spark repo — Command Spec

Manage multiple GitHub repositories under a directory via a registry file (not submodules).

Parent

spark repo

No arguments, no flags.

Registry file format:

# Project registry
# Synced by spark repo scan

projects:
    - name: spark-cli
      repo: https://github.com/variableway/spark-cli.git
      path: tooling/spark-cli
      desc: Spark CLI tool

spark repo scan

Recursively scan a directory for Git repositories (only those with an origin remote) and write them to registry_<folder>.yaml in the current directory.

spark repo scan [folder-name]

No flags.

Behavior:

• Recursively locates .git directories and .git files (including gitdir: redirection), parsing the url under [remote "origin"]
• Skips hidden directories and node_modules/venv/.venv/__pycache__/dist/build
• SSH URLs are normalized to HTTPS form before being stored
• The registry file name uses the base name of the directory’s absolute path: registry_<folder>.yaml
• Re-scanning merges with the existing registry: matches by path first, then by repo URL, preserves old name/desc, drops entries whose directory no longer exists, and appends newly discovered repos

spark repo clone

Clone repositories from a registry file into the folder directory. The target folder name is derived from the registry file name (registry_<folder>.yaml → <folder>).

spark repo clone [-r <repo-name>] -f <registry-file>

No arguments.

Behavior:

• Skips when .git already exists or the directory exists (SKIP ...)
• Runs git clone <repo> <target> for each repository
• Prints a cloned / skipped / failed summary; returns an error when failed > 0

spark repo list

List the repositories declared in a registry file (name, remote URL and relative path).

spark repo list -f <registry-file>

No arguments.

Output: one name<TAB>repo<TAB>path line per repository; prints No repositories in registry. when empty.