]> git.ipfire.org Git - thirdparty/sqlalchemy/sqlalchemy.git/log
thirdparty/sqlalchemy/sqlalchemy.git
17 years agoadd pydoc for from_statement().
Mike Bayer [Sun, 9 Dec 2007 17:50:24 +0000 (17:50 +0000)] 
add pydoc for from_statement().

17 years ago- added a warning when a relation() is added to an inheriting mapper that is present...
Mike Bayer [Sun, 9 Dec 2007 17:26:16 +0000 (17:26 +0000)] 
- added a warning when a relation() is added to an inheriting mapper that is present on a super-mapper; multiple DependencyProcessors are not expected during the flush process
- found an uncovered line in uow, was "covered" by one particular breaking test

17 years agoadd some updates too
Mike Bayer [Sun, 9 Dec 2007 16:13:27 +0000 (16:13 +0000)] 
add some updates too

17 years ago- added a test for boolean saves/retrieves
Mike Bayer [Sun, 9 Dec 2007 16:11:21 +0000 (16:11 +0000)] 
- added a test for boolean saves/retrieves

17 years ago- fix up the fixtures comparator
Mike Bayer [Sun, 9 Dec 2007 15:56:37 +0000 (15:56 +0000)] 
- fix up the fixtures comparator
- strengthened memory profiling test

17 years agoonly report true for source change if added + deleted, dont pick up inserts
Mike Bayer [Sun, 9 Dec 2007 15:04:14 +0000 (15:04 +0000)] 
only report true for source change if added + deleted, dont pick up inserts

17 years agomapper uses attributes to get non-cached history
Mike Bayer [Sun, 9 Dec 2007 05:24:07 +0000 (05:24 +0000)] 
mapper uses attributes to get non-cached history

17 years agoremove redundant identity map set
Mike Bayer [Sun, 9 Dec 2007 05:14:40 +0000 (05:14 +0000)] 
remove redundant identity map set

17 years ago- mutable primary key support is added. primary key columns can be
Mike Bayer [Sun, 9 Dec 2007 05:00:12 +0000 (05:00 +0000)] 
- mutable primary key support is added. primary key columns can be
changed freely, and the identity of the instance will change upon
flush. In addition, update cascades of foreign key referents (primary
key or not) along relations are supported, either in tandem with the
database's ON UPDATE CASCADE (required for DB's like Postgres) or
issued directly by the ORM in the form of UPDATE statements, by setting
the flag "passive_cascades=False".

17 years ago- added new methods to TypeDecorator, process_bind_param() and
Mike Bayer [Sat, 8 Dec 2007 23:03:22 +0000 (23:03 +0000)] 
- added new methods to TypeDecorator, process_bind_param() and
process_result_value(), which automatically take advantage of the processing
of the underlying type.  Ideal for using with Unicode or Pickletype.
TypeDecorator should now be the primary way to augment the behavior of any
existing type including other TypeDecorator subclasses such as PickleType.

17 years agosome updates, remove an old test (really crufty old stuff here)
Mike Bayer [Sat, 8 Dec 2007 21:49:03 +0000 (21:49 +0000)] 
some updates, remove an old test (really crufty old stuff here)

17 years ago[ticket:888] continued, synonym add_prop
Mike Bayer [Sat, 8 Dec 2007 19:56:11 +0000 (19:56 +0000)] 
[ticket:888] continued, synonym add_prop

17 years agook found it
Mike Bayer [Sat, 8 Dec 2007 19:25:10 +0000 (19:25 +0000)] 
ok found it

17 years agofixed test which didnt pass along 'allitems' collection to the sort...
Mike Bayer [Sat, 8 Dec 2007 19:15:12 +0000 (19:15 +0000)] 
fixed test which didnt pass along 'allitems' collection to the sort...

