Johannes Rudolph
Johannes Rudolph
It's not about the problems it solves but about how much time is saved by preventing to execute extra tests. But I agree about the complexity, so if we can...
> @jrudolph @mdedetrich do we think this is needed for 1.0.0-RC2? No, this is just an improvement to avoid overlooking issues in those modules in the future.
There are several complications here: * our HTTP/2 client does not allow send invalid URIs (`Raw-Request-URI` not yet supported) * there's currently no way to signal non-fatal parsing errors (should...
Yes, this is a bug but only a minor issue. Here's a stack trace (though, wrapped with another exception to understand how that exception can be triggered): ``` [info] java.lang.IllegalArgumentException:...
A simple but lazy fix would be to reject/ignore `Accept-Charset` headers like that if we cannot do anything with them. A more involved and potentially performance-impacting fix would be to...
This basically means a rewrite of the complete backend, including * TCP (QUIC reimplements many TCP algorithms on top of UDP) * TLS (QUIC does not stack on top of...
For triaging it would be good to gather good motivation about why to do this in the first place. Also, whether client or server side might be more important. In...
Weird that it expects port 1025 to be available, maybe something to fix... Not sure if someone already tested if the build runs on JDK 20. Maybe you can try...
Can you check again after the `h2spec` fix, @lomigmegard ?
Ah, forgot that even the newest version does not support Arm yet.