Introduces the filter `pre_attachment_url_to_postid` to allow developers to short-circuit the function `attachment_url_to_postid()`.
The return values are expected to be an attachment ID, zero (`0`) to indicate no attachment was found or `null` to indicate the function should proceed as usual.
The function performs an expensive database query so developers making use of the function frequently may wish to use a custom table with appropriate indexes to reduce the load on their database server.
Props antpb, apermo, audrasjb, joedolson.
Fixes#61383.
Built from https://develop.svn.wordpress.org/trunk@59118
git-svn-id: http://core.svn.wordpress.org/trunk@58514 1a063a9b-81f0-0310-95a4-ce76da25c4cd
Adds a check for `ABSPATH` to the top of the `/wp-includes/blocks/index.php` file and prevents the file from loading if it is not defined.
This prevents the file from throwing errors when accessed directly.
Props khokansardar, mukesh27.
Fixes#62108.
See #60352.
Built from https://develop.svn.wordpress.org/trunk@59117
git-svn-id: http://core.svn.wordpress.org/trunk@58513 1a063a9b-81f0-0310-95a4-ce76da25c4cd
Introduces post format support for both the Query Block with the new parameter `format`. In the `build_query_vars_from_query_block()` function, this is converted to a `post_format` taxonomy query passed to `WP_Query`.
Also introduces the `format` parameter to the REST API's Posts controller to support the feature in the Query block. The parameter type is an enumerated string accepted the post formats supported by each post type.
Props poena, mukesh27, mamaduka, noisysocks, TimothyBlynJacobs.
Fixes#62014.
Built from https://develop.svn.wordpress.org/trunk@59115
git-svn-id: http://core.svn.wordpress.org/trunk@58511 1a063a9b-81f0-0310-95a4-ce76da25c4cd
This aims to make it easier to enter digits in datetime text fields on mobile devices, instead of opening the full keyboard.
Follow-up to [43], [1506], [2998], [3648], [6078], [6989], [7285], [7338], [20168], [20217], [28730], [32945], [49283].
Props sabernhardt.
Fixes#62109.
Built from https://develop.svn.wordpress.org/trunk@59114
git-svn-id: http://core.svn.wordpress.org/trunk@58510 1a063a9b-81f0-0310-95a4-ce76da25c4cd
Change posts and comments pagination `nav` regions to have an aria-label of "Posts pagination" and "Comments pagination", respectively. This provides improved context for the numeric links contained within these regions and differentiates between posts navigation regions and posts pagination regions.
Props michaelbourne, sabernhardt, afercia, xyulex, webmandesign, tirth03, joedolson.
Fixes#54260.
Built from https://develop.svn.wordpress.org/trunk@59113
git-svn-id: http://core.svn.wordpress.org/trunk@58509 1a063a9b-81f0-0310-95a4-ce76da25c4cd
Add `get_user()` as an alias for `get_user_by( 'id', $user_id )`. Add `get_user()` to `user.php` so it is not a pluggable function. Bring parity with other standard `get_` functions for a more intuitive developer experience.
Props sc0ttkclark, kushang78, joedolson, peterwilsoncc.
Fixes#35124.
Built from https://develop.svn.wordpress.org/trunk@59111
git-svn-id: http://core.svn.wordpress.org/trunk@58507 1a063a9b-81f0-0310-95a4-ce76da25c4cd
PHP 8.4 deprecates the use of `trigger_errror()` with `E_USER_ERROR` as the error level, as there are a number of gotchas to this way of creating a `Fatal Error` (`finally` blocks not executing, destructors not executing).
The recommended replacements are either to use exceptions or to do a hard `exit`.
As this is a test-only class, do not have to take BC-breaks into account.
Also, as this is a test helper, throwing a exception is the most appropriate solution.
Reference:
* https://wiki.php.net/rfc/deprecations_php_8_4#deprecate_passing_e_user_error_to_trigger_error
Follow-up to [25002].
Props jrf.
See #62061.
