]> git.ipfire.org Git - thirdparty/sqlalchemy/sqlalchemy.git/log
thirdparty/sqlalchemy/sqlalchemy.git
13 years ago- do a straight __subclasses__ traversal here, so that we aren't
Mike Bayer [Sun, 28 Oct 2012 23:07:56 +0000 (19:07 -0400)] 
- do a straight __subclasses__ traversal here, so that we aren't
iterating through all mappers in memory when a new event is tacked
onto an unmapped superclass, also removes the strong ref that was
breaking some GC teardown tests

13 years ago- remove remote_foreign annotation
Mike Bayer [Sun, 28 Oct 2012 22:23:57 +0000 (18:23 -0400)] 
- remove remote_foreign annotation
- support annotations on Column where name isn't immediately present

13 years ago- updated relationship docs, others
Mike Bayer [Sun, 28 Oct 2012 22:23:24 +0000 (18:23 -0400)] 
- updated relationship docs, others

13 years ago- add a missing reference handler to handle the inheriting names that aren't in the...
Mike Bayer [Sun, 28 Oct 2012 22:21:53 +0000 (18:21 -0400)] 
- add a missing reference handler to handle the inheriting names that aren't in the docs

13 years agomore migration notes
Mike Bayer [Sun, 28 Oct 2012 18:16:50 +0000 (14:16 -0400)] 
more migration notes

13 years ago- store only MultipleClassMarkers inside of ModuleMarker, then
Mike Bayer [Sun, 28 Oct 2012 18:15:56 +0000 (14:15 -0400)] 
- store only MultipleClassMarkers inside of ModuleMarker, then
store ModuleMarkers for multiple roots, one for each token in a
module path.  this allows partial path resolution.
- docs to this effect

13 years ago- cover additional cases for [ticket:2585], where events are applied to base class
Mike Bayer [Sun, 28 Oct 2012 16:55:36 +0000 (12:55 -0400)] 
- cover additional cases for [ticket:2585], where events are applied to base class
after subclasses are already mapped

13 years agoglossary updates
Mike Bayer [Sun, 28 Oct 2012 16:27:32 +0000 (12:27 -0400)] 
glossary updates

13 years agotypo
Mike Bayer [Sun, 28 Oct 2012 03:16:50 +0000 (23:16 -0400)] 
typo

13 years ago- more docs
Mike Bayer [Sat, 27 Oct 2012 00:59:21 +0000 (20:59 -0400)] 
- more docs
- it appears we can get rid of all those "XYZ_toplevel" names and use :doc:.

13 years ago- add class_ to AliasedInsp
Mike Bayer [Fri, 26 Oct 2012 23:25:23 +0000 (19:25 -0400)] 
- add class_ to AliasedInsp
- redefine inspect(Class.attrname).parent to be always an inspectable
target; either Mapper or AliasedInsp
- add most major features to 08 migration, document, link

13 years ago- some naming changes on PropComparator, Comparator:
Mike Bayer [Fri, 26 Oct 2012 00:28:03 +0000 (20:28 -0400)] 
- some naming changes on PropComparator, Comparator:

  1. all Comparators now have "parent" which is always the parent mapper
  or AliasedClass instance

  2. only RelationshipProperty.Comparator has "mapper" now, which
  is the target mapper

  3. The names "parententity" and "parentmapper" are underscored

also improved the message with the "neither comparator nor instruentedattribute...."
to include the classname + attribute name

13 years agoupdates
Mike Bayer [Thu, 25 Oct 2012 23:54:04 +0000 (19:54 -0400)] 
updates

13 years agoversionadded
Mike Bayer [Thu, 25 Oct 2012 20:02:11 +0000 (16:02 -0400)] 
versionadded

13 years agoremove some whitespace
Mike Bayer [Thu, 25 Oct 2012 19:58:35 +0000 (15:58 -0400)] 
remove some whitespace

13 years ago0.7 updates, links
Mike Bayer [Thu, 25 Oct 2012 19:49:08 +0000 (15:49 -0400)] 
0.7 updates, links

13 years agoadd changelogs
Mike Bayer [Thu, 25 Oct 2012 19:20:36 +0000 (15:20 -0400)] 
add changelogs

13 years ago The long-deprecated and non-functional ``assert_unicode`` flag on
Mike Bayer [Thu, 25 Oct 2012 16:07:38 +0000 (12:07 -0400)] 
    The long-deprecated and non-functional ``assert_unicode`` flag on
        :func:`.create_engine` as well as :class:`.String` is removed.