17 years ago- flush() refactor merged from uow_nontree branch r3871-r3885
Mike Bayer [Sat, 8 Dec 2007 18:58:03 +0000 (18:58 +0000)] 
- flush() refactor merged from uow_nontree branch r3871-r3885
- topological.py cleaned up, presents three public facing functions which
return list/tuple based structures, without exposing any internals.  only
the third function returns the "hierarchical" structure.  when results
include "cycles" or "child" items, 2- or 3- tuples are used to represent
results.
- unitofwork uses InstanceState almost exclusively now.  new and deleted lists
are now dicts which ref the actual object to provide a strong ref for the
duration that they're in those lists.  IdentitySet is only used for the public
facing versions of "new" and "deleted".
- unitofwork topological sort no longer uses the "hierarchical" version of the sort
for the base sort, only for the "per-object" secondary sort where it still
helps to group non-dependent operations together and provides expected insert
order.  the default sort deals with UOWTasks in a straight list and is greatly
simplified.  Tests all pass but need to see if svilen's stuff still works,
one block of code in _sort_cyclical_dependencies() seems to not be needed anywhere
but i definitely put it there for a reason at some point; if not hopefully we
can derive more test coverage from that.
- the UOWEventHandler is only applied to object-storing attributes, not
scalar (i.e. column-based) ones.  cuts out a ton of overhead when setting
non-object based attributes.
- InstanceState also used throughout the flush process, i.e. dependency.py,
mapper.save_obj()/delete_obj(), sync.execute() all expect InstanceState objects
in most cases now.
- mapper/property cascade_iterator() takes InstanceState as its argument,
but still returns lists of object instances so that they are not dereferenced.
- a few tricks needed when dealing with InstanceState, i.e. when loading a list
of items that are possibly fresh from the DB, you *have* to get the actual objects
into a strong-referencing datastructure else they fall out of scope immediately.
dependency.py caches lists of dependent objects which it loads now (i.e. history
collections).
- AttributeHistory is gone, replaced by a function that returns a 3-tuple of
added, unchanged, deleted.  these collections still reference the object
instances directly for the strong-referencing reasons mentiontioned, but
it uses less IdentitySet logic to generate.

17 years agochanged the anonymous numbering scheme to be more appealing
Mike Bayer [Sat, 8 Dec 2007 18:38:18 +0000 (18:38 +0000)] 
changed the anonymous numbering scheme to be more appealing
got tests running

17 years agocorrected for current output...
Mike Bayer [Sat, 8 Dec 2007 18:04:46 +0000 (18:04 +0000)] 
corrected for current output...

17 years agoadded a mention about `eagerload_all()` [ticket:897]
Mike Bayer [Sat, 8 Dec 2007 17:07:40 +0000 (17:07 +0000)] 
added a mention about `eagerload_all()` [ticket:897]

17 years agomapper.instances() is deprecated
Mike Bayer [Fri, 7 Dec 2007 20:14:10 +0000 (20:14 +0000)] 
mapper.instances() is deprecated

17 years agoa little refinement to topological options, more to come
Mike Bayer [Fri, 7 Dec 2007 17:27:21 +0000 (17:27 +0000)] 
a little refinement to topological options, more to come

17 years agofix to unique bind params, you *can* use the same unique bindparam multiple times
Mike Bayer [Fri, 7 Dec 2007 16:47:00 +0000 (16:47 +0000)] 
fix to unique bind params, you *can* use the same unique bindparam multiple times
in a statement.  the collision check is strictly detecting non-unique's that happen to have
the same name.

17 years ago- also with dynamic, implemented correct count() behavior as well
Mike Bayer [Fri, 7 Dec 2007 16:31:12 +0000 (16:31 +0000)] 
- also with dynamic, implemented correct count() behavior as well
as other helper methods.

17 years ago- added "cascade delete" behavior to "dynamic" relations just like
Mike Bayer [Fri, 7 Dec 2007 16:13:19 +0000 (16:13 +0000)] 
- added "cascade delete" behavior to "dynamic" relations just like
that of regular relations.  if passive_deletes flag (also just added)
is not set, a delete of the parent item will trigger a full load of
the child items so that they can be deleted or updated accordingly.

17 years agofixed key error when no pks could be located
Mike Bayer [Fri, 7 Dec 2007 15:28:20 +0000 (15:28 +0000)] 
fixed key error when no pks could be located

17 years ago- generation of "unique" bind parameters has been simplified to use the same
Mike Bayer [Thu, 6 Dec 2007 22:23:10 +0000 (22:23 +0000)] 
- generation of "unique" bind parameters has been simplified to use the same
"unique identifier" mechanisms as everything else.  This doesn't affect
user code, except any code that might have been hardcoded against the generated
names.  Generated bind params now have the form "<paramname>_<num>",
whereas before only the second bind of the same name would have this form.