Built from https://develop.svn.wordpress.org/trunk@59109
git-svn-id: http://core.svn.wordpress.org/trunk@58505 1a063a9b-81f0-0310-95a4-ce76da25c4cd
PHP 8.4 deprecates the use of `trigger_errror()` with `E_USER_ERROR` as the error level, as there are a number of gotchas to this way of creating a `Fatal Error` (`finally` blocks not executing, destructors not executing).
The recommended replacements are either to use exceptions or to do a hard `exit`.
As this is a test-only class, do not have to take BC-breaks into account.
Also, as this is a test helper, throwing a exception is the most appropriate solution.
Reference:
* https://wiki.php.net/rfc/deprecations_php_8_4#deprecate_passing_e_user_error_to_trigger_error
Follow-up to [49230].
Props jrf.
See #62061.
Built from https://develop.svn.wordpress.org/trunk@59108
git-svn-id: http://core.svn.wordpress.org/trunk@58504 1a063a9b-81f0-0310-95a4-ce76da25c4cd
PHP 8.4 deprecates the use of `trigger_errror()` with `E_USER_ERROR` as the error level, as there are a number of gotchas to this way of creating a `Fatal Error` (`finally` blocks not executing, destructors not executing). The recommended replacements are either to use exceptions or to do a hard `exit`.
WP has its own `wp_trigger_error()` function, which under the hood calls `trigger_error()`. If passed `E_USER_ERROR` as the `$error_level`, this will hit the PHP 8.4 deprecation.
Now, there were basically three options:
* Silence the deprecation until PHP 9.0 and delay properly solving this until then. This would lead to an awkward solution, as prior to PHP 8.0, error silencing would apply to all errors, while, as of PHP 8.0, it will no longer apply to fatal errors. It also would only buy us some time and wouldn't actually solve anything.
* Use `exit($status)` when `wp_trigger_error()` is called with `E_USER_ERROR`. This would make the code untestable and would disable handling of these errors via custom error handlers, which makes this an undesirable solution.
* Throw an exception when `wp_trigger_error()` is called with `E_USER_ERROR`. This makes for the most elegant solution with the least BC-breaking impact, though it does open it up to the error potential being "caught" via a `try-catch`. That's not actually a bad thing and is likely to only happen for those errors which can be worked around, in which case, it's a bonus that that's now possible.
The third option is implemented which:
* Introduces a new `WP_Exception` class.
* Starts using `WP_Exception` in the `wp_trigger_error()` function when the `$error_level` is set to `E_USER_ERROR`.
This change is covered by pre-existing tests, which have been updated to expect the exception instead of a PHP error.
Why not use `WP_Error`?
Well, for one, this would lead to completely different behaviour (BC).
As `WP_Error` doesn't extend `Exception`, the program would not be stopped, but would continue running, which would be a much bigger breaking change and carries security risks. `WP_Error` also doesn't natively trigger displaying/logging of the error message, so in that case, it would still need an `exit` with the error message, bringing us back to point 2 above.
Introducing `WP_Exception` provides (essentially) the same behaviour in that it retains the fatal error and error message displaying/logging behaviors. It also introduces a base Exception class, from which future exception classes can extend.
References:
* https://wiki.php.net/rfc/deprecations_php_8_4#deprecate_passing_e_user_error_to_trigger_error
* https://www.php.net/manual/en/migration80.incompatible.php
Follow-up to [56530].
Props jrf, hellofromTonya.
See #62061.
Built from https://develop.svn.wordpress.org/trunk@59107
git-svn-id: http://core.svn.wordpress.org/trunk@58503 1a063a9b-81f0-0310-95a4-ce76da25c4cd
PHP 8.4 deprecates the use of `trigger_errror()` with `E_USER_ERROR` as the error level, as there are a number of gotchas to this way of creating a `Fatal Error` (`finally` blocks not executing, destructors not executing).
The recommended replacements are either to use exceptions or to do a hard `exit`.
This is an unmaintained external dependency; thus, the fix is made in the WP specific copy of the dependency.