13 years agocallcounts
Mike Bayer [Thu, 25 Oct 2012 05:08:37 +0000 (01:08 -0400)] 
callcounts

13 years agoAdded :meth:`.ColumnOperators.notin_`,
Mike Bayer [Wed, 24 Oct 2012 20:01:17 +0000 (16:01 -0400)] 
Added :meth:`.ColumnOperators.notin_`,
:meth:`.ColumnOperators.notlike`,
:meth:`.ColumnOperators.notilike` to :class:`.ColumnOperators`.
[ticket:2580]

13 years agoversion note
Mike Bayer [Wed, 24 Oct 2012 19:39:39 +0000 (15:39 -0400)] 
version note

13 years agoFixed bug where keyword arguments passed to
Mike Bayer [Wed, 24 Oct 2012 19:37:06 +0000 (15:37 -0400)] 
Fixed bug where keyword arguments passed to
:meth:`.Compiler.process` wouldn't get propagated
to the column expressions present in the columns
clause of a SELECT statement.  In particular this would
come up when used by custom compilation schemes that
relied upon special flags. [ticket:2593]

13 years ago- move most/all operator specific tests into test_operator, convert fully to TOT
Mike Bayer [Wed, 24 Oct 2012 18:29:52 +0000 (14:29 -0400)] 
- move most/all operator specific tests into test_operator, convert fully to TOT

13 years ago- get 100% lint/pep8 happening for test_compiler; next we will begin
Mike Bayer [Wed, 24 Oct 2012 18:02:37 +0000 (14:02 -0400)] 
- get 100% lint/pep8 happening for test_compiler; next we will begin
cutting up tests and removing old ones
- move test_in() to test_operators
- slice up migrated operator tests into TOT

13 years agoaaaarrrgg
Mike Bayer [Tue, 23 Oct 2012 20:02:00 +0000 (16:02 -0400)] 
aaaarrrgg

13 years agotweaks
Mike Bayer [Tue, 23 Oct 2012 18:22:56 +0000 (14:22 -0400)] 
tweaks

13 years ago Added a new method :meth:`.Engine.execution_options`
Mike Bayer [Tue, 23 Oct 2012 16:08:20 +0000 (12:08 -0400)] 
  Added a new method :meth:`.Engine.execution_options`
      to :class:`.Engine`.  This method works similarly to
      :class:`.Connection.execution_options` in that it creates
      a copy of the parent object which will refer to the new
      set of options.   The method can be used to build
      sharding schemes where each engine shares the same
      underlying pool of connections.   The method
      has been tested against the horizontal shard
      recipe in the ORM as well.

13 years ago- fix regression from 0.7 where calling get_history with passive
Mike Bayer [Mon, 22 Oct 2012 17:54:39 +0000 (13:54 -0400)] 
- fix regression from 0.7 where calling get_history with passive
on a never-set collection would fail; made this act just like
scalars for now and added tests.  I would think that HISTORY_BLANK
would be more appropriate here but it's too late in the game
to mess with that.

13 years ago- support multiple top-level cats in a changeset
Mike Bayer [Mon, 22 Oct 2012 17:41:56 +0000 (13:41 -0400)] 
- support multiple top-level cats in a changeset
- include ticket in the md5

13 years ago The auto-correlation feature of :func:`.select`, and
Mike Bayer [Mon, 22 Oct 2012 17:29:12 +0000 (13:29 -0400)] 
 The auto-correlation feature of :func:`.select`, and
      by proxy that of :class:`.orm.Query`, will not
      take effect for a SELECT statement that is being
      rendered directly in the FROM list of the enclosing
      SELECT.  Correlation in SQL only applies to column
      expressions such as those in the WHERE, ORDER BY,
      columns clause. [ticket:2595]

13 years ago- add md5-hashed links to changelog entries
Mike Bayer [Sun, 21 Oct 2012 22:36:17 +0000 (18:36 -0400)] 
- add md5-hashed links to changelog entries

