On Fri, Mar 17, 2023 at 12:03 PM Arnout Engelen <enge...@apache.org> wrote: > I haven't refreshed my memory on the details, but it is possible the > only reason that there are two Function types there is because we > wanted to change it, but had to introduce a second type to avoid > breaking (binary?) compatibility in code already using the 'old' > variation.
That's how I vaguely remember it as well. That's also backed by how in Seeta's PR above the issue was fixed by moving from the long deprecated pekko-http bindAndHandleAysnc API which used the old interface to the modern variant which uses the new one. > Do we want to 'clean up' such situations in Pekko? Does that need to > happen before our first release or can we keep some 'stabilization > period' before we start promising binary compatibility? I think we want to clean up but not for 1.0.0 mostly because of our decision to keep 1.0.0 as close to Akka as possible (but also because doing it comprehensively is quite some work). I'd say for 1.1.0 we should not strive for strict binary compatibility with 1.0.0 but do whatever is needed (within reason) to clean up old cruft (Akka 2.5 stuff, deprecations, maybe other removals as discussed before etc). Johannes --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@pekko.apache.org For additional commands, e-mail: dev-h...@pekko.apache.org
