]> git.ipfire.org Git - thirdparty/sqlalchemy/alembic.git/log
thirdparty/sqlalchemy/alembic.git
10 years ago0.7.7 rel_0_7 gh/rel_0_7 rel_0_7_7
Mike Bayer [Wed, 22 Jul 2015 17:25:37 +0000 (13:25 -0400)] 
0.7.7

10 years ago- fix this name for the 0.7 branch
Mike Bayer [Wed, 22 Jul 2015 16:53:56 +0000 (12:53 -0400)] 
- fix this name for the 0.7 branch

10 years ago- Fixed critical issue where a complex series of branches/merges would
Mike Bayer [Wed, 22 Jul 2015 16:37:35 +0000 (12:37 -0400)] 
- Fixed critical issue where a complex series of branches/merges would
bog down the iteration algorithm working over redundant nodes for
millions of cycles.   An internal adjustment has been
made so that duplicate nodes are skipped within this iteration.
fixes #310

10 years ago- Implemented support for :meth:`.BatchOperations.create_primary_key`
Mike Bayer [Fri, 3 Jul 2015 21:29:17 +0000 (17:29 -0400)] 
- Implemented support for :meth:`.BatchOperations.create_primary_key`
and :meth:`.BatchOperations.create_check_constraint`.
fixes #305
- table keyword arguments are copied from the original reflected table,
such as the "mysql_engine" keyword argument.

10 years ago- add a changelog for the legacy_schema_aliasing warning issue,
Mike Bayer [Fri, 26 Jun 2015 22:09:50 +0000 (18:09 -0400)] 
- add a changelog for the legacy_schema_aliasing warning issue,
the fix was committed in 229f8672.

10 years ago- add clarification that this tutorial is against the generic template
Mike Bayer [Wed, 17 Jun 2015 18:54:34 +0000 (14:54 -0400)] 
- add clarification that this tutorial is against the generic template

10 years ago- The :meth:`.MigrationContext.stamp` method, added as part of the
Mike Bayer [Tue, 9 Jun 2015 15:40:34 +0000 (11:40 -0400)] 
- The :meth:`.MigrationContext.stamp` method, added as part of the
versioning refactor in 0.7 as a more granular version of
:func:`.command.stamp`, now includes the "create the alembic_version
table if not present" step in the same way as the command version,
which was previously omitted.
fixes #300

10 years ago- turn on the mssql legacy_schema_aliasing flag to remove
Mike Bayer [Tue, 26 May 2015 14:55:36 +0000 (10:55 -0400)] 
- turn on the mssql legacy_schema_aliasing flag to remove
warnings

10 years ago- Fixed bug where foreign key options including "onupdate",
Mike Bayer [Tue, 19 May 2015 00:21:49 +0000 (20:21 -0400)] 
- Fixed bug where foreign key options including "onupdate",
"ondelete" would not render within the ``op.create_foreign_key()``
directive, even though they render within a full
``ForeignKeyConstraint`` directive.
fixes #298

10 years ago- add test for FKC with multiple kwargs, reference #92
Mike Bayer [Mon, 18 May 2015 23:48:09 +0000 (19:48 -0400)] 
- add test for FKC with multiple kwargs, reference #92

10 years ago0.7.6 rel_0_7_6
Mike Bayer [Tue, 5 May 2015 21:02:26 +0000 (17:02 -0400)] 
0.7.6

10 years agoMerged in jdanjou/alembic/jd/typo-doc-comparator (pull request #44)
Mike Bayer [Mon, 4 May 2015 16:57:13 +0000 (12:57 -0400)] 
Merged in jdanjou/alembic/jd/typo-doc-comparator (pull request #44)

Fix typo in autogenerate documentation

10 years agoFix typo in autogenerate documentation
Julien Danjou [Mon, 4 May 2015 15:57:42 +0000 (17:57 +0200)] 
Fix typo in autogenerate documentation

