]> git.ipfire.org Git - thirdparty/sqlalchemy/sqlalchemy.git/log
thirdparty/sqlalchemy/sqlalchemy.git
13 years ago- [bug] Fixed bug whereby TypeDecorator would
Mike Bayer [Mon, 28 Nov 2011 17:44:56 +0000 (12:44 -0500)] 
- [bug] Fixed bug whereby TypeDecorator would
return a stale value for _type_affinity, when
using a TypeDecorator that "switches" types,
like the CHAR/UUID type.

13 years agofix typo [ticket:2334]
Mike Bayer [Sat, 26 Nov 2011 18:17:21 +0000 (13:17 -0500)] 
fix typo [ticket:2334]

13 years agocontinue removing misleading/old sections, fix things that are just wrong
Mike Bayer [Sat, 26 Nov 2011 18:16:47 +0000 (13:16 -0500)] 
continue removing misleading/old sections, fix things that are just wrong

13 years agomore fixes, more coming
Mike Bayer [Sat, 26 Nov 2011 16:32:40 +0000 (11:32 -0500)] 
more fixes, more coming

13 years ago- shorten classical mapping, remove some not very clear things
Mike Bayer [Sat, 26 Nov 2011 06:26:22 +0000 (01:26 -0500)] 
- shorten classical mapping, remove some not very clear things
- separate the "prefix" thing

13 years agosome more ignores
Mike Bayer [Fri, 25 Nov 2011 22:00:19 +0000 (17:00 -0500)] 
some more ignores

13 years agodoc styling rework
Mike Bayer [Fri, 25 Nov 2011 21:17:16 +0000 (16:17 -0500)] 
doc styling rework

13 years ago- doc updates in loading
Mike Bayer [Thu, 24 Nov 2011 15:47:10 +0000 (10:47 -0500)] 
- doc updates in loading
- de-emphasize contains_alias() which is pretty much a non-use case
- add docs for immediateload which have been missing all this time
- fix links, formatting

13 years agocall the simply police
Mike Bayer [Wed, 23 Nov 2011 00:01:21 +0000 (19:01 -0500)] 
call the simply police

13 years agoalso add support for onupdate as we'd like this to fire off if an UPDATE actually
Mike Bayer [Tue, 22 Nov 2011 23:46:45 +0000 (18:46 -0500)] 
also add support for onupdate as we'd like this to fire off if an UPDATE actually
happens on the table

13 years agochangelog
Mike Bayer [Tue, 22 Nov 2011 23:15:51 +0000 (18:15 -0500)] 
changelog

13 years agomerge tip
Mike Bayer [Tue, 22 Nov 2011 23:05:20 +0000 (18:05 -0500)] 
merge tip

13 years agofixes to actually get tests to pass
Mike Bayer [Tue, 22 Nov 2011 23:05:05 +0000 (18:05 -0500)] 
fixes to actually get tests to pass

13 years ago- commit the unit tests
Mike Bayer [Tue, 22 Nov 2011 22:27:35 +0000 (17:27 -0500)] 
- commit the unit tests
- lots of doc updates to all three of update/insert/delete

13 years agotypo
Mike Bayer [Tue, 22 Nov 2011 19:47:31 +0000 (14:47 -0500)] 
typo

13 years agocleanup
Mike Bayer [Tue, 22 Nov 2011 03:10:49 +0000 (22:10 -0500)] 
cleanup

13 years agopasses for all three, includes multi col system with mysql
Mike Bayer [Tue, 22 Nov 2011 03:00:50 +0000 (22:00 -0500)] 
passes for all three, includes multi col system with mysql

13 years agosort of muscling this out, mysql a PITA
Mike Bayer [Tue, 22 Nov 2011 01:40:31 +0000 (20:40 -0500)] 
sort of muscling this out, mysql a PITA

13 years ago- [bug] further tweak to the fix from [ticket:2261],
Mike Bayer [Sun, 20 Nov 2011 04:28:01 +0000 (23:28 -0500)] 
- [bug] further tweak to the fix from [ticket:2261],
so that generative methods work a bit better
off of cloned (this is almost a non-use case though).
In particular this allows with_only_columns()
to behave more consistently.   Added additional
documentation to with_only_columns() to clarify
expected behavior, which changed as a result
of [ticket:2261].  [ticket:2319]
- document the crap out of with_only_columns, include caveats about
the change, etc.

