Make it mine's copying was inline in App and untested; it is now a pure
function with a test, so a change to either store can't make the copy
silently arrive empty. The question's buttons now say what they answer
when heard out of context. The prefs block and the spec now match what
shipped.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A reload can be served the browser's cached feed for minutes. The button
beside the data's age asks the server past that copy, swaps in a newer
feed without reloading, and says honestly what came back. Newer is
judged on events and sources, never the build stamp. The age now says
"just now" rather than "ended ago" when a fresh feed is a moment ahead
of the page's clock.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The changelog has two tiers, each tied to git: short entries shown after
a reload, and release notes summarised from the history and specs.
Ridu Weekly (1,300) and the Weekly Routine (10) are now sourced. The
first guesses of 2,100 and 10 were checked against the owner's links.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Review found that import and edit rebuilt a task from this build's parse.
That stripped any field it didn't know, and import dropped a newer
version's task entirely. Both now keep the record as stored and validate
only to display it.
An event whose goal is points can't be ticked, so it no longer counts as
waiting today: it counts with its week. The month-day field reset
itself to 1 when emptied; it now keeps a draft.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The product, storage and design records now cover F17. They say why the
task and amount stores, unlike the custom-event store, never filter on
load, which is the rule a future store would most likely copy wrongly.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Task status reuses the goal status shape, so one chip, one count and one
tally serve chores, events and tasks alike.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Review found three ordinary actions that deleted a goal this build could
not read: "It repeats daily", pressing an already-pressed Off, and "not
daily" without saying so. Only "not daily" clears a goal now, and its
warning keys off what is stored, not what parses.
The count field was bound straight to the goal, so emptying it snapped
the old number back and typing appended to it. It now keeps a draft.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The product, storage and design records now say what a goal is, why it
never marks an event done, and why an unreadable goal is kept rather than
stripped. That last rule is the one a future cleanup would most likely
undo by mistake.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The goal control lives on the checklist only: "It repeats daily" already
reaches it from a non-daily event, so a second entry point added nothing.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The goals draft stopped at three open questions; the owner answered them,
and widened the scope to weekly and monthly chores with numeric targets.
Both share one period engine, so they are specified together and built
goals-first.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Rewrites the reset table for every corrected game, records the rule that a
game's offset must make its day key name the source's calendar day, and
names Game-Time-Master as the evidence for server clocks a source does not
state (transcribed, never vendored: it is GPL-3.0).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A game's servers need not agree on the reset hour any more than on the
offset. Reverse: 1999 and Nikke spell their 05:00 out for all three
regions, so no reader's day key moves.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Forty minutes left of a two-week event is under half a tick, so the
meter rounded it down to an empty strip. The most urgent rows and the
headline looked like events that were already over.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
They are reasoning for whoever touches the interface next, which is
what docs/ holds, and the table in AGENTS.md is how an agent finds it.
The design tooling resolves both from docs/ without configuration.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
readableHue returned every dark hue untouched on the claim that all of
them already cleared 4.5:1 there. Five did not, and Fate's navy was
1.8:1, so its name under the headline deadline was nearly invisible.
Those five are now lifted in lightness only, keeping hue and saturation;
the fifteen that pass are unchanged. The owner approved relaxing the
"dark theme does not move" rule to this, and AGENTS.md and PRD F15 say so.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Document the forge responsibilities across AGENTS.md, README.md, ARCHITECTURE.md, INGESTION.md, and the CI workflow headers: GitHub Actions only builds and serves (Pages deployment) without fetching data sources, while Gitea Actions exclusively runs the scheduled refresh and publishes the container image.
The Blue Archive Wiki (bluearchive.wiki/wiki/Events) redesigned its
schedule, replacing the per-version MediaWiki tabber with a unified
#eventtable displaying JP and Global periods side-by-side.
- Parse #eventtable resolving GL Period, event titles, and article links
from .event-region-gl, with fallback to .event-region-jp.
- Maintain backwards compatibility with legacy tabber schedule format.
- Add pinned fixture fixtures/ba/bawiki-events-2026-09-17 and tests.