10 years ago- Fixed bug where the case of multiple mergepoints that all
Mike Bayer [Mon, 4 May 2015 00:17:19 +0000 (20:17 -0400)] 
- Fixed bug where the case of multiple mergepoints that all
have the identical set of ancestor revisions would fail to be
upgradable, producing an assertion failure.   Merge points were
previously assumed to always require at least an UPDATE in
alembic_revision from one of the previous revs to the new one,
however in this case, if one of the mergepoints has already
been reached, the remaining mergepoints have no row to UPDATE therefore
they must do an INSERT of their target version.
fixes #297

10 years ago- add tox targets for 3.4
Mike Bayer [Fri, 1 May 2015 18:54:44 +0000 (14:54 -0400)] 
- add tox targets for 3.4

10 years ago- Added support for type comparison functions to be not just per
Mike Bayer [Thu, 30 Apr 2015 15:33:58 +0000 (11:33 -0400)] 
- Added support for type comparison functions to be not just per
environment, but also present on the custom types themselves, by
supplying a method ``compare_against_backend``.
Added a new documentation section :ref:`compare_types` describing
type comparison fully.
fixes #296

10 years agoMerged in rtaranu/alembic (pull request #43)
Mike Bayer [Tue, 21 Apr 2015 17:04:50 +0000 (13:04 -0400)] 
Merged in rtaranu/alembic (pull request #43)

- fixed spelling mistake in docs

10 years ago- fixed spelling mistake in docs
Raul Taranu [Tue, 21 Apr 2015 13:22:40 +0000 (13:22 +0000)] 
- fixed spelling mistake in docs

10 years ago- Added a new option
Mike Bayer [Tue, 7 Apr 2015 16:36:51 +0000 (12:36 -0400)] 
- Added a new option
:paramref:`.EnvironmentContext.configure.literal_binds`, which
will pass the ``literal_binds`` flag into the compilation of SQL
constructs when using "offline" mode.  This has the effect that
SQL objects like inserts, updates, deletes as well as textual
statements sent using ``text()`` will be compiled such that the dialect
will attempt to render literal values "inline" automatically.
Only a subset of types is typically supported; the
:meth:`.Operations.inline_literal` construct remains as the construct
used to force a specific literal representation of a value.
The :paramref:`.EnvironmentContext.configure.literal_binds` flag
is added to the "offline" section of the ``env.py`` files generated
in new environments.
fixes #255
- enhance the op_fixture as well as MigrationContext._stdout_connection()
 so that it uses the real DefaultImpl
and MigrationContext fully in tests.

10 years ago- repair name of starting_rev argument, fixes #290
Mike Bayer [Sun, 29 Mar 2015 13:33:27 +0000 (09:33 -0400)] 
- repair name of starting_rev argument, fixes #290

10 years ago- dont rely on inpsect import, not in SQLA 0.7
Mike Bayer [Sat, 28 Mar 2015 16:51:45 +0000 (12:51 -0400)] 
- dont rely on inpsect import, not in SQLA 0.7

10 years ago- don't use op.f(), only works with later SQLA versions
Mike Bayer [Sat, 28 Mar 2015 15:55:50 +0000 (11:55 -0400)] 
- don't use op.f(), only works with later SQLA versions

10 years ago- Fully implemented the
Mike Bayer [Fri, 27 Mar 2015 22:55:00 +0000 (18:55 -0400)] 
- Fully implemented the
:paramref:`~.Operations.batch_alter_table.copy_from` parameter for
batch mode, which previously was not functioning.  This allows
"batch mode" to be usable in conjunction with ``--sql``.
fixes #289
- sqlite dialect checks for "create_index" and "drop_index" as exceptions
for "recreate" in batch mode, the same way as "add_column", so that
unnecessary table recreates don't emit for index-only operations

10 years ago- next version
Mike Bayer [Fri, 27 Mar 2015 22:52:20 +0000 (18:52 -0400)] 
- next version

10 years ago- Repaired support for the :meth:`.BatchOperations.create_index`
Mike Bayer [Fri, 27 Mar 2015 19:15:55 +0000 (15:15 -0400)] 
- Repaired support for the :meth:`.BatchOperations.create_index`
directive, which was mis-named internally such that the operation
within a batch context could not proceed.
fixes #287

