]> git.ipfire.org Git - thirdparty/sqlalchemy/sqlalchemy.git/log
thirdparty/sqlalchemy/sqlalchemy.git
17 years agomore searching for equiv columns
Mike Bayer [Fri, 9 Nov 2007 16:54:47 +0000 (16:54 +0000)] 
more searching for equiv columns

17 years ago- fixed error where Query.add_column() would not accept a class-bound
Mike Bayer [Fri, 9 Nov 2007 16:36:46 +0000 (16:36 +0000)] 
- fixed error where Query.add_column() would not accept a class-bound
attribute as an argument; Query also raises an error if an invalid
argument was sent to add_column() (at instances() time) [ticket:858]

17 years ago- query doesn't throw an error if you use distinct() and an order_by()
Mike Bayer [Fri, 9 Nov 2007 16:25:28 +0000 (16:25 +0000)] 
- query doesn't throw an error if you use distinct() and an order_by()
containing UnaryExpressions (or other) together [ticket:848]

17 years agoa tweak, still requiring test coverage, to add more "foreign key equivalents" to...
Mike Bayer [Fri, 9 Nov 2007 15:11:38 +0000 (15:11 +0000)] 
a tweak, still requiring test coverage, to add more "foreign key equivalents" to the equivalent_columns collection

17 years agomore changes to traverse-and-clone; a particular element will only be cloned once...
Mike Bayer [Thu, 8 Nov 2007 18:06:21 +0000 (18:06 +0000)] 
more changes to traverse-and-clone; a particular element will only be cloned once and is
then re-used.  the FROM calculation of a Select normalizes the list of hide_froms against all
previous incarnations of each FROM clause, using a tag attached from cloned clause to
previous.

17 years ago- identified some cases where Alias needs to be cloned; but still cant clone
Mike Bayer [Thu, 8 Nov 2007 00:26:23 +0000 (00:26 +0000)] 
- identified some cases where Alias needs to be cloned; but still cant clone
when its an alias of a Table; added some test coverage for one particular
case from the doctests
- fixed "having" example in doctests, updated eager load example

17 years agoupdated zoomark test_1a_populate callcounts. if the calls keep being removed like...
Ants Aasma [Thu, 8 Nov 2007 00:11:46 +0000 (00:11 +0000)] 
updated zoomark test_1a_populate callcounts. if the calls keep being removed like this then soon there will be none left.

17 years ago- fixed remainder of [ticket:853]
Mike Bayer [Wed, 7 Nov 2007 22:25:01 +0000 (22:25 +0000)] 
- fixed remainder of [ticket:853]
- bindparam 'shortname' is deprecated
- fixed testing.assert_compile() to actually generate bind param dict before asserting
- added bind param assertions to CRUDTest.test_update

17 years ago - <Engine|Connection>._execute_clauseelement becomes a public method
Mike Bayer [Wed, 7 Nov 2007 21:56:23 +0000 (21:56 +0000)] 
  - <Engine|Connection>._execute_clauseelement becomes a public method
    Connectable.execute_clauseelement

17 years ago - fix to compiled bind parameters to not mistakenly populate None
Mike Bayer [Wed, 7 Nov 2007 21:48:16 +0000 (21:48 +0000)] 
  - fix to compiled bind parameters to not mistakenly populate None
    [ticket:853]

17 years agofixed the previous TLTransaction checkin
Mike Bayer [Wed, 7 Nov 2007 21:46:48 +0000 (21:46 +0000)] 
fixed the previous TLTransaction checkin

17 years ago- fixed the close() method on Transaction when using strategy='threadlocal'
Mike Bayer [Wed, 7 Nov 2007 18:31:35 +0000 (18:31 +0000)] 
- fixed the close() method on Transaction when using strategy='threadlocal'

17 years agoadjusted "blank out primary key" rule to check for "allow_null_pks" on target mapper...
Mike Bayer [Wed, 7 Nov 2007 01:36:16 +0000 (01:36 +0000)] 
adjusted "blank out primary key" rule to check for "allow_null_pks" on target mapper.  this was revealed by
recent attributes.py change in r3695 that causes a value of "None" to register as part of the attribute history's
added_items() collection (i.e. since AttributeHistory compares against NO_VALUE instead of None).

