]> git.ipfire.org Git - thirdparty/sqlalchemy/sqlalchemy.git/log
thirdparty/sqlalchemy/sqlalchemy.git
12 years ago- just comment out this test, we may have to figure out if the whole
Mike Bayer [Tue, 22 Oct 2013 03:28:19 +0000 (23:28 -0400)] 
- just comment out this test, we may have to figure out if the whole
"serializer" can just be ditched, or if somehow these deterministic issues
can be fixed (everyone implements __reduce__ maybe?)

12 years ago- use a different approach here since oracle isn't doing it either, just round it
Mike Bayer [Tue, 22 Oct 2013 03:26:08 +0000 (23:26 -0400)] 
- use a different approach here since oracle isn't doing it either, just round it

12 years ago- oursql returns a pure FP here which isn't exact
Mike Bayer [Tue, 22 Oct 2013 03:21:11 +0000 (23:21 -0400)] 
- oursql returns a pure FP here which isn't exact

12 years agotry to get this to pass on slow environmnets
Mike Bayer [Tue, 22 Oct 2013 03:17:36 +0000 (23:17 -0400)] 
try to get this to pass on slow environmnets

12 years agotypos
Mike Bayer [Mon, 21 Oct 2013 22:33:59 +0000 (18:33 -0400)] 
typos

12 years ago- A :func:`.bindparam` construct with a "null" type (e.g. no type
Mike Bayer [Mon, 21 Oct 2013 20:49:46 +0000 (16:49 -0400)] 
- A :func:`.bindparam` construct with a "null" type (e.g. no type
specified) is now copied when used in a typed expression, and the
new copy is assigned the actual type of the compared column.  Previously,
this logic would occur on the given :func:`.bindparam` in place.
Additionally, a similar process now occurs for :func:`.bindparam` constructs
passed to :meth:`.ValuesBase.values` for a :class:`.Insert` or
:class:`.Update` construct. [ticket:2850]

12 years ago- Fixed bug where :func:`.type_coerce` would not interpret ORM
Mike Bayer [Mon, 21 Oct 2013 19:06:41 +0000 (15:06 -0400)] 
- Fixed bug where :func:`.type_coerce` would not interpret ORM
elements with a ``__clause_element__()`` method properly.
[ticket:2849]

12 years ago- The typing system now handles the task of rendering "literal bind" values,
Mike Bayer [Sun, 20 Oct 2013 20:59:56 +0000 (16:59 -0400)] 
- The typing system now handles the task of rendering "literal bind" values,
e.g. values that are normally bound parameters but due to context must
be rendered as strings, typically within DDL constructs such as
CHECK constraints and indexes (note that "literal bind" values
become used by DDL as of :ticket:`2742`).  A new method
:meth:`.TypeEngine.literal_processor` serves as the base, and
:meth:`.TypeDecorator.process_literal_param` is added to allow wrapping
of a native literal rendering method. [ticket:2838]
- enhance _get_colparams so that we can send flags like literal_binds into
INSERT statements
- add support in PG for inspecting standard_conforming_strings
- add a new series of roundtrip tests based on INSERT of literal plus SELECT
for basic literal rendering in dialect suite

12 years ago- add a type_coerce() step within Enum, Boolean to the CHECK constraint,
Mike Bayer [Sun, 20 Oct 2013 20:25:46 +0000 (16:25 -0400)] 
- add a type_coerce() step within Enum, Boolean to the CHECK constraint,
so that the custom type isn't exposed to an operation that is against the
"impl" type's constraint, [ticket:2842]
- this change showed up as some recursion overflow in pickling with labels,
add a __reduce__() there....pickling of expressions is less and less something
that's very viable...

