Commit Graph
160 Commits
Author SHA1 Message Date
kristianwindandClaude Opus 5 d8f5f6e236 feat(noter): nævn en person i en note, og den er delt med dem
At skrive nogens navn i en note *er* at ville have dem til at læse den.
Indtil nu var det to handlinger — skriv navnet, åbn panelet, find
personen, del — og den anden blev glemt, så noten stod med en besked til
en, der ikke kunne se den.

Nu deler `@Navn` noten med den, der bliver nævnt, og siger det til dem.
Editoren foreslår personer efter `@`, ligesom `#` foreslår projekter.

Slået op i navnene frem for tolket ud af teksten. Et navn kan have
mellemrum i sig, og et udtryk, der skulle fange dem alle, ville også
fange sætningen bagefter: `@Sofie i morgen` ville blive til en person,
der hed noget i den retning. Her er der ikke noget at gætte — der står et
navn, eller der gør ikke. Det er også dét, der gør det forsvarligt at
dele på: adgang må ikke afhænge af, hvor langt et regulært udtryk valgte
at læse. En mailadresse er ikke en omtale, det længste navn vinder, og
ejefaldets s hører til sætningen.

Kun nye omtaler, sammenlignet med teksten før gemningen på samme måde som
links: en note gemmes ved hver tastepause, så uden sammenligningen var
hver gemning en ny deling og en ny besked om det samme navn.

Kun ejerens omtaler deler. En medredaktør må skrive i noten, men ikke
give den videre — ellers kunne den, ejeren har delt med, skrive et navn
og dermed dele ejerens note med en, ejeren aldrig har hørt om. Det er
den samme regel, panelet følger.

Læser, ikke medredaktør: en omtale siger "se på det her", og det mindste,
der opfylder sætningen, er det rigtige at give. En rolle, ejeren har
hævet i panelet, sænkes ikke igen af en omtale.

Og én besked pr. handling: den, der lige er blevet nævnt, får "nævnte
dig" og ikke også "rettede en note" om den samme gemning.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-08 12:08:12 +02:00
kristianwindandClaude Opus 5 8ee9fbedfb feat(noter): del en note med en, der ikke har en konto endnu
En note kunne deles med dem, der allerede var på instansen — og kun med
dem. Skulle en anden læse med, skulle man først invitere dem ind, vente
på at de oprettede sig, finde noten frem og så dele den. Fire skridt og
en ventetid for det, der er én sætning: lad Sofie se den her.

Nu er der ét felt. Man skriver adressen, og serveren ved, om der er en
konto bag: er der det, bliver det en deling med det samme, og ellers en
invitation, der bliver til delingen i samme øjeblik, kontoen oprettes.
Hvem der har en konto, er ikke noget, man skal vide for at dele en note.

Rækken er den, der allerede fandtes. En invitation bar i forvejen en
e-mail, en rolle og et forseglet token; det eneste, der manglede, var at
kunne sige, hvad den var en invitation *til*. note_id ved siden af
project_id, med fremmednøglen — dén, der rydder invitationen væk, når
noten slettes, inden nogen når at tage imod en invitation til noget, der
ikke er der længere.

Det er ikke ny magt til nogen: enhver, der ejer et projekt, kunne i
forvejen invitere en fremmed ind på instansen pr. e-mail. Kun ejeren kan
invitere til en note, som kun ejeren kan dele den.

De inviterede står i panelet ved siden af dem, der er kommet, og kan
trækkes tilbage. Uden det er en invitation sendt i går usynlig, og så
sender man den igen — endnu et link til den samme indbakke, og kun det
ene af dem til at gøre ugyldigt.

Og delingen siger selv til: den, der får en note delt, får besked om, at
den er der. Klokken sorterer nu på id efter tid, fordi tiden er sekunder
— en note, der bliver delt og rettet i samme håndbevægelse, stod ellers
i den rækkefølge, basen nu engang fandt dem i.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-08 11:58:47 +02:00
kristianwindandClaude Opus 5 ab77024ec9 feat(beskeder): klokken siger til, når nogen rører noget, der er dit
En delt note blev rettet af en anden, og man opdagede det kun ved at åbne
den. En opgave blev givet til en, og den lå bare der. Det, andre gør ved
det, man deler, skal man kunne se uden at lede efter det.

Klokken i hovedet: ulæste beskeder om noter, der er blevet rettet, delt
med en, eller opgaver, man har fået. Sætningen skrives i browseren, ikke
på serveren, så et engelsk interface får en engelsk klokke.

Beskeder om den samme note bliver slået sammen, mens de er ulæste. En
note gemmes på en tastepause, så en række pr. gem ville lave én person,
der skriver i fem minutter, om til tyve ens linjer. Er beskeden læst, er
den næste rettelse en ny ting at få at vide — ikke en gentagelse.

note_id får sin egen søjle ved siden af project_id og task_id frem for et
generisk (kind, id)-par: de tre er ikke det samme slags nul, og
fremmednøglen — dén, der rydder beskeden op, når noten slettes — kan et
generisk par ikke have.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-08 11:40:43 +02:00
kristianwindandClaude Opus 5 916a0c0224 feat(noter): en delt note tager de noter med, den peger på
Den anden halvdel af den melding, der begyndte det her: linkene i en delt note
førte ingen steder hen, fordi det, de pegede på, ikke var delt. At dele en note
er at dele det, den *siger*, og et link er en del af det.

note_shares får en `linked`-søjle. 0 er en deling, nogen har valgt; 1 er en, der
er udledt af et link. Skellet er hele modellen. Fladen kan sige *hvorfor* en
person har en note — "fulgte med" er en anden sætning end "du delte den" — og de
udledte rækker kan regnes forfra efter hver ændring, uden at røre dem, ejeren
selv har sat.

