I didn't try anything other than what I implemented. It worked and was
quite simple, so I left it at that.
Bill Eldridge wrote:
> Ian Bicking wrote:
> > On Sat, 2002-05-18 at 11:25, Geoffrey Talvola wrote:
> > > - I modified it to use a ThreadedHTTPServer class I found
> in an old
> > > comp.lang.python post. So now it actually handles
> requests on separate
> > > threads, so it will support concurrent requests.
> Don't know if you ran across this thread, which answers
> a similar issue with:
> : Fred> Off to a bad start, I made the following
> simple-minded change
> : Fred> in class
> BaseHTTPRequestHandler(SocketServer.StreamRequestHandler) :
> : Fred> in def handle(self) :
> : Fred> the last line : method() # generally self.do_GET()
> : Fred> becomes : thread.start_new_thread(method, () )
> :I think you can just create a suitable subclass of
> :In my XML-RPC server, I did:
> : class GenericServer(SocketServer.ThreadingMixIn,
> : """generic XML-RPC server class"""
> : ...
> :Each request to my server is handled in a separate thread.
I didn't look at your code, just doing a knee-jerk reaction.
> I suppose this should get rid of the problem of blocking requests that
> someone mentioned on the discuss list.
> > - I also removed support for POST requests without a content-length
> Okay... I'm realizing that input should actually be read for any method,
> if a content-length exists. PUT is the the other traditional method,
> but many DAV methods take content -- HTTPAdapter doesn't need to
> understand DAV, it just needs to know to read data when content-length
> is present. I guess just taking out the line "if self.requestMethod ==
> 'POST':" would be enough.
> Hundreds of nodes, one monster rendering program.
> Now that's a super model! Visit http://clustering.foundries.sf.net/
> Webware-discuss mailing list
Radio Free Asia
Hundreds of nodes, one monster rendering program.
Now that's a super model! Visit http://clustering.foundries.sf.net/
Webware-discuss mailing list