13 years ago- clarify the purpose of "timezone", alleviate confusion in [ticket:2326]
Mike Bayer [Tue, 15 Nov 2011 00:24:08 +0000 (19:24 -0500)] 
- clarify the purpose of "timezone", alleviate confusion in [ticket:2326]

13 years ago- add tests to try to find the case that [ticket:2221] is looking for.
Mike Bayer [Sun, 13 Nov 2011 19:30:04 +0000 (14:30 -0500)] 
- add tests to try to find the case that [ticket:2221] is looking for.
So far, can't find a test where removing that check makes things better.
Easy to find tests where removing it makes things worse.

13 years ago- [bug] Unicode adjustments allow latest pymysql
Mike Bayer [Wed, 2 Nov 2011 16:34:55 +0000 (12:34 -0400)] 
- [bug] Unicode adjustments allow latest pymysql
(post 0.4) to pass 100% on Python 2.

13 years agofix this for PG, other on update cascade DBs
Mike Bayer [Sun, 30 Oct 2011 19:31:03 +0000 (15:31 -0400)] 
fix this for PG, other on update cascade DBs

13 years ago- attribute system gets a pop() method.
Mike Bayer [Sun, 30 Oct 2011 19:10:56 +0000 (15:10 -0400)] 
- attribute system gets a pop() method.
- remove() on a scalar object will raise if the object
removed is not what was present.
- InstanceState can be pickled if obj() is None; this
to support the other changes in this commit
- only use trackparent flag on attributes if
single_parent or ONETOMANY; otherwise we can skip this
overhead
- attribute hasparent()/sethasparent() check that trackparent
is set, else their usage is invalid
- [bug] Fixed backref behavior when "popping" the
value off of a many-to-one in response to
a removal from a stale one-to-many - the operation
is skipped, since the many-to-one has since
been updated.  [ticket:2315]

- [bug] After some years of not doing this, added
more granularity to the "is X a parent of Y"
functionality, which is used when determining
if the FK on "Y" needs to be "nulled out" as well
as if "Y" should be deleted with delete-orphan
cascade.   The test now takes into account the
Python identity of the parent as well its identity
key, to see if the last known parent of Y is
definitely X.   If a decision
can't be made, a StaleDataError is raised.  The
conditions where this error is raised are fairly
rare, requiring that the previous parent was
garbage collected, and previously
could very well inappropriately update/delete
a record that's since moved onto a new parent,
though there may be some cases where
"silent success" occurred previously that will now
raise in the face of ambiguity.
Expiring "Y" resets the "parent" tracker, meaning
X.remove(Y) could then end up deleting Y even
if X is stale, but this is the same behavior
as before; it's advised to expire X also in that
case.  [ticket:2264]

13 years ago- [bug] Postgresql dialect memoizes that an ENUM of a
Mike Bayer [Sat, 29 Oct 2011 21:38:56 +0000 (17:38 -0400)] 
- [bug] Postgresql dialect memoizes that an ENUM of a
particular name was processed
during a create/drop sequence.  This allows
a create/drop sequence to work without any
calls to "checkfirst", and also means with
"checkfirst" turned on it only needs to
check for the ENUM once.  [ticket:2311]

13 years ago- [bug] Fixed bug whereby a subclass of a subclass
Mike Bayer [Fri, 28 Oct 2011 21:46:28 +0000 (17:46 -0400)] 
- [bug] Fixed bug whereby a subclass of a subclass
using concrete inheritance in conjunction with
the new ConcreteBase or AbstractConcreteBase
would fail to apply the subclasses deeper than
one level to the "polymorphic loader" of each
base  [ticket:2312]

- [bug] Fixed bug whereby a subclass of a subclass
using the new AbstractConcreteBase would fail
to acquire the correct "base_mapper" attribute
when the "base" mapper was generated, thereby
causing failures later on.  [ticket:2312]

13 years ago- [bug] fixed inappropriate evaluation of user-mapped
Mike Bayer [Fri, 28 Oct 2011 15:53:32 +0000 (11:53 -0400)] 
- [bug] fixed inappropriate evaluation of user-mapped
object in a boolean context within query.get()
[ticket:2310].

