Struct GitContext

pub struct GitContext<'a> { /* private fields */ }

Holds the shared context needed for publishing Atoms.

Implementations

impl<'a> GitContext<'a>

pub fn tree_search(&self, path: &Path) -> GitResult<Option<Entry<'a>>>

Helper function to return an entry by path from the repo tree

Errors

This function will return an error if the call to [gix::object::tree::Tree::lookup_entry] fails.

pub fn transport(&mut self) -> &mut Box<dyn Transport + Send>

return a mutable reference to the transport

impl<'a> GitContext<'a>

pub async fn await_pushes(&self, errors: &mut Vec<Error>)

A method used to await the results of the concurrently running Git pushes, which were offloaded to a seperate thread of execution of Tokio’s runtime.

An errors that occurred will be collected into a Vec.

pub fn tree(&self) -> Tree<'a>

Return a reference to the git tree object of the commit the Atom originates from.

pub fn remote(&self) -> Remote<'a>

Return a reference to the git remote.

Trait Implementations

impl<'a> Publish<Root> for GitContext<'a>

fn publish<C>( &self, paths: C, remotes: HashMap<AtomTag, (Version, ObjectId)>, ) -> Vec<GitResult<GitOutcome>>

where C: IntoIterator<Item = PathBuf>,

Publishes atoms.

This function processes a collection of paths, each representing an atom to be published. The publishing process includes path normalization, existence checks, and actual publishing attempts.

Path Normalization

• First attempts to interpret each path as relative to the caller’s current location inside the repository.
• If normalization fails (e.g., in a bare repository), falls back to treating the path as already relative to the repo root.
• The normalized path is used to search the Git history, not the file system.

Publishing Process

For each path:

1. Normalizes the path (as described above).
2. Checks if the atom already exists in the repository.
   • If it exists, the atom is skipped, and a log message is generated.
3. Attempts to publish the atom.
   • If successful, the atom is added to the repository.
   • If any error occurs during publishing, the atom is skipped, and an error is logged.

Error Handling

• The function aims to process all provided paths, even if some fail.
• Errors and skipped atoms are collected as results but do not halt the overall process.
• The function continues until all the atoms have been processed.

Return Value

Returns a vector of results types (Vec<Result<PublishOutcome<T>, Self::Error>>), where the outter result represents whether an atom has failed, and the inner result determines whether an atom was safely skipped, e.g. because it already exists..

type Error = Error

The error type returned by the publisher.

type Id = ObjectId

The type representing the machine-readable identity for a specific version of an atom.

fn publish_atom<P: AsRef<Path>>( &self, path: P, remotes: &HashMap<AtomTag, (Version, ObjectId)>, ) -> GitResult<GitOutcome>

Publish an Atom.