Fix csvfile traceback on Python3 (#37625)

* Fix csvfile traceback on Python3

The csvfile lookup uses some custom iterators.  These needed to be
ported to handle the python3 iterator protocol.  In addition, the
csvfile module takes an iterator of byte strings in Python2 and an
iterator of text strings in Python3

Fixes #36808

lib/ansible/plugins/lookup/csvfile.py

@@ -53,6 +53,7 @@ from collections import MutableSequence
 
 from ansible.errors import AnsibleError, AnsibleAssertionError
 from ansible.plugins.lookup import LookupBase
+from ansible.module_utils.six import PY2
 from ansible.module_utils._text import to_bytes, to_native, to_text
 
 
@@ -66,8 +67,10 @@ class CSVRecoder:
     def __iter__(self):
         return self
 
-    def next(self):
-        return self.reader.next().encode("utf-8")
+    def __next__(self):
+        return next(self.reader).encode("utf-8")
+
+    next = __next__   # For Python 2
 
 
 class CSVReader:
@@ -77,13 +80,19 @@ class CSVReader:
     """
 
     def __init__(self, f, dialect=csv.excel, encoding='utf-8', **kwds):
-        f = CSVRecoder(f, encoding)
+        if PY2:
+            f = CSVRecoder(f, encoding)
+        else:
+            f = codecs.getreader(encoding)(f)
+
         self.reader = csv.reader(f, dialect=dialect, **kwds)
 
-    def next(self):
-        row = self.reader.next()
+    def __next__(self):
+        row = next(self.reader)
         return [to_text(s) for s in row]
 
+    next = __next__  # For Python 2
+
     def __iter__(self):
         return self
 
@@ -93,8 +102,8 @@ class LookupModule(LookupBase):
     def read_csv(self, filename, key, delimiter, encoding='utf-8', dflt=None, col=1):
 
         try:
-            f = open(filename, 'r')
-            creader = CSVReader(f, delimiter=to_bytes(delimiter), encoding=encoding)
+            f = open(filename, 'rb')
+            creader = CSVReader(f, delimiter=to_native(delimiter), encoding=encoding)
 
             for row in creader:
                 if row[0] == key: