This repo is a stress test to check if we are able to download only certain files or directories from a git repository: https://stackoverflow.com/questions/600079/how-do-i-clone-a-subdirectory-only-of-a-git-repository/52269934#52269934

This repo is designed such that if you clone anything outside of small/ or small2/ or subdir/small/, then you get at least 100 MB extra pseudorandom data, so it will be very noticeable on the network and disk usage.

Contents:

• big/: 10x pseudorandom 10 MB files (~100 MB), small filenames

• big0 - big9: 10x pseudorandom 10 MB files (~100 MB). These were added to ensure that toplevel objects are not fetched.

• big_tree: 10x directories with 50k empty files each, each with filenames of 200 pseudorandom bytes long (~100 MB across 10 large tree objects). To ensure that unnecessary tree objects are not fetched.

  GitHub refuses the upload because the object is too large if we don't split the tree into 10 smaller ones.

• small/: 1000x 1 B files, small filenames. The usual download target.

• small2/: same as small/. To see if the method supports downloading multiple directories at once together with small

• subdir/small/: same as small/. Because git sparse-checkout set --no-cone small works on basenames not full paths and also fetches subdir/small/, for full paths you need /small

There is also a branch bigissmall in which:

• big/ is renamed to bigdir/
• small/ is renamed to big/

to check that cloning from a different branch does not touch master in any way.

This repository is automatically generated with the generate.sh script. Just run:

./generate.sh

and it will nuke your old .git and create a clean new repo.