The entry is 798 kB (259 kB gzipped) instead of 927 kB (298 kB). Signup's
`export default` kept it unsplit, and login, recover and reset-password left a
bare `z.object(...)` statement in their route stubs (reset-password also read
its search through Zod), which held all of Zod and react-hook-form in the entry.
The schemas are factories now, reset-password validates its search by hand.
The shell routes keep their component in the entry on purpose
(`codeSplitGroupings: []`, no `export default`): splitting them out measured
about 60 chunks per signed-in screen and a slower flow list. The Zod
`@__PURE__` comment warning is filtered and the chunk limit is 850 kB, each with
a comment saying why.
At 4x CPU on slow 4G the signed-in screens load 5 to 9% sooner, view 6%; login
13% later (the form chunk is one more hop). On a LAN they are within noise.
Node source (Python) and the JSON dialog share one CodeMirror editor with
find and replace, autocomplete, folding, indent guides, comment toggle and a
JSON lint, themed from the CSS tokens so light and dark follow the .dark class
without a second palette. Programmatic value changes no longer fire onChange,
blur ignores focus moving into the find bar, readOnly keeps selection and copy.
The editor chunk is 470 kB (157 kB gzipped) against 2.7 MB (704 kB) for Monaco,
which also pulled in a 1.2 MB JSON mode, two workers and a font. At 4x CPU on
slow 4G a node panel is ready in 5.7 s instead of 10.8 s, the JSON dialog in
1.5 s instead of 6.7 s. Typing without a key delay keeps every character.
monaco-editor, @monaco-editor/react and the dompurify override are removed.
Specs: typed code reaches the node source, a JSON setting saves only text that
parses. Docs name Tab and Ctrl-M.
The channel sends through the Bot API with httpx only: HTML with escaped text,
cut to fit a message, silent for a warning, a short 429 waited out once, and
errors that say what to do about a blocked bot, a lost chat or a bad token and
never quote the request URL. The bot token is a stored secret: saving a
channel refuses a literal one, the card's "Store token" writes the secret and
a reference, and no endpoint returns it.
A chat id is never typed. Connect starts a ten minute code, and the chat that
sends `/start <code>` back through the bot is the one connected, so it is a
chat somebody opened a link for. The engine fetches updates only while a code
is waiting (no standing poll, no webhook, no inbound address) and ignores
everything else without a reply. Connect needs a saved form, and the form
takes the connected chat in without touching what was typed since.
The token is part of every Bot API URL, so the httpx logger gets the same
filter the request lines have, which now also masks an httpx.URL argument.
The alert manager kept its cooldowns and its ceiling per hour in memory, so a
node that stayed down was announced again by every restart. Each delivered
alert is now a row in the new alert_sent table (additive migration), loaded
for the last seven days when the manager starts and pruned as new ones come
in. A ledger that fails never stops an alert. The manager's clock is wall
time for the same reason.
A node_health "down" alert also waits 60 s (HEALTH_GRACE_S) in a pending task
per node; a later event that is not "down" cancels it, as does stopping the
manager, so a broker restarting with the stack alerts nobody.
The cooldown used to be stamped before the ceiling per hour was checked, so
an alert the ceiling held back still spent its cooldown. It is stamped only
once the alert is sent now.
The summary read degraded for 15 minutes after any single node_error, even
once the node had run well again and including one bad input from a visitor.
The controller now counts a node's failures since it last ran well and notes
when it recovered; both carry across rebuilds and an acknowledgement resets
the count. Two in a row inside the 15 minute window is "failing" (a problem),
one is listed under warnings, and nodes.failing keeps meaning "failing now".
Any signed-in account could save an ntfy channel with its own `server` and a
`{"$secret": ...}` token and press Test, which sent the stored secret to that
host. The config GET/PUT, the test route and the global search's alert entries
now answer to superusers, like /secrets. The webpush key and subscribe routes
stay open to every account, and the Alerts page shows a non-superuser only the
button for its own browser. The events table in the docs gains run_finished.
