As part of the roadmap unveiled at WordCamp EU earlier this year, WordPress.com has started rolling out Gutenberg to a subset of users.
According to a WordPress.com Happiness Engineer, the team is testing the implementation to determine the best way and time to enable it. Users will not be able to use Gutenberg unless their theme is updated to support blocks and the various alignment options.
Theme Wranglers are already in the process of adding support to WordPress.com’s nearly 100 free themes.
A quick search of the WordPress.com support forums for Gutenberg provides some insight into what users think about the new editor. For example, this user provided feedback on the use of so many icons without displaying their textual equivalent.
For now, Gutenberg is opt-in but eventually will be opt-out. Once Gutenberg is made available to a wider audience, support documents and official blog posts will be published to inform users about the new editor.
Gutenberg 3.8 is available for download. This release features a full screen mode that hides both the admin bar and the menu. Unlike previous versions of Distraction-Free-Writing mode where things would fade in and out of view, these two items stay hidden until full-screen mode is disabled.
User Interface elements have been added to manage reusable blocks in bulk. Theme authors are now able to register editor styles for blocks by targeting the blocks themselves. This avoids combating CSS specificity and doesn’t require knowledge of the internal DOM structure for the editor.
The block settings icon has been moved from the right side of blocks to the toolbar. This change sets the foundation for refactoring the toolbar and it reduces clutter by keeping the icons together.
Gutenberg 3.8 also contains a significant increase in performance thanks to a new hand-coded default block parser.
Having a formal specification of the Gutenberg block grammar has allowed us both to maintain a stable core during the almost 40 releases of the plugin and lately to allow competing parser implementation to evolve and be compared in terms of performance and correctness.
In concrete terms, we are shipping a new default implementation that is hundreds of times faster than the spec and has been stress tested with really long posts (including Moby Dick). These tests are also available for anyone to run against. Memory consumption has also gone down dramatically for server side operations. I’d like to specially thank Dennis Snell and Ivan Enderlin for their great work improving this area.
To see a complete list of changes along with links to their corresponding pull requests, check out the release post.