A printer kept dark between uses showed a black live view and sent black
notification photos (#1655). Settings → Camera can turn the light on for
all printers or for selected ones, for the live view, notification
photos, the finish photo and the snapshot API that Home Assistant uses,
and turns it off 15 seconds after the last use. It only turns off a
light it turned on, and switching the light by hand hands it back.
Snapshots wait an adjustable delay after the light comes on. Timelapse,
the Obico check and the in-print frame bank don't switch it, since they
would make it flash.
The camera card no longer lists every printer: external cameras are
added from a searchable printer picker and only those printers are
listed, so large farms get a short page.
Fetch a signed feed.json from the public bambuddy-notifications repo on
GitHub at startup and every 6 hours. Nothing about the install is sent;
targeting (version, beta channel, install type) is decided locally.
- Ed25519 against a key built into the app; an older serial is refused so a
withdrawn message can't come back. The feed replaces the stored list, and
a failed or rejected fetch keeps the last good one.
- Sidebar entry above System with an unread count, a slide-over list, and a
banner for unread important/critical messages. Read state per user.
- Admins by default; Settings > General > Updates can show them to all
users or switch them off, which also stops the fetch.
- Plain text only; links to github.com and bambuddy.cool only.
A PUT /settings with null for a number setting stored the literal
"None". From then on int()/float() raised inside the response builder,
and every settings read returned 503 until the row was fixed by hand.
Explicit null for any boolean or numeric setting is now refused with a
422 that names the keys, and nothing in that request is saved. A numeric
row that does not parse reads back as its default with a warning, so an
install that already stored one recovers. The typed-key lists moved to
module constants so the save check and the read path share them.
POST /notifications/app-message delivers an app's message to every channel
with the new "Messages from connected apps" switch on (off by default),
through quiet hours, the digest and the log. API keys need the new "Send
notifications" permission, and their owner notifications:update; plain text,
http(s) links, 20 messages a minute per key. The electricity-price door and
this one now share one scoped-key check. /queue?batch=<id> opens and
highlights one batch order.
print_archives.library_file_id -> library_files.folder_id ->
library_folders.archive_id -> print_archives. Three nullable SET NULL
links, each reasonable alone, that together made a loop
metadata.sorted_tables could not sort: it dropped those edges, warned on
every backup and every restore, and could return an order placing a
child before its parent -- which once imported library_files ahead of
library_folders and killed a restore on a ForeignKeyViolation.
The restore no longer depends on that order (it strips every foreign key
before importing and adds them back after), but the backup export sorts
the same way, and the warning ends with "may raise an error in a future
release" -- which would break backup and restore on one upgrade.
Marking one edge use_alter removes it from the sort graph, not from the
database: PostgreSQL emits it as ALTER TABLE ADD CONSTRAINT, as it
already did for every constraint on these three tables, and SQLite
inlines it into CREATE TABLE, so ON DELETE SET NULL holds on both.
Verified against PostgreSQL 16 and SQLite.
cost_centers:read_own exists so a non-admin can see their own wallet, balance
and cost-centre spend, and the Finance page honoured it -- typing the URL
worked and rendered their balance. The sidebar never offered the entry.
It decides whether to show Finance by reading billing_enabled from
GET /settings, which requires SETTINGS_READ. A non-admin gets 403 there, so
the value arrived undefined, `undefined !== true` held, and the entry was
hidden from precisely the users the permission was written for. The permission
map and the route guard were both already right; only discovery was broken.
Three more fields came from that same 403, and one of them failed the other way
up. The Notifications gate tests `=== false`, which undefined never satisfies,
so an administrator who switched user notifications off still left the entry
showing to the non-admins it governs. Nobody reported that one, and no
administrator could have reproduced either: administrators can read /settings.
The remaining two were quieter -- the sponsor prompt fell back to EUR whatever
the install uses, and the update check ran where it had been turned off.
SETTINGS_READ cannot be the price of knowing whether billing is on. It also
grants sight of the SMTP, LDAP and MQTT credentials, which is the reason
/settings/ui-preferences exists at all.
So: a second endpoint, GET /settings/ui-flags, carrying those four fields and
asking only that the caller be signed in, via the existing
require_auth_if_enabled. Layout drops its /settings query altogether, which
closes the class rather than the two instances that happened to be visible.
