When the `ms_files_rewriting` flag is enabled, `ms_upload_constants()` is required to properly set upload directory constants. Once this fires, it is impossible to clean up for a non `ms_files_rewriting` test by turning the option back off.
Excluding these tests by default offer a more consistent environment overall. Any tests written for uploaded files in multisite should ideally have a correspondign test in this area.
This commit also moves existing `ms_files_rewriting` tests for `test_switch_upload_dir()`.
Fixes#30256
Built from https://develop.svn.wordpress.org/trunk@30286
git-svn-id: http://core.svn.wordpress.org/trunk@30286 1a063a9b-81f0-0310-95a4-ce76da25c4cd
* Variables in DocBlocks should be backtick-escaped
* Parameter and return types should be as specific as possible
* `@param` types and variables should align with each other, but not intentionally with the `@return` description
See #30230.
Built from https://develop.svn.wordpress.org/trunk@30284
git-svn-id: http://core.svn.wordpress.org/trunk@30284 1a063a9b-81f0-0310-95a4-ce76da25c4cd
Sharing these fixtures results in a speed improvement of almost one minute per
run of the test suite.
My hope is that future WordPress developers will spend this extra minute with
their loved ones, for life on this earth is short, my friends, and the moments
you spend watching WP generate test data can never again be reclaimed from the
grizzled clutches of Time, and none of us are really getting younger, I mean,
geez, have you looked in the mirror lately, Gandalf?
See #30017.
Built from https://develop.svn.wordpress.org/trunk@30277
git-svn-id: http://core.svn.wordpress.org/trunk@30277 1a063a9b-81f0-0310-95a4-ce76da25c4cd
Most of our unit tests that involve permalinks expect non-pretty permalinks;
this is the default setting for WP. However, when the test suite installs
multisite, `populate_network()` is run before WP recogizes itself as multisite,
which leads to pretty permalinks being set. This breaks a number of unit
tests in ways that are difficult to trace.
Built from https://develop.svn.wordpress.org/trunk@30267
git-svn-id: http://core.svn.wordpress.org/trunk@30267 1a063a9b-81f0-0310-95a4-ce76da25c4cd
This forcing appears to have been introduced to save unnecessary queries, but
(a) in some cases it appeared to be causing *more* queries, and (b) it caused
incorrect results when the 'exclude_tree' param of `get_terms()` called
`get_terms()` on each item in the tree using the 'child_of' param.
Fixes#30275.
Built from https://develop.svn.wordpress.org/trunk@30265
git-svn-id: http://core.svn.wordpress.org/trunk@30265 1a063a9b-81f0-0310-95a4-ce76da25c4cd
Our call to `$wp_roles->_init()` relied on the `__call()` method in `WP_Roles` to handle the link to the protected method. This works back to PHP 5.2.9, when a bug was fixed allowing access to protected methods through this exact approach.
`install_blog()` needs a fresh `$wp_roles` object after `populate_roles()` resets everything in its path. We can create this new object from scratch, effectively doing the same thing with the call to `_init()` via the constructor.
Fixes#29692 for trunk.
Built from https://develop.svn.wordpress.org/trunk@30242
git-svn-id: http://core.svn.wordpress.org/trunk@30242 1a063a9b-81f0-0310-95a4-ce76da25c4cd
When updating an existing taxonomy term that shares its `term_id` with
another term, we generate a new row in `wp_terms` and associate the updated
term_taxonomy_id with the new term. This separates the terms, such that
updating the name of one term does not change the name of any others.
Note that this term splitting only occurs on installations whose database
schemas have been upgraded to version 30133 or higher. Note also that shared
terms are only split when run through `wp_update_term()`, as on edit-tags.php;
we will wait until a future release of WordPress to force the splitting of all
shared taxonomy terms.
Props boonebgorges, rmccue, greuben, garyc40, wonderboymusic, imath, jesin.
Fixes#5809.
Built from https://develop.svn.wordpress.org/trunk@30241
git-svn-id: http://core.svn.wordpress.org/trunk@30241 1a063a9b-81f0-0310-95a4-ce76da25c4cd
A "shared" term occurs when two entries in the `wp_term_taxonomy` table share a
single `term_id`, and thereby correspond to the same row in `wp_terms`. This
changeset stops the practice of creating shared terms: each new row in
`wp_term_taxonomy` will receive its own row in `wp_terms`. The new strategy
for term creation depends on whether the installation's database schema is up
to date for 4.1:
* If so, terms are allowed to be created with the same slug as an existing term, as long as they are in different taxonomies and do not share a parent. Thus, a new tag with the slug 'wordpress' can exist alongside a category with the slug 'wordpress'.
* If not, new terms will be forced to have unique slugs. Thus, on an installation containing a category with the slug 'wordpress', a new tag 'WordPress' will get the slug 'wordpress-2'.
Fixes#21950. See #5809.
Built from https://develop.svn.wordpress.org/trunk@30240
git-svn-id: http://core.svn.wordpress.org/trunk@30240 1a063a9b-81f0-0310-95a4-ce76da25c4cd
In case of edge cases where a duplicate term might exist in a replicated
database for a split second, these explicit query clauses ensure that
`term_exists()` will always recognize the oldest matched term as the
canonical one. See [30238] for background.
Props pento.
See #22023, #5809.
Built from https://develop.svn.wordpress.org/trunk@30239
git-svn-id: http://core.svn.wordpress.org/trunk@30239 1a063a9b-81f0-0310-95a4-ce76da25c4cd
In [30056], the UNIQUE index was removed from the 'slug' column of `wp_terms`.
While we have numerous checks in place to avoid the creation of unwanted
duplicate term+term_taxonomy pairs, it's possible that in certain edge cases -
such as during the lag caused by database replication, or a race condition
involving near-simultaneous creation of more than one term - we'll end up
unwittingly inserting two identical rows.
The current changeset minimizes this risk by introducing a failsafe mechanism
into `wp_insert_term()`. After a term and term_taxonomy are INSERTed, we check
to see whether the term just created is a duplicate of an existing term; if so,
we delete the new one and keep the old one. This prevents problems caused by
replication lag, because SELECT queries that take place after an INSERT will
hit the master database; it mitigates race conditions by enforcing that the
term that was created first (ie, the one with the lowest `term_id`) is
always the "canonical" one.
Props nacin, markjaquith.
See #22023, #5809.
Built from https://develop.svn.wordpress.org/trunk@30238
git-svn-id: http://core.svn.wordpress.org/trunk@30238 1a063a9b-81f0-0310-95a4-ce76da25c4cd
