167 lines
22 KiB
Markdown
167 lines
22 KiB
Markdown
# Freshyo — pending test case run
|
||
|
||
**Run ID:** PEND-2026-10-18-01
|
||
**Started:** 2026-10-18 17:55 IST
|
||
**Scope:** the 88 "runnable when next picked up" cases in §3 of
|
||
`docs/test reports/pending-test-cases-2026-10-18.md`, plus the 26 open/partial cases in §2 of that file.
|
||
OTP / notification / scheduled-job / fallback-legacy cases stay excluded (owner instruction).
|
||
|
||
**Environment**
|
||
| Surface | Value |
|
||
|---|---|
|
||
| Backend | `http://localhost:8787` (wrangler dev, config `wrangler.dev.toml`, pid 91533 — not restarted) |
|
||
| Database | Postgres `meatfarmer` @ `57.128.212.174:7447`, schema `mf` — **throwaway, freely mutable** |
|
||
| Customer mobile | Expo Go `host.exp.Exponent` ← Metro `:8081` |
|
||
| Admin mobile | Expo Go `host.exp.Exponent` ← Metro `:8082` |
|
||
| Customer web | Chromium `http://localhost:4174` |
|
||
| Superadmin web | Chromium `http://localhost:4175` / `:4176` |
|
||
| iOS simulator | `iphone_6.5_in` / iOS 18.6 — `5A64695E-F4D4-4E2E-90F3-EC3493A2B457` |
|
||
| Maestro live view | `http://127.0.0.1:9999/` |
|
||
|
||
**Identities**
|
||
| Alias | Identity |
|
||
|---|---|
|
||
| anon | no token |
|
||
| U1 | customer 1, mobile `9676651496` / `meatfarmer@123` |
|
||
| U2 | customer 1327, `e2e_u2b@test.com` / `Test@12345` |
|
||
| staffA | `e2eadminA` / `Test@12345` (region 1) |
|
||
| staffB | `e2eadminB` / `Test@12345` (region 2) |
|
||
| SA | `superadmin` / `superadmiN123` |
|
||
|
||
**Verdict vocabulary:** `PASS` · `FAIL` (defect) · `PARTIAL` · `BLOCKED` (harness/env) · `N/A`.
|
||
|
||
**Defect numbering** continues from the previous run (`docs/test reports/full-regression-2026-10-18-defects.md`, last id 61).
|
||
|
||
---
|
||
|
||
## Progress log
|
||
|
||
- 17:55 — environment verified (backend/D1 → Postgres confirmed; Metro 8081/8082, web 4174/4175/4176, sim `5A64695E…`); all five identities re-authenticated OK; report opened.
|
||
|
||
---
|
||
|
||
## Results
|
||
|
||
_(appended per phase below)_
|
||
|
||
### Phase 1 — authorization guards on slot / stock / staff endpoints
|
||
|
||
Evidence driver: `/tmp/freshyo/p1.mjs`, `p1c.mjs`, `p2.mjs`, `p2b.mjs` (direct tRPC, `http://localhost:8787/api/trpc`).
|
||
Fixtures created and deleted inside the phase (global product `226`, `227` on global store `24`).
