euri10
euri10
it's not about uvicorn reload strategy, it's about the fact our UvicornWorker is disabling a signal that gunicorn uses to reload its workers and trigger logrotate, see https://docs.gunicorn.org/en/latest/deploy.html?highlight=logrotate#logging
I've no idea sorry !
so if I follow correctly the stream of things here we have `def shutdown` in the ws protocols or http protocols being called in server.py by ``` # Request shutdown...
it does not seem that simple to me, but again take this comments with a grain of salt, it's quite hard to follow the flow of things I find here...
sorry for the long time @aviramha , if @florimondmanca could also take a lookk that would be awesome, my little brain tends to collapse easily in callbacks
> I think this is looking good. > > @euri10, any opinion by chance? > > @MarioIshac Would it be possible to add a test for this, eg by passing...
I have kind of the feeling this could be resurrected for 3.10
Yeah I wasn't even aware of this, after some digging this exists : https://github.com/bjoluc/pytest-reraise should we use this ?
Thanks for the detailed report @stefanw . What would help would be to know if this happens without gunicorn, just to make sure we throw that out of the equation....
seems like the exception check was added in https://github.com/encode/uvicorn/commit/3cab13283c849037e67b2c3afd64c57da3757f2c#diff-5f163724db1eaba811cc5e4ed55e529af488e351ee52ccfb7c1ad0329a8f0d92R65-R66 and I dont get the rationale given the context, I know it's been a while but do you remember why @almarklein...