]> git.ipfire.org Git - thirdparty/sqlalchemy/sqlalchemy.git/log
thirdparty/sqlalchemy/sqlalchemy.git
12 years ago- move to git dotfiles
Mike Bayer [Sat, 25 May 2013 20:20:59 +0000 (16:20 -0400)] 
- move to git dotfiles

12 years agoFixed a bug where the routine to detect the correct kwargs
Mike Bayer [Thu, 23 May 2013 16:59:53 +0000 (12:59 -0400)] 
Fixed a bug where the routine to detect the correct kwargs
being sent to :func:`.create_engine` would fail in some cases,
such as with the Sybase dialect.
[ticket:2732]

12 years agoRegression from this ticket caused the unsupported keyword
Mike Bayer [Wed, 15 May 2013 19:46:29 +0000 (15:46 -0400)] 
Regression from this ticket caused the unsupported keyword
"true" to render, added logic to convert this to 1/0
for SQL server.
[ticket:2682]

12 years agoFixed a regression from 0.7 caused by this ticket, which
Mike Bayer [Mon, 13 May 2013 20:13:15 +0000 (16:13 -0400)] 
Fixed a regression from 0.7 caused by this ticket, which
made the check for recursion overflow in self-referential
eager joining too loose, missing a particular circumstance
where a subclass had lazy="joined" or "subquery" configured
and the load was a "with_polymorphic" against the base.
[ticket:2481]

12 years ago- moderinzed the docs for the foreign_keys parameter a bit,
Mike Bayer [Fri, 10 May 2013 05:06:10 +0000 (01:06 -0400)] 
- moderinzed the docs for the foreign_keys parameter a bit,
[ticket:2725]

12 years ago- cleanup
Mike Bayer [Sun, 5 May 2013 05:39:35 +0000 (01:39 -0400)] 
- cleanup
- move the timeout here to 14 seconds as jenkins still chokes occasionally

12 years agocleanup
Mike Bayer [Sat, 4 May 2013 20:26:10 +0000 (16:26 -0400)] 
cleanup

12 years agocleanup
Mike Bayer [Sat, 4 May 2013 19:35:47 +0000 (15:35 -0400)] 
cleanup

12 years agocleanup
Mike Bayer [Sat, 4 May 2013 19:08:08 +0000 (15:08 -0400)] 
cleanup

12 years agoformatting stuff
Mike Bayer [Sat, 4 May 2013 18:23:29 +0000 (14:23 -0400)] 
formatting stuff

12 years ago- disable tests failing with pypy2, [ticket:2719]
Mike Bayer [Wed, 1 May 2013 14:09:34 +0000 (10:09 -0400)] 
- disable tests failing with pypy2, [ticket:2719]

12 years ago- Fixed a regression from 0.7 where the contextmanager feature
Mike Bayer [Tue, 30 Apr 2013 14:02:49 +0000 (10:02 -0400)] 
- Fixed a regression from 0.7 where the contextmanager feature
of :meth:`.Session.begin_nested` would fail to correctly
roll back the transaction when a flush error occurred, instead
raising its own exception while leaving the session still
pending a rollback.  [ticket:2718]

12 years agomissing comma
Mike Bayer [Tue, 30 Apr 2013 04:15:36 +0000 (00:15 -0400)] 
missing comma

12 years agobump
Mike Bayer [Mon, 29 Apr 2013 23:49:59 +0000 (19:49 -0400)] 
bump

12 years agoUpdated mysqlconnector dialect to check for disconnect based
Mike Bayer [Mon, 29 Apr 2013 23:49:28 +0000 (19:49 -0400)] 
Updated mysqlconnector dialect to check for disconnect based
on the apparent string message sent in the exception; tested
against mysqlconnector 1.0.9.