- bindparam() objects themselves can be used as keys for execute(), i.e.
statement.execute({bind1:'foo', bind2:'bar'})

17 years agowarn if query.get() used with existing criterion
Mike Bayer [Thu, 6 Dec 2007 19:31:59 +0000 (19:31 +0000)] 
warn if query.get() used with existing criterion

17 years ago- query.get() and query.load() do not take existing filter or other
Mike Bayer [Thu, 6 Dec 2007 19:15:04 +0000 (19:15 +0000)] 
- query.get() and query.load() do not take existing filter or other
criterion into account; these methods *always* look up the given id
in the database or return the current instance from the identity map,
disregarding any existing filter, join, group_by or other criterion
which has been configured. [ticket:893]

17 years agoassert_unicode=True replaced with default of assert_unicode='warn'
Mike Bayer [Wed, 5 Dec 2007 22:21:37 +0000 (22:21 +0000)] 
assert_unicode=True replaced with default of assert_unicode='warn'

17 years ago- ordering of cols in pks_by_table and cols_by_table is significant;
Mike Bayer [Wed, 5 Dec 2007 22:04:27 +0000 (22:04 +0000)] 
- ordering of cols in pks_by_table and cols_by_table is significant;
particularly for pks_by_table the ordering is expected to match the ordering
of pk columns in the table for usage in query.get() as well as identity key
generation

17 years agomore assertion hints...
Mike Bayer [Wed, 5 Dec 2007 21:51:16 +0000 (21:51 +0000)] 
more assertion hints...

17 years agoimproved assertions, test is failing on the buildbot only (not osx or linux)
Mike Bayer [Wed, 5 Dec 2007 21:34:22 +0000 (21:34 +0000)] 
improved assertions, test is failing on the buildbot only (not osx or linux)

17 years agosome more omit schemas for [ticket:890]
Mike Bayer [Wed, 5 Dec 2007 21:16:02 +0000 (21:16 +0000)] 
some more omit schemas for [ticket:890]

17 years ago- tables with schemas can still be used in sqlite, firebird,
Mike Bayer [Wed, 5 Dec 2007 21:14:09 +0000 (21:14 +0000)] 
- tables with schemas can still be used in sqlite, firebird,
schema name just gets dropped [ticket:890]

17 years ago- fixed wrong varname in session exception throw
Mike Bayer [Wed, 5 Dec 2007 20:55:33 +0000 (20:55 +0000)] 
- fixed wrong varname in session exception throw
- fixed vertical example to just use a scoped session

17 years ago- a major behavioral change to collection-based backrefs: they no
Mike Bayer [Wed, 5 Dec 2007 20:43:16 +0000 (20:43 +0000)] 
- a major behavioral change to collection-based backrefs: they no
longer trigger lazy loads !  "reverse" adds and removes
are queued up and are merged with the collection when it is
actually read from and loaded; but do not trigger a load beforehand.
For users who have noticed this behavior, this should be much more
convenient than using dynamic relations in some cases; for those who
have not, you might notice your apps using a lot fewer queries than
before in some situations. [ticket:871]

17 years ago- basic framework for generic functions, [ticket:615]
Mike Bayer [Wed, 5 Dec 2007 03:07:21 +0000 (03:07 +0000)] 
- basic framework for generic functions, [ticket:615]
- changed the various "literal" generation functions to use an anonymous
bind parameter.  not much changes here except their labels now look
like ":param_1", ":param_2" instead of ":literal"
- from_obj keyword argument to select() can be a scalar or a list.

17 years ago- fixed backref bug where you could not del instance.attr if attr
Mike Bayer [Tue, 4 Dec 2007 17:55:25 +0000 (17:55 +0000)] 
- fixed backref bug where you could not del instance.attr if attr
was None

17 years ago- relaxed rules on column_property() expressions having labels; any
Mike Bayer [Tue, 4 Dec 2007 17:06:55 +0000 (17:06 +0000)] 
- relaxed rules on column_property() expressions having labels; any
ColumnElement is accepted now, as the compiler auto-labels non-labeled
ColumnElements now.  a selectable, like a select() statement, still
requires conversion to ColumnElement via as_scalar() or label().

