Commit Graph

942 Commits

Author SHA1 Message Date
Daniele Varrazzo
74d2c4bef9 Fixed idempotence check changing connection characteristics 2018-01-11 02:23:27 +00:00
Daniele Varrazzo
e0226fc46a 'cursor.mogrify()' can be called on closed cursors
Fix #579.
2018-01-11 08:55:31 +09:00
Daniele Varrazzo
bad9b8b383 Merge branch 'macro-accessors' 2018-01-10 22:47:59 +00:00
Daniele Varrazzo
77c703395b Moved datatime compatibility macros with others 2018-01-10 22:29:16 +00:00
Ronan Amicel
b8f2f71819
Fix typo in comment in utils.c 2017-12-14 14:00:36 +01:00
Glyph
3fcb035126 define a "polyfill" inline for python 2 compatibility 2017-12-12 03:15:21 -08:00
Glyph
852884e086 use accessor macros for pypy3 compatibility 2017-12-12 03:06:18 -08:00
Jon Dufresne
e335d6d223 Trim trailing whitespace from all files throughout project
Many editors automatically trim whitespace on save. By trimming all
 files in one go, makes future diffs cleaner without extraneous
 whitespace changes.
2017-12-01 21:42:14 -08:00
Daniele Varrazzo
65ed5478d1 Collect rowcount in executemany even when discarding results
Closes #633.
2017-11-29 15:42:04 +00:00
Hugo
60b1517c55 Add news and update version check 2017-11-28 16:11:02 +00:00
Hugo
ffcc65d4f0 Drop support for EOL Python 2.6 2017-11-28 16:02:12 +00:00
Daniele Varrazzo
1959d2403f Merge branch 'solaris-support' 2017-11-28 03:16:04 +00:00
Federico Di Gregorio
5983b96c55 Don't cast point arrays to float arrays (fixes: #613) 2017-11-16 10:07:27 +01:00
My Karlsson
f54783ae6e Emulate timeradd and timersub on Solaris
Solaris does not have timeradd and timersub. Add solaris_support.c which
provides emulated versions of them on Solaris.
2017-10-26 20:21:30 +02:00
Daniele Varrazzo
775de754cf Added back timestamptz[] default cast
It was registered as side effect of an excessive definition that got
cleaned up in 338dbe70a6.

Looking at other removed redundant type oids, this was the only one
missing from the `string_types` map.

Close #578.
2017-07-24 14:23:36 +01:00
Daniele Varrazzo
1a97445471 Accept Composable in start_replication_expert()
Close #554
2017-06-17 03:30:44 +01:00
Daniele Varrazzo
49ce622a86 Fixed parsing interval from micros on 32 bit
Using integers the wrong size. Faithfully segfaulting since 1970.
2017-06-17 01:36:39 +01:00
Daniele Varrazzo
315f72862c Parse a number as microseconds when casting interval
Should close #558, but I'm curious to know if a number is returned
for interval < 1 day too (which wouldn't trigger the overflow, but will
finish parsing with part=0).
2017-06-16 00:54:38 +01:00
Daniele Varrazzo
338dbe70a6 Dropped wrong associations of PG types to datetime object
The wrong associations are overwritten in the typecaster map by the
right ones, so they have been harmless to date, but only because of the
order of creation of the adapters.
2017-06-15 18:24:25 +01:00
Daniele Varrazzo
602c74faa6 Check return code from decoding the connection status
It fails on Py3 after receiving a SIGABRT. Because we don't handle it
here it will resurface later with nonsense such as:

    SystemError: <some function> returned a result with an error set

Close #551
2017-06-15 17:33:05 +01:00
Daniele Varrazzo
6e5abf33f2 Merge branch 'fix-547' 2017-04-19 01:34:39 +01:00
Daniele Varrazzo
a7e3f46431 Merge remote-tracking branch 'fix_lobject_factory' 2017-04-19 01:06:24 +01:00
Daniele Varrazzo
248e653c9e Fixed args parsing in ReplicationCursor.consume_stream()
Close #547.
2017-04-19 01:01:59 +01:00
Frazer McLean
7b3ea43e92 Handle lobject mode=None correctly 2017-04-16 03:20:31 +02:00
Frazer McLean
38cd720369 Fix name of lobject keyword argument 2017-04-16 03:12:18 +02:00
Daniele Varrazzo
a66c34a6d0 Don't clobber a Python exception with an unknown error
Close #410
2017-04-05 14:54:07 +01:00
Daniele Varrazzo
7214c6652e Return objects with timezone parsing infinity timestamptz
Close #536.
2017-03-22 03:03:02 +00:00
Daniele Varrazzo
ba0329fb40 replication connection init refactored to use psyco_make_dsn
Some extra bonus refactoring to improve the function readability (don't
reuse names for variables with different refcount rules, don't pass
separate obj/self, async pass-through...)
2017-03-16 03:55:22 +00:00
Daniele Varrazzo
9f160fd820 Obscure the password on url dsn too
Note that we don't leak anymore the password length.

Fix #528
2017-03-16 03:53:40 +00:00
Daniele Varrazzo
103655d670 Password scrubbing refactored in a separate function 2017-03-15 16:04:45 +00:00
Greg Ward
12317557da Always raise OperationalError when connection was closed externally.
From the DB-API (https://www.python.org/dev/peps/pep-0249/):

  OperationalError

  Exception raised for errors that are related to the database's
  operation and not necessarily under the control of the programmer,
  e.g. an unexpected disconnect occurs, [...]

