← All posts
Engineering1 September 2026· 4 min read

Re-reviewing a pull request should not mean reviewing it again

Push three commits to an open PR and a naive reviewer comments on the same code three times. How incremental review works, when it doesn't apply, and why re-running an unchanged PR costs full price.

Pull requests are not reviewed once. You open one, get feedback, push a fix, push another, and the branch sits open for two days. A reviewer that re-reads the entire diff on every push produces the same comments repeatedly and burns your budget doing it.

Reviewing the delta

Pullora records the last reviewed commit SHA on each pull request. When new commits arrive, it fetches the diff between that SHA and the new head, and reviews only what changed.

That means no duplicate comments on code that was already reviewed and already commented on, and a much smaller review to pay for — a three-line fix costs like a three-line fix, not like the 400-line pull request it belongs to.

Deduplication on top

Findings are fingerprinted, so even when the same issue is legitimately re-detected — because the surrounding code moved, say — it is not posted twice.

The case it does not cover

If you re-review a pull request with no new commits, there is no delta, so the whole diff is reviewed again at full cost. That is the honest behaviour rather than the convenient one: nothing changed, so nothing can be skipped, and the second review genuinely costs what the first did.

If your intent is to re-read a review you already ran, open the existing review rather than starting a new one — the findings are all still there.

See a review before you install anything

Paste any public GitHub pull request URL and read the full review — no app installed, no repository access, nothing posted to the PR.

Review a public PR →