10 years ago- didn't name the internal number correctly, stil rel_0_7_5_post2
Mike Bayer [Fri, 20 Mar 2015 19:22:17 +0000 (15:22 -0400)] 
- didn't name the internal number correctly, stil
don't have this right.  up to post2

10 years ago- 0.7.5.post1 due to pypi deletion issue rel_0_7_5_post1
Mike Bayer [Thu, 19 Mar 2015 21:18:14 +0000 (17:18 -0400)] 
- 0.7.5.post1 due to pypi deletion issue

10 years ago0.7.5 rel_0_7_5
Mike Bayer [Thu, 19 Mar 2015 20:34:56 +0000 (16:34 -0400)] 
0.7.5

10 years ago- changelog for #266
Mike Bayer [Thu, 19 Mar 2015 19:51:11 +0000 (15:51 -0400)] 
- changelog for #266
- use exception fixture
- look directly at context.as_sql as that's where
the "sql mode" is most authoritative
- fixes #266

10 years agoMerge branch 'issue266' of https://bitbucket.org/jerdfelt/alembic into pr39
Mike Bayer [Thu, 19 Mar 2015 19:40:38 +0000 (15:40 -0400)] 
Merge branch 'issue266' of https://bitbucket.org/jerdfelt/alembic into pr39

10 years ago- Fixed bug where the mssql DROP COLUMN directive failed to include
Mike Bayer [Wed, 11 Mar 2015 15:34:40 +0000 (11:34 -0400)] 
- Fixed bug where the mssql DROP COLUMN directive failed to include
modifiers such as "schema" when emitting the DDL.
fixes #284

10 years ago- an adjustment to the logic for #282 - if the index we see here
Mike Bayer [Tue, 10 Mar 2015 20:14:21 +0000 (16:14 -0400)] 
- an adjustment to the logic for #282 - if the index we see here
*is* in conn_indexes_by_name, then obviously we should leave it in.

10 years ago- Postgresql "functional" indexes are necessarily skipped from the
Mike Bayer [Tue, 10 Mar 2015 15:26:43 +0000 (11:26 -0400)] 
- Postgresql "functional" indexes are necessarily skipped from the
autogenerate process, as the SQLAlchemy backend currently does not
support reflection of these structures.   A warning is emitted
both from the SQLAlchemy backend as well as from the Alembic
backend for Postgresql when such an index is detected.
fixes #282

10 years ago- add support for assertion of warnings emitted
Mike Bayer [Tue, 10 Mar 2015 15:26:21 +0000 (11:26 -0400)] 
- add support for assertion of warnings emitted

10 years ago- make the repo dynamic
Mike Bayer [Tue, 10 Mar 2015 15:26:07 +0000 (11:26 -0400)] 
- make the repo dynamic

10 years agoRaise exception if autogenerate is tried with as_sql=True
Johannes Erdfelt [Thu, 15 Jan 2015 15:37:20 +0000 (07:37 -0800)] 
Raise exception if autogenerate is tried with as_sql=True

This configuration is nonsensical since autogenerate needs to query
the database for schema information.

Fixes issue #266

10 years ago- Fixed bug where MySQL backend would report dropped unique indexes
Mike Bayer [Sat, 21 Feb 2015 00:20:53 +0000 (19:20 -0500)] 
- Fixed bug where MySQL backend would report dropped unique indexes
and/or constraints as both at the same time.  This is because
MySQL doesn't actually have a "unique constraint" construct that
reports differently than a "unique index", so it is present in both
lists.  The net effect though is that the MySQL backend will report
a dropped unique index/constraint as an index in cases where the object
was first created as a unique constraint, if no other information
is available to make the decision.  This differs from other backends
like Postgresql which can report on unique constraints and
unique indexes separately.
fixes #276

