Reflog

The Reflog gives a chronological account of how HEAD has moved. Each time it moves, there is a new entry in the Reflog.

The reflog is a list of chronological hashes, which represent where you have been during commits, all without regard to branches.

Since branches don't matter, we are able to recover dangling commits.
Each time a branch is updated to point to a new reference (ie. HEAD changes), an entry is written in the reflog to say where you were.
Since the branch is updated whenever you commit, the git reflog has a nice effect of storing your local developer's history.

It serves as Git's safety net, recording every change made in the repo (regardless of whether it was committed or not)

The listed commit hash represents the HEAD after that action.

Alongside the caret (^) and tilde (~) operators, git provides the @ for referencing reflog entries in a relative way.

git show HEAD@{5} gets us the value of HEAD 5 movements ago.
we can also see where the HEAD was by day:
- git show HEAD@{yesterday}
- git show HEAD@{1.day.ago}

While git log shows a history of the commits, we can think of git reflog as showing us a history of everything (what branch we checked out, what commands we ran etc.)

Becuase it keeps a full history, even if we were to git reset, we can still access the commit SHA

ex. we made some commits, then reset. If we decide that we want to undo the reset, all we need to do is checkout the commit we want with git checkout HEAD@{1}

ad8621a HEAD@{0}: reset: moving to HEAD~3
298eb9f HEAD@{1}: commit: Some other commit message
bbe9012 HEAD@{2}: commit: Continue the feature
9cb79fa HEAD@{3}: commit: Start a new feature

This puts us in detached head state. All we need to do is create a new branch, and continue working on our feature.

Data Recovery

Most of the time, the reflog is our friend in the circumstance where we want to recover data that has been "lost"
- we can either run git reflog, or use git log -g which gives us normal log output for the reflog
Imagine we hard reset a number of commits, effectively "erasing" them from the git log. All we need to do is find that commit with git log -g, create a new branch with its SHA (git branch recover-branch <SHA>)
If the data we are looking for is not in the reflog, we can try using git fsck --full, which will list all objects that aren't pointed to by another object.

Data Recovery