]> git.ipfire.org Git - thirdparty/sqlalchemy/sqlalchemy.git/log
thirdparty/sqlalchemy/sqlalchemy.git
14 years ago- Added a fully descriptive error message for the
Mike Bayer [Mon, 28 Feb 2011 17:47:04 +0000 (12:47 -0500)] 
- Added a fully descriptive error message for the
case where Column is subclassed and _make_proxy()
fails to make a copy due to TypeError on the
constructor.   The method _constructor should
be implemented in this case.

14 years ago- establish an "insert" option for events to control ordering if needed (not needed...
Mike Bayer [Fri, 25 Feb 2011 18:20:43 +0000 (13:20 -0500)] 
- establish an "insert" option for events to control ordering if needed (not needed yet tho)
- render foreign key constraints in the order in which they were cerated

14 years agoAdded tag rel_0_7b2 for changeset e6d00f8234d6f31113447098181eb5a722085c33
Mike Bayer [Sat, 19 Feb 2011 19:32:14 +0000 (14:32 -0500)] 
Added tag rel_0_7b2 for changeset e6d00f8234d6f31113447098181eb5a722085c33

14 years ago0.7b2 rel_0_7b2
Mike Bayer [Sat, 19 Feb 2011 18:11:34 +0000 (13:11 -0500)] 
0.7b2

14 years agomerge everything all over again...
Mike Bayer [Sat, 19 Feb 2011 18:10:46 +0000 (13:10 -0500)] 
merge everything all over again...

14 years ago - Fix error message referencing old @classproperty
Mike Bayer [Sat, 19 Feb 2011 17:59:12 +0000 (12:59 -0500)] 
  - Fix error message referencing old @classproperty
    name to reference @declared_attr [ticket:2061]

14 years ago- pulling out the PG type impls in create table knocked the counts out a bit (I am...
Mike Bayer [Sat, 19 Feb 2011 02:32:39 +0000 (21:32 -0500)] 
- pulling out the PG type impls in create table knocked the counts out a bit (I am assuming/guessing).

14 years ago-adjust the fix for [ticket:2065] to not rely upon type affinity, revert
Mike Bayer [Fri, 18 Feb 2011 02:05:24 +0000 (21:05 -0500)] 
-adjust the fix for [ticket:2065] to not rely upon type affinity, revert
the _type_affinity attribute of SmallInteger, BigInteger

14 years ago- use easy to recognize names for backref events [ticket:2066]
Mike Bayer [Fri, 18 Feb 2011 01:52:07 +0000 (20:52 -0500)] 
- use easy to recognize names for backref events [ticket:2066]

14 years ago- Fixed regression from 0.6 where SMALLINT and
Mike Bayer [Fri, 18 Feb 2011 01:43:17 +0000 (20:43 -0500)] 
- Fixed regression from 0.6 where SMALLINT and
BIGINT types would both generate SERIAL
on an integer PK column, instead of
SMALLINT and BIGSERIAL [ticket:2065]

14 years ago- The Column.references() method now returns True
Mike Bayer [Fri, 18 Feb 2011 01:27:35 +0000 (20:27 -0500)] 
- The Column.references() method now returns True
if it has a foreign key referencing the
given column exactly, not just it's parent
table.  [ticket:2064]

14 years ago- changelog for [ticket:2059]
Mike Bayer [Fri, 18 Feb 2011 01:02:48 +0000 (20:02 -0500)] 
- changelog for [ticket:2059]

14 years ago- rename EngineEvents to ConnectionEvents
Mike Bayer [Fri, 18 Feb 2011 00:59:45 +0000 (19:59 -0500)] 
- rename EngineEvents to ConnectionEvents
- simplify connection event model to be inline inside Connection, don't use ad-hoc
subclasses (technically would leak memory for the app that keeps creating engines
and adding events)
- not doing listen-per-connection yet.  this is closer.  overall things
are much simpler now (until we put listen-per-connection in...)