13 years agoadd support to link to changelogs and pullrequests
Mike Bayer [Sun, 21 Oct 2012 21:26:10 +0000 (17:26 -0400)] 
add support to link to changelogs and pullrequests

13 years agochangelog
Mike Bayer [Sun, 21 Oct 2012 21:00:36 +0000 (17:00 -0400)] 
changelog

13 years agobranch merge
Mike Bayer [Sun, 21 Oct 2012 20:57:21 +0000 (16:57 -0400)] 
branch merge

13 years agoMerged in agilevic/sqlalchemy (pull request #23)
Mike Bayer [Sun, 21 Oct 2012 20:56:33 +0000 (16:56 -0400)] 
Merged in agilevic/sqlalchemy (pull request #23)

13 years ago- converted beaker demo to dogpile.cache, [ticket:2589]
Mike Bayer [Sun, 21 Oct 2012 20:54:42 +0000 (16:54 -0400)] 
- converted beaker demo to dogpile.cache, [ticket:2589]

13 years agoupdated changelog thing
Mike Bayer [Sun, 21 Oct 2012 01:24:00 +0000 (21:24 -0400)] 
updated changelog thing

13 years ago- genericize the changelog plugin a bit, lets see if it works in other projects
Mike Bayer [Sun, 21 Oct 2012 00:41:44 +0000 (20:41 -0400)] 
- genericize the changelog plugin a bit, lets see if it works in other projects

13 years agolink fix
Mike Bayer [Sat, 20 Oct 2012 23:01:22 +0000 (19:01 -0400)] 
link fix

13 years agofix missing/incorrect version ids
Mike Bayer [Sat, 20 Oct 2012 22:57:52 +0000 (18:57 -0400)] 
fix missing/incorrect version ids

13 years agoupdates
Mike Bayer [Sat, 20 Oct 2012 22:35:19 +0000 (18:35 -0400)] 
updates

13 years ago- move all CHANGES files into the sphinx docs. this is rudimental now
Mike Bayer [Sat, 20 Oct 2012 22:28:02 +0000 (18:28 -0400)] 
- move all CHANGES files into the sphinx docs.    this is rudimental now
but puts the changelog in a more data driven format, so that we
can begin linking version directives with them as well
as with migration documents, once those are also moved into sphinx.

13 years ago- clarify what the link is
Mike Bayer [Fri, 19 Oct 2012 23:37:48 +0000 (19:37 -0400)] 
- clarify what the link is

13 years ago- rework the sphinx customizations into distinct modules
Mike Bayer [Fri, 19 Oct 2012 23:20:18 +0000 (19:20 -0400)] 
- rework the sphinx customizations into distinct modules
- build a new Sphinx extension that allows dialect info
to be entered as directives which is then rendered consistently
throughout all dialect/dbapi sections
- break out the "empty_strings" requirement for oracle test

13 years ago- errant maxdb
Mike Bayer [Fri, 19 Oct 2012 14:26:28 +0000 (10:26 -0400)] 
- errant maxdb
- errant pdbs

13 years ago- move out maxdb
Mike Bayer [Thu, 18 Oct 2012 22:24:15 +0000 (18:24 -0400)] 
- move out maxdb
- begin consolidating docs for dialects to be more self contained
- add a separate section for "external" dialects
- not sure how we're going to go with this yet.

13 years ago - [bug] TypeDecorator now includes a generic repr()
Mike Bayer [Thu, 18 Oct 2012 16:29:41 +0000 (12:29 -0400)] 
  - [bug] TypeDecorator now includes a generic repr()
    that works in terms of the "impl" type by default.
    This is a behavioral change for those TypeDecorator
    classes that specify a custom __init__ method; those
    types will need to re-define __repr__() if they need
    __repr__() to provide a faithful constructor representation.
    [ticket:2594]