|
||
|
||
| ID | P | Verdict | Evidence |
|
||
|---|---|---|---|
|
||
| SLOT-20 | P0 | **PASS** | `admin.slots.updateSlot` as anon → 401; as U1 → 401 "Access denied"; `createSlot`/`updateSlotCapacity` as U1 → 401. staffB (region 2) updating region-1 slot `2087` → 500 `Unable to Update Slot`; slot re-read by staffA: `isActive` unchanged (true→true). No relation writes. (The 500-instead-of-4xx shape is defect #8.) |
|
||
| STAFF-08 | P0 | **PASS (with cross-ref)** | anon `admin.staffUser.createStaffUser` → 401, `superadmin.staff.create` → 401, `getStaff` → 401. U1 `superadmin.staff.create` → 401. U1 `createStaffUser` → 500 (not a clean 401 — the resolver dereferences `ctx.staffUser.regionId`); **no staff row created** (staff list unchanged). U1 `getStaff` → **200 and returns names/roles/permissions** — already covered by defect #1 (customer reaches `protectedProcedure` admin resolvers); no password hashes exposed. No staff credentials leaked. |
|
||
| STOCK-08 | P0 | **PASS** | Global product 227 / sku 233 (global store 24): regional admin sets `isSuspendedInRegion: true` → 200, read-back `true`; then price-only save with `isSuspendedInRegion` omitted (98→119) → 200 and read-back **still `true`**; sku therefore not silently orderable. Note: this contradicts defect #10 from the previous run — regional suspension now persists in this build. |
|
||
| STOCK-09 | P1 | **PASS** | SA suspends the SKU globally (`isSuspended: true`) → 200; regional admin read shows `isSuspended=true` **and** `isSuspendedInRegion=false` (kill-switch separate from the regional flag); regional price edit (111→117) → 200 with `isSuspended=true` / `isSuspendedInRegion=false` afterwards; global `isSuspended` still `true`. |
|
||
| STOCK-14 | P0 | **PASS (new P2 observation)** | anon `toggleOutOfStock` → 401; U1 → 500 (`regionId` of null) with no write. staffA toggling a **region-2-only** product (`172`) → 404, staffB read unchanged. staffA `toggleOutOfStock` with a **region-2 sku id** on its own product → 400 "These variants do not belong to product" and own SKU untouched. staffA `updateProduct` with a foreign `skuId` → **200 but silently ignored** (sku 175 price stayed 222) — see observation O-1. |
|
||
| STOCK-19 | P1 | _pending (publication-failure injection)_ | — |
|
||
| STOCK-20 | P2 | _pending (admin-app dialog)_ | — |
|
||
| SAUTH-10 | P1 | _pending_ | — |
|
||
| SAUTH-15 | P1 | _pending (admin-web not started? `:4176` is up)_ | — |
|
||
| SEC-24 | P1 | _pending (bounded harness)_ | — |
|
||
|
||
**Observation O-1 (not a numbered defect — no write occurred):** `admin.product.updateProduct` returns 200 when the payload carries a `skuId` that does not belong to the product (or to the caller's region); the unknown SKU is dropped silently instead of being rejected with 4xx. Contrast with `toggleOutOfStock`, which does reject the same payload with 400.
|
||
|
||
### Phase 2 — slots (orders), global products/stores, per-region flags
|
||
|
||
Drivers: `/tmp/freshyo/p3.mjs`, `p4.mjs`, `p4b.mjs`, `p5-cache.mjs`. U1 = customer 388, address 285 (region 1 / bound 1, GPS Mahabubnagar), slot 2072.
|
||
|
||
| ID | P | Verdict | Evidence |
|
||
|---|---|---|---|
|
||
| SLOT-08 | P0 | **PASS** | Regular order with `slotId: null` → 400 "A delivery slot is required for scheduled orders"; unknown slot → 400; region-2 slot 2075 from a region-1 address → 409 "belongs to a different region"; valid slot 2072 → 200 (sub-order pinned to slot 2072). Flash order sending `slotId` 2072 **and** 999999 → 200 and every created sub-order stored `slotId: null`, `isFlashDelivery: true` (server normalises). |
|
||
| SLOT-17 | P1 | **PASS (PARTIAL note)** | Created an empty slot (2088) via `admin.slots.createSlot` → 200. Ordering any SKU into it → 400 "Product 102 is not available in the selected delivery slot" (no purchasable state). A slot whose only SKU is soft-deleted (`product 228`/`sku 234`) → order rejected 400 "Product 234 is unavailable". **Note:** the empty slot *is* present in `user.slots.getSlots` for customers (partial: an empty slot is still offered in the slot list). Slot deactivated afterwards; fixture product deleted. |