- Update snapshot, AGENTS.md, ARCHITECTURE.md, and INGESTION.md.
Update README, AGENTS.md, ARCHITECTURE.md, and INGESTION.md to reflect
that the scheduled refresh and container image publishing run on Gitea
Actions under .gitea/workflows/, while GitHub Actions under .github/
builds with BASE_PATH and deploys to GitHub Pages when synced.
The pipeline change made a set of sentences false, and they are the only
record of why things are the way they are, so they move with it: the CI
section and deploy target in README, the cron and BASE_PATH notes in
ARCHITECTURE, the documented User-Agent in INGESTION, and the contact
URL in SOURCES.
Two comments in source were false in the same way and are fixed here
rather than left for the next reader. `serve.ts` said compression was
"the whole difference between the Docker image and the deployed site" —
the image *is* the deployed site now, so that negotiation is the only
compression any reader gets. `robots.ts` justified product-token
matching with "our contact URL contains the string StereotypicalCat",
which it no longer does; the hazard is the arbitrary substring, not that
one, so the comment now says so and the regression test keeps the URL
that actually caused it.
What is deliberately NOT rewritten: game8.co's refusal stays in the past
tense with the Gitea result marked unverified, in AGENTS.md, README,
SOURCES and FEEDBACK alike. A new runner address is the variable that
failure turns on, so it may well be served — but "may well be" is not
evidence, and the four places that would have to change are named
together so they change together once a cycle has actually run.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
Add karendar.com as the event source for Punishing: Gray Raven (PGR).
Karendar publishes a dedicated Global event calendar with exact UTC
timestamps in clean server-rendered HTML.
- Register 'pgr' in GameId and GAMES registry (#CC292B, Kuro Games).
- Add parseWeekdayDayMonthYearUtc in src/ingest/dates.ts with unit tests.
- Implement karendarParser in src/ingest/parsers/karendar.ts reading
'week', 'ongoing', and 'upcoming' sections, mapping indefinite/permanent
ends to endsAt: null, and categorizing tags to EventType.
- Register pgr-karendar-events in src/ingest/adapters/index.ts.
- Add pinned fixture, expected output, and comprehensive test suites.
- Update documentation in AGENTS.md, README.md, docs/INGESTION.md,
and docs/SOURCES.md.
Add genshin-impact.fandom.com as a second source for Genshin Impact,
which has been fixture-backed in CI because Game8 CloudFront-blocks the
Actions runner with 202s.
- Add fifth Fandom template parser in src/ingest/parsers/fandom.ts
extracting titles from caption display text, stripping dated subpage
suffixes, and checking currency via latestBoundaryMs.
- Register genshin-fandom-events in src/ingest/adapters/index.ts and
set genshin-game8-events priority: 10 so existing event IDs survive
near-match ties, preventing localStorage completion marks from
orphaning.
- Explicitly sort adapter groups by priority descending in
scripts/build-feed.ts to align with parseGame.
- Pin fixtures and add test coverage in test/adapters/game8.test.ts.
- Update AGENTS.md, README.md, docs/INGESTION.md, and docs/SOURCES.md.
Patch cycles are ~6 weeks, so a span over 180 days previously indicated
a misread year. However, real year-long events exist in this domain:
Genshin's anniversary 5-star selection runs 336 days (2025-10-22 to
2026-09-23), correctly dated on its wiki, and Fire Emblem Heroes runs a
seven-month new-player banner (docs/SOURCES.md § 12b).
Widening the ceiling to 365 days admits genuinely year-long events while
still catching misread years, which place the end 12 months out (span +
~365, or 405+ days for a 6-week event).
Update test/adapters/game8.test.ts duration test, AGENTS.md domain rules,
and docs/INGESTION.md sanity checks.
Infinity Nikki's wiki prints "There are no Events in this category"
between versions, and the refresh runner has honoured that since
2026-09-03 — it stores the empty parse as the source's real answer
rather than letting a quiet lane reach the broken tier. The feed build
never asked, so the same page arrived at CI as parsedCount 0,
indistinguishable from a parser that has stopped reading a redesigned
page, and brokenSources failed every build while every refresh stayed
green.