14 years ago- the dictionary at the end of the __table_args__
Mike Bayer [Thu, 17 Feb 2011 17:06:56 +0000 (12:06 -0500)] 
- the dictionary at the end of the __table_args__
tuple is now optional.  [ticket:1468]

14 years agoadd a close to this as PG appears to be hanging
Mike Bayer [Mon, 14 Feb 2011 22:17:29 +0000 (17:17 -0500)] 
add a close to this as PG appears to be hanging

14 years ago- Fixed regression whereby composite() with
Mike Bayer [Mon, 14 Feb 2011 17:51:53 +0000 (12:51 -0500)] 
- Fixed regression whereby composite() with
Column objects placed inline would fail
to initialize.  The Column objects can now
be inline with the composite() or external
and pulled in via name or object ref.
[ticket:2058]

14 years ago- Beaker example now takes into account 'limit'
Mike Bayer [Mon, 14 Feb 2011 03:30:00 +0000 (22:30 -0500)] 
- Beaker example now takes into account 'limit'
and 'offset', bind params within embedded
FROM clauses (like when you use union() or
from_self()) when generating a cache key.

14 years agofix arg in ddl events
Mike Bayer [Mon, 14 Feb 2011 01:21:58 +0000 (20:21 -0500)] 
fix arg in ddl events

14 years ago- Association proxy now has correct behavior for
Mike Bayer [Mon, 14 Feb 2011 01:20:34 +0000 (20:20 -0500)] 
- Association proxy now has correct behavior for
any(), has(), and contains() when proxying
a many-to-one scalar attribute to a one-to-many
collection (i.e. the reverse of the 'typical'
association proxy use case)  [ticket:2054]

14 years ago- The Sequence() construct, when passed a MetaData()
Mike Bayer [Mon, 14 Feb 2011 00:32:00 +0000 (19:32 -0500)] 
- The Sequence() construct, when passed a MetaData()
object via its 'metadata' argument, will be
included in CREATE/DROP statements within
metadata.create_all() and metadata.drop_all(),
including "checkfirst" logic.  [ticket:2055]

14 years agofix this directive for py3k
Mike Bayer [Sun, 13 Feb 2011 03:29:36 +0000 (22:29 -0500)] 
fix this directive for py3k

14 years ago- Fixed bug whereby Session.merge() would call the
Mike Bayer [Sun, 13 Feb 2011 03:20:47 +0000 (22:20 -0500)] 
- Fixed bug whereby Session.merge() would call the
load() event with one too few arguments.
[ticket:2053]

- Added logic which prevents the generation of
events from a MapperExtension or SessionExtension
from generating do-nothing events for all the methods
not overridden. [ticket:2052]

14 years agoAdded tag rel_0_7b1 for changeset 53aab63b6c24dfdcc249ada5c9d712dc42e5ae40
Mike Bayer [Sat, 12 Feb 2011 23:30:46 +0000 (18:30 -0500)] 
Added tag rel_0_7b1 for changeset 53aab63b6c24dfdcc249ada5c9d712dc42e5ae40

14 years ago- remove some remaining sys.modules shenanigans rel_0_7b1
Mike Bayer [Sat, 12 Feb 2011 23:00:45 +0000 (18:00 -0500)] 
- remove some remaining sys.modules shenanigans
- The "sqlalchemy.exceptions" alias in sys.modules
has been removed.   Base SQLA exceptions are
available via "from sqlalchemy import exc".
The "exceptions" alias for "exc" remains in
"sqlalchemy" for now, it's just not patched into
sys.modules.

