Improve local git source repository handling

[frol]

Right now, either you use a remote git repo or a local one, conda build executes git clone and that is it. This works in some cases, it works slowly in other cases, and it may not work at all on access-restricted repos or while you are offline.

Once you have a submodule (obviously listed in .gitmodules with an external link), git clone of a local repo will require network access, "read" permissions to the submodule, and some network traffic. There are a few use-cases when this will be a show-stopper:

1. CI server, which clones private repositories with its own credentials and you cannot re-clone the submodules on conda build stage due to the lack of credentials.
2. Offline mode development is not possible without manual modifications (e.g. changing git_url to path in meta.yaml).

Even if it is not a show-stopper for you, you may suffer from a slow internet connection, so downloading all submodules on every build will be frustrating and completely unnecessary.

I used to use source/path option, but it was "fixed" in #745. I have also noticed #753, but it doesn't solve the issue at all since it will only work for the cases when you don't need submodules.

/cc @ilanschnell @jasongrout @stuarteberg @gmarkall @jakirkham @groutr

[stuarteberg]

I used to use source/path option, but it was "fixed" in #745.

IIUC, I think the behavior of source/path within git repos was recently improved by @groutr via #801. I haven't tested the new behavior myself, though.

[frol]

@stuarteberg The PR makes the code look much better. Yet, it doesn't change the fact that having GIT_DESCRIBE_* env variables requires source/git_url setting, which results in git clone.

Since we have our automated build scripts on top of conda-build, I have managed to auto-discover GIT_DESCRIBE_* information before conda-build run, and pass it via env variables to conda-build execution, so we continue to use GIT_DESCRIBE_* with source/path setting in meta.yaml.

[groutr]

@frol, if source/path isn't defining GIT_* variables, then it's very likely a bug that needs to be fixed. The test for source/path specifically checks that these GIT_* variables are defined when source/path is used.

[frol]

@groutr This "bug" was introduced as a fix in #745. (I have mentioned this in the first message.)

[stuarteberg]

This "bug" was introduced as a fix in #745

As I mentioned above, I think it was probably fixed by @groutr in #801. Can you confirm, @frol?

[github-actions]

Hi there, thank you for your contribution!

This issue has been automatically marked as stale because it has not had recent activity. It will be closed automatically if no further activity occurs.

If you would like this issue to remain open please:

1. Verify that you can still reproduce the issue at hand
2. Comment that the issue is still reproducible and include:
   - What OS and version you reproduced the issue on
   - What steps you followed to reproduce the issue

NOTE: If this issue was closed prematurely, please leave a comment.

Thanks!