As `trigger_error()` call looks to be a remnant of the PHP 4 era before a class could be declared as `abstract`, fixed by making both the method as well as the class `abstract` and removing the call to `trigger_error()`.
Ref:
* https://wiki.php.net/rfc/deprecations_php_8_4#deprecate_passing_e_user_error_to_trigger_error
* https://www.php.net/manual/en/migration80.incompatible.php
Follow-up to [7747].
Props jrf.
See #62061.
Built from https://develop.svn.wordpress.org/trunk@59106
git-svn-id: http://core.svn.wordpress.org/trunk@58502 1a063a9b-81f0-0310-95a4-ce76da25c4cd
PHP 8.4 deprecates the use of `trigger_errror()` with `E_USER_ERROR` as the error level, as there are a number of gotchas to this way of creating a `Fatal Error` (`finally` blocks not executing, destructors not executing). The recommended replacements are either to use exceptions or to do a hard `exit`.
This is an unmaintained external dependency; thus, the fix is made in the WP specific copy of the dependency.
Now, there were basically three options:
* Silence the deprecation until PHP 9.0 and delay properly solving this until then.
This would lead to an awkward solution, as prior to PHP 8.0, error silencing would apply to all errors, while, as of PHP 8.0, it will no longer apply to fatal errors.
It also would only buy us some time and wouldn't actually solve anything.
* Use `exit($status)`.
This would make the code untestable and would disable handling of these errors via custom error handlers, which makes this an undesirable solution.
* Throw an exception.
This makes for the most elegant solution with the least BC-breaking impact.
The third option is implemented which:
* Introduces a new `Text_Exception` class.
* Starts using that in the `Text_Diff::_check()` method in all applicable places.
* Adds tests for the first two error conditions.
References:
* https://wiki.php.net/rfc/deprecations_php_8_4#deprecate_passing_e_user_error_to_trigger_error
* https://www.php.net/manual/en/migration80.incompatible.php
Follow-up to [59070], [52978], [7747].
Props jrf.
See #62061.
Built from https://develop.svn.wordpress.org/trunk@59105
git-svn-id: http://core.svn.wordpress.org/trunk@58501 1a063a9b-81f0-0310-95a4-ce76da25c4cd
In the Patterns registry, use `apply_block_hooks_to_content` (introduced in [58291]) instead of the `WP_Block_Patterns_Registry` class's private `get_content` method. (The latter is removed as part of this changeset.)
In a similar vein, use `apply_block_hooks_to_content` in the `_build_block_template_result_from_file` and `_build_block_template_result_from_post` functions, respectively.
For that to work, `apply_block_hooks_to_content` is amended to inject the `theme` attribute into Template Part blocks, even if no hooked blocks are present.
This kind of centralization is required as a preparation for #61902.
Props bernhard-reiter, jonsurrell.
See #61902.
Built from https://develop.svn.wordpress.org/trunk@59101
git-svn-id: http://core.svn.wordpress.org/trunk@58497 1a063a9b-81f0-0310-95a4-ce76da25c4cd
This is the seventh part in a larger modularization of the data in `WP_Debug_Data`. Previously this was a single massive method drawing in debug data from various groups of related data, where the groups were independent from each other.
This patch separates the seventh of twelve groups, the `wp-dropins` info, into a separate method focused on that data.
This work precedes changes to make the `WP_Debug_Data` class more extensible for better use by plugin and theme code.
Developed in https://github.com/wordpress/wordpress-develop/pull/7418
Discussed in https://core.trac.wordpress.org/ticket/61648
Props apermo.
See #61648.
Built from https://develop.svn.wordpress.org/trunk@59100
git-svn-id: http://core.svn.wordpress.org/trunk@58496 1a063a9b-81f0-0310-95a4-ce76da25c4cd
When encountering inline SVG and MathML content in an HTML document, there are certain "integration points" which transition back into the HTML parsing ruleset. Previously, the HTML API was incorrectly switching into the namespace of the element transitioning into that ruleset.
In this patch, the correct transition is made, where all integration points refer to HTML rules, while non-integration points refer to the rules of the namespace corresponding to the token itself.
