anyMeta 3.1.2 released

It has been a madhouse, we have been releasing four anyMeta versions in less than three weeks. Why so many releases? We had real deadlines for some real life projects.

What did we improve and change?

In the last couple of versions we added and/or fixed:

• Better article creation. When you make a new article (or attachment, etc.), we used to clone a default and then let you adapt it. Now we do the cloning after you posted the form. This prevents non-edited things floating around in the database and on the site.
• Changed publication flow. We checked and double checked the complete content creation flow. It is now possible to publish the individual language versions of things, instead of all at once. And we fine tuned the differences between users who are allowed to publish and users who are not allowed to publish.
• Secure delete. Normally, when you delete something we only change the publish state to deleted, enabling undelete by re-publishing the thing. Now it is also possible to directly delete the whole thing, without any traces and undelete possibilities. Use with care!
• Searching in the admin interface. We fixed a problem in which it was not possible to search using double quotes in the /admin interface.
• Internet Explorer 7 support. Starting next month or so Microsoft will change the way Internet Explorer handles certain multimedia content. We changed the way we show that content to be compatible with the future changes in IE.
• Wiki fixes. We made some fixes in the wiki translation, especially concerning mailto links.
• Session handling. We now start a session for every visitor, also for anonymous visitors. Of course we don't give crawlers a session :)
• Session storage. We use the memcache to offload the database for storing session data. We also changed the serialisation of the session handling, enabling parallel requests on the same session.
• Page caching. The page caching now stores two versions of the pages, for browsers with and without support for deflated content.
• Compressed Javascript and CSS files. We changed the static content server and the handling code on the site of anyMeta to be able to compress (deflate) css and js files. Together with some comment and whitespace stripping. An overal size reduction of 80% has been achieved!
• Loadbalancing support. We can now host an anyMeta on different webservers at the same time. We have a pool of three loadbalanced webservers for the in-house hosted sites.
• Memcache. We use the memcached deamon for storing frequently used data. We adapted the PHP extension and the memcached deamon to support hierarchically structured keys. More about this extension another time.
• Cache synchronisation. The flush cache button now really flushes the caches, including the compiled template caches on the webservers. It also flushes the Access Control List cache, a feature much requested.
• Access Control List extension. We added an extension to the access control lists, so that it is possible to define exemptions from the general acl rules and from the trust permissions. In this way you can really appoint people to be able to edit specific (on id by id basis) things in the system.
• Image Quality. You can now define a separate quality for large and small images. In this way we have an adaptive compression, enabling higher quality thumbnail images.
• Error logging. We added a central logging table, in which we log all errors, warnings and notices. This will form the basis of a pro-active error signaling system for anyMeta sites. All PHP errors are inserted in this table.
• Sharing. We improved the syndication between anyMeta systems. Now you can define the depth to which we will fetch related items.
• XML-RPC support We added XML-RPC access to all XmlHttpRequest API functions. It is now very easy to let two anyMeta systems communicate with each other.
• Module system. The module system now supports uninstalling
• Captcha Module. We added a module to combat the growing spam problem. This module shows an image which has to be typed over by the person submitting the form. Giving a first line of defense against spam bots.
• Jury Module A module for signaling questionable content. The members of the community will decide if content should be removed or should stay on the site.
• LogonThrottle Module. A simple module that will deny logons to the system when the load is too high.
• Maximum attachment size. We now check the size of all attachments that are uploaded using an url instead of a file. Attachments that are bigger that the value configured in the php configuration are refused.
• Language handling. We fixed some bugs that resulted in the wrong language to be selected. Primary reason was an incompatibility between our authentication module and the PEAR Auth module version 1.3 and upwards.
• Crawler handling. Sometimes the search engine crawlers like anyMeta sites just a bit too much. We now implemented the If-Modified-Since headers for crawlers and keep them out if the load is too high. Sorry, come back later.. somehow we still value our human visitors above crawlers.

Besides this all we fixed more bugs and layed the foundation for huge future performance improvements.