17 years agotypo
Gaëtan de Menten [Tue, 4 Dec 2007 16:22:44 +0000 (16:22 +0000)] 
typo

17 years agofixed replacement of existing column properties with synonyms, [ticket:888]
Mike Bayer [Tue, 4 Dec 2007 15:53:23 +0000 (15:53 +0000)] 
fixed replacement of existing column properties with synonyms, [ticket:888]

17 years ago- moved class-level attributes placed by the attributes package into a _class_state
Mike Bayer [Sun, 2 Dec 2007 00:31:26 +0000 (00:31 +0000)] 
- moved class-level attributes placed by the attributes package into a _class_state
variable attached to the class.
- mappers track themselves primarily using the "mappers" collection on _class_state.
ClassKey is gone and mapper lookup uses regular dict keyed to entity_name; removes
a fair degree of WeakKeyDictionary overhead as well as ClassKey overhead.
- mapper_registry renamed to _mapper_registry; is only consulted by the
compile_mappers(), mapper.compile() and clear_mappers() functions/methods.

17 years ago- several ORM attributes have been removed or made private:
Mike Bayer [Sat, 1 Dec 2007 23:00:05 +0000 (23:00 +0000)] 
- several ORM attributes have been removed or made private:
mapper.get_attr_by_column(), mapper.set_attr_by_column(),
mapper.pks_by_table, mapper.cascade_callable(),
MapperProperty.cascade_callable(), mapper.canload()
- refinements to mapper PK/table column organization, session cascading,
some naming convention work

17 years agoadded test to ensure two conflicting m2m + backrefs raise an error
Mike Bayer [Fri, 30 Nov 2007 21:25:42 +0000 (21:25 +0000)] 
added test to ensure two conflicting m2m + backrefs raise an error

17 years agoremove a little cruft
Mike Bayer [Fri, 30 Nov 2007 03:34:06 +0000 (03:34 +0000)] 
remove a little cruft

17 years ago- added support for version_id_col in conjunction with inheriting mappers.
Mike Bayer [Fri, 30 Nov 2007 00:40:56 +0000 (00:40 +0000)] 
- added support for version_id_col in conjunction with inheriting mappers.
version_id_col is typically set on the base mapper in an inheritance
relationship where it takes effect for all inheriting mappers.
[ticket:883]
- a little rearrangement of save_obj()

17 years ago- adjustment to the previous checkin regarding inheritance to not conflict with globals
Mike Bayer [Thu, 29 Nov 2007 19:37:05 +0000 (19:37 +0000)] 
- adjustment to the previous checkin regarding inheritance to not conflict with globals
- fix to self-referential eager loading such that if the same mapped
instance appears in two or more distinct sets of columns in the same
result set, its eagerly loaded collection will be populated regardless
of whether or not all of the rows contain a set of "eager" columns for
that collection.  this would also show up as a KeyError when fetching
results with join_depth turned on.

17 years ago- fixed bug where Query would not apply a subquery to the SQL when LIMIT
Mike Bayer [Thu, 29 Nov 2007 15:36:13 +0000 (15:36 +0000)] 
- fixed bug where Query would not apply a subquery to the SQL when LIMIT
was used in conjunction with an inheriting mapper where the eager
loader was only in the parent mapper.

17 years agocleanup
Mike Bayer [Wed, 28 Nov 2007 23:44:14 +0000 (23:44 +0000)] 
cleanup

17 years agofixed unicode-ness for Unicode values
Mike Bayer [Wed, 28 Nov 2007 22:09:57 +0000 (22:09 +0000)] 
fixed unicode-ness for Unicode values

17 years ago- fixed bug which could arise when using session.begin_nested() in conjunction
Mike Bayer [Wed, 28 Nov 2007 22:03:14 +0000 (22:03 +0000)] 
- fixed bug which could arise when using session.begin_nested() in conjunction
with more than one level deep of enclosing session.begin() statements

17 years agonew synonym() behavior, including auto-attribute gen, attribute decoration,
Mike Bayer [Wed, 28 Nov 2007 21:13:35 +0000 (21:13 +0000)] 
new synonym() behavior, including auto-attribute gen, attribute decoration,
and auto-column mapping implemented; [ticket:801]

17 years agodefault value of assert_unicode is None on String, False on create_engine(), and...
Mike Bayer [Wed, 28 Nov 2007 00:44:16 +0000 (00:44 +0000)] 
default value of assert_unicode is None on String, False on create_engine(), and True on Unicode type.

17 years ago- column labels in the form "tablename.columname", i.e. with a dot, are now
Mike Bayer [Tue, 27 Nov 2007 16:57:56 +0000 (16:57 +0000)] 
- column labels in the form "tablename.columname", i.e. with a dot, are now
supported.

