Mike Bayer [Tue, 13 Sep 2022 15:00:46 +0000 (11:00 -0400)]
Add type awareness to evaluator
Fixed regression where using ORM update() with synchronize_session='fetch'
would fail due to the use of evaluators that are now used to determine the
in-Python value for expressions in the the SET clause when refreshing
objects; if the evaluators make use of math operators against non-numeric
values such as PostgreSQL JSONB, the non-evaluable condition would fail to
be detected correctly. The evaluator now limits the use of math mutation
operators to numeric types only, with the exception of "+" that continues
to work for strings as well. SQLAlchemy 2.0 may alter this further by
fetching the SET values completely rather than using evaluation.
For 1.4 this also adds "concat_op" as evaluable; 2.0 already has
more string operator support
Fixed regression caused by the fix for :ticket:`8231` released in 1.4.40
where connection would fail if the user does not have permission to query
the dm_exec_sessions or dm_pdw_nodes_exec_sessions system view when trying
to determine the current transaction isolation level.
Mike Bayer [Wed, 31 Aug 2022 15:07:23 +0000 (11:07 -0400)]
run update_subclass anytime we add new clslevel dispatch
Fixed event listening issue where event listeners added to a superclass
would be lost if a subclass were created which then had its own listeners
associated. The practical example is that of the :class:`.sessionmaker`
class created after events have been associated with the
:class:`_orm.Session` class.
Mike Bayer [Tue, 30 Aug 2022 14:47:24 +0000 (10:47 -0400)]
include TableClause.schema in cache key
Fixed issue where use of the :func:`_sql.table` construct, passing a string
for the :paramref:`_sql.table.schema` parameter, would fail to take the
"schema" string into account when producing a cache key, thus leading to
caching collisions if multiple, same-named :func:`_sql.table` constructs
with different schemas were used.
Mike Bayer [Tue, 30 Aug 2022 14:25:47 +0000 (10:25 -0400)]
implement event for merge/load=False for mutable state setup
Fixed issue in :mod:`sqlalchemy.ext.mutable` extension where collection
links to the parent object would be lost if the object were merged with
:meth:`.Session.merge` while also passing :paramref:`.Session.merge.load`
as False.
The event added here is currently private for expediency, but
is acceptable to become a public event at some point.
Mike Bayer [Tue, 30 Aug 2022 13:50:03 +0000 (09:50 -0400)]
apply consistent ORM mutable notes for all mutable SQL types
in
https://github.com/sqlalchemy/sqlalchemy/discussions/8447
I was surprised that we didnt have any notes about using Mutable
for ARRAY classes, since we have them for HSTORE and JSON.
Add a consistent topic box for these so we have something to
point towards.
Mike Bayer [Mon, 29 Aug 2022 14:43:36 +0000 (10:43 -0400)]
refine ruleset to determine when poly adaption should be used
Fixed regression appearing in the 1.4 series where a joined-inheritance
query placed as a subquery within an enclosing query for that same entity
would fail to render the JOIN correctly for the inner query. The issue
manifested in two different ways prior and subsequent to version 1.4.18
(related issue #6595), in one case rendering JOIN twice, in the other
losing the JOIN entirely. To resolve, the conditions under which
"polymorphic loading" are applied have been scaled back to not be invoked
for simple joined inheritance queries.
Mike Bayer [Tue, 23 Aug 2022 13:28:06 +0000 (09:28 -0400)]
integrate connection.terminate() for supporting dialects
Integrated support for asyncpg's ``terminate()`` method call for cases
where the connection pool is recycling a possibly timed-out connection,
where a connection is being garbage collected that wasn't gracefully
closed, as well as when the connection has been invalidated. This allows
asyncpg to abandon the connection without waiting for a response that may
incur long timeouts.
Mike Bayer [Mon, 22 Aug 2022 17:43:59 +0000 (13:43 -0400)]
reformat functions.rst; document coalsce
this file was all over the place autodocumenting all the
contents of functions.py with no regards to the heading
paragraph which seemed to be introducing the generic functions.
