#![allow(unused)] use std::path::{Path, PathBuf}; use crate::{File, State, decode, extension}; mod error { /// The error returned by [File::at()][super::File::at()]. #[derive(Debug, thiserror::Error)] #[expect(missing_docs)] pub enum Error { #[error("An IO error occurred while opening the index")] Io(#[from] std::io::Error), #[error(transparent)] Decode(#[from] crate::decode::Error), #[error(transparent)] LinkExtension(#[from] crate::extension::link::decode::Error), } } pub use error::Error; /// Initialization impl File { /// Try to open the index file at `path` with `options`, assuming `object_hash` is used throughout the file, or create a new /// index that merely exists in memory and is empty. `skip_hash` will increase the performance by a factor of 2, at the cost of /// possibly not detecting corruption. /// /// Note that the `path` will not be written if it doesn't exist. pub fn at_or_default( path: impl Into, object_hash: gix_hash::Kind, skip_hash: bool, options: decode::Options, ) -> Result { let path = path.into(); Ok(match Self::at(&path, object_hash, skip_hash, options) { Ok(f) => f, Err(Error::Io(err)) if err.kind() == std::io::ErrorKind::NotFound => { File::from_state(State::new(object_hash), path) } Err(err) => return Err(err), }) } /// Open an index file at `path` with `options`, assuming `object_hash` is used throughout the file. If `skip_hash` is `true`, /// we will not get or compare the checksum of the index at all, which generally increases performance of this method by a factor /// of 2 or more. /// /// Note that the verification of the file hash depends on `options`, and even then it's performed after the file was read and not /// before it is read. That way, invalid files would see a more descriptive error message as we try to parse them. pub fn at( path: impl Into, object_hash: gix_hash::Kind, skip_hash: bool, options: decode::Options, ) -> Result { let _span = gix_features::trace::detail!("gix_index::File::at()"); let path = path.into(); let (data, mtime) = { let mut file = std::fs::File::open(&path)?; // SAFETY: we have to take the risk of somebody changing the file underneath. Git never writes into the same file. #[expect(unsafe_code)] let data = unsafe { memmap2::MmapOptions::new().map_copy_read_only(&file)? }; if !skip_hash { // Note that even though it's trivial to offload this into a thread, which is worth it for all but the smallest // index files, we choose more safety here just like git does and don't even try to decode the index if the hashes // don't match. // Thanks to `skip_hash`, we can get performance and it's under caller control, at the cost of some safety. let expected = gix_hash::ObjectId::from_bytes_or_panic(&data[data.len() - object_hash.len_in_bytes()..]); if !expected.is_null() { let _span = gix_features::trace::detail!("gix::open_index::hash_index", path = ?path); let meta = file.metadata()?; let num_bytes_to_hash = meta.len() - object_hash.len_in_bytes() as u64; gix_hash::bytes( &mut file, num_bytes_to_hash, object_hash, &mut gix_features::progress::Discard, &Default::default(), ) .map_err(|err| match err { gix_hash::io::Error::Io(err) => Error::Io(err), gix_hash::io::Error::Hasher(err) => Error::Decode(err.into()), })? .verify(&expected) .map_err(decode::Error::from)?; } } (data, filetime::FileTime::from_last_modification_time(&file.metadata()?)) }; let (state, checksum) = State::from_bytes(&data, mtime, object_hash, options)?; let mut file = File { state, path, checksum }; if let Some(mut link) = file.link.take() { link.dissolve_into(&mut file, object_hash, skip_hash, options)?; } Ok(file) } /// Consume `state` and pretend it was read from `path`, setting our checksum to `null`. /// /// `File` instances created like that should be written to disk to set the correct checksum via `[File::write()]`. pub fn from_state(state: State, path: impl Into) -> Self { File { state, path: path.into(), checksum: None, } } }