Upstream changes:
0.143 2021-03-28
- fix RT #134147: Failed test 'SQL::Abstract - no limit' / Thanks to Slaven Rezić and Thibault Duponchelle
0.142 2020-11-22
- module adopted by Alexander Becker
- set up GitHub repository. See this repository for detailed changes.
- fix RT #34847: Avoid double quoting of table name. Fix provided by @dekimsey / Daniel Kimsey
- fix RT #50795: SQL::Abstract properties were ignored for select() if no limit was given. Thanks to Gunnar Hansson
- fix RT #73444: Offset argument wasn't handled properly if the value was 0. Thanks to ALEXBYK
- fix RT #62671: Properly handle 0 values in condition.
- remove Build.PL, switch to EUMM
Avoids build failure if monogb 4.x is installed.
Someone can always add a mongodb option to options.mk if they want, but
this is just to ensure reliable builds
- Updated all the live engine tests, aside from Oracle, to test with
unique registry names, so as to avoid conflicts when multiple
instances are being tested at once. Thanks to Slaven Rezić for the
report (#597).
- Removed `local` directory from the distribution, accidentally included
in v1.2.0. Thanks to gregor herrmann for the report (#600).
5.51.0 (2021-12-01)
* Make eager loading via tactical_eager_loading no longer modify objects who
already have a cached value for the association (jeremyevans)
* Make association cloning handle cases where clone association sets
different :class option than cloned association (jeremyevans)
* Make column schema entries on MySQL include an :extra entry for the Extra
column in DESCRIBE output (bschmeck) (#1791)
* Update mock PostgreSQL adapter to default to PostgreSQL 14 instead of
PostgreSQL 9.5 (jeremyevans)
* Support Dataset#with_recursive :search and :cycle options on PostgreSQL
14+ for result ordering and cycle detection (jeremyevans)
* Avoid method redefined verbose mode warnings in lazy_attributes plugin
(jeremyevans)
Version 1.3.0 - 2021-09-14
* Add GROUPS frame to Window
* Add exclude to Window
* Add method for each type of join
* Remove support for Python older than 3.5
* Support Select queries in conditionals and functions
Version 1.2.2 - 2021-05-16
* Escape table and column names containing double quote
* Use alias on the main table used in UPDATE and INSERT
* Add support for Python 3.9
Version 1.2.1 - 2020-10-08
* Fix AttributeError in Select.params with windows
Version 1.2.0 - 2020-10-05
* Add explicit Windows to Select
* Fix missing Windows definitions in nested expressions
Version 1.1.0 - 2020-01-25
* Add ORDER BY clause to aggregate functions
* Add support for Python 3.8
* Add distinct on Select
Version 1.0.0 - 2018-09-30
* Add Flavor filter_ to fallback to case expression
* Allow to use expression in AtTimeZone
* Fix Select query in returning
* Add comparison predicates
* Add COLLATE
kVS: ----------------------------------------------------------------------
SQLite Release 3.37.0 On 2021-11-27
STRICT tables provide a prescriptive style of data type management, for developers who prefer that kind of thing.
When adding columns that contain a CHECK constraint or a generated column containing a NOT NULL constraint, the ALTER TABLE ADD COLUMN now checks new constraints against preexisting rows in the database and will only proceed if no constraints are violated.
Added the PRAGMA table_list statement.
CLI enhancements:
Add the .connection command, allowing the CLI to keep multiple database connections open at the same time.
Add the --safe command-line option that disables dot-commands and SQL statements that might cause side-effects that extend beyond the single database file named on the command-line.
Performance improvements when reading SQL statements that span many lines.
Added the sqlite3_autovacuum_pages() interface.
The sqlite3_deserialize() does not and has never worked for the TEMP database. That limitation is now noted in the documentation.
The query planner now omits ORDER BY clauses on subqueries and views if removing those clauses does not change the semantics of the query.
The generate_series table-valued function extension is modified so that the first parameter ("START") is now required. This is done as a way to demonstrate how to write table-valued functions with required parameters. The legacy behavior is available using the -DZERO_ARGUMENT_GENERATE_SERIES compile-time option.
Added new sqlite3_changes64() and sqlite3_total_changes64() interfaces.
Added the SQLITE_OPEN_EXRESCODE flag option to sqlite3_open_v2().
Use less memory to hold the database schema.
