I have a couple of PRs up which I hope will be considered for inclusion in 0.8. One fixes a straight up https://github.com/pedestal/pedestal/pull/958, the other https://github.com/pedestal/pedestal/issues/955 for the interceptor route.
Sounds great! I’ll review them Tuesday.
I've also noticed a strange behaviour of :fast? io.pedestal.service.resources/resource-routes both during dev and from a jar: if I reload my page in a browser (with caching disabled) ~3 times (< 40 requests per reload), resource requests hang, never completing. When this happens I cannot reload the page again, but must close and re-open it. I can repro this reliably and it goes away with :fast? false.
I'm seeing a similar symptom with rc-2. If I open dev tools->network in a Chrome browser, I see sporadic "pending" requests that do not seem to get fulfilled. I'm not sure that it is just for resources. I've not done a great deal of testing, but I haven't been able to reproduce this issue in rc-1. I'm gonna downgrade to rc-1 for now.
Am using the jetty connector, btw.
Cljdoc issue: https://github.com/cljdoc/cljdoc/issues/1081
I think we might be well served by moving Jett back to 12.0.x from 12.1.
I think that's likely. I've verified that the get handler seems to return without issue. And in case it's relevant: I've determined only one file (> 3 MB) per page load actually qualifies for streaming.
This is troubling; have you been able to extract a stack trace from when the request hangs?
Seems like the jetty 12.1 migration guide is, at the time of this writing, https://github.com/jetty/jetty.project/issues/13498. I see https://github.com/jetty/jetty.project/pull/13376/files#diff-e1320510d79afa6cf58da59133607ae0a6c2ff3d4111329f5ead15f5016990d9. Could that be relevant to our issue?
No. Dev interceptors are not picking anything up. I'll try to poke around myself.
Might be hanging inside Jetty thread rather than Pedestal, ie after Pedestal sets up async response.