git hash-object

Where the SHA-1/SHA-256 magic happens.

🎯 What & why

git hash-object computes the object ID Git would assign to a piece of content, and (with -w) actually writes that blob into the object database. It's the SHA factory.

🧠 Mental model

Git hashes 'header + NUL + content' where header is e.g. 'blob 1234'. Same bytes -> same SHA, deterministically. -w just stores the result under .git/objects/.

🛠️ Synopsis

git hash-object [-t <type>] [-w] [--path=<file>] [--no-filters] [--stdin] [--] <file>...
git hash-object [-t <type>] [-w] [--stdin-paths] [--no-filters]
git hash-object --literally [-t <type>] [-w] --stdin

🎚️ Switches & options

Flag	What it does
`-t <type>`	Object type (blob, tree, commit, tag); default blob.
`-w`	⚠️ Write the object into the object DB, not just print its SHA.
`--stdin`	Read content from stdin instead of a file argument.
`--stdin-paths`	Read a list of paths from stdin and hash each.
`--no-filters`	Skip clean filters and CRLF normalization; hash bytes as-is.
`--path=<file>`	Pretend the content lives at <file> for filter/attribute purposes.
`--literally`	⚠️ Allow malformed/oversize objects; for forensic or test use only.

💡 Use cases

Inject blobs into a repo without staging (build pipelines, lfs-like flows).
Verify what SHA a file would produce before/after a filter.
Reproduce object IDs for fixtures and tests.
Inspect filter behavior by comparing hashes with and without --no-filters.

🧪 Examples

Hash without writing

git hash-object README.md

Write a blob into the DB

git hash-object -w big.bin

Hash from stdin, treat as file path

echo hello | git hash-object --stdin --path=greet.txt

Batch hash a file list

find src -type f | git hash-object --stdin-paths -w

🎓 Recommendations

Pair with git cat-file -p <sha> to immediately verify what you wrote.
Use --no-filters when you need byte-exact hashing across machines with different autocrlf settings.
Treat -w as mutating: every blob you write sticks around until GC prunes unreachable objects.

🪤 Common pitfalls

Without -w you only get the SHA; the object is not stored. Many users assume otherwise.
CRLF/clean filters can change the SHA between OSes; --path and --no-filters affect this.
--literally lets you create broken objects that can corrupt fsck output; reserve for plumbing tests.

🔗 Related modules

📝 Quiz

Hit each option, then Check answers. Score is recorded; Next is always open.