MySQL 8.0.17 deprecated the display width attribute for integer data types:
> As of MySQL 8.0.17, the `ZEROFILL` attribute is deprecated for numeric data types, as is the display width attribute for integer data types. You should expect support for `ZEROFILL` and display widths for integer data types to be removed in a future version of MySQL. Consider using an alternative means of producing the effect of these attributes. For example, applications can use the `LPAD()` function to zero-pad numbers up to the desired width, or they can store the formatted numbers in `CHAR` columns.
In practice, this means that display width is removed for integer types when creating a table:
* `BIGINT(20)` → `BIGINT`
* `INT(11)` → `INT`
* `MEDIUMINT(9)` → `MEDIUMINT`
* `SMALLINT(6)` → `SMALLINT`
* `TINYINT(4)` → `TINYINT`
Note: This only applies specifically to MySQL 8.0.17 or later. In MariaDB, display width for integer types is still available and expected.
This commit ensures that `dbDelta()`, which relies on the `DESCRIBE` SQL command to get the existing table structure and field types, when running on MySQL 8.0.17 or later, does not unnecessarily attempt to convert `BIGINT` fields back to `BIGINT(20)`, `INT` back to `INT(11)`, etc. When comparing the field type in the query with the existing field type, if display width is the only difference, it can be safely ignored to match MySQL behavior.
The change is covered by existing `dbDelta()` unit tests:
* A test for not altering `wp_get_db_schema()` queries on an existing install using MySQL 8.0.17+ now passes.
* More than twenty tests which previously failed on PHP 8.0.x + MariaDB due to incorrect expectations, caused by MariaDB version reporting not being consistent between PHP versions, now pass.
References:
* [https://dev.mysql.com/doc/refman/8.0/en/numeric-type-attributes.html MySQL: Nymeric Type Attributes]
* [https://mariadb.com/kb/en/data-types-numeric-data-types/ MariaDB: Numeric Data Types]
Follow-up to [1575], [18899], [37525], [47183], [47184].
Props SergeyBiryukov, pbearne, leewillis77, JavierCasares, desrosj, costdev, johnbillion.
Fixes#49364. See #51740.
git-svn-id: https://develop.svn.wordpress.org/trunk@53897 602fd350-edb4-49c9-b593-d223f7449a82
Adds a `public` visibility to test fixtures, tests, data providers, and callbacks methods.
Adds a `private` visibility to helper methods within test classes.
Renames callbacks and helpers that previously started with a `_` prefix. Why? For consistency and to leverage using the method visibility. Further naming standardizations is beyond the scope of this commit.
Props costdev, jrf, hellofromTonya.
Fixes#54177.
git-svn-id: https://develop.svn.wordpress.org/trunk@52010 602fd350-edb4-49c9-b593-d223f7449a82
> PHPUnit 8.0.0 introduced a `void` return type declaration to the "fixture" methods – `setUpBeforeClass()`, `setUp()`, `tearDown()` and `tearDownAfterClass()`. As the `void` return type was not introduced until PHP 7.1, this makes it more difficult to create cross-version compatible tests when using fixtures, due to signature mismatches.
>
> The `Yoast\PHPUnitPolyfills\TestCases\TestCase` overcomes the signature mismatch by having two versions. The correct one will be loaded depending on the PHPUnit version being used.
>
> When using this TestCase, if an individual test, or another TestCase which extends this TestCase, needs to overload any of the "fixture" methods, it should do so by using a snake_case variant of the original fixture method name, i.e. `set_up_before_class()`, `set_up()`, `assert_pre_conditions()`, `assert_post_conditions()`, `tear_down()`, and `tear_down_after_class()`.
>
> The snake_case methods will automatically be called by PHPUnit.
>
> > IMPORTANT: The snake_case methods should not call the PHPUnit parent, i.e. do not use `parent::setUp()` from within an overloaded `set_up()` method. If necessary, DO call `parent::set_up()`.
Reference: https://github.com/Yoast/PHPUnit-Polyfills#testcases
This commit renames all declared fixture methods, and calls to parent versions of those fixture methods, from camelCase to snake_case.
