What I ended up doing was using a command-line switch and dicts to abstract away the manual switching I was doing before:

if __name__ == ‘__main__’:
current_dir = os.path.dirname(os.path.abspath(__file__))

opts = OptionParser()
opts.add_option(’-e’, ‘–environment’,
dest=’environment’, default=’development’)

options, args = opts.parse_args()

engine = cherrypy.engine

base_config = {
‘log.access_file’: ’site_access.log’,
‘log.error_file’: ’site_errors.log’,
‘log.screen’: True
}

#functions for setting up different environments.
def development():
base_config.update({’server.socket_port’: 3030})
def production():
base_config.update({’server.socket_port’: 80,
’server.socket_host’: ‘0.0.0.0′})
dp = plugins.DropPrivileges(engine,
uid=”nobody”,
gid=”nogroup”)

dp.subscribe()

#dispatch based on environment
environments = {’development’: development,
‘dev’: development,
‘production’ : production,
‘prod’: production}
try:
environments[options.environment]()
except KeyError:
print “”"You specified an invalid environment option\n valid options are
([development | dev] | [production | prod])”"”
sys.exit(1)

cherrypy.config.update(base_config)

Thanks for the helpful article(s)!

I’ll definitely look more into this.

There’s a slightly easier way to do the above: change or add a new entry in _cp_config.environments:

thisdir = os.path.abspath(”.”)
from cherrypy import _cp_config
_cp_config.environments['production'].update({
‘error_page.default’: “%s/public/index.html” % thisdir,
‘error_page.401′: “%s/public/errors/401.html” % thisdir,
‘error_page.404′: “%s/public/errors/404.html” % thisdir,
‘error_page.500′: “%s/public/errors/500.html” % thisdir,
‘log.error_file’: “%s/logs/production_error.log” % thisdir,
‘log.access_file’: “%s/logs/production_access.log” % thisdir,
})

The _cp_config.environments dict works like a set of macros, allowing you to apply a set of config entries via a single config entry; that is, given the above, you may apply them all now via the single config entry: {’environment’: ‘production’}, or via the -e argument to cherryd.

Even though _cp_config looks private (since it starts with an underscore), this is the recommended way to collect entries together and name them. No need to invent your own environments on top of this system–it’s all there already.

Once you have created your own _cp_config.environments, then cherryd really can be used as a single point of entry.