Currently supports dBase III, Clipper, FoxPro, and Visual FoxPro
tables. Text is returned as unicode, and codepage settings in tables
are honored. Memos and Null fields are supported. Documentation
needs work, but author is very responsive to e-mails.
Not supported: index files (but can create tempory non-file indexes),
auto-incrementing fields, and Varchar fields.
- Fixed partitioned search for deployed changes on Oracle and Exasol to
correctly support plans with more than 250 changes. Thanks to @Nicqu
for the report (#521).
- DBI connections to the registry will now be set to trace level one
when a verbosity level of three or higher is passed to Sqitch
(i.e., `sqitch -vvv`). Thanks to @wkoszek for the suggestion (#155).
- Renamed the "master" branch to "main" and updated all relevant
references.
- Fixed the parsing of the Snowflake account name from the target URL
to that it no longer strips out the region and cloud platform parts.
Also deprecated the use of the region, host, and port config and
environment variables. This is to comply with the new account name
format. Thanks to @ldsingh00 for the report (#544).
- The checkout command will now show a usage statement when no branch
name is passed to it, rather than a series of warnings and a Git error.
Thanks to François Beausoleil for the report (#561).
- The checkout command now works when operating on a Sqitch project in
a subdirectory of a Git project. Thanks to François Beausoleil for the
report and suggested fix (#560).
- Fixed a failing bundle test when a top directory was configured in the
user or system configuration. Thanks to @CodingMinds for the spot
(#587).
- Added support to the Exasol engine for the `AUTHMETHOD` query parameter
(to allow Open ID authentication) and the `SSLCERTIFICATE=SSL_VERIFY_NONE`
query parameter to disable SSL verification. Thanks to Torsten Glunde
for the report (#588).
- Fixed "Use of uninitialized value $engine in concatenation" error when
Sqitch is unable to determine the engine when writing a plan file. Only
happens in the rare case that the core.engine value is not set.
- Improved the error message when attempting to update a plan file and no
project pragma is present. Thanks to Laurentiu Diaconu for the report
(#577).
- Fixed the error message when attempting to deploy a change that has
already been deployed to display the name of the change, rather than a
memory address. Thanks to Neil Mayhew for the report (#579).
- Added destination locking, currently implemented for PostgresQL and
MySQL. On starting a deploy or revert, Sqitch attempts to "lock the
destination" using advisory locks, to ensure that only one instance of
Sqitch makes changes to the database at any one time. This complements
the existing locking, which applies as each change is deployed or
reverted, as that pattern led to failures when multiple instances of
Sqitch were working at once. Thanks to Neil Mayhew for the report
(#579).
- Added the --lock-timeout option to the deploy, revert, rebase, and
checkout commands, to allow for shorter or longer times to wait for an
exclusive target database advisory lock for engines that support it.
Defaults to 60 seconds.
- Fixed the behavior of the plan command's `--max-count` option when used
with `--reverse` to show the proper items. Thanks to Adrian Klaver for
the report (#517).
- Fixed an issue that could cause bugs with the `check` command on
Firebird, Exasol, Oracle, and Vertica. Broken since the `check` command
was added in v1.1.0.
v0.25.0
Changes
Improve SSL option compatibility in URIs
Add Pool methods to determine its min, max, current and idle size
Make it possible to specify a statement name in Connection.prepare()
Implement support for multirange types
Fixes
Make sure timeout callbacks always get cleaned up
Update __all__ statements to a simpler form that is better supported by typecheckers
Fix test_timetz_encoding on Python 3.10
Fix a bunch of ResourceWarnings in the test suite
Fix SSLContext deprecation warnings
Fix the description of the database argument to connect()
Fix parsing of IPv6 addresses in the connection URI
Improve diagnostics of invalid executemany() input
PostgreSQL 14.1, 13.5, 12.9, 11.14, 10.19, and 9.6.24
Security Issues
CVE-2021-23214: Server processes unencrypted bytes from man-in-the-middle
Versions Affected: 9.6 - 14. The security team typically does not test unsupported versions, but this problem is quite old.
When the server is configured to use trust authentication with a clientcert requirement or to use cert authentication, a man-in-the-middle attacker can inject arbitrary SQL queries when a connection is first established, despite the use of SSL certificate verification and encryption.
The PostgreSQL project thanks Jacob Champion for reporting this problem.