Deliberately not four more fields on /ui-preferences. That endpoint is served
to anyone at all on the recorded grounds that its contents are "public defaults
that ship with the app" (test_route_auth_coverage.py), and its field set is
pinned by a test written to make anyone adding to it stop and think. These
fields are not defaults -- they say how this deployment is configured -- so
they get their own endpoint at their own trust level instead of stretching that
charter to fit them. require_auth_if_enabled also keeps the auth-disabled case
that /ui-preferences was ungated for: "works when there is no auth" and
"readable by anyone" are different statements, and conflating them is what put
a settings read in front of a permission that never needed one.
Twelve tests. Backend pins that the operator can read the flags, that the same
operator still gets 403 from /settings, that an anonymous caller is refused
when auth is on, that it answers when auth is off, the exact field set, that no
credential ever appears, and that the public endpoint did not quietly gain
these fields. Frontend pins Finance visible for cost_centers:read_own with
/settings returning 403, and Notifications hidden when the flag is off -- each
waiting on a positive signal before asserting an absence, so the negative cases
cannot pass before the query resolves.
Reported by @lonix, who traced it to the queryKey and the route gate.
Turning Spoolman mode on ran an unfiltered delete(SpoolAssignment) across every
printer. Turning it straight back off cleared the other table instead, so the
two directions were symmetric in code and one-way in effect, and the setting
auto-saves on a 500 ms debounce with no save button and no confirmation.
Opening the settings page to see what the option did was enough to destroy the
configuration: the reporter's log shows four toggles in 85 seconds, which is
someone looking and reverting, and the assignments never came back.
The deletion was not careless. Checks that read both assignment tables would
otherwise let a row in the mode you are not using answer for the mode you are,
which is how #1473 was fixed, and emptying the inactive table made that
impossible by construction. The cost was that the guarantee was bought with the
user's data. That decision belongs to the readers -- the mode is a property of
the install, not of the rows -- so spoolman_owns_assignments now answers it and
nothing is deleted on a toggle. Each mode keeps its own assignments and
switching is reversible. Existing installs need no migration: their inactive
table is already empty, because it was being emptied.
Six sites had to be told which mode they meant, and only two of them are the
reads you would guess at, the missing-assignment notification and the queue cost
estimate. The per-slot K-profile lookup consults the built-in table first and,
on a hit with no matching profile, deliberately stops rather than falling
through to Spoolman, so a leftover row would have shadowed the Spoolman binding
for that slot -- the symptom #1556 reported from the other direction.
configure_ams_slot *writes* a K-profile against whichever table answers first,
so the same leftover would have filed a calibration against a spool the printer
is not drawing on and never written the local one, leaving a calibration that
appeared to succeed and then did not apply.
The auto-unlink pass in on_ams_change is the one that would have made this
change worthless. It drops any assignment whose tray no longer matches the
fingerprint it recorded, and it ends in db.delete. Ungated, it would have
removed the preserved rows one slot at a time as the AMS contents changed under
the other mode -- the same loss, arriving slowly enough not to be connected to
the toggle that caused it.
The sixth is the built-in remaining-weight fallback inside the Spoolman AMS
sync, and it is deliberately left inert rather than woken up. It could never
fire while the table it reads was being emptied, it is keyed by slot rather than
by spool, and create_spool writes remaining_weight unconditionally where the
update path does not -- so preserving the rows would have seeded a stale figure
into a brand new Spoolman spool the first time a tray reported an unusable
remain%. The query stays, gated off, so the intent survives for whoever
revisits the cross-mode fallback.
Separately, a print that could not debit a spool said nothing about it, and that
is what turned a mis-click into lost filament. The reporter's print was already
running when they toggled. At completion it resolved its 3MF, read its
per-filament grams, resolved its tray, and then skipped the debit because the
assignment row no longer existed -- logged at INFO, invisible under the default
log level, while the completion notification fired as usual. 65.49 g was never
deducted and they only noticed because a spool's remaining weight looked wrong.
_resolve_spool_id_for_tray has no tag or fingerprint fallback, so there was
nothing else to catch it.