17 years ago - fixed very hard-to-reproduce issue where by the FROM clause of Query
Mike Bayer [Wed, 7 Nov 2007 00:52:49 +0000 (00:52 +0000)] 
  - fixed very hard-to-reproduce issue where by the FROM clause of Query
    could get polluted by certain generative calls [ticket:852]

17 years agofixed/added coverage for list extension; [ticket:855]
Mike Bayer [Tue, 6 Nov 2007 19:19:30 +0000 (19:19 +0000)] 
fixed/added coverage for list extension; [ticket:855]

17 years agoAdded missing InternalError exception wrapper [ticket:854]
Jason Kirtland [Tue, 6 Nov 2007 18:21:28 +0000 (18:21 +0000)] 
Added missing InternalError exception wrapper [ticket:854]

17 years agoCreate a storage field for arbitrary info on tables/columns; ticket #573
Paul Johnston [Tue, 6 Nov 2007 10:41:40 +0000 (10:41 +0000)] 
Create a storage field for arbitrary info on tables/columns; ticket #573

17 years agoAdded note about EXT_CONTINUE to 3.x -> 4.x migration guide
Rick Morrison [Mon, 5 Nov 2007 22:52:21 +0000 (22:52 +0000)] 
Added note about EXT_CONTINUE to 3.x -> 4.x migration guide

17 years agoMigrated maxdb behavioral assumptions from unsupported to fails_on
Jason Kirtland [Mon, 5 Nov 2007 22:11:30 +0000 (22:11 +0000)] 
Migrated maxdb behavioral assumptions from unsupported to fails_on

17 years agoAdded testing.fails_on('db') failure-asserter.
Jason Kirtland [Mon, 5 Nov 2007 22:02:00 +0000 (22:02 +0000)] 
Added testing.fails_on('db') failure-asserter.

17 years ago- figured out a way to get previous oracle behavior back. the ROWID thing
Mike Bayer [Mon, 5 Nov 2007 19:38:01 +0000 (19:38 +0000)] 
- figured out a way to get previous oracle behavior back.  the ROWID thing
is still a pretty thorny issue.

17 years ago- adjustments to oracle ROWID logic...recent oid changes mean we have to
Mike Bayer [Mon, 5 Nov 2007 19:23:08 +0000 (19:23 +0000)] 
- adjustments to oracle ROWID logic...recent oid changes mean we have to
use "rowid" against the select itself (i.e. its just...'rowid', no table name).
seems to work OK but not sure if issues will arise
- fixes to oracle bind param stuff to account for recent removal of ClauseParameters object.

17 years ago- oid_column proxies more intelligently off of Select, CompoundSelect - fixes platfor...
Mike Bayer [Mon, 5 Nov 2007 18:30:30 +0000 (18:30 +0000)] 
- oid_column proxies more intelligently off of Select, CompoundSelect - fixes platform-affected bugs in missing the correct "oid" column
- locate_all_froms() is expensive; added an attribute-level cache for it
- put a huge warning on all select.append_XXX() methods stating that derived collections like locate_all_froms() may become invalid if
already initialized

17 years agoWhitespace cleanup
Jason Kirtland [Mon, 5 Nov 2007 17:26:17 +0000 (17:26 +0000)] 
Whitespace cleanup

17 years ago- base_columns on ColumnElement becomes a list; as usual, because columns in Compound...
Mike Bayer [Mon, 5 Nov 2007 17:18:21 +0000 (17:18 +0000)] 
- base_columns on ColumnElement becomes a list; as usual, because columns in CompoundSelects
may extend from more than one root column.
- keys_ok argument from corresponding_column() removed.  no more name-based matching of columns anywhere.
- DictDecorator is gone.  all row translators provided by orm.util.create_row_adapter().  Mapper
and contains_alias() cache the adapters on target mapper to avoid re-computation of adapters.
- create_row_adapter() accepts an "equivalent_columns" map as produced by Mapper, so that
row adapters can take join conditions into account (as usual again, to help with the CompoundSelects
produced by polymorphic_union).
- simplified TableSingleton to just provide lookup; moved all initialization into Table.
- the "properties" accessor on Mapper is removed; it now throws an informative
exception explaining the usage of mapper.get_property() and
mapper.iterate_properties

17 years ago- SHOW CREATE TABLE output is slightly different if msyql is in ANSI mode
Jason Kirtland [Mon, 5 Nov 2007 17:15:41 +0000 (17:15 +0000)] 
- SHOW CREATE TABLE output is slightly different if msyql is in ANSI mode

