Checkout flow regression — Stripe → LemonSqueezy migration

Chronological findings — 8 logged across 2h 14m

1. !

   Discount code lost on Stripe → LemonSqueezy hand-offSeverity: high

   00:42:17

   Applied promo `LAUNCH20` is dropped when LemonSqueezy receives the cart payload. Steps + payload diff in §3.2.

   1. Open /checkout in Chrome 126 (macOS); add cart item
   2. Apply promo LAUNCH20; verify line-item discount visible
   3. Click Pay → observe LS payload missing discount field
2. !

   Address autofill clobbers manually-entered ZIPSeverity: med

   01:08:54

   Reproduced on Chrome 126 (macOS) + Safari 17 (iOS). Form re-submit fails silently when ZIP mismatches autofill.

   1. Trigger Chrome address autofill on shipping form
   2. Manually edit ZIP to a different valid value
   3. Submit → form silently rejects with no error visible
3. ✓

   Apple Pay round-trip — greenSeverity: low

   01:31:09

   End-to-end Apple Pay sandbox checkout passes on iOS 17 / Safari 17. Receipt emails arrive within SLA.

   1. Initiate Apple Pay sheet from /checkout
   2. Authenticate via sandbox account on iOS 17
   3. Verify receipt email arrives within 60s
4. !

   Webhook retry storm on 503Severity: high

   02:02:42

   When the LemonSqueezy webhook receiver returns 503, the retry policy fires 6 attempts within 11 seconds (expected: exponential backoff).

   1. Force webhook endpoint to return 503 via fault injection
   2. Observe retry log — 6 attempts in 11s
   3. Compare expected exponential backoff (max 3 attempts in 30s)
5. i

   Tax calculation latency p95 = 1.4sSeverity: low

   02:08:11

   Avalara API call adds 1.4s p95 latency on cart update. Within SLO but worth caching at edge.

   1. Trigger cart update with new tax-jurisdiction ZIP
   2. Measure p95 over 50 calls
   3. Compare with cached jurisdiction lookup baseline
6. ✓

   Refund flow — green across all 3 providersSeverity: low

   02:11:25

   Stripe + LS + Apple Pay all process partial refunds correctly; balance ledger reconciles on next webhook.

   1. Issue partial refund via admin dashboard for each provider
   2. Verify ledger balance updates on next webhook
   3. Confirm customer notification email arrives
7. !

   Cart total off-by-1¢ on multi-currency convertSeverity: med

   02:13:02

   Floating-point rounding produces 1¢ delta when cart switches USD → EUR mid-session. Server total ≠ client total by 0.01.

   1. Add USD item to cart
   2. Switch locale to EU → convert to EUR
   3. Compare client display vs server total — observe 0.01 delta
8. ✓

   All 6 SOC 2-required audit logs presentSeverity: low

   02:14:00

   Every checkout state transition emits structured audit log with actor + delta + timestamp.

   1. Walk through checkout end-to-end on staging
   2. Tail audit log; verify 6 required transitions logged
   3. Cross-check actor + delta + timestamp fields

Full report sections

This is what a real Test4All report looks like — structured, reproducible, hand-written by senior QA engineers.