]> git.ipfire.org Git - thirdparty/sqlalchemy/sqlalchemy.git/log
thirdparty/sqlalchemy/sqlalchemy.git
15 years ago78 chars
Mike Bayer [Wed, 7 Jul 2010 14:20:43 +0000 (10:20 -0400)] 
78 chars

15 years agomerge
Mike Bayer [Wed, 7 Jul 2010 00:00:39 +0000 (20:00 -0400)] 
merge

15 years ago78 chars for sybase
Mike Bayer [Wed, 7 Jul 2010 00:00:13 +0000 (20:00 -0400)] 
78 chars for sybase

15 years agoAdded tag rel_0_6_2 for changeset 7e72edc9d39966721f2731ef2dbf508ca4a03e4c
Mike Bayer [Tue, 6 Jul 2010 21:57:41 +0000 (17:57 -0400)] 
Added tag rel_0_6_2 for changeset 7e72edc9d39966721f2731ef2dbf508ca4a03e4c

15 years ago- updated obsolete docs regarding sequences, RETURNING for pg rel_0_6_2
Mike Bayer [Tue, 6 Jul 2010 21:39:20 +0000 (17:39 -0400)] 
- updated obsolete docs regarding sequences, RETURNING for pg
- 78 chars for postgresql/base.py, engine/default.py

15 years agoup the variance on these as the call counts are very low and
Mike Bayer [Tue, 6 Jul 2010 21:09:42 +0000 (17:09 -0400)] 
up the variance on these as the call counts are very low and
things wiggle around even between 2.6.4/2.6.5

15 years ago- latest distribute
Mike Bayer [Mon, 5 Jul 2010 23:54:46 +0000 (19:54 -0400)] 
- latest distribute
- added caveats to unittest README encountered with Py2.7 + current nose 0.11.3
- call counts for py2.7.   all tests pass for sqlite + mysql-python + psycopg2

15 years agoforce __doc__ to transfer over from the decorated funciton.
Mike Bayer [Mon, 5 Jul 2010 21:17:34 +0000 (17:17 -0400)] 
force __doc__ to transfer over from the decorated funciton.
2.6 does this anyway, 2.4 and 2.5 do not.

15 years agomerge
Mike Bayer [Sun, 4 Jul 2010 22:52:56 +0000 (18:52 -0400)] 
merge

15 years agosome prep for the 'viewcode' plugin, which works, though not sure if its helpful
Mike Bayer [Sun, 4 Jul 2010 22:52:34 +0000 (18:52 -0400)] 
some prep for the 'viewcode' plugin, which works, though not sure if its helpful

15 years agosome linkage
Mike Bayer [Sun, 4 Jul 2010 21:17:59 +0000 (17:17 -0400)] 
some linkage

15 years agotransfer docstrings from @classproperty to props
Mike Bayer [Sun, 4 Jul 2010 16:06:19 +0000 (12:06 -0400)] 
transfer docstrings from @classproperty to props

15 years ago- idle 78-char adjustments
Mike Bayer [Sat, 3 Jul 2010 22:55:29 +0000 (18:55 -0400)] 
- idle 78-char adjustments

15 years ago- Added support for @classproperty to provide
Mike Bayer [Sat, 3 Jul 2010 18:53:37 +0000 (14:53 -0400)] 
- Added support for @classproperty to provide
any kind of schema/mapping construct from a
declarative mixin, including columns with foreign
keys, relationships, column_property, deferred.
This solves all such issues on declarative mixins.
An error is raised if any MapperProperty subclass
is specified on a mixin without using @classproperty.
[ticket:1751] [ticket:1796] [ticket:1805]

- un-anglicized the declarative docs

15 years ago- verbiage expressing that pool_size==0 means no limit; NullPool is
Mike Bayer [Fri, 2 Jul 2010 18:16:08 +0000 (14:16 -0400)] 
- verbiage expressing that pool_size==0 means no limit; NullPool is
used to disable pooling. [ticket:1164]

15 years ago- The 'default' compiler is automatically copied over
Mike Bayer [Fri, 2 Jul 2010 18:07:42 +0000 (14:07 -0400)] 
- The 'default' compiler is automatically copied over
when overriding the compilation of a built in
clause construct, so no KeyError is raised if the
user-defined compiler is specific to certain
backends and compilation for a different backend
is invoked. [ticket:1838]