13 years ago - [feature] The Query can now load entity/scalar-mixed
Mike Bayer [Tue, 16 Oct 2012 00:07:13 +0000 (20:07 -0400)] 
  - [feature] The Query can now load entity/scalar-mixed
    "tuple" rows that contain
    types which aren't hashable, by setting the flag
    "hashable=False" on the corresponding TypeEngine object
    in use.  Custom types that return unhashable types
    (typically lists) can set this flag to False.
    [ticket:2592]
  - [bug] Applying a column expression to a select
    statement using a label with or without other
    modifying constructs will no longer "target" that
    expression to the underlying Column; this affects
    ORM operations that rely upon Column targeting
    in order to retrieve results.  That is, a query
    like query(User.id, User.id.label('foo')) will now
    track the value of each "User.id" expression separately
    instead of munging them together.  It is not expected
    that any users will be impacted by this; however,
    a usage that uses select() in conjunction with
    query.from_statement() and attempts to load fully
    composed ORM entities may not function as expected
    if the select() named Column objects with arbitrary
    .label() names, as these will no longer target to
    the Column objects mapped by that entity.
    [ticket:2591]

13 years ago- allow a __clause_element__() to be passed to query.filter() also
Mike Bayer [Mon, 15 Oct 2012 22:13:33 +0000 (18:13 -0400)] 
- allow a __clause_element__() to be passed to query.filter() also

13 years ago- break out unicode tests to a varchar + a text test, add a requires
Mike Bayer [Mon, 15 Oct 2012 21:36:17 +0000 (17:36 -0400)] 
- break out unicode tests to a varchar + a text test, add a requires

13 years ago - [feature] "scalar" selects now have a WHERE method
Mike Bayer [Mon, 15 Oct 2012 21:21:38 +0000 (17:21 -0400)] 
  - [feature] "scalar" selects now have a WHERE method
    to help with generative building.  Also slight adjustment
    regarding how SS "correlates" columns; the new methodology
    no longer applies meaning to the underlying
    Table column being selected.  This improves
    some fairly esoteric situations, and the logic
    that was there didn't seem to have any purpose.
  - [feature] Some support for auto-rendering of a
    relationship join condition based on the mapped
    attribute, with usage of core SQL constructs.
    E.g. select([SomeClass]).where(SomeClass.somerelationship)
    would render SELECT from "someclass" and use the
    primaryjoin of "somerelationship" as the WHERE
    clause.   This changes the previous meaning
    of "SomeClass.somerelationship" when used in a
    core SQL context; previously, it would "resolve"
    to the parent selectable, which wasn't generally
    useful.  Related to [ticket:2245].

13 years agofix for pypy/jython gc_collect
Mike Bayer [Mon, 15 Oct 2012 18:10:09 +0000 (14:10 -0400)] 
fix for pypy/jython gc_collect

13 years ago- lets default propagate to True here since it was essentially always
Mike Bayer [Fri, 12 Oct 2012 21:28:25 +0000 (17:28 -0400)] 
- lets default propagate to True here since it was essentially always
propagated previously

13 years ago - [feature] Improvements to event listening for
Mike Bayer [Fri, 12 Oct 2012 21:21:08 +0000 (17:21 -0400)] 
  - [feature] Improvements to event listening for
    mapped classes allows that unmapped classes
    can be specified for instance- and mapper-events.
    The established events will be automatically
    set up on subclasses of that class when the
    propagate=True flag is passed, and the
    events will be set up for that class itself
    if and when it is ultimately mapped.
    [ticket:2585]

  - [bug] The instrumentation events class_instrument(),
    class_uninstrument(), and attribute_instrument()
    will now fire off only for descendant classes
    of the class assigned to listen().  Previously,
    an event listener would be assigned to listen
    for all classes in all cases regardless of the
    "target" argument passed.  [ticket:2590]

13 years ago- add some UPDATE..FROM tests and also some new reqs
Mike Bayer [Fri, 12 Oct 2012 17:20:56 +0000 (13:20 -0400)] 
- add some UPDATE..FROM tests and also some new reqs

13 years ago- start adding ORM-level tests for plain UPDATE..FROM scenarios.
Mike Bayer [Fri, 12 Oct 2012 16:17:03 +0000 (12:17 -0400)] 
- start adding ORM-level tests for plain UPDATE..FROM scenarios.

13 years agoadd beaker_data to hgignore..
Mike Bayer [Thu, 11 Oct 2012 21:46:17 +0000 (17:46 -0400)] 
add beaker_data to hgignore..

