twisted.web.wsgi usage example
Hi, I'm trying to come up with an example on how to use the spanking new twisted.web.wsgi module. Here it is. Any comments? 8<-------------------------------------------------------------------------------------------- from twisted.internet import reactor from twisted.application import service, internet from twisted.web import server, static, wsgi from twisted.python import threadpool def wsgi_app(environ, start_response): status = '200 OK' response_headers = [('Content-type', 'text/plain')] start_response(status, response_headers) return ['Hello from twisted.web.wsgi'] class WSGIService(service.Service): def __init__(self, minPoolSize=5, maxPoolSize=20): self.pool = threadpool.ThreadPool(minPoolSize, maxPoolSize) def startService(self): self.pool.start() service.Service.startService(self) def stopService(self): self.pool.stop() service.Service.stopService(self) application = service.Application('wsgi') root = static.File("www/documents") site = server.Site(root) internet.TCPServer(8080, site).setServiceParent(application) serv = WSGIService() root.putChild('wsgi', wsgi.WSGIResource(reactor, serv.pool, wsgi_app)) serv.setServiceParent(application)
Just because I've found myself doing this several times, you might give an example of how to run an existing WSGI application in twisted. I would suggest Trac, since you can just import trac.web.main.dispatch_request and pass it as your WSGI application object, plus it's something that I know I often end up using in tandem with web projects. -phil On Dec 9, 2008, at 12:38 AM, Waldemar Osuch wrote:
Hi, I'm trying to come up with an example on how to use the spanking new twisted.web.wsgi module. Here it is. Any comments?
8 < --------------------------------------------------------------------------------------------
from twisted.internet import reactor from twisted.application import service, internet from twisted.web import server, static, wsgi from twisted.python import threadpool
def wsgi_app(environ, start_response): status = '200 OK' response_headers = [('Content-type', 'text/plain')] start_response(status, response_headers) return ['Hello from twisted.web.wsgi']
class WSGIService(service.Service): def __init__(self, minPoolSize=5, maxPoolSize=20): self.pool = threadpool.ThreadPool(minPoolSize, maxPoolSize)
def startService(self): self.pool.start() service.Service.startService(self)
def stopService(self): self.pool.stop() service.Service.stopService(self)
application = service.Application('wsgi')
root = static.File("www/documents") site = server.Site(root) internet.TCPServer(8080, site).setServiceParent(application)
serv = WSGIService() root.putChild('wsgi', wsgi.WSGIResource(reactor, serv.pool, wsgi_app)) serv.setServiceParent(application)
_______________________________________________ Twisted-web mailing list Twisted-web@twistedmatrix.com http://twistedmatrix.com/cgi-bin/mailman/listinfo/twisted-web
On Tue, 9 Dec 2008 14:40:13 -0500, Phil Christensen <phil@bubblehouse.org> wrote:
Just because I've found myself doing this several times, you might give an example of how to run an existing WSGI application in twisted.
I would suggest Trac, since you can just import trac.web.main.dispatch_request and pass it as your WSGI application object, plus it's something that I know I often end up using in tandem with web projects.
This is pretty easy: twistd web --wsgi trac.web.main.dispatch_request
-phil
On Dec 9, 2008, at 12:38 AM, Waldemar Osuch wrote:
Hi, I'm trying to come up with an example on how to use the spanking new twisted.web.wsgi module. Here it is. Any comments?
Just one, I think
8 < --------------------------------------------------------------------------------------------
from twisted.internet import reactor from twisted.application import service, internet from twisted.web import server, static, wsgi from twisted.python import threadpool
def wsgi_app(environ, start_response): status = '200 OK' response_headers = [('Content-type', 'text/plain')] start_response(status, response_headers) return ['Hello from twisted.web.wsgi']
class WSGIService(service.Service): def __init__(self, minPoolSize=5, maxPoolSize=20): self.pool = threadpool.ThreadPool(minPoolSize, maxPoolSize)
def startService(self): self.pool.start() service.Service.startService(self)
def stopService(self): self.pool.stop() service.Service.stopService(self)
self.pool.stop() will block until all the threads in the pool finish whatever job they're on. This is probably okay for now, but if Twisted's WSGI server ever changes to deliver request body bytes to WSGI applications incrementally then this could cause a deadlock. In order to fix this, there needs to be a way to shut down the threadpool without blocking the reactor. This would make a good enhancement request for ThreadPool. Another approach would be to stop the HTTP port so that new requests cannot be made and then wait for all existing requests to finish before trying to stop the pool.
application = service.Application('wsgi')
root = static.File("www/documents") site = server.Site(root) internet.TCPServer(8080, site).setServiceParent(application)
serv = WSGIService() root.putChild('wsgi', wsgi.WSGIResource(reactor, serv.pool, wsgi_app)) serv.setServiceParent(application)
Jean-Paul
participants (3)
-
Jean-Paul Calderone -
Phil Christensen -
Waldemar Osuch