← Back to gallery

U7.2 — Session time range on Daily Sales Report PDF PROD

2026-05-27 Gate 2 ship. Third follow-up touch to the closing-session PDF today (after U7 + U7.1). Narong: "under As of Date, can we include From Time → To Time?". The "As of MM/DD/YYYY" boxed cell now has a second smaller line below showing the session's open/close window — HH:MM → HH:MM for closed sessions, HH:MM → (open) for X (mid-session) reports.

Summary

Status: 12/12 prod · 51/51 regression · 63/63 total · awaiting Gate 2 approval
Commit: nix-cafe 36b4520
Files: 2 nix-cafe files modified · ~40 LOC net · no migration · no backend · no schema
Source: Narong's Telegram 2026-05-27 after U7.1 ship.

What shipped

"As of" boxed cell is now two lines. Date stays the primary top line ("As of MM/DD/YYYY", size 10). New secondary line below renders the session time range smaller + muted (size 8.5). Box grew from 22pt × 140pt → 34pt × 160pt.
Closed sessions render the actual close time: e.g. 09:00 → 17:30. 24-hour format via Intl.DateTimeFormat("en-GB", { hour: "2-digit", minute: "2-digit", hour12: false }).
Open sessions render the upper bound as the literal "(open)": e.g. 09:00 → (open). Using "now" or the PDF generation time would mislead — for an X (mid-session) report the close time genuinely isn't known yet, so the literal is the most honest signal.
Same timezone as the date: time range uses the same tzForDate = shopTimezone || tenant.timezone resolution as U7.1, so the date and the time range stay consistent. Both reflect the BRANCH's local clock for multi-country tenants.
Arrow encoding fallback: pdf-lib's StandardFonts WinAnsi encoding doesn't include U+2192 ("→") natively. The builder's safe() sanitizer would strip it. The new drawAsOfCell checks for arrow-loss after safe() and substitutes ASCII "->" when needed. (On prod, pdf-lib's text encoder embeds the arrow correctly, so the visible PDF shows "→".)

Files shipped

MOD nix-cafe/lib/reports/closing_session_pdf.ts — sessionTimeRange: string on input interface; drawAsOfCell grew 22pt → 34pt to fit two lines + new timeRange param + arrow encoding fallback
MOD nix-cafe/app/api/cafe/sessions/[sessionId]/closing-session-pdf/route.ts — new fmtHHMM helper formats both opened_at + closed_at via Intl en-GB 24h with timeZone: tzForDate; composes sessionTimeRange = `${fromTime} → ${toTime}`; open sessions branch to literal "(open)"

12/12 prod checks ✓

✓	Find a closed lumiere session w/ orders + its branch (Z test) — session `b967eb21`, shop "Lumière BKK1"
✓	SSO-login lumiere owner
✓	Z test: fetch closing-session PDF (200 / application/pdf / closing-session- filename / %PDF magic)
✓	Closed session PDF: title 'Daily Sales Report Z' + section headings
✓	U7.1: PDF header shows shop name + tenant name (branch labeled)
✓	U7.2: PDF As of cell shows session time range — at least 2 HH:MM endpoints extracted from PDF text for the closed session
✓	U7.1: Content-Disposition filename includes shop code prefix (`-bkk1-`)
✓	Closed session PDF: legacy 'Order#' / 'Methods' table headers gone
✓	Counts in PDF body match raw SQL: Number of transactions: 1 + Transactions Refunded: 0
✓	Find an open lumiere session for the X test
✓	X test: open-session PDF has title 'Daily Sales Report X' AND U7.2: shows "(open)" upper bound
✓	No 5xx HTTP responses during the suite

Real rendered PDFs from prod (with U7.2 time range)

Fetched live from lumiere-coffee.nixtech.app. The "As of" cell in the top-right corner of each PDF now shows two lines: date primary, session time range smaller below.

Closed session — `HH:MM → HH:MM`

Open session — `HH:MM → (open)`

Regression sweep 51/51

51/51 regression checks green — no regressions from this push.

test-phase1-prod.mjs	11/11
test-phase2-sso-outdoor-prod.mjs	6/6
test-m1-prod.mjs	10/10
test-r7-prod.mjs	14/14
test-r8-prod.mjs	4/4
test-phase2-cafe-multishop-prod.mjs (solo)	6/6

narongix chain ran sequentially and was first-attempt green this time (no r7 flake retry needed — clean first run). phase2-cafe-multishop ran solo per feedback_phase2_cafe_multishop_solo_retry — first-attempt green (10th validation today).

Mid-Gate-2 note: deploy-race retest pattern

First prod test attempt ran ~75s after push and failed with "expected at least 2 HH:MM time stamps; found 1" + "'(open)' upper bound missing" — the new U7.2 features were missing from the rendered PDF. wrangler deployments list confirmed the CF Workers Build hadn't landed yet (last deploy was U7.1 at 06:24 UTC; my U7.2 push was 06:51 UTC). Used a backgrounded until [ ... ]; do sleep 15; done loop polling wrangler deployments list for the new timestamp prefix, then auto-reran the test. Deploy landed within ~30s of the loop starting and the retest came back 12/12. Clean recovery path — no manual intervention.