I’ve been using WordPress 3.0 beta for about 2 weeks and I’ve only run into one odd behavior.
When I upload or upgrade a plugin, the old plugin gets deactivated and the new plugin version gets downloaded. What is not happening is that the page is not loading up the next steps of re-activating the new version (or just recently downloaded) plugin. I did not catch this and was running without the upgraded plugin for a day or two.
I’m keeping my copies current using SVN and I’m convinced it’s my wonky setup. This server is running NGINX and PHP-FPM. I have the NGINX running as www-data and PHP-FPM running as a separate userid (why? because I can). The NGINX and PHP5 are both installed using non-Ubuntu repos.
I don’t mind this situation and I’ll get it sorted out by running the beta on another box running Apache to get an understanding of what’s going on.
Over on Otto’s blog, he discusses the idea of holding off upgrading to 3.0 when it comes out. Given the hysteria on the support forum whenever a new version comes out, I think that’s an excellent idea. What would be fun is when the 3.0 version is released, and the automatic notification does not get triggered across the globe. Instead the blog gets updated and the notification does not get triggered until 3.1 comes out.
It’s wishful thinking on my part because too many people treat their self-hosted WordPress installation as a service and not as something they are responsible for. I think the majority of WordPress.ORG users are pretty savvy but the minority sure is vocal.
While talking about WordPress, the W3 Total Cache plugin is amazing. I’m playing with it while running the beta and it appears to be speeding up my site without using a disk cache. I have used WP Super Cache but this seems approach the problem from a different angle. Instead of creating HTML static pages, it has the option to use PHP’s opcache. I’m using APC and the speed looks like it’s working. It’s a very cool idea and the implementation looks really polished.
Once WordPress 3.0 is officially released, I’ll run two test blogs on two different boxes and compare the output with and without W3 Total Cache. This way I should be able to get almost scientific results for comparison.