SyncLinkedShares regner billedet om fra bunden efter enhver af ejerens egne
handlinger: en deling, en fjernet deling, en gemning der har ændret links eller
titel, en sletning. Hele kæden følger med, ikke kun første led, med et loft på
hundrede noter — "del denne ene note" må ikke kunne blive til fire hundrede.
Rollen arves fra den note, man kom fra, og stærkest vinder, når en note kan nås
ad to veje.

Kun ejerens egne noter følger med, og det er ikke en optimering: et link peger
på en titel, og en medredaktør må skrive titler. Kunne et link hente hvad som
helst, kunne den, man har delt med, skrive [[Bankoplysninger]] og læse med.
Derfor slås der kun op blandt ejerens egne noter, og derfor regner kun ejerens
egne gemninger om. Begge dele har en Go-prøve.

Det lukker ikke hullet helt: skriver medredaktøren linket, og gemmer ejeren
bagefter sin egen note, følger den note med dér. Det, der står imod, er
synlighed — linket står som tekst i ejerens egen note, og panelet viser hver
note, der er fulgt med, og hvem der har den.

Udregningen åbner hver af ejerens noter, fordi titlerne er forseglet. Derfor to
værn: har man ikke delt noget, svares der med det samme ud fra ét EXISTS, og en
gemning regner kun om, når links eller titel faktisk har flyttet sig. Uden det
første kostede "Ny note" halvfems millisekunder, og i det hul nåede e2e-prøven
for monospace at taste sin titel ind i den forrige note, som editoren så tegnede
væk.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013r4DxpUZcSShxfCZZaeMGZ
2026-09-07 22:28:58 +02:00
kristianwindandClaude Opus 5 08d703e3af fix(noter): følg note-links i en delt note, og lad </> være kode
To ting meldt af en, der havde en note delt med sig.

Note-linkene i den delte note gjorde ingenting. Opslaget faldt tilbage på det
første søgeresultat, når ingen titel passede — og i en delt note er det første
resultat næsten altid noten selv, fordi den nævner den titel, den peger på. Så
åbnede linket den note, man allerede stod i. Kun en titel, der passer helt,
tæller nu; og rammer den forbi, siger beskeden den rigtige grund: en note, der
deles, deler ikke de noter, den peger på, så modtageren kan se linket uden at
kunne åbne det.

Og `</>` sad på kildevisningen med teksten "Vis som kode" på. Tegnet betyder
inline kode i ethvert andet program, så et markeret ord plus et tryk gav hele
noten som Markdown. Nu gør `</>` det, det siger — ⌘E, en kontakt, og ordet under
markøren når der ikke er markeret noget — og kilden har fået Markdown-mærket og
sin egen tekst.

Kode indsættes som en node frem for med execCommand: Chromium renser den HTML
gennem indsætningsvejen og gav et <span> med skrifttypen skrevet ind i sig, som
htmlToMarkdown ikke kan skrive ned som kode.

To e2e-prøver, som begge fejler på koden før den her.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013r4DxpUZcSShxfCZZaeMGZ
2026-09-07 21:50:12 +02:00
kristianwindandClaude Opus 5 094ea5cdd8 feat(upcoming): look up to 30 days ahead, and choose how far
Kommendes liste has always been the next seven days and nothing else. Seven days
is the right walk to take in the morning, and it is still what the view opens on
— but "when does this month get busy" is a different question, and the month grid
answers it with shape rather than in order. So the list now takes a horizon: 7,
14 or 30 days, chosen with a switcher beside the view buttons.

Only the list gets one. A week grid is a week and a month grid is a month; there
is no fortnight for them to show, so the switcher is hidden in those modes.

The choice lives in localStorage next to the mode, under its own key. It is the
same kind of choice as list-or-grid — how you want to look, not a fact about the
work — and it should survive a reload without following you to a phone where
thirty day headings is a very long scroll.

Once the strip is longer than a week the day headings carry the week number.
Thirty days repeat every weekday four times, and "mandag 6. okt" alone does not
say which mandag.

Server side the endpoint already took `?days=` and capped it at 31, so the window
needed nothing — but its row limit was a constant 1000 written for a week. A
month asked for with a week's ceiling would truncate the last days silently, and
a day cut off the end of the answer looks exactly like a day with nothing due, so
the limit now follows the window.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SA1TDmgQZw3AufJqkbwezx
2026-09-07 14:46:49 +02:00
Kristian Wind 4d5f1a56b1 Merge pull request #56 from kristianwind/gmail-unstar
CI / OpenAPI is valid (push) Successful in 2m55s
CI / End-to-end (push) Successful in 8m3s
CI / Docker image builds (push) Successful in 1m46s
Release / Build and publish image (push) Failing after 7s
CI / Go (push) Failing after 13m23s
feat(gmail): take the star off a mail once it has become a task
2026-09-04 09:06:31 +02:00
kristianwindandClaude Opus 4.8 84ca506961 feat(gmail): take the star off a mail once it has become a task
A starred Gmail message became a task and kept its star, so a starred inbox only
grew — the IMAP side has always unflagged what it imports, and Gmail could not,
because the scope was gmail.readonly. This gives Gmail the same one write: after
the task is made, the STARRED label comes off the message, in one batchModify for
the whole sweep. Nothing else is touched — not the body, not other labels, and a
label trigger's label stays where it is, since that label is the person's own.

The scope goes from gmail.readonly to gmail.modify, which is the narrowest scope
Google offers that can remove a label. An existing connection keeps working on its
old read-only grant — the sync still reads, the unstar is a best-effort 403 that is
logged, not fatal — and starts clearing stars once the account is reconnected and
the new grant is in place. It self-heals: a mail that could not be unstarred stays
starred, stays in the next query, and is tried again.