Developed in https://github.com/wordpress/wordpress-develop/pull/7425
Discussed in https://core.trac.wordpress.org/ticket/61576
Props dmsnell, jonsurrell.
See #61576.
Built from https://develop.svn.wordpress.org/trunk@59099
git-svn-id: http://core.svn.wordpress.org/trunk@58495 1a063a9b-81f0-0310-95a4-ce76da25c4cd
Moves the 'loading' and 'loaded' i18n strings for the `interactivity-router` to the script module data via the `script_module_data_@wordpress/interactivity-router` filter.
Key changes:
- Add the `filter_script_module_interactivity_router_data()` method, hooked into the `script_module_data_@wordpress/interactivity-router` filter, to set the `i18n` data with the 'loading' and 'loaded' messages.
- Rename the `print_router_loading_and_screen_reader_markup()` method to `print_router_markup()` and remove the screen reader markup from it because it's no longer needed.
- Remove the `loading` and `loaded` strings from the `core/router` store state because they're no longer needed.
- Initialize the `core/router` store with a minimal navigation object to prevent errors in the interactivity-router script module when the store is not properly initialized.
- Update corresponding unit tests to reflect these changes.
This change ensures that the `interactivity-router` i18n messages are localized in a single place and removes the need to initialize them in the `core/router` store state.
Props jonsurrell, swissspidy, czapla.
See #60647.
Built from https://develop.svn.wordpress.org/trunk@59097
git-svn-id: http://core.svn.wordpress.org/trunk@58493 1a063a9b-81f0-0310-95a4-ce76da25c4cd
Fixes an issue with the image block when using pattern overrides, where the image block with overrides enabled was not outputting all the expected image attributes. Ensures that the `process_block_bindings` method returns any updates to the block's binding metadata along with other computed attributes.
Props talldanwp, cbravobernal, santosguillamot, mukesh27, gziolo.
Fixes#62069.
Built from https://develop.svn.wordpress.org/trunk@59095
git-svn-id: http://core.svn.wordpress.org/trunk@58491 1a063a9b-81f0-0310-95a4-ce76da25c4cd
Ensures that the default values defined in the schema for block attributes are used when rendering the output of the block with `get_block_wrapper_attributes` helper.
Props gziolo, jonsurrell, youknowriad, ryelle.
Fixes#62114.
Built from https://develop.svn.wordpress.org/trunk@59093
git-svn-id: http://core.svn.wordpress.org/trunk@58489 1a063a9b-81f0-0310-95a4-ce76da25c4cd
Perform a canonical redirect for an invalid pagination request of a static front page.
When a site has a static front page assigned and that page has a `<!--nextpage-->` within its content, previously accessing non-existing pages (e.g. `example.com/page/3/`) did not redirect or return a 404 or 301. This changeset resolves that issue by performing a canonical redirect.
Unit tests are also included for this specific use case and to ensure the fix does not affect a blog listing home page.
Follow-up to [47738], [47727], [34492].
Props dd32, audrasjb, chaion07, hellofromTonya, joemcgill, lukecarbis, Mte90, mukesh27, peterwilsoncc, rajinsharwar, SergeyBiryukov.
Fixes#50163.
See meta#5184.
Built from https://develop.svn.wordpress.org/trunk@59091
git-svn-id: http://core.svn.wordpress.org/trunk@58487 1a063a9b-81f0-0310-95a4-ce76da25c4cd
The Script Module has the same API as the `wp-a11y` WP Script.
Key changes:
- Add `@wordpress/a11y` to the list of Script and Module dual packages.
- Update `script-modules-packages.min.php` to include the a11y module.
- Modify `WP_Script_Modules` class to track and handle a11y module availability.
- Add method to print required HTML markup for a11y `speak()` functionality.
See #60647.
Props jonsurrell, gziolo, czapla.
