Also quale is not maintained either, would be nice to replace it with which.

Thank you for the nice suggestion!

I'll change from quale to which.
byte-unit seems to be nice to replace unbytify.

I can't imagine the benefit to make dockworker optional.
Why do you think it is better that dockworker is optional?

I can't imagine the benefit to make dockworker optional.

1. The docker is not really working with cgroupsv2 (which is default for Fedora)
2. podman is what we use in Fedora and recommend everyone to use it instead of docker
3. That crate has dependency on hyper 0.12 which is not packaged anymore (we switched everything to 0.13), so it is making harder to package procs in Fedora

I understand it.
I'll add a feature for dockwolker.

I plan the feature is default for other distributions and macOS and can be disabled for Fedora.
Is it OK?

just create a cargo feature for it and enable it on macOS builds.

Each packaging script can adjust features for each distribution,
but I think that user which install by cargo install will forget to add non-default features.
So I want to enable it by default.

Sure, any way works for me. Just need to be able to build without it :)

I released v0.9.11.
Now build without docker can be done by cargo build --no-default-features.

@dalance one more thing, would it be feasible to replace getch with something else? That dependency is not really maintained.

I can't find the replacement of getch.
Is this critical problem?
If so, I can fork and maintain it.

No, not critical. I've already managed to package it in Fedora. So it was more like an RFE to replace it with something else.

That said, I think all missing bits are now packaged in Fedora and I can get procs properly there.

Though there is one test which fails:

---- tests::test_run_all stdout ----
thread 'tests::test_run_all' panicked at 'called `Result::unwrap()` on an `Err` value: Error { inner: ErrorInner { kind: Custom, line: Some(190), col: 0, at: Some(2850), message: "unknown variant `Docker`, expected one of `Command`, `ContextSw`, `CpuTime`, `Eip`, `Esp`, `Gid`, `GidFs`, `GidReal`, `GidSaved`, `Group`, `GroupFs`, `GroupReal`, `GroupSaved`, `MajFlt`, `MinFlt`, `Nice`, `Pid`, `Policy`, `Ppid`, `Priority`, `Processor`, `ReadBytes`, `RtPriority`, `Separator`, `ShdPnd`, `Slot`, `SigBlk`, `SigCgt`, `SigIgn`, `SigPnd`, `Ssb`, `StartTime`, `State`, `TcpPort`, `Threads`, `Tree`, `Tty`, `UdpPort`, `Uid`, `UidFs`, `UidLogin`, `UidReal`, `UidSaved`, `UsageCpu`, `UsageMem`, `User`, `UserFs`, `UserLogin`, `UserReal`, `UserSaved`, `Username`, `VmData`, `VmExe`, `VmHwm`, `VmLib`, `VmLock`, `VmPeak`, `VmPin`, `VmPte`, `VmRss`, `VmSize`, `VmStack`, `VmSwap`, `Wchan`, `WriteBytes`", key: ["columns", "kind"] } }', src/libcore/result.rs:1188:5

Thanks.
I fixed the failure.

Why is the nix crate outdated?
I don't know if this will fix my problem but procs is broken on aarch64 linux musl:

error[E0425]: cannot find value `MAP_32BIT` in crate `libc`
  --> /host/cargo/registry/src/github.com-1ecc6299db9ec823/nix-0.11.1/src/sys/mman.rs:53:9
   |
53 |         MAP_32BIT;
   |         ^^^^^^^^^ not found in `libc`
   Compiling url v1.7.2
error: aborting due to previous error

It seems to be an unfixed issue of nix.

nix-rust/nix#951

You can build without nix by cargo build --no-default-features because nix is used by docker feature.

Sorry. It seems to be fixed at nix v0.12.0.
I'll request nix version up to the upstream crate.

@dalance You never updated nix (still on 0.11.1)