- Trotted out util.pending_deprecation, replacing some 'TODO: deprecate's
- Big session docstring content edit fiesta
- session.py line length and whitespace non-fiesta
- @unsupported now only accepts a single target and demands a reason
for not running the test.
- @exclude also demands an exclusion reason
- Greatly expanded @testing.requires.<feature>, eliminating many
decorators in the suite and signficantly easing integration of
multi-driver support.
- New ORM test base class, and a featureful base for mapped tests
- Usage of 'global' for shared setup going away, * imports as well
- removes everything to do with select_table, which remains as a keyword argument synonymous with
with_polymorphic=('*', select_table).
- all "polymorphic" selectables find their way to Query by way of _set_select_from() now, so that
all joins/aliasing/eager loads/etc. is handled consistently. Mapper has methods for producing
polymorphic selectables so that Query and eagerloaders alike can get to them.
- row aliasing simplified, so that they don't need to nest. they only need the source selectable
and adapt to whatever incoming columns they get.
- Query is more egalitarian about mappers/columns now. Still has a strong sense of "entity zero",
but also introduces new unpublished/experimental _values() method which sets up a columns-only query.
- Query.order_by() and Query.group_by() take *args now (also still take a list, will likely deprecate
in 0.5). May want to do this for select() as well.
- the existing "check for False discriminiator" "fix" was not working completely, added coverage
- orphan detection was broken when the target object was a subclass of the mapper with the orphaned
relation, fixed that too.
- session.refresh() and session.expire() raise an error when
called on instances which are not persistent within the session
- session._validate_persistent() properly raises an error for false check
- your cries have been heard: removing a pending item
from an attribute or collection with delete-orphan
expunges the item from the session; no FlushError is raised.
Note that if you session.save()'ed the pending item
explicitly, the attribute/collection removal still knocks
it out.
an existing table (such as a table that was already
reflected) using the 'useexisting=True' flag, which now
takes into account the arguments passed along with it.
- fixed one element of [ticket:910]
- refactored reflection test
and update() constructs which return a new object with
criterion joined to existing criterion via AND, just
like select().where().
- compile assertions use assertEquals()
- added 2.4-style binops to util.Set on 2.3
- OrderedSets pickle on 2.3
- more lib/sqlalchemy set vs Set corrections
- fixed InstrumentedSet.discard for 2.3
- set, sorted compatibility for test suite
- added testing.fails_if decorator
- Part one of test suite fixes to run on 2.3
Lots of failures still around sets; sets.Set differs from __builtin__.set
particularly in the binops. We depend on set extensively now and may need to
provide a corrected sets.Set subclass on 2.3.
- Importing testenv has no side effects- explicit functions provide similar behavior to the old immediate behavior of testbase
- testing.db has the configured db
- Fixed up the perf/* scripts
the tree can now run the full suite of tests out of the box.
- Migrated most @supported to @fails_on, @fails_on_everything_but, or (last
resort) @unsupported. @fails_on revealed a slew of bogus test skippage,
which was corrected.
- Added @fails_on_everything_but. Yes, the first usage *was*
"fails_on_everything_but('postgres')". How did you guess!
- Migrated @supported in dialect/* to the new test-class attribute __only_on__.
- Test classes can also have __unsupported_on__ and __excluded_on__.
- removed ClauseParameters object; compiled.params returns a regular dictionary
now, as well as result.last_inserted_params()/last_updated_params().
- various code trimming, method removals.
- reduced complexity of parameter handling during execution; __distill_params does all
parameter munging, executioncontext.parameters always holds a list of parameter structures
(lists, tuples, or dicts).
"column_keys". the parameters sent to execute() only interact with the
insert/update statement compilation process in terms of the column names
present but not the values for those columns.
produces more consistent execute/executemany behavior, simplifies things a
bit internally.
2. compiler names changed to be less verbose, unused classes removed.
3. Methods on Dialect which return compilers, schema generators, identifier preparers
have changed to direct class references, typically on the Dialect class itself
or optionally as attributes on an individual Dialect instance if conditional behavior is needed.
This takes away the need for Dialect subclasses to know how to instantiate these
objects, and also reduces method overhead by one call for each one.
4. as a result of 3., some internal signatures have changed for things like compiler() (now statement_compiler()), preparer(), etc., mostly in that the dialect needs to be passed explicitly as the first argument (since they are just class references now). The compiler() method on Engine and Connection is now also named statement_compiler(), but as before does not take the dialect as an argument.
5. changed _process_row function on RowProxy to be a class reference, cuts out 50K method calls from insertspeed.py
- changed testing connection closer to work on _ConnectionFairy instances, resulting in
pool checkins, not actual closes
- disabled session two phase test for now, needs work
- added some two-phase support to TLEngine, not tested
- TLTransaction is now a wrapper
of convert_bind_param() and convert_result_value() to callable-returning
bind_processor() and result_processor() methods. if no callable is
returned, no pre/post processing function is called.
- hooks added throughout base/sql/defaults to optimize the calling
of bind param/result processors so that method call overhead is minimized.
special cases added for executemany() scenarios such that unneeded "last row id"
logic doesn't kick in, parameters aren't excessively traversed.
- new performance tests show a combined mass-insert/mass-select test as having 68%
fewer function calls than the same test run against 0.3.
- general performance improvement of result set iteration is around 10-20%.
Jason Kirtland