🚃 lib for CLI utilities, printing, and error handling

axo

Last update: Apr 4, 2023

Related tags

Overview

axocli

Common code for setting up a CLI App and handling errors/printing.

Example

See examples/axoapp.rs for a walkthrough/example.

Some various interesting example invocations to play with:

# clap help
cargo run --example axoapp -- --help

# success
cargo run --example axoapp -- 5
cargo run --example axoapp -- 5 --output-format=json

# normal error
cargo run --example axoapp -- 2
cargo run --example axoapp -- 2 --output-format=json

# panic (setting RUST_BACKTRACE=1 here is also interesting)
cargo run --example axoapp -- 0
cargo run --example axoapp -- 0 --output-format=json

# clap error
cargo run --example axoapp --

What Does It Do?

First off, it handles setting up:

a tracing subscriber (currently the kind that cargo-dist uses, more work needed for oranda's use)
a miette formatter (so we can have a shared Look And Feel)
a panic handler (to get more control over output if the app panics)
(not implemented but likely in the future) a tokio runtime

It also handles printing top-level errors, notably including a json_errors mode where the error gets formatted to json and printed to stdout, while still printing a human-friendly version to stderr. This is done for both an error returned from real_main and panics. So anything expecting machine-readable output from our apps will not freak out and get something unparseable when things error/panic. It will also set the process exit code on error (with std::process::exit, on the assumption that all cleanup was done when we returned/panicked out of real_main).

It also exposes the json diagnostic formatting machinery so you can Write them wherever or turn them into serde_json::Values. This is useful for returning a larger result with diagnostics nested inside of it (say, for reporting warnings).

License

Licensed under either of

Apache License, Version 2.0, (LICENSE-APACHE or apache.org/licenses/LICENSE-2.0)
MIT license (LICENSE-MIT or opensource.org/licenses/MIT)

at your option.

Rust lib for fetching official protoc (Protocol Buffer compiler) releases

protoc-fetcher Rust library for fetching official Protocol Buffer compiler (protoc) releases, pegged to a specific version. protoc-fetcher downloads a

2 Sep 5, 2022

Uma lib para a API do Brasil API (para o Rust)

Uma lib para a API do BrasilAPI (para o Rust) Features CEP (Zip code) DDD Bank CNPJ IBGE Feriados Nacionais Tabela FIPE ISBN Registros de domínios br

6 Dec 13, 2022

Simple time handling in Rust

time Documentation: latest release main branch book Minimum Rust version policy The time crate is guaranteed to compile with any release of rustc from

680 Dec 31, 2022

A simple endian-handling library for Rust

endi Yet another endian handling library for Rust. The approach is very similar to that of byteordered crate with its Endianness enum, except that end

5 Dec 11, 2023

A cli tool to write your idea in terminal

Ideas ideas is a cli tools to write your idea in your terminal. Demo Features tagged idea, contains tips, idea, todo status switch ascii icon write yo

12 Jun 22, 2022

clone of grep cli written in Rust. From Chapter 12 of the Rust Programming Language book

minigrep is a clone of the grep cli in rust Minigrep will find a query string in a file. To test it out, clone the project and run cargo run body poem

1 Dec 14, 2021

CLI program written in Rust to anonymize DICOM files

dicom-anonymizer CLI program written in Rust to anonymize DICOM files anonymizer 0.1.0 Domenic Melcher USAGE: anonymizer [OPTIONS] FILE ARGS:

2 May 30, 2022

Rust app that creates cli quiz for you!

quizer rust app that creates cli quiz for you! Installation Download binary release or build from source using cargo Make binary executable chmod +x /

1 Sep 5, 2022

Simple CLI tool to create dummy accounts with referral links to give yourself free Plus.

Free Duolingo Plus A simple CLI tool to create dummy accounts with referral links to give yourself free Plus (max 24/41 weeks depending on whether you

23 Apr 27, 2023

Comments

testing mode

Right now I don't think the the design is well-suited for tests.

For global init in tests it's typical to have a global lazy_static that all tests try to use, ensuring global setup happens once. However we currently make setup and running a single transactional operation. Ideally we could have an alternate way of starting things that lets each test share global init that can be shared AND let each have their own "sandbox". Like it would be great if you could ask for logs that were generated by test1 without test2 messing with it.

idk how much of a pipe dream the full sandboxing stuff is, but at very least we want to be able to get the primary output/error of each test individually.
enhancement

opened by Gankra 0
tokio runtime setup

oranda wants this, and i'm sure we'll have other apps that use tokio that want to share the logic

The entered runtime could be passed as part of CliApp, this was factored into the design.
enhancement

opened by Gankra 0
log files

oranda needs this

orogene has a super complicated/robust solution:

https://github.com/orogene/orogene/blob/main/src/lib.rs#L145-L255 and https://github.com/orogene/orogene/blob/main/src/lib.rs#L295-L306 set up logging, and also automatically always logging to a log file in the cache, then if there's an error, it says "here's where the log file is, if you want it. It also keeps the most recent 5 logs for you in case you wanna go back and see.

oh and it also has all the stuff you need so you can provide the really fancy EnvFilter strings to --loglevel, instead of just warn/info/etc (--loglevel node_maintainer=trace,warn)

enhancement

opened by Gankra 0
prettier backtraces

We do some custom backtrace stuff in our panic hook (the relevant details being in panic.rs) and I don't really like the current output. It's very noisy, especially with printing the address of a frame even if we symbolicated it. That said I need to test it out in "production" environments where symbols are worse (no local debuginfo sitting in target/) to figure out what can be trimmed or not.

The situation isn't 1:1 (not using the backtrace crate) but minidump_processor has much prettier output with this more complex impl: https://github.com/rust-minidump/rust-minidump/blob/main/minidump-processor/src/process_state.rs#L737
enhancement

opened by Gankra 0