Skip to main content

FileCategorizer

limnifs_write::file_categorizer

Trait FileCategorizer 

Source 
pub trait FileCategorizer: Sync + Send {
    // Required methods
    fn name(&self) -> &'static str;
    fn categories(&self) -> &'static [&'static str];
    fn categorize(&self, path: &Path, data: &[u8]) -> Option<Categorization>;

    // Provided method
    fn first_byte_hint(&self) -> Option<&'static [u8]> { ... }
}
Expand description

One file-level categorizer.

Implementations should be:

  • Pure-functional: same input → same output, no I/O.
  • Deterministic: no clocks, no RNG, no system state.
  • Cheap to refuse: header parsing should bail on the first mismatched magic byte, not scan the whole file.

Categorizers are tried in registration order. The first one to return Some(Categorization) wins; later categorizers are not consulted. Order matters: register specific categorizers before generic ones.

Required Methods§

Source

fn name(&self) -> &'static str

Unique name for logging/diagnostics.

Source

fn categories(&self) -> &'static [&'static str]

Categories this categorizer can emit. Used for diagnostic dumps; does not affect dispatch.

Source

fn categorize(&self, path: &Path, data: &[u8]) -> Option<Categorization>

Categorize a file by its path and full contents.

Returns Some(Categorization) if this categorizer claims the file, None to defer to the next categorizer in the registry (or to the FastCDC fallback path).

Implementations should not read the file from disk — data is already in hand. Path is provided for extension-based hints when magic-byte detection is ambiguous.

Provided Methods§

Source

fn first_byte_hint(&self) -> Option<&'static [u8]>

The set of first bytes that this categorizer can possibly match. The registry uses this to skip categorizers without a function call when data[0] isn’t in the set.

Return None (default) to opt out of the early-exit optimisation — the categorizer is always tried. Return Some(&[bytes]) to enable early-exit: the registry checks data[0] and skips this categorizer if it’s not in the set.

Example: ELF categorizer returns Some(&[0x7F]) — it can only match files whose first byte is 0x7F.

Dyn Compatibility§

This trait is dyn compatible.

In older versions of Rust, dyn compatibility was called "object safety".

Implementors§