The skip is now a warning naming the grams, and a completed print that failed to
charge a tray it drew from raises the missing-spool-assignment notification. The
print-start check cannot cover this and was right to stay quiet: the assignments
existed when it ran. The two are different statements -- the first says the
weight may not be tracked, the second says it was not -- so a print warned at
start will notify twice, which is the right trade. Collected across the print
rather than fired per slot, and given the caller's session, because this runs
inside on_print_complete's transaction and opening a second one to read the
printer's name would deadlock against it on SQLite. This is independent of the
toggle and catches any other cause of an assignment disappearing mid-print.
Restoring a SQLite backup into PostgreSQL died part-way with
insert or update on table "library_files" violates foreign key
constraint "library_files_folder_id_fkey"
DETAIL: Key (folder_id)=(1) is not present in table "library_folders".
The import recreates the schema and is supposed to create every table
without foreign keys, so the order rows arrive in cannot matter; the
constraints are added back once the data has landed. Phase 1 did that by
discarding each ForeignKeyConstraint from table.constraints before
create_all -- which only suppresses the inline REFERENCES clause.
Table.foreign_key_constraints is derived from the columns' ForeignKey
objects and was never touched, and when create_all meets a dependency
cycle it cannot sort, it falls back to emitting those tables' keys as
separate ALTER TABLE ... ADD FOREIGN KEY statements read from exactly
that property.
library_files, library_folders and print_archives form such a cycle, so
twelve constraints survived across the three of them -- measured against
a real PostgreSQL by running the old phase verbatim. The same cycle also
costs those tables their place in sorted_tables, so they were imported
alphabetically, putting library_files ahead of the library_folders rows
its folder_id references.
Phase 1 now creates the tables normally and drops every foreign key from
pg_constraint afterwards, in the same transaction, scoped to contype 'f'
in the public schema. That is indifferent to how create_all chose to
emit them, so a future cycle between other tables cannot bring this
back. Phase 3 is unchanged.
This also removes a second fault: the keys were stripped from the
process-wide Base.metadata and only restored after the drop/create
transaction, so a failure in between left the running app without them
until restart. The metadata is no longer modified at all.
Verified end to end against a real PostgreSQL -- a backup whose child
rows import before their parents restores cleanly, with all 90
constraints back afterwards. Four regression tests added.
Back-to-back prints in chamber-heated materials (ASA, ABS, PA, PC) each paid
a full heat-soak from cold, even when the print that just finished had left
the chamber at temperature. Two changes remove that cost.
Keep bed warm between prints
While a printer sits in FINISH awaiting plate-clear and the next queued item
needs chamber heat, hold the bed hot so the chamber does not cool during the
bed-clearing window. The bed is the chamber's heating element here, not a
print surface, so the hold runs at the new `queue_keep_warm_bed_temp`
(default 90C, which also satisfies bed-threshold-linked aftermarket chamber
heaters), raised to the item's own bed temperature when that is higher.
Gated on `queue_keep_bed_warm` AND `require_plate_clear` AND
`preheat_enabled`, all re-checked in the backend so a stale UI cannot leave
the feature running. `queue_keep_warm_max_minutes` (default 120) bounds the
hold: when it elapses the bed is switched off and the hold latches until the
printer is next a candidate, so a plate nobody clears cannot leave the bed
hot indefinitely. The hold is also released when the item is deleted, the
queue empties, or a gate is toggled off mid-hold, and never when firmware
reports a target other than the one it set — a temperature the user or a
print changed is left alone. Publishing is idempotent.
Smart soak reduction from chamber history
The scheduler samples each connected printer's chamber temperature every tick
into a 2h rolling history. Preheat credits time the chamber has already spent
at temperature against the configured soak, shortening or skipping it.
Credit starts no earlier than the newest sample, the most recent unbroken run
of samples, or the end of the last real dip below target. A dip only counts
once it outlasts a grace period: an enclosed chamber cannot lose and regain
several degrees quickly (measured on an X1C, cooling from 55C to below 48C
takes 23-73 minutes, ~0.2 C/min), so a brief low reading is a door opening or
sensor noise rather than lost soak — and a plate swap, which is exactly when
keep-warm runs, produces one. A stale history credits nothing: at that
cooling rate the chamber can cross the threshold unobserved, so the full soak
runs instead.