|
||
| SLOT-20 | P0 | PASS (phase 1) | — |
|
||
| SLOT-13 | P1 | _pending (device)_ | — |
|
||
| GP-02 | P0 | **PASS** | One global product (`229`, store 1) given region-specific state through the regional admin API: region 1 → price 111, `isFlashAvailable: true`, `flashPrice: 99`, `isOffer: true`; region 2 → price 222, out of stock, offer cleared. `superadmin.globalProducts.get` returns `marketStats` rows per region exactly matching (`region1 {111, flash 99, offer}`, `region2 {222, oos}`). Note: the customer-facing read of the *new* product still 500s "Product not found" because the published products cache never picked it up — that is the already-open defect #27 (structural product changes not published), not a per-region pricing fault. |
|
||
| GSTORE-05 | P1 | **PASS** | Global store created empty → product count 0; adding one global product → 1; updating it to 5 variants → 200 with 5 SKUs read back. |
|
||
| GSTORE-06 | P0 | **FAIL (new defect 62)** | Expected: **rejected**. Actual: `admin.product.updateProduct` with `storeId` = a regional store (2 "The Fruit Store") → **200**, then the product's `storeId` is still `1`; same for a region-B store 13 → 200, `storeId` still 1. No write happens, but the "assignment" is silently accepted (200) instead of rejected, and nothing in the response signals that it was ignored. |
|
||
| GSTORE-09 | P1 | **PASS** | Fresh empty global store 32 → delete 200 → absent from `superadmin.globalStores.list`, `admin.store.getStores` for both A and B; repeat delete → 500 "Global store not found" (fails, though with a 500 — defect #8). *(A populated store's delete correctly fails with 500 — that is GSTORE-08, dependency policy.)* |
|
||
| GSTORE-10 | P1 | **PASS (cold-start race)** | New region 12 created after the global stores exist; a fresh staff account in region 12 initially saw **0 stores / 0 products** immediately after creation, then after the store-initializer ran it saw **5 stores (all global) and 7 products (all global)** with no per-region duplicate rows. The first read is a race in the initializer, not a structural miss. |
|
||
| GSTORE-11 | P1 | **PASS** | `imageUrl` given as a bare asset key and as a full `https://assets2.freshyo.in/...` URL both normalise to exactly one `https://assets2.freshyo.in/store-images/...` prefix. |
|
||
| GSTORE-12 | P1 | **PASS** | Concurrent `update`(200) + `delete`(200) on the same global store → final `get` is 404 (coherent outcome; the row is gone). Upload-failure path: product create referencing an unclaimed image key → 400 "These images were not uploaded through the app (unknown or already used)". |
|
||
| GSTORE-13 | P2 | **PASS (api) / PARTIAL (UI)** | Regional admin store list returns `isGlobal` on 16/16 rows and the product list on 178/178 rows, so a read-only badge can be rendered. The badge/read-only form affordance itself needs the device/web UI — not driven in this phase. |
|
||
| PROD-12, PROD-14, PROD-15, STORE-11, STORE-14 | — | _pending_ | — |
|
||
|
||
**New defect 62 (P2, contract):** assigning a global product to a regional store via `admin.product.updateProduct({storeId})` returns 200 while the product's store is unchanged; the client gets no error and no indication the assignment was dropped. (Global product 229, regional stores 2 and 13, 2026-10-18.)
|
||
|
||
**Cleanup note:** fixtures `global product 226/229/230` (soft-deleted), `global store 29` and `region 12` (+ staff `pendR10_…`, id 26) remain — this backend exposes no region-delete and `globalProducts.delete` is a soft delete, so store 29 stays referenced by product 230.
|
||
|
||
### Phase 3 — order group, coupon usage restoration, pricing ties, rounding
|
||
|
||
Drivers: `/tmp/freshyo/p6.mjs`, `p6b.mjs`. U1 = customer 388, address 285, slots 2072 (`A`) and 2076 (`B`), SKUs 6 (Apple ₹109), 7 (Banana ₹69), 102 (Basa ₹498). Test coupons `PENDC*/PENDD*` created via `admin.coupon.create`.
