I'm working on a series of enhancements to Eventframe. Today's:
1. I've completely restructured Eventframe's code layout. Nodes are now in their own folder. This should make it easier for others to dive in and add new node types.
If you are upgrading an Eventframe installation, please delete *.pyc as a safety measure as the corresponding .py files may have moved.
2. I've been finding ever more reasons to regret the /folder/node architecture decision. Today I fixed one of the problems it had caused. Nodes can specify custom URLs to handle /folder/node/custom endpoints. This was used by the Event node to specify the /rsvp endpoint for participant responses.
The problem was with an Event in the root folder, which then has a /event/rsvp path, rather than the usual /folder/event/rsvp. Eventframe would look for an "event" folder, not find it, and abort with a 404. Today's patch makes it look for a node if a folder can't be found, then look for a custom URL handler in the node.
I'm not happy with the implementation as it runs five db queries to get to the node. The only sensible way out of this appears to be to make Folder a type of Node and have all nodes hold a direct reference to their website (rather than the current node -> folder -> website). This will be a non-trivial db migration for existing installations.
The PyCon website should upgrade to the latest commit of Eventframe as (IIRC)
in.pycon.org/2012 links to a root folder in Eventframe. This will make the Event frame usable for taking workshop registrations.
Both today's changes are backward compatible. Installations can upgrade without issues (just remember to remove *.pyc).
Best,
Kiran