14 years ago- Added an explicit check for the case that the name
Mike Bayer [Sat, 12 Feb 2011 22:33:15 +0000 (17:33 -0500)] 
- Added an explicit check for the case that the name
'metadata' is used for a column attribute on a
declarative class. [ticket:2050]

14 years ago0.7 will be beta classifier to start
Mike Bayer [Sat, 12 Feb 2011 22:14:34 +0000 (17:14 -0500)] 
0.7 will be beta classifier to start

14 years agoweird non-ascii char
Mike Bayer [Sat, 12 Feb 2011 22:09:59 +0000 (17:09 -0500)] 
weird non-ascii char

14 years agough. OK let is_modified accept true/false
Mike Bayer [Fri, 11 Feb 2011 23:24:36 +0000 (18:24 -0500)] 
ugh. OK let is_modified accept true/false

14 years ago- fix long lines in test_attributes
Mike Bayer [Fri, 11 Feb 2011 22:58:49 +0000 (17:58 -0500)] 
- fix long lines in test_attributes
- add deprecation warning to get_history() when passed
True or False, convert
- rearrange symbol() so we can get sphinx autodata to work

14 years ago- replace all usage of True and False for passive with PASSIVE_NO_INITIALIZE,
Mike Bayer [Fri, 11 Feb 2011 20:37:44 +0000 (15:37 -0500)] 
- replace all usage of True and False for passive with PASSIVE_NO_INITIALIZE,
PASSIVE_OFF, now expresed as non-boolean symbols
- make "passive" available positionally on all get_history() methods,
call it like that

14 years ago- Additional tuning to "many-to-one" relationship
Mike Bayer [Fri, 11 Feb 2011 19:41:29 +0000 (14:41 -0500)] 
- Additional tuning to "many-to-one" relationship
loads during a flush().   A change in version 0.6.6
([ticket:2002]) required that more "unnecessary" m2o
loads during a flush could occur.   Extra loading modes have
been added so that the SQL emitted in this
specific use case is trimmed back, while still
retrieving the information the flush needs in order
to not miss anything.  [ticket:2049]

14 years agomake it more explicit in tests which dialect we want to use for things
Mike Bayer [Fri, 11 Feb 2011 16:24:54 +0000 (11:24 -0500)] 
make it more explicit in tests which dialect we want to use for things

14 years ago- need to limit the list of oracle fn's that don't get parens to a
Mike Bayer [Fri, 11 Feb 2011 05:49:28 +0000 (00:49 -0500)] 
- need to limit the list of oracle fn's that don't get parens to a
fixed list.  window functions need parens

14 years agofix sqlite test
Mike Bayer [Fri, 11 Feb 2011 05:24:38 +0000 (00:24 -0500)] 
fix sqlite test

14 years agofix some tests
Mike Bayer [Fri, 11 Feb 2011 04:37:15 +0000 (23:37 -0500)] 
fix some tests

14 years ago- The concept of associating a ".bind" directly with a
Mike Bayer [Fri, 11 Feb 2011 03:27:45 +0000 (22:27 -0500)] 
- The concept of associating a ".bind" directly with a
ClauseElement has been explicitly moved to Executable,
i.e. the mixin that describes ClauseElements which represent
engine-executable constructs.  This change is an improvement
to internal organization and is unlikely to affect any
real-world usage.  [ticket:2048]

14 years agofix some really egregious long lines
Mike Bayer [Fri, 11 Feb 2011 03:09:26 +0000 (22:09 -0500)] 
fix some really egregious long lines

14 years ago- Added over() function, method to FunctionElement
Mike Bayer [Fri, 11 Feb 2011 02:57:44 +0000 (21:57 -0500)] 
- Added over() function, method to FunctionElement
classes, produces the _Over() construct which
in turn generates "window functions", i.e.
"<window function> OVER (PARTITION BY <partition by>,
ORDER BY <order by>)".
[ticket:1844]

14 years ago- oursql dialect accepts the same "ssl" arguments in
Mike Bayer [Thu, 10 Feb 2011 23:04:54 +0000 (18:04 -0500)] 
- oursql dialect accepts the same "ssl" arguments in
create_engine() as that of MySQLdb.  [ticket:2047]

14 years ago- typo fix [ticket:2017]
Mike Bayer [Thu, 10 Feb 2011 20:24:26 +0000 (15:24 -0500)] 
- typo fix [ticket:2017]