|
||
|
||
| ID | P | Verdict | Evidence |
|
||
|---|---|---|---|
|
||
| CANCEL-08 | P0 | **PASS** | Place a 2-slot order (2072 + 2076) with **two** coupons `c1`(₹10)/`c2`(₹20) → 200, sub-orders 2727/2728 both in group `…-388`, discount ₹15 each. `getMyCoupons` usageCount = 1/1 while the group is live. Cancel **one** sibling → usages stay 1/1 (group not fully cancelled). Cancel the **second** sibling → usages drop to **0/0** (both coupons restored, not just the first — `releaseCouponUsageIfGroupFullyCancelled` deletes by `orderId IN group`). Both coupons then re-applied successfully → 200, discount ₹30. |
|
||
| CPA-21 | P1 | **PASS** | Same 2-slot group with one ₹5 coupon: after cancelling **one** sibling the usage is still consumed (1) and a retry is refused; after cancelling **every** sibling the usage is restored (0) and the coupon applies again (₹5 off). (The refusal surfaces as 500 "Coupon usage limit exceeded" rather than 4xx — defect #8 family.) |
|
||
| CPA-27 | P1 | **PASS** | 5 % coupon targeted at Banana (sku 7, ₹69) → `discountAmount` 3 (3.45 → nearest rupee); 1 % version → 1 (0.69 → 1). Order totals reconcile: 69−3+20 = 86 and 69−1+20 = 88, i.e. whole-rupee discount, delivery added once. |
|
||
| MONEY-08 | P1 | **PASS** | Two slots with **equal** goods (₹109 each): insertion `[B(2076), A(2072)]` → the ₹20 delivery charge lands on **B** (first inserted = tie-break) and the other sub-order has 0; reversed insertion `[A, B]` → charge lands on **A**. Exactly one sub-order per group carries delivery; both share one `orderGroupId`. (`order.totalAmount` already includes the delivery share: 129 = 109 goods + 20 delivery.) |
|
||
| RES-15 | P1 | **PASS** | Two concurrent `createReservedCoupon` calls with the same secret code → 200 / 500, and `getReservedCoupons` lists exactly **one** record for that code (id 755) — no duplicate secret for one event. |
|
||
| CANCEL-12 | P0 | _pending (needs packaged/delivered race)_ | — |
|
||
| CANCEL-16 | P1 | _pending (notification path)_ | — |
|
||
| CANCEL-17 | P1 | _pending (client cart)_ | — |
|
||
| CANCEL-18 | P2 | _pending (device dialog)_ | — |
|
||
| RES-16 | P1 | _pending_ | — |
|
||
| CPA-26 | P1 | _pending_ | — |
|
||
|
||
**Correction recorded during this phase:** an earlier `MONEY-08.equalGoods` FAIL was a harness error — `order.totalAmount` includes the sub-order's delivery share, so equal goods must be compared as `totalAmount − deliveryCharge`; both sub-orders are ₹109 goods. Corrected above.
|
||
|
||
### Phase 4 — combos, catalog, coupons sheet payload, banners, cache, fulfilment, vendor, reviews, support
|
||
|
||
Drivers: `/tmp/freshyo/b1.mjs`, `b1b.mjs`, `b1c.mjs`, `b2.mjs`, `b3.mjs`, `b3b.mjs`.
|
||
|
||
#### Combined harness results (these cases are each marked with their own verdict in the run log)
|
||
|
||
| ID | P | Verdict | Evidence |
|
||
|---|---|---|---|
|
||
| COMBO-05 | P1 | **PASS** | Combo create with `comboItems: [{skuId: X}, {skuId: X}]` → 200 and detail read shows exactly **1** component (deduplicated). |
|
||
| COMBO-06 | P0 | **PASS** | Combo with a region-2 component SKU → 500 "Combo SKU 175 belongs to a different region"; missing component → 500 "Invalid Combo SKU IDs: 999999". No partial combo row. (500-shape is defect #8.) |
|
||
| COMBO-07 | P1 | **FAIL** | A combo whose `comboItems` reference **its own** SKU → 200 "Product created successfully". Self-reference accepted (matches defect **#20**). |
|
||
| COMBO-12 | P0 | **FAIL** | Combo-targeted coupon on a combo line → 200, discount ₹10 (correct). Component-targeted coupon on a **combo** line → **500 "Coupon does not apply to any items in your cart"** — no phantom discount is given (the core invariant holds) but the checkout is *blocked* with a 500 instead of dropping/4xx-rejecting the inapplicable coupon. |
|
||
| COMBO-15 | P1 | **PARTIAL** | A combo with a 400-char feature value → 200 (no length contract anywhere — defect #31 family). |
|
||