/search gains a `setting` category, `<node>.<setting>` under its flow with the
node type and the value in a new `detail` field, so Ctrl-K finds which node
uses a broker or a timeout and opens the node. `setting_text` is the one place
that decides what may be said: a reference to a stored secret shows its name
(to superusers only, as the secrets list does), a plain value in an `x-secret`
field or under a key like password, token or api_key is masked and never
indexed, nested credentials in headers or query are masked, long values are
clipped. The global search ranks what contains the query as typed ahead of
loose spellings before it caps a group, which long setting values would
otherwise have crowded out.
An input row now starts from its message and ends as the argument the code
takes it as ("as prompt"); an output keeps its key on top and the message
below. The name the code uses is muted text with a pencil while it follows the
message's last segment, and a field only when it differs. Device Follows ports
get the same order (message, then the port driven). A list port's item type
moves to the second row so it no longer squeezes the name, two ports deriving
one name get a warning, and the rows rebuilt use 44px touch targets on phones.
The stored node spec is unchanged.
The command line no longer imports a research module, so import-time GPU work
ends with a short-lived child instead of holding a CUDA context for the length
of a followed run. fluksio.sdk.declare starts the child, which writes each
flow's document and shims as JSON; Declared stands in for Flow in sync().
Errors keep their type and message, a child's traceback reaches the user.
save_artifact/load_artifact outside a node use the worker's own store code in
$FLUKSIO_ARTIFACT_DIR or <tmp>/fluksio-artifacts, so a node can be called by
hand. --follow prints status changes and each node's resource wait; the
placer's waiting snapshot now carries the run. fluksio cancel <run>...
- InputInspector hangs off the label's element and EdgeInspector off a point in flow coordinates (ViewportPortal), both re-measured every frame while open and hidden once the anchor leaves the view. A label that leaves the canvas closes its popover.
- Dock: below md a bar across the width, flush with the bottom edge (bg-card, top border, safe-area inset, no shadow); from md the floating pill as before. The flow dock drops the Run caption and the dashboard editor clears the bar with its own padding, so the editing state fits one row from 360px.
- Tests: the search spec checks the popover rides along with a pan; the mobile spec checks both docks are a full-width bar flush with the bottom.
- LiveEdge: a value inside a running pulse restarts the animation by hand, since the class is already on and would show nothing.
- Dock, FlowDock: the bar is centred on left-1/2, where a shrink-wrapped box gets only the half of the screen to the right and wrapped into a 195px pill. w-max with the viewport cap keeps one row whenever the buttons fit. The dashboard editor's dock had the same squeeze.
- FlowEditor: arriving on an input drawn on the canvas centres its label and opens the label popover once the view is fit, as clicking it does. An input the canvas does not draw still opens the flow settings.
- Tests: the phone dock may not be confined to half the screen; the search spec reads the popover first and reaches the settings through its pencil.
_serve_once treated a cancelled call future as finished, so an event loop
cancelling all tasks at once (shutdown) never sent the worker a cancel and the
call's thread kept asyncio.run waiting for SILENCE_S. The cameras audio test
now removes its voice flow and device, the unpaired-frame test waits for its
first frame before the unpair, and test_connector no longer inherits the
controller an app test bound to demand.
Per the Microcopy rule: size and duration errors, the flavor errors, run
parameter errors, the Resources field descriptions, 401/409/400 details,
secret errors, the websocket node description and its on_connect setting.
The generated client follows.
The phone header carries a single speech bubble: Control on Home, the
assistant elsewhere. Chips above the input offer the user's most sent
messages per tab, then starters; a tap or a double-click on a sent or
heard message fills the input without sending.
POST/GET /flows/{name}/nodes/{id}/train submit wren_train on the node's
targets, examples and tests and read its latest run (run.for_node,
cause "train"); "Use this model" writes `model` into the draft.
