v0.53 release

[thomaseizinger]

Tracking issue for the upcoming v0.53 release to make sure we don't forget anything.

PRs to include in final patch release

Beta Give feedback

1. fix(swarm): prevent overflow in keep-alive computation #4559
   send-it +1
2. fix(upnp): do not return failure events repeatedly #4569
   send-it +1
3. feat(websocket): add WebSocket for WASM environments #4102
   send-it +1
4. feat(request-response): deprecate Config::set_connection_keep_alive #4029
   internal-change send-it +2
5. feat(swarm): deprecate KeepAlive::Until #4656
   send-it +1
6. fix(kad): re-export NodeStatus #4645
   send-it +1
7. fix(SwarmBuilder): prioritize relay, then websocket, then any other #4672
   internal-change send-it +2
8. feat(identity): make rand optional #4349
   internal-change send-it +2

Tasks

Beta Give feedback

1. Mention transition to tracing in release notes on GitHub
2. Mention API improvements like PollParameters & FromSwarm
3. https://github.com/libp2p/rust-libp2p/milestone/7
4. Remove all deprecated items
5. Mention more aggressive KeepAlive
6. feat(swarm): add #[non_exhaustive] to key enums #4581
   internal-change send-it +2
   

[thomaseizinger]

Changelog: I added #4672 which attempts to fix a quite drastic bug in the new SwarmBuilder.

[thomaseizinger]

@mxinden @umgefahren What should we do about the API changes to Transport in regards to this release? I don't really want to block this release on all of AutoNATv2 being implemented but we also said to not merge the changes until we have validated the API.

Can we somehow get the best of both worlds and say with high certainty that the API designed in #4568 is what we need, merge that as part of v0.53 and ship AutoNATv2 as a patch-release then?

[umgefahren]

We could, but we'd have to do all the necessary refactors over all transports.
However I suggested first testing if the Transport trait updates work with the AutoNATv2 implementation. I would leave the decision to @mxinden and @thomaseizinger

[thomaseizinger]

However I suggested first testing if the Transport trait updates work with the AutoNATv2 implementation. I would leave the decision to @mxinden and @thomaseizinger

I am in favor of that too. It will mean that likely, AutoNATv2 will have to wait for the next breaking release though. There are a couple more PRs open that I'd like to merge but otherwise it is pretty much ready to go out.

The v0.52 release dragged on for about a month which I really want to avoid this time.

[umgefahren]

I'm fine with that, depending on when the next breaking release is scheduled.

[mxinden]

I suggest we don't block v0.53 on AutoNATv2. I suggest we aim for AutoNATv2 and related changes in v0.54.

[dhuseby]

@thomaseizinger once this goes out should I update #4371 to ship as a non-breaking change? Or should we delay a day or two while I get it cleaned up to include in this release?

[thomaseizinger]

@thomaseizinger once this goes out should I update #4371 to ship as a non-breaking change? Or should we delay a day or two while I get it cleaned up to include in this release?

Best to coordinate with @StemCll what their plans are on working on it.

I don't see any reason to block the release on it. I am writing the release notes today and then we should be good to go :)

[thomaseizinger]

I've drafted some release notes here: https://hackmd.io/@thomaseizinger/SkA-nMG7a/edit

Feedback is welcome, feel free to just edit it :)

@mxinden Once #4794 is merged, feel free to cut a release with the above notes.

[mxinden]

I've drafted some release notes here: https://hackmd.io/@thomaseizinger/SkA-nMG7a/edit

Very well written. Lots of attention to details. Made a couple of minor changes. Thank you @thomaseizinger. This will be an entertaining read for users.

@mxinden Once #4794 is merged, feel free to cut a release with the above notes.

Will do 🎉

[thomaseizinger]

Couple of learnings from the making the release post (I'll send a docs PR later):

• Should not forget to check the "create discussion for release" button at the bottom of the page (see https://github.blog/changelog/2021-04-06-releases-support-comments-and-reactions-with-discussion-linking/).
• The markdown inside the release notes is not "commonmark", meaning single line breaks are actually rendered as line breaks and thus need to be removed for the text to flow nicely.
• We should directly tag people using @ instead of writing markdown links as I did in the notes above.