CVE-2021-23222: libpq processes unencrypted bytes from man-in-the-middle
Versions Affected: 9.6 - 14. The security team typically does not test unsupported versions, but this problem is quite old.
A man-in-the-middle attacker can inject false responses to the client's first few queries, despite the use of SSL certificate verification and encryption.
If more preconditions hold, the attacker can exfiltrate the client's password or other confidential data that might be transmitted early in a session. The attacker must have a way to trick the client's intended server into making the confidential data accessible to the attacker. A known implementation having that property is a PostgreSQL configuration vulnerable to CVE-2021-23214.
As with any exploitation of CVE-2021-23214, the server must be using trust authentication with a clientcert requirement or using cert authentication. To disclose a password, the client must be in possession of a password, which is atypical when using an authentication configuration vulnerable to CVE-2021-23214. The attacker must have some other way to access the server to retrieve the exfiltrated data (a valid, unprivileged login account would be sufficient).
The PostgreSQL project thanks Jacob Champion for reporting this problem.
Bug Fixes and Improvements
This update fixes over 40 bugs that were reported in the last several months. The issues listed below affect PostgreSQL 14. Some of these issues may also affect other supported versions of PostgreSQL.
Some of these fixes include:
Fix physical replication for cases where the primary crashes after shipping a WAL segment that ends with a partial WAL record. When applying this update, update your standby servers before the primary so that they will be ready to handle the fix if the primary happens to crash.
Fix parallel VACUUM so that it will process indexes below the min_parallel_index_scan_size threshold if the table has at least two indexes that are above that size. This problem does not affect autovacuum. If you are affected by this issue, you should reindex any manually-vacuumed tables.
Fix causes of CREATE INDEX CONCURRENTLY and REINDEX CONCURRENTLY writing corrupt indexes. You should reindex any concurrently-built indexes.
Fix for attaching/detaching a partition that could allow certain INSERT/UPDATE queries to misbehave in active sessions.
Fix for creating a new range type with CREATE TYPE that could cause problems for later event triggers or subsequent executions of the CREATE TYPE command.
Fix updates of element fields in arrays of a domain that is a part of a composite.
Disallow the combination of FETCH FIRST WITH TIES and FOR UPDATE SKIP LOCKED.
Fix corner-case loss of precision in the numeric power() function.
Fix restoration of a Portal's snapshot inside a subtransaction, which could lead to a crash. For example, this could occur in PL/pgSQL when a COMMIT is immediately followed by a BEGIN ... EXCEPTION block that performs a query.
Clean up correctly if a transaction fails after exporting its snapshot. This could occur if a replication slot was created then rolled back, and then another replication slot was created in the same session.
Fix for "overflowed-subtransaction" wraparound tracking on standby servers that could lead to performance degradation.
Ensure that prepared transactions are properly accounted for during promotion of a standby server.
Ensure that the correct lock level is used when renaming a table.
Avoid crash when dropping a role that owns objects being dropped concurrently.
Disallow setting huge_pages to on when shared_memory_type is sysv
Fix query type checking in the PL/pgSQL RETURN QUERY.
Several fixes for pg_dump, including the ability to dump non-global default privileges correctly.
Use the CLDR project's data to map Windows time zone names to IANA time zones.
This update also contains tzdata release 2021e for DST law changes in Fiji, Jordan, Palestine, and Samoa, plus historical corrections for Barbados, Cook Islands, Guyana, Niue, Portugal, and Tonga.
Also, the Pacific/Enderbury zone has been renamed to Pacific/Kanton. Also, the following zones have been merged into nearby, more-populous zones whose clocks have agreed with them since 1970: Africa/Accra, America/Atikokan, America/Blanc-Sablon, America/Creston, America/Curacao, America/Nassau, America/Port_of_Spain, Antarctica/DumontDUrville, and Antarctica/Syowa. In all these cases, the previous zone name remains as an alias.
1.7.5
Released: November 11, 2021
bug
[bug] [tests]
Adjustments to the test suite to accommodate for error message changes occurring as of SQLAlchemy 1.4.27.
1.7.4
Released: October 6, 2021
bug
[bug] [regression]
Fixed a regression that prevented the use of post write hooks on python version lower than 3.9
[bug] [environment]
Fixed issue where the MigrationContext.autocommit_block() feature would fail to function when using a SQLAlchemy engine using 2.0 future mode.
1.7.3
Released: September 17, 2021
bug
[bug] [mypy]
Fixed type annotations for the “constraint_name” argument of operations create_primary_key(), create_foreign_key(). Pull request courtesy TilmanK.
1.7.2
Released: September 17, 2021
bug
[bug] [typing]
Added missing attributes from context stubs.
[bug] [mypy]
Fixed an import in one of the .pyi files that was triggering an assertion error in some versions of mypy.
[bug] [ops] [regression]
Fixed issue where registration of custom ops was prone to failure due to the registration process running exec() on generated code that as of the 1.7 series includes pep-484 annotations, which in the case of end user code would result in name resolution errors when the exec occurs. The logic in question has been altered so that the annotations are rendered as forward references so that the exec() can proceed.
1.7.1
Released: August 30, 2021
bug
[bug] [installation]
Corrected “universal wheel” directive in setup.cfg so that building a wheel does not target Python 2. The PyPi files index for 1.7.0 was corrected manually. Pull request courtesy layday.
[bug] [pep484]
Fixed issue in generated .pyi files where default values for Optional arguments were missing, thereby causing mypy to consider them as required.
[bug] [batch] [regression]
Fixed regression in batch mode due to 883 where the “auto” mode of batch would fail to accommodate any additional migration directives beyond encountering an add_column() directive, due to a mis-application of the conditional logic that was added as part of this change, leading to “recreate” mode not being used in cases where it is required for SQLite such as for unique constraints.
1.7.0
Released: August 30, 2021
changed
[changed] [installation]
Alembic 1.7 now supports Python 3.6 and above; support for prior versions including Python 2.7 has been dropped.
[changed] [installation]
Make the python-dateutil library an optional dependency. This library is only required if the timezone option is used in the Alembic configuration. An extra require named tz is available with pip install alembic[tz] to install it.
[changed] [installation]
The dependency on pkg_resources which is part of setuptools has been removed, so there is no longer any runtime dependency on setuptools. The functionality has been replaced with importlib.metadata and importlib.resources which are both part of Python std.lib, or via pypy dependency importlib-metadata for Python version < 3.8 and importlib-resources for Python version < 3.9 (while importlib.resources was added to Python in 3.7, it did not include the “files” API until 3.9).
feature
[feature] [environment]
Enhance version_locations parsing to handle paths containing spaces. The new configuration option version_path_separator specifies the character to use when splitting the version_locations string. The default for new configurations is version_path_separator = os, which will use os.pathsep (e.g., ; on Windows).
[feature] [tests]
Created a “test suite” similar to the one for SQLAlchemy, allowing developers of third-party dialects to test their code against a set of Alembic tests that have been specially selected to exercise back-end database operations. At the time of release, third-party dialects that have adopted the Alembic test suite to verify compatibility include CockroachDB and SAP ASE (Sybase).
[feature] [general]
pep-484 type annotations have been added throughout the library. Additionally, stub .pyi files have been added for the “dynamically” generated Alembic modules alembic.op and alembic.config, which include complete function signatures and docstrings, so that the functions in these namespaces will have both IDE support (vscode, pycharm, etc) as well as support for typing tools like Mypy. The files themselves are statically generated from their source functions within the source tree.
usecase
[usecase] [batch]
Named CHECK constraints are now supported by batch mode, and will automatically be part of the recreated table assuming they are named. They also can be explicitly dropped using op.drop_constraint(). For “unnamed” CHECK constraints, these are still skipped as they cannot be distinguished from the CHECK constraints that are generated by the Boolean and Enum datatypes.
Note that this change may require adjustments to migrations that drop or rename columns which feature an associated named check constraint, such that an additional op.drop_constraint() directive should be added for that named constraint as there will no longer be an associated column for it; for the Boolean and Enum datatypes, an existing_type keyword may be passed to BatchOperations.drop_constraint as well.
See also Changing the Type of Boolean, Enum and other implicit CHECK datatypes
Including CHECK constraints
bug
[bug] [operations]
Fixed regression due to 803 where the .info and .comment attributes of Table would be lost inside of the DropTableOp class, which when “reversed” into a CreateTableOp would then have lost these elements. Pull request courtesy Nicolas CANIART.
[bug] [batch] [sqlite]
Batch “auto” mode will now select for “recreate” if the add_column() operation is used on SQLite, and the column itself meets the criteria for SQLite where ADD COLUMN is not allowed, in this case a functional or parenthesized SQL expression or a Computed (i.e. generated) column.
[bug] [commands]
Re-implemented the python-editor dependency as a small internal function to avoid the need for external dependencies.
[bug] [postgresql]
Fixed issue where usage of the PostgreSQL postgresql_include option within a Operations.create_index() would raise a KeyError, as the additional column(s) need to be added to the table object used by the construct internally. The issue is equivalent to the SQL Server issue fixed in 513. Pull request courtesy Steven Bronson.
1.6.5
Released: May 27, 2021
bug
[bug] [autogenerate]
Fixed issue where dialect-specific keyword arguments within the DropIndex operation directive would not render in the autogenerated Python code. As support was improved for adding dialect specific arguments to directives as part of 803, in particular arguments such as “postgresql_concurrently” which apply to the actual create/drop of the index, support was needed for these to render even in a drop index operation. Pull request courtesy Jet Zhou.
1.6.4
Released: May 24, 2021
bug
[bug] [op directives] [regression]
Fixed regression caused by just fixed 844 that scaled back the filter for unique=True/index=True too far such that these directives no longer worked for the op.create_table() op, this has been fixed.
1.6.3
Released: May 21, 2021
bug
[bug] [autogenerate] [regression]
Fixed 1.6-series regression where UniqueConstraint and to a lesser extent Index objects would be doubled up in the generated model when the unique=True / index=True flags were used.
[bug] [autogenerate]
Fixed a bug where paths defined in post-write hook options would be wrongly escaped in non posix environment (Windows).
[bug] [regression] [versioning]
Fixed regression where a revision file that contained its own down revision as a dependency would cause an endless loop in the traversal logic.
1.6.2
Released: May 6, 2021
bug
[bug] [regression] [versioning]
Fixed additional regression nearly the same as that of 838 just released in 1.6.1 but within a slightly different codepath, where “alembic downgrade head” (or equivalent) would fail instead of iterating no revisions.
1.6.1
Released: May 6, 2021
bug
[bug] [regression] [versioning]
Fixed regression in new revisioning traversal where “alembic downgrade base” would fail if the database itself were clean and unversioned; additionally repairs the case where downgrade would fail if attempting to downgrade to the current head that is already present.
1.6.0
Released: May 3, 2021
feature
[feature] [autogenerate]
Fix the documentation regarding the default command-line argument position of the revision script filename within the post-write hook arguments. Implement a REVISION_SCRIPT_FILENAME token, enabling the position to be changed. Switch from str.split() to shlex.split() for more robust command-line argument parsing.
[feature]
Implement a .cwd (current working directory) suboption for post-write hooks (of type console_scripts). This is useful for tools like pre-commit, which rely on the working directory to locate the necessary config files. Add pre-commit as an example to the documentation. Minor change: rename some variables from ticket 819 to improve readability.
bug
[bug] [autogenerate]
Refactored the implementation of MigrateOperation constructs such as CreateIndexOp, CreateTableOp, etc. so that they no longer rely upon maintaining a persistent version of each schema object internally; instead, the state variables of each operation object will be used to produce the corresponding construct when the operation is invoked. The rationale is so that environments which make use of operation-manipulation schemes such as those those discussed in Fine-Grained Autogenerate Generation with Rewriters are better supported, allowing end-user code to manipulate the public attributes of these objects which will then be expressed in the final output, an example is some_create_index_op.kw["postgresql_concurrently"] = True.
Previously, these objects when generated from autogenerate would typically hold onto the original, reflected element internally without honoring the other state variables of each construct, preventing the public API from working.
[bug] [environment]
Fixed regression caused by the SQLAlchemy 1.4/2.0 compatibility switch where calling .rollback() or .commit() explicitly within the context.begin_transaction() context manager would cause it to fail when the block ended, as it did not expect that the transaction was manually closed.
[bug] [autogenerate]
Improved the rendering of op.add_column() operations when adding multiple columns to an existing table, so that the order of these statements matches the order in which the columns were declared in the application’s table metadata. Previously the added columns were being sorted alphabetically.