The consent screen now asks for more than it used to; the settings copy, the docs
and the scope table all say modify rather than read-only in the same change.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-09-04 08:57:07 +02:00
kristianwindandClaude Opus 4.8 57689ddfbb feat(offline): hold through a short network drop with a caching worker
The service worker did Web Push and nothing else, with a note that a cache
serving yesterday's data was worse than an honest "cannot connect". This adds a
cache that does not do that: network-first for everything that can go stale, so a
reader online always gets the live answer and the cache is reached only when the
network cannot answer. The task and note lists stay live; they just no longer go
blank the moment the signal does — a reload through a tunnel, a lift or a
base-station handover comes back to the app instead of the browser's offline page.

Writes are untouched and still go straight to the network; api.js already retries
them across the same short drop. The shell cache is versioned so a deploy lands in
a fresh cache and the old one is dropped on activate; the data cache is cleared on
sign-out so the next person on the browser does not inherit it.

The old hand-registered /sw.js becomes SvelteKit's src/service-worker.js, which it
registers itself, and the push handlers move across unchanged.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-09-03 22:46:35 +02:00
kristianwindandClaude Opus 4.8 55b8e438ad fix(rows): open on the first tap on iPhone, and round the hover highlight
On a touch screen a :hover on the tap target makes iOS spend the first tap
applying the hover and only open on the second — the double-tap Kristian hit
on tasks (and notes). The row hover is gated behind @media (hover: hover)
now, so a touch has no hover to apply and the first tap opens. The note
row's action buttons, which the hover used to reveal, are shown outright on
touch instead. And the task-row hover highlight is a rounded box (border
radius on the row), the same corner as a note in the list.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-09-03 22:15:23 +02:00
kristianwindandClaude Opus 4.8 8a5dfef41a fix(theme): lighter day separators in Charcoal
In Charcoal the note-list day separators used --ground, the darkest colour
in the theme, so they read as dark bars on the lighter list. They take the
menu's own resting grey (#2d3133) now — the colour "Projekter" sits on —
via a --note-group-bg fallback token, so every other theme keeps --ground.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-09-03 20:46:51 +02:00
kristianwindandClaude Opus 4.8 c2d4ce0179 fix(notes): open a note on the first tap, not the second
On a phone the list/editor switch keyed on the loaded note, which arrives a
moment after the tap — so the first tap looked like nothing happened and
people tapped again. It keys on the selected id now, set the instant a row
is tapped, so the pane switches at once; the back button is there
immediately and the body fills in when it arrives, with a quiet pane rather
than the "pick a note" prompt in the half-second between.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-09-03 20:27:28 +02:00
kristianwindandClaude Opus 4.8 8a2c46f5ea fix(notes): show the note on a phone — one pane at a time
On a phone the note list and the editor stacked in one column, so a tapped
note opened below a full-height list, out of sight — "kan vælge en note men
kan ikke se indholdet". Now it is master-detail: the list fills the screen
until a note is opened, the editor takes over once one is, and an iOS-style
"‹ Noter" at the top of the editor goes back. The cards drop their border
and corners on one column, where there is nothing beside them.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-09-03 19:07:02 +02:00
kristianwindandClaude Opus 4.8 905ceebc06 fix(theme): match Charcoal to Apple Notes' sampled colours
Kristian sent the exact values; the first Charcoal was a neutral grey guess
and missed them. Now: three panel greys, lightest sidebar (#2d3133) to
darkest editor (#1e1e1e) with the list (#242b2e) between — the reverse of
the usual sunken-is-darkest order, so the editor takes its own --editor-bg
rather than sharing the list's surface. The accent is amber: the selected
menu item is solid #ffce35 with dark text, a selected note is gold #a78c26
with dark text, and the section headings drop to a quiet grey (#8a8e90).

Each per-element colour is a fallback token only Charcoal sets, so every
other theme reads through to the shared token unchanged. Verified the
computed values land on Kristian's samples exactly.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-09-03 17:42:55 +02:00
Kristian Wind 87c557941a Merge pull request #49 from kristianwind/theme-charcoal-polish
CI / OpenAPI is valid (push) Successful in 9s
CI / End-to-end (push) Successful in 8m47s
CI / Docker image builds (push) Successful in 1m47s
Release / Build and publish image (push) Failing after 5s
CI / Go (push) Failing after 14m25s
Charcoal theme, elegant selection, rounded note list, confirm sign-out
2026-09-03 16:37:44 +02:00
kristianwindandClaude Opus 4.8 b81f8f67cc feat(ui): Charcoal theme, elegant selection, rounded note list, confirm sign-out
- A new Charcoal theme — neutral dark greys on the menu, note list and
  editor instead of the default's faint green cast, with matching inks;
  the green accent stays. It joins the theme picker on its own (the picker
  is data-driven from THEMES).
- The text selection is a translucent wash of the accent, not the old
  near-invisible tint and not a solid slab — visible on white and on the
  dark surfaces, keeping the text's own colour.
- The note list is a rounded card now, the same corner and border as the
  editor beside it, with a few pixels more air around the menu and the list.
- Signing out asks first: a second click within a few seconds confirms, so
  a button the pointer passes over does not drop you at the login screen.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-09-03 16:27:55 +02:00
kristianwindandClaude Opus 4.8 f62ffc8d92 feat(2fa): show a scannable QR code when enabling two-factor auth
The setup screen showed the secret and the otpauth URI as text and asked
you to scan — with nothing to scan. It renders the URI as a QR now, an SVG
(one dark path on a white card, so it reads in either theme) generated from
the barcode encoder already in the module graph via pquerna/otp. The secret
stays below it for a device that cannot scan.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-09-03 16:27:11 +02:00
kristianwindandClaude Opus 4.8 b2fcc0b104 feat(offline): ride out short connection drops instead of losing the save
A save made while the connection blips for a few seconds is retried rather
than reverted: the request layer keeps trying a network failure for a
fifteen-second window, waking the instant the OS reports a link again, and
only falls back to the existing error path if the drop outlasts it. A quiet
"reconnecting" pill appears if the wait passes about a second, and leaves
the moment it heals — so a two-second blip is felt as nothing at all.