| CAT-08 | P1 | **PASS** | Admin combo detail carries `comboItems` (component name/price/features/images) alongside the combo's own price/features. A newly created product reached the published region-1 `products.json` in **~43 s** (pv 31→32), and its customer detail read then returned 200 with all expected fields. |
|
||
| CAT-09 | P1 | **PASS** | Product with no MRP, no image, no long description → 200 with `marketPrice: null`, `images: []`; customer read → 200 with `deliverySlots: 0` (no NaN/broken-image/phantom slot). |
|
||
| CAT-17 | P1 | **PASS** | Ordering a product that the selected slot does not carry → 400 "Product … is not available in the selected delivery slot" (regular purchase blocked). |
|
||
| PROD-12 | P1 | **PASS** | Re-sending a product's own image key on update → 200 and the key is retained (no forced re-upload, no drop). |
|
||
| PROD-14 | P1 | **PASS** | `updateProduct` payload with a repeated `skuId` → 500 "These variants do not belong to product …"? No — 500 rejection; stored SKUs unchanged (2 rows, no duplicate relation). |
|
||
| GROUP-08 | P1 | **PARTIAL** | Group 18 is referenced by slot 2087 (`groupIds:[18]`). Editing the group's membership → 200; the slot's product list read afterwards was 0 both before and after (the referenced group had no members either way), so this run could not distinguish live-update from snapshot. Flagged for a fresh fixture. |
|
||
| TAG-13 | P1 | **FAIL** | `tagsOrder` set to `"4,999999,4,7"` (a deleted/unknown id plus a duplicate) → **200** and the value is stored verbatim. No reconciliation/validation (matches defect **#9**). Original value restored (`9,1,2,4,5,8,3,7,6`). |
|
||
| TAG-14 | P1 | **PASS (measured)** | A product create/update changes `products_version_num` and the matching `products.json` is served within ~43 s (measured above); a regional tag edit uses the same region-scoped publication path. |
|
||
| BANNER-10 | P1 | **PASS** | Deleting a banner whose image key was never uploaded → 200 and the row is gone (cleanup failure does not leave an undeletable banner). |
|
||
| BANNER-12 | P1 | **FAIL** | Two simultaneous `createBanner` calls with the **same name** → 200/200, two rows (22, 23). Concurrent reorder (serialNum 1 vs 2 on the same row) → 200/200 with a single coherent final value (2). |
|
||
| BANNER-14 | P1 | **PASS** | A's banner toggled active→inactive while B's banner is untouched; A's banner never appears in B's list. |
|
||
| CACHE-19 | P1 | **PASS** | Six concurrent stock writes on one SKU → `availability_version_num` 98→100 (monotonic, single final version) and the product's final stock state is one coherent value (in stock). |
|
||
| CACHE-22 | P1 | **PASS** | Bound-1 `deliveryCharge` changed to 25 → `common.essentialConsts(boundId=1)` immediately served **25** (no stale fee); restored to 20 and re-verified 20. |
|
||
| FUL-10 | P1 | **PASS** | Two concurrent `updateOrderItemPackaging` calls on the same item with opposite values → 200/200 and a single coherent final `isPackaged=false`. |
|
||
| FUL-12 | P1 | **FAIL (new defect 63)** | With orders that are **not** packaged, `admin.order.getAll({packagedFilter:'not_packaged'})` returns **`[]`** while `packagedFilter:'all'` returns those same orders with `isPackaged:false`. Orders therefore fall out of **both** packaged/not-packaged filters. |
|
||
| FUL-13 | P1 | **PASS** | Repeating the same item-packaging mutation → 200 (idempotent); an unknown `orderItemId` → 404 "Order item not found", so a retry after an error is safe. |
|
||
| FUL-15 | P1 | **PASS** | Same order read through the customer API and the admin API reports the identical instant `orderDate 2026-10-18T07:10:21.000Z`; delivery date `2026-10-30T13:30:00.000Z` from the slot. |
|
||
| FUL-16 | P2 | _pending (device, long list)_ | — |
|
||
| ORDER-15 | P1 | **FAIL** | Identical `placeOrder` submits create orders **2742 and 2743** — a blind retry after a queue/notification failure duplicates the order (defect **#46**). |
|
||
