Git Notes

Git is Content-Addressible Storage

Git is a CAS, a kind of database, with a VCS on top of it. The CAS commands are plumbing commands that deal with this underlying database. The usual git {add,rm,commit} commands are compound porcelain commands that build on top of these lower level commands.

This means that you can actually use the plumbing commands to manually add, remove, commit things. The “Git Internals” section of the handbook has great examples.

But I used git cat-file to really investigate a lot of things. git cat-file -p will determine the type of object before pretty catting it.

There are four Types of Objects: blob, tree, tag, and commit.

Areas

There are three areas: Working, Staging, and Repository.

Working -> Staging

• When you do git diff you are comparing Working to Staging.
• When you git add you are adding stuff from the Working to Staging.
• git reset does the reverse: moves changes back to Working.

Staging -> Respository

• git diff --staged compares Staging to Repository
• When you git commit you are adding stuff from the Staging to Repository.

When you git rm you are ‘adding’ the deletion from the Working to the Staging area. This is a small head-scratcher until you see the tree object and note that the file(s) you just deleted are no longer referenced. That’s ultimately what happens. The file blob is not deleted; it’s part of git forever (until you use some tool to parse history and remove it manually.)

Other Trivia

Everything git writes is just a blob. There’s no binary/text types.

The reflog does not contain all commits.

Snippets

List the Largest Files in a Repo

# Here's one way. It's supposed to be *BLAZINGLY FAST*
git rev-list --objects --all |
  git cat-file --batch-check='%(objecttype) %(objectname) %(objectsize) %(rest)' |
  sed -n 's/^blob //p' |
  sort --numeric-sort --key=2 |
  cut -c 1-12,41- |
  $(command -v gnumfmt || echo numfmt) --field=2 --to=iec-i --suffix=B --padding=7 --round=nearest

# Here's another. It's simpler to read.
git rev-list --objects --all \
  | grep "$(git verify-pack -v .git/objects/pack/*.idx \
           | sort -k 3 -n \
           | tail -10 \
           | awk '{print$1}')"

Large File Storage (LFS)

Initialize with

git lfs install

Then add a .gitattributes file:

# Image files
*.bmp filter=lfs diff=lfs merge=lfs -text
*.gif filter=lfs diff=lfs merge=lfs -text
*.jpeg filter=lfs diff=lfs merge=lfs -text
*.jpg filter=lfs diff=lfs merge=lfs -text
*.png filter=lfs diff=lfs merge=lfs -text
*.tiff filter=lfs diff=lfs merge=lfs -text
*.webp filter=lfs diff=lfs merge=lfs -text

# Audio files
*.aac filter=lfs diff=lfs merge=lfs -text
*.flac filter=lfs diff=lfs merge=lfs -text
*.mp3 filter=lfs diff=lfs merge=lfs -text
*.wav filter=lfs diff=lfs merge=lfs -text

# Video files
*.avi filter=lfs diff=lfs merge=lfs -text
*.mkv filter=lfs diff=lfs merge=lfs -text
*.mov filter=lfs diff=lfs merge=lfs -text
*.mp4 filter=lfs diff=lfs merge=lfs -text

# Document files
*.doc filter=lfs diff=lfs merge=lfs -text
*.docx filter=lfs diff=lfs merge=lfs -text
*.pdf filter=lfs diff=lfs merge=lfs -text
*.ppt filter=lfs diff=lfs merge=lfs -text
*.pptx filter=lfs diff=lfs merge=lfs -text

# Archive files
*.zip filter=lfs diff=lfs merge=lfs -text
*.tar filter=lfs diff=lfs merge=lfs -text
*.gz filter=lfs diff=lfs merge=lfs -text
*.rar filter=lfs diff=lfs merge=lfs -text

# Font files
*.woff filter=lfs diff=lfs merge=lfs -text
*.woff2 filter=lfs diff=lfs merge=lfs -text
*.eot filter=lfs diff=lfs merge=lfs -text

# Other binary files
*.exe filter=lfs diff=lfs merge=lfs -text
*.dll filter=lfs diff=lfs merge=lfs -text
*.bin filter=lfs diff=lfs merge=lfs -text
*.iso filter=lfs diff=lfs merge=lfs -text

Migrate existing repos with

git lfs migrate import --everything --include="*.bmp,*.gif,*.jpeg,*.jpg,*.png,*.tiff,*.webp,*.aac,*.flac,*.mp3,*.wav,*.avi,*.mkv,*.mov,*.mp4,*.doc,*.docx,*.pdf,*.ppt,*.pptx,*.zip,*.tar,*.gz,*.rar,*.woff,*.woff2,*.eot,*.exe,*.dll,*.bin,*.iso"

Now the .git folder might be larger than before, which sort of beats the whole point. This might be because:

• When you migrate files to LFS, Git keeps the original history alongside the new LFS pointers until garbage collection runs. You essentially have both versions temporarily stored.
• Git LFS maintains its own metadata structures in the .git/lfs directory.
• The migration process creates entries in Git’s reflog, which tracks reference changes.

You can run Garbage Collection with

git gc --aggressive --prune=now

This has worked for me pretty well.

👉 To completely remove old binary files from history (!!! REWRITES HISTORY !!!)

git filter-branch --force --index-filter "git rm --cached --ignore-unmatch path/to/files/*" --prune-empty --tag-name-filter cat -- --all
git reflog expire --expire=now --all
git gc --aggressive --prune=now

I’d make a backup.

References

• Advanced Git: Graphs, Hashes, and Compression, Oh My!