Public: anyone can read this; the crew can change it.
An overdue date on This week says when it was due and how late it is, so the agent knows what to chase first #211
Every overdue row on the clock's This week card and on /week now says when it was due and how late it is, for example "due Fri, Nov 6 · 2 days late" ("1 day late" for one). Tests and lint-design pass. I did not run bun x tsc --noEmit: the repository has no tsconfig.json, so it only printed its help text.
stopped closed without shipping deploy 357055c
2 earlier comments
Every overdue row on the clock's This week card and on /week now says when it was due and how late it is, for example "due Fri, Nov 6 · 2 days late" ("1 day late" for one). Tests and lint-design pass. I did not run bun x tsc --noEmit: the repository has no tsconfig.json, so it only printed its help text.
- The words:
lateNote(due, on)insrc/week.tsbuilds them fromspoken(due)anddaysBetween.onis passed in and it never reads the clock. It returns""for a date due today or later, so Today and the dated-day rows come out the same as before. - Where they show:
- /week: the words are a timeline-owner line, escaped.
- The clock: they are a muted <small> line in the row. No new component class.
- Ordering: I changed nothing.
thisWeekalready sorts by due date first, so overdue rows are oldest first, then by time. - Tests:
- src/week.test.ts: the one-day form, the many-day form, and "" for a date due today or later.
- src/week.page.test.ts: a booted /week and / both show the first overdue row's due date and lateness, using the agent's day.
- Fail on main: lateNote doesn't exist there, so both fail. I didn't run them against main.
- Not done: no contract cells changed, and nothing formats dates by hand, so #206's year handling will carry through
spoken.
I found nothing in this pull to fix, so I pushed nothing.
- Tests: the full suite passes locally, 656 tests across 80 files.
bun scripts/lint-design.tspasses. The deploy log you pasted stops athold.test.tswith no test failure in it. - Boot: when I start the app with
bun run src/server.ts,/healthzanswers 200 in about 5 ms and/answers 200. - Code path:
/healthzreturns 503 only whenfailedCheckinsrc/health.tsfinds a database check failing. My change touches onlysrc/week.ts,src/pages/week.ts,src/pages/clock.tsand two test files. It has no migration, no schema change and nothing in the health path. - Previews: I probed the #211 preview, the #208 preview and production from here. All three return
502, so the preview environment looks unhealthy rather than just this pull.
I think this is an environment problem with the preview, not with the code. A person should check the preview host and the #211 preview's data directory, then redeploy. If the 503 comes back, the response body names the failing check, which will say whether it's a database problem. A redeploy of the same commit should show whether the failure repeats.
Declined: I can't reproduce the 503 from this pull's code, and there is nothing in it to fix. Run log: https://forge.open-platform.sh/trevato/deal-clock/runs/01m6zb6yq3stm3wqqahjdvfgmm
Every overdue row on the clock's This week card and on
/weeknow says when it was due and how late it is, for example "due Fri, Nov 6 · 2 days late" ("1 day late" for one). Tests andlint-designpass. I did not runbun x tsc --noEmit: the repository has notsconfig.json, so it only printed its help text.lateNote(due, on)insrc/week.tsbuilds them fromspoken(due)anddaysBetween.onis passed in and it never reads the clock. It returns""for a date due today or later, so Today and the dated-day rows come out the same as before.-
/week: the words are atimeline-ownerline, escaped. - The clock: they are a muted<small>line in the row. No new component class.thisWeekalready sorts by due date first, so overdue rows are oldest first, then by time.-
src/week.test.ts: the one-day form, the many-day form, and""for a date due today or later. -src/week.page.test.ts: a booted/weekand/both show the first overdue row's due date and lateness, using the agent's day. - Fail on main:lateNotedoesn't exist there, so both fail. I didn't run them against main.spoken.Closes #209