13 years ago- add support for integer indexes and slices to PathRegistry
Mike Bayer [Thu, 11 Oct 2012 21:44:16 +0000 (17:44 -0400)] 
- add support for integer indexes and slices to PathRegistry
- write a full test suite for PathRegistry

13 years ago- add coverage for merge_result() [ticket:2588]
Mike Bayer [Thu, 11 Oct 2012 20:59:02 +0000 (16:59 -0400)] 
- add coverage for merge_result() [ticket:2588]
- pre-determine keys for the keyed tuples

13 years ago- keep dialect at default for type compile
Mike Bayer [Thu, 11 Oct 2012 15:15:29 +0000 (11:15 -0400)] 
- keep dialect at default for type compile
- python2.5/3 fixes

13 years agopython 3 division
Mike Bayer [Thu, 11 Oct 2012 00:19:29 +0000 (20:19 -0400)] 
python 3 division

13 years ago- create new sqlalchemy.testing.suite.test_types, starting with Unicode round trip...
Mike Bayer [Thu, 11 Oct 2012 00:12:25 +0000 (20:12 -0400)] 
- create new sqlalchemy.testing.suite.test_types, starting with Unicode round trip tests.
- clean out existing test_types:UnicodeTest to be strictly individual unit tests
against Unicode(), String(), etc. with no database access.

13 years ago - [feature] Added "collation" parameter to all
Mike Bayer [Wed, 10 Oct 2012 23:34:29 +0000 (19:34 -0400)] 
  - [feature] Added "collation" parameter to all
    String types.  When present, renders as
    COLLATE <collation>.  This to support the
    COLLATE keyword now supported by several
    databases including MySQL, SQLite, and Postgresql.
    [ticket:2276]

  - [change] The Text() type renders the length
    given to it, if a length was specified.

13 years ago - [feature] Added TIME type to mysql dialect,
Mike Bayer [Wed, 10 Oct 2012 18:25:21 +0000 (14:25 -0400)] 
  - [feature] Added TIME type to mysql dialect,
    accepts "fst" argument which is the new
    "fractional seconds" specifier for recent
    MySQL versions.  The datatype will interpret
    a microseconds portion received from the driver,
    however note that at this time most/all MySQL
    DBAPIs do not support returning this value.
    [ticket:2534]
- attempted to modernize the types tests in test_mysql a little, though has a long
way to go

13 years ago- changelog update, added 0.7's version of [ticket:2561] and also
Mike Bayer [Wed, 10 Oct 2012 17:26:00 +0000 (13:26 -0400)] 
- changelog update, added 0.7's version of [ticket:2561] and also
r74618cabb1cb referred to the wrong ticket.

13 years ago - [feature] Added "read_timeout" flag to MySQLdb
Mike Bayer [Wed, 10 Oct 2012 17:15:16 +0000 (13:15 -0400)] 
  - [feature] Added "read_timeout" flag to MySQLdb
    dialect.  [ticket:2554]

13 years ago - [feature] Added "raise_on_warnings" flag to OurSQL
Mike Bayer [Wed, 10 Oct 2012 17:13:41 +0000 (13:13 -0400)] 
  - [feature] Added "raise_on_warnings" flag to OurSQL
    dialect.  [ticket:2523]

13 years agoupdates
Mike Bayer [Wed, 10 Oct 2012 05:41:01 +0000 (01:41 -0400)] 
updates

13 years agothis is an RST now
Mike Bayer [Wed, 10 Oct 2012 05:25:14 +0000 (01:25 -0400)] 
this is an RST now

13 years ago- new dialect development README
Mike Bayer [Wed, 10 Oct 2012 05:24:58 +0000 (01:24 -0400)] 
- new dialect development README

13 years ago- remove old hardcoded rule about chained being limited to certain load types
Mike Bayer [Mon, 8 Oct 2012 21:42:20 +0000 (17:42 -0400)] 
- remove old hardcoded rule about chained being limited to certain load types

13 years agomissing with_statement
Mike Bayer [Mon, 8 Oct 2012 18:54:54 +0000 (14:54 -0400)] 
missing with_statement

13 years ago- add a nifty contextmanager form of assert_execution/assert_statement_count
Mike Bayer [Mon, 8 Oct 2012 18:53:13 +0000 (14:53 -0400)] 
- add a nifty contextmanager form of assert_execution/assert_statement_count

13 years ago - [feature] Various API tweaks to the "dialect"
Mike Bayer [Mon, 8 Oct 2012 18:41:35 +0000 (14:41 -0400)] 
  - [feature] Various API tweaks to the "dialect"
    API to better support highly specialized
    systems such as the Akiban database, including
    more hooks to allow an execution context to
    access type processors.