13 years ago- [bug] the value of a composite attribute is now
Mike Bayer [Wed, 26 Oct 2011 16:41:18 +0000 (12:41 -0400)] 
- [bug] the value of a composite attribute is now
expired after an insert or update operation, instead
of regenerated in place.  This ensures that a
column value which is expired within a flush
will be loaded first, before the composite
is regenerated using that value.  [ticket:2309]

- [bug] The fix in [ticket:2309] also emits the
"refresh" event when the composite value is
loaded on access, even if all column
values were already present, as is appropriate.
This fixes the "mutable" extension which relies
upon the "load" event to ensure the _parents
dictionary is up to date, fixes [ticket:2308].
Thanks to Scott Torborg for the test case here.

13 years ago - [feature] Added new support for remote "schemas":
Mike Bayer [Sun, 23 Oct 2011 20:57:48 +0000 (16:57 -0400)] 
  - [feature] Added new support for remote "schemas":
    - MetaData() accepts "schema" and "quote_schema"
      arguments, which will be applied to the same-named
      arguments of a Table
      or Sequence which leaves these at their default
      of ``None``.
    - Sequence accepts "quote_schema" argument
    - tometadata() for Table will use the "schema"
      of the incoming MetaData for the new Table
      if the schema argument is explicitly "None"
    - Added CreateSchema and DropSchema DDL
      constructs - these accept just the string
      name of a schema and a "quote" flag.
    - When using default "schema" with MetaData,
      ForeignKey will also assume the "default" schema
      when locating remote table.  This allows the "schema"
      argument on MetaData to be applied to any
      set of Table objects that otherwise don't have
      a "schema".
    - a "has_schema" method has been implemented
      on dialect, but only works on Postgresql so far.
    Courtesy Manlio Perillo, [ticket:1679]

13 years agofix the 2to3 here, again
Mike Bayer [Sun, 23 Oct 2011 19:49:39 +0000 (15:49 -0400)] 
fix the 2to3 here, again

13 years agoBacked out changeset 8ef278ec7557a5b3446807e9bc16624630ffd034
Mike Bayer [Sun, 23 Oct 2011 19:48:51 +0000 (15:48 -0400)] 
Backed out changeset 8ef278ec7557a5b3446807e9bc16624630ffd034

13 years agofix the 2to3 here
Mike Bayer [Sun, 23 Oct 2011 17:42:02 +0000 (13:42 -0400)] 
fix the 2to3 here

13 years ago- Added accessor to types called "python_type",
Mike Bayer [Sun, 23 Oct 2011 17:05:56 +0000 (13:05 -0400)] 
- Added accessor to types called "python_type",
returns the rudimentary Python type object
for a particular TypeEngine instance, if known,
else raises NotImplementedError.  [ticket:77]

13 years ago - Cls.column.collate("some collation") now
Mike Bayer [Sun, 23 Oct 2011 16:02:58 +0000 (12:02 -0400)] 
   - Cls.column.collate("some collation") now
     works.  [ticket:1776]  Also in 0.6.9

13 years ago- Added missing comma to PASSIVE_RETURN_NEVER_SET
Mike Bayer [Thu, 20 Oct 2011 19:00:17 +0000 (15:00 -0400)] 
- Added missing comma to PASSIVE_RETURN_NEVER_SET
symbol [ticket:2304]

13 years ago- Fixed bug in history_meta.py example where
Mike Bayer [Mon, 17 Oct 2011 17:26:26 +0000 (13:26 -0400)] 
- Fixed bug in history_meta.py example where
the "unique" flag was not removed from a
single-table-inheritance subclass which
generates columns to put up onto the base.

