.
>>
>> I'm not familiar with how Wicket and Tapestry solved the issue? Did they 
>> reimplement ResourceBundle?
>>
>> The next Click release will have a pluggable messages Map factory to change 
>> the default
>> implementation. Still I'd be interested to know if there is an easy way to 
>> achieve UTF-8 support.
>>
>> There is this hack[1], but hacks make me nervous.
>>
>> Kind regards
>>
>> Bob
>>
>> [1]: http://www.thoughtsabout.net/blog/archives/000044.html
>>
> 

Reply via email to