12 years ago- The :meth:`.Table.tometadata` method now produces copies of
Mike Bayer [Sat, 19 Oct 2013 00:01:45 +0000 (20:01 -0400)] 
- The :meth:`.Table.tometadata` method now produces copies of
all :attr:`.SchemaItem.info` dictionaries from all :class:`.SchemaItem`
objects within the structure including columns, constraints,
foreign keys, etc.   As these dictionaries
are copies, they are independent of the original dictionary.
Previously, only the ``.info`` dictionary of :class:`.Column` was transferred
within this operation, and it was only linked in place, not copied.
[ticket:2716]

12 years ago- attempt to clarify what mutable extension does for HSTORE, [ticket:2803]
Mike Bayer [Fri, 18 Oct 2013 23:34:58 +0000 (19:34 -0400)] 
- attempt to clarify what mutable extension does for HSTORE, [ticket:2803]

12 years ago- add a "limitations of reflection" section, [ticket:2233]
Mike Bayer [Fri, 18 Oct 2013 23:23:40 +0000 (19:23 -0400)] 
- add a "limitations of reflection" section, [ticket:2233]

12 years agouse a set here also
Mike Bayer [Fri, 18 Oct 2013 23:03:21 +0000 (19:03 -0400)] 
use a set here also

12 years ago- Removed a 128-character truncation from the reflection of the
Mike Bayer [Fri, 18 Oct 2013 23:00:05 +0000 (19:00 -0400)] 
- Removed a 128-character truncation from the reflection of the
server default for a column; this code was original from
PG system views which truncated the string for readability.
[ticket:2844]

12 years ago- The change in :ticket:`2721`, which is that the ``deferrable`` keyword
Mike Bayer [Fri, 18 Oct 2013 18:44:01 +0000 (14:44 -0400)] 
- The change in :ticket:`2721`, which is that the ``deferrable`` keyword
of :class:`.ForeignKeyConstraint` is silently ignored on the MySQL
backend, will be reverted as of 0.9; this keyword will now render again, raising
errors on MySQL as it is not understood - the same behavior will also
apply to the ``initially`` keyword.  In 0.8, the keywords will remain
ignored but a warning is emitted.   Additionally, the ``match`` keyword
now raises a :class:`.CompileError` on 0.9 and emits a warning on 0.8;
this keyword is not only silently ignored by MySQL but also breaks
the ON UPDATE/ON DELETE options.

To use a :class:`.ForeignKeyConstraint`
that does not render or renders differently on MySQL, use a custom
compilation option.  An example of this usage has been added to the
documentation, see :ref:`mysql_foreign_keys`.
[ticket:2721] [ticket:2839]

12 years ago- Added support for rendering ``SMALLSERIAL`` when a :class:`.SmallInteger`
Mike Bayer [Tue, 15 Oct 2013 23:06:21 +0000 (19:06 -0400)] 
- Added support for rendering ``SMALLSERIAL`` when a :class:`.SmallInteger`
type is used on a primary key autoincrement column, based on server
version detection of Postgresql version 9.2 or greater.
[ticket:2840]

12 years agoThe MySQL :class:`.mysql.SET` type now features the same auto-quoting
Mike Bayer [Mon, 14 Oct 2013 20:12:54 +0000 (16:12 -0400)] 
The MySQL :class:`.mysql.SET` type now features the same auto-quoting
behavior as that of :class:`.mysql.ENUM`.  Quotes are not required when
setting up the value, but quotes that are present will be auto-detected
along with a warning.  This also helps with Alembic where
the SET type doesn't render with quotes. [ticket:2817]

12 years agoThe ``.unique`` flag on :class:`.Index` could be produced as ``None``
Mike Bayer [Mon, 14 Oct 2013 18:53:28 +0000 (14:53 -0400)] 
The ``.unique`` flag on :class:`.Index` could be produced as ``None``
if it was generated from a :class:`.Column` that didn't specify ``unique``
(where it defaults to ``None``).  The flag will now always be ``True`` or
``False``. [ticket:2825]

12 years agoworkaround for #2838 here. still need to figure out why an ENUM test is suddenly...
Mike Bayer [Mon, 14 Oct 2013 14:56:11 +0000 (10:56 -0400)] 
workaround for #2838 here.  still need to figure out why an ENUM test is suddenly hitting this.

