The A of ITA had lost its diagonal: at 260px the hairlines of a Didone
fall below a pixel, downscaling averages them into grey, and the
panel's mid-grey cut then discarded them. What reached the shop window
was a detached serif and something closer to a lambda.
Letting more grey count as ink keeps them whole. 165 was not enough —
the diagonal came back dotted — and 195 holds it solid without
thickening the stems into mud. The number belongs to this particular
drawing, so `npm run brand:variants` renders the grid of widths and
thresholds it was chosen from, magnified, for the day the logo changes.
Width stays at 260: the header shares its line with the date, and the
main block is narrower now the week runs down the right-hand side.
The panel will redraw once when it next wakes, since the image and
therefore its content hash have changed.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012cSY9pVhZmJUKNN7wf1Myd
A bottom strip gave each day a 114px column, and that width was what
capped the type size. A side band gives 142px of usable width and a 68px
row, so the hours go from 19pt to 22pt, the day labels from 20 to 23,
and each day fits on one line — "Mar 11:00–18:00" reads at a glance
where a column had to be read downwards.
Right rather than left, for two reasons the renders make plain: on the
left the logo is pushed into the middle and loses its place as a
signature, and the eye meets the week before the day's status. OUVERT is
what has to be read first from the pavement.
The times stay whole. Only the spaces around the dash go, which is
enough to hold 22pt without abbreviating anything: a shop's opening hour
is not a detail to shorten.
The idea came from the shop owner, and testing it surfaced a defect that
had been in production all along. Rendering the worst case the data
model allows — three periods in a day, none on a round hour — showed the
day's hours wrapping, pushing the layout down, and the English line of
the banner falling off the bottom edge. A panel does not scroll and does
not reflow: what does not fit is lost.
Type size is now fitted to the space available, with a floor below which
it will not go, and nothing may leave its box. The width estimate is
deliberately crude: measuring glyphs would mean loading the font in that
module and would still be an estimate, since the renderer applies its
own kerning. Erring small costs a point of type, erring large costs a
broken screen.
Three periods a day will reportedly never happen. The protection stays
anyway — it costs nothing and removes the need to remember.
`npm run screen:fonts` renders every arrangement and both alternative
typefaces, including the worst case, because choosing type for a 1-bit
panel is done by looking at it thresholded.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012cSY9pVhZmJUKNN7wf1Myd
Twenty-six tests across two viewports, covering what the spec asks for:
changing today's hours reaches the panel, a message survives the
translation service being unavailable, a closure period closes the days
it covers — plus the authorisation paths and the sign-out regression.
They run against the standalone build served the way the container
serves it, not `next start`, which refuses to work with standalone
output anyway. The suite therefore exercises the artifact that ships
rather than a second arrangement that could drift from it.
Sign-in mints the session cookie Auth.js would have issued rather than
driving Authentik. What is under test is the application's behaviour for
a given role; the handshake itself is verified against the live provider
separately, and standing up an identity provider per run would trade a
lot of machinery for coverage of somebody else's code. The secret lives
in one module imported by both the config and the fixtures — when it
differed, every signed-in test failed at once while looking like an
authorisation bug.
Database access goes through plain SQL rather than the Prisma client,
whose generated module format Playwright's loader and Next's bundler
disagree about. That traded one problem for a subtler one: node-postgres
parses a DATE column into a local-midnight Date, so reading it back
shifted the day at UTC+2. Dates are read as text now.
The mobile profile runs on Chromium: WebKit needs system packages only
root can install, and a suite nobody can run locally is a suite nobody
runs. The config says how to switch to the real engine.
Two real defects surfaced, both found by the tests rather than by
reading. The seven "Ouvert" checkboxes on the hours page were
indistinguishable to a screen reader; each now names its day. And on a
phone the signed-in address appeared nowhere at all — the header hides
it to save room — so nobody could tell which account was about to sign
an audit entry on a device the shop shares. It is on the dashboard now.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012cSY9pVhZmJUKNN7wf1Myd
Multi-stage build on node:22-alpine, standalone output, non-root user,
healthcheck on /api/health, and migrations applied by the entrypoint
before the first request. A failed migration stops the container rather
than serving an inconsistent database.
Getting the Prisma CLI into the runtime image took three attempts and
the reasoning is recorded in the Dockerfile. Copying it out of the build
stage leaves its transitive dependencies behind; patching them in one at
a time is a losing game. It now gets its own stage and its own tree,
with the schema and prisma.config.ts beside it, and the entrypoint runs
from there so every import resolves locally. The version is read from
our own package.json so it cannot drift from the generated client.
Two things had to change to build without a database, which a build
container rightly does not have. prisma.config.ts no longer reads the
URL through prisma's env() helper, which throws on a missing variable
even for `generate`. And lib/db.ts creates the client on first use
rather than on import: Next imports every route module while collecting
page data, so a module that threw on import failed the build with an
error naming whichever route was analysed first, which says nothing
useful. The failure now lands on the first query, where it belongs.