15 years ago- add typing/operator converts to Float/Numeric to get Oracle tests going again,
Mike Bayer [Fri, 2 Jul 2010 17:55:24 +0000 (13:55 -0400)] 
- add typing/operator converts to Float/Numeric to get Oracle tests going again,
as we have an unusually high level of dependence on SQLA typing now
- force oracle seqs to return integers

15 years ago- If server_version_info is outside the usual
Mike Bayer [Fri, 2 Jul 2010 16:57:15 +0000 (12:57 -0400)] 
- If server_version_info is outside the usual
range of (8, ), (9, ), (10, ), a warning is emitted
which suggests checking that the FreeTDS version
configuration is using 7.0 or 8.0, not 4.2.
[ticket:1825]

15 years agokeep ordering of columns with mixins constant
Chris Withers [Fri, 2 Jul 2010 13:57:19 +0000 (14:57 +0100)] 
keep ordering of columns with mixins constant

15 years ago- Query.join() will check for a call of the
Mike Bayer [Fri, 2 Jul 2010 00:50:57 +0000 (20:50 -0400)] 
- Query.join() will check for a call of the
form query.join(target, clause_expression),
i.e. missing the tuple, and raise an informative
error message that this is the wrong calling form.

15 years ago - Oracle's "native decimal" metadata begins to return
Mike Bayer [Thu, 1 Jul 2010 20:57:02 +0000 (16:57 -0400)] 
  - Oracle's "native decimal" metadata begins to return
    ambiguous typing information about numerics
    when columns are embedded in subqueries as well
    as when ROWNUM is consulted with subqueries, as we
    do for limit/offset.  We've added these ambiguous
    conditions to the cx_oracle "convert to Decimal()"
    handler, so that we receive numerics as Decimal
    in more cases instead of as floats.  These are
    then converted, if requested, into Integer
    or Float, or otherwise kept as the lossless
    Decimal [ticket:1840].

15 years ago- An instance which is moved to "transient", has
Mike Bayer [Wed, 30 Jun 2010 20:12:32 +0000 (16:12 -0400)] 
- An instance which is moved to "transient", has
an incomplete or missing set of primary key
attributes, and contains expired attributes, will
raise an InvalidRequestError if an expired attribute
is accessed, instead of getting a recursion overflow.

- make_transient() removes all "loader" callables from
the state being made transient, removing any
"expired" state - all unloaded attributes reset back
to undefined, None/empty on access.

15 years agoreplace all occurences of "closes the connection" "closes the transaction"
Mike Bayer [Tue, 29 Jun 2010 23:02:49 +0000 (19:02 -0400)] 
replace all occurences of "closes the connection" "closes the transaction"
with explicit verbiage regarding "returns the connection to the connection
pool", "rolls back the transaction"

15 years agoadd make_transient
Mike Bayer [Tue, 29 Jun 2010 15:24:15 +0000 (11:24 -0400)] 
add make_transient

15 years agofix oracle char length test for newer descriptors
Mike Bayer [Sun, 27 Jun 2010 17:05:05 +0000 (13:05 -0400)] 
fix oracle char length test for newer descriptors

15 years ago- Fixed bug that would prevent overridden clause
Mike Bayer [Sat, 26 Jun 2010 21:20:17 +0000 (17:20 -0400)] 
- Fixed bug that would prevent overridden clause
compilation from working for "annotated" expression
elements, which are often generated by the ORM.

15 years ago- Won't generate "CREATE TYPE" / "DROP TYPE" if
Mike Bayer [Fri, 25 Jun 2010 17:11:06 +0000 (13:11 -0400)] 
- Won't generate "CREATE TYPE" / "DROP TYPE" if
using types.Enum on a PG version prior to 8.3 -
the supports_native_enum flag is fully
honored.  [ticket:1836]

15 years agofix test
Mike Bayer [Thu, 24 Jun 2010 16:45:37 +0000 (12:45 -0400)] 
fix test