12 years ago- fix non-deterministic ordering point
Mike Bayer [Mon, 14 Oct 2013 14:35:54 +0000 (10:35 -0400)] 
- fix non-deterministic ordering point

12 years ago- fix test warning here (oursql)
Mike Bayer [Mon, 14 Oct 2013 14:34:03 +0000 (10:34 -0400)] 
- fix test warning here (oursql)

12 years agoMerge pull request #35 from alex/less-dicts
mike bayer [Mon, 14 Oct 2013 14:27:47 +0000 (07:27 -0700)] 
Merge pull request #35 from alex/less-dicts

Perform fewer dict lookups - tiny microoptimization

12 years agoPerform fewer dict lookups - tiny microoptimization
Alex Gaynor [Mon, 14 Oct 2013 11:03:15 +0000 (13:03 +0200)] 
Perform fewer dict lookups - tiny microoptimization

12 years agothese are in alphabetical order
Mike Bayer [Sun, 13 Oct 2013 21:17:00 +0000 (17:17 -0400)] 
these are in alphabetical order

12 years agomigration guide
Mike Bayer [Sun, 13 Oct 2013 21:16:36 +0000 (17:16 -0400)] 
migration guide

12 years agoname
Mike Bayer [Sun, 13 Oct 2013 20:43:57 +0000 (16:43 -0400)] 
name

12 years ago- Added new option to :func:`.relationship` ``distinct_target_key``.
Mike Bayer [Sun, 13 Oct 2013 20:42:32 +0000 (16:42 -0400)] 
-  Added new option to :func:`.relationship` ``distinct_target_key``.
This enables the subquery eager loader strategy to apply a DISTINCT
to the innermost SELECT subquery, to assist in the case where
duplicate rows are generated by the innermost query which corresponds
to this relationship (there's not yet a general solution to the issue
of dupe rows within subquery eager loading, however, when joins outside
of the innermost subquery produce dupes).  When the flag
is set to ``True``, the DISTINCT is rendered unconditionally, and when
it is set to ``None``, DISTINCT is rendered if the innermost relationship
targets columns that do not comprise a full primary key.
The option defaults to False in 0.8 (e.g. off by default in all cases),
None in 0.9 (e.g. automatic by default).   Thanks to Alexander Koval
for help with this. [ticket:2836]

12 years agoMySQL-connector dialect now allows options in the create_engine
Mike Bayer [Sun, 13 Oct 2013 18:09:35 +0000 (14:09 -0400)] 
MySQL-connector dialect now allows options in the create_engine
query string to override those defaults set up in the connect,
including "buffered" and "raise_on_warnings".
[ticket:2515]

12 years agoParenthesis will be applied to a compound SQL expression as
Mike Bayer [Sun, 13 Oct 2013 00:21:18 +0000 (20:21 -0400)] 
Parenthesis will be applied to a compound SQL expression as
rendered in the column list of a CREATE INDEX statement.
[ticket:2742]

12 years ago- Fixed bug in default compiler plus those of postgresql, mysql, and
Mike Bayer [Sun, 13 Oct 2013 00:04:55 +0000 (20:04 -0400)] 
- Fixed bug in default compiler plus those of postgresql, mysql, and
mssql to ensure that any literal SQL expression values are
rendered directly as literals, instead of as bound parameters,
within a CREATE INDEX statement. [ticket:2742]
- don't need expression_as_ddl(); literal_binds and include_table
take care of this functionality.

12 years agouse accepts scalar loader here so we deal with _ProxyImpl correctly
Mike Bayer [Fri, 11 Oct 2013 20:49:36 +0000 (16:49 -0400)] 
use accepts scalar loader here so we deal with _ProxyImpl correctly

12 years ago- fix bug due to regression from #2793, make sure we only go to load
Mike Bayer [Fri, 11 Oct 2013 20:41:52 +0000 (16:41 -0400)] 
- fix bug due to regression from #2793, make sure we only go to load
scalar attributes here and not relationships, else we get an error if there's no
actual scalars to load

12 years agoThe ``default`` argument of :class:`.Column` now accepts a class
Mike Bayer [Fri, 11 Oct 2013 20:16:08 +0000 (16:16 -0400)] 
The ``default`` argument of :class:`.Column` now accepts a class
or object method as an argument, in addition to a standalone function;
will properly detect if the "context" argument is accepted or not.