Use specific autoclass/autofunc docs as automodule is generally
unworkable.
Mike Bayer [Fri, 19 Aug 2022 15:12:41 +0000 (11:12 -0400)]
remove narrative "reconstructor" document
this event hook is not commonly used and this page does not
fit into the current narrative very well. We should possibly
write a new paragraph regarding how instances load
at some point though the best place to put it is not clear.
Mike Bayer [Fri, 19 Aug 2022 13:49:25 +0000 (09:49 -0400)]
fix some inaccuracies in with_variant doc
* the table wont create on mysql/mariadb b.c. user_name had
no length
* "utf-8" is not recognized by mysql/mariadb, use "utf8"
* mysql or mariadb name match is determined by the URL, not the
actual DB that is detected (I know I made it work that way but
I forgot)
* for the 1.4 backport only, will remove the "mariadb" part
as we dont support that API, #8408
Mike Bayer [Wed, 17 Aug 2022 17:06:51 +0000 (13:06 -0400)]
fill out all distinguising fields for AliasedInsp
Hardened the cache key strategy for the :func:`_orm.aliased` and
:func:`_orm.with_polymorphic` constructs. While no issue involving actual
statements being cached can easily be demonstrated (if at all), these two
constructs were not including enough of what makes them unique in their
cache keys for caching on the aliased construct alone to be accurate.
Mike Bayer [Tue, 16 Aug 2022 18:25:12 +0000 (14:25 -0400)]
refine transfer of cached ORM options for selectin, lazy
Fixed issue involving :func:`_orm.with_loader_criteria` where a closure
variable used as bound parameter value within the lambda would not carry
forward correctly into additional relationship loaders such as
:func:`_orm.selectinload` and :func:`_orm.lazyload` after the statement
were cached, using the stale originally-cached value instead.
This change brings forth a good refinement where we finally realize
we shouldn't be testing every ORM option with lots of switches, we
just let the option itself be given "here is your uncached version,
you are cached, tell us what to do!". the current decision is
that strategy loader options used the cached in all cases as they
always have, with_loader_criteria uses the uncached, because the
uncached will have been invoked with new closure state that we
definitely need. The only
edge that might not work is if with_loader_criteria referenced
an entity that is local to the query, namely a specific AliasedInsp,
however that's not a documented case for this. if we had to do that,
then we perhaps would introduce a more complex reconcilation
logic, and this would also give us the hook to do that.
For this approach to work in 1.4, state.load_options has to
be ordered, so, this does the switch of load_options from set->tuple,
which has been in 2.0 for a long time. if this change is not
feasbile, due to affecting other areas, we may have to scale
back this fix a bit, but for here, it's just two collections
without any deep impacts.
Mike Bayer [Tue, 9 Aug 2022 13:50:57 +0000 (09:50 -0400)]
improve names for datatypes section
Continuing along #8362, if we look at the document here without
"this is the same text I've read for 14 years", we begin to see
that the title "Column and Data types" makes no sense at all,
is there a "column type" and a "Data type"? I guess what I was
thinking at that time is that a type can be placed on a
Column, or it can be used whenever you have, you know, "data".
The phrase "SQL expression" wasn't discovered yet.
"SQL Datatype" is not spectacular but at least it's one term.
the new intro then is focused on the hierarchy layout so let's
name it that.
not amazing, but better than the duplicate name that made
no sense before.
Mike Bayer [Fri, 5 Aug 2022 21:25:05 +0000 (17:25 -0400)]
deep compare CTEs before considering them conflicting
Fixed issue where referencing a CTE multiple times in conjunction with a
polymorphic SELECT could result in multiple "clones" of the same CTE being
constructed, which would then trigger these two CTEs as duplicates. To
resolve, the two CTEs are deep-compared when this occurs to ensure that
they are equivalent, then are treated as equivalent.