15 years ago- The argument to "ESCAPE" of a LIKE operator or similar
Mike Bayer [Thu, 24 Jun 2010 16:19:15 +0000 (12:19 -0400)] 
- The argument to "ESCAPE" of a LIKE operator or similar
is passed through render_literal_value(), which may
implement escaping of backslashes.  [ticket:1400]
- Postgresql render_literal_value() is overridden which escapes
backslashes, currently applies to the ESCAPE clause
of LIKE and similar expressions.
Ultimately this will have to detect the value of
"standard_conforming_strings" for full behavior.
[ticket:1400]
- MySQL render_literal_value() is overridden which escapes
backslashes, currently applies to the ESCAPE clause
of LIKE and similar expressions.   This behavior
is derived from detecting the value of
NO_BACKSLASH_ESCAPES.  [ticket:1400]

15 years ago- Fixed @memoized_property and @memoized_instancemethod
Mike Bayer [Sat, 19 Jun 2010 18:08:20 +0000 (14:08 -0400)] 
- Fixed @memoized_property and @memoized_instancemethod
decorators so that Sphinx documentation picks up
these attributes and methods, such as
ResultProxy.inserted_primary_key. [ticket:1830]

15 years ago- Firebird dialect adds CHAR, VARCHAR types which
Mike Bayer [Sat, 19 Jun 2010 17:51:55 +0000 (13:51 -0400)] 
- Firebird dialect adds CHAR, VARCHAR types which
accept a "charset" flag, to support Firebird
"CHARACTER SET" clause.  [ticket:1813]

15 years ago- Fixed ora-8 compatibility flags such that they
Mike Bayer [Sat, 19 Jun 2010 17:39:32 +0000 (13:39 -0400)] 
- Fixed ora-8 compatibility flags such that they
don't cache a stale value from before the first
database connection actually occurs. [ticket:1819]

15 years agofix accessors, [ticket:1816]
Mike Bayer [Sat, 19 Jun 2010 17:29:16 +0000 (13:29 -0400)] 
fix accessors, [ticket:1816]

15 years ago- a mixin class can now define a column that matches
Mike Bayer [Sat, 19 Jun 2010 17:25:37 +0000 (13:25 -0400)] 
- a mixin class can now define a column that matches
one which is present on a __table__ defined on a
subclass.  It cannot, however, define one that is
not present in the __table__, and the error message
here now works.  [ticket:1821]

15 years ago- MySQL dialect doesn't emit CAST() for MySQL version
Mike Bayer [Sat, 19 Jun 2010 17:21:19 +0000 (13:21 -0400)] 
- MySQL dialect doesn't emit CAST() for MySQL version
detected < 4.0.2.  This allows the unicode
check on connect to proceed. [ticket:1826]

15 years ago- moved "post update" functionality from _save_obj() into
Mike Bayer [Fri, 18 Jun 2010 23:12:52 +0000 (19:12 -0400)] 
- moved "post update" functionality from _save_obj() into
its own method, which also groups updates into executemanys.
[ticket:1831]

15 years ago- cleanup of new post update, changelog
Mike Bayer [Fri, 18 Jun 2010 19:42:01 +0000 (15:42 -0400)] 
- cleanup of new post update, changelog

15 years agoadd a test for the actual feature, that multiple post updates are batched
Mike Bayer [Fri, 18 Jun 2010 19:11:32 +0000 (15:11 -0400)] 
add a test for the actual feature, that multiple post updates are batched

15 years agothis one is actually doing it. removed the cruft we don't need from the old approach.
Mike Bayer [Fri, 18 Jun 2010 17:31:51 +0000 (13:31 -0400)] 
this one is actually doing it.  removed the cruft we don't need from the old approach.
not sure if one-to-many with elaborate self-refs, etc., but we appear to be
as good as we were before.

15 years agoadds some tests, refines out the m2o approach.
Mike Bayer [Fri, 18 Jun 2010 17:11:19 +0000 (13:11 -0400)] 
adds some tests, refines out the m2o approach.

15 years agothis approach seems to allow many-to-one post updates to occur as a single action...
Mike Bayer [Fri, 18 Jun 2010 15:36:35 +0000 (11:36 -0400)] 
this approach seems to allow many-to-one post updates to occur as a single action.   for one-to-many, we
have it working the old way still. as usual the huge job is the tests.   also I am not yet certain
that post updates can always be a "full mapper" operation, i.e. are never involved in per-state dependencies,
or even if the old approach supports that.