17 years ago- rewrote and simplified the system used to "target" columns across
Mike Bayer [Mon, 5 Nov 2007 00:59:19 +0000 (00:59 +0000)] 
- rewrote and simplified the system used to "target" columns across
selectable expressions.  On the SQL side this is represented by the
"corresponding_column()" method. This method is used heavily by the ORM
to "adapt" elements of an expression to similar, aliased expressions,
as well as to target result set columns originally bound to a
table or selectable to an aliased, "corresponding" expression.  The new
rewrite features completely consistent and accurate behavior.
- the "orig_set" and "distance" elements as well as all associated
fanfare are gone (hooray !)
- columns now have an optional "proxies" list which is a list of all
columns they are a "proxy" for; only CompoundSelect cols proxy more than one column
(just like before).  set operations are used to determine lineage.
- CompoundSelects (i.e. unions) only create one public-facing proxy column per
column name.  primary key collections come out with just one column per embedded
PK column.
- made the alias used by eager load limited subquery anonymous.

17 years ago - func. objects can be pickled/unpickled [ticket:844]
Mike Bayer [Sun, 4 Nov 2007 22:04:22 +0000 (22:04 +0000)] 
  - func. objects can be pickled/unpickled [ticket:844]

17 years ago- eager loading with LIMIT/OFFSET applied no longer adds the primary
Mike Bayer [Sat, 3 Nov 2007 23:17:34 +0000 (23:17 +0000)] 
- eager loading with LIMIT/OFFSET applied no longer adds the primary
table joined to a limited subquery of itself; the eager loads now
join directly to the subquery which also provides the primary table's
columns to the result set.  This eliminates a JOIN from all eager loads
with LIMIT/OFFSET.  [ticket:843]

17 years ago- rewritten ClauseAdapter merged from the eager_minus_join branch; this is a much...
Mike Bayer [Sat, 3 Nov 2007 22:13:17 +0000 (22:13 +0000)] 
- rewritten ClauseAdapter merged from the eager_minus_join branch; this is a much simpler
and "correct" version which will copy all elements exactly once, except for those which were
replaced with target elements.  It also can match a wider variety of target elements including
joins and selects on identity alone.

17 years agoAdded a profiled benchmark for orm attribute modification & flush
Jason Kirtland [Sat, 3 Nov 2007 22:03:18 +0000 (22:03 +0000)] 
Added a profiled benchmark for orm attribute modification & flush

17 years ago- Swap operator.eq for lambda x,y: x==y
Jason Kirtland [Sat, 3 Nov 2007 21:59:39 +0000 (21:59 +0000)] 
- Swap operator.eq for lambda x,y: x==y
- Some docstring re-formatting

17 years agoFormatting for 0.4.1
Jason Kirtland [Sat, 3 Nov 2007 20:53:37 +0000 (20:53 +0000)] 
Formatting for 0.4.1

17 years agoAdded some more notes for maxdb
Jason Kirtland [Sat, 3 Nov 2007 20:42:07 +0000 (20:42 +0000)] 
Added some more notes for maxdb

17 years ago- Removed equality, truth and hash() testing of mapped instances. Mapped
Jason Kirtland [Sat, 3 Nov 2007 20:23:26 +0000 (20:23 +0000)] 
- Removed equality, truth and hash() testing of mapped instances. Mapped
  classes can now implement arbitrary __eq__ and friends. [ticket:676]

17 years ago- removed "name" attribute from FromClause, Join, Select, CompoundSelect. its needless
Mike Bayer [Sat, 3 Nov 2007 20:12:59 +0000 (20:12 +0000)] 
- removed "name" attribute from FromClause, Join, Select, CompoundSelect.  its needless
and led to some very strange anonymous label names
- removed what was apparently cruft in some column-targeting code

17 years agoadded small clarification on single-table inheritance mapper def
Rick Morrison [Sat, 3 Nov 2007 16:11:41 +0000 (16:11 +0000)] 
added small clarification on single-table inheritance mapper def

17 years agodoc updates for save_on_init=False, merge(...dont_save=True)
Mike Bayer [Fri, 2 Nov 2007 21:02:22 +0000 (21:02 +0000)] 
doc updates for save_on_init=False, merge(...dont_save=True)

17 years ago- extra merge test
Mike Bayer [Fri, 2 Nov 2007 18:05:06 +0000 (18:05 +0000)] 
- extra merge test
- merge with dont_load also propagates _state.modified

17 years ago- merge() includes a keyword argument "dont_load=True". setting this flag will cause
Mike Bayer [Fri, 2 Nov 2007 17:58:20 +0000 (17:58 +0000)] 
- merge() includes a keyword argument "dont_load=True".  setting this flag will cause
the merge operation to not load any data from the database in response to incoming
detached objects, and will accept the incoming detached object as though it were
already present in that session.  Use this to merge detached objects from external
caching systems into the session.

