(lots of) changelog updates for 2.0.5

Change-Id: Iea2ca2cd1eb44b4e9139f2bb6c3d6239f67aeb56

13 files changed, 60 insertions, 33 deletions

diff --git a/doc/build/changelog/changelog_20.rst b/doc/build/changelog/changelog_20.rst
index eeee375b8..9c430bfe0 100644
--- a/doc/build/changelog/changelog_20.rst
+++ b/doc/build/changelog/changelog_20.rst
@@ -217,7 +217,7 @@
         lost.
 
     .. change::
-        :tags: bug, orm, regression
+        :tags: bug, orm, ression
         :tickets: 9232
 
         Fixed obscure ORM inheritance issue caused by :ticket:`8705` where some
diff --git a/doc/build/changelog/unreleased_20/5648.rst b/doc/build/changelog/unreleased_20/5648.rst
index acc1251b1..84539ecd3 100644
--- a/doc/build/changelog/unreleased_20/5648.rst
+++ b/doc/build/changelog/unreleased_20/5648.rst
@@ -3,7 +3,7 @@
     :tickets: 5648
 
     The support for pool ping listeners to receive exception events via the
-    :meth:`.ConnectionEvents.handle_error` event added in 2.0.0b1 for
+    :meth:`.DialectEvents.handle_error` event added in 2.0.0b1 for
     :ticket:`5648` failed to take into account dialect-specific ping routines
     such as that of MySQL and PostgreSQL. The dialect feature has been reworked
     so that all dialects participate within event handling.   Additionally,
diff --git a/doc/build/changelog/unreleased_20/8853.rst b/doc/build/changelog/unreleased_20/8853.rst
index a10ec43a6..fd0786973 100644
--- a/doc/build/changelog/unreleased_20/8853.rst
+++ b/doc/build/changelog/unreleased_20/8853.rst
@@ -6,7 +6,7 @@
     name to be fully qualified regardless of whether or not
     ``from __annotations__ import future`` were present. This issue first fixed
     in 2.0.0b3 confirmed that this case worked via the test suite, however the
-    test suite apparently was not testing the behavior for the name ``Mapped``
-    not being locally present at all; string resolution has been updated to
-    ensure the ``Mapped`` symbol is locatable as applies to how the ORM uses
-    these functions.
+    test suite apparently was not testing the behavior for the name
+    :class:`_orm.Mapped` not being locally present at all; string resolution
+    has been updated to ensure the :class:`_orm.Mapped` symbol is locatable as
+    applies to how the ORM uses these functions.
diff --git a/doc/build/changelog/unreleased_20/9343.rst b/doc/build/changelog/unreleased_20/9343.rst
index f6cf95e13..8c82ec1fa 100644
--- a/doc/build/changelog/unreleased_20/9343.rst
+++ b/doc/build/changelog/unreleased_20/9343.rst
@@ -2,7 +2,7 @@
     :tags: engine, performance
     :tickets: 9343
 
-    A small optimization to the Cython implementation of :class:`.ResultProxy`
+    A small optimization to the Cython implementation of :class:`.Result`
     using a cdef for a particular int value to avoid Python overhead. Pull
     request courtesy Matus Valo.
 
diff --git a/doc/build/changelog/unreleased_20/9349.rst b/doc/build/changelog/unreleased_20/9349.rst
index 957ac5687..7e708f379 100644
--- a/doc/build/changelog/unreleased_20/9349.rst
+++ b/doc/build/changelog/unreleased_20/9349.rst
@@ -1,6 +1,7 @@
 .. change::
-    :tags: bug, schema, postgresql
-    :tickets: 9332
+    :tags: bug, postgresql
+    :tickets: 9349
 
-    ExcludeConstraint correctly uses literal compile when compiling
-    expression ddl.
+    Fixed issue in PostgreSQL :class:`_postgresql.ExcludeConstraint` where
+    literal values were being compiled as bound parameters and not direct
+    inline values as is required for DDL.
diff --git a/doc/build/changelog/unreleased_20/9367.rst b/doc/build/changelog/unreleased_20/9367.rst
index 9355ffdea..883e5a215 100644
--- a/doc/build/changelog/unreleased_20/9367.rst
+++ b/doc/build/changelog/unreleased_20/9367.rst
@@ -2,10 +2,11 @@
     :tags: bug, ext
     :tickets: 9367
 
-    Fixed issue in automap where calling ``.prepare()`` from one of the mapped
-    classes would not use the correct base class when automap detected new
-    tables, instead using the given class, leading to mappers trying to
-    configure inheritance. While one should normally call ``.prepare()`` from
-    the base in any case, it shouldn't misbehave that badly when called from a
-    subclass.
+    Fixed issue in automap where calling :meth:`_automap.AutomapBase.prepare`
+    from a specific mapped class, rather than from the
+    :class:`_automap.AutomapBase` directly, would not use the correct base
+    class when automap detected new tables, instead using the given class,
+    leading to mappers trying to configure inheritance. While one should
+    normally call :meth:`_automap.AutomapBase.prepare` from the base in any
+    case, it shouldn't misbehave that badly when called from a subclass.
 
diff --git a/doc/build/changelog/unreleased_20/9376.rst b/doc/build/changelog/unreleased_20/9376.rst
index 6d63d09d4..1bec16485 100644
--- a/doc/build/changelog/unreleased_20/9376.rst
+++ b/doc/build/changelog/unreleased_20/9376.rst
@@ -2,7 +2,7 @@
     :tags: bug, typing
     :tickets: 9376
 