12 years agoThe "name" attribute is set on :class:`.Index` before the "attach"
Mike Bayer [Fri, 11 Oct 2013 19:55:57 +0000 (15:55 -0400)] 
The "name" attribute is set on :class:`.Index` before the "attach"
events are called, so that attachment events can be used to dynamically
generate a name for the index based on the parent table and/or
columns. [ticket:2835]

12 years ago- The method signature of :meth:`.Dialect.reflecttable`, which in
Mike Bayer [Fri, 11 Oct 2013 16:48:46 +0000 (12:48 -0400)] 
- The method signature of :meth:`.Dialect.reflecttable`, which in
all known cases is provided by :class:`.DefaultDialect`, has been
tightened to expect ``include_columns`` and ``exclude_columns``
arguments without any kw option, reducing ambiguity - previously
``exclude_columns`` was missing. [ticket:2748]

12 years agoA :func:`.select` that is made to refer to itself in its FROM clause,
Mike Bayer [Wed, 9 Oct 2013 00:06:58 +0000 (20:06 -0400)] 
A :func:`.select` that is made to refer to itself in its FROM clause,
typically via in-place mutation, will raise an informative error
message rather than causing a recursion overflow.
[ticket:2815]

12 years agoFixed bug where using an annotation such as :func:`.remote` or
Mike Bayer [Tue, 8 Oct 2013 23:47:13 +0000 (19:47 -0400)] 
Fixed bug where using an annotation such as :func:`.remote` or
:func:`.foreign` on a :class:`.Column` before association with a parent
:class:`.Table` could produce issues related to the parent table not
rendering within joins, due to the inherent copy operation performed
by an annotation. [ticket:2813]

12 years ago0.8 changelog
Mike Bayer [Tue, 8 Oct 2013 23:25:55 +0000 (19:25 -0400)] 
0.8 changelog

12 years agoThe erroneous kw arg "schema" has been removed from the :class:`.ForeignKey`
Mike Bayer [Tue, 8 Oct 2013 23:20:09 +0000 (19:20 -0400)] 
The erroneous kw arg "schema" has been removed from the :class:`.ForeignKey`
object. this was an accidental commit that did nothing; a warning is raised
in 0.8.3 when this kw arg is used. [ticket:2831]

12 years ago- fix a bunch of test failures
Mike Bayer [Tue, 8 Oct 2013 14:42:13 +0000 (10:42 -0400)] 
- fix a bunch of test failures

12 years ago- add an option to Bundle single_entity=True to allow for single
Mike Bayer [Mon, 7 Oct 2013 16:53:04 +0000 (12:53 -0400)] 
- add an option to Bundle single_entity=True to allow for single
entity returns without otherwise changing much [ticket:2824]

12 years ago- fix a glitch with how the default thing worked out
Mike Bayer [Mon, 7 Oct 2013 16:51:38 +0000 (12:51 -0400)] 
- fix a glitch with how the default thing worked out

12 years ago- add some tests for propagate of wildcard lazyload
Mike Bayer [Mon, 7 Oct 2013 01:07:20 +0000 (21:07 -0400)] 
- add some tests for propagate of wildcard lazyload

12 years ago11th hour realization that Load() needs to do the _chop_path() thing as
Mike Bayer [Mon, 7 Oct 2013 00:12:28 +0000 (20:12 -0400)] 
11th hour realization that Load() needs to do the _chop_path() thing as
well.  this probably has some bugs

12 years ago- merge ticket_1418 branch, [ticket:1418]
Mike Bayer [Mon, 7 Oct 2013 00:29:08 +0000 (20:29 -0400)] 
- merge ticket_1418 branch, [ticket:1418]
- The system of loader options has been entirely rearchitected to build
upon a much more comprehensive base, the :class:`.Load` object.  This
base allows any common loader option like :func:`.joinedload`,
:func:`.defer`, etc. to be used in a "chained" style for the purpose
of specifying options down a path, such as ``joinedload("foo").subqueryload("bar")``.
The new system supersedes the usage of dot-separated path names,
multiple attributes within options, and the usage of ``_all()`` options.
-  Added a new load option :func:`.orm.load_only`.  This allows a series
of column names to be specified as loading "only" those attributes,
deferring the rest.