Built from https://develop.svn.wordpress.org/trunk@59089
git-svn-id: http://core.svn.wordpress.org/trunk@58485 1a063a9b-81f0-0310-95a4-ce76da25c4cd
This allows other users of the WordPress unit test suite framework to run their own unit tests without needing the WordPress Importer plugin, which should only be a requirement if running core tests.
Follow-up to [59085].
Props bjorsch.
Fixes#62106.
Built from https://develop.svn.wordpress.org/trunk@59086
git-svn-id: http://core.svn.wordpress.org/trunk@58482 1a063a9b-81f0-0310-95a4-ce76da25c4cd
If a hard requirement for the test suite is not fulfilled, running the tests should be blocked from the test bootstrap. A test should only fail when it doesn't produce the expected result.
Since the WordPress Importer plugin is considered a hard requirement for the test suite at this time, this commit moves the check whether the plugin is installed from individual tests to the test bootstrap.
Includes defining a global constant for the path to the file for reuse in the tests.
Reference: [https://make.wordpress.org/core/handbook/contribute/git/#unit-tests Core Contributor Handbook: The Code Repository (Git): Unit Tests].
Follow-up to [40531], [40532], [41090], [41169], [48592], [49535], [49571].
Props jrf, hellofromTonya.
See #61530.
Built from https://develop.svn.wordpress.org/trunk@59085
git-svn-id: http://core.svn.wordpress.org/trunk@58481 1a063a9b-81f0-0310-95a4-ce76da25c4cd
In [50030] we enforced that Application Passwords have unique names. This was done with the assumption that applications would not connect to a user multiple times. However, in practice we've seen applications run into issues with the unique name constraint. Depending on the app, they may not know if they've been authorized before, or they may intentionally allow connecting multiple times. To prevent friction, App developers need to make their App Name unique, and in doing so often include things like the current date & time, which is already included in the App Passwords list table.
This commit removes this requirement to simplify usage of the Authorize Application flow.
Props mark-k, Boniu91, timothyblynjacobs, peterwilsoncc.
Fixes#54213.
Built from https://develop.svn.wordpress.org/trunk@59084
git-svn-id: http://core.svn.wordpress.org/trunk@58480 1a063a9b-81f0-0310-95a4-ce76da25c4cd
This is a companion to https://github.com/WordPress/gutenberg/pull/65460 that requires syncing in WordPress Core. Namely, the block-library changes require registration with their updated script module IDs so that the blocks continue to work correctly.
They key improvement is script modules registration is handled in one central place, and a combined asset file is used to improve the performance by avoiding multiple disk operations for every individual file.
Props jonsurrell, gziolo, wildworks, noisysocks.
See #60647, #59462.
Built from https://develop.svn.wordpress.org/trunk@59083
git-svn-id: http://core.svn.wordpress.org/trunk@58479 1a063a9b-81f0-0310-95a4-ce76da25c4cd
Composer 1.10.0 introduced a `lock` config option, which, when set to `false` will prevent a `composer.lock` file from being created and will ignore it when one exists.
This is a useful option for packages like WordPress where the `lock` file has no meaning.
It also makes life more straightforward for contributors as they don't have to remember that for this repo they should use `composer update` instead of `composer install`. Both will now work the same.
Reference: [https://getcomposer.org/doc/06-config.md#lock Composer Documentation: Config: lock].
Follow-up to [51543].
Props jrf.
See #61530.
Built from https://develop.svn.wordpress.org/trunk@59082
git-svn-id: http://core.svn.wordpress.org/trunk@58478 1a063a9b-81f0-0310-95a4-ce76da25c4cd
Removes query alteration from `build_comment_query_vars_from_block` by introducing a new way to pass the `$page` as argument to functions handling pagination for the comments.
Props cybr, santosguillamot, bernhard-reiter, gziolo.
Fixes#60806.
Built from https://develop.svn.wordpress.org/trunk@59081
git-svn-id: http://core.svn.wordpress.org/trunk@58477 1a063a9b-81f0-0310-95a4-ce76da25c4cd
Extends block context during block bindings processing. This implies that the context is extended ONLY for the blocks where bindings are defined and only when rendered on the page.