This is deliberately not offline mode: the app is a live view of a shared
list, so there is no cache to work in with no server and no write log to
reconcile. It is only the blip, ridden out. A server error (4xx/5xx) is the
server answering and is not retried; an aborted request (a superseded
search) is left to propagate.

Kristian: an offline function for when the net drops for short periods,
seconds. Shared items stay server-authoritative — nothing is edited against
a stale copy for longer than the blip.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-09-03 15:26:46 +02:00
kristianwindandClaude Opus 4.8 05b7dac713 feat(notes): three-line rows like Apple Notes, with a filed line and dividers
Each note is three lines now: the title, then the date and the excerpt,
then where the note is filed — a folder like the menu's groups and the
project's name, the owner's when it is someone else's, or simply "Noter"
when it belongs to no project. The date leads the second line, left
aligned, instead of being pushed to the right of the title. A faint
hairline separates one note from the next, and a month heading is
capitalised ("August 2026", not "august 2026").

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-09-03 14:52:56 +02:00
Kristian Wind cfd382e3df Merge pull request #44 from kristianwind/settings-link-style
CI / OpenAPI is valid (push) Successful in 11s
CI / End-to-end (push) Successful in 10m3s
CI / Docker image builds (push) Successful in 2m8s
Release / Build and publish image (push) Failing after 6s
CI / Go (push) Failing after 14m58s
Style the unstyled "open completed tasks" link in settings
2026-09-03 14:35:20 +02:00
kristianwindandClaude Opus 4.8 01001d0980 fix(settings): the "open completed tasks" link was unstyled
It carried class="ghost", which nothing in the data page styles — so it
fell back to the browser's blue underline while every link beside it is
the accent-coloured .link. Now it matches them.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-09-03 14:27:20 +02:00
kristianwindandClaude Opus 4.8 4ac7180cda feat(notes): drop the disclosure arrow from the note-list headers too
Same as the sidebar: the box folds on a click, the star and people icon
are the marks that mean something, and a date group keeps an empty spacer
so every label still starts at the same place.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-09-03 14:25:55 +02:00
kristianwindandClaude Opus 4.8 e1683ef09d feat(notes): match the note-list headers to the new heading design
The note list still wore the old faint-uppercase group header beside the
sidebar's bold plain-case one, so the two columns read as two apps. The
group headers are bold, plainly cased and in the reading ink now, at
menu size, and their separator boxes carry the same rounded corner as an
active menu item — the hard rule underneath is gone, the box is the
separation. The disclosure mark stays: unlike a sidebar section, a
time-group has no icon, so it is the only thing that says "this folds".

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-09-03 14:22:19 +02:00
kristianwindandClaude Opus 4.8 ec147cd1ba feat(sidebar): drop the fold arrows from headers and groups
The chevron beside every section and group was saying "clickable" a
second time next to an icon and a bold name that already said it. Gone
now: a system section folds from anywhere on its heading, and a group
folds from its folder icon while the name still opens the group's page.
No arrow on the rail.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-09-03 13:34:33 +02:00
kristianwindandClaude Opus 4.8 6cf63113df feat(calendar): lighter grid, per-calendar pastels, today marked at top
The calendar was a grey grid where every event was the same sunken block
edged in a colour you had to already know. Now the day itself is the
brightest surface and each event is a soft, dusty wash of its own
calendar's colour — so a screen of them reads as colour laid on white,
and two calendars are told apart at a glance. Tasks stay outlined cards,
the thing you can pick up, against the solid washes you cannot.

Today is marked where the day is named: its number sits in a filled disc
at the top of the column, the same one the month grid gives it, over a
faint wash down the whole column.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-09-03 13:09:07 +02:00
kristianwindandClaude Opus 4.8 1c223a2c26 fix(sidebar): scrollbar flush to the divider, and match the group test
The scrollbar thumb centred a 4px bar in a 10px track behind a
transparent border, so it floated three pixels shy of the right edge.
The thumb is the track now — a thin 6px rounded bar sitting flush
against the divider line, where a scrollbar belongs.

The group heading is a heading now, not a project row: the smoke test
measured a group against the projects filed under it and read the
section icon as a dot. It measures a group against "Projekter" — the
tier it shares — and reads the icon both headings carry.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-09-03 12:54:05 +02:00
kristianwindandClaude Opus 4.8 a9c90ee6f1 feat(ui): unify section and group headers with icons
System sections (Projects, Filters, Labels) and user groups (Work,
Home) now share one design: a bold label at menu size, a leading
folder/funnel/tag/people icon, and an always-visible fold chevron.
Groups drop their colour dot for the same folder icon, so the sidebar
reads as one consistent hierarchy instead of two visual languages.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-09-03 12:34:07 +02:00
kristianwindandClaude Opus 4.8 378980de0c fix(sidebar): the scrollbar sits at the edge, thin and dim
It sat a pad-width inside the right edge because the scroller lives inside the
sidebar's padding, so it read as floating in the middle. Pull the scroller out
to the border and pad the content back — the rows do not move — and make the
bar thin and dim so it is present without drawing the eye.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-09-03 12:14:39 +02:00
kristianwindandClaude Opus 4.8 2f00246d6f feat(ui): colour the fixed-view icons, lift the notes surfaces, fix the create header
Following the Things pass, four bits of polish:

- The fixed views carry their own colour — amber star for Today, blue tray
  for the Inbox, red Upcoming, violet Waiting on others, teal Notes, indigo
  Calendar — and keep it whether or not they are the active row.
- The active row was all but invisible on the dark sidebar (a half-step of
  surface); it is now an accent-tinted selection that reads in both themes.
- The Projekter heading — the only place to add a project or a group — no
  longer scrolls away with a long list; it is pinned to the top of the
  scroller.
- The note list and the editor lift off the page: white in light, a nuance
  lighter than the ground in dark, so the two read as one writing surface.

