Repositories

Bringing in a GitHub repository at one commit, browsing and searching its code, and keeping up with newer commits.

A GitHub repository can be a source like any paper. The workbench stores it at one commit, and records it as software you can cite at that commit.

Bringing one in

1. Choose a folder, then press Add a repository.
2. In Repository, paste its page address, its clone address, or just owner/name.
3. Optionally, name a branch, tag or commit. Left empty, the default branch is used.
4. Tick Search its code as well as its documentation if you want code search.
5. Press Bring it in.

The repository is downloaded only when you ask, and only then. It goes into its own folder under a top-level Repositories folder, which opens in the Code view.

The Code view

• A breadcrumb trail, Download all (.zip) and Show as documents.
• Where it came from: name, owner, commit, commit date and licence.
• A file browser, and a viewer with Copy, Download and As a document. The README is shown formatted.
• Search its code / Stop searching its code. Code is found by its words and by the names of functions and classes, so IsLeapYear, is_leap_year and leap year all match. With the code model installed it is found by meaning too (see Searching).

Newer commits

Check for updates looks for a newer commit on the default branch. The desktop workbench also checks every repository once a day by itself. When there is one:

• Bring in the new commit adds it beside the old one. The old commit stays as it is, with everything that refers to it.
• See what changed opens the comparison on GitHub.

When a repository is held at several commits, the folder tree marks each one latest or older. The start page lists repositories with A newer commit to bring in.