Props santosguillamot, gziolo, artemiosans, cbravobernal.
Fixes#61642.
Built from https://develop.svn.wordpress.org/trunk@59080
git-svn-id: http://core.svn.wordpress.org/trunk@58476 1a063a9b-81f0-0310-95a4-ce76da25c4cd
HTML often appears in ways that are unexpected. It may be missing implicit tags, may have unquoted, single-quoted, or double-quoted attributes, may contain duplicate attributes, may contain unescaped text content, or any number of other possible invalid constructions. The HTML API understands all fo these inputs, but downline parsers may not, and HTML snippets which are safe on their own may introduce problems when joined with other HTML snippets.
This patch introduces the `serialize()` method on the HTML Processor, which prints a fully-normative HTML output, eliminating invalid markup along the way. It produces a string which contains every missing tag, double-quoted attributes, and no duplicates. A `normalize()` static method on the HTML Processor provides a convenient wrapper for constructing a fragment parser and immediately serializing.
Subclasses relying on the `serialize_token()` method may perform structural HTML modifications with as much security as the upcoming `\Dom\HTMLDocument()` parser will, though these are not
able to provide the full safety that will eventually appear with `set_inner_html()`.
Further work may explore serializing to XML (which involves a number of other important transformations) and adding constraints to serialization (such as only allowing inline/flow/formatting elements and text).
Developed in https://github.com/wordpress/wordpress-develop/pull/7331
Discussed in https://core.trac.wordpress.org/ticket/62036
Props dmsnell, jonsurrell, westonruter.
Fixes#62036.
Built from https://develop.svn.wordpress.org/trunk@59076
git-svn-id: http://core.svn.wordpress.org/trunk@58472 1a063a9b-81f0-0310-95a4-ce76da25c4cd
Previously, there were a few cases where the modifiable text read from an HTML comment differs slightly from the parsed value of its inner text in a browser. This is due to the specific way that invalid HTML syntax tokens become "bogus comments."
This patch introduces a new method to the Tag Processor to allow differentiating these specific cases, such as when copying or serializing HTML from one source to another. Similar code has already been in use in the html5lib tests, and this patch simplifies the test runner, evidencing the fact that this method was already needed.
Developed in https://github.com/wordpress/wordpress-develop/pull/7342
Discussed in https://core.trac.wordpress.org/ticket/62036
Props dmsnell, jonsurrell.
See #62036.
Built from https://develop.svn.wordpress.org/trunk@59075
git-svn-id: http://core.svn.wordpress.org/trunk@58471 1a063a9b-81f0-0310-95a4-ce76da25c4cd
This ensures that the function applies the `wp_editor_settings` filter and merges the resulting array with the rest of TinyMCE init settings.
Includes a unit test to verify that the settings are merged correctly after adding the assignment of `array_merge()` result that was missed in the initial commit.
Follow-up to [44265], [59033].
Props kkmuffme, akshat2802, davidbaumwald, SergeyBiryukov.
Fixes#61754.
Built from https://develop.svn.wordpress.org/trunk@59074
git-svn-id: http://core.svn.wordpress.org/trunk@58470 1a063a9b-81f0-0310-95a4-ce76da25c4cd
Expands tests to ensure the version number of packages updated via NPM matches the version number used for registering the script in the script loader.
This adds tests for (by their registered name in WordPress):
* backbone
* clipboard
* hoverIntent
* hoverintent-js
* imagesloaded
* jquery-color
* jquery-core
* jquery-form
* masonry
* react-jsx-runtime
* underscore
* wp-polyfill-dom-rect
* wp-polyfill-element-closest
* wp-polyfill-fetch
* wp-polyfill-formdata
* wp-polyfill-inert
* wp-polyfill-node-contains
* wp-polyfill-object-fit
* wp-polyfill-url
This expands on the earlier tests introduced for:
* lodash
* moment
* react
* react-dom
* regenerator-runtime
An additional test is added to ensure that the data provider for these tests is maintained as libraries are added via package.json.