Screenshots regenerated.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-09-03 11:47:17 +02:00
kristianwindandClaude Opus 4.8 cf69cac410 feat(sidebar): icons for the fixed views, bold areas, one straight edge
Toward Things' calm: the fixed views carry icons instead of dots — a star
for Today, a calendar for Upcoming, a person for Waiting on others, a tray
for the Inbox. Areas (the user's groups) are set in bold, plainly cased, so
the eye lands on them the way it lands on an area in Things; the system
sections (Projekter, Filtre, Etiketter) stay quiet. And projects sit at the
same left margin whether or not they are in an area — one straight edge down
the sidebar rather than a staircase, with the bold heading doing the
grouping the indent used to.

Screenshots regenerated to match.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-09-03 10:02:53 +02:00
kristianwindandClaude Opus 4.8 8dfab618f7 feat(settings): menu and body-text font size, and a bolder menu by default
The sidebar read as muted 13px; the menu items now use their own size at
medium weight so they carry at a glance (Todoist-ish). Both sizes are
choices under Settings → Appearance:

- Menu font size drives a dedicated --menu-size the sidebar uses.
- Body text size re-scales the shared type tokens (the list and the notes),
  the same mechanism a look uses.

The two are independent on purpose — one is a glance down a list of names,
the other a paragraph you read. Kept in localStorage like theme and look,
and applied in app.html before first paint so a chosen size does not flash
in and resize. Four steps each (Small/Default/Large/Largest); Default
carries no attribute so the base tokens stand.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-09-02 15:30:30 +02:00
kristianwindandClaude Opus 4.8 4103e1b0ca fix(tasks): open the task by clicking its text, links carved out
The first cut put an open-button overlay over the row, which intercepted
every click on the title text — nine smoke tests (and real users) open a
task by clicking its text, and the overlay stole those clicks. Make the
body a role=button instead, with each link stopping its own click from
bubbling: plain text opens the task, a link opens the page.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-08-27 16:33:30 +02:00
kristianwindandClaude Opus 4.8 9d4a13c2c2 feat(tasks): make a URL in a task's title or description clickable
A task is plain text — the title is an input, the description a textarea —
so a pasted URL just sat there as words. It is now caught at display time
and shown as a link in the list and on the board, without the stored text
ever being anything but plain text.

The link cannot live inside the row's open-button (an <a> in a <button> is
invalid), so the body is a container with the open action as an overlay
behind it: plain text and the meta row still open the task, links are
raised above and take their own click. Conservative about what counts —
an explicit http(s):// or a bare www. host, not every dotted word.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-08-27 16:17:56 +02:00
kristianwindandClaude Opus 4.8 272c4918b0 feat(settings): a quiet Buy me a coffee link
A donate link belongs in Settings and nowhere in the way — so it sits last,
understated, below the device list. A plain anchor with an inline SVG cup:
no third-party script, no webfont, no tracking, so it works offline and
cannot phone home. Wording says the app is MIT with nothing held back and
that nothing changes if you don't pay.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-08-27 14:49:24 +02:00
Kristian Wind 2b76a172e8 Merge pull request #30 from kristianwind/add-delegated-marker
Release / Build and publish image (push) Failing after 10s
feat(tasks): mark a delegated task with a hand-off arrow
2026-08-27 12:43:13 +02:00
kristianwindandClaude Opus 4.8 7dcd475036 feat(tasks): mark a delegated task with a hand-off arrow
A task assigned to somebody who is not you is out of your hands and into
theirs, but the row only showed their name — the same weight as any
other meta. The assignee is now a bordered chip led by an outgoing arrow
in the accent colour, so a list reads at a glance as which tasks are
yours to do and which you are waiting on. Covers the board too: its
cards render a TaskRow.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-08-27 12:27:52 +02:00
kristianwindandClaude Opus 4.8 fc4877c3bd fix(mobile): stop the task drawer's priority/due row overflowing
On a phone the detail drawer's field grid stayed two columns, so each
field got half the viewport. Prioritet collapsed to "Ing" and the due
row — a date input plus fixed-width time and duration fields — ran off
the right edge. Below 560px the grid now stacks to one column per field
and the due row wraps.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-08-27 12:24:29 +02:00
kristianwindandClaude Opus 5 9e3f20b050 feat(tasks): snooze — park a task at the bottom until a time
Snooze puts a task out of the way without finishing it or filing it: it sinks to
the bottom of its list and greys, and comes back on its own when its time passes.
No job wakes it — "snoozed" is just snoozed_until being still in the future, read
against the clock each time the list is ordered.

- `snoozed_until` on tasks (migration 0031); the list order sinks a still-snoozed
  task below everything, and it rejoins its place the moment the time passes.
- POST /tasks/{id}/snooze {until} — an RFC3339 time to snooze to, or empty to wake.
  Nothing else about the task moves; a snooze is about when you want to see it, not
  when it is due.
- The task drawer gains quick picks (this evening / tomorrow / next week) and a
  custom time, and a Wake now once it is snoozed. The row dims and carries a
  "snoozed until X" mark, and the list keeps snoozed tasks at the foot — the client
  sorts the same way the server does, so a drag-sorted list does not jump on reload.

Tests: snooze sinks a task below one made after it and waking returns it (Go); and
from the drawer the row greys, shows the mark, and un-greys on wake (e2e). Full
suites green.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-25 10:59:34 +02:00
kristianwindandClaude Opus 5 a70ca93102 feat(notes): an Archive button in the open note's footer
Archiving a single note meant finding its row in the list and its hover-only
button. With a note open to read or write, the one place you would reach for it —
its own footer, beside Gem and Slet — had no way to put it away.