12 years ago- fix long-outdated documentation for sql_mode/ansiquotes,
Mike Bayer [Mon, 29 Apr 2013 23:21:44 +0000 (19:21 -0400)] 
- fix long-outdated documentation for sql_mode/ansiquotes,
closes [ticket:1552]

12 years agomerge rel_0_8
Mike Bayer [Sat, 27 Apr 2013 21:51:52 +0000 (17:51 -0400)] 
merge rel_0_8

12 years agoAdded tag rel_0_8_1 for changeset 2cf49d898d3a10d1c7a5ac87dd015691d51f9248
Mike Bayer [Sat, 27 Apr 2013 21:36:18 +0000 (17:36 -0400)] 
Added tag rel_0_8_1 for changeset 2cf49d898d3a10d1c7a5ac87dd015691d51f9248

12 years ago0.8.1 rel_0_8_1
Mike Bayer [Sat, 27 Apr 2013 21:27:13 +0000 (17:27 -0400)] 
0.8.1

12 years agomerge default
Mike Bayer [Sat, 27 Apr 2013 21:24:23 +0000 (17:24 -0400)] 
merge default

12 years agomodernize some more examples
Mike Bayer [Sat, 27 Apr 2013 21:21:31 +0000 (17:21 -0400)] 
modernize some more examples

12 years ago- add documentation note for [ticket:1841]
Mike Bayer [Sat, 27 Apr 2013 19:35:13 +0000 (15:35 -0400)] 
- add documentation note for [ticket:1841]

12 years agothe callcount went up here, still less than with no C extensions, not prepared
Mike Bayer [Fri, 26 Apr 2013 23:51:25 +0000 (19:51 -0400)] 
the callcount went up here, still less than with no C extensions, not prepared
to worry about it too much

12 years agooracle doesn't have TRUE/FALSE, put 1/0 here, see how that goes
Mike Bayer [Fri, 26 Apr 2013 23:49:48 +0000 (19:49 -0400)] 
oracle doesn't have TRUE/FALSE, put 1/0 here, see how that goes

12 years agoneed to test for (list, tuple) here and not hasattr("__iter__")
Mike Bayer [Fri, 26 Apr 2013 23:48:43 +0000 (19:48 -0400)] 
need to test for (list, tuple) here and not hasattr("__iter__")
since Py3K strings have __iter__

12 years agofix here, was only failing in the py3k version
Mike Bayer [Fri, 26 Apr 2013 23:45:06 +0000 (19:45 -0400)] 
fix here, was only failing in the py3k version

12 years agomerge default
Mike Bayer [Fri, 26 Apr 2013 19:51:45 +0000 (15:51 -0400)] 
merge default

12 years agoFixes to the ``sqlalchemy.ext.serializer`` extension, including
Mike Bayer [Fri, 26 Apr 2013 19:51:29 +0000 (15:51 -0400)] 
Fixes to the ``sqlalchemy.ext.serializer`` extension, including
that the "id" passed from the pickler is turned into a string
to prevent against bytes being parsed on Py3K, as well as that
``relationship()`` and ``orm.join()`` constructs are now properly
serialized. [ticket:2698] and some other observed issues.

12 years agomerge default
Mike Bayer [Fri, 26 Apr 2013 19:04:54 +0000 (15:04 -0400)] 
merge default

12 years agochangelog for [ticket:2714]
Mike Bayer [Fri, 26 Apr 2013 19:04:42 +0000 (15:04 -0400)] 
changelog for [ticket:2714]

12 years agoand a lot of tests. five fail without the change
Mike Bayer [Fri, 26 Apr 2013 17:08:28 +0000 (13:08 -0400)] 
and a lot of tests.  five fail without the change

12 years ago- further cleanup. in particular, the left clause is never None. But we still
Mike Bayer [Thu, 25 Apr 2013 18:44:37 +0000 (14:44 -0400)] 
- further cleanup.  in particular, the left clause is never None.  But we still
can have errors when we do the auto-join, these are both covered in test_joins

12 years agowe can always adapt to right also. suppose if right
Mike Bayer [Thu, 25 Apr 2013 18:29:48 +0000 (14:29 -0400)] 
we can always adapt to right also.  suppose if right
were an alias of a table, should add tests for that.