`@wordpress/*` scripts are excluded from these tests as wp-scripts generates a version number automatically based on the file's contents.
Additionally, the version of element-closest listed in package.json is updated to use a fixed version rather than a range. This reflects the current practice of WordPress to define the specific version in core. For the avoidance of doubt, this does not affect the version shipped in WordPress.
Follow up to [57185].
Props peterwilsoncc, jorbin.
Fixes#61855.
Built from https://develop.svn.wordpress.org/trunk@59071
git-svn-id: http://core.svn.wordpress.org/trunk@58467 1a063a9b-81f0-0310-95a4-ce76da25c4cd
On the first `foreach` loop in Text_Diff::_check()`, `$prevtype` is `null`. As `instanceof` requires the class name term to be an object or string, a fatal error is thrown:
>Fatal error: Uncaught Error: Class name must be a valid object or a string on line 279
This change:
* Adds a simple test for the `Text_Diff::_check()` method, which is how the bug was discovered as the test could never pass with the code as-is.
* Adds a defensive guard to protect against the fatal. It checks if `$prevtype` is not `null` as a pre-condition to for checking the instance. This bugfix also resolves the failing test.
Follow-up to [49194], [7747].
Props jrf, hellofromTonya.
See #62083.
Built from https://develop.svn.wordpress.org/trunk@59070
git-svn-id: http://core.svn.wordpress.org/trunk@58466 1a063a9b-81f0-0310-95a4-ce76da25c4cd
The `mysqli_ping()` function is deprecated as of PHP 8.4, though, in reality, the function wasn't working according to spec anymore since PHP 8.2 when the `libmysql` driver was dropped in favour of `libmysqlnd`, which was already the default (and recommended) driver since PHP 5.4.
The `mysqli_ping()` method was also not really correctly named as its functionality was to reconnect to the database, not just ping.
The alternative is to "manually" ping the database by sending a `DO 1` query (the cheapest possible SQL query).
Adding a PHP version based toggle was considered, but as mentioned above, the default driver has been `libmysqlnd` since PHP 5.4 and in that case, the function never worked anyway, so in reality `mysqli_ping()` was only really functional for the odd custom PHP compilation where `mysqli` was build against `libmysql` AND `reconnect` was not disabled.
With this in mind, this change replaces the call to `mysqli_ping()` with the `DO 1` query completely. If that query succeeds, it concludes the database connection is still alive. This solution should be the most stable as it will work for both PHP 7.2 <= 8.1, independently of which driver `mysqli` was compiled with, as well as for PHP 8.2+.
Note: It could also be considered to remove the function call to `mysqli_ping()` completely and rely on standard error handling in case the connection would have dropped, as after all, the fact that the connection existed at the moment the "ping" happened, is no guarantee that the connection will still exist when the next query is send.... this approach was not chosen so as WP has custom error handling and does not use the PHP native mysqli exceptions for this, which would make implementing this more awkward.
Includes a test to verify that the connection check works when there is a valid connection (this was previously not covered by tests).
Refs:
* https://wiki.php.net/rfc/deprecations_php_8_4#mysqli_ping_and_mysqliping
* https://github.com/php/php-src/pull/11912#issuecomment-1671762583
* https://stackoverflow.com/questions/2546868/cheapest-way-to-determine-if-a-mysql-connection-is-still-alive/2546922#2546922
* php/php-src#11945
* https://wiki.php.net/rfc/mysqli_support_for_libmysql
* https://www.php.net/mysqli_ping
Follow-up to [56475], [27250], [27075].
Props jrf, hellofromTonya.
See #62061.
Built from https://develop.svn.wordpress.org/trunk@59069
git-svn-id: http://core.svn.wordpress.org/trunk@58465 1a063a9b-81f0-0310-95a4-ce76da25c4cd
Peter Wilson
Sergey Biryukov
joedolson
hellofromTonya
dmsnell
Bernhard Reiter
czapla
cbravobernal
audrasjb
gziolo
Andrew Ozz
TimothyBlynJacobs
noisysocks