Optimize UTF8 and Latin1 decoding

Cache a pointer to a fast decoding function when the connection encoding
is set so skip a repeated codec lookup for every string.

3 files changed, 40 insertions, 3 deletions

diff --git a/psycopg/connection.h b/psycopg/connection.h
index ec10742..b925bd4 100644
--- a/psycopg/connection.h
+++ b/psycopg/connection.h
@@ -122,6 +122,9 @@ struct connectionObject {
     int autocommit;
 
     PyObject *cursor_factory;    /* default cursor factory from cursor() */
+
+    /* Pointer to a decoding function, e.g. PyUnicode_DecodeUTF8 */
+    PyObject *(*cdecoder)(const char *, Py_ssize_t, const char *);
 };
 
 /* map isolation level values into a numeric const */
diff --git a/psycopg/connection_int.c b/psycopg/connection_int.c
index 43d0fda..62976d4 100644
--- a/psycopg/connection_int.c
+++ b/psycopg/connection_int.c
@@ -361,6 +361,31 @@ exit:
     return rv;
 }
 
+
+/* set fast access functions according to the currently selected codec
+ */
+void
+conn_set_fast_codec(connectionObject *self)
+{
+    Dprintf("conn_set_fast_codec: codec=%s", self->codec);
+
+    if (0 == strcmp(self->codec, "utf_8")) {
+        Dprintf("conn_set_fast_codec: PyUnicode_DecodeUTF8");
+        self->cdecoder = PyUnicode_DecodeUTF8;
+        return;
+    }
+
+    if (0 == strcmp(self->codec, "iso8859_1")) {
+        Dprintf("conn_set_fast_codec: PyUnicode_DecodeLatin1");
+        self->cdecoder = PyUnicode_DecodeLatin1;
+        return;
+    }
+
+    Dprintf("conn_set_fast_codec: no fast codec");
+    self->cdecoder = NULL;
+}
+
+
 /* Read the client encoding from the connection.
  *
  * Store the encoding in the pgconn->encoding field and the name of the
@@ -402,6 +427,8 @@ conn_read_encoding(connectionObject *self, PGconn *pgconn)
     self->codec = codec;
     codec = NULL;
 
+    conn_set_fast_codec(self);
+
     rv = 0;
 
 exit:
@@ -1243,6 +1270,8 @@ conn_set_client_encoding(connectionObject *self, const char *enc)
         codec = NULL;
     }
 
+    conn_set_fast_codec(self);
+
     Dprintf("conn_set_client_encoding: set encoding to %s (codec: %s)",
             self->encoding, self->codec);
 
diff --git a/psycopg/typecast_basic.c b/psycopg/typecast_basic.c
index a31047f..760555e 100644
--- a/psycopg/typecast_basic.c
+++ b/psycopg/typecast_basic.c
@@ -93,12 +93,17 @@ typecast_STRING_cast(const char *s, Py_ssize_t len, PyObject *curs)
 static PyObject *
 typecast_UNICODE_cast(const char *s, Py_ssize_t len, PyObject *curs)
 {
-    char *enc;
+    connectionObject *conn;
 
     if (s == NULL) { Py_RETURN_NONE; }
 
-    enc = ((cursorObject*)curs)->conn->codec;
-    return PyUnicode_Decode(s, len, enc, NULL);
+    conn = ((cursorObject*)curs)->conn;
+    if (conn->cdecoder) {
+        return conn->cdecoder(s, len, NULL);
+    }
+    else {
+        return PyUnicode_Decode(s, len, conn->codec, NULL);
+    }
 }
 
 /** BOOLEAN - cast boolean value into right python object **/