14 years ago- mssql VARBINARY emits 'max' for length when no length specified, as is
Mike Bayer [Thu, 10 Feb 2011 20:22:07 +0000 (15:22 -0500)] 
- mssql VARBINARY emits 'max' for length when no length specified, as is
the case already for VARCHAR, NVARCHAR [ticket:1833]

14 years ago- The path given as the location of a sqlite database is now
Mike Bayer [Thu, 10 Feb 2011 20:03:11 +0000 (15:03 -0500)] 
- The path given as the location of a sqlite database is now
normalized via os.path.abspath(), so that directory changes
within the process don't affect the ultimate location
of a relative file path.  [ticket:2036]

14 years ago- Query.distinct() now accepts column expressions
Mike Bayer [Thu, 10 Feb 2011 19:17:08 +0000 (14:17 -0500)] 
- Query.distinct() now accepts column expressions
  as *args, interpreted by the Postgresql dialect
  as DISTINCT ON (<expr>). [ticket:1069]
- select.distinct() now accepts column expressions
  as *args, interpreted by the Postgresql dialect
  as DISTINCT ON (<expr>).  Note this was already
  available via passing a list to the `distinct`
  keyword argument to select(). [ticket:1069]
- select.prefix_with() accepts multiple expressions
  (i.e. *expr), 'prefix' keyword argument to select()
  accepts a list or tuple.
- Passing a string to the `distinct` keyword argument
  of `select()` for the purpose of emitting special
  MySQL keywords (DISTINCTROW etc.) is deprecated -
  use `prefix_with()` for this.
- put kw arguments to select() in order
- restore docs for _SelectBase, renamed from _SelectBaseMixin

14 years agoformatting fix
Mike Bayer [Thu, 10 Feb 2011 16:43:04 +0000 (11:43 -0500)] 
formatting fix

14 years ago- An exception is raised in the unusual case that an
Mike Bayer [Thu, 10 Feb 2011 16:30:23 +0000 (11:30 -0500)] 
- An exception is raised in the unusual case that an
append or similar event on a collection occurs after
the parent object has been dereferenced, which
prevents the parent from being marked as "dirty"
in the session.  Will commit as a warning in 0.6.
[ticket:2046]

14 years ago- py3k fix regarding new exception system
Mike Bayer [Thu, 10 Feb 2011 00:18:33 +0000 (19:18 -0500)] 
- py3k fix regarding new exception system
- this test keeps throwing a TNS error on Oracle on the buildbot only, runs locally,
seems to be related to some scaling/memory type of issue on the bot

14 years ago- Non-DBAPI errors which occur in the scope of an `execute()`
Mike Bayer [Wed, 9 Feb 2011 23:11:40 +0000 (18:11 -0500)] 
- Non-DBAPI errors which occur in the scope of an `execute()`
call are now wrapped in sqlalchemy.exc.StatementError,
and the text of the SQL statement and repr() of params
is included.  This makes it easier to identify statement
executions which fail before the DBAPI becomes
involved.  [ticket:2015]

14 years ago- figured out the ::autodata directive, can move the docstring for
Mike Bayer [Wed, 9 Feb 2011 21:30:49 +0000 (16:30 -0500)] 
- figured out the ::autodata directive, can move the docstring for
expression.func into the .py module
- added a note about logging only being checked on new connections,
as one user had this issue awhile back, and I suspect it for
a current ML user issue

14 years ago- The compiler extension now supports overriding the default
Mike Bayer [Wed, 9 Feb 2011 20:45:15 +0000 (15:45 -0500)] 
- The compiler extension now supports overriding the default
compilation of expression._BindParamClause including that
the auto-generated binds within the VALUES/SET clause
of an insert()/update() statement will also use the new
compilation rules. [ticket:2042]

14 years ago- Fixed bug where a column with a SQL or server side default
Mike Bayer [Wed, 9 Feb 2011 20:14:54 +0000 (15:14 -0500)] 
- Fixed bug where a column with a SQL or server side default
that was excluded from a mapping with include_properties
or exclude_properties would result in UnmappedColumnError.
[ticket:1995]

