]> git.ipfire.org Git - thirdparty/sqlalchemy/alembic.git/log
thirdparty/sqlalchemy/alembic.git
10 years ago- add a type here so that to account for pre-0.9 FK/type support
Mike Bayer [Sun, 9 Nov 2014 17:40:35 +0000 (12:40 -0500)] 
- add a type here so that to account for pre-0.9 FK/type support

10 years agoMerge branch 'master' into batch_alter
Mike Bayer [Sun, 9 Nov 2014 17:38:40 +0000 (12:38 -0500)] 
Merge branch 'master' into batch_alter

10 years ago- adjust these to account for some MySQL deprecations regarding
Mike Bayer [Sun, 9 Nov 2014 17:38:16 +0000 (12:38 -0500)] 
- adjust these to account for some MySQL deprecations regarding
duplicate indexes against the same columns

10 years ago- add some connection cleanup
Mike Bayer [Sun, 9 Nov 2014 17:33:26 +0000 (12:33 -0500)] 
- add some connection cleanup

10 years ago- more docs
Mike Bayer [Sun, 9 Nov 2014 04:20:10 +0000 (23:20 -0500)] 
- more docs

10 years agodocumenting
Mike Bayer [Sun, 9 Nov 2014 03:55:51 +0000 (22:55 -0500)] 
documenting

10 years ago- add a little step to get PG to work rudimentally, however
Mike Bayer [Sun, 9 Nov 2014 03:42:07 +0000 (22:42 -0500)] 
- add a little step to get PG to work rudimentally, however
the drop + recreate routine still needs a solution for refernential
integrity for it to be of general use

10 years ago- start docs
Mike Bayer [Sat, 8 Nov 2014 23:59:28 +0000 (18:59 -0500)] 
- start docs
- run tests against mysql/PG to make sure auto mode works at least, try
a proof of concept recreate for MySQL.  recreate doesn't work on PG
as constraint names are global (ugh).  Will have to figure something
out on that.

10 years ago- round trip tests
Mike Bayer [Sat, 8 Nov 2014 23:44:43 +0000 (18:44 -0500)] 
- round trip tests
- fixes to add column

10 years agoMerge branch 'master' into batch_alter
Mike Bayer [Sat, 8 Nov 2014 23:06:47 +0000 (18:06 -0500)] 
Merge branch 'master' into batch_alter

10 years ago- The :class:`~sqlalchemy.schema.Table` object is now returned when
Mike Bayer [Sat, 8 Nov 2014 23:06:26 +0000 (18:06 -0500)] 
- The :class:`~sqlalchemy.schema.Table` object is now returned when
the :meth:`.Operations.create_table` method is used.  This ``Table``
is suitable for use in subsequent SQL operations, in particular
the :meth:`.Operations.bulk_insert` operation.
fixes #205

10 years ago- finish up most features
Mike Bayer [Sat, 8 Nov 2014 22:51:34 +0000 (17:51 -0500)] 
- finish up most features

10 years agomove this
Mike Bayer [Sat, 8 Nov 2014 16:24:47 +0000 (11:24 -0500)] 
move this

10 years ago- testing approaches for BatchOperationsImpl and ApplyBatchImpl
Mike Bayer [Sat, 8 Nov 2014 16:22:34 +0000 (11:22 -0500)] 
- testing approaches for BatchOperationsImpl and ApplyBatchImpl

10 years agoproof of concept
Mike Bayer [Fri, 7 Nov 2014 23:45:33 +0000 (18:45 -0500)] 
proof of concept

10 years agodocstring
Mike Bayer [Fri, 7 Nov 2014 22:20:33 +0000 (17:20 -0500)] 
docstring

10 years ago- add tests for batch autogenerate
Mike Bayer [Fri, 7 Nov 2014 22:06:38 +0000 (17:06 -0500)] 
- add tests for batch autogenerate

10 years ago- add support for autogenerate to include "batch"
Mike Bayer [Fri, 7 Nov 2014 00:37:49 +0000 (19:37 -0500)] 
- add support for autogenerate to include "batch"

10 years ago- use pyprinter for autogen so that we get indentation logic
Mike Bayer [Thu, 6 Nov 2014 23:49:24 +0000 (18:49 -0500)] 
- use pyprinter for autogen so that we get indentation logic

10 years agoMerge branch 'master' into batch_alter
Mike Bayer [Thu, 6 Nov 2014 23:29:34 +0000 (18:29 -0500)] 
Merge branch 'master' into batch_alter

