Tanguy

nim-libp2p can already be used as a hole punching server (autonat & relay are available), but cannot hole punch itself (missing the dctur for that)

Using this branch, I got another kind of race condition: "Process does not exist" It seem that in some cases, you can't use "waitForExit" just after "startProcess" finished

Sorry, when I say "blocking", I mean in the context of async, but I should say "closing is not instantaneous" So with the current setup of `nim-websock`, if an application...

Discussed at today's meeting, In the end it's just a naming misunderstanding: for me `close` should be instantaneous, and there should be a `closeWait` which actually waits for the connection...

> so it's like an "aborted return" instruction as the `yield` says "do continue afterwards" I don't think yield is supposed to "abort a return" or "do continue afterwards" Otherwise:...

> The semantics are "after a return run the finally section including its yield instructions". That seems pretty consistent and reasonable to me. Indeed. However, if an exception is raised...

> Why we should keep references to other projects if this projects do not keep references to chronos? libp2p mentions chronos as part of the [first tutorial on how to...

@arnetheduck @cheatfate other remarks on this PR? I would like to have at least the website up, and then we can work on better documentation content

Seems to be a bearssl limitation: https://github.com/status-im/BearSSL/blob/acc70b1be60a6f321e2da618cd35d901b1a598a4/src/x509/x509_minimal.t0#L750-L756 The presence of "server_name" is used as a flag of "verify server name", so we can't send the server_name without verifying it

> this error should not happen in the wild. We've seen it a few times, for instance on macos or whatever > Thus, while this could result in a busy...