13 years agoplus a period. seriously. should i just watch TV today instead of screwing more...
Mike Bayer [Mon, 17 Oct 2011 17:14:04 +0000 (13:14 -0400)] 
plus a period.  seriously.   should i just watch TV today instead of screwing more things up

13 years agoand i screwed it up
Mike Bayer [Mon, 17 Oct 2011 16:58:52 +0000 (12:58 -0400)] 
and i screwed it up

13 years agoupdated conventions on mixins
Mike Bayer [Mon, 17 Oct 2011 16:33:51 +0000 (12:33 -0400)] 
updated conventions on mixins

13 years agoAdded tag rel_0_7_3 for changeset 8aa229cf84cea6934bd1a041f86dadd0304a58b6
Mike Bayer [Sun, 16 Oct 2011 18:30:37 +0000 (14:30 -0400)] 
Added tag rel_0_7_3 for changeset 8aa229cf84cea6934bd1a041f86dadd0304a58b6

13 years agolink to declarative concrete helper doc rel_0_7_3
Mike Bayer [Sun, 16 Oct 2011 18:24:29 +0000 (14:24 -0400)] 
link to declarative concrete helper doc

13 years ago0.7.3 prep
Mike Bayer [Sun, 16 Oct 2011 18:14:45 +0000 (14:14 -0400)] 
0.7.3 prep

13 years agosome mssql stuff, though unicode is really not working still...
Mike Bayer [Sun, 16 Oct 2011 15:34:48 +0000 (11:34 -0400)] 
some mssql stuff, though unicode is really not working still...

13 years agofix this for oracle
Mike Bayer [Sat, 15 Oct 2011 21:17:33 +0000 (17:17 -0400)] 
fix this for oracle

13 years agofix this test for oursql
Mike Bayer [Sat, 15 Oct 2011 19:22:21 +0000 (15:22 -0400)] 
fix this test for oursql

13 years ago- Added client_encoding parameter to create_engine()
Mike Bayer [Sat, 15 Oct 2011 19:05:51 +0000 (15:05 -0400)] 
- Added client_encoding parameter to create_engine()
when the postgresql+psycopg2 dialect is used;
calls the psycopg2 set_client_encoding() method
with the value upon connect.  [ticket:1839]

13 years ago- Reinstated "comparator_factory" argument to
Mike Bayer [Sat, 15 Oct 2011 18:31:02 +0000 (14:31 -0400)] 
- Reinstated "comparator_factory" argument to
composite(), removed when 0.7 was released.
[ticket:2248]

13 years ago- Added an exception catch + warning for the
Mike Bayer [Sat, 15 Oct 2011 16:10:31 +0000 (12:10 -0400)] 
- Added an exception catch + warning for the
"return unicode detection" step within connect,
allows databases that crash on NVARCHAR to
continue initializing, assuming no NVARCHAR
type implemented.  [ticket:2299]

13 years agofix typos [ticket:2294] [ticket:2300]. these are both due to reworkings of 0.6 docs...
Mike Bayer [Sat, 15 Oct 2011 15:24:43 +0000 (11:24 -0400)] 
fix typos [ticket:2294] [ticket:2300].  these are both due to reworkings of 0.6 docs that don't have these typos

13 years ago- Improved query.join() such that the "left" side
Mike Bayer [Wed, 12 Oct 2011 19:15:28 +0000 (15:15 -0400)] 
- Improved query.join() such that the "left" side
can more flexibly be a non-ORM selectable,
such as a subquery.   A selectable placed
in select_from() will now be used as the left
side, favored over implicit usage
of a mapped entity.
If the join still fails based on lack of
foreign keys, the error message includes
this detail.  Thanks to brianrhude
on IRC for the test case.  [ticket:2298]

13 years ago - Added new flag expire_on_flush=False to column_property(),
Mike Bayer [Tue, 11 Oct 2011 18:15:09 +0000 (14:15 -0400)] 
  - Added new flag expire_on_flush=False to column_property(),
    marks those properties that would otherwise be considered
    to be "readonly", i.e. derived from SQL expressions,
    to retain their value after a flush has occurred, including
    if the parent object itself was involved in an update.

13 years ago- adjust some tests and such to work better with a mysql 5.5 install
Mike Bayer [Tue, 11 Oct 2011 04:15:43 +0000 (00:15 -0400)] 
- adjust some tests and such to work better with a mysql 5.5 install
  - Added mysql_length parameter to Index construct,
    specifies "length" for indexes.  [ticket:2293]

13 years ago - Fixed bug related to [ticket:2141] whereby the
Mike Bayer [Wed, 5 Oct 2011 15:47:19 +0000 (11:47 -0400)] 
  - Fixed bug related to [ticket:2141] whereby the
    same modified index behavior in PG 9 affected
    primary key reflection on a renamed column.
    [ticket:2291].  Also in 0.6.9.