10 years ago- skip unique constraint tests on 0.7
Mike Bayer [Wed, 5 Nov 2014 16:17:23 +0000 (11:17 -0500)] 
- skip unique constraint tests on 0.7

10 years ago- Indexes and unique constraints are now included in the
Mike Bayer [Sun, 2 Nov 2014 23:33:51 +0000 (18:33 -0500)] 
- Indexes and unique constraints are now included in the
:paramref:`.EnvironmentContext.configure.include_object` hook.
Indexes are sent with type ``"index"`` and unique constraints with
type ``"unique_constraint"``.
fixes #203

10 years ago- Bound parameters are now resolved as "literal" values within the
Mike Bayer [Sun, 2 Nov 2014 17:46:20 +0000 (12:46 -0500)] 
- Bound parameters are now resolved as "literal" values within the
SQL expression inside of a CheckConstraint(), when rendering the SQL
as a text string; supported for SQLAlchemy 0.8.0 and forward.
fixes #219

10 years ago- Added a workaround for SQLAlchemy issue #3023 (fixed in 0.9.5) where
Mike Bayer [Sun, 2 Nov 2014 17:35:47 +0000 (12:35 -0500)] 
- Added a workaround for SQLAlchemy issue #3023 (fixed in 0.9.5) where
a column that's part of an explicit PrimaryKeyConstraint would not
have its "nullable" flag set to False, thus producing a false
autogenerate.  Also added a related correction to MySQL which will
correct for MySQL's implicit server default of '0' when a NULL integer
column is turned into a primary key column. fixes #199

10 years ago- Repaired issue related to the fix for #208 and others; a composite
Mike Bayer [Sun, 2 Nov 2014 15:34:58 +0000 (10:34 -0500)] 
- Repaired issue related to the fix for #208 and others; a composite
foreign key reported by MySQL would cause a KeyError as Alembic
attempted to remove MySQL's implicitly generated indexes from the
autogenerate list.
fixes #240

10 years ago- If the "alembic_version" table is present in the target metadata,
Mike Bayer [Thu, 30 Oct 2014 01:01:31 +0000 (21:01 -0400)] 
- If the "alembic_version" table is present in the target metadata,
autogenerate will skip this also.  Pull request courtesy
Dj Gilcrease. fixes #28
- The :paramref:`.EnvironmentContext.configure.version_table`
and :paramref:`.EnvironmentContext.configure.version_table_schema`
arguments are now honored during the autogenerate process, such that
these names will be used as the "skip" names on both the database
reflection and target metadata sides. fixes #77

10 years ago- sketch of how "batch operations" would work
Mike Bayer [Wed, 22 Oct 2014 23:12:17 +0000 (19:12 -0400)] 
- sketch of how "batch operations" would work

10 years ago- the original rationale for defaulting the user-defined namespace
Mike Bayer [Tue, 21 Oct 2014 17:14:06 +0000 (13:14 -0400)] 
- the original rationale for defaulting the user-defined namespace
to "sa." was to force users to deal with making sure their custom
types came from a fixed module somewhere.  However, it's not worth
defending this rationale.

The default value of the
:paramref:`.EnvironmentContext.configure.user_module_prefix`
parameter is **no longer the same as the SQLAlchemy prefix**.
When omitted, user-defined types will now use the ``__module__``
attribute of the type class itself when rendering in an
autogenerated module.

fixes #229

10 years ago- amending d81619b50b9df7ff4458:
Mike Bayer [Mon, 20 Oct 2014 18:24:41 +0000 (14:24 -0400)] 
- amending d81619b50b9df7ff4458:

Revision files are now written out using the ``'wb'`` modifier to
``open()``, since Mako reads the templates with ``'rb'``, thus preventing
CRs from being doubled up as has been observed on windows.  The encoding
of the output now defaults to 'utf-8', which can be configured using
a newly added config file parameter ``output_encoding``.
fixes #234

10 years ago- Revision files are now written out using the ``'wb'`` modifier to
Mike Bayer [Mon, 20 Oct 2014 17:02:32 +0000 (13:02 -0400)] 
- Revision files are now written out using the ``'wb'`` modifier to
``open()``, since Mako reads the templates with ``'rb'``, thus preventing
CRs from being doubled up as has been observed on windows.
fixes #234