There is one now: Arkivér between Save and Delete, which reads "bring it back" once
the note is already archived. Owner only, like delete, and it runs the same archive
the row button does — the note leaves the list without leaving the account.

e2e: archiving from the footer takes the note off the list, the archive holds it
with a bring-back button, and bringing it back returns it. The existing bulk-archive
test is scoped to the selection bar, now that "Arkivér" also lives in the footer.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-24 15:37:17 +02:00
kristianwindandClaude Opus 5 8022502015 feat(notes): [[note]] is an inline link, not bracketed text
Choosing a note from the [[ completion wrote [[Title]] as plain text. It made a
real link behind the scenes — it showed in the panel under the note — but in the
note itself it stayed brackets, and "it doesn't become a link" is exactly how that
reads. A wiki link is a link where it is written.

Now a [[note]] renders inline as the title, in the accent colour, and a click on it
opens the note it points at. It is atomic — contenteditable="false" — so it moves
and deletes as one thing and a click is a click, not a caret landing mid-word. And
it appears the moment it is chosen: the completion drops the finished link in,
brackets and all swallowed, rather than text waiting to become a link when the note
is next opened.

The seam holds both ways: markdownToHtml lifts [[title]] into the link the same way
it already lifts an image, and htmlToMarkdown reads the link back to [[title]] by
the title it carries — so what is on disk is still the same Markdown, and a note
opened, saved and reopened is unchanged.

e2e: choosing a note writes an inline link, it survives the round-trip through
Markdown and a reload, and clicking it opens the note it names. Full suite green.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-24 13:41:04 +02:00
kristianwindandClaude Opus 5 de95df44c1 fix(notes): a [[note]] link is one you can follow
Picking a note from the [[ completion wrote [[Title]] and the server made a real
link from it — but the link only ever showed in the panel under the note as plain
text. It said the note pointed somewhere and gave no way to go there, which reads
as "it didn't become a link at all".

The links in that panel are buttons now. A [[note]] resolves by its title to the
note — the one already on the list, or looked up if not — and opens it in place; a
#project goes to its page; a task link to the task. Titles with punctuation work,
brackets and parentheses included, because the match is on the whole title, not a
parse of it.

e2e: after linking a note, clicking its chip opens the note it names.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-24 13:11:46 +02:00
kristianwindandClaude Opus 5 6116ed2e21 fix(notes): [[ completes a note link, the way # completes a project
Typing [[ did nothing. The toolbar hint offered "[[ another note" and the parser
turned a fully-typed [[Title]] into a link on save, but the completion that makes
it usable — the one # has for projects — was never there, so you had to know and
type a note's title exactly, brackets and all.

Now [[ opens the same picker # does, over the notes the page has loaded: the
current note is left out (a note linking to itself is a loop nobody meant), and
one without a title yet cannot be pointed at. Choosing a note closes its own
brackets — the opener is already in the text — and leaves a space, so the next
word is a word and not more of the tag. On save the link is real and shows in the
note's links, not merely as text.

The picker's two kinds are told apart where a project wears its colour: a note
suggestion wears the brackets that open it.

e2e: [[ offers a note by title, choosing it writes [[Title]] into the note, and
the link lands in the note's links. Full Chromium suite green.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-24 12:22:38 +02:00
kristianwindandClaude Opus 5 4f25905d81 feat(calendar): a task's length is drawn, resizable, and editable
A timed task was drawn at a fixed half hour whatever its `duration_min` said, so a
length the data model already carried was invisible and uneditable. Three parts:

- The calendar draws a task at its real length — a quarter-hour floor so a
  zero-minute task is still grabbable, the day's end as the ceiling — instead of a
  flat thirty minutes. A ninety-minute task now looks like ninety minutes.

- A grip along the foot of a timed task drags its length. Pointer events, not the
  HTML5 drag the body uses, and stopPropagation between them — otherwise making a
  task longer would pick it up and move it instead. It snaps to the quarter hour,
  follows the handle live, and writes once on release through a new store `resize`.
  Moving a task (the existing day/time drag) leaves the length alone.

- The task drawer gains a length field beside the time — minutes, shown only with
  a time to hang it on, since a duration without a start is a length of nothing and
  the calendar has nowhere to draw it. So a length can be set off the calendar too.

e2e: dragging the foot of a task makes it longer both on screen and in what the
server stores, and the time does not move. Full Chromium suite green.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-24 10:14:39 +02:00
kristianwindandClaude Opus 5 97e6be3afa fix(notes): gather sharing behind a Del button so the footer stays tidy
The sharing controls were laid out across the note's footer — the project select,
the whole "Del med personer" panel and the Save/Delete buttons all on one row. It
crammed: "Læg i projekt" wrapped to two lines and the people panel ran into the
rest.

Now one Del button in the footer opens a popover that holds both ways a note is
shared — a project and named people — stacked in a small card. The footer is a
status on the left and Del / Gem / Slet on the right, and nothing wraps. The
button carries a count when the note is shared with someone, so "shared" is
legible without opening it.

The popover closes on a click outside, on Escape, and on switching to another
note — the last tracked by note id, so a save returning a fresh copy of the same
note does not slam it shut mid-use. The inner "add someone" button is now Tilføj,
so it does not read as a second Del.

e2e updated to open the popover before reaching the controls; full Chromium suite
green.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-24 09:58:02 +02:00
kristianwindandClaude Opus 5 f0d18bb800 fix(notes): let a monospace line be left — on return and through the menu
A <pre> is preformatted, so the browser answered Enter by inserting a <br> and
keeping the caret inside it. A monospace last line therefore trapped everything
typed after it in monospace, with no way back to body by typing — and the style
menu could not get out either: `formatBlock('p')` on a line inside a <pre> nests a
<p> *inside* the block (`<pre>…<p>body</p></pre>`), so the text stayed monospace
and the note was left malformed. Reported as "monospace keeps going and I can
neither delete it nor switch to body".