Mike Bayer [Wed, 3 Aug 2022 16:08:54 +0000 (12:08 -0400)]
send in the dragons on async_scoped_session
make it clear that async_scoped_session.remove() must
be called, else memory will build up. Generally
discourage the whole pattern as well, as this is a
"framework" pattern and we don't really want to be supporting
frameworks. Also indicate that scopefunc must be idempotent
and lightweight.
Fixed issue where the SQL Server dialect's query for the current isolation
level would fail on Azure Synapse Analytics, due to the way in which this
database handles transaction rollbacks after an error has occurred. The
initial query has been modified to no longer rely upon catching an error
when attempting to detect the appropriate system view. Additionally, to
better support this database's very specific "rollback" behavior,
implemented new parameter ``ignore_no_transaction_on_rollback`` indicating
that a rollback should ignore Azure Synapse error 'No corresponding
transaction found. (111214)', which is raised if no transaction is present
in conflict with the Python DBAPI.
Mike Bayer [Tue, 2 Aug 2022 18:51:49 +0000 (14:51 -0400)]
reword yield_per a bit more
I'm still not satisified with this section as it is still
too wordy and dense, but at least let's put a better description
of what yield_per actually is and why one might use it at the top.
Mike Bayer [Tue, 2 Aug 2022 15:34:35 +0000 (11:34 -0400)]
fix up SAVEPOINT docs
these contained a factual error that the entire session is
expired, which is no longer the case (I can't find exactly
when this was changed). Additionally, added a PostgreSQL
specific example w/ IntegrityError as this is the most
common case for this. Tried to tighten up other language
and make it as clear as possible.
Mike Bayer [Mon, 1 Aug 2022 14:29:13 +0000 (10:29 -0400)]
repair psycopg2 (and psycopg) multiple hosts format
Fixed issue in psycopg2 dialect where the "multiple hosts" feature
implemented for :ticket:`4392`, where multiple ``host:port`` pairs could be
passed in the query string as
``?host=host1:port1&host=host2:port2&host=host3:port3`` was not implemented
correctly, as it did not propagate the "port" parameter appropriately.
Connections that didn't use a different "port" likely worked without issue,
and connections that had "port" for some of the entries may have
incorrectly passed on that hostname. The format is now corrected to pass
hosts/ports appropriately.
As part of this change, maintained support for another multihost style that
worked unintentionally, which is comma-separated
``?host=h1,h2,h3&port=p1,p2,p3``. This format is more consistent with
libpq's query-string format, whereas the previous format is inspired by a
different aspect of libpq's URI format but is not quite the same thing.
If the two styles are mixed together, an error is raised as this is
ambiguous.
Mike Bayer [Tue, 19 Jul 2022 14:50:05 +0000 (10:50 -0400)]
check for TypeDecorator when handling getitem
Fixed issue where :class:`.TypeDecorator` would not correctly proxy the
``__getitem__()`` operator when decorating the :class:`.ARRAY` datatype,
without explicit workarounds.
Federico Caselli [Fri, 17 Jun 2022 21:12:39 +0000 (23:12 +0200)]
add shield() in aexit
Added ``asyncio.shield()`` to the connection and session release process
specifically within the ``__aexit__()`` context manager exit, when using
:class:`.AsyncConnection` or :class:`.AsyncSession` as a context manager
that releases the object when the context manager is complete. This appears
to help with task cancellation when using alternate concurrency libraries
such as ``anyio``, ``uvloop`` that otherwise don't provide an async context
for the connection pool to release the connection properly during task
cancellation.
Mike Bayer [Sun, 17 Jul 2022 15:32:27 +0000 (11:32 -0400)]
use concat() directly for contains, startswith, endswith
Adjusted the SQL compilation for string containment functions
``.contains()``, ``.startswith()``, ``.endswith()`` to force the use of the
string concatenation operator, rather than relying upon the overload of the
addition operator, so that non-standard use of these operators with for
example bytestrings still produces string concatenation operators.
To accommodate this, needed to add a new _rconcat operator function,
which is private, as well as a fallback in concat_op() that works
similarly to Python builtin ops.