10 years ago- fail this before 0.9
Mike Bayer [Sat, 11 Oct 2014 21:30:44 +0000 (17:30 -0400)] 
- fail this before 0.9

10 years agoMerged in ltvolks/alembic/docs-153 (pull request #28)
Mike Bayer [Fri, 10 Oct 2014 19:25:15 +0000 (15:25 -0400)] 
Merged in ltvolks/alembic/docs-153 (pull request #28)

Update documentation for create_index parameters

10 years ago- add tests for unique and quote flag on create_index()
Mike Bayer [Fri, 10 Oct 2014 18:05:09 +0000 (14:05 -0400)] 
- add tests for unique and quote flag on create_index()

10 years agoUpdate documentation for create_index parameters
Lucas Taylor [Fri, 10 Oct 2014 19:21:18 +0000 (14:21 -0500)] 
Update documentation for create_index parameters

10 years ago- add tests for unique and quote flag on create_index()
Mike Bayer [Fri, 10 Oct 2014 18:05:09 +0000 (14:05 -0400)] 
- add tests for unique and quote flag on create_index()

10 years ago- Added support for use of the :class:`~sqlalchemy.sql.elements.quoted_name`
Mike Bayer [Fri, 3 Oct 2014 15:12:19 +0000 (11:12 -0400)] 
- Added support for use of the :class:`~sqlalchemy.sql.elements.quoted_name`
construct when using the ``schema`` argument within operations.  This
allows a name containing a dot to be fully quoted, as well as to
provide configurable quoting on a per-name basis.
fixes #230

10 years ago- add changelog for bugfix re: match + SQLA 0.7
Mike Bayer [Mon, 22 Sep 2014 18:36:44 +0000 (14:36 -0400)] 
- add changelog for bugfix re: match + SQLA 0.7

10 years agoinspect
Mike Bayer [Sat, 20 Sep 2014 23:13:58 +0000 (19:13 -0400)] 
inspect

10 years ago- compat for sqla 0.7
Mike Bayer [Sat, 20 Sep 2014 22:33:00 +0000 (18:33 -0400)] 
- compat for sqla 0.7

10 years agoadd missing util lib
Mike Bayer [Sat, 20 Sep 2014 22:16:17 +0000 (18:16 -0400)] 
add missing util lib

10 years ago- fix this to be a definite non native boolean dialect
Mike Bayer [Sat, 20 Sep 2014 21:25:16 +0000 (17:25 -0400)] 
- fix this to be a definite non native boolean dialect

10 years ago- Added a routine by which the Postgresql Alembic dialect inspects
Mike Bayer [Sat, 20 Sep 2014 21:00:21 +0000 (17:00 -0400)] 
- Added a routine by which the Postgresql Alembic dialect inspects
the server default of INTEGER/BIGINT columns as they are reflected
during autogenerate for the pattern ``nextval(<name>...)`` containing
a potential sequence name, then queries ``pg_catalog`` to see if this
sequence is "owned" by the column being reflected; if so, it assumes
this is a SERIAL or BIGSERIAL column and the server default is
omitted from the column reflection as well as any kind of
server_default comparison or rendering, along with an INFO message
in the logs indicating this has taken place. This allows SERIAL/BIGSERIAL
columns to keep the SEQUENCE from being unnecessarily present within
the autogenerate operation.
fixes #73

10 years ago- changelog for #197, fixes #197.
Mike Bayer [Sat, 20 Sep 2014 19:06:05 +0000 (15:06 -0400)] 
- changelog for #197, fixes #197.

10 years ago- get tests passing for 0.8, 0.7
Mike Bayer [Sat, 20 Sep 2014 18:54:37 +0000 (14:54 -0400)] 
- get tests passing for 0.8, 0.7

10 years ago- move tests erroneously local to the "named" tests
Mike Bayer [Sat, 20 Sep 2014 18:45:10 +0000 (14:45 -0400)] 
- move tests erroneously local to the "named" tests
- add tests for CAST within index

10 years ago- more tweaks for server defaults re: comparison, test results
Mike Bayer [Sat, 20 Sep 2014 18:30:00 +0000 (14:30 -0400)] 
- more tweaks for server defaults re: comparison, test results
- in particular we've changed the server_default for order.amount
to text('0'); SQlite and Postgresql now both report the server
default as '0', whereas previously, when we were emitting '0'
with the quotes in the CREATE TABLE, SQlite reported it as
"'0'" and PG as '0::numeric'.

