A worked example, rendered from real sample data. Sign in to run the tool on your own input.
diff --git a/src/auth.ts b/src/auth.ts
index 1a2b3c4..5d6e7f8 100644
--- a/src/auth.ts
+++ b/src/auth.ts
@@ -1,6 +1,8 @@
import { verify } from './jwt';
-export function check(token: string) {
- return verify(token);
+export function check(token: string, audience: string) {
+ console.log('checking', token);
+ return verify(token, { audience });
}
═══ What this tool did ═══
ℹ Every number below was counted from the diff you pasted. This page runs offline and makes no outbound network calls. — no repository was read.
═══ Totals ═══
Files changed: 1
Insertions: +3
Deletions: -2
Net: +1 lines
Hunks: 1
Add/remove ratio: 1.50
═══ Composition ═══
source: 1 file, +3/-2 (100% of the churn)
⚠ Source files changed but no test file did.
═══ Files (1 of 1, largest first) ═══
src/auth.ts
modified · typescript · 1 hunk · +3/-2
++++++++++++++----------
═══ Signals found in the added lines ═══
⚠ console logging was added — 1 occurrence (src/auth.ts:4)
═══ Review effort (heuristic, formula shown) ═══
Reviewable churn: 5 lines (lockfiles excluded)
Formula: ceil(churn / 60 + files x 0.4) = 1 minute
ℹ This is arithmetic on the diff, not a measurement of anything. Treat it as a size class, not a schedule.
✓ Under the 800-line mark where review quality collapses.
═══ Useful follow-ups ═══
git diff --stat
git diff --numstat | awk '{a+=$1;d+=$2} END {print a" added, "d" deleted"}'
git diff -- . ':(exclude)*lock*' review without generated churn
git diff --diff-filter=D --name-only what this change deletes
git log --oneline -20 -- <path> history of the hottest file above
Parse a pasted unified diff for real insertion/deletion counts, per-file churn, test coverage of the change and risk signals in added lines. Part of the DevTools Surf developer suite. Browse more tools in the Developer Utilities collection.