[bug] [versioning]
The algorithm used for calculating downgrades/upgrades/iterating revisions has been rewritten, to resolve ongoing issues of branches not being handled consistently particularly within downgrade operations, as well as for overall clarity and maintainability. This change includes that a deprecation warning is emitted if an ambiguous command such as “downgrade -1” when multiple heads are present is given.
In particular, the change implements a long-requested use case of allowing downgrades of a single branch to a branchpoint.
Huge thanks to Simon Bowly for their impressive efforts in successfully tackling this very difficult problem.
[bug] [batch] ¶
Added missing batch_op.create_table_comment(), batch_op.drop_table_comment() directives to batch ops.
DBD::MariaDB is the Perl5 Database Interface driver for MariaDB
and MySQL databases. In other words: DBD::MariaDB is an interface
between the Perl programming language and the MariaDB/MySQL
programming API that comes with the MariaDB/MySQL relational database
management system. Most functions provided by this programming API
are supported. Some rarely used functions are missing, mainly
because no-one ever requested them. :-)
=== 5.50.0 (2021-11-01)
* Make Migrator :allow_missing_migration_files also allow down migrations
where the current database version is greater than the last migration file
version (francisconeves97) (#1789)
* Fix Model#freeze in composition, serialization, and
serialization_modification_detection plugins to return self (jeremyevans)
(#1788)
* Fix typecasting of lazy columns when using lazy_attributes plugin in model
where dataset selects from subquery (jeremyevans)
* Add :before_preconnect Database option, for configuring extensions loaded
via :preconnect_extensions (MarcPer, jeremyevans) (#1786)
* Change Dataset#columns! to use a LIMIT 0 query instead of a LIMIT 1 query
(jeremyevans)
* Add sql_log_normalizer extension for normalizing logged SQL, helpful for
analytics and sensitive data (jeremyevans)
* Add support for range_merge, multirange, and unnest, and PGMultiRange#op
to pg_range_ops extension (jeremyevans)
* Add pg_multirange extension with support for PostgreSQL 14+ multirange
types (jeremyevans)
=== 5.49.0 (2021-10-01)
* Switch block_given? usage to defined?(yield) (jeremyevans)
* Support table aliases for JOIN USING columns on PostgreSQL 14+
(jeremyevans)
* Support calling PostgreSQL procedures without arguments (jeremyevans)
* Support hstore subscripts in pg_hstore_ops on PostgreSQL 14+, for updating
only part of an hstore value (jeremyevans)
* Support JSONB subscripts in pg_json_ops on PostgreSQL 14+, for updating
only part of a JSONB value (jeremyevans)
* Support SQL::Expression#sequel_ast_transform for custom AST transforms on
arbitrary expressions (jeremyevans)
* Add Database#create_trigger :replace option on PostgreSQL 14+ for CREATE
OR REPLACE TRIGGER (jeremyevans)
* Make auto_validations plugin automatically setup no_null_byte validations
(jeremyevans)
* Add Model#validates_no_null_byte to validation_helpers plugin
(jeremyevans)
A PostgreSQL SQL syntax beautifier that can work as a console program or as a
CGI.
This SQL formatter/beautifier supports keywords from SQL-92, SQL-99, SQL-2003,
SQL-2008, SQL-2011 and PostgreSQL specifics keywords. Works with any other
databases too.
Changes from the submitted version in the PR:
- p5 prefix for the package name
- clean up of github logic
- REPLACE_PERL
- updated distinfo with BLAKE2s
Version 1.22
* Fix file header validation
* Fix key verification in sequential access
* Fix testing with DejaGNU 1.6.3
* Fix stack overflow in print_usage
* Fix a leak of avail entry on pushing a new avail block
The leak would occur if the original avail table had odd number of entries.
* New gdbmtool variables: errorexit, errormask, trace, timing
"Errorexit" and "errormask" control which GDBM errors would cause the
program termination and emitting a diagnostic message,
correspondingly. Both variables are comma-delimited lists of error
codes.
The "trace" variable enables tracing of the gdbmtool commands.
The "timing" variable, when set, instructs gdbmtool to print time
spent in each command it runs.
* New gdbmtool options: -t (--trace), and -T (--timing)
3.14.8
Back-out all changes to automatically use RETURNING for `SqliteExtDatabase`,
`CSqliteExtDatabase` and `APSWDatabase`. The issue I found is that when a
RETURNING cursor is not fully-consumed, any parent SAVEPOINT (and possibly
transaction) would not be able to be released. Since this is a
backwards-incompatible change, I am going to back it out for now.