10 years agopep8
Mike Bayer [Sat, 20 Sep 2014 17:56:13 +0000 (13:56 -0400)] 
pep8

10 years ago- restore _render_server_default_for_compare() to compare.py using
Mike Bayer [Sat, 20 Sep 2014 17:53:11 +0000 (13:53 -0400)] 
- restore _render_server_default_for_compare() to compare.py using
the method previously used in _render_server_default() in render;
these two functions become more specific to compare vs. render now
so are separated out

10 years ago- check SQLA version for drop w/ enums in PG
Mike Bayer [Sat, 20 Sep 2014 17:37:28 +0000 (13:37 -0400)] 
- check SQLA version for drop w/ enums in PG

10 years agoMerge remote-tracking branch 'github/pr/16' into pr16
Mike Bayer [Sat, 20 Sep 2014 17:17:33 +0000 (13:17 -0400)] 
Merge remote-tracking branch 'github/pr/16' into pr16

10 years ago- py3k fix
Mike Bayer [Mon, 15 Sep 2014 01:06:27 +0000 (21:06 -0400)] 
- py3k fix

10 years ago- coverage was not working (and needs to be fixed in SQLAlchemy too),
Mike Bayer [Sun, 14 Sep 2014 19:49:04 +0000 (15:49 -0400)] 
- coverage was not working (and needs to be fixed in SQLAlchemy too),
go back to using a "bootstrap" system where we load in pytestplugin/
noseplugin via file importing, plugin/ is not used as a package.
- identify the pytest_sessionstart hook as the best place to do
plugin_base.post_begin() and actually begin importing main modules;
this is where coverage has actually started.
- we're now targeting 1.0.0 as this has to be ported to SQLAlchemy.
- start using .coveragerc

10 years ago- discover # noqa
Mike Bayer [Sun, 14 Sep 2014 16:04:37 +0000 (12:04 -0400)] 
- discover # noqa

10 years ago- dont do py3k testing for 0.7, 0.8
Mike Bayer [Sun, 14 Sep 2014 15:54:38 +0000 (11:54 -0400)] 
- dont do py3k testing for 0.7, 0.8

10 years agofixes for the nose runner
Mike Bayer [Sun, 14 Sep 2014 15:51:04 +0000 (11:51 -0400)] 
fixes for the nose runner

10 years ago- add plugin directory
Mike Bayer [Sun, 14 Sep 2014 15:45:04 +0000 (11:45 -0400)] 
- add plugin directory

10 years ago- move pretty much all of sqlalchemy.testing over for now, as we'd
Mike Bayer [Sun, 14 Sep 2014 15:37:50 +0000 (11:37 -0400)] 
- move pretty much all of sqlalchemy.testing over for now, as we'd
like to run tests against 0.8 and even late 0.7 versions with the same
capabilities, as well as run parallel testing against all of them.
we need a consistent system to get that all to work, so for now
we have the whole SQLA system shoved into here, not ideal but we have
a very good testing situation for now.  Once we target 0.9.4 at the lowest
we should be able to move all this out.
- re-support 0.7, starting at 0.7.6 which is where things mostly work.
All tests, taking into account known skips and fails which are
added here for 0.7, early 0.8s, pass on 0.7.9.

10 years agochangelog gh/upgrade_testing
Mike Bayer [Sat, 13 Sep 2014 21:10:34 +0000 (17:10 -0400)] 
changelog

10 years ago- bootstrap support
Mike Bayer [Sat, 13 Sep 2014 21:08:22 +0000 (17:08 -0400)] 
- bootstrap support
- some workarounds to get 0.8 mostly working

10 years ago- finish up
Mike Bayer [Sat, 13 Sep 2014 20:53:57 +0000 (16:53 -0400)] 
- finish up

10 years agopostgresql
Mike Bayer [Sat, 13 Sep 2014 20:17:09 +0000 (16:17 -0400)] 
postgresql

10 years agodev
Mike Bayer [Sat, 13 Sep 2014 20:14:39 +0000 (16:14 -0400)] 
dev

10 years agotest command
Mike Bayer [Sat, 13 Sep 2014 19:49:53 +0000 (15:49 -0400)] 
test command

10 years agobulk insert
Mike Bayer [Sat, 13 Sep 2014 19:47:30 +0000 (15:47 -0400)] 
bulk insert

