So I guess result is that there is probably more users which will be interested and some of them can try to make some funding for this job. So next question is how much time it can take and how much funding this project can require? I am not sure about crowdfunding services as they tend to require more time so I guess just simple donations or invoices from Unbit can with work easier way.
Btw I think there is quite a lot of users which use https/http or fast routers as they are much simpler to setup if you already have uwsgi installed. st 17. 10. 2018 v 20:44 odesÃlatel Aleks <al-uw...@none.at> napsal: > Hi. > > Am 17.10.2018 um 16:13 schrieb Roberto De Ioris: > > > >> Hello guys, > >> > >> I would like to initiate again discussion about supporting HTTP2 router > as > >> SPDY cannot be used and HTTP2 is missing. I guess there was previously > >> some > >> talk about this, but I think there was lack of interest. > >> > >> If there will be more interest > we were thinking, we could ask > developers > >> to restart work on this and if we could also try to make some funding > >> project for this. > >> > >> Reason, why we want this is that having NGINX makes our stack more > >> complicated and I really do not like that for each feature I need > another > >> service. > >> > >> Also this could make uWSGI more interesting option again and make it > more > >> promising for users. > >> > >> -- > >> Regards, > >> > >> Zdenek > >> Web: www.pripravto.cz > >> _______________________________________________ > >> > > > > Hi, yes the SPDY code is basically useless (note to me and Riccardo: time > > to remove it ;), and i constantly underestimate how many users runs the > > http/fast routers. HTTP/2 (albeit based on SPDY) is pretty different, and > > implementing it in a solid way would require lot of manpower. If some > > company want to sponsor its development (like Lincoln Loop with the > > pyuwsgi stuff) can drop a mail to i...@unbit.it. > > As I followed the nginx H2 implementation and follow the haproxy H2 > implementation I can fully agree that this isn't a easy task. > > Even the h2 implementation of the two popular servers there are still some > bugs > based on client behavior. Maybe it helps to learn from the current > implementations for the own implementation. > > http://git.haproxy.org/?p=haproxy.git;a=blob;f=src/h2.c;hb=HEAD > http://hg.nginx.org/nginx/file/tip/src/http/v2 > > There is also a "new" player in the h2 game called h2o > > https://h2o.examp1e.net/ > https://github.com/h2o/h2o/tree/master/lib/http2 > > As I plan to move to uwsgi for php to get rid of php-fpm I would also > prefer > that uwsgi have the HTTP2 feature. > > @Roberto: Do you have any Idea about time and money to add this feature in > the > uwsgi? > > > Just as a side note, all of the new features (or massive refactoring) of > > the uWSGI code in the last 3 years have been sponsored by various > > companies while Unbit simply manages commercial support. This is why the > > release cycle slowed down. > > > > Personally my focus now (with now > 1 year :) is on asyncio (it is not a > > mistery i have never been a fan of the WSGI standard), but i am still not > > sure a cooperation with asyncio and uWSGI can be made without heavily > > refactoring uWSGI itself. > > > > Thanks > > Best regards > Aleks > _______________________________________________ > uWSGI mailing list > uWSGI@lists.unbit.it > http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi >
_______________________________________________ uWSGI mailing list uWSGI@lists.unbit.it http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi
