status.md

Status

Repository::status performs the same conceptual comparisons as Git status:

HEAD tree  ->  index  ->  worktree
              staged     unstaged

The result is sorted by raw path bytes. Each StatusEntry independently reports an index change and a worktree change as added, modified, deleted, type-changed, or unmerged. Untracked files are worktree-only entries. An unborn HEAD is treated as an empty tree.

# use git_rs::{InitOptions, MemoryFileSystem, Repository, StatusOptions};
let repository = Repository::init(
    MemoryFileSystem::new(),
    "project",
    &InitOptions::default(),
)?;
let status = repository.status(&StatusOptions::default())?;
assert!(status.is_clean());
# Ok::<(), git_rs::Error>(())

Worktree comparisons hash file or symlink bytes when determining modifications; stat-cache data is not treated as proof of equality. This prioritizes correctness for adapters whose timestamps or inode fields are unavailable or have coarse resolution. Assume-valid and skip-worktree entries retain their Git semantics and are omitted from worktree comparison.

Gitlink directories are treated as tracked boundaries rather than recursively reported as untracked content. Submodule HEAD/dirty-state inspection will be provided by the submodule layer.

Git source comparisons

• wt-status.c organizes index-versus-HEAD and worktree-versus-index changes.
• diff-lib.c:run_diff_index defines staged comparison behavior.
• diff-lib.c:run_diff_files defines worktree comparison and unmerged stages.
• dir.c:read_directory defines untracked traversal boundaries.
• read-cache.c:ce_match_stat_basic informs mode and gitlink comparisons.