Opinionated starting point for building small/medium Rust command line applications.
The clap crate with derive feature is used as the primary description of the command line structure.
The module structure has the cli module containing the global_args.rs, mod.rs, and to_args.rs, with the cli::command module containing additional modules to handle the actual commands.
The to_args implementation and GlobalArgs fields will come in use for other desired outcomes.
The tracing and tracing-subscriber crates are used for logging. Passing --debug to the application will set the global args debug field to true which affects the tracing setup.
The tracing-subscriber configuration changes depending if it's a release or debug build.
By having the ToArgs trait implemented for our commands, we can easily relaunch the program when detected in an unelevated state.
The console handle is reused when relaunching to ensure that output goes to the terminal that the user expects, with the unelevated process exiting after the elevated one to ensure the console behaves as desired.
The eyre and color-eyre crates are used for handling error cases.
The fn main is responsible for early installation of the color-eyre behaviour.
All fallible functions return the fully-qualified eyre::Result<T> signature; we never use use eyre::Result; imports and we always qualify.
- One import per line
- Imports grouped into one big block
- See
./rustfmt.toml
These are some crates I like to use.
- itertools
- nucleo
- ratatui
- tachyonfx
- throbber-widgets-tui
- tui-textarea
- uom
- rayon
- strum
- memmap2
- humansize
- humantime
- chrono
- cloud_terrastodon_user_input
- serde
- serde_json