The runner's verdict cannot travel on its own: only a parser has seen
the page, and by the time brokenSources runs there is nothing left but
the feed. So the fact rides on SourceHealth, defaulted so an older feed
the service worker cached keeps validating and reads as the strict
answer.
Both ends now ask it the same way — of an empty parse only, from the
page's own words only — because a redesign yields zero rows too, and
excusing that is the silently emptied calendar the gate exists for.
The rule sits in a module rather than in build-feed.ts, which writes
public/ and so runs a build if a test imports it. That is how the two
ends drifted apart with nothing to catch it.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
The two preceding changes make sentences in docs/ false. INGESTION.md
described `canParse` as a structural check without saying that "structural"
excludes the presence of a row — which is exactly the drift that let this
bug in — and described the Stage 1 gate as rejecting every empty parse.
SOURCES.md § 11 left a reader to conclude that an empty Infinity Nikki lane
means the source has died.
AGENTS.md gets the rule in both places somebody would look: § Working on
parsers, next to the bullet about a source resting on one row, and § Fandom
where the fourth template is described.
The general lesson is worth more than the incident. Any page that can list
nothing can produce this, so the guidance is about what identifies a page
versus what happens to be on it — not about this wiki.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
The footer's staleness notice is an instruction, and the only remedy it
offers is "go and check that game's source page". That is not something a
reader can act on for a game they turned off and cannot see a single row
of — and with nineteen lanes, naming eleven they do not play buries the one
they do. A new lane arrives switched off (F8), so an untouched install was
being warned about most of the catalogue.
Scoped in two places, not one: the named list, and the count the
summarising branch measures against. Without the second, that branch would
never fire for a reader with most of the calendar off and they would get a
list where a sentence was the readable answer.
The notice now also states whose games it counted, in the words NextUp
already uses for the same set. Narrowing what the footer measured while
still printing "nothing has refreshed" would turn a claim about four games
into one about eighteen, in the one paragraph on the page whose job is
being trusted about age.
The headline age and the credits stay whole: the first is a fact about the
feed and feeds the bug form, the second is attribution owed to every source
we read regardless of what is on screen.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
Review pointed out that `dailyGroups` was well covered and the wiring to it
was not: passing the wrong pref from App, binding the checkbox to a
neighbour, or hardcoding `true` in Dailies each shipped with the suite
green. Three of those are now caught by rendering the strip and the settings
panel rather than the function behind them — verified by performing the
mutations, not by reading the code.
The fourth, App handing Dailies the wrong pref, is still uncovered: no test
renders App, and closing it needs a feed and a click library this project
does not have.
Also pins the default. Flipping that one line silently empties Today's
dailies for every existing reader, which is the whole argument of the
comment above it, and nothing was watching — `defaults` is exported so the
value and the upgrade path can both be asserted.
And DATA-MODEL's enumeration of the prefs blob had gone stale; AGENTS.md
names it as the doc to update when a stored key moves.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
Review found the new settings row opening nothing for a rule whose `until`
precedes its `startsAt`. That parses — nothing ties the two together — and
yields no occurrence at all, so nextOccurrences and the anchor fallback both
come back empty, nearestOccurrence returns null, and openRow cannot resolve
a bare rule id. It was the undeletable record this index exists to rescue,
now with a button that lies about it.
So resolution moves out of App into displayEventFor, with the rule itself as
the floor: a stored id always names something the reader can edit and
delete, whatever the rule does or does not generate. Exported because
nothing here can click and no test renders App, which is exactly how a dead
button shipped — the chain is now covered without a DOM.
Two of the caption tests could not fail, proven by mutation rather than
read: deleting the whole "starts" branch and removing the null-end guard
both left the suite green, because `not.toContain("ended")` was never
watching the branch at risk. They assert what the caption says now.
And a repeating series that has stopped said only how often it repeats — in
the one place whose job is explaining why an event is on no other surface. A
healthy cadence explains nothing; it says when it stopped.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
An event of theirs that had ended was on no surface at all. Every list and
the board drop a row once its end has passed, and settings only counted the
events a game held rather than naming them — so a one-off became unreachable
the day it finished: impossible to edit, and impossible to delete out of a
store nothing else can see. Repeating events escaped only because their
occurrences roll forward.
