Skip to content

Commit b87b135

Browse files
author
James William Pye
committed
Only list changes in the branch.
1 parent 1b8ed13 commit b87b135

2 files changed

Lines changed: 6 additions & 109 deletions

File tree

postgresql/documentation/changes.txt

Lines changed: 0 additions & 104 deletions
Original file line numberDiff line numberDiff line change
@@ -19,107 +19,3 @@ Changes
1919
* Increased default recvsize and chunksize for improved performance.
2020
* 'D' messages are special cased as builtins.tuples instead of
2121
protocol.element3.Tuple
22-
23-
0.9.3 released on 2010-01-01
24-
----------------------------
25-
26-
* Fix inconsistency in savepoint identification by pq3.Transaction.
27-
Savepoints via db.xact() were broken.
28-
[Reported by Elvis Pranskevichus]
29-
30-
0.9.2 released on 2009-12-14
31-
----------------------------
32-
33-
* Alter protocol message type comparison method.
34-
In 0.9.1 and before, types were compared using 'is' because all
35-
message type objects were loaded from a tuple created by the
36-
postgresql.protocol.message_types module. In some environments,
37-
sys.modules is cleared. For pure-Python installations, this is unlikely
38-
to cause a problem, but when the optimized module is available, it will hold
39-
a reference to the "old" version of the message_types data. Subsequently, the
40-
use of the 'is' operator is then broken resulting in superfluous protocol
41-
errors. The fix is to use the '==' operator. This does not add any
42-
significant overhead, but use of 'is' may return if some permanent storage
43-
method becomes available. [Reported by Radomir Stevanovic]
44-
* Fix a bug revealed by a no route to host connection failure. Ended in an
45-
ambiguous KeyError due there being no message string for that particular
46-
errno. Add an appropriate message string, and fallback on a default message
47-
when a specific one doesn't exist.
48-
* Correct types.Array binary serialization (8.5 added extra checks)
49-
* Add more encoding name mapping entries.
50-
* Documentation improvements.
51-
* Fix slice() support in Row.__getitem__.
52-
* DEPRECATION: 2PC interfaces are being deprecated. Warnings in 1.0,
53-
removal in 1.1.
54-
55-
0.9.1 released on 2009-08-12
56-
----------------------------
57-
58-
* Fix misaligned reads in C extensions [Reported by Leonardo Francalanci]
59-
* Don't be strict about keys(field type) in Notice messages per specification:
60-
http://www.postgresql.org/docs/8.4/static/protocol-message-formats.html
61-
* Fix test_protocol on ubuntu linux systems.
62-
* Stop waiting for cluster startup on postgresql.exceptions.ProtocolError.
63-
* DEAD BRANCH: 0.8 is no longer supported.
64-
65-
0.9.0 released on 2009-06-13
66-
----------------------------
67-
68-
* Performance improvements.
69-
* Extend `postgresql.open` to take keyword arguments.
70-
* Implement query libraries and connection categories[#1010581 & #1010618].
71-
* Implement ``clone()`` on Connections, Statements, and Cursors.
72-
* Display the line and mark the location of the POSITION for syntax errors.
73-
* Refactor ``ps.load()`` into ``ps.load_chunks()`` and ``ps.load_rows()``.
74-
* Refactor `postgresql.api.InterfaceElement`.
75-
* Refactor driver.pq3.Connection to use protocol.client3.Connection.
76-
* Refactor driver.pq3.Cursor into types selected by PreparedStatements.
77-
* Fix memory leak due to circular references and __del__. [Reported by Valentine Gogichashvili]
78-
Additionally, try to avoid circular references at that level.
79-
* Correct StoredProcedure's statement production. It was falsely using a
80-
a mechanism that would use SQL predefined type names instead of
81-
qualified names. [Reported by Dallas Morisett]
82-
* Make DB-API connect() inherit defaults like postgresql.open()
83-
* Make DB-API extend PG-API so that DB-API connections will now have direct
84-
access to PG-API features.
85-
* Fix DB-API setting of rowcount after execute. [Reported by Mike Bayer; 1010643]
86-
* Document msghook attributes, and use postgresql.sys.errformat for
87-
postgresql.exceptions.Error.__str__.
88-
89-
0.8.2 released on 2009-06-13
90-
----------------------------
91-
92-
* Correct StoredProcedure's statement production. It was falsely using a
93-
a mechanism that would use SQL predefined type names instead of
94-
qualified names. [Reported by Dallas Morisett]
95-
* Fix DB-API setting of rowcount after execute. [Reported by Mike Bayer; 1010643]
96-
97-
0.8.1 released on 2009-04-30
98-
----------------------------
99-
100-
* Fix memory leak due to circular references and __del__. [Reported by Valentine Gogichashvili]
101-
* Fix encoding normalization. [Reported by Marc Silver]
102-
* Fix "method" decorator to return callable when val is None.
103-
* Fix startup of clusters in silent mode.
104-
105-
0.8.0 released on 2009-04-03
106-
----------------------------
107-
108-
Differences from the original pg/python projects.
109-
110-
* Integrate projects into a single package. Remove "codenames".
111-
* pg_greentrunk evolved into PG-API, postgresql.api.
112-
* Add new execution methods to statement objects: chunks, rows, declare.
113-
* Refactor connection negotiation code to use a generator. (xact3.Negotiation)
114-
* Remove sixbit confusion from postgresql.exceptions.
115-
* Integrate fcrypt module for crypt authentication.
116-
* Remove pytz dependency, always use UTC for timestamptz.
117-
* Remove netaddr dependency.
118-
* Improve DB-API interface by allowing subjective paramstyle. (psycopg2 compat)
119-
* Provide autocommit support on DB-API connections.
120-
* Add support for binary numeric.
121-
* Add proper support for sslmode and connect_timeout.
122-
* Conditionally DECLARE certain cursors WITH HOLD when outside of blocks.
123-
* Add typed support for fetches from db.cursor_from_id().
124-
* Change the test infrastructure to automatically create a cluster.
125-
* More improvements than I can remember.

postgresql/driver/pq3.py

Lines changed: 6 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -271,11 +271,12 @@ def _process_tuple_chunk_Row(self, x,
271271
proc = process_chunk,
272272
from_seq = Row.from_sequence,
273273
):
274-
attmap = self._output_attmap
275-
return [
276-
from_seq(attmap, y)
277-
for y in proc(self._output_io, x, self._raise_column_tuple_error)
278-
]
274+
return proc(self._output_io, x, self._raise_column_tuple_error)
275+
#attmap = self._output_attmap
276+
#return [
277+
# from_seq(attmap, y)
278+
# for y in proc(self._output_io, x, self._raise_column_tuple_error)
279+
#]
279280

280281
# Process the elemnt.Tuple messages in `x` for chunks()
281282
def _process_tuple_chunk(self, x, proc = process_chunk):

0 commit comments

Comments
 (0)