On Mon, Jan 23, 2023 at 11:39 AM Mark Thomas <ma...@apache.org> wrote: > > > > On 21/01/2023 14:33, Christopher Schultz wrote: > > Mark, > > > > On 1/20/23 14:11, Mark Thomas wrote: > >> Hi all, > >> > >> You will have seen I've just committed the updated priorities > >> implementation for HTTP/2. The open question is whether or not we want > >> to back-port this and if so, how far. > >> > >> Thoughts? > > > > Shall we let a release or two of Tomcat 11 go by to see if there are any > > problems reported in the wild before back-porting? > > Assuming the next release vote for 11 passes, I'd be fine with that. I'm > not sure there is much pick-up of Tomcat 11 yet. I'd probably leave a > gap between the 10.1.x to 9.0.x back-port as well. > > I'm likely being over cautious since I'd be surprised if this code was > used much in the real world anyway. Priotising requests is hard in > normal circumstances and even harder in the Tomcat case where the > content is dymanically generated and the client likely has little idea > of the relative costs of genertaing the responses. In essence, all this > code does is in the case of connection window exhaustion, it uses > priorities to determine which responses to advance when the next window > update arrives.
Yes, I have read the new priority specification, which is much nicer and does a good job at explaining everything. I agree with you: the big problem is it is still extremely powerful and at the same time figuring out how to best use it seems still as impossible as before. It has been demonstrated that h2 now lowers latency consistently in the real world (JF can finally ace his demo these days !), but it would be much harder to use the priority system to actually net a further benefit. > > Or is this the kind > > of thing that will be much easier to back-port today than, say, in > > March/April? > > The HTTP/2 is pretty consistent between all versions. I'm not expecting > any major issues back-porting this in a few months time. +1 Rémy > Mark > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org > For additional commands, e-mail: dev-h...@tomcat.apache.org > --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
