A magazine site moving off an emoji subdomain came apart under wp search-replace. WordPress Importer 0.9.5 rewrites URLs during the import instead, using the block markup parsers in core to tell a real URL from a lookalike string.
WordPress 6.9 ships live on stage during State of the Word on December 2. The timeline behind it, from the dry run 26 hours out to the T-30m handoff, is worth borrowing for your own launches.
An official WordPress GitHub repo is not a support promise. Roughly 20 repos were archived across the WordPress and bbPress orgs, which is a good reason to go looking for orphaned code in the sites you maintain.
A tiny patch to a custom gateway plugin broke checkout for about 15% of one store’s customers. Release Confirmation, tracked in #8009-meta, holds auto-updates back for 24 hours so a mistake like that stays small.
A client wanted their block plugin hosted under the WordPress GitHub organization and found the path opaque. Here are the actual criteria: community ownership, a sponsoring Make team, more than one maintainer, and what happens when maintenance stops.
A client’s store slowed to a crawl right after someone clicked update to WooCommerce 10.4. Here is what the release actually changes in the REST API, the Store API and the block editor, and which custom code you should review first.
Plugins have always exposed their features however they liked, so integration meant writing a wrapper per plugin and repairing it after every update. The Abilities API in WordPress 6.9 replaces that with one registration format.
A WooCommerce shop inherited custom plugins that broke on every update and dragged the admin down. The fix was not more hotfixes, it was the basics: prefixed functions, core hooks, sanitized input and capability checks.
Custom capabilities get messy fast on big WordPress sites. The Abilities API, in development for 6.9 as part of Core-AI, gives a structured way to define and manage them.
A client’s custom roles suddenly locked users out after WordPress 6.9 shipped its Abilities API. The fix was to register capabilities and map_meta_cap the way core now expects, not to brute-force add_cap calls.