10 years ago- move towards sqlalchemy test base. autogenerate tests so far
Mike Bayer [Sat, 13 Sep 2014 19:38:53 +0000 (15:38 -0400)] 
- move towards sqlalchemy test base.  autogenerate tests so far

10 years ago- updates
Mike Bayer [Sat, 13 Sep 2014 18:14:21 +0000 (14:14 -0400)] 
- updates

10 years ago- start using tox 1.8 so that we can easily get multiple sqla versions off the ground
Mike Bayer [Sat, 13 Sep 2014 16:12:57 +0000 (12:12 -0400)] 
- start using tox 1.8 so that we can easily get multiple sqla versions off the ground

10 years ago0.7 start
Mike Bayer [Sat, 13 Sep 2014 16:12:36 +0000 (12:12 -0400)] 
0.7 start

10 years ago0.6.7 rel_0_6 gh/rel_0_6 rel_0_6_7
Mike Bayer [Tue, 9 Sep 2014 21:50:09 +0000 (17:50 -0400)] 
0.6.7

10 years ago- update naming convention link to point to official docs gh/cleanup
Mike Bayer [Tue, 9 Sep 2014 20:54:25 +0000 (16:54 -0400)] 
- update naming convention link to point to official docs

10 years ago- finish flake8 on tests
Mike Bayer [Tue, 9 Sep 2014 20:51:29 +0000 (16:51 -0400)] 
- finish flake8 on tests

10 years ago- fix long lines within package
Mike Bayer [Tue, 9 Sep 2014 20:35:40 +0000 (16:35 -0400)] 
- fix long lines within package

10 years ago- do an autopep8 pass for just about everything other
Mike Bayer [Tue, 9 Sep 2014 16:24:43 +0000 (12:24 -0400)] 
- do an autopep8 pass for just about everything other
than line length

10 years ago- changelog for pr bitbucket:26
Mike Bayer [Tue, 9 Sep 2014 16:13:48 +0000 (12:13 -0400)] 
- changelog for pr bitbucket:26

10 years agoMerge https://bitbucket.org/utek/alembic/branch/master into pr26
Mike Bayer [Tue, 9 Sep 2014 16:10:39 +0000 (12:10 -0400)] 
Merge https://bitbucket.org/utek/alembic/branch/master into pr26

10 years ago- add a tox.ini file
Mike Bayer [Tue, 9 Sep 2014 16:10:10 +0000 (12:10 -0400)] 
- add a tox.ini file
- update the readme.unittests a bit

10 years agomore test fixes
Łukasz Bołdys [Tue, 9 Sep 2014 09:43:05 +0000 (11:43 +0200)] 
more test fixes

10 years agofixed test
Łukasz Bołdys [Tue, 9 Sep 2014 08:16:06 +0000 (10:16 +0200)] 
fixed test

10 years agoMS SQL is using sp_rename instead of Alter table to rename table
Łukasz Bołdys [Tue, 9 Sep 2014 08:12:12 +0000 (10:12 +0200)] 
MS SQL is using sp_rename instead of Alter table to rename table

10 years agoapplied mikes patch
Carlos Eduardo Rivera [Mon, 8 Sep 2014 22:43:08 +0000 (17:43 -0500)] 
applied mikes patch

10 years agoreplacing eval in multidb template with globals() function lookup
Andreas Grapentin [Thu, 28 Aug 2014 19:53:37 +0000 (21:53 +0200)] 
replacing eval in multidb template with globals() function lookup

10 years agoMerged in andreasgrapentin/alembic/template_eval (pull request #25)
Mike Bayer [Thu, 28 Aug 2014 20:00:10 +0000 (16:00 -0400)] 
Merged in andreasgrapentin/alembic/template_eval (pull request #25)

replacing eval in multidb template with globals() function lookup

10 years agoreplacing eval in multidb template with globals() function lookup
Andreas Grapentin [Thu, 28 Aug 2014 19:53:37 +0000 (21:53 +0200)] 
replacing eval in multidb template with globals() function lookup

10 years ago- Added support for functional indexes when using the
Mike Bayer [Wed, 27 Aug 2014 23:22:42 +0000 (19:22 -0400)] 
- Added support for functional indexes when using the
:meth:`.Operations.create_index` directive.   Within the list of columns,
the SQLAlchemy ``text()`` construct can be sent, embedding a literal
SQL expression; the :meth:`.Operations.create_index` will perform some hackery
behind the scenes to get the :class:`.Index` construct to cooperate.
This works around some current limitations in :class:`.Index`
which should be resolved on the SQLAlchemy side at some point.
fixes #222