Three supporting changes to preheat itself:
* Cancelling or deleting a queued item now stops a preheat already running
for it. Those routes only write `status` to the database, which a dispatch
coroutine parked in `asyncio.sleep` cannot observe, so the heaters ran for
the rest of max_wait + soak — 45 minutes at the default settings — and the
printer stayed in `busy_printers`, blocking every other queued item behind
a print that was not happening. The routes now signal the scheduler
directly, and the stage sleeps in slices so it notices promptly and
abandons the dispatch, letting the existing rollback shut the heaters off.
* A chamber-heated print whose slicer metadata carries no bed temperature
(common for Orca-exported 3MFs) used to skip preheat entirely and start
with a cold chamber. It now heats the bed to `queue_keep_warm_bed_temp`.
A parsed bed temperature still wins, and a print with no chamber
requirement still skips — no bed temperature is invented for the print
itself. Preheat's bed target is transient regardless: the print's own
gcode issues its M140/M190 at start.
* Preheat records which commands it sent (bed, chamber, airduct) and unwinds
them if the dispatch aborts before the print starts — a failed upload, a
cancelled item, an exception — instead of leaving the printer heating for
a job that is not happening.
The photo fired the moment layer_num reached total_layer_num. That edge is
where the printer *starts* its final layer, not where it finishes it: the
reporter's H2C capture shows it arriving at 92% with mc_remaining_time=2,
three minutes and seventeen seconds and one filament change before the print
actually ended, so the frame caught the toolhead mid-print over the model.
The trigger also latched _finish_photo_captured, which locked out both the
stage-22 and FINISH triggers for the rest of the print — so on firmware that
never reports an end-of-print filament unload (H2C and A1 Mini confirmed)
nothing could replace the bad frame.
Remove the last-layer trigger. The photo is now taken at the FINISH-state
trigger, which every model sends and which lands after the toolhead parks.
Since Bambu's end G-code drops the plate ~100mm just before that, restore the
framing before capturing: absolute G90/G1 Z to max_z_height + 10mm clearance,
settle, capture, then drop it back so the print is as reachable as the printer
left it. Absolute is the safety argument — that Z is a height the toolhead
occupied seconds earlier, so it is inside the travel limits by construction and
leaves the nozzle above the part, and it is unambiguous across model families
because Z is the nozzle-to-bed gap whether the bed moves or the toolhead does.
M211 is never touched (#2579). This is what #1145, #1397 and #1565 asked for.
The height is only trusted when two independent sources agree: the archive is
matched by the finished print's subtask_name by equality (not LIKE, so "Cube"
cannot resolve to "Cube v2"), and its layer count from the 3MF must match the
layer count the printer reported over MQTT. Matching on "most recent archive
for this printer" was not safe — on_print_complete pops the _active_prints
binding concurrently, and a print Bambuddy failed to archive would have
resolved to its predecessor. A wrong height is the one failure that could drive
the nozzle into the model.
The move is additionally skipped when the print height is unknown, when a queue
item is pending for the printer, when the printer has left FINISH, and when the
new finish_photo_restore_plate setting is off.
for every FINISH-state capture — which is what shipped the mid-print photo —
the bank is used only when the dispatcher recorded that it injected End G-code
into this print, since a SwapMod snippet may have ejected the plate. The flag is
handed over in two steps (mark_pending at dispatch, adopt at print start) so it
can never outlive its print: a job started from the slicer or SD card adopts
False rather than inheriting its predecessor's answer. Those prints also skip
the plate move outright, bank or no bank.
The bank now refreshes on mc_percent advances as well as layer changes, via a
new on_print_progress callback. Layer changes stop the instant the final layer
begins, which left the #1867 fallback frame stale by the whole length of that
layer; progress keeps ticking there and freezes before the End G-code runs, so
a swapped plate still cannot reach the bank. The last-layer throttle exemption
is dropped, since it would now fire a grab on every percent tick.
On the timelapse path the moment producer returns early, so the consumer does
the restore itself before its live-grab fallback — the documented usual outcome
on P1-series, where the video has not transferred by the time the notification
goes out and the shipped photo was of an already-dropped plate. The two waits
are now derived from the settle window and the video poll timeout rather than
hardcoded; at the old flat 75s that fallback was guaranteed to be cut off
mid-settle.
extract_max_z_height_from_3mf reads only a bounded prefix of the plate G-code,
since a sliced plate is routinely tens of megabytes and the header is ~40 lines.
It returns None for missing, unparseable, zero and negative values so callers
must treat "don't know" as such rather than defaulting.