15 years ago- query.order_by() now accepts False, which cancels
Mike Bayer [Thu, 17 Jun 2010 19:53:07 +0000 (15:53 -0400)] 
- query.order_by() now accepts False, which cancels
any existing order_by() state on the Query, allowing
subsequent generative methods to be called which do
not support ORDER BY.  This is not the same as the
already existing feature of passing None, which
suppresses any existing order_by() settings, including
those configured on the mapper.  False will make it
as though order_by() was never called, while
None is an active setting.

15 years agofix import
Mike Bayer [Thu, 17 Jun 2010 18:32:32 +0000 (14:32 -0400)] 
fix import

15 years agoadd a test for query.order_by(None)
Mike Bayer [Thu, 17 Jun 2010 14:13:23 +0000 (10:13 -0400)] 
add a test for query.order_by(None)

15 years ago- Column.copy() takes along the "unique" attribute
Mike Bayer [Tue, 15 Jun 2010 21:56:17 +0000 (17:56 -0400)] 
- Column.copy() takes along the "unique" attribute
among others, fixes [ticket:1829] regarding declarative
mixins

15 years agoversion_info is the tuple
Mike Bayer [Tue, 15 Jun 2010 18:05:02 +0000 (14:05 -0400)] 
version_info is the tuple

15 years ago- Can now call make_transient() on an instance that
Mike Bayer [Tue, 15 Jun 2010 18:04:13 +0000 (14:04 -0400)] 
- Can now call make_transient() on an instance that
is referenced by parent objects via many-to-one,
without the parent's foreign key value getting
temporarily set to None - this was a function
of the "detect primary key switch" flush handler.
It now ignores objects that are no longer
in the "persistent" state, and the parent's
foreign key identifier is left unaffected.

15 years ago- hypothesize that newer pysqlites are also growing a warning log on non-unicdoe...
Mike Bayer [Tue, 15 Jun 2010 00:15:27 +0000 (20:15 -0400)] 
- hypothesize that newer pysqlites are also growing a warning log on non-unicdoe strings

15 years agofix variable name
Mike Bayer [Tue, 15 Jun 2010 00:11:02 +0000 (20:11 -0400)] 
fix variable name

15 years ago- Query.statement, Query.subquery(), etc. now transfer
Mike Bayer [Tue, 15 Jun 2010 00:07:44 +0000 (20:07 -0400)] 
- Query.statement, Query.subquery(), etc. now transfer
the values of bind parameters, i.e. those specified
by query.params(), into the resulting SQL expression.
Previously the values would not be transferred
and bind parameters would come out as None.

- Subquery-eager-loading now works with Query objects
which include params(), as well as get() Queries.

15 years ago- Modified the internals of "column annotation" such that
Mike Bayer [Mon, 14 Jun 2010 23:39:26 +0000 (19:39 -0400)] 
- Modified the internals of "column annotation" such that
a custom Column subclass can safely override
_constructor to return Column, for the purposes of
making "configurational" column classes that aren't
involved in proxying, etc.

15 years ago- Fixed bug regarding flushes on self-referential
Mike Bayer [Sun, 13 Jun 2010 20:25:26 +0000 (16:25 -0400)] 
- Fixed bug regarding flushes on self-referential
bi-directional many-to-many relationships, where
two objects made to mutually reference each other
in one flush would fail to insert a row for both
sides.  Regression from 0.5. [ticket:1824]

15 years ago- The warning emitted by the Unicode and String types
Mike Bayer [Sun, 13 Jun 2010 20:09:13 +0000 (16:09 -0400)] 
- The warning emitted by the Unicode and String types
with convert_unicode=True no longer embeds the actual
value passed.   This so that the Python warning
registry does not continue to grow in size, the warning
is emitted once as per the warning filter settings,
and large string values don't pollute the output.
[ticket:1822]

15 years ago- added documentation for Inspector [ticket:1820]
Mike Bayer [Sun, 13 Jun 2010 19:49:27 +0000 (15:49 -0400)] 
- added documentation for Inspector [ticket:1820]
- fixed "toplevel" regexp to include tokens of the form foo-bar-toplevel (i.e. multiple dashes)
- some CSS adjustment so that all API headings are in gray.

15 years ago- Inspector hits bind.connect() when invoked to ensure
Mike Bayer [Fri, 11 Jun 2010 15:38:06 +0000 (11:38 -0400)] 
- Inspector hits bind.connect() when invoked to ensure
initialize has been called.  the internal name ".conn"
is changed to ".bind", since that's what it is.

