Francesco Nigro

@vietj how this look bud?

@vietj I've addressed your comments and added some tests for the additional unsynchronized methods.

@vietj this is fine than? So we can bring it in the next quarkus release

FYI: https://github.com/quarkusio/quarkus/blob/main/independent-projects/resteasy-reactive/server/vertx/src/main/java/org/jboss/resteasy/reactive/server/vertx/ResteasyReactiveOutputStream.java#L53 and check ` Http1xServerRequest` as well, which is still synchronizing over `conn` -> the connection

@vietj the best way would be to try running some of the tests on Quarkus which relies on it, by using a vertx which version relax these properties (ie this...

> which others do you mean ? Http1xServerRequest, for example

Let's wait till you sync with @cescoffier about the new API and the time to backport it: this one is pretty small and contained, so, depending for how long we...

@vietj if merged I have to fix https://github.com/quarkusio/quarkus/blob/4e6908f522190127a28af7c96806c06cc0b0a60e/extensions/vertx/runtime/src/main/java/io/quarkus/vertx/core/runtime/VertxLocalsHelper.java#L25 which is wrongly allocating a local map, although not really required.

It is a nice approach @cescoffier which will give performance and flexibility based on the usage, I like it!

I have some mixed feeling after re-reading what @cescoffier rightly said, and remind me about something I've worked recently eg https://github.com/quarkusio/quarkus/pull/38761 which has a very similar problem ie at build...