13 years agoadd a test to confirm alter table of pk column gets reflected
Mike Bayer [Tue, 4 Oct 2011 16:30:23 +0000 (12:30 -0400)] 
add a test to confirm alter table of pk column gets reflected

13 years agoundo that change with the _, it doesn't break round trips
Mike Bayer [Mon, 3 Oct 2011 19:24:26 +0000 (15:24 -0400)] 
undo that change with the _, it doesn't break round trips

13 years ago - a CREATE TABLE will put the COLLATE option
Mike Bayer [Mon, 3 Oct 2011 19:19:07 +0000 (15:19 -0400)] 
  - a CREATE TABLE will put the COLLATE option
    after CHARSET, which appears to be part of
    MySQL's arbitrary rules regarding if it will actually
    work or not.  [ticket:2225]

  - reflecting a MySQL table will ensure that the
    options added to the Table at the table.kwargs
    level have spaces converted to underscores.
    This is a slight behavioral change specifically
    to the "mysql_default_charset" option which
    previously would not be symmetrical.

13 years ago - Added "postgresql_using" argument to Index(), produces
Mike Bayer [Sun, 2 Oct 2011 21:17:46 +0000 (17:17 -0400)] 
  - Added "postgresql_using" argument to Index(), produces
    USING clause to specify index implementation for
    PG.  [ticket:2290].  Thanks to Ryan P. Kelly for
    the patch.

13 years ago - Enhanced the instrumentation in the ORM to support
Mike Bayer [Wed, 28 Sep 2011 17:04:42 +0000 (13:04 -0400)] 
  - Enhanced the instrumentation in the ORM to support
    Py3K's new argument style of "required kw arguments",
    i.e. fn(a, b, *, c, d), fn(a, b, *args, c, d).
    Argument signatures of mapped object's __init__
    method will be preserved, including required kw rules.
    [ticket:2237]

13 years ago- Fixed bug whereby mapper.order_by attribute would
Mike Bayer [Wed, 28 Sep 2011 13:10:23 +0000 (09:10 -0400)] 
- Fixed bug whereby mapper.order_by attribute would
be ignored in the "inner" query within a
subquery eager load.  [ticket:2287].
Also in 0.6.9.

13 years ago - SQLSoup will not be included in version 0.8
Mike Bayer [Mon, 26 Sep 2011 21:58:30 +0000 (17:58 -0400)] 
  - SQLSoup will not be included in version 0.8
    of SQLAlchemy; while useful, we would like to
    keep SQLAlchemy itself focused on one ORM
    usage paradigm.  SQLSoup will hopefully
    soon be superseded by a third party
    project.  [ticket:2262]

13 years agochangelog
Mike Bayer [Mon, 26 Sep 2011 21:45:09 +0000 (17:45 -0400)] 
changelog

13 years agowarn when a subclass' base uses @declared_attr for a regular column -
Mike Bayer [Mon, 26 Sep 2011 21:43:53 +0000 (17:43 -0400)] 
warn when a subclass' base uses @declared_attr for a regular column -
does not propagate to subclasses.  [ticket:2283]

13 years agoimproved docs on select_from() [ticket:2274]
Mike Bayer [Mon, 26 Sep 2011 21:33:01 +0000 (17:33 -0400)] 
improved docs on select_from() [ticket:2274]

13 years ago - Adjusted dictlike-polymorphic.py example
Mike Bayer [Mon, 26 Sep 2011 21:22:52 +0000 (17:22 -0400)] 
  - Adjusted dictlike-polymorphic.py example
    to apply the CAST such that it works on
    PG, other databases.  [ticket:2266]

13 years agofix test
Mike Bayer [Sat, 24 Sep 2011 16:13:45 +0000 (12:13 -0400)] 
fix test

13 years agofix tests here, mssql/mysql are mutually exclusive about x IN/= <subquery> at least...
Mike Bayer [Sat, 24 Sep 2011 15:16:54 +0000 (11:16 -0400)] 
fix tests here, mssql/mysql are mutually exclusive about x IN/= <subquery> at least when LIMIT is present