17 years agoMore improvements to testlib's mapper decorator
Jason Kirtland [Fri, 2 Nov 2007 00:27:40 +0000 (00:27 +0000)] 
More improvements to testlib's mapper decorator

17 years agoTweaks for assert_unordered_result
Jason Kirtland [Thu, 1 Nov 2007 23:13:30 +0000 (23:13 +0000)] 
Tweaks for assert_unordered_result

17 years agoAdded AssertMixin.assert_unordered_result
Jason Kirtland [Thu, 1 Nov 2007 20:59:30 +0000 (20:59 +0000)] 
Added AssertMixin.assert_unordered_result

17 years ago- session checks more carefully when determining "object X already in another session";
Mike Bayer [Thu, 1 Nov 2007 20:12:36 +0000 (20:12 +0000)] 
- session checks more carefully when determining "object X already in another session";
e.g. if you pickle a series of objects and unpickle (i.e. as in a Pylons HTTP session
or similar), they can go into a new session without any conflict
- added stricter checks around session.delete() similar to update()
- shored up some old "validate" stuff in session/uow

17 years agoAdded rowset() testing helper.
Jason Kirtland [Thu, 1 Nov 2007 19:40:08 +0000 (19:40 +0000)] 
Added rowset() testing helper.

17 years agoFixed truth-evaluation of mapped objects (part of [ticket:676]
Jason Kirtland [Thu, 1 Nov 2007 16:36:50 +0000 (16:36 +0000)] 
Fixed truth-evaluation of mapped objects (part of [ticket:676]

17 years ago- merged factor_down_bindparams branch.
Mike Bayer [Thu, 1 Nov 2007 03:58:21 +0000 (03:58 +0000)] 
- merged factor_down_bindparams branch.
- removed ClauseParameters object; compiled.params returns a regular dictionary
  now, as well as result.last_inserted_params()/last_updated_params().
- various code trimming, method removals.

17 years agoAdded command line options to add tripwires for __hash__, __eq__ and __nonzero__...
Jason Kirtland [Wed, 31 Oct 2007 23:49:09 +0000 (23:49 +0000)] 
Added command line options to add tripwires for __hash__, __eq__ and __nonzero__ on mapped classes.

17 years ago- A more efficient IdentitySet
Jason Kirtland [Wed, 31 Oct 2007 19:53:27 +0000 (19:53 +0000)] 
- A more efficient IdentitySet

17 years ago- split ScalarInstrumentedAttribute into a "scalar" and an "object" version.
Mike Bayer [Wed, 31 Oct 2007 19:11:22 +0000 (19:11 +0000)] 
- split ScalarInstrumentedAttribute into a "scalar" and an "object" version.
The "object" version loads the existing value on set/del, fires events,
 and handles trackparent operations; the "scalar" version does not.
- column loaders now use the "scalar" version of InstrumentedAttribute, so that
event handlers etc. don't fire off for regular column attribute operations.
- some adjustments to AttributeHistory to work properly for non-loaded attributes
- deferred column attributes no longer trigger a load operation when the
attribute is assigned to.  in those cases, the newly assigned
value will be present in the flushes' UPDATE statement unconditionally.

