The sr_RS locale in glibc corresponds to the Cyrillic script, while the 
agreed upon locale for the Latin alphabet is sr_RS@latin. Unfortunately, 
the locale python module crashes when trying to parse this locale. Here 
is the traceback:

File "/usr/lib/python2.6/locale.py", line 497, in get locale
  return _parse_localename(localename)
File "usr/lib/python2.6/locale.py", line 410, in _parse_localename
  raise ValueError, 'unknown locale: %s' % localename

Looking at the code, it only checks for the @euro modifier and ignores 
all other modifiers (including @latin). This is all when I set 
LANG=sr_RS@latin. If I use LANG=sr_RS.UTF8@latin (also allowed by 
glibc), Python interprets this as the sr_RS locale (which is wrong, as 
that is for Cyrillic).