Verified by running the image against a real database: migrations
applied, cron scheduled in Europe/Zurich, a device paired, and the panel
image served as a genuine 1-bit 800x480 BMP — so satori, resvg and the
vendored fonts all work on musl. The image hash came out identical to
the one produced on the glibc host, which is the reproducibility the
vendored fonts were for.
The production overlay publishes through an existing Traefik, drops the
host port, mounts the filesystem read-only, and adds a nightly dump kept
for a fortnight.
README and DEPLOY are in French and cover what actually bites: the panel
receives nothing and only updates when it wakes; the issuer must match
to the character; the captive portal URL takes no trailing slash; a
rollback across a migration needs the dump, because Prisma does not
undo one.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012cSY9pVhZmJUKNN7wf1Myd
Mirrors the button on the settings page, for verifying a deployment from
the host before anyone signs in, and for watching the cache work.
Measured against the live service, which settles a question the spec had
guessed at: a short phrase takes between 4.6 and 10.7 seconds, because
the endpoint really does start a Claude Code process. The ten-second
timeout the spec assumed would have failed intermittently on exactly the
kind of message this shop writes. Thirty seconds stands.
Cached answers come back in 3 milliseconds, so the cache is worth
roughly three thousand times its complexity.
Quality checked on real phrasing: times and dates survive unchanged
("Ouverture retardée à 14:00" to "Delayed opening at 14:00 today"),
which is what the prompt asks for and what the screen needs.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012cSY9pVhZmJUKNN7wf1Myd
The rolling twelve-month window is fetched from openholidaysapi.org each
night at 03:00 local, and can be triggered from the page, from
POST /api/admin/holidays/sync, or from `npm run holidays:sync` for the
first run after a deployment.
The calendar is fetched twice, once per language, and the two answers
joined on the entry id. Holiday names are proper nouns with established
English forms — "Jeûne genevois" is not something a translation model
should be improvising, and this costs one extra HTTP call.
Two properties are load-bearing and tested against a real database.
The sync is idempotent: running it twice leaves exactly what running it
once did, verified live as well as against a mock. And it never touches
`isAutoClosed` on an existing row — that is the shop's decision, not the
API's, and a nightly job quietly reopening a day the owner had closed
would be invisible until someone found the door locked.
When the API is down the local cache is left untouched and the failure
is recorded with its timestamp, so the page can say how stale the
calendar is rather than showing nothing. Retries widen the gap between
attempts; the nightly job can afford to wait, the shop cannot afford a
stale calendar for a day.
node-cron runs inside the application process rather than an external
cron hitting an endpoint: one container, one shop, no second instance to
coordinate with, and no trigger endpoint to protect and document. The
reasoning is recorded next to the schedule.
Verified against the live API: nine Geneva holidays, both languages,
including the cantonal Restauration de la République.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012cSY9pVhZmJUKNN7wf1Myd
The application is published at horaires.ita-ito.com. One origin serves
both the panel and the browser, so the OIDC callback, the image URL the
device is handed and the documentation all follow from this single name.
Port 3000 is already taken by the facture_ocr stack on the development
machine, so the app listens on 3010 there and the reason is written next
to the setting rather than left to be rediscovered.
The fixtures and the frozen payload snapshot move to the real domain too:
a contract snapshot carrying a hostname that never existed is a small
puzzle left for whoever reads it next.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012cSY9pVhZmJUKNN7wf1Myd
With the move to BYOS there is no TRMNL cloud to interpret a Liquid
template, so the application draws the screen itself.
The pipeline is satori (flexbox to SVG) then resvg (SVG to pixels) then
a hand-written 1-bit encoder. Chromium was the alternative and was
rejected: half a gigabyte and a real memory appetite in the runtime
image, for a screen of eight blocks. Playwright still handles the E2E
tests, in its own pinned image, never in the application one.
The payoff is testability. The layout is asserted on the element tree
and the geometry on the SVG text, so a rendering regression shows up as
a readable diff instead of a pixel comparison. The BMP and PNG encoders
are verified field by field against their specifications, including an
independent CRC-32 for the PNG: a device rejecting a malformed image is
expensive to debug from a shop window.
Two properties are pinned because the battery depends on them: the same
payload must produce byte-identical output, and changed hours must
produce different output. The filename handed to the device is a hash of
these bytes, and the firmware skips the redraw when it is unchanged.
The fonts are vendored into public/fonts and the logo into public/brand,
both committed. Rendering must not depend on an install tree, a CDN or
the network, or the bytes drift and the panel wakes for nothing.
`npm run screen:preview` writes a real 1-bit image plus a magnified view,
which is where clipping and thin strokes give themselves up. That is how
the week strip was caught clipping and condensed.
`npm run brand` rebuilds the assets: it locates the wordmark band in the
shop logo rather than hard-coding offsets a future revision would break,
and thresholds it with the panel's own encoder. sharp is a devDependency
used only there; nothing at runtime needs it.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012cSY9pVhZmJUKNN7wf1Myd
vliaudatandClaude Opus 5