15 years agofix the enum for MySQL after native_enum added to adaption keys
Mike Bayer [Wed, 9 Jun 2010 22:27:43 +0000 (18:27 -0400)] 
fix the enum for MySQL after native_enum added to adaption keys

15 years ago- Fixed incorrect signature in do_execute() for firebird, access, error
Mike Bayer [Wed, 9 Jun 2010 15:07:57 +0000 (11:07 -0400)] 
- Fixed incorrect signature in do_execute() for firebird, access, error
introduced in 0.6.1. [ticket:1823]

15 years ago- Fixed bug in Enum type which blew away native_enum
Mike Bayer [Tue, 8 Jun 2010 23:25:15 +0000 (19:25 -0400)] 
- Fixed bug in Enum type which blew away native_enum
flag when used with TypeDecorators or other adaption
scenarios.

15 years ago- tweak to work with sphinx 1.0b2
Mike Bayer [Tue, 8 Jun 2010 19:53:13 +0000 (15:53 -0400)] 
- tweak to work with sphinx 1.0b2
- more open ended language for bound metadata

15 years agoAdded tag rel_0_6_1 for changeset 1bd54fd9c854044d8153dba0f6dd261bf6bb6bff
Mike Bayer [Mon, 31 May 2010 19:22:55 +0000 (15:22 -0400)] 
Added tag rel_0_6_1 for changeset 1bd54fd9c854044d8153dba0f6dd261bf6bb6bff

15 years ago0.6.1 rel_0_6_1
Mike Bayer [Mon, 31 May 2010 18:47:07 +0000 (14:47 -0400)] 
0.6.1

15 years agofix this test for oracle
Mike Bayer [Mon, 31 May 2010 18:46:21 +0000 (14:46 -0400)] 
fix this test for oracle

15 years agomapper.py is entirely 79 char lines now
Mike Bayer [Mon, 31 May 2010 16:23:46 +0000 (12:23 -0400)] 
mapper.py is entirely 79 char lines now

15 years agochange the weakkeydict to be just an LRU cache. Add tests
Mike Bayer [Mon, 31 May 2010 15:56:08 +0000 (11:56 -0400)] 
change the weakkeydict to be just an LRU cache.   Add tests
for the "many combinations of UPDATE keys" issue.

15 years agoshorten lines to 78 char on _save_obj, _delete_obj, related. much more to go for...
Mike Bayer [Mon, 31 May 2010 01:52:24 +0000 (21:52 -0400)] 
shorten lines to 78 char on _save_obj, _delete_obj, related.   much more to go for mapper.py

15 years agoreduce the calls to execution_options and weakkeydict.setdefault
Mike Bayer [Mon, 31 May 2010 01:23:59 +0000 (21:23 -0400)] 
reduce the calls to execution_options and weakkeydict.setdefault

15 years ago- Pool classes will reuse the same "pool_logging_name" setting
Mike Bayer [Mon, 31 May 2010 00:24:08 +0000 (20:24 -0400)] 
- Pool classes will reuse the same "pool_logging_name" setting
after a dispose() occurs.

- Engine gains an "execution_options" argument and
update_execution_options() method, which will apply to
all connections generated by this engine.

- Added more aggressive caching to the mapper's usage of
UPDATE, INSERT, and DELETE expressions.  Assuming the
statement has no per-object SQL expressions attached,
the expression objects are cached by the mapper after
the first create, and their compiled form is stored
persistently in a cache dictionary for the duration of
the related Engine.

- change #3 required change #1 so that we could test
a set of mappers operating over the course of many engines without
memory usage increase.

15 years ago79 char lines for base.py
Mike Bayer [Sun, 30 May 2010 17:36:40 +0000 (13:36 -0400)] 
79 char lines for base.py

15 years ago- Fixed concatenation of constraints when "PRIMARY KEY"
Mike Bayer [Thu, 27 May 2010 18:46:44 +0000 (14:46 -0400)] 
- Fixed concatenation of constraints when "PRIMARY KEY"
constraint gets moved to column level due to SQLite
AUTOINCREMENT keyword being rendered.  [ticket:1812]
- remove some extra space in between constraint DDL
- added alias() to binary comparison test, fixing pg + mysql failures