12 years agoheh dont need that either
Mike Bayer [Thu, 25 Apr 2013 18:16:54 +0000 (14:16 -0400)] 
heh dont need that either

12 years agodont need this
Mike Bayer [Thu, 25 Apr 2013 18:14:23 +0000 (14:14 -0400)] 
dont need this

12 years agocleanup
Mike Bayer [Thu, 25 Apr 2013 18:13:13 +0000 (14:13 -0400)] 
cleanup

12 years agoeverything passes with this!!!!!!! holy crap !!!!! and its the simplest of all
Mike Bayer [Thu, 25 Apr 2013 17:54:40 +0000 (13:54 -0400)] 
everything passes with this!!!!!!! holy crap !!!!!  and its the simplest of all

12 years ago- attempt to replace the whole idea of "join_to_left" with a more
Mike Bayer [Wed, 24 Apr 2013 22:58:09 +0000 (18:58 -0400)] 
- attempt to replace the whole idea of "join_to_left" with a more
fundamental and general purpose heuristic.   this initial approach
has about 60 tests failing but seems to have gone pretty far

12 years agobreak into multiple tests
Mike Bayer [Wed, 24 Apr 2013 22:10:16 +0000 (18:10 -0400)] 
break into multiple tests

12 years agoformatting
Mike Bayer [Wed, 24 Apr 2013 21:49:24 +0000 (17:49 -0400)] 
formatting

12 years ago- merge exists() patch + changelog
Mike Bayer [Wed, 24 Apr 2013 00:06:04 +0000 (20:06 -0400)] 
- merge exists() patch + changelog

12 years agomerge default
Mike Bayer [Tue, 23 Apr 2013 17:24:02 +0000 (13:24 -0400)] 
merge default

12 years agoAdded a conditional to the unpickling process for ORM
Mike Bayer [Tue, 23 Apr 2013 17:23:48 +0000 (13:23 -0400)] 
Added a conditional to the unpickling process for ORM
mapped objects, such that if the reference to the object
were lost when the object was pickled, we don't
erroneously try to set up _sa_instance_state - fixes
a NoneType error.

12 years agomerge default
Mike Bayer [Tue, 23 Apr 2013 17:07:50 +0000 (13:07 -0400)] 
merge default

12 years ago- only search in the exception before the first newline, to avoid
Mike Bayer [Tue, 23 Apr 2013 17:07:36 +0000 (13:07 -0400)] 
- only search in the exception before the first newline, to avoid
false positives for SQL statements containing certain text

12 years agomerge default
Mike Bayer [Mon, 22 Apr 2013 23:33:49 +0000 (19:33 -0400)] 
merge default

12 years agoOpened up the checking for "disconnect" with psycopg2/libpq
Mike Bayer [Mon, 22 Apr 2013 23:33:39 +0000 (19:33 -0400)] 
Opened up the checking for "disconnect" with psycopg2/libpq
to check for all the various "disconnect" messages within
the full exception hierarchy.  Specifically the
"closed the connection unexpectedly" message has now been
seen in at least three different exception types.
[ticket:2712]

12 years agomerge default
Mike Bayer [Mon, 22 Apr 2013 23:13:00 +0000 (19:13 -0400)] 
merge default

12 years agoFully implemented the IS and IS NOT operators with
Mike Bayer [Mon, 22 Apr 2013 23:12:47 +0000 (19:12 -0400)] 
Fully implemented the IS and IS NOT operators with
regards to the True/False constants.  An expression like
``col.is_(True)`` will now render ``col IS true``
on the target platform, rather than converting the True/
False constant to an integer bound parameter.
This allows the ``is_()`` operator to work on MySQL when
given True/False constants.
[ticket:2682]

12 years agomerge default
Mike Bayer [Mon, 22 Apr 2013 21:24:42 +0000 (17:24 -0400)] 
merge default

12 years agotweak this some more to handle the array being empty again
Mike Bayer [Mon, 22 Apr 2013 21:24:31 +0000 (17:24 -0400)] 
tweak this some more to handle the array being empty again

12 years agomerge default
Mike Bayer [Mon, 22 Apr 2013 21:08:11 +0000 (17:08 -0400)] 
merge default

12 years ago- change to [ticket:2681], pre-coerce the array to list
Mike Bayer [Mon, 22 Apr 2013 21:08:02 +0000 (17:08 -0400)] 
- change to [ticket:2681], pre-coerce the array to list
unconditonally instead so that it works in all cases.

