Shibboleth SP attributes character encoding

[Peter Schober]

* Robert Führicht <Robert.Fuehricht at jku.at> [2015-09-01 15:47]:
> > Does this help?
> > http://wiki.apache.org/tomcat/FAQ/CharacterEncoding
> 
> Yes, I'm guessing but not in the way you meant it, Peter. Just setting
> the request charset to UTF-8 with a ServletFilter doesn't change a thing.

My (limited) understanding was that Tomcat must be missing some info
as to the actual encoding used, hence the reference to (more specifically):
http://wiki.apache.org/tomcat/FAQ/CharacterEncoding#Q8

One thing I don't understand is why noone else commented on that yet.
AFAIU everyone running e.g. a shibbolized Confluence etc. would run
into this problem?
(I don't have access to our own instance anymore and the people
running it nowadays are still on vacation.)

> So I used a ServletFilter as Peter suggested and Bytebuddy to create 
> a proxy class around the request object and inject it back in the processing
> chain to fix the data when the application asks for it
> (i.e. decorating getAttribute/getHeader).
> 
> It's not pretty, but it works and it's confined to the application
> that needs the fix.

That sounds very much worthy of documentation of some sort.
If you don't have a place to publish that feel free to add somewhere
below to the Contributions page in the Shib wiki.
Cheers,
-peter