17 years agodont call up new session unless save_on_init
Mike Bayer [Wed, 31 Oct 2007 17:15:30 +0000 (17:15 +0000)] 
dont call up new session unless save_on_init

17 years ago- also added proxying of save_or_update to scoped sessions.
Ants Aasma [Wed, 31 Oct 2007 16:30:06 +0000 (16:30 +0000)] 
- also added proxying of save_or_update to scoped sessions.
- session.update() raises an error when updating an instance that is already in the session with a different identity
- adjusted zoomarks lower limits so I can get a nice clean run

17 years agoadded REAL to reflection list
Mike Bayer [Wed, 31 Oct 2007 15:52:33 +0000 (15:52 +0000)] 
added REAL to reflection list

17 years agoAdded proxying of __contains__ and __iter__ methods for scoped sessions
Ants Aasma [Wed, 31 Oct 2007 12:14:22 +0000 (12:14 +0000)] 
Added proxying of __contains__ and __iter__ methods for scoped sessions

17 years agoFixup sp_columns call
Paul Johnston [Wed, 31 Oct 2007 10:48:30 +0000 (10:48 +0000)] 
Fixup sp_columns call

17 years agoFixed a truncation error when re-assigning a subset of a collection
Jason Kirtland [Wed, 31 Oct 2007 09:21:22 +0000 (09:21 +0000)] 
Fixed a truncation error when re-assigning a subset of a collection
(obj.relation = obj.relation[1:]) [ticket:834]

17 years agoAdded util.IdentitySet to support [ticket:676] and [ticket:834]
Jason Kirtland [Wed, 31 Oct 2007 09:13:12 +0000 (09:13 +0000)] 
Added util.IdentitySet to support [ticket:676] and [ticket:834]

17 years agoFormatting tweaks.
Jason Kirtland [Wed, 31 Oct 2007 05:44:17 +0000 (05:44 +0000)] 
Formatting tweaks.

17 years ago- Removed unused util.hash()
Jason Kirtland [Wed, 31 Oct 2007 03:25:04 +0000 (03:25 +0000)] 
- Removed unused util.hash()
- Fixed __hash__ for association proxy collections

17 years ago- Don't re-save objects in these tests (post r3681)
Jason Kirtland [Wed, 31 Oct 2007 03:17:51 +0000 (03:17 +0000)] 
- Don't re-save objects in these tests (post r3681)

17 years ago- Refinements for maxdb's handling of SERIAL and FIXED columns
Jason Kirtland [Wed, 31 Oct 2007 00:28:53 +0000 (00:28 +0000)] 
- Refinements for maxdb's handling of SERIAL and FIXED columns
- Expanded maxdb's set of paren-less functions

17 years ago- Added the NUMERIC sql type alias
Jason Kirtland [Wed, 31 Oct 2007 00:21:36 +0000 (00:21 +0000)] 
- Added the NUMERIC sql type alias

17 years ago- fix to "row switch" behavior, i.e. when an INSERT/DELETE is combined into a
Mike Bayer [Tue, 30 Oct 2007 18:04:00 +0000 (18:04 +0000)] 
- fix to "row switch" behavior, i.e. when an INSERT/DELETE is combined into a
  single UPDATE; many-to-many relations on the parent object update properly.
  [ticket:841]
- it's an error to session.save() an object which is already persistent
  [ticket:840]
- changed a bunch of repr(obj) calls in session.py exceptions to use mapperutil.instance_str()

17 years ago- restored MapperExtension functionality for [ticket:829], added test coverage
Mike Bayer [Mon, 29 Oct 2007 16:12:29 +0000 (16:12 +0000)] 
- restored MapperExtension functionality for [ticket:829], added test coverage
- changed naming convention in mapper.py tests to test_<testname>

17 years ago- fixed INSERT statements w.r.t. primary key columns that have SQL-expression
Mike Bayer [Sun, 28 Oct 2007 21:28:53 +0000 (21:28 +0000)] 
- fixed INSERT statements w.r.t. primary key columns that have SQL-expression
  based default generators on them; SQL expression executes inline as normal
  but will not trigger a "postfetch" condition for the column, for those DB's
  who provide it via cursor.lastrowid

17 years ago- fixed expression translation of text() clauses; this repairs various
Mike Bayer [Sun, 28 Oct 2007 18:25:37 +0000 (18:25 +0000)] 
- fixed expression translation of text() clauses; this repairs various
  ORM scenarios where literal text is used for SQL expressions

17 years ago- a little cleanup to deferred inheritance loading
Mike Bayer [Sun, 28 Oct 2007 04:36:36 +0000 (04:36 +0000)] 
- a little cleanup to deferred inheritance loading

