git grep

Faster than grep on a repo. Plays well with revspecs.

🎯 What & why

Search for a pattern in tracked content. Faster than POSIX grep because it walks the index, not the filesystem; and it can search any historical revision, not just the current working tree.

🧠 Mental model

Grep takes a pattern and (optionally) a tree-ish or pathspec. Without a tree-ish, it searches the working tree for tracked files. With a tree-ish, it greps inside that tree's blobs.

🛠️ Synopsis

git grep [<options>] <pattern> [<tree-ish>...] [[--] <pathspec>...]

🎚️ Switches & options

Flag	What it does
`-n`	Show line numbers.
`-l`	List only matching files.
`-c`	Count matches per file.
`-i`	Case-insensitive.
`-E`	Extended regex.
`-P`	PCRE.
`--cached`	Search the index, not the working tree.
`<tree-ish>`	Search inside that revision.

💡 Use cases

Quick code search instead of grep -r.
Searching historical states of the codebase.
Finding all callers of a function before refactoring.

🧪 Examples

Find a function definition in the working tree.

$ git grep -n 'def parse_request'

Find every place a variable was used in v1.2.

$ git grep -n 'session_token' v1.2

Count matches across the codebase.

$ git grep -c 'TODO'

Limit by path.

$ git grep 'malloc' -- '*.c'

🎓 Recommendations

git grep over grep -r in any Git repo. It honors .gitignore, ignores .git/, and is faster.
Use -P for advanced regex when you need it.

🪤 Common pitfalls

By default, untracked files are skipped. Pass --untracked if you want them included.
PCRE (-P) requires Git built with PCRE support — most distros do, some don't.

🔗 Related modules

📝 Quiz

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