[Dovecot] Full text search improvements

[Michael M Slusarz]

Quoting Timo Sirainen <tss at iki.fi>:

> 1. Support for multiple languages. Use textcat while indexing to  
> guess the language of the indexed data.

FWIW, you could probably use the Content-Language header (if it  
exists) to at least give a hint.  No guarantee it is correct, but it's  
a better starting place than simply scanning all languages.

And, for that matter, you could leverage Accept-Language also (again,  
if it exists).  Which might be more useful, since it lists all the  
languages the user recognizes.

michael