10 years ago- additional fix which impacts #267. fix filtered_heads() to accommodate
Mike Bayer [Mon, 9 Feb 2015 00:27:42 +0000 (19:27 -0500)] 
- additional fix which impacts #267.  fix filtered_heads() to accommodate
being given "heads" as the target so that it will in fact match when all heads
are given.  fixes #267

10 years ago- move resolution of "starting rev" for --sql mode into
Mike Bayer [Tue, 3 Feb 2015 16:40:40 +0000 (11:40 -0500)] 
- move resolution of "starting rev" for --sql mode into
get_current_heads() directly; therefore we don't need to
do this in alembic.command, which we were doing for stamp but
not downgrade/upgrade.  The slight change here is that the
context.get_starting_revision_argument() method will
return an abbreviated starting rev as abbreviated in
all cases, including the stamp command, where we previously
were converting a stamp argument first, but not for the
upgrade or downgrade commands.
- Fixed bug where using a partial revision identifier as the
"starting revision" in ``--sql`` mode in a downgrade operation
would fail to resolve properly.  fixes #269

10 years agoremove errant pdb
Mike Bayer [Sat, 24 Jan 2015 01:09:46 +0000 (20:09 -0500)] 
remove errant pdb

10 years ago- Added a new feature :attr:`.Config.attributes`, to help with the use
Mike Bayer [Sat, 24 Jan 2015 01:05:02 +0000 (20:05 -0500)] 
- Added a new feature :attr:`.Config.attributes`, to help with the use
case of sharing state such as engines and connections on the outside
with a series of Alembic API calls; also added a new cookbook section
to describe this simple but pretty important use case.

10 years agofix missing changelog directive
Mike Bayer [Sat, 24 Jan 2015 00:41:44 +0000 (19:41 -0500)] 
fix missing changelog directive

10 years ago- Added support for "alembic stamp" to work when given "heads" as an
Mike Bayer [Fri, 23 Jan 2015 23:46:53 +0000 (18:46 -0500)] 
- Added support for "alembic stamp" to work when given "heads" as an
argument, when multiple heads are present.
fixes #267

10 years ago- set up for next rev
Mike Bayer [Fri, 23 Jan 2015 23:46:32 +0000 (18:46 -0500)] 
- set up for next rev

10 years ago- fix all flake8
Mike Bayer [Fri, 23 Jan 2015 23:31:17 +0000 (18:31 -0500)] 
- fix all flake8

10 years ago- the enum render is an 0.9 regression, apparently we used
Mike Bayer [Wed, 14 Jan 2015 01:19:17 +0000 (20:19 -0500)] 
- the enum render is an 0.9 regression, apparently we used
a different repr() scheme in 0.7.9->0.8 that didn't omit native_enum

10 years ago- 0.7.4 rel_0_7_4
Mike Bayer [Mon, 12 Jan 2015 19:16:28 +0000 (14:16 -0500)] 
- 0.7.4

10 years ago- happy new year
Mike Bayer [Mon, 12 Jan 2015 19:15:35 +0000 (14:15 -0500)] 
- happy new year

10 years agochangelog for #241, fixes #241
Mike Bayer [Mon, 12 Jan 2015 18:59:15 +0000 (13:59 -0500)] 
changelog for #241, fixes #241

10 years agoMerge https://bitbucket.org/dtheodor/alembic/branch/master into pr37
Mike Bayer [Mon, 12 Jan 2015 18:49:16 +0000 (13:49 -0500)] 
Merge https://bitbucket.org/dtheodor/alembic/branch/master into pr37

10 years agochecked test failure
Dimitris Theodorou [Mon, 12 Jan 2015 16:33:34 +0000 (17:33 +0100)] 
checked test failure

10 years agoremove redundant Float type check
Dimitris Theodorou [Mon, 12 Jan 2015 03:16:14 +0000 (04:16 +0100)] 
remove redundant Float type check

10 years agoEnum autogenerate renders the native_enum flag as well
Dimitris Theodorou [Mon, 12 Jan 2015 01:37:19 +0000 (02:37 +0100)] 
Enum autogenerate renders the native_enum flag as well

