WordPress 6.9 and the heading CSS specificity fix
WordPress 6.9 is at RC3 with a launch set for December 2nd. The heading specificity fix, the Post Editor iframe and the new DataViews all change what a custom theme has to do.
WordPress 6.9 is at RC3 with a launch set for December 2nd. The heading specificity fix, the Post Editor iframe and the new DataViews all change what a custom theme has to do.
How automated unit tests fit into WordPress plugin work: scaffolding a suite with WP-CLI, choosing what to assert about a custom post type, and the regressions that manual click-through testing keeps missing.
The December 10 dev chat pushed 6.9.1 planning into the new year, which says something about how 6.9 is holding up. Here is how I stage a major core update for client sites, and why the Gutenberg 22.2 labels matter.
Gutenberg 22.3 pulls font management out of the Global Styles panels and into its own Fonts page, makes the core Grid block responsive when columns are set, and stops the image cropper from losing your aspect ratio.
WordPress goes back to three major releases a year in 2026, on a four-month cycle. What that changes for maintenance, plus a small logger that catches deprecation notices before the next update lands.
I almost sold a client a custom Python bridge between OpenAI and WordPress, then State of the Word 2025 made it pointless. Notes on what the Abilities API and the WP AI Client actually change for developers and site owners.
A store was throwing random 500 errors at checkout because a deleted flash-sale product was still sitting in customer carts. WooCommerce 10.3.6 fixes that class of crash, along with a PHP 8.4 failure in the Helper updater.
A law firm got a WordPress site out of an AI builder in ten minutes, then came back two weeks later with a broken contact form and no search traffic. How I treat AI output as scaffolding instead of a finished build.
A WooCommerce client hit editor errors after updating to 6.9, with the official fix scheduled for January 2026. Why rolling back was the wrong call and what we filtered instead.
A 15k-product WooCommerce store had red LCP scores and broken layouts after stacking speed plugins. What actually helped: enqueueing block styles properly, plus the Safari LCP support and Compression Dictionaries work from the core performance chat.