12 years ago- fix propagation of quote flag within _gen_label() so that if the
Mike Bayer [Sun, 6 Oct 2013 03:47:49 +0000 (23:47 -0400)] 
- fix propagation of quote flag within _gen_label() so that if the
name is already an instance of _anonymous_label(), we don't downgrade
it to a plain quoted_name - fixes regression from [ticket:2812].
[ticket:2834]

12 years agofix entity_zero resolution
Mike Bayer [Fri, 4 Oct 2013 19:22:43 +0000 (15:22 -0400)] 
fix entity_zero resolution

12 years ago- A new construct :class:`.Bundle` is added, which allows for specification
Mike Bayer [Thu, 3 Oct 2013 21:06:55 +0000 (17:06 -0400)] 
-  A new construct :class:`.Bundle` is added, which allows for specification
of groups of column expressions to a :class:`.Query` construct.
The group of columns are returned as a single tuple by default.  The
behavior of :class:`.Bundle` can be overridden however to provide
any sort of result processing to the returned row.  One example included
is :attr:`.Composite.Comparator.bundle`, which applies a bundled form
of a "composite" mapped attribute.
[ticket:2824]
- The :func:`.composite` construct now maintains the return object
when used in a column-oriented :class:`.Query`, rather than expanding
out into individual columns.  This makes use of the new :class:`.Bundle`
feature internally.  This behavior is backwards incompatible; to
select from a composite column which will expand out, use
``MyClass.some_composite.clauses``.

12 years ago- add monetdb
Mike Bayer [Tue, 1 Oct 2013 17:43:26 +0000 (13:43 -0400)] 
- add monetdb
- break out into "production" and "experimental"

12 years ago- add support for removal of instance methods as event listeners, taking
Mike Bayer [Tue, 1 Oct 2013 16:14:34 +0000 (12:14 -0400)] 
- add support for removal of instance methods as event listeners, taking
into account the id() of the function itself and self, [ticket:2832]

12 years ago- put exact version string in the test
Mike Bayer [Sun, 29 Sep 2013 21:24:29 +0000 (17:24 -0400)] 
- put exact version string in the test
- use match with a .* preceding instead of search

12 years agochangelog
Mike Bayer [Sun, 29 Sep 2013 21:21:35 +0000 (17:21 -0400)] 
changelog

12 years agoMerge branch 'master' of github.com:zzzeek/sqlalchemy
Mike Bayer [Sun, 29 Sep 2013 21:19:51 +0000 (17:19 -0400)] 
Merge branch 'master' of github.com:zzzeek/sqlalchemy

12 years agoMerge pull request #26 from hackop/ticket_2819
mike bayer [Sun, 29 Sep 2013 21:19:44 +0000 (14:19 -0700)] 
Merge pull request #26 from hackop/ticket_2819

re.match to re.search

12 years agoadd test for upcoming pullreq
Mike Bayer [Sun, 29 Sep 2013 21:19:25 +0000 (17:19 -0400)] 
add test for upcoming pullreq

12 years agoits a feature
Mike Bayer [Sun, 29 Sep 2013 21:16:57 +0000 (17:16 -0400)] 
its a feature

12 years agochangelog
Mike Bayer [Sun, 29 Sep 2013 21:09:25 +0000 (17:09 -0400)] 
changelog

12 years agoMerge pull request #25 from gthb/ticket_2821
mike bayer [Sun, 29 Sep 2013 21:03:46 +0000 (14:03 -0700)] 
Merge pull request #25 from gthb/ticket_2821