10 years agoChange single-quoting of floats in PostgreSQL compare_server_default
Dimitris Theodorou [Sun, 11 Jan 2015 23:41:59 +0000 (00:41 +0100)] 
Change single-quoting of floats in PostgreSQL compare_server_default

Do not wrap string defaults with single quotes when comparing against
columns of type float or numeric.
This fixes the crash occuring when the default of a float column is
an integer value (e.g., DEFAULT 5), while the Python server_default is
a string (e.g., server_default="5.0"). This results in the query
used in the comparison to throw a DataError ('SELECT 5 = '5.0').

10 years ago- The rendering of a :class:`~sqlalchemy.schema.ForeignKeyConstraint`
Mike Bayer [Sat, 10 Jan 2015 22:08:03 +0000 (17:08 -0500)] 
- The rendering of a :class:`~sqlalchemy.schema.ForeignKeyConstraint`
will now ensure that the names of the source and target columns are
the database-side name of each column, and not the value of the
``.key`` attribute as may be set only on the Python side.
This is because Alembic generates the DDL for constraints
as standalone objects without the need to actually refer to an in-Python
:class:`~sqlalchemy.schema.Table` object, so there's no step that
would resolve these Python-only key names to database column names.
fixes #259

10 years ago- Fixed bug in foreign key autogenerate where if the in-Python table
Mike Bayer [Sat, 10 Jan 2015 21:32:17 +0000 (16:32 -0500)] 
- Fixed bug in foreign key autogenerate where if the in-Python table
used custom column keys (e.g. using the ``key='foo'`` kwarg to
``Column``), the comparison of existing foreign keys to those specified
in the metadata would fail, as the reflected table would not have
these keys available which to match up.  Foreign key comparison for
autogenerate now ensures it's looking at the database-side names
of the columns in all cases; this matches the same functionality
within unique constraints and indexes.
fixes #260

10 years ago- make this test use legal table designs
Mike Bayer [Sat, 10 Jan 2015 21:23:10 +0000 (16:23 -0500)] 
- make this test use legal table designs

10 years ago- add/remove seemed backwards here
Mike Bayer [Sat, 10 Jan 2015 21:18:30 +0000 (16:18 -0500)] 
- add/remove seemed backwards here

10 years ago- changelog: Fixed issue in autogenerate type rendering where types that belong
Mike Bayer [Fri, 9 Jan 2015 16:26:36 +0000 (11:26 -0500)] 
- changelog:  Fixed issue in autogenerate type rendering where types that belong
to modules that have the name "sqlalchemy" in them would be mistaken
as being part of the ``sqlalchemy.`` namespace.  Pull req courtesy
Bartosz Burclaf.  fixes #261

10 years agoFixes user module namespace for custom types
Bartosz Burclaf [Fri, 9 Jan 2015 16:14:29 +0000 (17:14 +0100)] 
Fixes user module namespace for custom types

10 years ago0.7.3 rel_0_7_3
Mike Bayer [Tue, 30 Dec 2014 17:14:44 +0000 (12:14 -0500)] 
0.7.3

10 years ago- Fixed regression in new versioning system where upgrade / history
Mike Bayer [Tue, 30 Dec 2014 17:12:58 +0000 (12:12 -0500)] 
- Fixed regression in new versioning system where upgrade / history
operation would fail on AttributeError if no version files were
present at all.
fixes #258

10 years ago- put sqlalchemy master here as SQLA issue https://bitbucket.org/zzzeek/sqlalchemy...
Mike Bayer [Fri, 19 Dec 2014 17:59:31 +0000 (12:59 -0500)] 
- put sqlalchemy master here as SQLA issue https://bitbucket.org/zzzeek/sqlalchemy/issue/3218
is blowing up our sphinx build

10 years ago0.7.2 rel_0_7_2
Mike Bayer [Thu, 18 Dec 2014 20:30:44 +0000 (15:30 -0500)] 
0.7.2

