Description of problem:
With at least LANG=cs_CZ.UTF-8 and LANG=en_US.UTF-8 local -a output contains data that can't be decoded as utf-8. This breaks grep an other tools (the output for example can't be decoded by the Python utf-8 codec).


Version-Release number of selected component (if applicable):
glibc-common-2.20-7.fc21.x86_64

How reproducible:
always

Steps to Reproduce:
1. LANG=en_US.UTF-8
2. locale -a | grep bok

Actual results:
Grep returns only:
Binary file (standard input) matches

Not the actual line matching the input string.


Expected results:
Grep returns the lines matching the input string.


Additional info:
If LANG=c is set, grep returns the matching lines:

export LANG=c
locale -a | grep bok
bokmal
bokm�l

Also the offending strings that contains the non-utf-8 binary data seem to be bokm�l and fran�ais.