Hide password in URL and Engine __repr__

12 years agoMerge pull request #30 from alex/visitor-pattern
mike bayer [Wed, 25 Sep 2013 17:33:34 +0000 (10:33 -0700)] 
Merge pull request #30 from alex/visitor-pattern

Replace a big loop + dict lookup in Connection.execute() with a simple visitor pattern

12 years agoReplace a big loop + dict lookup in Connection.execute() with a simple visitor
Alex Gaynor [Wed, 25 Sep 2013 17:29:52 +0000 (10:29 -0700)] 
Replace a big loop + dict lookup in Connection.execute() with a simple visitor
pattern

12 years ago- Removed some now unneeded version checks [ticket:2829] courtesy alex gaynor
Mike Bayer [Mon, 23 Sep 2013 00:35:40 +0000 (20:35 -0400)] 
- Removed some now unneeded version checks [ticket:2829] courtesy alex gaynor

12 years agofix typo [ticket:2827]
Mike Bayer [Tue, 17 Sep 2013 22:48:48 +0000 (17:48 -0500)] 
fix typo [ticket:2827]

12 years ago render "backports"/ "forward ports" automatically, only write out
Mike Bayer [Tue, 17 Sep 2013 20:03:15 +0000 (15:03 -0500)] 
render "backports"/ "forward ports" automatically, only write out
    changelog once

12 years ago- modify what we did in [ticket:2793] so that we can also set the
Mike Bayer [Sat, 7 Sep 2013 01:39:36 +0000 (21:39 -0400)] 
- modify what we did in [ticket:2793] so that we can also set the
version id programmatically outside of the generator.   using this system,
we can also leave the version id alone.

12 years agore.match to re.search
Scott Schaefer [Sat, 7 Sep 2013 00:14:26 +0000 (18:14 -0600)] 
re.match to re.search

Convert to re.search to eliminate the restriction on only matching the
beginning of the string

12 years agoHide password in URL and Engine __repr__
Gunnlaugur Þór Briem [Fri, 6 Sep 2013 17:55:19 +0000 (17:55 +0000)] 
Hide password in URL and Engine __repr__

Fixes #2821

12 years agoFixed Query.exists() method for the case, when query doesn't have any filters applied.
Vladimir Magamedov [Wed, 4 Sep 2013 08:43:40 +0000 (11:43 +0300)] 
Fixed Query.exists() method for the case, when query doesn't have any filters applied.

12 years agoadd caveats regarding RETURNING
Mike Bayer [Mon, 2 Sep 2013 16:33:49 +0000 (12:33 -0400)] 
add caveats regarding RETURNING

12 years agouse a different col here to keep oracle happy
Mike Bayer [Mon, 2 Sep 2013 16:11:24 +0000 (12:11 -0400)] 
use a different col here to keep oracle happy

12 years agoplus some more adjustments for mysql, or in general if an Index refers to
Mike Bayer [Wed, 28 Aug 2013 21:31:40 +0000 (17:31 -0400)] 
plus some more adjustments for mysql, or in general if an Index refers to
in-python only cols

12 years agoFixed bug where using the ``column_reflect`` event to change the ``.key``
Mike Bayer [Wed, 28 Aug 2013 21:25:44 +0000 (17:25 -0400)] 
Fixed bug where using the ``column_reflect`` event to change the ``.key``
of the incoming :class:`.Column` would prevent primary key constraints,
indexes, and foreign key constraints from being correctly reflected.
Also in 0.8.3. [ticket:2811]

12 years ago- fix bug here in profiling.py
Mike Bayer [Wed, 28 Aug 2013 14:19:11 +0000 (10:19 -0400)] 
- fix bug here in profiling.py
- callcount

12 years agorepair missing oracle skip
Mike Bayer [Wed, 28 Aug 2013 14:16:52 +0000 (10:16 -0400)] 
repair missing oracle skip

12 years agocomment this alternative approach...
Mike Bayer [Wed, 28 Aug 2013 14:16:28 +0000 (10:16 -0400)] 
comment this alternative approach...