12 years agomerge default
Mike Bayer [Mon, 22 Apr 2013 20:57:54 +0000 (16:57 -0400)] 
merge default

12 years agoThe operators for the Postgresql ARRAY type supports
Mike Bayer [Mon, 22 Apr 2013 20:57:15 +0000 (16:57 -0400)] 
The operators for the Postgresql ARRAY type supports
input types of sets, generators, etc. but only when a dimension
is specified for the ARRAY; otherwise, the dialect
needs to peek inside of "arr[0]" to guess how many
dimensions are in use.  If this occurs with a non
list/tuple type, the error message is now informative
and directs to specify a dimension for the ARRAY.
[ticket:2681]

12 years ago- changelog
Mike Bayer [Sun, 21 Apr 2013 21:23:03 +0000 (17:23 -0400)] 
- changelog
- just do a fetchone() here, no need for len() etc.

12 years agomerge cymysql branch
Mike Bayer [Sun, 21 Apr 2013 21:19:07 +0000 (17:19 -0400)] 
merge cymysql branch

12 years agomerge into cymysql branch...
Mike Bayer [Sun, 21 Apr 2013 21:18:49 +0000 (17:18 -0400)] 
merge into cymysql branch...

12 years agomerge default
Mike Bayer [Sun, 21 Apr 2013 21:10:01 +0000 (17:10 -0400)] 
merge default

12 years ago- Improvements to the operation of the pymysql dialect on
Mike Bayer [Sun, 21 Apr 2013 21:09:45 +0000 (17:09 -0400)] 
- Improvements to the operation of the pymysql dialect on
Python 3, including some important decode/bytes steps.
Issues remain with BLOB types due to driver issues.
Courtesy Ben Trofatter.
- start using util.py3k, we will eventually remove the
sa2to3 fixer entirely

12 years agoMerged in bentrofatter/sqlalchemy-2663 (pull request #49)
Mike Bayer [Sun, 21 Apr 2013 20:58:55 +0000 (16:58 -0400)] 
Merged in bentrofatter/sqlalchemy-2663 (pull request #49)

Fixed PyMySQL problems for Python 2.x and mitigated some issues with Python 3.x

12 years ago- merge default
Mike Bayer [Sun, 21 Apr 2013 15:48:01 +0000 (11:48 -0400)] 
- merge default

12 years ago- identify [ticket:2710] as a regression from [ticket:2229], doesn't
Mike Bayer [Sun, 21 Apr 2013 15:47:43 +0000 (11:47 -0400)] 
- identify [ticket:2710] as a regression from [ticket:2229], doesn't
need to be in 0.7.

12 years agoFixed bug where many-to-many relationship with uselist=False
Mike Bayer [Sun, 21 Apr 2013 15:31:29 +0000 (11:31 -0400)] 
Fixed bug where many-to-many relationship with uselist=False
would fail to delete the association row and raise an error
if the scalar attribute were set to None.  Also in 0.7.11.
[ticket:2710]

12 years ago- remove lots of old tests here that are covered
Mike Bayer [Sun, 21 Apr 2013 15:09:51 +0000 (11:09 -0400)] 
- remove lots of old tests here that are covered
elsewhere, cleanup

12 years ago- establish ordering here for the hstore repr test
Mike Bayer [Sat, 20 Apr 2013 15:47:25 +0000 (11:47 -0400)] 
- establish ordering here for the hstore repr test

12 years agoorder the properties here to get the SQL to be deterministic
Mike Bayer [Sat, 20 Apr 2013 15:44:27 +0000 (11:44 -0400)] 
order the properties here to get the SQL to be deterministic

12 years ago- these pickle the assoc proxy by itself tests can't really pass
Mike Bayer [Sat, 20 Apr 2013 09:34:40 +0000 (05:34 -0400)] 
- these pickle the assoc proxy by itself tests can't really pass
now without strong ref on the parent
- fix message compare for py3k

12 years agomerge from default
Hajime Nakagami [Sat, 20 Apr 2013 08:10:23 +0000 (17:10 +0900)] 
merge from default

12 years agotesting.crashes on cymysql >= 0.6
Hajime Nakagami [Sat, 20 Apr 2013 07:33:32 +0000 (16:33 +0900)] 
testing.crashes on cymysql >= 0.6