Returning clause can still be specified for Sqlite, however it just needs to be
done so manually rather than having it applied automatically.
3.14.7
Fix bug in APSW extension with Sqlite 3.35 and newer, due to handling of last
insert rowid with RETURNING.
3.14.6
Fix pesky bug in new `last_insert_id()` on the `SqliteExtDatabase`.
3.14.5
This release contains a number of bug-fixes and small improvements.
* Only raise `DoesNotExist` when `lazy_load` is enabled on ForeignKeyField.
* Add missing convenience method `ModelSelect.get_or_none()`
* Allow `ForeignKeyField` to specify a custom `BackrefAccessorClass`,
* Ensure foreign-key-specific conversions are applied on INSERT and UPDATE.
* Add handling of MySQL error 4031 (inactivity timeout) to the `ReconnectMixin`
helper class.
* Support specification of conflict target for ON CONFLICT/DO NOTHING.
* Add `encoding` parameter to the DataSet `freeze()` and `thaw()` methods.
* Fix bug which prevented `DeferredForeignKey` from being used as a model's
primary key.
* Ensure foreign key's related object cache is cleared when the foreign-key is
set to `None`.
* Allow specification of `(schema, table)` to be used with CREATE TABLE AS...,
* Allow reusing open connections with DataSet.
* Add `highlight()` and `snippet()` helpers to Sqlite `SearchField`, for use
with full-text search extension.
* Preserve user-provided aliases in column names.
* Add support for Sqlite 3.37 strict tables.
* Ensure database is inherited when using `ThreadSafeDatabaseMetadata`, and
also adds an implementation in `playhouse.shortcuts` along with basic unit
tests.
* Better handling of Model's dirty fields when saving.
* Add basic support for MariaDB connector driver in `playhouse.mysql_ext`.
* Begin a basic implementation for a psycopg3-compatible pg database.
* Add provisional support for RETURNING when using the appropriate versions of
Sqlite or MariaDB.
All checksums have been double-checked against existing RMD160 and
SHA512 hashes
The following distfiles could not be fetched (some may be only fetched
conditionally):
./databases/cstore/distinfo D6.data.ros.gz
./databases/cstore/distinfo cstore0.2.tar.gz
./databases/cstore/distinfo data4.tar.gz
Upstream changes:
1.59 2021-02-22T13:25:52Z
- Added error detection for when a result set's callback doesn't return
an arrayref or arrayrefs
- Moved to GitLab for issue tracking
1.58 2020-11-02T13:34:48Z
- Added the ability for mock result sets to set custom attributes for
statement handles. Thanks to Erik Huelsmann for testing the new feature.
Upstream changes:
1.05 - Todd Rinaldo <toddr@cpan.org> 2020-12-11
- Additional tests for known corner cases.
- Use Perl_warn not warn in XS.
- Test cleanup. Abstract helper calls.
- Macro cleanup in XS for code readability.
1.04 - Todd Rinaldo <toddr@cpan.org> 2020-12-10
- add utf8 mode for read/write of database
- Major re-factor to remove SVs from the CDB internals. SVs are only generated at return.
This was in attempt to address a known memory leak in the previous implementation.
- CDB_File will warn if undef keys or values are passed to insert. This is meant to be
consistent with how Perl handles the situation.
- Consistently tidy CDB_File.xs and remove tabs
- Make unit tests use File::Temp so we don't leave temp files laying around
- update Devel::PPPort to 3.62
1.03 - Todd Rinaldo <toddr@cpan.org> 2020-11-13
WARNING: due to a bug (#12) on 1.02, CDB databases generated by 1.02 will
not be legible by all other versions and vice-versa. It is recommended you
re-generate any databases generated by that version.
- #12 Fix incompatibility with CDB files due to accidental algorithm change.
Add xt tests for 0.96 databases to assure this does not happen again.
- Fixup Changelog for metacpan
- #11 Add additional benchmark test in git for.
- Provide the CDB_File algorithm in POD in the event the original document is lost.
- Update Devel::PPPort to 3.54
Version 3.12.1 fixes a number of bugs:
- Fixed a bug that caused a multi-document transaction to fail when the first
operation was large bulk write (>48MB) that required splitting a batched
write command (`PYTHON-2915`_).