Every output is a target and every input besides prompt a reading. A
targets setting gives descriptions, values, ranges and per-target
guards, model takes a trained pack's artifact digest, min_confidence
defaults to the model's threshold. A command is applied whole or not at
all, and targets the model was not trained on read degraded. The home
path (area/kind ports, groups, brightness, temperature answers) is gone.
wren drops a call on an area the command does not name (other than the
user's or, with none named, the living room): "switch on the appliances"
turned on the bathroom light at 0.97. The assistant's `values` setting gives
a str target its only values, explains them in the tool and confirms in the
word set for each ("Canopy open.").
The clipboard spec selected text in the previous node's panel while it was
still sliding out after the paste, so the selection collapsed when it
unmounted and the canvas took the copy. The values spec read the panel's box
mid-slide-in, where the translate leaves the width a float off 400.
The portal's copy of the dashboard is the `appshell` target of
backend/Dockerfile (18 MB instead of the 252 MB nginx image); the nginx
frontend Dockerfile, its confs and csp.check.ts are gone, since the app and
the hub both hash the CSP from the index.html they serve.
The runtime image copies the 4 MB git binary alone: the store's init, add,
commit, gc and rev-parse are builtins, and a build-time round trip fails the
build if that ever changes. About 100 MB of perl and helpers less.
The bundled Redis's overcommit warning and FastMCP's lifespan warning no
longer greet every `docker logs`.
The backend image now builds the SPA and serves it on the API's own origin
(fluksio/spa.py), so an instance is a single container with no proxy, DNS,
port 80 or .env: `docker compose up -d` with the new compose.yml runs it on
localhost:8080. It starts on fluksio serve (persisted secret key, first
admin printed once) and brings its own Redis when REDIS_HOST is unset.
Slimmer: python:3.13-slim, no dev group, uv once, fastapi without the cloud
CLI. Optional features are per-feature extras (mqtt, influx, email, sentry,
ml, server for all) picked with the EXTRAS build argument; numpy moved to
`ml`. ffmpeg left the image for imageio-ffmpeg via the Modules screen.
The operator stack runs the same image with app. and api. routed to it; the
nginx frontend service, the LAN nginx proxy and DOMAIN in the backend are
gone, and the MCP issuer is FRONTEND_HOST.
Copy widget in the widget panel or mod+c keeps one widget in localStorage;
a paste button in the dock and mod+v insert it below the other widgets with
a fresh id. Placement is shared with adding a widget.
frontend/scripts/no-em-dash.ts walks src/ with the TypeScript parser and flags
an em-dash in JSX text, string literals and template text. A bare "—"
placeholder passes, separators go in no-em-dash.allow.json. The script is
identical to index's copy.
ai/client.py streams the completion and puts the chunks back together
(Assembler), covering OpenAI's piecewise tool-call arguments, Ollama's whole
calls, OpenRouter's comment lines, unnumbered calls and in-stream errors.
The turn emits a delta event per piece and a message per round as before,
and the chat grows the last answer as deltas arrive.
react-markdown, lazy-loaded with the first answer (37 kB gzip chunk, main
bundle unchanged). No raw HTML, images dropped, links open in a new tab.
The system prompt asks for lists, inline code and fenced code blocks.
faster-whisper and piper as plain Function nodes, the mic -> STT -> logic ->
TTS -> speaker wiring, the Modules manifest they need (uv pip compile, av<19)
and measured sizes, memory and CPU times. Modules docs now say uv pip sync
installs no dependencies.
run_coroutine_threadsafe refused it and it was never awaited, a RuntimeWarning
under pytest 9. The cancel still logs that it could not reach the worker.
WebKit takes *.localhost to loopback through GLib and ignores /etc/hosts, so
the container shares Traefik's network namespace. The webkit project only
exists with PLAYWRIGHT_WEBKIT=1, so test-frontend stays Chromium only.
The suite creates and deletes flows, dashboards and users, and ran against
whatever stack was up, which is the dev instance's data. make test-stack starts
the compose project fluksio-test (compose.test.yml over the local stack's
files) with its own containers, volumes and image tags, served by the local
Traefik on app.test.localhost and api.test.localhost, and leaves it running.
