A TUI for Rust docs that aims to improve the UX on tree view and generic code.

This tool utilizes the nightly compiler to compile JSON doc for a package. Therefore, make sure the nightly toolchain is usable on your machine.

It's recommended to have rust-docs-json too, so you can run:

rustup rustup toolchain install nightly --component rust-docs-json

# or add the component if nightly toolchain lacks it
rustup component add --toolchain nightly rust-docs-json
rustup update nightly

This tool is very immature, and far from the target for now. It only provides the basic ability to compile and cache docs, and view markdown docs.

Key shortcuts or usage help can be found via F1 key press.

More screen shots can be found here.

Roadmap

• item outline
  • expand / fold (based on module tree)
    • expand zero level items (i.e. items in root module with sub modules folded)
    • expand to first level items
    • focus on the latest module only (but with all other level modules folded)
    • expand all public items
  • features related
• doc content
  • text wrapping
  • syntax highlighting in codeblocks
  • recognize rustdoc syntax attributes on codeblocks
    • in links
    • in codeblock (default to rust, hide lines, etc)
• navigation
  • markdown outline
  • item's associated items/fields outline
• package source / DashBoard Popup
  • local
    • local registry src dirs
      • fuzzing search
      • select pkgs to compile docs and cache the artifacts in local db files
    • caches in database (json docs that have been generated will be cached in local db)
      • cache raw JSON output and compress it via xz
      • cache parsed output for faster loading and compress it via xz
      • Sorting the cache list for all items or in groups
    • local paths to Cargo.toml: low priority
  • non-local (i.e. download pkgs from the web): low priority
• configuration
  • theme: low priority
  • keybind: low priority
• fuzzing search
  • by item name
  • by all documentation contents
  • by function/method signature
    • on concrete types
    • on generic types
    • on trait bounds
  • by crate features
• generic types enhancement
  • generic type parameters
    • list concrete candidate types that meet the trait bounds
      • from within the current pkg
      • from within the caches in database
    • list the functions/methods that
      • return generic types that hold the same trait bounds
      • return concrete candidate types
    • list the function/methods that
      • accept generic types that hold the same trait bounds
      • accept concrete candidate types
  • lifetime parameters
    • variance (lack of this info in json docs, but maybe not hard to have it)
• concrete types
  • list methods in which the concrete Type and its ownership variants &Type / &mut Type is
    • receiver type
    • argument type
    • return type
• traits
  • classify trait implementors
    • by ownership (impl Trait for Type vs &mut Type vs &Type vs Box<Type>)
    • by concrete vs generic

Misc/Basics

• data access policy on crates.io
  • <index.crates.io> can be accessed without rate limits to query crates' history versions, features and dependencies
• local registry cache:
  • ~/.cargo/registry/src/ contains the official crates.io registry and other mirror/custom registries
  • ~/.cargo/registry/index/ contains the API URLs to query or download crates from these registries

id rules

Mainly steal from id_from_item_inner

[IMPL:]CRATE_ID:ITEM_ID[:NAME_ID][-EXTRA]:

• impl
  • a: for auto impls
  • b: for blanket impls
  • empty if others, like non-impls, inherent impls, normal trait impls
• name is the item's name if available (it's not for impl blocks for example).
• extra is used for reexports: it contains the ID of the reexported item. It is used to allow to have items with the same name but different types to both appear in the generated JSON.