nip

nip is a git remote helper that'll put your repo's objects on IPFS - i.e. Nowhere In Particular.

Installation

Like with most Rust packages, the easiest way to install will be using Cargo:

$ cargo install nip

Usage

Important: Before you try to use nip please make sure that your local IPFS instance is running on its standard port.

Pushing an existing repo to a nip remote for the first time

$ git remote add nip nip::new-ipfs # Use a magic placeholder URL representing a new IPFS repo
$ git push --all nip # Push all refs to a brand new repo

Cloning a repo from nip

$ git clone nip::/ipfs/QmZq47khma5nP7DjHUPoERhKnfNUPqkr5pVwmS8A6TQSeN some_repo

Repo administration with nipctl (WIP)

nip comes with nipctl - a utility for nip repo administration. It's nowhere near ready yet, but you can view the list of planned features here. Suggestions for additional features are very welcome.

How it all works a.k.a. FAQ

How does git talk to nip?

nip implements what is called a git remote helper - a new remote transport backend that can be used by git for pushes and fetches of remote git repositories. In fact not so long ago the HTTP transport in git used to be a separate binary taking advantage of this API. You can read more about the exact remote helper operation in gitremote-helpers(1).

Under the hood, the stuff above means that upon a git push to or git fetch from a nip remote, git will run the git-remote-nip binary and exchange information about local/remote states via stdio. Then the binary is expected to carry out a state sync as per the specification of the push/fetch.

How does nip interact with git repos and IPFS?

Local repo

Locally, nip takes advantage of git2-rs which is a set of Rust bindings to libgit2. libgit2 is then used to scoop out or instantiate git objects - depending on whether a push or fetch operation is requested by git.

IPFS storage

For IPFS storage nip uses a fairly thin CBOR-encoded format comprised of two datatypes: NIPIndex and NIPObject. NIPIndex is what every top-level nip repo IPFS link points to and effectively the face of every nip remote - it stores information about all git objects available in a given remote as well as where branch tips and tags should resolve to. NIPObject on the other hand captures the actual git object tree topology of the repo.

Every NIPObject is comprised of two parts:

• An IPFS link to the raw bytes of the underlying git object - this data isn't inlined within the data structure to maximize data deduplication, including objects produced by different nip versions or even different IPFS git backends that choose to operate in the same manner.
• git object-specific metadata - this is done via a helper enum type where the variants contain differently arranged git hashes depending on object type:
  • commits - parent hash(es), tree hash
  • trees - children hashes (pointing to another nested tree or a blob)
  • blobs - this variant is purely symbolic, the raw bytes link is sufficient since blobs are always leaf nodes in git
  • tag objects - target object hash of the tag; only used for annotated/signed tags

A note on object tree edges in nip

An important fact about NIPObject metadata is that the references to other NIPObjects are git hashes and not IPFS ones - it is done that way so that the Rust code can check if the local git repo already contains a given git object without making any additional requests to the local IPFS node (it looks them up in the NIPIndex which is always downloaded first). Also, this practice makes the format less forgiving and therefore less prone to being incorrectly used.

How does nip intend to stay backwards-compatible?

Internally, nip prepends every serialized NIPIndex and NIPObject with a very simple 8-byte header. It starts with a b"NIPNIP" magic followed by a big-endian 16-bit number denoting the version of the data format a given object uses. This ensures that even when the serialization format is changed or even if serde is no longer used, nip will still be able to find out in time.

Development

If you'd like to hack on nip, the dev_bootstrap.sh script is where you should start. It symlinks nipctl and git-remote-nip as nipdevctl and git-remote-nipdev in ~/.cargo/bin, respectively. As a result, git will pick git-remote-nipdev for every remote that has a nipdev::<hash_or_mode> address.

Limitations

• Repo pinning and git push notifications - people interested in keeping track of remote repo's progress have no way of knowing about pushes made to it. See this issue for progress on the solution.
• Submodules - nip doesn't understand how to push/pull submodule pins yet.
• Disk space - by design local git objects need to have IPFS counterparts which are kept in your local IPFS node's data store. In practice this means that every local object pushed to a nip repo needs to be stored on your disk again in a form that IPFS understands. However, nip guarantees object deduplication for all repos you use with it, which means a given git object is stored on IPFS only once, no matter the repo it comes from.
• Object size - nip doesn't know yet how to stream objects into/out of the local repository and will attempt to load them into RAM, this increases the memory footprint substantially for repos that posess large objects. Tracked here.
• Descriptor limits - Because of improper tokio use, currently nip may exceed descriptor limits because of redundant tokio runtime instances. Tracked here. Easily solved by issuing ulimit -n unlimited just before using git with a nip repo.