Maximilian Hils

> (Basically what I said before, by keeping the original data around.) Maybe I'm misunderstanding this, but this sounds like a (partial) rewrite that takes quite a bit of effort....

> How do I tell the tests to use the updated `mitmproxy_rs`? This currently requires us to ship a new release of mitmproxy_rs. There is an upper bound in https://github.com/mitmproxy/mitmproxy/blob/main/pyproject.toml#L45...

Regarding this PR: Could we just expose a `get_system_dns_server()` from Rust (that uses https://docs.rs/hickory-resolver/latest/hickory_resolver/system_conf/fn.read_system_conf.html internally) instead? We can then treat DNS mode just like we treat reverse DNS mode. (We...

> > that uses https://docs.rs/hickory-resolver/latest/hickory_resolver/system_conf/fn.read_system_conf.html internally > > This is available for UNIX only The docs are misleading here - it's available for Windows as well if you look at...

https://github.com/mitmproxy/mitmproxy/pull/6975#discussion_r1676522249 (unresolved comment above), otherwise LGTM :)

No worries, this is why we squash! :)

Thank you for the report. Do you have any better ideas how this can be reproduced? What happens if you run it with a RUST_BACKTRACE env var set to 1?

Thanks! This looks good and bad at the same time - it looks like the redirect task is crashing, but we don't capture the process output from that. This is...

We used to have eslint, but apparently I removed it in https://github.com/mitmproxy/mitmproxy/commit/0e9206ceb11b1c4a3ea298e6b5bda54b9baf08ee as part of the transition to esbuild. IIRC it wasn't that easy to set up with TypeScript back...

When adopting eslint, we should also run it with `--fix` as part of our autofixer. :)