10 years ago- Added support for functional indexes when using the
Mike Bayer [Wed, 27 Aug 2014 23:22:42 +0000 (19:22 -0400)] 
- Added support for functional indexes when using the
:meth:`.Operations.create_index` directive.   Within the list of columns,
the SQLAlchemy ``text()`` construct can be sent, embedding a literal
SQL expression; the :meth:`.Operations.create_index` will perform some hackery
behind the scenes to get the :class:`.Index` construct to cooperate.
This works around some current limitations in :class:`.Index`
which should be resolved on the SQLAlchemy side at some point.
fixes #222

10 years agoRendering expressions for indexes
Carlos Eduardo Rivera [Wed, 27 Aug 2014 19:06:22 +0000 (14:06 -0500)] 
Rendering expressions for indexes

10 years agoadding mike's patch
Carlos Eduardo Rivera [Wed, 27 Aug 2014 10:35:47 +0000 (05:35 -0500)] 
adding mike's patch

10 years agomoving textclause logic into _render_server_default
Carlos Eduardo Rivera [Mon, 25 Aug 2014 21:21:41 +0000 (16:21 -0500)] 
moving textclause logic into _render_server_default

10 years agoAdding TextClause support to server_default
Carlos Eduardo Rivera [Mon, 25 Aug 2014 19:06:45 +0000 (14:06 -0500)] 
Adding TextClause support to server_default

11 years ago0.6.6 rel_0_6_6
Mike Bayer [Thu, 7 Aug 2014 20:14:56 +0000 (16:14 -0400)] 
0.6.6

11 years ago- ignore __init__.py in versions directory, fixes #95.
Mike Bayer [Thu, 24 Jul 2014 22:48:17 +0000 (18:48 -0400)] 
- ignore __init__.py in versions directory, fixes #95.

11 years ago- changelog for pullreq bitbucket:23
Mike Bayer [Thu, 7 Aug 2014 18:26:23 +0000 (14:26 -0400)] 
- changelog for pullreq bitbucket:23

11 years agoMerged in paradoxxxzero/alembic/paradoxxxzero/fix-sorted-of-none-values-in-compare...
Mike Bayer [Thu, 7 Aug 2014 18:22:31 +0000 (14:22 -0400)] 
Merged in paradoxxxzero/alembic/paradoxxxzero/fix-sorted-of-none-values-in-compare-for-1403597500844 (pull request #23)

Fix sorted of None values in compare for python 3

11 years agoSupport autogenerating migrations for tables with 255 or more columns
Ryan P. Kelly [Thu, 5 Jun 2014 01:29:47 +0000 (21:29 -0400)] 
Support autogenerating migrations for tables with 255 or more columns

11 years agochangelog for pr github:14
Mike Bayer [Thu, 7 Aug 2014 18:03:17 +0000 (14:03 -0400)] 
changelog for pr github:14

11 years agoMerge remote-tracking branch 'github/pr/14' into pr14
Mike Bayer [Thu, 7 Aug 2014 18:00:26 +0000 (14:00 -0400)] 
Merge remote-tracking branch 'github/pr/14' into pr14

11 years agoAlter autogenerate tests to show python3 problem on already existing tables sort
Florian Mounier [Wed, 25 Jun 2014 13:32:40 +0000 (15:32 +0200)] 
Alter autogenerate tests to show python3 problem on already existing tables sort

11 years agothis is 0.6.6
Mike Bayer [Tue, 24 Jun 2014 18:26:14 +0000 (14:26 -0400)] 
this is 0.6.6

11 years agoFix sorted of None values in compare for python 3
paradoxxxzero [Tue, 24 Jun 2014 08:13:05 +0000 (08:13 +0000)] 
Fix sorted of None values in compare for python 3

11 years ago- Some deep-in-the-weeds fixes to try to get "server default" comparison
Mike Bayer [Mon, 23 Jun 2014 17:49:44 +0000 (13:49 -0400)] 
- Some deep-in-the-weeds fixes to try to get "server default" comparison
working better across platforms and expressions, in particular on
the Postgresql backend, mostly dealing with quoting/not quoting of various
expressions at the appropriate time and on a per-backend basis.
Repaired and tested support for such defaults as Postgresql interval
and array defaults.
fixes #212