Fixed #4664 -- Forced the client character set encoding to UTF-8 for PostgreSQL

(via the psycopg backend). The previous version was causing problems on some
setups, particularly PostgreSQL 7.x. Current code should work with 7.x and 8.x,
no matter what the default client encoding is.


git-svn-id: http://code.djangoproject.com/svn/django/branches/unicode@5535 bcc190cf-cafb-0310-a4f2-bffc1f526a37

2 files changed, 4 insertions, 60 deletions

diff --git a/django/db/backends/postgresql/base.py b/django/db/backends/postgresql/base.py
index ea19f65839..e2051beaa2 100644
--- a/django/db/backends/postgresql/base.py
+++ b/django/db/backends/postgresql/base.py
@@ -6,7 +6,6 @@ Requires psycopg 1: http://initd.org/projects/psycopg1
 
 from django.utils.encoding import smart_str, smart_unicode
 from django.db.backends import util
-from django.db.backends.postgresql.encodings import ENCODING_MAP
 try:
     import psycopg as Database
 except ImportError, e:
@@ -64,7 +63,6 @@ class UnicodeCursorWrapper(object):
             return getattr(self.cursor, attr)
 
 postgres_version = None
-client_encoding = None
 
 class DatabaseWrapper(local):
     def __init__(self, **kwargs):
@@ -94,15 +92,8 @@ class DatabaseWrapper(local):
         cursor = self.connection.cursor()
         if set_tz:
             cursor.execute("SET TIME ZONE %s", [settings.TIME_ZONE])
-        cursor.execute("SHOW client_encoding")
-        encoding = ENCODING_MAP[cursor.fetchone()[0]]
-        cursor = UnicodeCursorWrapper(cursor, encoding)
-        global client_encoding
-        if not client_encoding:
-            # We assume the client encoding isn't going to change for random
-            # reasons.
-            Database.register_type(Database.new_type(Database.types[1043].values, 'STRING', typecast_string))
-            client_encoding = encoding
+        cursor.execute("SET client_encoding to 'UNICODE'")
+        cursor = UnicodeCursorWrapper(cursor, 'utf-8')
         global postgres_version
         if not postgres_version:
             cursor.execute("SELECT version()")
@@ -289,7 +280,7 @@ def typecast_string(s):
     """
     if not s:
         return s
-    return smart_unicode(s, client_encoding)
+    return smart_unicode(s)
 
 # Register these custom typecasts, because Django expects dates/times to be
 # in Python's native (standard-library) datetime/time format, whereas psycopg
@@ -302,6 +293,7 @@ Database.register_type(Database.new_type((1083,1266), "TIME", util.typecast_time
 Database.register_type(Database.new_type((1114,1184), "TIMESTAMP", util.typecast_timestamp))
 Database.register_type(Database.new_type((16,), "BOOLEAN", util.typecast_boolean))
 Database.register_type(Database.new_type((1700,), "NUMERIC", util.typecast_decimal))
+Database.register_type(Database.new_type(Database.types[1043].values, 'STRING', typecast_string))
 
 OPERATOR_MAPPING = {
     'exact': '= %s',
diff --git a/django/db/backends/postgresql/encodings.py b/django/db/backends/postgresql/encodings.py
deleted file mode 100644
index 934cc6e6de..0000000000
--- a/django/db/backends/postgresql/encodings.py
+++ /dev/null
@@ -1,48 +0,0 @@
-# Mapping between PostgreSQL encodings and Python codec names. This mapping
-# doesn't exist in psycopg, so we have to maintain it by hand (using
-# information from section 21.2.1 in the PostgreSQL manual).
-ENCODING_MAP = {
-    "ALT": 'cp866',
-    "BIG5": 'big5-tw',
-    "EUC_CN": 'gb2312',
-    "EUC_JP": 'euc_jp',
-    "EUC_KR": 'euc_kr',
-    "GB18030": 'gb18030',
-    "GBK": 'gbk',
-    "ISO_8859_5": 'iso8859_5',
-    "ISO_8859_6": 'iso8859_6',
-    "ISO_8859_7": 'iso8859_7',
-    "ISO_8859_8": 'iso8859_8',
-    "JOHAB": 'johab',
-    "KOI8": 'koi18_r',
-    "KOI18R": 'koi18_r',
-    "LATIN1": 'latin_1',
-    "LATIN2": 'iso8859_2',
-    "LATIN3": 'iso8859_3',
-    "LATIN4": 'iso8859_4',
-    "LATIN5": 'iso8859_9',
-    "LATIN6": 'iso8859_10',
-    "LATIN7": 'iso8859_13',
-    "LATIN8": 'iso8859_14',
-    "LATIN9": 'iso8859_15',
-    "SJIS": 'shift_jis',
-    "SQL_ASCII": 'ascii',
-    "TCVN": 'cp1258',
-    "UHC": 'cp949',
-    "UNICODE": 'utf-8',
-    "UTF8": 'utf-8',
-    "WIN": 'cp1251',
-    "WIN866": 'cp866',
-    "WIN874": 'cp874',
-    "WIN1250": 'cp1250',
-    "WIN1251": 'cp1251',
-    "WIN1252": 'cp1252',
-    "WIN1256": 'cp1256',
-    "WIN1258": 'cp1258',
-
-    # Unsupported (no equivalents in codecs module):
-    # EUC_TW
-    # LATIN10
-    # MULE_INTERNAL
-}
-