12 years ago- these tests now fail on a normalize name DB, as they should! because we're quoting...
Mike Bayer [Wed, 28 Aug 2013 04:11:15 +0000 (00:11 -0400)] 
- these tests now fail on a normalize name DB, as they should! because we're quoting the lowercase name.

12 years ago- rework the profile thing to just rewrite all failing numbers when --write-profiles...
Mike Bayer [Wed, 28 Aug 2013 04:04:38 +0000 (00:04 -0400)] 
- rework the profile thing to just rewrite all failing numbers when --write-profiles is set
- some sqlite callcounts

12 years agocallcounts
Mike Bayer [Wed, 28 Aug 2013 03:51:34 +0000 (23:51 -0400)] 
callcounts

12 years ago- cx_oracle seems to have a bug here though it is hard to track down
Mike Bayer [Wed, 28 Aug 2013 01:37:22 +0000 (21:37 -0400)] 
- cx_oracle seems to have a bug here though it is hard to track down
- cx_oracle dialect doesn't use normal col names, lets just not rely on that for now

12 years agomore seealsos...
Mike Bayer [Wed, 28 Aug 2013 00:46:18 +0000 (20:46 -0400)] 
more seealsos...

12 years ago- A rework to the way that "quoted" identifiers are handled, in that
Mike Bayer [Wed, 28 Aug 2013 00:43:22 +0000 (20:43 -0400)] 
- A rework to the way that "quoted" identifiers are handled, in that
instead of relying upon various ``quote=True`` flags being passed around,
these flags are converted into rich string objects with quoting information
included at the point at which they are passed to common schema constructs
like :class:`.Table`, :class:`.Column`, etc.   This solves the issue
of various methods that don't correctly honor the "quote" flag such
as :meth:`.Engine.has_table` and related methods.  The :class:`.quoted_name`
object is a string subclass that can also be used explicitly if needed;
the object will hold onto the quoting preferences passed and will
also bypass the "name normalization" performed by dialects that
standardize on uppercase symbols, such as Oracle, Firebird and DB2.
The upshot is that the "uppercase" backends can now work with force-quoted
names, such as lowercase-quoted names and new reserved words.
[ticket:2812]

12 years agocorrect for missing fail() methods which were lost when we removed unittest.TestCase
Mike Bayer [Mon, 26 Aug 2013 16:30:33 +0000 (12:30 -0400)] 
correct for missing fail() methods which were lost when we removed unittest.TestCase

12 years agoclean up formatting and other linting issues
Mike Bayer [Mon, 26 Aug 2013 16:29:27 +0000 (12:29 -0400)] 
clean up formatting and other linting issues

12 years ago- ensure rowcount is returned for an UPDATE with no implicit returning
Mike Bayer [Sun, 25 Aug 2013 21:37:59 +0000 (17:37 -0400)] 
- ensure rowcount is returned for an UPDATE with no implicit returning
- modernize test for that
- use py3k compatible next() in test_returning/test_versioning

12 years ago- fix server-side non-returning example
Mike Bayer [Sun, 25 Aug 2013 19:00:54 +0000 (15:00 -0400)] 
- fix server-side non-returning example
- mention eager_defaults

12 years agomigration for version id feature, eager_defaults
Mike Bayer [Sun, 25 Aug 2013 18:37:35 +0000 (14:37 -0400)] 
migration for version id feature, eager_defaults

12 years ago- restore constructor docs blown away in the merge
Mike Bayer [Sun, 25 Aug 2013 18:13:03 +0000 (14:13 -0400)] 
- restore constructor docs blown away in the merge