13 years ago - [bug] Continuing [ticket:2566] regarding extra
Mike Bayer [Sat, 6 Oct 2012 16:46:02 +0000 (12:46 -0400)] 
  - [bug] Continuing [ticket:2566] regarding extra
    state post-flush due to event listeners;
    any states that are marked as "dirty" from an
    attribute perspective, usually via column-attribute
    set events within after_insert(), after_update(),
    etc., will get the "history" flag reset
    in all cases, instead of only those instances
    that were part of the flush.  This has the effect
    that this "dirty" state doesn't carry over
    after the flush and won't result in UPDATE
    statements.   A warning is emitted to this
    effect; the set_committed_state()
    method can be used to assign attributes on objects
    without producing history events. [ticket:2582]

13 years ago- break session tests into smaller suites. still some ambiguity what
Mike Bayer [Sat, 6 Oct 2012 16:10:01 +0000 (12:10 -0400)] 
- break session tests into smaller suites.  still some ambiguity what
some of these tests are testing

13 years agoit is sensible to add NCHAR and NVARCHAR as recognized data types for sqlite dialect...
Victor Olex [Fri, 5 Oct 2012 20:31:20 +0000 (13:31 -0700)] 
it is sensible to add NCHAR and NVARCHAR as recognized data types for sqlite dialect because they are widely used and because of the type affinity mechanism of sqlite

13 years ago- add back __engine_options__
Mike Bayer [Thu, 4 Oct 2012 22:26:55 +0000 (18:26 -0400)] 
- add back __engine_options__
- break test_insert tests into explicitly get_lastrowid() vs. implicit_returning tests,
fix up requirements to split them out

13 years ago- connect immediately on _create_testing_engine to initialize dialect capabilities
Mike Bayer [Thu, 4 Oct 2012 21:37:44 +0000 (17:37 -0400)] 
- connect immediately on _create_testing_engine to initialize dialect capabilities

13 years ago- don't call get_lastrowid() on explicit returning
Mike Bayer [Thu, 4 Oct 2012 21:36:02 +0000 (17:36 -0400)] 
- don't call get_lastrowid() on explicit returning
- don't hardwire "subqueries" requirement in the base, mysql < 4.1 isn't working anyway
- don't need explicit FB/PG exclusions in test_returning
- hit db.connect() for the returning requirement

13 years ago - [feature] The Session will produce warnings
Mike Bayer [Thu, 4 Oct 2012 17:50:36 +0000 (13:50 -0400)] 
  - [feature] The Session will produce warnings
    when unsupported methods are used inside the
    "execute" portion of the flush.   These are
    the familiar methods add(), delete(), etc.
    as well as collection and related-object
    manipulations, as called within mapper-level
    flush events
    like after_insert(), after_update(), etc.
    It's been prominently documented for a long
    time that  SQLAlchemy cannot guarantee
    results when the Session is manipulated within
    the execution of the flush plan,
    however users are still doing it, so now
    there's a warning.   Maybe someday the Session
    will be enhanced to support these operations
    inside of the flush, but for now, results
    can't be guaranteed.

13 years ago- fix declarative __init__
Mike Bayer [Thu, 4 Oct 2012 14:32:14 +0000 (10:32 -0400)] 
- fix declarative __init__

13 years ago- test updates
Mike Bayer [Thu, 4 Oct 2012 14:28:21 +0000 (10:28 -0400)] 
- test updates

13 years ago- requirements
Mike Bayer [Wed, 3 Oct 2012 19:41:34 +0000 (15:41 -0400)] 
- requirements

13 years ago- add one more #2583 test to cover the "multiple PK switch" use case
Mike Bayer [Wed, 3 Oct 2012 15:10:42 +0000 (11:10 -0400)] 
- add one more #2583 test to cover the "multiple PK switch" use case