14 years ago- add connection and cursor to is_disconnect(). We aren't using it yet,
Mike Bayer [Wed, 9 Feb 2011 20:06:32 +0000 (15:06 -0500)] 
- add connection and cursor to is_disconnect().  We aren't using it yet,
but we'd like to.   Most DBAPIs don't give us anything we can do with it.
Some research was done on psycopg2 and it still seems like they give us
no adequate method (tried connection.closed, cursor.closed, connection.status).
mxodbc claims their .closed attribute will work (but I am skeptical).
- remove beahvior in pool that auto-invalidated a connection when
the cursor failed to create.  That's not the pool's job.  we need the conn
for the error logic.  Can't get any tests to fail, curious why that
behavior was there, guess we'll find out (or not).
- add support for psycopg2 version detection.  even though we have
no use for it yet...
- adjust one of the reconnect tests to work with oracle's
horrendously slow connect speed

14 years ago- Added an additional libpq message to the list of "disconnect"
Mike Bayer [Wed, 9 Feb 2011 15:58:16 +0000 (10:58 -0500)] 
- Added an additional libpq message to the list of "disconnect"
exceptions, "could not receive data from server"
[ticket:2044]

14 years ago- Adjusted flush accounting step to occur before
Mike Bayer [Mon, 7 Feb 2011 21:12:24 +0000 (16:12 -0500)] 
- Adjusted flush accounting step to occur before
the commit in the case of autocommit=True.  This allows
autocommit=True to work appropriately with
expire_on_commit=True, and also allows post-flush session
hooks to operate in the same transactional context
as when autocommit=False.  [ticket:2041]

14 years ago- Session constructor emits a warning when autoflush=True
Mike Bayer [Mon, 7 Feb 2011 18:30:27 +0000 (13:30 -0500)] 
- Session constructor emits a warning when autoflush=True
or expire_on_commit=True when autocommit=True.
[ticket:2041]

14 years agoshow sequence members
Mike Bayer [Sat, 5 Feb 2011 21:51:24 +0000 (16:51 -0500)] 
show sequence members

14 years agotypo
Mike Bayer [Sat, 5 Feb 2011 21:50:52 +0000 (16:50 -0500)] 
typo

14 years ago - Session weak_instance_dict=False is deprecated.
Mike Bayer [Sat, 5 Feb 2011 21:48:17 +0000 (16:48 -0500)] 
  - Session weak_instance_dict=False is deprecated.
    [ticket:1473]

14 years ago- use proper directive here
Mike Bayer [Sat, 5 Feb 2011 21:10:14 +0000 (16:10 -0500)] 
- use proper directive here

14 years ago- A single contains_eager() call across
Mike Bayer [Sat, 5 Feb 2011 21:09:49 +0000 (16:09 -0500)] 
- A single contains_eager() call across
multiple entities will indicate all collections
along that path should load, instead of requiring
distinct contains_eager() calls for each endpoint
(which was never correctly documented).
[ticket:2032]

- The "name" field used in orm.aliased() now renders
in the resulting SQL statement.

14 years ago- apply optimizations to alternate row proxies, [ticket:1787]
Mike Bayer [Fri, 4 Feb 2011 23:33:49 +0000 (18:33 -0500)] 
- apply optimizations to alternate row proxies, [ticket:1787]
- add check to fetchmany() for None, don't send argument if not present,
helps DBAPIs which don't accept "None" for default (ie. pysqlite, maybe others)
- add tests to test_execute to provide 100% coverage for the three alternate
result proxy classes

14 years ago- When explicit sequence execution derives the name
Mike Bayer [Fri, 4 Feb 2011 22:37:36 +0000 (17:37 -0500)] 
- When explicit sequence execution derives the name
of the auto-generated sequence of a SERIAL column,
which currently only occurs if implicit_returning=False,
now accommodates if the table + column name is greater
than 63 characters using the same logic Postgresql uses.
[ticket:1083]

