P-2.3 Version Control: Commits, History and Undo

Work through these

• A repository, a commit, and why a commit is a snapshot

  Recording a commit stores the state of every file at that moment, and the differences you see are computed between snapshots rather than stored. Once that lands, moving around history stops feeling dangerous.

  Pro Git (2nd edition) · Reference

• The staging area: choosing what goes into the next commit

  Changes are selected before they are recorded, which lets one messy afternoon become three commits that each do one thing. This is the step most people skip and later wish they had not.

• Writing a commit message that is worth reading in a year

  The message explains why the change was made, because the code already shows what changed. A history of messages saying update is a history nobody can use.

• Reading history: what changed, when, and by whom

  Being able to find the change that introduced a behaviour turns a mysterious defect into a dated, attributed, reviewable event. This is the payoff for every message you wrote carefully.

• Undoing safely: restoring a file, moving back, and what is recoverable

  Almost everything recorded can be recovered, and the small number of operations that genuinely discard work are worth knowing by name before you need them.