10 years ago- Adjusted the SQLite backend regarding autogen of unique constraints
Mike Bayer [Sun, 14 Dec 2014 17:10:56 +0000 (12:10 -0500)] 
- Adjusted the SQLite backend regarding autogen of unique constraints
to work fully with the current SQLAlchemy 1.0, which now will report
on UNIQUE constraints that have no name.
- fix named foreign key test requirements for SQLAlchemy 1.0 sqlite
FK reflection

10 years ago- identify a test that can't pass on 0.8 for now
Mike Bayer [Sun, 14 Dec 2014 02:15:22 +0000 (21:15 -0500)] 
- identify a test that can't pass on 0.8 for now

10 years ago- also add an explicit schema test for the multi fk condition
Mike Bayer [Sat, 13 Dec 2014 19:16:34 +0000 (14:16 -0500)] 
- also add an explicit schema test for the multi fk condition

10 years ago- Fixed bug in batch where if the target table contained multiple
Mike Bayer [Sat, 13 Dec 2014 19:02:14 +0000 (14:02 -0500)] 
- Fixed bug in batch where if the target table contained multiple
foreign keys to the same target table, the batch mechanics would
fail with a "table already exists" error.  Thanks for the help
on this from Lucas Kahlert. fixes #254

10 years ago-Repaired a regression in both the MSSQL and Oracle dialects whereby
Mike Bayer [Fri, 12 Dec 2014 17:55:08 +0000 (12:55 -0500)] 
-Repaired a regression in both the MSSQL and Oracle dialects whereby
the overridden ``_exec()`` method failed to return a value, as is
needed now in the 0.7 series. fixes #253
- add __backend__ to UpdateRevTest which does a great job testing
_exec() pathing for all backends

10 years ago- adjust test here so that it applies to all backends
Mike Bayer [Thu, 4 Dec 2014 18:36:40 +0000 (13:36 -0500)] 
- adjust test here so that it applies to all backends
- skip index at the level of the index, not the columns inside it
- changelog
- bump for 0.7.2

10 years agoMerge branch 'mysql-unique-index-removed' of https://bitbucket.org/jerdfelt/alembic...
Mike Bayer [Thu, 4 Dec 2014 18:30:46 +0000 (13:30 -0500)] 
Merge branch 'mysql-unique-index-removed' of https://bitbucket.org/jerdfelt/alembic into pr35

10 years agoMySQL: Ignore unique indexes when removing implicit indexes
Johannes Erdfelt [Thu, 4 Dec 2014 17:42:46 +0000 (09:42 -0800)] 
MySQL: Ignore unique indexes when removing implicit indexes

MySQL will implicitly create indexes when using foreign keys. Alembic
attempts to remove those implicit indexes so they don't appear as
removes when comparing metadata.

However, unique indexes with the same name as a column are considered
as possibly implicitly created causing alembic to emit a spurious
'remove_constraint'.

Since MySQL will never implicitly create unique indexes, they can be
safely ignored when removing the implicit indexes.

Fixes #251

10 years ago0.7.1 rel_0_7_1
Mike Bayer [Wed, 3 Dec 2014 20:50:37 +0000 (15:50 -0500)] 
0.7.1

10 years ago- The ``render_as_batch`` flag was inadvertently hardcoded to ``True``,
Mike Bayer [Wed, 3 Dec 2014 20:49:58 +0000 (15:49 -0500)] 
- The ``render_as_batch`` flag was inadvertently hardcoded to ``True``,
so all autogenerates were spitting out batch mode...this has been
fixed so that batch mode again is only when selected in env.py.