17 years agoopened up the test for "reflection with convert_unicode=True". this is since convert...
Mike Bayer [Tue, 27 Nov 2007 16:23:02 +0000 (16:23 +0000)] 
opened up the test for "reflection with convert_unicode=True".  this is since convert_unicode by default has assert_unicode, want to ensure that other dialects (at least oracle) support this (i.e. not unicode schema names themselves, just that they dont sent thru bytestrings to a String).

if maxdb or sybase *should* be able to handle this too though I can't test on this end.

17 years agoadded None to support zope __provides__, [ticket:882]
Mike Bayer [Tue, 27 Nov 2007 16:15:32 +0000 (16:15 +0000)] 
added None to support zope __provides__, [ticket:882]

17 years agoadded a test to validate ResultProxy truncation behavior
Mike Bayer [Tue, 27 Nov 2007 16:13:45 +0000 (16:13 +0000)] 
added a test to validate ResultProxy truncation behavior

17 years agoun-screw up the attribute manager checkin
Mike Bayer [Tue, 27 Nov 2007 05:44:16 +0000 (05:44 +0000)] 
un-screw up the attribute manager checkin

17 years agofixed reflection of unicode, [ticket:881]
Mike Bayer [Tue, 27 Nov 2007 05:37:57 +0000 (05:37 +0000)] 
fixed reflection of unicode, [ticket:881]

17 years agoAttributeManager class and "cached" state removed....attribute listing
Mike Bayer [Tue, 27 Nov 2007 05:15:13 +0000 (05:15 +0000)] 
AttributeManager class and "cached" state removed....attribute listing
is tracked from _sa_attrs class collection

17 years agoFix: MSSQL set identity_insert and errors [ticket:538]
Paul Johnston [Sun, 25 Nov 2007 23:56:38 +0000 (23:56 +0000)] 
Fix: MSSQL set identity_insert and errors [ticket:538]

17 years ago- check for NoneType too with unicode....
Mike Bayer [Sun, 25 Nov 2007 23:46:05 +0000 (23:46 +0000)] 
- check for NoneType too with unicode....
- fixed ORM tests to have proper unicode

17 years agoMSSQL doesn't support subqueries in insert values; disable test
Paul Johnston [Sun, 25 Nov 2007 23:36:20 +0000 (23:36 +0000)] 
MSSQL doesn't support subqueries in insert values; disable test

17 years agoFix: test_decimal on MSSQL - use a value that is accurately represented as a float...
Paul Johnston [Sun, 25 Nov 2007 23:34:39 +0000 (23:34 +0000)] 
Fix: test_decimal on MSSQL - use a value that is accurately represented as a float, and make when asdecimal=False, convert Decimal to float

17 years agoedits
Mike Bayer [Sun, 25 Nov 2007 23:34:26 +0000 (23:34 +0000)] 
edits

17 years agoFix: MSSQL concatenate operator is + not || [ticket:879]
Paul Johnston [Sun, 25 Nov 2007 23:17:02 +0000 (23:17 +0000)] 
Fix: MSSQL concatenate operator is + not || [ticket:879]

17 years ago- added new flag to String and create_engine(), assert_unicode=(True|False|None).
Mike Bayer [Sun, 25 Nov 2007 23:14:03 +0000 (23:14 +0000)] 
- added new flag to String and create_engine(), assert_unicode=(True|False|None).
When convert_unicode=True, this flag also defaults to `True`, and results in all
unicode conversion operations raising an exception when a non-unicode bytestring
is passed as a bind parameter.  It is strongly advised that all unicode-aware
applications make proper use of Python unicode objects (i.e. u'hello' and
not 'hello').

17 years agoAvoid doubling quoting of identifier in MSSQL reflection
Paul Johnston [Sun, 25 Nov 2007 23:10:17 +0000 (23:10 +0000)] 
Avoid doubling quoting of identifier in MSSQL reflection

17 years agoMake function a reserved word in MSSQL
Paul Johnston [Sun, 25 Nov 2007 22:58:14 +0000 (22:58 +0000)] 
Make function a reserved word in MSSQL

17 years agoMSSQL/PyODBC no longer has a global set nocount on
Paul Johnston [Sun, 25 Nov 2007 22:46:22 +0000 (22:46 +0000)] 
MSSQL/PyODBC no longer has a global set nocount on

17 years agoChange to make PyODBC result fetching a bit more reliable
Paul Johnston [Sun, 25 Nov 2007 17:40:49 +0000 (17:40 +0000)] 
Change to make PyODBC result fetching a bit more reliable