12 years ago- The ``version_id_generator`` parameter of ``Mapper`` can now be specified
Mike Bayer [Sun, 25 Aug 2013 18:03:54 +0000 (14:03 -0400)] 
- The ``version_id_generator`` parameter of ``Mapper`` can now be specified
to rely upon server generated version identifiers, using triggers
or other database-provided versioning features, by passing the value
``False``.  The ORM will use RETURNING when available to immediately
load the new version identifier, else it will emit a second SELECT.
[ticket:2793]
- The ``eager_defaults`` flag of :class:`.Mapper` will now allow the
newly generated default values to be fetched using an inline
RETURNING clause, rather than a second SELECT statement, for backends
that support RETURNING.
- Added a new variant to :meth:`.ValuesBase.returning` called
:meth:`.ValuesBase.return_defaults`; this allows arbitrary columns
to be added to the RETURNING clause of the statement without interfering
with the compilers usual "implicit returning" feature, which is used to
efficiently fetch newly generated primary key values.  For supporting
backends, a dictionary of all fetched values is present at
:attr:`.ResultProxy.returned_defaults`.
- add a glossary entry for RETURNING
- add documentation for version id generation, [ticket:867]

12 years agoforgot to add system to the copy() method
Mike Bayer [Sun, 25 Aug 2013 16:33:37 +0000 (12:33 -0400)] 
forgot to add system to the copy() method

Conflicts:
lib/sqlalchemy/schema.py

12 years agoadded "system=True" to Column, so that we generally don't have to bother
Mike Bayer [Sun, 25 Aug 2013 16:28:47 +0000 (12:28 -0400)] 
added "system=True" to Column, so that we generally don't have to bother
with CreateColumn rules

12 years agoMerge pull request #23 from yoloseem/master
mike bayer [Sun, 25 Aug 2013 16:08:00 +0000 (09:08 -0700)] 
Merge pull request #23 from yoloseem/master

Fixed two typos.

12 years agofix typos
Hyunjun Kim [Sat, 24 Aug 2013 16:59:47 +0000 (01:59 +0900)] 
fix typos

12 years ago- The :class:`.CreateColumn` construct can be appled to a custom
Mike Bayer [Sat, 24 Aug 2013 17:55:14 +0000 (13:55 -0400)] 
- The :class:`.CreateColumn` construct can be appled to a custom
compilation rule which allows skipping of columns, by producing
a rule that returns ``None``.  Also in 0.8.3.

12 years ago- "primary mapper entity" is now an attribute on Query,
Mike Bayer [Sat, 24 Aug 2013 15:27:44 +0000 (11:27 -0400)] 
- "primary mapper entity" is now an attribute on Query,
making _MapperEntity slightly less dependent on a particular parent
Query (in theory more shareable by multiple Query objects in different contexts)
- remove some comments that have been misunderstanding what _mapper_entities
does, or perhaps forgot to get removed
- simplify _mapper_entities

12 years agoformatting
Mike Bayer [Fri, 23 Aug 2013 22:58:09 +0000 (18:58 -0400)] 
formatting

12 years ago- update the test times, include pypy, clean up the script
Mike Bayer [Wed, 21 Aug 2013 23:18:18 +0000 (19:18 -0400)] 
- update the test times, include pypy, clean up the script

12 years agoadd FAQ entry on large inserts
Mike Bayer [Wed, 21 Aug 2013 23:09:01 +0000 (19:09 -0400)] 
add FAQ entry on large inserts

12 years agofix link
Mike Bayer [Wed, 21 Aug 2013 22:57:00 +0000 (18:57 -0400)] 
fix link

12 years agomove FAQ to the docs, [ticket:2133]
Mike Bayer [Wed, 21 Aug 2013 22:48:34 +0000 (18:48 -0400)] 
move FAQ to the docs, [ticket:2133]

12 years agoyikes return the modname if no lookup found...
Mike Bayer [Wed, 21 Aug 2013 21:14:04 +0000 (17:14 -0400)] 
yikes return the modname if no lookup found...

12 years agoFixed bug where list instrumentation would fail to represent a
Mike Bayer [Tue, 20 Aug 2013 17:27:27 +0000 (13:27 -0400)] 
Fixed bug where list instrumentation would fail to represent a
setslice of ``[0:0]`` correctly, which in particular could occur
when using ``insert(0, item)`` with the association proxy.  Due
to some quirk in Python collections, the issue was much more likely
with Python 3 rather than 2.  Also in 0.8.3, 0.7.11.
[ticket:2807]