Marc Brinkmann

@dcuddeback That interface is deprecated. It's a bit confusing, because the `termios-rs` crate uses a confusing label for the datastructure (see https://github.com/dcuddeback/termios-rs/pull/8) I found an blog post here: https://www.downtowndougbrown.com/2013/11/linux-custom-serial-baud-rates/ ,...

> Also, a complete solution needs to deal with getting the custom baud rate from read_settings() as well. Should be fixed in 42e2bcf

Well, I can't do more at the moment, other than the PRs and issues in `serial-rs`, `termio-rs` and `ioctl-rs` that I have made so far (I would recommend starting with...

An alternative that is less flexible would be adding support for the `tcsendbreak` syscall, which achieves the above without funky bitrate changing tricks (although the required pause `MARK-AFTER-BREAK` must then...

Heya, this is a bit weird but my current practice for all of my packages is to _not_ include docs and tests inside the PyPI release tarball, but publish docs...

Same issue on elm-reactor 0.18 on Linux.

A first draft is in PR #112, for your consideration ;) Some comments beyond: > While it's impossible to cover every conceivable sophisticated attack strategy, we should send at least...

Looks good to me, so far. I'll scrap the V1 write-up in favor of this, I don't think we need to write it up any more formal?

Yes, `proptest` will run the same test every time if it is a "known regression", stored in `tests/net_dynamic_hb.proptest-regressions` for example. Running with an identical random seed and unchanged test will...

My remaining two gripes with this are that "faulty node" is no longer a meaningful term and "faulty participant" does not have the same ring to it.