The port name follows the message until it is typed by hand, and clearing
it follows again. The second line holds the message, so a list port's item
type no longer squeezes the name.
The viewport effect marked the selection as centred before the layout
had placed anything, so an arrival at ?node= fitted the whole flow.
It is recorded once the fit has run.
/search lists each input and batch output under its qualified message,
and picking one opens the flow on the label's id, which the editor now
treats like a click on the label: it opens the flow's settings.
The session db fixture now disposes the engine before deleting the file, since a
module under a no-op db (tests/sdk, tests/flow) may have connected first; the
sdk import tests forget the study packages they import; the supervision tests
keep the real quarantine length so it cannot expire mid-assertion.
`Run.passed` (migration d5a3c8e1f246) keeps the names the caller chose, so
the run detail and the comparison table can mark a value as the flow's own
again; rows from before it fall back to the names their params lack. Retries
hand on their parent's list, imports record their entry's.
Also: a cancel of a queued run no longer leaves its id in `_cancelled`, a
linked comparison no longer crashes before the list loads, and the pre-commit
failures are fixed (stale client, `text-[10px]`, worker formatting).
A claimed run only joined the leased set once its pipeline was built, so a
build waiting for a busy worker pool got the run swept as abandoned. Cancel
now also stops such a run, and a terminal status being overwritten is logged.
A worker over FLOW_WORKER_MAX_RSS_MB (--worker-max-rss) when its node
returns is retired instead of kept warm, in the shared pool and in every
declared one; 0, the default, keeps them all. The engine now accounts its
own memory (FLOW_RAM / --ram, else the box or the cgroup limit less 1 GB),
so a node's `ram` declaration holds locally. A declared environment the
shared pool already has is that pool rather than a duplicate set of warm
workers.
A provider set under Settings (OpenAI-compatible; key encrypted in
flow-data/ai.json) turns on an Assistant panel in the sidebar. Each turn runs
the FastMCP tools in-process as the person chatting, streams the rounds over
SSE and keeps one transcript per user. The panel floats on any screen and
shares the editor's lane: a settings panel keeps the top half, the chat the
bottom. The MCP client is now built once, for the endpoint and the chat.
A history recorded before the move (MLflow experiments) had nowhere to
go, since runs were only made by running a flow. Each entry is stored as
it ended: parameters completed and checked like a submit, cause
"import", empty commit and code digest, metrics and artifacts written
as given, nothing queued or announced. Run.external_id (new migration)
makes a second import a no-op and is an export column so a row joins
back. The request is all or nothing and capped at 200 entries;
Client.import_runs pages a longer history.
`fluksio run` walked the working directory and imported every module in
it, which ran a research repository's analysis scripts before the run was
submitted. A flow's origin now records the file and module name that
declared it, and run/sweep import just that, under the same name so the
generated node bodies do not change. A flow with no record still walks,
now skipping what .gitignore covers.
Every code-defined flow reports `flow` changed once on its next sync,
since the origin gained two fields.
A serve that lost the port race had already signed in, so client.json
named a port another engine held and every client talked to that one.
The socket is now bound first and handed to uvicorn.
A test of the declarations had to import them from fluksio.sdk while the
guide teaches `from fluksio import ...`. The worker's inert module gets the
same two names so a module importing them still loads inside a node.
An aborted Playwright run never reaches `afterAll`, so its flows and
dashboards stayed in the store until somebody deleted them by hand. A
setup-project test now removes them before the next run, which is the
only point that is reached even when the run before it died.
The suite runs against whatever stack is up — somebody's own instance
with real flows in it — so the predicate matches only the shape every
fixture has, `test_<spec>_<base36 Date.now()>` with an optional suffix,
and only once that timestamp is an hour old.
Also correct the coverage-html fallback hint: `make clean` cannot clear
a root-owned htmlcov either, for the same reason `coverage html` cannot
write it.
`node_health` fires once at the transition, while `HEALTH_FLOOR` still
hides the node, and nothing is published when the floor expires 60 s
later. Schedule a per-node re-read for that moment, cancelled when the
node recovers and on socket teardown. Health now also invalidates the
brain graph, whose key is a sibling of the flow's rather than a child,
so Home's fault rings never moved on a health flip at all.