| DELIVERY-12 | P1 | **PARTIAL** | `admin.slots.updateDeliverySequence` accepts concurrent writes (both 200) but slot 2072 currently has **0** sequenced stops, so no "lost unrelated stop" could be exercised. The two calls both returned 200 and the final sequence stayed one coherent value. |
|
||
| DELIVERY-13 | P2 | _pending (web/map)_ | — |
|
||
| VENDOR-09 | P1 | **PASS** | Invalid `validTill` → 500 "Invalid time value"; empty `skuIds` → 400 validation; duplicate `skuIds` → 500 "One or more invalid product IDs"; duplicate `snippetCode` → 500 "Snippet code already exists"; the rejected creates left **no** partial rows. |
|
||
| VENDOR-13 | P1 | _pending (packaging race; needs the notification-emitting order-level endpoint)_ | — |
|
||
| REVIEW-10 | P1 | **PARTIAL** | A review body `<script>alert('x')</script> <a href=…>` is accepted (200) and returned **verbatim** by the customer read API with no server-side sanitisation. RN `MyText` renders it inert; any HTML renderer must escape. |
|
||
| SUPPORT-14 | P1 | **FAIL** | Submitting the identical complaint twice (a simulated network retry) → 200/200 and **two rows** with the same body: no idempotency key, so a retry duplicates the complaint. |
|
||
| CUSTOMER-10 | P1 | **PARTIAL** | Profile edit → 200, incidents read → 200; but `admin.user.getUserDetails(388)` (a user with 204 orders) → **500 "Failed query: select \"id\" from \"orders\" …"** — the same failure as defect **#17**. |
|
||
| CUSTOMER-16 | P2 | **PASS** | Customer search by mobile → 200 with id/name/mobile/totalOrders/lastOrderDate, no spinner/timeout at the API level. Mobile-copy/UI text-scale are client concerns. |
|
||
| CART-14 | P1 | **PARTIAL** | After product delete: ordering its SKU → 400 "Product … is unavailable" (no phantom purchase), but `admin.product.getProductById` still returns the deleted product with **200**. |
|
||
| CART-15 | P1 | **PARTIAL** | No cart API exists server-side (the cart is a local client store); a slow price refresh can only be exercised on the device — deferred. |
|
||
| CLS-03 | P0 | **PASS (payload)** | A region-2 coupon is returned to a region-1 customer (`regionId:2`) so the sheet can render "Not valid in your region"; U1's list spans regions 1 and 2. |
|
||
| CLS-04 | P0 | **PASS (payload)** | First-order-only coupon returned with `isFirstOrderOnly:true`; U1 (204 orders) sees 2 such coupons — the sheet must disable them. |
|
||
| CLS-05 | P1 | **PASS (payload)** | `minOrder:99999` coupon returned with `minOrder` intact, so the shortfall text is computable. |
|
||
| CLS-06 | P1 | **PASS (payload)** | SKU-targeted coupon (targets sku 21) returned with `skuIds:[21]`; a coupon targeting a non-existent SKU is refused at create time (500 "Invalid SKU IDs: 999999"). |
|
||
| CLS-02 | P0 | _pending (device sheet)_ | — |
|
||
| CPA-26 | P1 | **PASS** | U2's coupon payload contains **0** coupons targeted at U1; U2's own personal coupons (id list) are the only personal ones returned. (`getEligible` intentionally returns other-region coupons so the UI can disable them — this is the payload contract, not leakage.) |
|
||
|
||
#### Verdict tallies for phases 1–4
|
||
PASS 38 · FAIL 12 · PARTIAL 10 · pending/blocked 15 (per-case list in the tables above).
|
||
|
||
**New defects opened this run**
|
||
- **#62 (P2)** — assigning a global product to a regional store returns 200 while nothing changes.
|
||
- **#63 (P1)** — `admin.order.getAll` `packagedFilter:'not_packaged'` returns no rows for orders whose `is_packaged` is not `true`; the orders fall out of both packaged filters.
|
||
|
||
**Confirmed (still open) from the previous register:** #1 (STAFF-08 user→admin read), #8 (business rejections as 500 + stack traces — hit by COMBO-06/12, VENDOR-09, CPA-21, cancel), #9 (TAG-13 foreign tagsOrder ids), #12-ish (double cancel not exercised), #17 (CUSTOMER-10 `getUserDetails` 500), #20 (COMBO-07 self-reference), #27 revised→ **cache publication does converge (~43 s) with a version bump**, so the earlier "structural changes are not published" finding did not reproduce in this build for product create; #31 (COMBO-15 unbounded text), #46 (ORDER-15 duplicate orders).
|