13 years ago - [bug] Fixed Session accounting bug whereby replacing
Mike Bayer [Wed, 3 Oct 2012 14:52:14 +0000 (10:52 -0400)] 
  - [bug] Fixed Session accounting bug whereby replacing
    a deleted object in the identity map with another
    object of the same primary key would raise a
   "conflicting state" error on rollback(),
    if the replaced primary key were established either
    via non-unitofwork-established INSERT statement
    or by primary key switch of another instance.
    [ticket:2583]

13 years ago- fix the fixture here that wasn't creating consistently
Mike Bayer [Mon, 1 Oct 2012 05:59:59 +0000 (01:59 -0400)] 
- fix the fixture here that wasn't creating consistently
- rewrite --dropfirst to be more industrial strength, includes views
- fix order_by="foreign_key" to maintain the same ordering as
metadata.sorted_tables.   Not ideal that this was the other way throughout
0.7 but this is still a little-used method, in contrast to metadata.sorted_tables.

13 years agoremove access
Mike Bayer [Mon, 1 Oct 2012 00:32:38 +0000 (20:32 -0400)] 
remove access

13 years ago - [feature] the MS Access dialect has been
Mike Bayer [Mon, 1 Oct 2012 00:17:49 +0000 (20:17 -0400)] 
  - [feature] the MS Access dialect has been
    moved to its own project on Bitbucket,
    taking advantage of the new SQLAlchemy
    dialect compliance suite.   The dialect is
    still in very rough shape and probably not
    ready for general use yet, however
    it does have *extremely* rudimental
    functionality now.

13 years agoallow variability here
Mike Bayer [Sun, 30 Sep 2012 23:59:40 +0000 (19:59 -0400)] 
allow variability here

13 years agomssql: - [bug] Fixed bug where reflection of primary key constraint
Mike Bayer [Sun, 30 Sep 2012 22:23:21 +0000 (18:23 -0400)] 
mssql:   - [bug] Fixed bug where reflection of primary key constraint
    would double up columns if the same constraint/table
    existed in multiple schemas.
- force returns_rows to False for inserts where we know rows shouldnt be returned;
allows post_exec() to use the cursor without issue

13 years agospacing
Mike Bayer [Sun, 30 Sep 2012 21:20:15 +0000 (17:20 -0400)] 
spacing

13 years ago - [bug] Fixed bug in over() construct whereby
Mike Bayer [Sun, 30 Sep 2012 21:18:14 +0000 (17:18 -0400)] 
  - [bug] Fixed bug in over() construct whereby
    passing an empty list for either partition_by
    or order_by, as opposed to None, would fail
    to generate correctly.
    Courtesy Gunnlaugur Por Briem.
    [ticket:2574]

13 years agoput this inside a main()
Mike Bayer [Sun, 30 Sep 2012 21:05:00 +0000 (17:05 -0400)] 
put this inside a main()

13 years ago - [bug] Fixed compiler bug whereby using a correlated
Mike Bayer [Sun, 30 Sep 2012 20:57:23 +0000 (16:57 -0400)] 
  - [bug] Fixed compiler bug whereby using a correlated
    subquery within an ORDER BY would fail to render correctly
    if the stament also used LIMIT/OFFSET, due to mis-rendering
    within the ROW_NUMBER() OVER clause.  Fix courtesy
    sayap [ticket:2538]

13 years ago - [bug] The CreateIndex construct in Oracle
Mike Bayer [Sun, 30 Sep 2012 20:00:46 +0000 (20:00 +0000)] 
  - [bug] The CreateIndex construct in Oracle
    will now schema-qualify the name of the index
    to be that of the parent table.  Previously this
    name was omitted which apparently creates the
    index in the default schema, rather than that
    of the table.

13 years ago- add a glossary b.c. hey its a good idea
Mike Bayer [Sun, 30 Sep 2012 15:54:40 +0000 (11:54 -0400)] 
- add a glossary b.c. hey its a good idea

13 years agomodernize session links
Mike Bayer [Sun, 30 Sep 2012 15:29:54 +0000 (11:29 -0400)] 
modernize session links

13 years ago- commit Priit Laes docstring fixes
Mike Bayer [Sun, 30 Sep 2012 15:13:52 +0000 (11:13 -0400)] 
- commit Priit Laes docstring fixes
- don't even talk about metadata.bind in declarative

13 years agofix the path here
Mike Bayer [Sun, 30 Sep 2012 15:03:27 +0000 (11:03 -0400)] 
fix the path here