17 years agocolumn.foreign_key -> foreign_keys in MSSQL
Paul Johnston [Sun, 25 Nov 2007 17:32:06 +0000 (17:32 +0000)] 
column.foreign_key -> foreign_keys in MSSQL

17 years agobump for py2.4
Mike Bayer [Sun, 25 Nov 2007 04:00:44 +0000 (04:00 +0000)] 
bump for py2.4

17 years agoOrderedSet to appease the unit tests....need to find a way to get rid of this
Mike Bayer [Sun, 25 Nov 2007 03:45:20 +0000 (03:45 +0000)] 
OrderedSet to appease the unit tests....need to find a way to get rid of this

17 years ago- named_with_column becomes an attribute
Mike Bayer [Sun, 25 Nov 2007 03:28:49 +0000 (03:28 +0000)] 
- named_with_column becomes an attribute
- cleanup within compiler visit_select(), column labeling
- is_select() removed from dialects, replaced with returns_rows_text(), returns_rows_compiled()
- should_autocommit() removed from dialects, replaced with should_autocommit_text() and
should_autocommit_compiled()
- typemap and column_labels collections removed from Compiler, replaced with single "result_map" collection.
- ResultProxy uses more succinct logic in combination with result_map to target columns

17 years ago- decruftify old visitors used by orm, convert to functions that
Mike Bayer [Sat, 24 Nov 2007 01:59:29 +0000 (01:59 +0000)] 
- decruftify old visitors used by orm, convert to functions that
use a common traversal function.
- TranslatingDict is finally gone, thanks to column.proxy_set simpleness...hooray !
- shoved "slice" use case on RowProxy into an exception case.  knocks noticeable time off of large result set operations.

17 years ago- all kinds of cleanup, tiny-to-slightly-significant speed improvements
Mike Bayer [Sat, 24 Nov 2007 00:55:39 +0000 (00:55 +0000)] 
- all kinds of cleanup, tiny-to-slightly-significant speed improvements

17 years agomore changes to merge(dont_load); since we now have a guarantee that
Mike Bayer [Fri, 23 Nov 2007 19:01:40 +0000 (19:01 +0000)] 
more changes to merge(dont_load); since we now have a guarantee that
all merged instances aren't dirty, we can copy the attribues without using
any append/replace events, and therefore don't have any concern of lazy loaders
firing off.  added tests to ensure that '_is_orphan()' doesn't get screwed up,
and also that the 'dirty' list on the new session stays empty, which is an
extra bonus of this approach.

17 years ago- some clarifications and fixes to merge(instance, dont_load=True).
Mike Bayer [Fri, 23 Nov 2007 05:24:32 +0000 (05:24 +0000)] 
- some clarifications and fixes to merge(instance, dont_load=True).
fixed bug where lazy loaders were getting disabled on returned instances.
Also, we currently do not support merging an instance which has uncommitted
changes on it, in the case that dont_load=True is used....this will
now raise an error.  This is due to complexities in merging the
"committed state" of the given instance to correctly correspond to the
newly copied instance.  Since the use case for dont_load=True is
caching, the given instances shouldn't have any uncommitted changes on them
anyway.

17 years agoTypo fix
Jason Kirtland [Tue, 20 Nov 2007 18:32:01 +0000 (18:32 +0000)] 
Typo fix

17 years ago- clarified the error message which occurs when you try to update()
Mike Bayer [Tue, 20 Nov 2007 15:55:36 +0000 (15:55 +0000)] 
- clarified the error message which occurs when you try to update()
an instance with the same identity key as an instance already present
in the session.
- opened up the recursive checks in session.merge() a little bit

17 years ago- fixed endless loop issue when using lazy="dynamic" on both
Mike Bayer [Mon, 19 Nov 2007 08:30:52 +0000 (08:30 +0000)] 
- fixed endless loop issue when using lazy="dynamic" on both
sides of a bi-directional relationship [ticket:872]

17 years agoadd a polymorphic get() test
Mike Bayer [Mon, 19 Nov 2007 04:08:48 +0000 (04:08 +0000)] 
add a polymorphic get() test

17 years agodoctest fixups rel_0_4_1
Mike Bayer [Sun, 18 Nov 2007 22:47:55 +0000 (22:47 +0000)] 
doctest fixups