10 years ago- use a much bigger stamp here as only three chars can collide quite easily
Mike Bayer [Mon, 1 Dec 2014 19:28:38 +0000 (14:28 -0500)] 
- use a much bigger stamp here as only three chars can collide quite easily

10 years ago0.7.1
Mike Bayer [Mon, 1 Dec 2014 03:05:35 +0000 (22:05 -0500)] 
0.7.1

10 years ago- repair the rendering of add_fk_constraint(), quoting there too
Mike Bayer [Mon, 1 Dec 2014 00:30:27 +0000 (19:30 -0500)] 
- repair the rendering of add_fk_constraint(), quoting there too

10 years ago- naming convention limited to SQLA 0.9.4 or greater
Mike Bayer [Mon, 1 Dec 2014 00:20:38 +0000 (19:20 -0500)] 
- naming convention limited to SQLA 0.9.4 or greater

10 years ago- fix drop_foreign_key to include that this only works with SQLite
Mike Bayer [Mon, 1 Dec 2014 00:08:01 +0000 (19:08 -0500)] 
- fix drop_foreign_key to include that this only works with SQLite
as stated

10 years ago- reword the first paragraph about offline mode to better flow
Mike Bayer [Mon, 1 Dec 2014 00:03:06 +0000 (19:03 -0500)] 
- reword the first paragraph about offline mode to better flow
from the preceding sections.

10 years ago- Added :paramref:`~.Operations.batch_alter_table.naming_convention`
Mike Bayer [Sun, 30 Nov 2014 23:55:29 +0000 (18:55 -0500)] 
- Added :paramref:`~.Operations.batch_alter_table.naming_convention`
argument to :meth:`.Operations.batch_alter_table`, as this is necessary
in order to drop foreign key constraints; these are often unnamed
on the target database, and in the case that they are named, SQLAlchemy
is as of the 0.9 series not including these names yet.
- rework the docs on batch + constraints, which remains subject
to a lot of caveats and problems, some to be resolved in SQLAlchemy 1.0

10 years ago- name the include_object hook "foreign_key_constraint"
Mike Bayer [Sun, 30 Nov 2014 23:50:03 +0000 (18:50 -0500)] 
- name the include_object hook "foreign_key_constraint"
- changelog and other doc updates, fixes #178
- fix drop_constraint() unit tests and add two more for FKs

10 years ago- refactor the FK merge a bit
Mike Bayer [Sun, 30 Nov 2014 23:02:57 +0000 (18:02 -0500)] 
- refactor the FK merge a bit
- getting at attributes of FKs varies a bit on SQLA versions,
so implement an _fk_spec() called for all FK inspection
- to enable include_object() filters and allow the FK constraint
code to flow like that of indexes/uniques, change the approach
so that we deal with an _fk_constraint_sig() object again which
contains the real ForeignKeyConstraint() within; we need this
anyway for include_object, but also allows us to use the standard
"drop_constraint" call for rendering.
- enhance tests in test_autogen_fks to support real FK databases like
Postgresql, MySQL, add in InnoDB flags and ensure that FKs refer
to real primary key constraints for PG support
- implement and test include_object() support for FKs
- inspectors all have get_foreign_keys(), no need to check
- repair the drop_constraint call to quote the "type" and table
name correctly, run all constraint drops through drop_constraint()
for rendering
- fix up schema identifiers for foreign key autogens

10 years ago- Merge branch 'add_fk_check' of https://bitbucket.org/akamyshnikova/alembic into...
Mike Bayer [Sun, 30 Nov 2014 19:20:40 +0000 (14:20 -0500)] 
- Merge branch 'add_fk_check' of https://bitbucket.org/akamyshnikova/alembic into pr32
- complete merge, get all tests passing
- use 'foreignkey' literal

Conflicts:
alembic/autogenerate/compare.py
tests/test_autogenerate.py

10 years ago- Fixed bug where the "source_schema" argument was not correctly passed
Mike Bayer [Sun, 30 Nov 2014 16:46:37 +0000 (11:46 -0500)] 
- Fixed bug where the "source_schema" argument was not correctly passed
when calling :meth:`.BatchOperations.create_foreign_key`.  Pull
request courtesy Malte Marquarding.
- add tests

10 years agoMerge branch 'master' of https://bitbucket.org/mmarquar/alembic
Mike Bayer [Sun, 30 Nov 2014 16:44:51 +0000 (11:44 -0500)] 
Merge branch 'master' of https://bitbucket.org/mmarquar/alembic