13 years ago- When an open Session is garbage collected, the objects
Mike Bayer [Sat, 24 Sep 2011 15:12:34 +0000 (11:12 -0400)] 
- When an open Session is garbage collected, the objects
within it which remain are considered detached again
when they are add()-ed to a new Session.
This is accomplished by an extra check that the previous
"session_key" doesn't actually exist among the pool
of Sessions.  [ticket:2281]

13 years ago - The String type now generates VARCHAR2 on Oracle
Mike Bayer [Sat, 24 Sep 2011 15:02:29 +0000 (11:02 -0400)] 
  - The String type now generates VARCHAR2 on Oracle
    which is recommended as the default VARCHAR.
    Added an explicit VARCHAR2 and NVARCHAR2 to the Oracle
    dialect as well.   Using NVARCHAR still generates
    "NVARCHAR2" - there is no "NVARCHAR" on Oracle -
    this remains a slight breakage of the "uppercase types
    always give exactly that" policy.  VARCHAR still
    generates "VARCHAR", keeping with the policy.   If
    Oracle were to ever define "VARCHAR" as something
    different as they claim (IMHO this will never happen),
    the type would be available.  [ticket:2252]

13 years ago- The integer "id" used to link a mapped instance with
Mike Bayer [Sat, 24 Sep 2011 14:32:39 +0000 (10:32 -0400)] 
- The integer "id" used to link a mapped instance with
its owning Session is now generated by a sequence
generation function rather than id(Session), to
eliminate the possibility of recycled id() values
causing an incorrect result, no need to check that
object actually in the session.  [ticket:2280]

13 years ago - The entry point resolution supported by
Mike Bayer [Sat, 24 Sep 2011 13:01:31 +0000 (09:01 -0400)] 
  - The entry point resolution supported by
    create_engine() now supports resolution of
    individual DBAPI drivers on top of a built-in
    or entry point-resolved dialect, using the
    standard '+' notation - it's converted to
    a '.' before being resolved as an entry
    point.  [ticket:2286]

13 years agomore specific
Mike Bayer [Sat, 24 Sep 2011 02:25:31 +0000 (22:25 -0400)] 
more specific

13 years agodocument CircularDependencyError. [ticket:2285]
Mike Bayer [Sat, 24 Sep 2011 02:17:18 +0000 (22:17 -0400)] 
document CircularDependencyError.  [ticket:2285]

13 years ago - The behavior of =/!= when comparing a scalar select
Mike Bayer [Sat, 24 Sep 2011 01:56:10 +0000 (21:56 -0400)] 
  - The behavior of =/!= when comparing a scalar select
    to a value will no longer produce IN/NOT IN as of 0.8;
    this behavior is a little too heavy handed (use in_() if
    you want to emit IN) and now emits a deprecation warning.
    To get the 0.8 behavior immediately and remove the warning,
    a compiler recipe is given at
    http://www.sqlalchemy.org/docs/07/dialects/mssql.html#scalar-select-comparisons
    to override the behavior of visit_binary().
    [ticket:2277]

13 years ago - Modified Column.copy() to use _constructor(),
Mike Bayer [Sat, 24 Sep 2011 00:51:58 +0000 (20:51 -0400)] 
  - Modified Column.copy() to use _constructor(),
    which defaults to self.__class__, in order to
    create the new object.  This allows easier support
    of subclassing Column.  [ticket:2284]

13 years ago - Fixed bug in unit of work whereby detection of
Mike Bayer [Thu, 22 Sep 2011 23:21:39 +0000 (19:21 -0400)] 
  - Fixed bug in unit of work whereby detection of
    "cycles" among classes in highly interlinked patterns
    would not produce a deterministic
    result; thereby sometimes missing some nodes that
    should be considered cycles and causing further
    issues down the road.  Note this bug is in 0.6
    also; not backported at the moment.
    [ticket:2282]

13 years agoremove inline imports from core events too
Mike Bayer [Thu, 22 Sep 2011 15:55:15 +0000 (11:55 -0400)] 
remove inline imports from core events too

