Working with git worktrees

[greened]

Just getting started with hatch after using setuptools.

I am not sure how to integrate hatch's project notion with the use of git worktrees. Typically, my workarea looks like this:

project
  main
  worktree1
  worktree2
  ...

In local mode hatch will pick up worktree1 as the project rather than project. In project mode I don't know how to specify a directory when any of project/* should map to project. Am I just out of luck?

[ofek]

As I am unfamiliar with gitworktrees, can you please explain a bit more your workflow?

[greened]

git worktrees create a new working directory from a single local clone. So instead of a single clone with all the files in the working directory you can create multiple workareas with the sources checked out on each branch. In my workflow, the local clone is effectively bare:

foo/
    .git/
    # No files checked out in `foo/`
    main/
        .git # Points to ../.git/
        pyproject.toml
    feature1/
        .git # Points to ../.git/
        pyproject.toml
    feature2/
        .git # Points to ../.git/
        pyproject.toml

For project foo, main, feature1 and feature2 are all git worktrees with branches main, feature1 and feature2 checked out. They are all different versions of project foo in different states of development. hatch's project heuristics mean that when I am working in main, hatch show displays:

[Project] - main (current directory)
[Location] - /path/to/foo/main
[Config] - /home/me/.config/hatch/config.toml

Does hatch reporting main as the project mean anything? Is there anything in hatch's internals that rely on this project name vs. the name given in pyproject.toml? What is the purpose of [Project] anyway? If eveything is described in pyproject.toml why doesn't hatch just use that to determine the project?

I can't use project mode to resolve this because there is no single path to "the project" that contains pyproject.toml.