Complete Guide to Reverting Git Repository to Previous Commits

Overview of Git Revert Operations

In Git version control system, "revert" operations encompass multiple meanings and implementations depending on developers' actual needs and current repository state. Understanding the differences between various revert methods is crucial for proper code history management.

Temporarily Switching to Historical Commits

When needing to temporarily inspect historical commit states or conduct experimental modifications, the git checkout command can be used. This approach doesn't modify commit history but temporarily switches the working directory to the specified commit state.

# Detach HEAD pointer, temporarily switch to specified commit
git checkout 0d1d7fc32

If new development work is required based on historical commits, a new branch can be created simultaneously:

# Create new branch based on historical commit
git checkout -b old-state 0d1d7fc32

After completing temporary operations, return to the original branch via git checkout. If modifications were made during this period, they need to be handled according to specific circumstances: discard changes with git reset, stash changes with git stash, or directly commit to the new branch.

Hard Reset for Unpublished Commits

For local commits not yet pushed to remote repositories, git reset --hard can be used to completely remove these commits, fully resetting the repository state to the specified commit.

# Hard reset to specified commit, destroy all subsequent modifications
git reset --hard 0d1d7fc32

If uncommitted modifications need preservation, combine with git stash:

# Save current modifications, reapply after reset
git stash
git reset --hard 0d1d7fc32
git stash pop

It's important to note that hard reset operations permanently delete all commits and modifications after the reset point. Before operation, ensure these contents are no longer needed. If mistakes occur, attempt recovery through Git's reflog.

Reverting Published Commits

For commits already pushed to remote repositories, reset operations are not recommended as they cause history rewriting, potentially affecting other collaborators. Instead, use the git revert command to undo historical changes by creating new reverse commits.

First, determine the commit range to revert. For regular commits, use the following command to list commits requiring reversion:

# List all non-merge commits between two commits
git log --no-merges --pretty=format:"%h" 0d1d7fc..HEAD | tr '\n' ' '

Revert operations should be executed in chronological order from newest to oldest:

# Revert specified commits individually
git revert a867b4af 25eee4ca 0766c053

# Use range syntax to revert multiple commits
git revert HEAD~2..HEAD

# Use commit hash range
git revert 0d1d7fc..a867b4a

Merge commit reversion requires special handling:

# Revert merge commit, -m parameter specifies parent commit
git revert -m 1 <merge_commit_sha>

An alternative method involves manually restoring file states and committing:

# Restore working directory and staging area to specified commit state
git checkout 0d1d7fc32 .

# Create new commit
git commit

Operation Selection Guidelines

When choosing revert methods, consider multiple factors: whether commits are published, need to preserve historical records, team collaboration requirements, etc. Temporary switching suits exploratory work, hard reset suits thorough cleanup of local unpublished commits, while commit reversion provides safe handling for published changes.

Each method has specific use cases and risks. Before actual operations, use git status and git log to confirm current state, and create backup branches before important operations to prevent accidental data loss.