So settings names them now, under the game they were filed against, each row
opening the same detail sheet a row on the front page does. Nothing about
how they are managed changes; what was missing was the way back to them.
Two things the index has to get right or it leaves the same hole it closes.
The lists hold occurrences, never rules, so a repeating rule's own id opens
nothing — nearestOccurrence bridges that, and answers for a finished series
too by falling back to the first occurrence, since a rule whose `until` has
passed would otherwise be exactly as stuck. And an event filed under a game
we track has no row of theirs to nest under; the form allows that, so those
get their own heading rather than trailing the list and reading as though
they belonged to whichever game came last.
Each row says why it is not on the front page — ended and when, or its
cadence — because a list of bare titles leaves you guessing which of two
entries is the dead one.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
Review found the headline feature unreachable. `valid` still demanded an end
date unconditionally, while every sibling rule had been gated on
`datedWindow`. A fresh form starts with endKnown true and endDate empty, so
picking weekly hid both the field and the checkbox and left Save disabled
with nothing on screen a reader could do about it. Editing an existing event
was fine, which is why no test caught it: every static render drives the form
through `initial`, and cadenceOf never returns a preset for an event that has
an end.
So the three states that needed a click to reach are now pure functions that
do not. endStated pins exactly that case. repeatOf takes the cadence and
answers for all five, which also makes `until` surviving the preset path
provable — mutating it away previously left the suite green.
Two more from the same review. The unknown-end note keyed off the cadence, so
custom with the repeat set to never promised a countdown that would never
arrive; it keys off the rule now, which is what actually decides. And a delay
whose start and end disagree about having a time of day can never land a
whole unit from the anchor — unstatable, not merely unstated — so it says so
rather than disabling Save in silence.
"State it myself" was also a one-way door: nothing ever cleared it, so the
measured reading could not be recovered without abandoning the form.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
The repeat controls sat after the dates and applied to whatever was there,
which meant a weekly chore still had to answer an end date it has no honest
answer to. Asking first inverts that: the cadence decides which dates are
even questions.
A preset carries no window. Weekly means the week is the window, so there is
no end to type and no ignorance to admit — five fields instead of eight, and
it stores exactly what the model already renders as back-to-back
occurrences. A dated recurring event is therefore a custom, which is where
the forever/delay control now lives; a one-off shows no repeat machinery at
all.
cadenceOf derives which of the five a saved event opens in, so a rule made
before this control existed opens in whichever answer describes it. Nothing
about the schema moved.
Switching cadence hides the end date rather than clearing it: hiding a field
and quietly discarding what is in it is how a form loses somebody's work
when they change their mind back.
The duplicate note was found by rendering the form and looking at it. Both
of the older notes explain the end-date field, so a preset — which has no
such field — was showing two sentences that said nearly the same thing.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
The PRD said a reader's rule could optionally stop on a date; the until
control was deliberately descoped from the form during planning, and the
PRD was written as though it shipped. Corrected the PRD to describe what
actually ships, and noted that the field still exists in the schema —
reachable today only by importing a file that carries one.
That gap had a second-order bug behind it: because the edit form rebuilt
`repeat` from scratch on every save, hard-coding `until: null`, a rule that
already carried a non-null `until` had it silently reset to eternal on any
edit at all, including a pure title fix. Extracted the `repeat`-building
logic into `repeatFrom`, which now carries `initial?.repeat?.until` forward,
and is exported so the behaviour is provable without a submit nothing in
this suite's renderToStaticMarkup tests can click.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
Three sentences were false once a rule could bound an occurrence. DATA-MODEL
gains the derived occurrence key and what a reschedule costs; F13 gains the
rule and the reason an occurrence need not state its end.
SOURCES' arustats note stays, with the answer beside it: the reason
scheduleBosses is unread was a design gap, that gap is closed, and reading
it is now a parser change. Nothing on the ingest side has moved — this note
is what stops the next person re-deriving why.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
The "ancient anchor" test expected three occurrences and omitted 31 August.