14 years ago- Also emits the correct WHERE criterion
Mike Bayer [Thu, 3 Feb 2011 00:03:20 +0000 (19:03 -0500)] 
- Also emits the correct WHERE criterion
when using single table inheritance. [ticket:2038]

14 years ago- Fixed bug where "middle" class in a polymorphic hierarchy
Mike Bayer [Wed, 2 Feb 2011 23:10:07 +0000 (18:10 -0500)] 
- Fixed bug where "middle" class in a polymorphic hierarchy
would have no 'polymorphic_on' column if it didn't also
specify a 'polymorphic_identity', leading to strange
errors upon refresh, wrong class loaded when querying
from that target. [ticket:2038]

14 years ago- Some adjustments so that Interbase is supported as well.
Mike Bayer [Tue, 1 Feb 2011 21:34:37 +0000 (16:34 -0500)] 
- Some adjustments so that Interbase is supported as well.
FB/Interbase version idents are parsed into a structure
such as (8, 1, 1, 'interbase') or (2, 1, 588, 'firebird')
so they can be distinguished. [ticket:1885]
- fixed relfection of the "autoincrement" flag against a default
placed on the column.

14 years agodocuemnt args for before_parent_attach, after_parent_attach
Mike Bayer [Tue, 1 Feb 2011 20:09:28 +0000 (15:09 -0500)] 
docuemnt args for before_parent_attach, after_parent_attach

14 years ago- schema.copy() functions will copy dispatch, but only for those events
Mike Bayer [Tue, 1 Feb 2011 14:26:36 +0000 (09:26 -0500)] 
- schema.copy() functions will copy dispatch, but only for those events
added with propagate=True.   needs tests
- oracle/fb test schema uses event w/ propagate=True to apply
Sequence to primary key columns with test_needs_autoincrement.
this does actually test the propagate=True flag a bit since it's
needed in the declarative mixin tests where the 'id' column
is copied.

14 years ago- SchemaItem, SchemaType now descend from common type
Mike Bayer [Mon, 31 Jan 2011 01:29:48 +0000 (20:29 -0500)] 
- SchemaItem, SchemaType now descend from common type
SchemaEventTarget, which supplies dispatch
- the dispatch now provides before_parent_attach(),
after_parent_attach(), events which generally bound the _set_parent()
event. [ticket:2037]
- the _on_table_attach mechanism now usually uses the
event dispatch
- fixed class-level event dispatch to propagate to all subclasses, not just
immediate subclasses
- fixed class-level event unpickling to handle more involved
inheritance hierarchies, needed by the new schema event dispatch.
- ForeignKeyConstraint doesn't re-call the column attach event
on ForeignKey objects that are already associated with the correct
Column
- we still need that ImportError on mysqldb CLIENT FLAGS to support
mock DBAPIs

14 years ago - Added some defs to the resultproxy.c extension so that
Mike Bayer [Sun, 30 Jan 2011 20:08:41 +0000 (15:08 -0500)] 
  - Added some defs to the resultproxy.c extension so that
    the extension compiles and runs on Python 2.4.
    [ticket:2023]

14 years ago - A warning is emitted when a joined-table inheriting mapper
Mike Bayer [Sun, 30 Jan 2011 19:18:01 +0000 (14:18 -0500)] 
  - A warning is emitted when a joined-table inheriting mapper
    has no primary keys on the locally mapped table
    (but has pks on the superclass table).  [ticket:2019]

14 years agoyikes, change that name on the Alias class too
Mike Bayer [Sun, 30 Jan 2011 19:12:03 +0000 (14:12 -0500)] 
yikes, change that name on the Alias class too

14 years ago- Added a `name` argument to `Query.subquery()`, to allow
Mike Bayer [Sun, 30 Jan 2011 18:39:08 +0000 (13:39 -0500)] 
- Added a `name` argument to `Query.subquery()`, to allow
a fixed name to be assigned to the alias object.
[ticket:2030]
- changed the kw name 'alias' to 'name' on the alias() standalone
function.
- fixed up some alias/join docstrings

