Build: use standard infrastructure for translations
    
    We now use standard framework generatedby "gettextize" utility.
    
    It has two limitations which I do not consider sufficiently important
    to invest into hand-made solution:
    
    1. It can automatically gather strings only from files which have some
       file extension like .c or .py. Right now we do not have any
       translatable strings in Python files without extensions. Given that these
       files will be removed from source tree and replaced with entry points
       from setuptools I do not see a reason to invest into supporting this.
    
    2. It does not automatically strip untranslated strings from po files.
       This is a manual step in mainteiner's in workflow anyway so I will
       add separate Makefile target for it later on.
    
    This commit contains gettextize instrastructure + filled-in files
    Makevars and POTFILES.in.
    
    https://fedorahosted.org/freeipa/ticket/6418
    
    Reviewed-By: Martin Basti <mbasti@redhat.com>
    Reviewed-By: Christian Heimes <cheimes@redhat.com>
    Reviewed-By: Lukas Slebodnik <lslebodn@redhat.com>