LoadingButton kept its own copy of buttonVariants, which had drifted past
`brand` and the icon sizes; it imports Button's now, so the Connect /
Re-pair button can show the loader instead of swapping its own label.
DashboardView mounted a second LookProvider inside the one CanvasSurface
already gives it and dropped `touch` on the way, so every widget on a
touch panel read `useLook().touch` as false. The provider is now mounted
once, by whatever owns the canvas root.
useCanvasRoot resolved the palette twice per render — once for the theme
class, once for the tokens — so useDashboardTheme takes the roles the
caller already has.
The bar assertions in widgets.spec.ts are scoped to the Load tile rather
than the page, which held only because the unbound spare bar draws no
rows.
`widgets.spec.ts` took a one-shot `boundingBox()` of a fill that is still
moving. Two ways that reads wrong. "a bar draws a row per reading" compared
two widths that are both zero until the socket has delivered the readings,
and a plain `expect` over a captured number does not retry. "a reading over
its scale fills the track and no more" allowed the fill one pixel past the
track, but the house spring is underdamped: `motion`'s own generator puts the
fluksio look at 100.66% of the track 276ms in (material 100.91%, glass
103.58%), which on this tile's ~240px track is ~1.6px — over the tolerance,
and reached only if the sample happens to land near the peak.
Neither is the widget. `fractionOf` clamps the reading to [0, 1] and
`.dui-bar-track` is `overflow: hidden`, so nothing past a full track is ever
drawn; a bounding box just is not clipped by an ancestor's overflow. So both
now poll: one until the shorter reading is drawn shorter, the other until the
fill is flush with its track — which is a stricter reading of "and no more"
than the pixel of slack was.
The live-socket test gains the other half of what that spin cost: an ASGI
task that never returns is one uvicorn's graceful shutdown waits on forever,
which is why `fluksio serve` could only be SIGKILLed and left `serve.pid`
behind.
"Add remote user" was the one call site swapping its label instead of
using LoadingButton, so it alone had no spinner.
Global search's couldMatch was a strict subsequence test while cmdk's
scorer steps over one query character for a transposition, so a typo it
would have scored never reached it. The prefilter now carries that one
skip, which recovers every single-transposition query in a sweep of the
index shapes and admits nothing cmdk scores zero.
A setting driven by a flow re-renders whatever reads it on every message.
`LookProvider` and `PaletteProvider` built a fresh context value each time,
and `CanvasSurface` called its render prop again, so a value on `theme`,
`palette`, `look`, `background`, `touch` or `locked` reconciled the whole
widget tree and re-ran every chart's `setData`.
The palette is now read as roles once per set of colours, the tokens and
both context values are memoised on it, and the canvas holds the tiles it
built while the callback and the scale stay put. `WidgetBody` is memoised
too: its two props are stable while the document is, so a grid rendered for
its own reasons no longer rebuilds every chart with it.
A connector that misses one poll turned the health badge amber and the node
on the canvas red until it recovered. `LoadedNode.health_since` marks when the
status last actually changed — a rotating error detail under an unchanged
status does not move it, or a connector retrying with a different errno each
poll would never debounce — and `unhealthy_nodes` applies a 60 s floor to it.
The floor is on the surfacing, not on the transition: the stored health is the
truth the moment a node reports it, so a flow reading node health is never
told a stale story. Only going down is held back; recovery clears at once.
A bound chord used to be `preventDefault`ed before its handler ran, so the
canvas took every ⌘C and ⌘V whether or not the node clipboard would act —
which is what blocked copying selected text out of the chrome floating over
the canvas. A binding may now return `false` to decline the key; ⌘C does so
when there is a text selection or no node selected, ⌘V when nothing of ours
is stored.
`asyncio.wait(..., FIRST_COMPLETED)` returns with at least one of the two
tasks done, so `receiver` missing from `done` already means `sender` is in
it. The `elif sender not in done: continue` could never fire.
The two runs were taken off the top of the unfiltered table, so the shot
only showed anything while the newest runs happened to be the ones the
`_results` dashboard reads. Another flow finishing a run since photographed
an empty grid — every tile "No curve was recorded", every stat blank. The
table is filtered to that dashboard's flow first.