17 years ago- merged path_based_options branch
Mike Bayer [Sun, 28 Oct 2007 01:51:43 +0000 (01:51 +0000)] 
- merged path_based_options branch
- behavior of query.options() is now fully based on paths, i.e. an option
  such as eagerload_all('x.y.z.y.x') will apply eagerloading to only
  those paths, i.e. and not 'x.y.x'; eagerload('children.children') applies
  only to exactly two-levels deep, etc. [ticket:777]
- removes old compiler()/schemagenerator()/schemadropper() methods from mysql dialect

17 years ago- inlined a couple of context variables
Mike Bayer [Sat, 27 Oct 2007 18:45:20 +0000 (18:45 +0000)] 
- inlined a couple of context variables
- PG two phase was calling text() without the correct bind param format, previous compiler checkin revealed issue

17 years ago- removed regular expression step from most statement compilations.
Mike Bayer [Sat, 27 Oct 2007 17:41:30 +0000 (17:41 +0000)] 
- removed regular expression step from most statement compilations.
  also fixes [ticket:833]
- inlining on PG with_returning() call
- extra options added for profiling

17 years agoclear MapperStub ArgSingletons on clear_mappers() too...
Mike Bayer [Sat, 27 Oct 2007 01:19:25 +0000 (01:19 +0000)] 
clear MapperStub ArgSingletons on clear_mappers() too...

17 years ago- fixed eager calc endless loop, introduced by r3646? this issue should have been...
Mike Bayer [Sat, 27 Oct 2007 00:50:20 +0000 (00:50 +0000)] 
- fixed eager calc endless loop, introduced by r3646?  this issue should have been present before 3646 though.

17 years ago- fixed clear_mappers() behavior to better clean up after itself
Mike Bayer [Sat, 27 Oct 2007 00:24:43 +0000 (00:24 +0000)] 
- fixed clear_mappers() behavior to better clean up after itself

17 years agoMake access dao detecting more reliable #828
Paul Johnston [Fri, 26 Oct 2007 18:17:14 +0000 (18:17 +0000)] 
Make access dao detecting more reliable #828

17 years agoremove unused method
Mike Bayer [Thu, 25 Oct 2007 19:30:43 +0000 (19:30 +0000)] 
remove unused method

17 years ago- removed needless 'continue'
Mike Bayer [Thu, 25 Oct 2007 16:58:02 +0000 (16:58 +0000)] 
- removed needless 'continue'

17 years ago- refactored _compile_properties/_compile_property, removed redundant code.
Mike Bayer [Thu, 25 Oct 2007 16:57:04 +0000 (16:57 +0000)] 
- refactored _compile_properties/_compile_property, removed redundant code.
still a little squirrely but much less complex.
- improved behavior of add_property() etc., fixed [ticket:831] involving
  synonym/deferred

17 years ago- de-cruftified backref configuration code, backrefs which step on existing
Mike Bayer [Wed, 24 Oct 2007 23:25:47 +0000 (23:25 +0000)] 
- de-cruftified backref configuration code, backrefs which step on existing
  properties now raise an error [ticket:832]

17 years agoA --db alias for max.
Jason Kirtland [Wed, 24 Oct 2007 07:55:18 +0000 (07:55 +0000)] 
A --db alias for max.

17 years ago- Added initial version of MaxDB dialect.
Jason Kirtland [Tue, 23 Oct 2007 07:38:07 +0000 (07:38 +0000)] 
- Added initial version of MaxDB dialect.
- All optional test Sequences are now optional=True

17 years agoFixed assert_sql_count exception logic.
Jason Kirtland [Tue, 23 Oct 2007 07:03:07 +0000 (07:03 +0000)] 
Fixed assert_sql_count exception logic.

17 years agoUpdated for r3655.
Jason Kirtland [Tue, 23 Oct 2007 01:57:40 +0000 (01:57 +0000)] 
Updated for r3655.

17 years agoAdded support for dialects that have both sequences and autoincrementing PKs.
Jason Kirtland [Tue, 23 Oct 2007 01:47:21 +0000 (01:47 +0000)] 
Added support for dialects that have both sequences and autoincrementing PKs.

17 years ago- The post_exec() hook now gets invoked before autocommit fires.
Jason Kirtland [Tue, 23 Oct 2007 01:16:27 +0000 (01:16 +0000)] 
- The post_exec() hook now gets invoked before autocommit fires.

17 years agoChanged some columns from TEXT to VARCHAR for sapdb.
Jason Kirtland [Mon, 22 Oct 2007 23:01:02 +0000 (23:01 +0000)] 
Changed some columns from TEXT to VARCHAR for sapdb.