-    Improved typing for the mapping passed to :meth:`.UpdateBase.values` to be
-    more open-ended about collection type, by indicating read-only ``Mapping``
-    instead of writeable ``Dict`` which would error out on too limited of a key
-    type.
+    Improved typing for the mapping passed to :meth:`.Insert.values` and
+    :meth:`.Update.values` to be more open-ended about collection type, by
+    indicating read-only ``Mapping`` instead of writeable ``Dict`` which would
+    error out on too limited of a key type.
diff --git a/doc/build/changelog/unreleased_20/9379.rst b/doc/build/changelog/unreleased_20/9379.rst
index 97e9b6c9b..9784639df 100644
--- a/doc/build/changelog/unreleased_20/9379.rst
+++ b/doc/build/changelog/unreleased_20/9379.rst
@@ -1,6 +1,8 @@
 .. change::
-    :tags: bug, sqlite
+    :tags: bug, sqlite, regression
     :tickets: 9379
 
-    Fixed bug that prevented SQLAlchemy to connect when using a very old
-    sqlite version (before 3.8.3) on python 3.8+.
+    Fixed regression for SQLite connections where use of the ``deterministic``
+    parameter when establishing database functions would fail for older SQLite
+    versions, those prior to version 3.8.3. The version checking logic has been
+    improved to accommodate for this case.
diff --git a/doc/build/changelog/unreleased_20/9391.rst b/doc/build/changelog/unreleased_20/9391.rst
index 99336a71c..33c42f265 100644
--- a/doc/build/changelog/unreleased_20/9391.rst
+++ b/doc/build/changelog/unreleased_20/9391.rst
@@ -2,6 +2,7 @@
     :tags: bug, typing
     :tickets: 9391
 
-    Added missing init overload to :class:`_sql.Numeric` to allow
-    type checkers to properly resolve the type var given the
-    ``asdecimal`` parameter.
+    Added missing init overload to the :class:`_types.Numeric` type object so
+    that pep-484 type checkers may properly resolve the complete type, deriving
+    from the :paramref:`_types.Numeric.asdecimal` parameter whether ``Decimal``
+    or ``float`` objects will be represented.
diff --git a/doc/build/changelog/unreleased_20/9401.rst b/doc/build/changelog/unreleased_20/9401.rst
index 48d88f56b..50232959b 100644
--- a/doc/build/changelog/unreleased_20/9401.rst
+++ b/doc/build/changelog/unreleased_20/9401.rst
@@ -1,5 +1,8 @@
 .. change::
-    :tags: bug, schema, postgresql
+    :tags: bug, postgresql
     :tickets: 9401
 
-    Fixed issue when copying ExcludeConstraint.
+    Fixed issue where the PostgreSQL :class:`_postgresql.ExcludeConstraint`
+    construct would not be copyable within operations such as
+    :meth:`_schema.Table.to_metadata` as well as within some Alembic scenarios,
+    if the constraint contained textual expression elements.
diff --git a/doc/build/changelog/unreleased_20/9423.rst b/doc/build/changelog/unreleased_20/9423.rst
index ddb6f9f4f..84dc5cb3d 100644
--- a/doc/build/changelog/unreleased_20/9423.rst
+++ b/doc/build/changelog/unreleased_20/9423.rst
@@ -2,5 +2,5 @@
     :tags: bug, engine
     :tickets: 9423
 
-    Fixed bug where :meth:`_engine.Row`s could not be
-    unpickled by other processes.
+    Fixed bug where :class:`_engine.Row` objects could not be reliably unpickled
+    across processes due to an accidental reliance on an unstable hash value.
diff --git a/doc/build/changelog/unreleased_20/metadata_schema.rst b/doc/build/changelog/unreleased_20/metadata_schema.rst
index 88664d1cb..fdb840965 100644
--- a/doc/build/changelog/unreleased_20/metadata_schema.rst
+++ b/doc/build/changelog/unreleased_20/metadata_schema.rst
@@ -1,5 +1,5 @@
 .. change::
     :tags: schema
 
-    Validate that when provided the :paramref:`_sql.MetaData.schema`
-    argument of :class:`_sql.MetaData` is a string.
+    Validate that when provided the :paramref:`_schema.MetaData.schema`
+    argument of :class:`_schema.MetaData` is a string.
diff --git a/lib/sqlalchemy/engine/events.py b/lib/sqlalchemy/engine/events.py
index b906389a9..9028c8c33 100644
--- a/lib/sqlalchemy/engine/events.py
+++ b/lib/sqlalchemy/engine/events.py
@@ -744,6 +744,25 @@ class DialectEvents(event.Events[Dialect]):
         make use of a native ``ping()`` method supplied by the DBAPI which does
         not make use of disconnect codes.
 
+        .. versionchanged:: 2.0.0 The :meth:`.DialectEvents.handle_error`
+           event hook participates in connection pool "pre-ping" operations.
+           Within this usage, the :attr:`.ExceptionContext.engine` attribute
+           will be ``None``, however the :class:`.Dialect` in use is always
+           available via the :attr:`.ExceptionContext.dialect` attribute.
+
+        .. versionchanged:: 2.0.5 Added :attr:`.ExceptionContext.is_pre_ping`
+           attribute which will be set to ``True`` when the
+           :meth:`.DialectEvents.handle_error` event hook is triggered within
+           a connection pool pre-ping operation.
+
+        .. versionchanged:: 2.0.5 An issue was repaired that allows for the
+           PostgreSQL ``psycopg`` and ``psycopg2`` drivers, as well as all
+           MySQL drivers, to properly participate in the
+           :meth:`.DialectEvents.handle_error` event hook during
+           connection pool "pre-ping" operations; previously, the
+           implementation was non-working for these drivers.
+
+
         A handler function has two options for replacing
         the SQLAlchemy-constructed exception into one that is user
         defined.   It can either raise this new exception directly, in