12 years agomerge default
Mike Bayer [Sat, 20 Apr 2013 07:09:22 +0000 (03:09 -0400)] 
merge default

12 years ago- this issue is a bug, mention the apply_labels issue sooner
Mike Bayer [Sat, 20 Apr 2013 07:09:10 +0000 (03:09 -0400)] 
- this issue is a bug, mention the apply_labels issue sooner

12 years ago- dont do a boolean check on the mapped object
Mike Bayer [Sat, 20 Apr 2013 06:59:58 +0000 (02:59 -0400)] 
- dont do a boolean check on the mapped object

12 years agoskip with cymysql 0.6
Hajime Nakagami [Sat, 20 Apr 2013 06:54:33 +0000 (15:54 +0900)] 
skip with cymysql 0.6

12 years agoImproved the behavior of instance management regarding
Mike Bayer [Sat, 20 Apr 2013 06:45:08 +0000 (02:45 -0400)] 
Improved the behavior of instance management regarding
the creation of strong references within the Session;
an object will no longer have an internal reference cycle
created if it's in the transient state or moves into the
detached state - the strong ref is created only when the
object is attached to a Session and is removed when the
object is detached.  This makes it somewhat safer for an
object to have a `__del__()` method, even though this is
not recommended, as relationships with backrefs produce
cycles too.  A warning has been added when a class with
a `__del__()` method is mapped.
[ticket:2708]

12 years agochangelog for pullreq 54 gaerdbms
Mike Bayer [Sat, 20 Apr 2013 01:41:11 +0000 (21:41 -0400)] 
changelog for pullreq 54 gaerdbms

12 years agoFix mysql+gaerdbms dialect for changed exception format
Dan Ring [Sat, 20 Apr 2013 01:01:39 +0000 (18:01 -0700)] 
Fix mysql+gaerdbms dialect for changed exception format

googleappengine v1.7.5 changed the exception format to be
incompatible with MySQLDialect_gaerdbms#_extract_error_code

This fix works for both old- and new-style exceptions.

Changes causing the breakage:
/trunk/python/google/storage/speckle/python/api/rdbms.py
at
https://code.google.com/p/googleappengine/source/detail?r=318

12 years agoFixed a long-standing bug in the caching example, where
Mike Bayer [Fri, 19 Apr 2013 00:11:08 +0000 (20:11 -0400)] 
Fixed a long-standing bug in the caching example, where
the limit/offset parameter values wouldn't be taken into
account when computing the cache key.  The
_key_from_query() function has been simplified to work
directly from the final compiled statement in order to get
at both the full statement as well as the fully processed
parameter list.

12 years ago- remove erroneous second RelationshipCache class
Mike Bayer [Thu, 18 Apr 2013 22:35:45 +0000 (18:35 -0400)] 
- remove erroneous second RelationshipCache class

12 years ago- this pymssql test needs to be against the pymssql dialect
Mike Bayer [Thu, 18 Apr 2013 21:41:30 +0000 (17:41 -0400)] 
- this pymssql test needs to be against the pymssql dialect
- Part of a longer series of fixes needed for pyodbc+
      mssql, a CAST to NVARCHAR(max) has been added to the bound
      parameter for the table name and schema name in all information schema
      queries to avoid the issue of comparing NVARCHAR to NTEXT,
      which seems to be rejected by the ODBC driver in some cases,
      such as FreeTDS (0.91 only?) plus unicode bound parameters being passed.
      The issue seems to be specific to the SQL Server information
      schema tables and the workaround is harmless for those cases
      where the problem doesn't exist in the first place.
[ticket:2355]

12 years agopython2.5 fix
Mike Bayer [Thu, 18 Apr 2013 19:45:15 +0000 (15:45 -0400)] 
python2.5 fix

12 years ago- test + changelog for [ticket:2691]
Mike Bayer [Thu, 18 Apr 2013 16:01:16 +0000 (12:01 -0400)] 
- test + changelog for [ticket:2691]

12 years ago- remove reference to _exc_info before reraise to reduce cycles
Mike Bayer [Thu, 18 Apr 2013 15:04:51 +0000 (11:04 -0400)] 
- remove reference to _exc_info before reraise to reduce cycles