Mike Bayer [Mon, 11 Jul 2022 01:24:17 +0000 (21:24 -0400)]
support "SELECT *" for ORM queries
A :func:`_sql.select` construct that is passed a sole '*' argument for
``SELECT *``, either via string, :func:`_sql.text`, or
:func:`_sql.literal_column`, will be interpreted as a Core-level SQL
statement rather than as an ORM level statement. This is so that the ``*``,
when expanded to match any number of columns, will result in all columns
returned in the result. the ORM- level interpretation of
:func:`_sql.select` needs to know the names and types of all ORM columns up
front which can't be achieved when ``'*'`` is used.
If ``'*`` is used amongst other expressions simultaneously with an ORM
statement, an error is raised as this can't be interpreted correctly by the
ORM.
Mike Bayer [Thu, 7 Jul 2022 15:44:09 +0000 (11:44 -0400)]
document using fetch() with Oracle
We implemented working FETCH support, but it's not
yet implied by limit/offset. The docs make no mention
that this is available which is very misleading including
to maintainers. Make it clear that fetch() support is
there right now, it's just not yet implicit with
limit/offset.
Mike Bayer [Wed, 6 Jul 2022 01:05:18 +0000 (21:05 -0400)]
generalize sql server check for id col to accommodate ORM cases
Fixed issues that prevented the new usage patterns for using DML with ORM
objects presented at :ref:`orm_dml_returning_objects` from working
correctly with the SQL Server pyodbc dialect.
Here we add a step to look in compile_state._dict_values more thoroughly
for the keys we need to determine "identity insert" or not, and also
add a new compiler variable dml_compile_state so that we can skip the
ORM's compile_state if present.
Restrict the GitHub token permissions only to the required ones; this way, even if the attackers will succeed in compromising your workflow, they won’t be able to do much.
- Included permissions for the action. https://github.com/ossf/scorecard/blob/main/docs/checks.md#token-permissions
[Keeping your GitHub Actions and workflows secure Part 1: Preventing pwn requests](https://securitylab.github.com/research/github-actions-preventing-pwn-requests/)
Mike Bayer [Thu, 30 Jun 2022 23:10:06 +0000 (19:10 -0400)]
repair yield_per for non-SS dialects and add new options
Implemented new :paramref:`_engine.Connection.execution_options.yield_per`
execution option for :class:`_engine.Connection` in Core, to mirror that of
the same :ref:`yield_per <orm_queryguide_yield_per>` option available in
the ORM. The option sets both the
:paramref:`_engine.Connection.execution_options.stream_results` option at
the same time as invoking :meth:`_engine.Result.yield_per`, to provide the
most common streaming result configuration which also mirrors that of the
ORM use case in its usage pattern.
Fixed bug in :class:`_engine.Result` where the usage of a buffered result
strategy would not be used if the dialect in use did not support an
explicit "server side cursor" setting, when using
:paramref:`_engine.Connection.execution_options.stream_results`. This is in
error as DBAPIs such as that of SQLite and Oracle already use a
non-buffered result fetching scheme, which still benefits from usage of
partial result fetching. The "buffered" strategy is now used in all
cases where :paramref:`_engine.Connection.execution_options.stream_results`
is set.
Added :meth:`.FilterResult.yield_per` so that result implementations
such as :class:`.MappingResult`, :class:`.ScalarResult` and
:class:`.AsyncResult` have access to this method.
Mike Bayer [Tue, 28 Jun 2022 22:55:19 +0000 (18:55 -0400)]
produce column copies up the whole hierarchy first
Fixed issue where a hierarchy of classes set up as an abstract or mixin
declarative classes could not declare standalone columns on a superclass
that would then be copied correctly to a :class:`_orm.declared_attr`
callable that wanted to make use of them on a descendant class.
Originally it looked like this would produce an ordering change,
however an adjustment to the flow for produce_column_copies
has avoided that for now.
projects / thirdparty / sqlalchemy / sqlalchemy.git / log