Additionally, psycopg2 was inconsistent, at least in the async case:
depending on how the "connection closed" error was reported from the
kernel to libpq, it would sometimes raise OperationalError and
sometimes DatabaseError. Now it always raises OperationalError.
2017-03-14 12:14:00 +00:00
Greg Ward
b203a7c775 Always detect when a connection is closed behind psycopg2's back.
There's a race condition that only seems to happen over Unix-domain
sockets. Sometimes, the closed socket is reported by the kernel to
libpq like this (captured with strace):

  sendto(3, "Q\0\0\0\34select pg_backend_pid()\0", 29, MSG_NOSIGNAL, NULL, 0) = 29
  recvfrom(3, "E\0\0\0mSFATAL\0C57P01\0Mterminating "..., 16384, 0, NULL, NULL) = 110
  recvfrom(3, 0x12d0330, 16384, 0, 0, 0)  = -1 ECONNRESET (Connection reset by peer)

That is, psycopg2/libpq sees no error when sending the first query
after the connection is closed, but gets an error reading the result.
In that case, everything worked fine.

But sometimes, the error manifests like this:

  sendto(3, "Q\0\0\0\34select pg_backend_pid()\0", 29, MSG_NOSIGNAL, NULL, 0) = -1 EPIPE (Broken pipe)
  recvfrom(3, "E\0\0\0mSFATAL\0C57P01\0Mterminating "..., 16384, 0, NULL, NULL) = 110
  recvfrom(3, "", 16274, 0, NULL, NULL)   = 0
  recvfrom(3, "", 16274, 0, NULL, NULL)   = 0

i.e. libpq received an error when sending the query. This manifests as
a slightly different exception from a slightly different place. More
importantly, in this case connection.closed is left at 0 rather than
being set to 2, and that is the bug I'm fixing here.

Note that we see almost identical behaviour for sync and async
connections, and the fixes are the same. So I added extremely similar
test cases.

Finally, there is still a bug here: for async connections, we
sometimes raise DatabaseError (incorrect) and sometimes raise
OperationalError (correct). Will fix that next.
2017-03-14 12:08:03 +00:00
Daniele Varrazzo
26bd5df048 Don't use PG_INT64_TYPE
Only defined in Postgres 9.3

Should close #520, but let's wait for the CI response (build on
Windows etc.)
2017-03-02 19:32:39 +00:00
Daniele Varrazzo
72f74d1f41 Dropped compiler warning on Windows 2017-02-24 13:33:42 +00:00
Daniele Varrazzo
691df4952b Handle overflow in interval parsing 2017-02-24 13:23:56 +00:00
Daniele Varrazzo
3b665d35d5 Make sure to use 64 bits in interval parsing accumulators 2017-02-24 12:10:54 +00:00
Daniele Varrazzo
834e9996da Parse interval only using integers
(almost... except for micros rounding)

While this is probably an improvement on the previous implementation,
I am largely waving a dead chicken at windows, which keeps failing to
pass the seconds overflow test. If it doesn't pass now either I'll start
blaming Python's timedelta.
2017-02-24 03:48:41 +01:00
Daniele Varrazzo
a15dfbbd4f Try to fix the same problem in windows
Previous commit doesn't pass on Windows: it looks like window's floor()
has an integer overflow.
2017-02-24 02:41:34 +01:00
Daniele Varrazzo
14fe3ad8c9 Fixed integer overflow in interval typecaster
Close #512.
2017-02-24 02:10:27 +01:00
Daniele Varrazzo
61101888e4 Revert default_transaction_* to default only if set
When moving from autocommit True -> False reset only the server
parameters that were actually specified by psycopg to honour the
serssion characteristics.
2017-02-16 13:15:14 +00:00
Daniele Varrazzo
d7bba865f3 Check for deferrable unsupported applied to attribute too 2017-02-16 13:14:45 +00:00
Daniele Varrazzo
d50ed48807 Added readonly and deferrable attributes 2017-02-16 12:46:35 +00:00
Daniele Varrazzo
b5d80b609d Revert pre-2.7b1 behaviour of silent rollback on conn.set_isolation_level()
Legacy method is legacy.
2017-02-16 12:46:35 +00:00
Daniele Varrazzo
20c9c17457 connection.isolation_level is now writable 2017-02-16 12:46:35 +00:00
Jason Erickson
32e81a27b5 Allowed inittype.h's defs all versions of MSVC
Apparently only MSVC 2015 x64 had defined these types.  Changed check to
work for all versions of MSVC.  Does not affect 2015x64.
2017-02-10 10:14:17 -07:00
Jason Erickson
7493ea24f2 Defined inttypes.h for older MSVC compilers
Older MSVC compilers do not support the C99 standard
2017-02-09 22:27:18 -07:00
Jason Erickson
d52e9100d2 Added support for MSVC compiler
Added function 'timeradd'.
Changed second parameter of 'gettimeofday' to void since not used in
function and MSVC timezone definition is not a struct).
2017-02-09 20:57:54 -07:00
Jason Erickson
83cf908c88 Fixed MSVC compiler error and warning
MSVC doesn't like referencing new initialized variables in the variable
declaration section.
2017-02-09 20:54:29 -07:00
Daniele Varrazzo
d2cd1236a8 Dropped compiler warning about signed/unsigned comparisons 2017-02-08 08:44:08 +00:00
Daniele Varrazzo
3e12522bc9 Don't convert '{}'::unknown into an empty list
Close #506.
2017-02-08 08:43:53 +00:00