stormshield-frb

Hi everyone. Sorry to barge in 3 days later, but we just saw this PR and might have some interesting thoughts to share. We totally agree with the initial issue...

Hi @thomaseizinger, you're welcome :wink: Glad you found it interesting. Indeed, it might be too much by default, and I do agree with your mantra. However, quoting you from an...

> have a simple state variable that tracks, whether we've already successfully run at least 1 bootstrap. Absolutely. This is a good idea. We actually do that in our code...

> Exponential backoffs are useful if simply having time pass might fix the problem (e.g. a remote being overloaded). In our case, all we need to wait for is another...

> As @PanGan21 said, I think it makes sense to focus this PR on (1) & (2). Yes, I totally agree. Part `3` can always be done later if needed,...

Since I did not know how to contribute to your work @PanGan21 without breaking everything or duplicating PR, I did a separate branch in my fork to showcase what I...

As @stormshield-pj50 mentioned some time ago, we had the same problem. We have resolved it in our fork with a patch allowing us to define the following fields constants: https://github.com/libp2p/rust-libp2p/blob/1aa016e1c7e3976748a726eab37af44d1c5b7a6e/protocols/mdns/src/lib.rs#L51-L57

I've updated the PR and fixed some of the checks but I cannot manage to find why the `semver` check is failing sorry :/

About moving away from long-lived stream, will it not be a costly to open and close streams with `TCP` or `Yamux` for example ? I understand that it is almost...

> This looks a bit hacky, wouldn't it be better to modify the bootstrap `Status` instead (e.g [`poll_next_bootstrap`](https://github.com/libp2p/rust-libp2p/blob/e5c072c04c1e8ae5899b46eb17bdf9a32c4eaeee/protocols/kad/src/bootstrap.rs#L95))? I'm not sure to understand what you mean. `on_started` and `on_finished` are...