CODE Git documentation
CODE Git is a managed .NET 10 library for acquiring read-only repository snapshots over Git Smart HTTP protocol v2 and HTTPS. It returns repository metadata and file streams without invoking Git, checking out a working tree, or using provider REST APIs.
This folder is the documentation entry point for developers and AI contributors. Its Markdown files are included as solution items under Docs in CODE.Git.slnx.
| Document | Use it for |
|---|---|
| Getting started | Build, run the sample, and acquire snapshots |
| Architecture | Component diagram, snapshot sequence, project map, and implementation responsibilities |
| Authentication | Existing Basic/Bearer support, custom credentials, and limitations |
| Azure DevOps example | Clone an authenticated Azure Repos snapshot using Entra or a PAT |
| Public API | Client operations, models, handlers, results, and exceptions |
| Security and limits | Network trust, disk safety, resource defaults, and unsupported content |
| Development | Tests, fixtures, contribution rules, and AI orientation |
For a first contribution, read architecture and development, then the topic relevant to your change. Read security before changing transport, pack parsing, paths, or file delivery. The root README remains the short consumer overview; public types also have XML documentation in CODE.Git.Client.
Documentation describes the checked-in implementation. Authentication exists already, but interactive login, automatic token acquisition, and credential-store integration do not. Provider compatibility depends on endpoint capabilities and object-ID fetch policy; authentication alone does not guarantee selective fetching will work.