- Fixed a bug that caused the ``tlsDisableOCSPEndpointCheck`` URI option to
be applied incorrectly (`PYTHON-2866`_).
Changes in MySQL 5.7.36
Security Notes
Bugs Fixed
Security Notes
The linked OpenSSL library for MySQL Server has been updated to version 1.1.1l. Issues fixed in the new OpenSSL version are described at https://www.openssl.org/news/cl111.txt and and http://www.openssl.org/news/vulnerabilities.html.
Bugs Fixed
Incompatible Change: For all SELECT statements on a view, the query digest was based on the view definition. As a result, different queries had the same digest and aggregated together in the Performance Schema table events_statements_summary_by_digest, so statistics in that table were not usable for distinguishing distinct SELECT statements.
The query digest for each SELECT statement on a view now is based on the SELECT, not the view definition. This enables distinguishing distinct SELECT statements in the events_statements_summary_by_digest table. However, tools that use query digests may need some adjustment to account for this change. For example, MySQL Enterprise Firewall and query rewrite plugins rely on query digests and existing rules for them that are associated with views may need to be updated.
InnoDB: With undo log truncation enabled (innodb_undo_log_truncate=ON), it was possible for a deadlock and eventual failure to occur when an undo log truncate operation was initiated after a version upgrade from MySQL 5.6 to MySQL 5.7.34 or earlier. A patch introduced in MySQL 5.7.35
[Note] InnoDB: Found duplicate reference rseg: 33 space: 1 page: 3
[Note] InnoDB: Reset pre-5.7.2 rseg: 1 after duplicate is found.
If pre-5.7.2 rollback segment slots have no undo data to purge, a message similar to the following is emitted:
[Note] InnoDB: Successfully reset 32 pre-5.7.2 rseg slots.
If undo data is found in pre-5.7.2 rollback segment slots, a message similar to the following is emitted recommending a slow shutdown and restart:
[Note] InnoDB: pre-5.7.2 rseg: 2 holds data to be purged.
History length: 1. Recommend slow shutdown with innodb_fast_shutdown=0 and restart
InnoDB: Truncation of an undo tablespace during use by an active transaction raised an assertion failure. The transaction was prematurely marked as complete, permitting the truncation operation.
InnoDB: Deleting or updating a row from a parent table initiated a cascading SET NULL operation on the child table that set a virtual column value to NULL. The virtual column value should have been derived from the base column value.
Thanks to Yin Peng at Tencent for the contribution.
InnoDB: The InnoDB recovery process did not recognize that page compression had been applied to data that was being recovered, causing the tablespace data file to increase in size during the redo log apply phase, which could lead to a recovery failure for systems approaching a disk-full state.
Replication: The error messages issued by MySQL Replication when GTIDs required for auto-positioning have been purged could be incorrectly assigned or scrambled in some situations.
Replication: The contents of the gtid_executed and gtid_purged GTID sets were not persisted after restoring a dump taken using mysqldump. The dump file sequence has now been changed so that the mysql schema (which contains the mysql.gtid_executed table) is not dropped after the gtid_purged GTID set is written. A new option --skip-mysql-schema is added for mysqldump which lets you choose not to drop the mysql schema at all.
JSON: Conversion of JSON values to text caused linear growth of the destination string, resulting in an unnecessarily high number of reallocations. Now this process uses exponential growth instead, to reduce the number of allocations required.
This fix originally appeared in MySQL 8.0 and was backported to MySQL 5.7 by Annirudh Prasad, whom we thank for the contribution.
Concurrent insert operations on multiple tables with full-text indexes caused a large number of full-text index synchronization requests, resulting in an out of memory condition.
When a query uses a temporary table for aggregation, the group by item is used as a unique constraint on the temporary table: If the item value is already present, the row is updated; otherwise, a new row is inserted into the temporary table. If the item has a result field or reference item, it it evaluated twice, once to check whether the result exists in the temporary table and, if not, again while constructing the row to be inserted. When the group by item was nondeterministic, the result value used to check for existence differed from that with which an insert was attempted, causing the insert to be rejected if the value already existed in the table.
We fix this by using the hash of any nondeterministic items as the unique constraint, so that the hash is evaluated once only.
Quote handling was improved for the SHOW GRANTS statement.