`receive_text` raises `WebSocketDisconnect` once the peer has gone and
`RuntimeError` on every call after that. The loop swallowed both — the
exception was checked only to decide whether the frame held a media
subscription — and immediately made another read, which raised again. A
closed tab therefore left a coroutine reading a dead socket at full tilt:
around 15,000 turns a second, one orphaned `Queue.get()` per turn parked on
the bus queue's waiter list. The event loop stalled behind it, `/health`
answered 503 on its loop-lag check, and Docker's probe plus Traefik took
the API out of the pool mid-request — which is what `make verify` was dying
on, with `TypeError: Failed to fetch` from the browser.
The spin ended only when an event finally reached the send path and the
socket raised there, so it lasted from milliseconds to minutes depending on
what the bus happened to be carrying. Four abandoned sockets logged 2,518
`Task was destroyed but it is pending!` lines and up to 2.6s of loop lag;
twenty now log none and cost 1.0ms.
A failed read is the disconnect, so it ends the stream. The getter created
beside it is cancelled rather than abandoned: nothing reads it after the
turn that made it, and left behind it would also take an event the live
reader wanted.
Redis runs with appendfsync everysec, so an unclean stop loses up to a
second of journaled entries. Name that where the docstrings promised a
crash costs nothing, and state the setting and its trade-off on the
configuration page.
`wanted_names` never read its `previous` argument — the caller reassigns
`wanted` from the return value — so the parameter goes rather than gaining
an underscore; its one call site and four test calls follow. Sorts the
imports in test_artifacts.py. Both were failing `make lint-backend` at HEAD.
Regenerates the client, which 8c097d9 left behind when it corrected the
revoke_client docstring.
Home's "recently modified" order was a proxy — drafts first, then the
version counter, then the name. Both summaries now carry `updated_at`,
read as the mtime of the working copy: every write in the store commits
immediately, so a file's mtime is its commit time, and `git log -1 --
<path>` costs ~990ms across this instance's 15 documents (it walks the
history back to the last commit touching each one, so it is slowest for
the stalest) against ~1.8ms for the stats. No cache needed.
`DashboardSummary` also carries `footprint`: each widget as its type plus
the placement a panel resolves, which is the whole of what the mosaic
draws. That drops the document fetch per tile and with it the cap of
eight, past which tiles showed a name and nothing else. Verified against
this instance's 8 dashboards: the blocks are identical to what the old
client-side derivation produced.
An MCP access token is a stateless JWT good until it expires, so deleting
the client row revoked nothing already handed out — on the MCP endpoint or
on the REST API, which takes the same token directly. Both doors now look
the client up by the `client_id` the token has always carried, so tokens
already in circulation are held to it too.
A dashboard's `locked` setting stopped the client drawing a control and
nothing else; the server took a publish from a panel showing it anyway. It
now bounds the panel's write scope, resolved live where a flow drives the
flag, exactly as the client resolves it. Reads are untouched — read-only is
not blind — and so is a querying chart's request, which is how that tile
reads rather than something anyone touched.
A run record is a record of what ran, so deleting a flow no longer sweeps its
history: `_forget_runs` is gone from the flow-delete path, and `DELETE /runs/{id}`
is the only thing that removes a run one at a time. The in-flight guard stays —
that is about work, not history.
`DELETE /runs?flow=` is the counterpart to the list's flow filter and what a
reseed needs, sharing the four statements with the single-run delete and
refusing the same way while a run of that flow is still going.
A run whose flow is gone reads as one: `useFlowInputs` reports the 404 rather
than an empty declaration set, so the run page says "flow deleted", explains it,
and disables Retry, which the route would refuse anyway.
The global search capped each category before cmdk had matched anything, so
nothing past the twentieth node or widget could be found at all — this
instance has 28 nodes and 97 widgets. The cap now trims the candidates the
query could reach rather than the raw index.
The admin teardown asked /users/ for limit=1000, which the route stopped
accepting when its bounds went in; a 422 body has no `data` to iterate, so the
hook threw and took the tests it was attributed to with it. It asks for the
500 the route allows.
And `submit` folds every declared initial into a run's params, so an input the
run never passed can no longer read as the flow's own. That assertion is gone;
what the panel does show — the value the run actually started from, passed or
not — is what the test checks.
vite.config.js and vite.config.d.ts are tsc output of vite.config.ts, emitted
next to the source when the composite tsconfig.node.json is built. Nothing
reads them — and vite picks a vite.config.js ahead of the .ts, so the committed
copy was quietly shadowing the real config. Gitignore both.