13 years agoremove all inline imports statements from events.py, [ticket:2279]
Mike Bayer [Thu, 22 Sep 2011 15:52:23 +0000 (11:52 -0400)] 
remove all inline imports statements from events.py, [ticket:2279]

13 years agoadd a document note regarding CAST and the workaround in [ticket:2242]
Mike Bayer [Wed, 21 Sep 2011 22:02:31 +0000 (18:02 -0400)] 
add a document note regarding CAST and the workaround in [ticket:2242]

13 years agono this is not in 0.6.9, it was an 0.7 regression
Mike Bayer [Wed, 21 Sep 2011 21:38:57 +0000 (17:38 -0400)] 
no this is not in 0.6.9, it was an 0.7 regression

13 years ago- Fixed ReturningResultProxy for zxjdbc dialect.
Mike Bayer [Wed, 21 Sep 2011 21:33:04 +0000 (17:33 -0400)] 
- Fixed ReturningResultProxy for zxjdbc dialect.
[ticket:2272].  Also in 0.6.9

13 years ago - Fixed bug whereby with_only_columns() method of
Mike Bayer [Wed, 21 Sep 2011 21:08:08 +0000 (17:08 -0400)] 
  - Fixed bug whereby with_only_columns() method of
    Select would fail if a selectable were passed.
    [ticket:2270].  Also in 0.6.9.

13 years ago - Adjusted the "importlater" mechanism, which is
Mike Bayer [Wed, 21 Sep 2011 20:56:14 +0000 (16:56 -0400)] 
   - Adjusted the "importlater" mechanism, which is
     used internally to resolve import cycles,
     such that the usage of __import__ is completed
     when the import of sqlalchemy or sqlalchemy.orm
     is done, thereby avoiding any usage of __import__
     after the application starts new threads,
     fixes [ticket:2279].  Also in 0.6.9.

13 years agoensure sa_pool_key isn't passed to connect
Mike Bayer [Wed, 21 Sep 2011 14:26:49 +0000 (10:26 -0400)] 
ensure sa_pool_key isn't passed to connect

13 years ago - Added optional "sa_pool_key" argument to
Mike Bayer [Tue, 20 Sep 2011 16:36:30 +0000 (12:36 -0400)] 
  - Added optional "sa_pool_key" argument to
    pool.manage(dbapi).connect() so that serialization
    of args is not necessary.

13 years agoplus synonyms accessed as string
Mike Bayer [Tue, 20 Sep 2011 02:13:24 +0000 (22:13 -0400)] 
plus synonyms accessed as string

13 years ago - added "adapt_on_names" boolean flag to orm.aliased()
Mike Bayer [Mon, 19 Sep 2011 20:48:39 +0000 (16:48 -0400)] 
  - added "adapt_on_names" boolean flag to orm.aliased()
    construct.  Allows an aliased() construct
    to link the ORM entity to a selectable that contains
    aggregates or other derived forms of a particular
    attribute, provided the name is the same as that
    of the entity mapped column.

13 years ago - Changes to attempt support of FreeTDS 0.91 with
Mike Bayer [Sun, 18 Sep 2011 19:33:12 +0000 (15:33 -0400)] 
  - Changes to attempt support of FreeTDS 0.91 with
    Pyodbc.  This includes that string binds are sent as
    Python unicode objects when FreeTDS 0.91 is detected,
    and a CAST(? AS NVARCHAR) is used when we detect
    for a table.   However, I'd continue
    to characterize Pyodbc + FreeTDS 0.91 behavior as
    pretty crappy, there are still many queries such
    as used in reflection which cause a core dump on
    Linux, and it is not really usable at all
    on OSX, MemoryErrors abound and just plain broken
    unicode support.   [ticket:2273]

13 years agoordering
Mike Bayer [Fri, 16 Sep 2011 02:26:12 +0000 (22:26 -0400)] 
ordering

13 years ago - Fixed a variety of synonym()-related regressions
Mike Bayer [Fri, 16 Sep 2011 00:18:26 +0000 (20:18 -0400)] 
  - Fixed a variety of synonym()-related regressions
    from 0.6:
        - making a synonym against a synonym now works.
        - synonyms made against a relationship() can
          be passed to query.join(), options sent
          to query.options(), passed by name
          to query.with_parent().
- revised the approach taken earlier to just
propagate "property" out from the proxied attr the same
way queryable attribute does.