Bed levelling, flow calibration, and nozzle-offset calibration were on/off
only, so the sole way to run bed levelling was to force a full level before
every print. Bambu Studio has always offered a third "Auto" state that lets
the printer skip the calibration when it was done recently -- the state most
users actually want. Make these three options tri-state (off/on/auto),
defaulting to auto, and leave vibration/layer-inspect/timelapse as on/off
(Bambu Studio exposes no auto for those).
Wire encoding follows Bambu Studio's source exactly: each option sends a JSON
bool (true only for "on") plus a companion int -- off=0, on=1, auto=2. The
bool fields stay booleans (the #1478 H2S regression); only the companion int
widened from {0,1} to {0,1,2}. #1721's observation that stage 8/39 stays
queued when sending 2 is the auto contract (queued, skipped at runtime if
recent), not a broken "off".
- schemas: TriState = Literal[off/on/auto] with a BeforeValidator coercing
legacy bool / 0-1 / true-false so old clients and un-migrated rows validate
- model + migration: boolean columns -> String; SQLite via column affinity +
data backfill, PostgreSQL via ALTER COLUMN TYPE guarded on information_schema
(verified on both dialects); settings rows normalised true/false -> on/off
- MQTT: start_print takes the tri-state strings and emits the paired bool+int
- Virtual Printer: reconstructs the slicer's auto/on/off from the int companion
(auto_bed_leveling / extrude_cali_flag) in both capture paths
- frontend: CalibrationMode type; off/auto/on segmented controls in the print
dialog, queue bulk-edit, and Settings -> Workflow; calibrationMode_* strings
in all 11 locales
The reporter's 19-printer farm started prints "one by one", up to an hour apart.
check_queue awaited each dispatch inline, and a dispatch includes the FTP upload,
so every printer queued behind every other printer's transfer despite being an
independent machine. His logs give the arithmetic: 40978500 bytes in 254.1s,
157 KB/s - a Bambu printer's SD write, not the network, is the bottleneck. Nineteen
of those in series is ~80 minutes, and the next upload started 131 ms after the
previous one finished. The delay is linear in fleet size, which is why it got worse
the more printers he selected.
Dispatch is now collected during the (still sequential) selection loop and run
concurrently afterwards, capped by queue_max_concurrent_uploads - Settings ->
Workflow -> Queue & Dispatch, default 4, 1 restores the old behaviour. Every gate
is untouched; only the transfers overlap. The pass still awaits its uploads before
returning: _start_print flips the row pending -> printing only after the upload,
so an early return would let the next tick re-dispatch the same rows.
FTP work moves to its own thread pool. It was on asyncio's default executor -
min(32, cpu+4), six threads on a 2-core NAS, shared with everything else - which
was survivable only while uploads were serial.
Two problems the same bundle exposed:
A printer that accepts project_file but never starts (#1678) was retried forever:
270s watchdog, revert to pending, re-upload the whole file, repeat. Hence his
"printer who, since the morning, still not launch" - and on a farm each lap also
eats an upload slot the other printers are waiting on. Attempts are now counted on
the queue item; after three it fails with a message pointing at the printer instead
of queueing a fourth re-upload.
The debug bundle we asked him for held 4m49s of history. The push_status dumps fired
on every frame rather than on change - several while their own comment claimed
otherwise - which is 27,727 of the bundle's 29,830 lines and rolls 5 MB in under five
minutes on 19 printers. They now log transitions only. The bundle also read just the
live log while three rotated backups sat next to it, under a byte budget four times
larger than the file it was reading.
Migration verified on SQLite and Postgres: idempotent, backfills legacy NULLs
(dispatch_attempts + 1 is NULL for a NULL row, which would silently disable the cap).
Tests: 6 on concurrent dispatch (overlap, cap honoured, 1 == serial, default applies
with no settings row, a failed printer does not cancel its siblings, no early return),
4 on the retry budget, 6 on the bundle's rotated-log span, 7 on the debug gating.
Each verified to fail against the unfixed code - the first end-to-end log assertion I
wrote passed without the fix and had to be tightened.
