git-learning-mode

git-learning-mode.sh configures the current repository only for studying Git's object model with minimal hidden automation.

The script disables background/automatic maintenance features that can mutate .git/objects behind the scenes (such as auto-GC, auto-packing, pruning, and maintenance tasks). It also disables pack delta behavior and some acceleration indexes so object storage and history traversal stay easier to inspect.

Why this exists

Git normally optimizes storage and performance automatically. That is great for day-to-day work, but it can make internals harder to observe while learning.

This script puts a repository into a "learning mode" where:

• Objects are less likely to be silently repacked.
• Unreachable objects are retained instead of pruned.
• Reflog history is preserved indefinitely.
• Auxiliary acceleration features are reduced.

Scope and safety

• Changes are written with git config --local.
• No global or system Git config is touched.
• You must run it from inside a Git repository.

Usage

From inside a repository:

path/to/git-learning-mode.sh

If successful, it prints a short summary and suggested inspection commands.

What gets configured

The script sets local config values including:

• gc.auto=0
• gc.autoPackLimit=0
• maintenance.auto=false
• maintenance.strategy=none (best-effort)
• core.commitGraph=false
• core.compression=0
• pack.window=0
• pack.depth=0
• pack.useBitmaps=false
• gc.pruneExpire=never
• gc.reflogExpire=never
• gc.reflogExpireUnreachable=never
• core.fsmonitor=false (best-effort)

Helpful follow-up commands

After running the script, inspect object internals directly:

find .git/objects -type f
git cat-file -t <hash>
git cat-file -p <hash>

Notes

• This mode favors transparency over performance.
• It is intended for local learning repos, not production workflows.