13 years ago - Another previously unknown feature from 0.6, synonyms
Mike Bayer [Thu, 15 Sep 2011 21:58:01 +0000 (17:58 -0400)] 
  - Another previously unknown feature from 0.6, synonyms
    of relationship() can again be passed to join().

13 years ago- Fixed previously untested function which regressed
Mike Bayer [Thu, 15 Sep 2011 20:23:15 +0000 (16:23 -0400)] 
- Fixed previously untested function which regressed
in 0.7, can now make a synonym() of a synonym()
again.

13 years ago - Changed the update() method on association proxy
Mike Bayer [Wed, 14 Sep 2011 15:31:33 +0000 (11:31 -0400)] 
  - Changed the update() method on association proxy
    dictionary to use a duck typing approach, i.e.
    checks for "keys", to discern between update({})
    and update((a, b)).   Previously, passing a
    dictionary that had tuples as keys would be misinterpreted
    as a sequence. [ticket:2275]

13 years agofix sentence
Mike Bayer [Sat, 10 Sep 2011 23:10:18 +0000 (19:10 -0400)] 
fix sentence

13 years ago - New event hook, MapperEvents.after_configured().
Mike Bayer [Sat, 10 Sep 2011 20:54:23 +0000 (16:54 -0400)] 
  - New event hook, MapperEvents.after_configured().
    Called after a configure() step has completed and
    mappers were in fact affected.   Theoretically this
    event is called once per application, unless new mappings
    are constructed after existing ones have been used
    already.

  - New declarative features:
        - __declare_last__() method, establishes an event
        listener for the class method that will be called
        when mappers are completed with the final "configure"
        step.
        - __abstract__ flag.   The class will not be mapped
        at all when this flag is present on the class.
        - New helper classes ConcreteBase, AbstractConcreteBase.
        Allow concrete mappings using declarative which automatically
        set up the "polymorphic_union" when the "configure"
        mapper step is invoked.
        - The mapper itself has semi-private methods that allow
        the "with_polymorphic" selectable to be assigned
        to the mapper after it has already been configured.
        [ticket:2239]

13 years agominor callcount bump resulting from r3876c9c0b9a4
Mike Bayer [Fri, 9 Sep 2011 21:18:10 +0000 (17:18 -0400)] 
minor callcount bump resulting from r3876c9c0b9a4

13 years ago - Reflection functions for Table, Sequence no longer
Mike Bayer [Fri, 9 Sep 2011 21:12:52 +0000 (17:12 -0400)] 
  - Reflection functions for Table, Sequence no longer
    case insensitive.  Names can be differ only in case
    and will be correctly distinguished.  [ticket:2256]

13 years ago- Calling class_mapper() and passing in an object
Mike Bayer [Fri, 9 Sep 2011 20:18:44 +0000 (16:18 -0400)] 
- Calling class_mapper() and passing in an object
that is not a "type" (i.e. a class that could
potentially be mapped) now raises an informative
ArgumentError, rather than UnmappedClassError.
[ticket:2196]

13 years ago - Identity map .discard() uses dict.pop(,None)
Mike Bayer [Fri, 9 Sep 2011 20:11:13 +0000 (16:11 -0400)] 
  - Identity map .discard() uses dict.pop(,None)
    internally instead of "del" to avoid KeyError/warning
    during a non-determinate gc teardown [ticket:2267]

13 years ago - Behavioral improvement: empty
Mike Bayer [Fri, 9 Sep 2011 20:03:23 +0000 (16:03 -0400)] 
  - Behavioral improvement: empty
    conjunctions such as and_() and or_() will be
    flattened in the context of an enclosing conjunction,
    i.e. and_(x, or_()) will produce 'X' and not 'X AND
    ()'. [ticket:2257].

13 years ago- Fixed bug whereby if __eq__() was
Mike Bayer [Fri, 9 Sep 2011 19:51:40 +0000 (15:51 -0400)] 
- Fixed bug whereby if __eq__() was
redefined, a relationship many-to-one lazyload
would hit the __eq__() and fail.  [ticket:2260]
Does not apply to 0.6.9.