Two fixes, both in the editor:

- Plain Enter in a <pre> now steps out: the tail after the caret becomes a
  paragraph below the block, and that is where the caret lands. Shift+Enter is
  untouched, so a soft line break still writes multi-line code inside the block —
  Enter to leave, Shift+Enter to stay.

- The style menu converts anything touching a <pre> by hand, replacing the element
  rather than reformatting inside it, so a monospace line goes cleanly back to body
  (or to a heading) instead of nesting.

The split can leave an empty text node in the new paragraph, and a block with no
element and no text is not a caret host in Chromium — the next keystroke falls back
into the block above. So an empty paragraph is given a <br> to stand in, which was
the difference between the caret landing in the new line and the text reappearing
in the code.

e2e: a monospace line left on Enter is body after it, Shift+Enter stays monospace,
the menu returns it to body, and the whole thing survives the round-trip through
Markdown and a reload.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-24 09:27:44 +02:00
kristianwindandClaude Opus 5 91671a4f97 feat(notes): share a note with a person, not only through a project
Until now a note was shared exactly one way: file it in a shared project, and
everyone who could read the project could read the note. That is right when the
note belongs with a body of work and wrong when it is just "let Sofie see this" —
it made you stand up a project to hand over a single page.

A note can now be shared with named people directly, at viewer or editor — the
same two roles a project membership uses, so there is one idea of a role in the
program and not two. The paths stack: a note can be loose, filed, shared with
people, or any mix, and access is the most any of them allow.

The store:
  - note_shares (migration 0030): note ↔ user ↔ role, one row per person, owner
    recorded. Visibility (AllNotes, and search via mayTouchNote) widens to include
    notes shared directly with the reader.
  - the effective role folds a direct share on top of the owner and project paths.

The rules, drawn where they protect the owner:
  - read = viewer, edit text = editor, on any path.
  - MOVE a note (change its project) is owner-only: an editor filing it into a
    project the owner is not in would take the owner's own note away from them.
  - DELETE is owner or project-editor — a direct share, however high its role,
    does not reach to throwing the note away.
  - SHARE and UNSHARE are the owner's alone, like a project's invitations.
  - the picker is everyone on the instance (a closed, invited set) minus who is
    already on the note; a made-up id and the owner themselves are refused.

The list:
  - a "Delt med mig" group at the top, collapsible, that appears only when it has
    something in it — the same rule Favoritter follows — so it costs nothing when
    you share nothing. A shared note lives there, not also in a time bucket.
  - every shared note carries a small chip in its owner's avatar colour, the
    project mark's shape said for a person.
  - no star or archive button on a note somebody else owns: those change the note
    itself, and it is not yours to put away.
  - the editor grows a "Del med personer" panel for the owner: who can see it, at
    what role, and an add row drawn from the instance's other people.

The old comment in the notes footer argued against a note having an access list
of its own. That was the right call until the case it did not cover — sharing
without a project — turned out to be the one actually asked for. Updated to match.

Person now carries json tags, so a note's owner serialises as id/name/avatar_color
like every other person in the API rather than as Go field names.

Tests: a shared note reaches the other person marked and owned; a viewer reads but
cannot write; an editor writes but cannot re-share or delete; unshare removes
access and is owner-only; the candidate list excludes who is already on and the
owner; and a made-up id and self are refused. OpenAPI covers the three new routes.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-24 07:36:55 +02:00
Kristian Wind b6cbc6cbc2 fix(ai): din egen model kunne ikke sættes op — tre fejl i én oplevelse
CI / OpenAPI is valid (push) Successful in 12s
CI / End-to-end (push) Failing after 9m31s
CI / Docker image builds (push) Successful in 1m56s
CI / Go (push) Failing after 14m40s
Rapporteret som "AI gemmer ikke mine oplysninger". Anden gang det symptom
meldes, og en helt anden årsag end sidst — dengang delte scopes i user_settings
én række og slettede hinanden.

Udbyderen hedder "OpenAI-kompatibel (Ollama, LM Studio, …)" og feltet foreslår
selv http://localhost:11434/v1. safedial nægter præcis den slags adresser. Så
den ene udbyder, der findes for at køre sin egen model, kunne ikke sættes op,
og formularen foreslog en adresse, serveren ville afvise.

Serveren svarede 422 med en fejl på base_url. Formularen havde ingen plads til
at vise en fejl på det felt, og fordi fejlobjektet ikke var tomt, blev der
heller ikke vist en toast. Beskeden forsvandt, og det så ud, som om der ikke
skete noget ved at trykke Gem.

Tre rettelser:

- Private adresser tillades for admins. Muren står, fordi en server, der henter
  en brugerangivet adresse, lader den, der angiver den, nå ting han ellers ikke
  kan. En admin på en selvhostet instans ejer allerede værten og dens netværk,
  så det giver ham intet nyt. For en inviteret bruger er den lukket som før.
  Tilladelsen regnes ud pr. forespørgsel, ikke gemt ved siden af adressen — den,
  der ikke længere er admin, skal ikke kunne nå indenfor på et gammelt flag.

- Tailnet er ikke "indenfor". Halvdelen virkede allerede ved et tilfælde: Gos
  IsPrivate dækker RFC1918 og ikke det CGNAT-område, Tailscale låner, så 100.x
  slap igennem uden at nogen havde besluttet det. IPv6-halvdelen gjorde ikke —
  fd7a:115c:a1e0::/48 ligger inde i fc00::/7. Samme maskine var altså
  tilgængelig eller ej alt efter adressefamilie. Begge områder er nu navngivet,
  hvilket gør IPv4-halvdelen bevidst frem for heldig.