15 years ago- Re-established support for Oracle 8 with cx_oracle,
Mike Bayer [Thu, 27 May 2010 17:02:17 +0000 (13:02 -0400)] 
- Re-established support for Oracle 8 with cx_oracle,
including that use_ansi is set to False automatically,
NVARCHAR2 and NCLOB are not rendered for Unicode,
"native unicode" check doesn't fail, cx_oracle
"native unicode" mode is disabled, VARCHAR() is emitted
with bytes count instead of char count. [ticket:1808]

15 years ago- FOR UPDATE is emitted in the syntactically correct position
Mike Bayer [Thu, 27 May 2010 15:50:16 +0000 (11:50 -0400)] 
- FOR UPDATE is emitted in the syntactically correct position
when limit/offset is used, i.e. the ROWNUM subquery.
However, Oracle can't really handle FOR UPDATE with ORDER BY
or with subqueries, so its still not very usable, but at
least SQLA gets the SQL past the Oracle parser.
[ticket:1815]

15 years ago- Columns of _Binary type (i.e. LargeBinary, BLOB, etc.)
Mike Bayer [Sat, 22 May 2010 23:08:48 +0000 (19:08 -0400)] 
- Columns of _Binary type (i.e. LargeBinary, BLOB, etc.)
will coerce a "basestring" on the right side into a
_Binary as well so that required DBAPI processing
takes place.

15 years ago- Fixed regression introduced in 0.6.0 unit of work refactor
Mike Bayer [Fri, 21 May 2010 21:44:56 +0000 (17:44 -0400)] 
- Fixed regression introduced in 0.6.0 unit of work refactor
that broke updates for bi-directional relationship()
with post_update=True. [ticket:1807]

15 years agonew link for pymssql #1810
Mike Bayer [Fri, 21 May 2010 21:05:06 +0000 (17:05 -0400)] 
new link for pymssql #1810

15 years agodocs on dispose
Mike Bayer [Sat, 15 May 2010 20:12:14 +0000 (16:12 -0400)] 
docs on dispose

15 years agodisclaimer for the "type=None" semi-feature
Mike Bayer [Sat, 15 May 2010 19:56:49 +0000 (15:56 -0400)] 
disclaimer for the "type=None" semi-feature

15 years ago- Added table.add_is_dependent_on(othertable), allows manual
Mike Bayer [Thu, 13 May 2010 19:53:18 +0000 (15:53 -0400)] 
- Added table.add_is_dependent_on(othertable), allows manual
placement of dependency rules between two Table objects
for use within create_all(), drop_all(), sorted_tables.
[ticket:1801]

15 years ago- Added internal warning in case an instance without a
Mike Bayer [Thu, 13 May 2010 19:45:56 +0000 (15:45 -0400)] 
- Added internal warning in case an instance without a
full PK happened to be expired and then was asked
to refresh. [ticket:1797]

15 years ago- fixed __setstate__ method of CollectionAdapter to not
Mike Bayer [Thu, 13 May 2010 19:36:22 +0000 (15:36 -0400)] 
- fixed __setstate__ method of CollectionAdapter to not
fail during deserialize where parent InstanceState not
yet unserialized.  [ticket:1802]

15 years ago- func.XXX() doesn't inadvertently resolve to non-Function
Mike Bayer [Thu, 13 May 2010 19:26:53 +0000 (15:26 -0400)] 
- func.XXX() doesn't inadvertently resolve to non-Function
classes (e.g. fixes func.text()).  [ticket:1798]

15 years agosome pep-8ing
Mike Bayer [Thu, 13 May 2010 17:33:22 +0000 (13:33 -0400)] 
some pep-8ing

15 years ago- the _make_proxy() method of ColumnClause and Column now use
Mike Bayer [Mon, 10 May 2010 15:37:48 +0000 (11:37 -0400)] 
- the _make_proxy() method of ColumnClause and Column now use
self.__class__ to determine the class of object to be returned
instead of hardcoding to ColumnClause/Column, making it slightly
easier to produce specific subclasses of these which work in
alias/subquery situations.

