Well, you can easily run CherryPy behind Apache (see http://trac.cherrypy.org/cgi-bin/trac.cgi/wiki/BehindApache). Since CherryPy provides a WSGI interface (although it's still experimental), you can also run your CherryPy app with any WSGI-compatible HTTP server (although I don't really see any advantage to doing this).
In the future when more WSGI environments are deployed, I think the benefits will be nice. For instance, under WSGIKit (http://svn.colorstudy.com/trunk/WSGIKit) you should be able to create a file:
# cherry_app.py:
from cherrypy import cpg, wsgiapp from my_cherry_app import root_instance cpg.root = root_instance wsgiapp.init() # This variable name is automatically picked up: application = wsgiapp.wsgiApp
Then your application would be available under cherry_app/ (wherever you happened to put that file in the URL space). Note that WSGIKit isn't a server itself, just a set of middleware that delegates to different applications.
It occurs to me that it would be considerably cleaner if it looked more like this:
from cherrypy import wsgiapp from my_cherry_app import root_instance application = wsgiapp.Application(my_cheery_app)
Otherwise you can only run one CherryPy application in a process...? That would be very limiting.
Anyway, the idea is you can deploy a CherryPy-based application fairly easily alongside other WSGI Python web applications. The particulars still need some work -- I don't like the little Python file used to put the pieces together -- but I think it's a good direction.
-- Ian Bicking / [EMAIL PROTECTED] / http://blog.ianbicking.org -- http://mail.python.org/mailman/listinfo/python-list