17 years ago- added tests for [ticket:768]
Mike Bayer [Sun, 18 Nov 2007 22:35:19 +0000 (22:35 +0000)] 
- added tests for [ticket:768]

17 years ago- added op() operator to instrumented attributes; i.e.
Mike Bayer [Sun, 18 Nov 2007 22:19:05 +0000 (22:19 +0000)] 
- added op() operator to instrumented attributes; i.e.
User.name.op('ilike')('%somename%') [ticket:767]

17 years ago- MSSQL anonymous labels for selection of functions made deterministic
Mike Bayer [Sun, 18 Nov 2007 21:32:13 +0000 (21:32 +0000)] 
- MSSQL anonymous labels for selection of functions made deterministic
- propagate correct **kwargs through mssql methods

17 years agorepaired FB functions, [ticket:862]
Mike Bayer [Sun, 18 Nov 2007 21:20:32 +0000 (21:20 +0000)] 
repaired FB functions, [ticket:862]

17 years agoTests for mysql casts and a couple adjustments.
Jason Kirtland [Sun, 18 Nov 2007 20:44:20 +0000 (20:44 +0000)] 
Tests for mysql casts and a couple adjustments.

17 years ago- Clarified use of python's Decimal
Jason Kirtland [Sun, 18 Nov 2007 19:01:42 +0000 (19:01 +0000)] 
- Clarified use of python's Decimal

17 years agoMigrated Connection.properties to Connection.info ('info' is the new standard name...
Jason Kirtland [Sun, 18 Nov 2007 18:19:52 +0000 (18:19 +0000)] 
Migrated Connection.properties to Connection.info ('info' is the new standard name for user-writable property collections that came out of [ticket:573]).  'properties' is now an alias, will be removed in 0.5.

17 years ago- added a little more checking for garbage-collection dereferences in
Mike Bayer [Sun, 18 Nov 2007 17:31:09 +0000 (17:31 +0000)] 
- added a little more checking for garbage-collection dereferences in
InstanceState.__cleanup() to reduce "gc ignored" errors on app
shutdown

17 years agoadded some ORDER BYs to appease the ever picky postgres
Mike Bayer [Sun, 18 Nov 2007 17:17:04 +0000 (17:17 +0000)] 
added some ORDER BYs to appease the ever picky postgres

17 years ago- PickleType will compare using `==` when set up with mutable=False,
Mike Bayer [Sun, 18 Nov 2007 16:32:47 +0000 (16:32 +0000)] 
- PickleType will compare using `==` when set up with mutable=False,
and not the `is` operator.  To use `is` or any other comparator, send
in a custom comparison function using PickleType(comparator=my_custom_comparator).

17 years agotest cases were not fully testing contains_eager() with regards to [ticket:777],...
Mike Bayer [Sun, 18 Nov 2007 16:18:54 +0000 (16:18 +0000)] 
test cases were not fully testing contains_eager() with regards to [ticket:777], fixed contains_eager() for more than one level deep

17 years agosome wide zoomark ranges...
Mike Bayer [Sun, 18 Nov 2007 07:23:00 +0000 (07:23 +0000)] 
some wide zoomark ranges...

17 years agocut down a good deal of Join construction overhead
Mike Bayer [Sun, 18 Nov 2007 06:51:17 +0000 (06:51 +0000)] 
cut down a good deal of Join construction overhead

17 years agologging fixes
Mike Bayer [Sun, 18 Nov 2007 05:46:45 +0000 (05:46 +0000)] 
logging fixes

17 years agofixed both group-deferred attributes and expired attributes to not
Mike Bayer [Sun, 18 Nov 2007 05:16:52 +0000 (05:16 +0000)] 
fixed both group-deferred attributes and expired attributes to not
blow away changes made on attributes before the load takes place

17 years ago- oracle will now reflect "DATE" as an OracleDateTime column, not
Mike Bayer [Sun, 18 Nov 2007 02:44:15 +0000 (02:44 +0000)] 
- oracle will now reflect "DATE" as an OracleDateTime column, not
OracleDate

- added awareness of schema name in oracle table_names() function,
fixes metadata.reflect(schema='someschema') [ticket:847]

17 years agoremoved is_expired() method since the meaning of "expired" is per-attribute now
Mike Bayer [Sun, 18 Nov 2007 02:22:41 +0000 (02:22 +0000)] 
removed is_expired() method since the meaning of "expired" is per-attribute now