The hostname is fixed, so it no longer comes off the running dev stack, and a
LAN stack makes no difference. The suite's default origin is the test stack.
Changing dashboard kept the grid mounted, so every new tile rose at once.
The grid is drawn afresh per dashboard and its tiles come in row by row:
from below, top row first, for one further down the rail, and from above,
bottom row first, for one further up, as a round panel's dashboards slide.
A phone stacks a dashboard instead of drawing it on a canvas, and the canvas
was what carried the ground, so a stack had only the palette's colour. The
panel and the editor now draw the look's ground behind the stack, held still
while it scrolls. Glass's drift stops under reduced motion as it does on a
canvas.
A widget that fell off the panel, after the canvas was made smaller, was
drawn under the clip and could only be re-packed with Reflow. Editing, the
canvas now extends to the lowest widget and scrolls, with the panel's own
edge dashed and the area past it hatched, so the widget can be dragged
back. The scale stays the panel's, and view mode is unchanged.
The canvas states the look of the dashboard shown, so a neighbour beside it
and the one a turn leaves wore that look too. Each page now restates its own
look, theme and palette. The ring's paint moves from rules under
[data-look] to variables each look sets on its root, which stop at the
nearer root where a descendant rule reached into both.
A matrix of the declared floor and the newest release in place of 3.13
alone. The job now syncs --all-extras like `make test-backend`, since the
mail, MQTT and Influx tests fail on a fresh venv without the server extra.
Both packages list 3.14 among their classifiers.
A face has no card behind its disc, so the tile button's white wash all but
vanished into a light ground. The look's root states the press's own wash,
tinted with the primary and casting the pane's shadow.
pytest 8+ collects files and directories together in name order, so crud/
now runs before test_cloud.py and its leftover superusers made adopting the
owner ambiguous.
The hello carries the fluksio-worker version. An agent refused for an older
protocol stays listed as Needs upgrade until it attaches again. Docs gain an
Upgrading workers section.
A portal identity whose email a local user already has is still refused
on arrival. POST /cloud/members/{portal_sub}/link lets a superuser map it
onto that user, never a superuser's, after the portal confirms the
membership. GET /cloud/members marks such members linkable, and Admin →
Users offers "Link portal account" for them, saying that a linked user is
deleted with a removal at the portal or a Disconnect.
A local Disconnect deletes every portal-mapped user except the account the
instance was enrolled as and superusers, so no inert rows stay behind. The
confirm dialog says so, and Admin → Users refreshes.
The connector now serves /mcp, its OAuth documents and /hooks/ besides
/api/v1/, refuses any dot segment (httpx resolves them, which carried
/api/v1/../../mcp onto the MCP mount), addresses calls to the hub's host,
and bridges served sockets by calling the app with a websocket scope.
A hook without a secret answers the tunnel with the 404 (or 1008) of a
wrong secret. Requests through the hub get the hub's issuer, resource and
consent page in the OAuth metadata and the 401, and the MCP host check
admits the hub while enrolled.
Every datetime the API emits for versions (flow summaries' updated_at,
runs' created/started/finished with flow_version) is already ISO 8601
with Z or +00:00. A test holds it and the API docs say so.
CACHE_SHAPE_SINCE marks the last change that moved every cache key. The
run service's periodic sweep clears cache_key on rows of runs created
before it, which nothing can match any more. Their outputs stay.
DELETE /runs?flow= removes every Run row of the flow but left its flow_run
rows, so a deleted flow kept appearing in the history. history=true now
drops those too, and the flow delete dialogs (editor and bulk) send it when
"Also delete its runs" is ticked.
Migration ccb21eacf5d5 deletes rows whose run is gone, then adds foreign keys
with ON DELETE CASCADE from run_node, run_metric and run_artifact to run.id.
Deleting runs is now the one statement. A node or metric written after its
run was deleted is dropped with an info line. Migrations run with foreign keys
off, since a batch rebuild of run would otherwise cascade into every child row.