Follow-up to [51559-51567].
Props jrf, hellofromTonya, johnbillion, netweb, dd32, pputzer, SergeyBiryukov.
See #46149.
git-svn-id: https://develop.svn.wordpress.org/trunk@51568 602fd350-edb4-49c9-b593-d223f7449a82
The parent methods should always be called before any test-specific functionality is set up, unless there is a specific and documented reason.
See #52625
git-svn-id: https://develop.svn.wordpress.org/trunk@50449 602fd350-edb4-49c9-b593-d223f7449a82
This ensures that not only the return values match the expected results, but also that their type is the same.
Going forward, stricter type checking by using `assertSame()` should generally be preferred to `assertEquals()` where appropriate, to make the tests more reliable.
Props johnbillion, jrf, SergeyBiryukov.
See #38266.
git-svn-id: https://develop.svn.wordpress.org/trunk@48937 602fd350-edb4-49c9-b593-d223f7449a82
This avoids the performance overhead of the function call every time `dirname( __FILE__ )` was used instead of `__DIR__`.
This commit also includes:
* Removing unnecessary parentheses from `include`/`require` statements. These are language constructs, not function calls.
* Replacing `include` statements for several files with `require_once`, for consistency:
* `wp-admin/admin-header.php`
* `wp-admin/admin-footer.php`
* `wp-includes/version.php`
Props ayeshrajans, desrosj, valentinbora, jrf, joostdevalk, netweb.
Fixes#48082.
git-svn-id: https://develop.svn.wordpress.org/trunk@47198 602fd350-edb4-49c9-b593-d223f7449a82
Since MySQL 5.7, InnoDB engine supports `FULLTEXT` indexes, so forcing MyISAM is unnecessary.
Follow-up to [47193], which addressed the issue in `Tests_dbDelta::setUp()`, but missed other tests.
Props xkon, joonasvanhatapio, SergeyBiryukov.
Fixes#49367.
git-svn-id: https://develop.svn.wordpress.org/trunk@47195 602fd350-edb4-49c9-b593-d223f7449a82
* MySQL 8.0.11 changed the `GeometryCollection` data type name to `GeomCollection`, with the latter being the preferred name.
* MySQL 8.0.17 removed support for the display width attribute for integer data types. Previously, default display width of 20 digits was used: `BIGINT(20)`.
The affected tests now check the MySQL server version and use the appropriate data types.
This leaves one unresolved failure on MySQL 8.0.17+ to be addressed in the future, caused by the same `BIGINT` display width discrepancy coming from `wp_get_db_schema()`.
Props kaggdesign, ottok, jeremyfelt, SergeyBiryukov.
Fixes#44384, #49344. See #49364.
git-svn-id: https://develop.svn.wordpress.org/trunk@47184 602fd350-edb4-49c9-b593-d223f7449a82
All PHP files in `/tests` now conform to the PHP coding standards, or have exceptions appropriately marked.
Travis now also runs `phpcs` on the `/tests` directory, any future changes to these files must conform entirely to the WordPress PHP coding standards. 🎉
See #47632.
git-svn-id: https://develop.svn.wordpress.org/trunk@45607 602fd350-edb4-49c9-b593-d223f7449a82
WPCS 1.0.0 includes a bunch of new auto-fixers, which drops the number of coding standards issues across WordPress significantly. Prior to running the auto-fixers, there were 15,312 issues detected. With this commit, we now drop to 4,769 issues.
This change includes three notable additions:
- Multiline function calls must now put each parameter on a new line.
- Auto-formatting files is now part of the `grunt precommit` script.
- Auto-fixable coding standards issues will now cause Travis failures.
Fixes#44600.
git-svn-id: https://develop.svn.wordpress.org/trunk@43571 602fd350-edb4-49c9-b593-d223f7449a82
WordPress' code just... wasn't.
This is now dealt with.
Props jrf, pento, netweb, GaryJ, jdgrimes, westonruter, Greg Sherwood from PHPCS, and everyone who's ever contributed to WPCS and PHPCS.