15 years ago- expr.in_() now accepts a text() construct as the argument.
Mike Bayer [Sat, 8 May 2010 20:25:30 +0000 (16:25 -0400)] 
- expr.in_() now accepts a text() construct as the argument.
Grouping parenthesis are added automatically, i.e. usage
is like `col.in_(text("select id from table"))`.
[ticket:1793]

15 years ago- session.merge() will not expire attributes on the returned
Mike Bayer [Sat, 8 May 2010 20:09:48 +0000 (16:09 -0400)] 
- session.merge() will not expire attributes on the returned
instance if that instance is "pending".  [ticket:1789]

15 years ago- func.sysdate() emits "SYSDATE()", i.e. with the ending
Mike Bayer [Sat, 8 May 2010 19:29:08 +0000 (15:29 -0400)] 
- func.sysdate() emits "SYSDATE()", i.e. with the ending
parenthesis, on MySQL.  [ticket:1794]

15 years ago- Added 'connection shutdown' to the list of exception strings
Mike Bayer [Sat, 8 May 2010 19:22:10 +0000 (15:22 -0400)] 
- Added 'connection shutdown' to the list of exception strings
which indicate a dropped connection.  [ticket:1646]

15 years agoadd a note about native unicode mode, [ticket:1792]
Mike Bayer [Sat, 8 May 2010 19:19:23 +0000 (15:19 -0400)] 
add a note about native unicode mode, [ticket:1792]

15 years ago- the SqlSoup constructor accepts a `base` argument which specifies
Mike Bayer [Sat, 8 May 2010 19:08:48 +0000 (15:08 -0400)] 
- the SqlSoup constructor accepts a `base` argument which specifies
the base class to use for mapped classes, the default being
`object`.  [ticket:1783]

15 years agoadd more insert examples #1791
Mike Bayer [Tue, 4 May 2010 02:18:27 +0000 (22:18 -0400)] 
add more insert examples #1791

15 years agomerge tip
Mike Bayer [Mon, 3 May 2010 23:32:34 +0000 (19:32 -0400)] 
merge tip

15 years ago- Fixed use_ansi=False mode, which was producing broken
Mike Bayer [Mon, 3 May 2010 23:24:52 +0000 (19:24 -0400)] 
- Fixed use_ansi=False mode, which was producing broken
WHERE clauses in pretty much all cases.  [ticket:1790]

15 years agoplace docs for DBAPI-agnostic transaction_isolation setting for sqlite, pg [ticket...
Mike Bayer [Fri, 30 Apr 2010 19:00:48 +0000 (15:00 -0400)] 
place docs for DBAPI-agnostic transaction_isolation setting for sqlite, pg [ticket:1784]

15 years agoremove docs for deprecated feature
Chris Withers [Fri, 30 Apr 2010 07:39:25 +0000 (08:39 +0100)] 
remove docs for deprecated feature

15 years agofix some 2to3 breakage on the cx_oracle dialect
Mike Bayer [Thu, 29 Apr 2010 19:33:19 +0000 (15:33 -0400)] 
fix some 2to3 breakage on the cx_oracle dialect

15 years agoremove a few superfluous imports
Mike Bayer [Thu, 29 Apr 2010 18:10:28 +0000 (14:10 -0400)] 
remove a few superfluous imports

15 years ago- Fixed bug in connection pool cursor wrapper whereby if a
Mike Bayer [Wed, 28 Apr 2010 20:16:56 +0000 (16:16 -0400)] 
- Fixed bug in connection pool cursor wrapper whereby if a
cursor threw an exception on close(), the logging of the
message would fail.  [ticket:1786]

15 years agofix sphinx reference
Chris Withers [Wed, 28 Apr 2010 18:22:55 +0000 (19:22 +0100)] 
fix sphinx reference

15 years ago- oracle_xe 5 doesn't accept a Python unicode object in
Mike Bayer [Wed, 28 Apr 2010 17:30:25 +0000 (17:30 +0000)] 
- oracle_xe 5 doesn't accept a Python unicode object in
its connect string in normal Python 2.x mode - so we coerce
to str() directly.  non-ascii characters aren't supported
in connect strings here since we don't know what encoding
we could use.  [ticket:1670]

15 years agochangelog fix for #1781
Gaëtan de Menten [Wed, 28 Apr 2010 07:33:47 +0000 (09:33 +0200)] 
changelog fix for #1781

15 years agotake2 on #1781.
Gaëtan de Menten [Tue, 27 Apr 2010 18:57:48 +0000 (20:57 +0200)] 
take2 on #1781.