Rewrite Git history via drag-and-drop

Original Article ↗ Hacker News Discussion ↗

Purpose and value of rewriting history

  • Many see value in rewriting only local or feature-branch history before merging: remove “oops/typo/WIP” commits, group related changes, and make each commit buildable and meaningful.
  • Clean history is viewed as critical for git bisect, git blame, and understanding why code changed; small, well-described commits help debugging years later.
  • Some use history rewriting to maintain internal forks of upstream projects, splitting local changes into well-isolated feature commits to ease future upstream updates.
  • Rewriting may be necessary in corporate setups when author name/email changes must be applied retroactively.

Skepticism and opposition

  • Others are uncomfortable with “rewriting history,” worrying it hides the real development path or weakens the audit trail.
  • Some prefer “messy but complete” logs, arguing that failed attempts and PR-fix commits can provide useful context or cultural insight.
  • There is disagreement on whether main should be perfectly clean vs. using merge commits and log filters while preserving all intermediate commits.

Retcon’s approach and technical aspects

  • Retcon is seen as a polished GUI wrapper around git rebase -i, with its main novelty being drag-and-drop reordering and rich undo/redo.
  • A key feature: when a move introduces conflicts, users can keep rearranging commits first and resolve conflicts later, all tracked in an in-memory “virtual history” before writing to Git.
  • Some want more: true DAG-based drag-and-drop rebasing, better conflict resolution, and powerful commit-splitting (down to line-level, via drag-and-drop).

Comparisons to existing tools

  • Alternatives mentioned: Sublime Merge, SmartGit, JetBrains IDEs, GitKraken, lazygit, jj/jujutsu (plus GUIs on top), GitButler, Fork, IntelliJ UI.
  • Several note that many of these already support interactive rebasing, drag-and-drop within limits, or advanced conflict handling; opinions differ on how smooth they are.

Pricing and adoption

  • Multiple commenters balk at a subscription (e.g., ~$10/month) for something used infrequently and seen as “just a nicer rebase UI,” preferring a modest one-time license.
  • The developer argues subscriptions help sustain long-term development; Retcon is also available via Homebrew.
  • Some doubt many will pay for a dedicated history-rewrite tool when free or existing paid tools already cover much of this workflow.