Today I learned that you don’t need to switch commits or enter detached HEAD mode just to look at an old version of a file. Git can show you exactly what a file looked like at any commit — without touching your working tree — because every version of every file is stored as a blob object.

All you need is:

git show <commit_sha>:<file_name>

git-show(1) normally displays commit details, but when you append <file> after <commit>:, Git resolves it to a blob object and prints exactly what that file looked like at that moment in history.

It’s a surprisingly clean way to time-travel a single file.

For example: imagine a random tuesday afternoon where you've just cloned the Git source and suddenly feel like reading the first version of read-cache.c written by Linus back in 2005.

1. Find the first commit.

$ git log --reverse --oneline

(Copy the first SHA you see.)

2. Use git show to view read-cache.c from that commit:

$ git show e83c516331:read-cache.c

And now you're looking at 2005 code — no checkouts, no detached HEAD, no messing with your working directory.