Fixes#41057.
git-svn-id: https://develop.svn.wordpress.org/trunk@42343 602fd350-edb4-49c9-b593-d223f7449a82
On newer versions of MySQL, an error was being thrown when creating a table with an index that we wanted to be silently truncated.
To avoid this, the test now tries to use a newer InnoDB file format where available, and skips the test when that happens.
Props pento, danielbachhuber, straussd.
Fixes#41716.
git-svn-id: https://develop.svn.wordpress.org/trunk@41818 602fd350-edb4-49c9-b593-d223f7449a82
In [31349], core `varchar` column key lengths were changed from 255 to 191 to support the 767 byte index size limit on standard utf8mb4 MySQL installs. This changes the DB schema tests to match.
Props caseypatrickdriscoll, clarinetlord.
Fixes#35958.
git-svn-id: https://develop.svn.wordpress.org/trunk@40339 602fd350-edb4-49c9-b593-d223f7449a82
If index lengths change in table definitions, we don't recreate the index - instead, we throw a database error, as `dbDelta()` tries to create a new index with the same name.
It's better to leave the index as is, MySQL doesn't have an efficient process for resizing indices, and dropping/creating is a slow process which we don't want to trigger automatically.
Fixes#34870.
git-svn-id: https://develop.svn.wordpress.org/trunk@39921 602fd350-edb4-49c9-b593-d223f7449a82
When comparing index definitions, normalise the index names to lower case, as they are not case sensitive within MySQL.
Fixes#34874.
git-svn-id: https://develop.svn.wordpress.org/trunk@38591 602fd350-edb4-49c9-b593-d223f7449a82
`dbDelta()` compares the index definitions against the result of `SHOW INDEX FROM $table_name`. This requires a specific format so indices are not unnecessarily re-created. This format wasn't ensured, until now.
* Parse the raw index definition to extract the type, name and columns so a normalized definition can be built (#20263, #34873).
* Standardize on uppercase types (#34871) and on 'KEY'. 'INDEX' is only a synonym for 'KEY'.
* Escape index names with backticks (#20263).
* Normalize columns: Ignore ASC and DESC definitions (#34959), remove whitespaces (#34869) and escape column names with backticks (#20263).
* Add backticks to all index change queries (#20263).
Props ocean90, pento, kurtpayne.
Fixes#20263, #34869, #34871, #34873, #34959.
git-svn-id: https://develop.svn.wordpress.org/trunk@37583 602fd350-edb4-49c9-b593-d223f7449a82
`dbDelta()` already supported spatial fields (by virtue of not checking field types), so it's nice to round that out with spatial key support, too.
Fixes#36948.
git-svn-id: https://develop.svn.wordpress.org/trunk@37574 602fd350-edb4-49c9-b593-d223f7449a82
When upgrading to `utf8mb4`, `TEXT` fields will be upgraded to `MEDIUMTEXT` (and likewise for all other `*TEXT` and `*BLOB` fields). This is to allow for the additional space requirements of `utf8mb4`.
On the subsequent upgrade, after the `utf8mb4` upgrade, `dbDelta()` would try and downgrade the fields to their original size again. At best, this it a waste of time, at worst, this could truncate any data larger than the original size. There's no harm in leaving them at their original size, so let's do that.
Fixes#36748.
git-svn-id: https://develop.svn.wordpress.org/trunk@37525 602fd350-edb4-49c9-b593-d223f7449a82
Test to make sure that dbDelta properly inserts a value into the DB.
Props tryon, jtsternberg, ebinnion, JPry, avnarun, kevkoeh, salcode.
Fixes#29020.
git-svn-id: https://develop.svn.wordpress.org/trunk@33175 602fd350-edb4-49c9-b593-d223f7449a82
Add new tests for more parts of dbDelta. This was cleaned up and prepared for commit at #wcphilly contributor day.
Props jdgrimes, tryon, jtsternberg, ebinnion, JPry, avnarun, kevkoeh, jorbin, salcode.
Fixes#29020.
git-svn-id: https://develop.svn.wordpress.org/trunk@32770 602fd350-edb4-49c9-b593-d223f7449a82
