> In my experience, filters are a bad API anyway. Every time I've ever > wanted to use a filter, it was for a particular URL, which isn't > really supported without lots of jiggering.
I use filters all the time for things like: - gzip compression (in prod mode) - automatic XHTML validating of my pages (in dev mode) - automatic encoding/decoding of the data - authentication - caching - sessions - ... For all of these I want the filter to apply to *lots* of URLs, but not always *all* of them ... Using filters allows me to easily specify which URLs I want the filter to apply to ... What's your way of doing this without using a filter ? RhubarbTart claims that they can probably get a similar functionality with decorators ... But will you have to specify 5 or 6 decorators for each method ? Finally, beware of the statement "RhubarbTart is a framework with a CherryPy interface" . A more accurate statement would be "RhubartTart is a framework that implements a very small portion of the CherryPy interface". CherryPy has been around for many years and is being used in many critical projects so a lot of the problems and bugs have been ironed out. The fact that it is now being used in other projects like Subway and Turbogears (as opposed to standalone) has created the need for a few small changes to CP but we've reacted quickly and implemented these changes (and will continue to do so). Remi.
