While the current idea with large binary files is that they're going to be fetched via the generic fetcher backend in cachi2 [1], git LFS was here long before that. Customers may already have valid setups based on LFS that would simply not work using hermetic builds because, well, LFS replaces the large blobs with text references and only fetches the LFS objects on demand. To make them available right after cloning one would have to explicitly pull the objects with git lfs pull. That seems like something that would fit into the git-clone task which would pull all these blobs in order to make them available for hermetic builds.
While the current idea with large binary files is that they're going to be fetched via the generic fetcher backend in cachi2 [1], git LFS was here long before that. Customers may already have valid setups based on LFS that would simply not work using hermetic builds because, well, LFS replaces the large blobs with text references and only fetches the LFS objects on demand. To make them available right after cloning one would have to explicitly pull the objects with
git lfs pull. That seems like something that would fit into thegit-clonetask which would pull all these blobs in order to make them available for hermetic builds.[1] https://github.com/containerbuildsystem/cachi2/releases/tag/0.14.0