10 years ago- Repaired the inspection, copying and rendering of CHECK constraints
Mike Bayer [Sat, 29 Nov 2014 21:31:30 +0000 (16:31 -0500)] 
- Repaired the inspection, copying and rendering of CHECK constraints
and so-called "schema" types such as Boolean, Enum within the batch
copy system; the CHECK constraint will not be "doubled" when the table is
copied, and additionally the inspection of the CHECK constraint for
its member columns will no longer fail with an attribute error.
fixes #249
- Added two new arguments
:paramref:`.Operations.batch_alter_table.reflect_args`
and :paramref:`.Operations.batch_alter_table.reflect_kwargs`, so that
arguments may be passed directly to suit the
:class:`~.sqlalchemy.schema.Table`
object that will be reflected.

10 years agoFix source_schema argument in BatchOperation.create_foreign_key
Malte Marquarding [Fri, 28 Nov 2014 01:26:21 +0000 (12:26 +1100)] 
Fix source_schema argument in BatchOperation.create_foreign_key

10 years agofix symbol
Mike Bayer [Wed, 26 Nov 2014 18:18:46 +0000 (13:18 -0500)] 
fix symbol

10 years agoFix failing unittests and update compare_fk method
Ann Kamyshnikova [Tue, 25 Nov 2014 07:07:53 +0000 (10:07 +0300)] 
Fix failing unittests and update compare_fk method

10 years ago0.7.0 rel_0_7_0
Mike Bayer [Mon, 24 Nov 2014 20:11:13 +0000 (15:11 -0500)] 
0.7.0

10 years agofix manifest
Mike Bayer [Mon, 24 Nov 2014 20:20:45 +0000 (15:20 -0500)] 
fix manifest

10 years ago- test runner updates
Mike Bayer [Mon, 24 Nov 2014 20:16:09 +0000 (15:16 -0500)] 
- test runner updates

10 years ago- add a missing @classmethod decorator...
Mike Bayer [Mon, 24 Nov 2014 20:09:39 +0000 (15:09 -0500)] 
- add a missing @classmethod decorator...

10 years ago- The ``alembic revision`` command accepts the ``--sql`` option to
Mike Bayer [Sun, 23 Nov 2014 20:35:12 +0000 (15:35 -0500)] 
- The ``alembic revision`` command accepts the ``--sql`` option to
suit some very obscure use case where the ``revision_environment``
flag is set up, so that ``env.py`` is run when ``alembic revision``
is run even though autogenerate isn't specified.   As this flag is
otherwise confusing, error messages are now raised if
``alembic revision`` is invoked with both ``--sql`` and
``--autogenerate`` or with ``--sql`` without
``revision_environment`` being set.
fixes #248

10 years ago- Relative revision identifiers as used with ``alembic upgrade``,
Mike Bayer [Sun, 23 Nov 2014 20:23:52 +0000 (15:23 -0500)] 
- Relative revision identifiers as used with ``alembic upgrade``,
``alembic downgrade`` and ``alembic history`` can be combined with
specific revisions as well, e.g. ``alembic upgrade ae10+3``, to produce
a migration target relative to the given exact version.

10 years ago- ensure we include for dependencies when we do a stamp, add an option
Mike Bayer [Sun, 23 Nov 2014 15:42:39 +0000 (10:42 -0500)] 
- ensure we include for dependencies when we do a stamp, add an option
to filter_for_lineage

10 years ago- use the correct method here; this needs to have some tests
Mike Bayer [Sun, 23 Nov 2014 15:33:47 +0000 (10:33 -0500)] 
- use the correct method here; this needs to have some tests

10 years agoyikes forgot batch mode
Mike Bayer [Sun, 23 Nov 2014 04:10:12 +0000 (23:10 -0500)] 
yikes forgot batch mode

10 years ago- spiff up autogenerate docs a bit
Mike Bayer [Sun, 23 Nov 2014 03:56:13 +0000 (22:56 -0500)] 
- spiff up autogenerate docs a bit