cool! it's in good shape, but let me run it for an extra day on my server to make sure it's working flawlessly, and I'll commit it. is there a policy on tests when committing such stuff?
On Mon, Feb 28, 2011 at 6:51 PM, Bryan Duxbury <[email protected]> wrote: > I'd have no problem committing this, if it's in good shape. Go ahead and > open a ticket. > > On Mon, Feb 28, 2011 at 8:26 AM, Dvir Volk <[email protected]> wrote: > > > Hi all, > > > > I've written a python preforking thrift server, based loosely on > > TThreadPoolServer. > > This server has both the advantage of avoiding GIL single CPU binding, > and > > of voiding the overhead of forking a new child for each client. > > > > It's of course much faster than TThreadPoolServer, and in situations > where > > clients connect and disconnect all the time and don't issue a lot of > > requests per connection, also also much faster than TForkingServer. > > > > If anyone is interested, or if the core developers want this class > > contributed back, I'd be happy to post it. > > > > Thanks > > > -- Dvir Volk System Architect, Do@, http://doat.com
