krb5 commit [krb5-1.12]: Use system dictionary for db2 tests again

[Tom Yu]

https://github.com/krb5/krb5/commit/dc6e116f341300fb79a5e1aba76f884d01ef378d
commit dc6e116f341300fb79a5e1aba76f884d01ef378d
Author: Greg Hudson <ghudson at mit.edu>
Date:   Tue Feb 18 01:14:01 2014 -0500

    Use system dictionary for db2 tests again
    
    The built-in word list is not long enough for all of the libdb2 tests
    to run properly.  Revert d21a86e47a7cda29225013e08d060095b94b2ee7 and
    go back to using the system dictionary if we find one.  However, omit
    any lines from the chosen word list which contain non-alphabetical
    characters.
    
    (cherry picked from commit dba768e873d3ae34cfb2ff9d9c2d3644981f23a5)
    
    ticket: 7860
    version_fixed: 1.12.2
    status: resolved

 src/plugins/kdb/db2/libdb2/test/run.test |    9 +++++----
 1 files changed, 5 insertions(+), 4 deletions(-)

diff --git a/src/plugins/kdb/db2/libdb2/test/run.test b/src/plugins/kdb/db2/libdb2/test/run.test
index 7d9f0fe..d99b42d 100644
--- a/src/plugins/kdb/db2/libdb2/test/run.test
+++ b/src/plugins/kdb/db2/libdb2/test/run.test
@@ -71,10 +71,11 @@ main()
 }
 
 getnwords() {
-	# Delete blank lines because the db code appears not to
-	# like empty keys.  On Debian Linux, $DICT appears to contain
-	# some non-ASCII characters, and "rev" chokes on them.
-	sed -e '/^$/d' < $DICT | cat -v | sed -e ${1}q
+	# Delete blank lines because the db code appears not to like
+	# empty keys.  Omit lines with non-alphanumeric characters to
+	# avoid shell metacharacters and non-ASCII characters which
+	# could cause 'rev' to choke.
+	LC_ALL=C sed -e '/^$/d' -e '/[^A-Za-z]/d' < $DICT | sed -e ${1}q
 }
 
 # Take the first hundred entries in the dictionary, and make them