12 years agoReworked internal exception raises that emit
Mike Bayer [Thu, 18 Apr 2013 15:00:12 +0000 (11:00 -0400)] 
Reworked internal exception raises that emit
a rollback() before re-raising, so that the stack
trace is preserved from sys.exc_info() before entering
the rollback.  This so that the traceback is preserved
when using coroutine frameworks which may have switched
contexts before the rollback function returns.
[ticket:2703]

12 years ago- additional test + correction for [ticket:2699]
Mike Bayer [Thu, 18 Apr 2013 14:34:59 +0000 (10:34 -0400)] 
- additional test + correction for [ticket:2699]

12 years agomerge default
Mike Bayer [Sun, 14 Apr 2013 23:33:06 +0000 (19:33 -0400)] 
merge default

12 years agoadd cymysql...
Mike Bayer [Sun, 14 Apr 2013 23:32:54 +0000 (19:32 -0400)] 
add cymysql...

12 years agomysql+cymysql dialect supports_sane_rowcount = False
Hajime Nakagami [Sat, 13 Apr 2013 09:19:05 +0000 (18:19 +0900)] 
mysql+cymysql dialect supports_sane_rowcount = False

12 years agoa bit revert
Hajime Nakagami [Sat, 13 Apr 2013 09:04:31 +0000 (18:04 +0900)] 
a bit revert

12 years agomerge from default
Hajime Nakagami [Sat, 13 Apr 2013 08:54:16 +0000 (17:54 +0900)] 
merge from default

12 years agomerge default
Mike Bayer [Fri, 12 Apr 2013 00:05:11 +0000 (20:05 -0400)] 
merge default

12 years agoFixed bug whereby ORM would run the wrong kind of
Mike Bayer [Fri, 12 Apr 2013 00:04:57 +0000 (20:04 -0400)] 
Fixed bug whereby ORM would run the wrong kind of
query when refreshing an inheritance-mapped class
where the superclass was mapped to a non-Table
object, like a custom join() or a select(),
running a query that assumed a hierarchy that's
mapped to individual Table-per-class.
[ticket:2697]

12 years agomerge default
Mike Bayer [Thu, 11 Apr 2013 23:39:04 +0000 (19:39 -0400)] 
merge default

12 years ago- unit test failing on python3.3 + psycopg2, add an order by
Mike Bayer [Thu, 11 Apr 2013 23:38:49 +0000 (19:38 -0400)] 
- unit test failing on python3.3 + psycopg2, add an order by

12 years agoFixed `__repr__()` on mapper property constructs
Mike Bayer [Thu, 11 Apr 2013 23:37:56 +0000 (19:37 -0400)] 
Fixed `__repr__()` on mapper property constructs
to work before the object is initialized, so
that Sphinx builds with recent Sphinx versions
can read them.

12 years agoThe _Binary base type now converts values through
Mike Bayer [Thu, 11 Apr 2013 23:34:46 +0000 (19:34 -0400)] 
The _Binary base type now converts values through
the bytes() callable when run on Python 3; in particular
psycopg2 2.5 with Python 3.3 seems to now be returning
the "memoryview" type, so this is converted to bytes
before return.

12 years agomerge default
Mike Bayer [Thu, 11 Apr 2013 23:10:19 +0000 (19:10 -0400)] 
merge default

12 years agoImprovements to Connection auto-invalidation
Mike Bayer [Thu, 11 Apr 2013 23:10:02 +0000 (19:10 -0400)] 
Improvements to Connection auto-invalidation
handling.  If a non-disconnect error occurs,
but leads to a delayed disconnect error within error
handling (happens with MySQL), the disconnect condition
is detected.  The Connection can now also be closed
when in an invalid state, meaning it will raise "closed"
on next usage, and additionally the "close with result"
feature will work even if the autorollback in an error
handling routine fails and regardless of whether the
condition is a disconnect or not.
[ticket:2695]

12 years agomerge default
Mike Bayer [Thu, 11 Apr 2013 20:51:40 +0000 (16:51 -0400)] 
merge default

12 years agofix category
Mike Bayer [Thu, 11 Apr 2013 20:51:27 +0000 (16:51 -0400)] 
fix category