- Feltfejl kan ikke længere forsvinde. base_url og model har fået et sted at
  vise en, og en fejl på et felt, der ikke har et, råbes op som en toast frem
  for at blive tabt.
2026-08-22 00:03:45 +02:00
Kristian WindandClaude Opus 5 8d5dd1f1c3 fix(notes): grupperne stod tomme — hver liste rullede for sig
Release / Build and publish image (push) Failing after 9m59s
Meldt ind som "noter vises ikke i den nye listeform": hver eneste overskrift stod
der med sit tal, og der var ikke en eneste note under nogen af dem.

Før var der ét `<ul>`, og *det* rullede. Da listen blev delt i grupper, blev hver
gruppe sit eget `<ul>` — og dermed sit eget flex-element med `overflow-y: auto`.
Flex-elementer skrumper som udgangspunkt, og et element, der skrumper og klipper,
viser ingenting. Med tyve måneder i søjlen fik hver liste nul at være høj i.

Rulningen hører til kassen om dem alle, ikke til hver af dem.

**Prøven fangede det ikke, og det er værd at forstå hvorfor.** Den lavede to noter,
altså tre grupper, i et vindue på 720 pixels — og med så få grupper er der plads
nok til, at skrumpningen ikke bider. Fejlen kræver pres. Prøven gør vinduet lavt
nu og kigger efter, at rækkerne kan *ses* og ikke bare findes, og den siger
desuden hvor rulningen ligger: på kassen, ikke på listerne. Den er set fejle med
den gamle stil.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-21 13:00:15 +02:00
Kristian WindandClaude Opus 5 2219f6af13 fix(i18n): en gentagelse siges nu på læserens sprog
Release / Build and publish image (push) Failing after 10m2s
`recurrence.Describe` skriver dansk. Fladen kunne navngive de fem almindelige
regler selv, og alt andet — "hver mandag", "hverdage", "hver anden uge tirsdag" —
faldt tilbage til serverens tekst, som er dansk uanset hvad kontoen er sat til.

Serveren kan ikke bare tage en locale med, og det er ikke dovenskab: `toTaskJSON`
har niogtyve kaldesteder, og mange af dem er WebSocket-udsendelser, hvor én
nyttelast når alle i et projekt på én gang. Der findes ikke ét sprog at skrive den
i.

Men reglen er den samme, uanset hvem der læser den. Den ligger allerede i
nyttelasten, så beskrivelsen hører hjemme dér, hvor læseren er — i fladen.

Serverens tekst bliver stående som reserve. Den er stadig den, et API-kald og en
MCP-forbindelse får, og de har ingen flade at oversætte i.

Ugedagen kommer fra `Intl` og ikke fra en tabel — ikke for at spare syv strenge pr.
sprog, men for at få de navne, sproget faktisk bruger. Dagene sorteres, som ugen
går, frem for som reglen tilfældigvis skrev dem: "mandag og torsdag" og "torsdag og
mandag" er den samme regel, og den ene af dem er den, folk siger.

Og to har sit eget ord på begge sprog — "hver anden", "every other". Det er ikke
tallet to sat ind i den samme sætning, så det er sin egen nøgle. Et sprog, der
bøjer enheden efter tallet, kan ikke sættes sammen af stumper; derfor står tallet
inde i teksten frem for at blive limet på.

Prøven bruger `hver mandag` og `hverdage`, fordi de netop *ikke* står på listen
over de fem, fladen altid har kunnet. En tredje vending, jeg først skrev prøven
med — "hver anden tirsdag" — viste sig ikke at være en, hurtig tilføjelse
overhovedet læser: den lagde "hver anden" i titlen og gjorde "tirsdag" til en dato.
Det er en anden sag, og den står nu skrevet ét sted, hvor den kan tages op.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-20 22:00:05 +02:00
Kristian WindandClaude Opus 5 868588de10 feat(projects): et projekts kalender kan vises som en uge
Ugen fandtes på Kommende og på Kalender, men ikke dér hvor et projekt bor — og det
er dér, "hvornår sker de her fire ting" oftest bliver spurgt.

Inde i kalendervisningen frem for som en fjerde `view_mode`. `view_mode` er
projektets og delt med alle, der kan se det, mens uge eller måned er et spørgsmål
om hvor bred skærmen er, foran hvilken der sidder én person — samme grund som
sidebjælkens bredde ligger i localStorage af. Og en fjerde værdi ville kræve en
tabelombygning: `CHECK (view_mode IN ('list','board','calendar'))` kan SQLite ikke
ændre på plads.

**Og så fandt den en fejl, der har ligget og ventet.** `cursor` blev sat med
`span === 'week' ? … : …`, hvilket læser `span` i det øjeblik komponenten bliver
til og fanger den værdi for altid. Byggeriet har advaret om det lige så længe. Så
længe hver visning havde sin egen komponent, gjorde det ingen forskel — og det
gjorde det i samme sekund, en vælger begyndte at skifte `span` på den samme.
Markøren stod på den 1. i måneden, og et skift til uge landede på ugen omkring den
1.: "27. jul.–2. aug." for en person, der kiggede på august.

Rettet, med den ene undtagelse skrevet ned: en månedsvisnings markør *er* den
første i måneden, så et skift til uge, mens man står i den indeværende måned,
skal lande på den uge, i dag ligger i — ikke på ugen omkring den 1.

Og en anden fælde, jeg selv gik i undervejs: en underklasses felter sættes *efter*
basens konstruktør er kørt. Nøglen til localStorage som et felt på en underklasse
betød, at valget blev gemt ét sted og læst et andet — en uge, man havde valgt, var
en måned igen efter en genindlæsning. Nøglen er et argument nu; et argument har
ingen rækkefølge at tage fejl af.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-08-20 21:45:38 +02:00