The biome hook selected by the frontend/ prefix and passed the staged paths on,
but `bun run lint` runs in frontend/, so biome saw frontend/frontend/… and
reported an internal io error for every staged file. The script already lints
the whole tree, so stop passing filenames and let the pattern — now the
extensions biome parses — decide only whether the hook runs.
The icon variant lost its last caller when the responsive variant stopped
carrying a second cropped image. runOverviewQueryOptions now takes the
refetch interval, defaulting to the 30s the Runs screen and the dashboard
panel want; Home passes 10s so a short run does not linger on the tile.
agents.md spells the API host as api.${DOMAIN}, which is what the rest of
the docs use.
Several stateless instructions where a single button is one — six presets are
otherwise six tiles to place, six titles to read, and the message they share
repeated six times. Nothing is read back, as for a single button: what these
send is an instruction, and the last one sent is not a state to draw.
Auto-fit columns rather than a configured count: a tile is resized in the
editor and scaled again to whatever panel it hangs on, so how many fit is not
something the document can know.
The editor reads `cfg.buttons` raw rather than through `buttonsOf`, which
drops the blanks — a row being typed into is blank until the first keystroke.
Also carried along by the hooks: the generated SDK was stale (it had no
`RunsReadMetricNames`) and one pre-existing block in `test_panels.py` was
unformatted.
A python node's ports and settings arrive as keyword arguments, so a declared
name its `process` does not take was a TypeError on every call — and a node
that loads fine and fails every time it runs is the quiet kind of broken: the
hosted demo did it 720 times an hour for two days and the health badge read
ok throughout. `_build_node` now reads a written body with `ast` and refuses
the mismatch at load, so the node is an error on the canvas and an issue on
publish. Skipped for `**kwargs`, a decorated or absent `process`, and the
template a new node opens with. The SDK's generated shim always takes
`**settings`, so synced flows are untouched.
`/observability/summary` names a node that has failed in the last fifteen
minutes and reads degraded while it does, which is what would have made the
badge amber. `nodes.failing` carries the count.
A share link admits somebody at the portal, so this instance first hears of
them when they arrive rather than when a superuser types their code in.
An unmapped portal identity is now checked once against the portal's own
list of who may reach this instance and given an ordinary local account
only if the portal vouches for it.
Asking rather than believing the token is the point: a token stays signed
and valid until it expires, so trusting its claims would let one rebuild
the account somebody deleted here and deleting a user would stop being the
whole of the revocation.
The account-making itself moved out of the route, since both ways in build
the same thing.
Each was a loose end recorded under `### SDK` in the notepad.
`serve` takes its own pidfile down on SIGTERM. uvicorn restores the handler it
found and re-raises the signal it stopped on, so the default handler ended the
process without unwinding and the `finally` never ran — which is what a stop
sends, and what left `serve.pid` behind.
`serve.log` is cut back past 5 MB by the engine rather than by the screen that
started it, so an adopted engine is bounded too. Gated on its own stdout being
an appended regular file, which is what makes the cut safe: the kernel then
puts the next write at the new end.
Cards are counted from `/dev/nvidia[0-9]*`, so `FLOW_GPUS`/`--gpus` of 0 means
"work it out" the way `FLOW_CPUS` always has. The engine counts, not the
accountant — a remote worker builds one of those from its own inventory, and
detecting there would hand it the engine host's cards. The worker counts last:
what a batch job says it was granted still wins.
`GET /runs/metrics/names` is the distinct over a selection that `--list` and
the terminal's metric picker were approximating by reading the newest run that
had measured anything, which missed a name only an older run ever wrote.
`MetricSink` announces each batch it has written (`run_metric`, carrying the
names). Not a per-point event: one covers up to 500 points or two seconds of
them, and the rows stay the record. The terminal comparison fills in as the
first readings land instead of staying blank until reopened, and the browser
refetches the run and any comparison rather than the list behind them.
`retry --group` pages the list route by `before` instead of stopping at 500.
The terminal dashboard takes the terminal's colours (`ansi-dark`), and the web
UI can re-pair from Settings without disconnecting first.
stroblme