On 21 July 2013 02:30, Some Developer <someukdevelo...@gmail.com> wrote:
> On 20/07/2013 14:02, Curtis Maloney wrote: > >> I'm more or less building atop flup as it is, however I plan to shed >> anything not related to FastCGI. >> >> For me it's a chance to get down and dirty with raw protocols again... I >> do agree there is a shorter path to just applying Django's "fixes" to a >> fork of flup. >> > > If you really want to get down and dirty with raw protocols have you > considered an implementation that used epoll or kqueue for the networking? > Combined with nginx as the front end HTTP server it should (theoretically) > result in a huge increase in performance. > Well, that remains to be seen... it's early days yet, but I hope to be able to support various work models. > Basing the work on one of the Python async networking libraries (Twisted, > Eventlet, gevent etc) should result in better performance and might well > make FastCGI competitive with uWSGI and WSGI again. I really think that > flup is the bottleneck when it comes to using FastCGI with Django. Having > said that I haven't had a chance to look at the Django side of the code yet > so there might be some optimisations that are possible there as well. I wasn't aware there was a particular performance issue, but I'll certainly keep it in mind. Mostly, I was hoping to raise a new project to bear the burden of maintenance, instead of Django core. From what I've seen of late there's various parts of the code which could be simplified if core assumes it's only ever a WSGI app to be published. -- Curtis -- You received this message because you are subscribed to the Google Groups "Django developers" group. To unsubscribe from this group and stop receiving emails from it, send an email to django-developers+unsubscr...@googlegroups.com. To post to this group, send email to django-developers@googlegroups.com. Visit this group at http://groups.google.com/group/django-developers. For more options, visit https://groups.google.com/groups/opt_out.