14 years ago - Column.copy(), as used in table.tometadata(), copies the
Mike Bayer [Sun, 30 Jan 2011 18:09:45 +0000 (13:09 -0500)] 
  - Column.copy(), as used in table.tometadata(), copies the
    'doc' attribute.  [ticket:2028]

14 years agofix this to work on 2.4
Mike Bayer [Wed, 26 Jan 2011 16:50:00 +0000 (11:50 -0500)] 
fix this to work on 2.4

14 years ago- New DBAPI support for pymysql, a pure Python port
Mike Bayer [Wed, 26 Jan 2011 16:18:03 +0000 (11:18 -0500)] 
- New DBAPI support for pymysql, a pure Python port
of MySQL-python.  [ticket:1991]

14 years ago- new dialect for Drizzle [ticket:2003]
Mike Bayer [Wed, 26 Jan 2011 15:44:00 +0000 (10:44 -0500)] 
- new dialect for Drizzle [ticket:2003]
- move mysqldb to a connector, can be shared among mysql/drizzle

14 years ago- fix oracle tests
Mike Bayer [Wed, 26 Jan 2011 03:37:18 +0000 (22:37 -0500)] 
- fix oracle tests
- fix some obnoxious type adaption issues resulting from the "adapt must copy" change

14 years agocallcount
Mike Bayer [Tue, 25 Jan 2011 19:08:58 +0000 (14:08 -0500)] 
callcount

14 years agoadd example usage for shard_id in context feature [ticket:2031]
Mike Bayer [Sun, 23 Jan 2011 22:08:57 +0000 (17:08 -0500)] 
add example usage for shard_id in context feature [ticket:2031]

14 years ago- Horizontal shard query places 'shard_id' in
Mike Bayer [Sun, 23 Jan 2011 22:03:19 +0000 (17:03 -0500)] 
- Horizontal shard query places 'shard_id' in
context.attributes where it's accessible by the
"load()" event. [ticket:2031]

14 years agofix some formatting here
Mike Bayer [Fri, 21 Jan 2011 22:16:47 +0000 (17:16 -0500)] 
fix some formatting here

14 years agoand try that again.....
Mike Bayer [Thu, 20 Jan 2011 20:36:40 +0000 (15:36 -0500)] 
and try that again.....

14 years ago- use types.MethodType here for python3 compat
Mike Bayer [Thu, 20 Jan 2011 19:25:15 +0000 (14:25 -0500)] 
- use types.MethodType here for python3 compat

14 years agorename 'frozendict' to 'immutabledict', since 'frozen' implies hashability
Mike Bayer [Thu, 20 Jan 2011 18:32:59 +0000 (13:32 -0500)] 
rename 'frozendict' to 'immutabledict', since 'frozen' implies hashability
like frozenset which isn't really the purpose of 'immutabledict' (could be someday,
in which case, we'd change the name back :) )

14 years agotypos
Mike Bayer [Tue, 18 Jan 2011 22:23:15 +0000 (17:23 -0500)] 
typos

14 years ago- Session.connection(), Session.execute() accept 'bind',
Mike Bayer [Tue, 18 Jan 2011 22:13:36 +0000 (17:13 -0500)] 
- Session.connection(), Session.execute() accept 'bind',
to allow execute/connection operations to participate
in the open transaction of an engine explicitly.
[ticket:1996]
- fix up the docs for connection, execute, which were awful.
- correct usage strings in hybrid

14 years ago- Session.merge() will check the version id of the incoming
Mike Bayer [Tue, 18 Jan 2011 21:34:34 +0000 (16:34 -0500)] 
- Session.merge() will check the version id of the incoming
state against that of the database, assuming the mapping
uses version ids and incoming state has a version_id
assigned, and raise StaleDataError if they don't
match.  [ticket:2027]

14 years ago- move maxdb notes from wiki to the maxdb module
Mike Bayer [Tue, 18 Jan 2011 21:02:02 +0000 (16:02 -0500)] 
- move maxdb notes from wiki to the maxdb module
- move exception docs down to the lower portion of core/orm