With a daily rule anchored at 09:00 and no stated end, the 31st runs to
09:00 on the 1st — inside a window that opens at midnight on the 1st. The
brief's own rule is that an occurrence overlapping either edge is included,
so four is right and three was the plan disagreeing with itself.
Found by the Task 4 implementer, which corrected the test and left the
implementation alone. Verified by running occurrencesOf directly.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
Task 2's version called occurrenceId twice with identical arguments, which
any pure function satisfies — occurrenceId never takes a title, so it could
not fail the thing it claimed to pin. Task 5 passes the whole rule into
asOccurrenceEvent, so a rename is a real input there and the assertion has
something to bite on.
Found by the Task 2 review.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
Pre-flight scan finding. The function is implemented in Task 2 and its four
tests sat in Task 4's block, so Task 2 would have shipped untested code and
Task 4 would have tested a function it did not write. Task 2's stated count
of 23 already assumed the corrected placement.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
Eleven tasks from the design, each with its own test cycle and commit.
Ordered so the schema field lands before anything derives from it and the
store migration is proved before a single row is expanded.
Three things the plan had to settle that the spec left to the implementer.
The overlap check is one exported predicate, because a form that restates
its schema's rule drifts from it and starts refusing saves that would
succeed. The timeline's expansion needs App to pass the same four scope
filters the lists use, so the predicate is extracted rather than copied.
And custom-ui.test.tsx renders statically with no testing-library in the
project at all, so anything needing a click is stated as a pure function
and tested directly rather than through a harness that does not exist.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
The app can express a window closing and a day repeating, and nothing in
between — daily.ts counts in days and stops there. That gap is already
costing parsed data: arustats scheduleBosses is read and discarded because
"a recurring rotation with no end is not a deadline" (docs/SOURCES.md:740).
A rule fixes exactly that. Content repeating every fourteen days ends, at
the latest, when its next occurrence opens — a boundary entailed by the
interval rather than invented for a form. So an occurrence may state no end
and still be a deadline, which is the distinction the spec turns on.
Proposes it first on the reader's own events (F13), where the rule is typed
rather than fetched: no parser, no review gate, no GachaEvent change, and a
tested recurrence model in shared/ for the ingest side to adopt later.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
SOURCES.md § 13 declined this game's only known source partly because every
boundary was a week bucket, and we have now built a different source with that
same objection standing. Leaving that contradiction unwritten is how a later
pass rediscovers the trade the expensive way, so § 14 records the decision
itself: the conduct evidence, which three of § 13's four objections arustats
answers, which one it does not, and that any source stating a date per event
retires it. § 13 is marked superseded rather than rewritten — its verdict on
marisaimpact.com still stands and was not re-tested.
AGENTS.md gets the same warning where a parser author will hit it, because the
estimate is not visible from the code: the numbers look like every other
day-precision date in the project.
Also corrects the counts this game made false — nine parsers, twenty sources,
nineteen games. The "stopped working at eighteen" passages are left alone: they
date the settings-panel redesign, and bumping them would falsify the history
rather than update a count.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
Two things the last two commits taught, written down where the next person
looking at a parser will hit them.
The date table now carries the slash-dated open range, and the prose beside it
says the year is required and why — that reader is tried last, so a guess there
reaches the feed unchallenged. It also names the second copy of the open-end
vocabulary, because the two lists being out of step is not a hypothetical
failure any more.
The second is the one worth having on record. `eventCount` in a snapshot's meta
is not a count of the page. Endfield's Game8 lane read 1 for a week — one row of
an *upcoming* table, while four live events sat in a card grid the parser cannot
read — so the first time the page moved, the lane read 0 and the gate reported a
shape change. Neither number described the page. A count far below what the page
shows is a shape being skipped, and it is worth finding before it reaches zero
and the gate has to guess on your behalf.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
Both files stated the h4 rule as a flat prohibition — an unrecognised h4
must never claim the event title — which is no longer what the code does.