17 years ago- Now guarding against broken DB-APIs when wrapping their exceptions.
Jason Kirtland [Mon, 22 Oct 2007 19:24:02 +0000 (19:24 +0000)] 
- Now guarding against broken DB-APIs when wrapping their exceptions.
- Added an explicit test for exception wrapping.

17 years ago- sqlite will reflect "DECIMAL" as a numeric column
Mike Bayer [Mon, 22 Oct 2007 15:23:08 +0000 (15:23 +0000)] 
- sqlite will reflect "DECIMAL" as a numeric column

17 years agoexception message, [ticket:826]
Mike Bayer [Mon, 22 Oct 2007 15:20:15 +0000 (15:20 +0000)] 
exception message, [ticket:826]

17 years agobumped sqlite version for empty insert test to 3.4
Mike Bayer [Sun, 21 Oct 2007 16:45:58 +0000 (16:45 +0000)] 
bumped sqlite version for empty insert test to 3.4

17 years ago- the whole OperationContext/QueryContext/SelectionContext thing greatly scaled back;
Mike Bayer [Sun, 21 Oct 2007 16:36:34 +0000 (16:36 +0000)] 
- the whole OperationContext/QueryContext/SelectionContext thing greatly scaled back;
all MapperOptions process the Query and that's it, one very simpliied QueryContext object gets passed
around at query.compile() and query.instances() time
- slight optimization to MapperExtension allowing the mapper to check for the presence of an extended method, takes 3000 calls off of masseagerload.py test (only a slight increase in speed though)
- attempting to centralize the notion of a "path" along mappers/properties, need to define what that is better.  heading towards [ticket:777]...

17 years agoExcluding older sqlite versions from the new insert tests. Currently conservative...
Jason Kirtland [Sat, 20 Oct 2007 17:14:46 +0000 (17:14 +0000)] 
Excluding older sqlite versions from the new insert tests.  Currently conservative- 2.8 definitely fails, 3.3 works.  The 3.0 and 3.1 binaries from sqlite.org segfault for me so the version check should be revisited when possible.

17 years ago- Added contains operator (which generate a "LIKE %<other>%" clause).
Gaëtan de Menten [Fri, 19 Oct 2007 10:27:06 +0000 (10:27 +0000)] 
- Added contains operator (which generate a "LIKE %<other>%" clause).

- Added test coverage for endswith operator

17 years agoFixed empty (zero column) sqlite inserts, allowing inserts on
Jason Kirtland [Fri, 19 Oct 2007 07:02:48 +0000 (07:02 +0000)] 
Fixed empty (zero column) sqlite inserts, allowing inserts on
autoincrementing single column tables.

17 years ago- added test coverage for unknown type reflection, fixed
Mike Bayer [Thu, 18 Oct 2007 17:56:18 +0000 (17:56 +0000)] 
- added test coverage for unknown type reflection, fixed
  sqlite/mysql handling of type reflection for unknown types

17 years agotypos, PassiveDefault likes "text()" otherwise the argument is a literal
Mike Bayer [Wed, 17 Oct 2007 19:10:16 +0000 (19:10 +0000)] 
typos, PassiveDefault likes "text()" otherwise the argument is a literal

17 years agoBump!
Jason Kirtland [Wed, 17 Oct 2007 18:04:33 +0000 (18:04 +0000)] 
Bump!

17 years agoput a little note at the top directing users to the more *exciting* changes rel_0_4_0
Mike Bayer [Wed, 17 Oct 2007 17:53:12 +0000 (17:53 +0000)] 
put a little note at the top directing users to the more *exciting* changes

17 years agoUpdate for 0.4.0
Jason Kirtland [Wed, 17 Oct 2007 17:41:59 +0000 (17:41 +0000)] 
Update for 0.4.0

17 years agotypo
Mike Bayer [Wed, 17 Oct 2007 17:40:04 +0000 (17:40 +0000)] 
typo

17 years ago- removed __len__ from "dynamic" collection as it would require issuing
Mike Bayer [Wed, 17 Oct 2007 16:49:15 +0000 (16:49 +0000)] 
- removed __len__ from "dynamic" collection as it would require issuing
  a SQL "count()" operation, thus forcing all list evaluations to issue
  redundant SQL [ticket:818]