14 years ago- tests for hybrid
Mike Bayer [Tue, 18 Jan 2011 01:05:09 +0000 (20:05 -0500)] 
- tests for hybrid
- documentation for hybrid
- rewrite descriptor, synonym, comparable_property documentation

14 years ago- remove old perf/ tests from pre-2007
Mike Bayer [Mon, 17 Jan 2011 01:24:05 +0000 (20:24 -0500)] 
- remove old perf/ tests from pre-2007

14 years ago- the zblog example is obsolete, the tests don't really test it
Mike Bayer [Mon, 17 Jan 2011 01:18:32 +0000 (20:18 -0500)] 
- the zblog example is obsolete, the tests don't really test it
and a key feature of its mapping (the deferred col outside of the select)
doesn't work anyway.
- add a token "deferred on selectable" test to test_mapper.

14 years ago- execution_options() on Connection accepts
Mike Bayer [Sun, 16 Jan 2011 22:04:07 +0000 (17:04 -0500)] 
- execution_options() on Connection accepts
"isolation_level" argument, sets transaction isolation
level for that connection only until returned to the
connection pool, for thsoe backends which support it
(SQLite, Postgresql) [ticket:2001]
- disallow the option on Engine (use isolation_level to create_engine()),
Executable (we don't want to check/set per statement)
- docs

14 years ago- default_isolation_level is equal to isolation_level since it's
Mike Bayer [Sun, 16 Jan 2011 18:19:14 +0000 (13:19 -0500)] 
- default_isolation_level is equal to isolation_level since it's
detected after the initial setting

14 years ago- add dialect.reset_isolation_level(), reverts
Mike Bayer [Sun, 16 Jan 2011 18:15:26 +0000 (13:15 -0500)] 
- add dialect.reset_isolation_level(), reverts
- base not implemented methods on Dialect

14 years ago- factor consistent set_isolation_level(), get_isolation_level()
Mike Bayer [Sun, 16 Jan 2011 18:05:18 +0000 (13:05 -0500)] 
- factor consistent set_isolation_level(), get_isolation_level()
per-connection methods for sqlite, postgresql, psycopg2 dialects
- move isolation test suite to test engines/test_transaction
- preparing for [ticket:2001]

14 years ago- oracle fixes
Mike Bayer [Sun, 16 Jan 2011 00:50:01 +0000 (19:50 -0500)] 
- oracle fixes

14 years agoRegarding reflection for this case, reflection of an int PK col
Mike Bayer [Sat, 15 Jan 2011 21:50:52 +0000 (16:50 -0500)] 
Regarding reflection for this case, reflection of an int PK col
with a server_default sets the "autoincrement" flag to False,
except in the case of a PG SERIAL col where we detected a
sequence default. [ticket:2020] [ticket:2021]

14 years ago- getting slightly more consistent behavior for the edge case of pk columns
Mike Bayer [Sat, 15 Jan 2011 21:42:29 +0000 (16:42 -0500)] 
- getting slightly more consistent behavior for the edge case of pk columns
with server default - autoincrement is now false with any server_default,
so these all return None, applies consistency to [ticket:2020], [ticket:2021].
if prefetch is desired a "default" should be used instead of server_default.

14 years agostupid workaround for pg8000
Mike Bayer [Sat, 15 Jan 2011 18:49:39 +0000 (13:49 -0500)] 
stupid workaround for pg8000

14 years ago- allow aliased() to call .alias() on a selectable, [ticket:2018]
Mike Bayer [Sat, 15 Jan 2011 18:48:58 +0000 (13:48 -0500)] 
- allow aliased() to call .alias() on a selectable, [ticket:2018]

14 years ago- composite now relates to its parent class in terms of MapperProperty,
Mike Bayer [Sat, 15 Jan 2011 18:25:03 +0000 (13:25 -0500)] 
- composite now relates to its parent class in terms of MapperProperty,
not Column.  This allows it to compose any mapped attributes, including
relationship().  [ticket:2024]