Tech debt

[rustworthy]

• Use ::connect() vs ::connect_to(&str) as per thread

• Job's reserve_for is now using Duration;

• We are adding public Wokrer::is_terminated so that they can check and not run a terminated worker;

• Refactoring: using tokio::time::interval in the heartbeat thread instead of tokio::time::sleep

---

This change is 

[codecov]

Codecov Report

Attention: Patch coverage is 59.15493% with 29 lines in your changes missing coverage. Please review.

Project coverage is 67.2%. Comparing base (3c4058a) to head (cf8f757).
Report is 3 commits behind head on main.

Files with missing lines	Patch %	Lines
src/tls/rustls.rs	0.0%	13 Missing ⚠️
src/tls/native_tls.rs	0.0%	10 Missing ⚠️
src/bin/loadtest.rs	0.0%	3 Missing ⚠️
src/worker/builder.rs	75.0%	2 Missing ⚠️
src/proto/single/utils.rs	95.2%	1 Missing ⚠️

Additional details and impacted files

Files with missing lines	Coverage Δ
src/proto/batch/mod.rs	91.6% <ø> (ø)
src/proto/client/mod.rs	87.1% <100.0%> (+0.2%)	⬆️
src/proto/single/mod.rs	94.2% <100.0%> (+0.1%)	⬆️
src/proto/single/resp.rs	87.6% <ø> (ø)
src/proto/utils.rs	74.5% <ø> (-3.1%)	⬇️
src/worker/health.rs	89.1% <100.0%> (+0.4%)	⬆️
src/worker/mod.rs	85.0% <100.0%> (+1.6%)	⬆️
src/worker/runner.rs	25.0% <ø> (ø)
src/proto/single/utils.rs	99.0% <95.2%> (-1.0%)	⬇️
src/worker/builder.rs	88.3% <75.0%> (+0.4%)	⬆️
... and 3 more

... and 1 file with indirect coverage changes

[rustworthy]

@jonhoo Sorry it took me a while this time. I was moving apartment and preparing some legal docs around my expat life 😅

So I have addressed - I think so - all the threads, but this one. I was guided by the compiler and the current implementation made it happy, and you unhappy 😄 I agree it looks and feels weird, but I think I will need your help with this.

Please let me know if you while reviewing you catch other things that we better refactor now if this will be breaking.

[jonhoo]

These changes all look great, thanks! A small nit is all.

No worries with the delay — I've had a similar delay now, and will likely be gone for the next month or so between work and holidays 😅

[jonhoo]

Feel free to merge this yourself once corrected 👍

I've also now given you publish rights to faktory on crates.io — you're now a full-fledged maintainer :)

[jonhoo]

On the Reconnect thing, it's probably not ultimately all that important. My main wonder is if we can replace

impl Reconnect for BufStream<TlsStream<tokio::net::TcpStream>>
impl Reconnect for BufStream<TlsStream<proto::BoxedConnection>>

with

impl<S> Reconnect for TlsStream<S> where S: Reconnect

I'm also a little suspicious of

impl Reconnect for BufStream<TlsStream<proto::BoxedConnection>>

in the first place since that will end up returning a BoxedConnection that in turn contains a BoxedConnection, meaning an extra (unnecessary) indirection.