They now state the narrower claim: an h4 may fill an empty event title but
never take one, with the reason each page needs it. Genshin's labels arrive
after its h3 has claimed the slot; Wuthering Waves' block sits under a
section heading that leaves the slot empty.
INGESTION.md gains the block as shape 9, and the part that is easy to get
wrong on the way back: its heading has to be *in* the section vocabulary
rather than merely fall through, because an unrecognised heading is read as
an event name and would spend the title on the section.
AGENTS.md also gains a bullet for where a missing blurb hides. Two
templates now put unlock conditions in the cell where a description
belongs, and in both the real prose is in a per-event section further down
the page — under an h3 on one and an h4 on the other.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
They were one group called "Reading", grouped on "both are how do I read
this?" — which summarised as "Europe · Dark", two unrelated answers joined
by a dot, under a name for neither of them, in a panel whose whole premise
is that a closed line answers its own group.
The region is not a reading preference. Region-scoped ends, a date printed
without a time, and every daily streak are all cut on that server's clock,
so it is the one control in this panel that can make a countdown wrong —
and it was filed behind a word for the theme. Split, each line answers one
question, and the region group says what it governs when opened.
The eyebrow above each pill row goes with it, since the group summary now
asks the question the eyebrow was repeating; aria-label keeps the
accessible name, which is what stops six unlabelled buttons in a row from
being all a screen reader hears. The summary assertions are looked up by
group name rather than by index, so splitting a group no longer moves every
assertion below it onto a neighbour.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
Looked at on a 1440px screen it was the one block on the page not using
the column it sits in. The rows were held to max-w-3xl inside an 1150px
shell, but the section's own rule spans the shell — so every hairline
inside the panel stopped a third of the column short of the one above it,
between a two-column checklist and a three-column footer that both go
full width. The measure was there to keep a name and its state reading as
one line, but that pairing is what "running now" already does with a
title and its countdown directly above, so full width is the house
pattern rather than a stretch. Prose inside a group keeps its measure.
The game list was the worse half: eighteen rows of four controls at
max-w-md is a thousand pixels of ribbon down the left with two thirds of
the screen empty beside it. It takes two columns past lg once there are
nine or more games, flowing down then across so the numbers still read in
a straight line and an arrow still swaps a row with its neighbour. Below
nine, and on a phone, it is unchanged.
The last group also loses its bottom rule, which was a second hairline a
few pixels above the footer's own.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
`fetchFeed` cannot start until main.js has downloaded, parsed and mounted React,
so the one request this page exists to make was the last one to leave. Nothing
about the URL depends on the bundle, so the preload scanner can have it in flight
in the first few milliseconds instead.
Measured on a 4G/4x-CPU profile with a cold cache and gzip on: the events reach
the screen at ~470ms instead of ~570ms, and the loading state stops being
something a reader can see rather than merely being brief. First contentful paint
is unchanged within noise — the win is that the feed is off the critical path
entirely, arriving at 216ms while the bundle is still downloading.
`crossorigin` on that link is not decoration. `as="fetch"` has to match the
credentials mode of the `fetch()` that follows, and a mismatch is not a no-op:
the browser throws the preload away and downloads the feed twice, which is worse
than not preloading at all. Verified by counting requests in a real browser,
because no test here can see it.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
The progress store merged like the mark stores do, keeping whichever copy had
the earlier timestamp. That is right for a mark — `at` is when the reader made
it, membership is the whole fact, and the oldest stamp is the truest one — and
wrong here, because this record *is* the data. A status, an effort, a note and
whether an event repeats all live in it, and `at` says when one of those last
changed.
So the earlier copy winning discarded every edit made after it, in both of the
directions an import actually happens in: restoring a backup taken before an
evening's work undid the evening, and importing an old file onto a device with
newer progress rolled the device back. Neither is recoverable — there is no
account and no server holding a second copy.
The later record wins now. Nothing is removed in either direction, and taking
the maximum of two timestamps is as order-independent and idempotent as the rule
it replaces. `mergeProgress` came out of the hook to be tested, which is also
how the ignored store's opposite rule is now written down rather than assumed.
Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>