On a PostgreSQL install, create_backup_zip() exports a portable SQLite copy
so backups move between engines. It rebuilt each table with only column name
+ type + PK, dropping NOT NULL, server_default/DEFAULT, foreign keys, and
unique constraints. Restore onto SQLite page-copies that schema straight onto
the live database, and post-restore init_db() can't repair it (create_all is
CREATE TABLE IF NOT EXISTS). So server_default columns like
spoolbuddy_devices.created_at (server_default=func.now()) ended up with no
DEFAULT: SQLAlchemy omits them on INSERT, the DB wrote NULL, and the next read
500'd on Pydantic validation. Every server_default column was exposed the same
way; the FK/unique loss followed from the same simplified CREATE TABLE.
Build the portable schema with Base.metadata.create_all() against a SQLite
engine instead of the hand-rolled loop, so it emits the exact DDL a native
SQLite install gets (NOT NULL, DEFAULT func.now() -> CURRENT_TIMESTAMP, FKs,
unique constraints, indexes). The data-export insert path is unchanged, and
the #1333 OIDC-icon guard is preserved automatically (LargeBinary -> BLOB),
which lets the now-redundant _sqlalchemy_type_to_sqlite_type() helper be
removed. Fixes newly-created backups; a backup from an older build still
carries the degraded schema, so re-take backups after upgrading.
Replace the #1333 type-mapping unit tests with three that inspect the real
backup schema via metadata.create_all + PRAGMA table_info: icon_data is BLOB,
created_at keeps its CURRENT_TIMESTAMP DEFAULT, a NOT NULL non-PK column stays
NOT NULL.
New scheduler stage that heats the bed (and the chamber, on supported
printers) and holds at temperature before each queued print starts —
the heat-soak engineering filaments need for adhesion and warp
control. Bambuddy waits between FTP upload and start_print, so the
soak runs while the printer is otherwise idle. M191 is silently
ignored by Bambu firmware, so doing this at the orchestration layer
is the only place it works.
Resolution order at dispatch:
1. PrintQueueItem.preheat_override ∈ {inherit, on, off}.
'off' skips entirely; 'inherit' falls back to the global
preheat_enabled toggle; 'on' forces the stage even when the
global is off.
2. chamber_target = item.preheat_chamber_target_override
?? max(filament_map[normalize(t.tray_type)] for loaded slots)
?? 0.
Mixed PA+PLA picks PA's 50 (max-across-slots — PA's chamber
requirement is binding, PLA doesn't suffer being warm). PLA-only
derives 0 and skips the chamber phase automatically.
3. Three hardware tiers for chamber heat:
- Active chamber heater (H2C/H2D/H2D Pro/H2S/X2D/X1E) → M141 +
chamber-sensor wait
- Chamber sensor only (X1C/P2S) → no M141, passive bed-radiation
wait with hard max-wait cap
- No chamber sensor (P1S/P1P/A1/A1 Mini) → bed + soak timer only
4. Airduct flap (H2C/H2D/H2D Pro/H2S/X2D/P2S) auto-switches to
match the chamber target — heating mode for engineering
filaments, cooling mode for PLA. Bambu firmware does NOT
auto-switch the flap with M141, so without this an ABS print
on a previously-cooling flap fights the open exhaust, and a
PLA print on a previously-hot flap recirculates ABS heat.
Idempotent: only fires set_airduct_mode when current ≠ desired.
Settings → Workflow → Queue & Dispatch → Preheat & Heat Soak card:
master enable toggle (default off — disabled installs see no change),
per-filament chamber-target editor (replaces a single global int that
shipped in the first cut and couldn't serve PA + PLA in the same
config), preheat_max_wait_seconds, preheat_soak_seconds. The Print
Options panel in PrintModal gets a Preheat sub-section with the
tri-state Inherit/On/Off control and an optional chamber-target
override input.
DB migration: PrintQueueItem gains preheat_override VARCHAR(10)
DEFAULT 'inherit' and preheat_chamber_target_override INTEGER NULL.
Idempotent via _safe_execute. Existing rows behave exactly as before
the migration.
Best-effort throughout: printer drops, refused M141 or set_airduct,
missing bed temp, lost MQTT state mid-wait all log and return cleanly.
Normal upload + start path runs after this returns regardless.
PR A/B turned the slice modal's preset bundle into a one-click dispatch
with a pinned target printer. PR C closes the original issue: operators
type in a number of copies, Bambuddy slices once and distributes prints
across a fleet per the pipeline's chosen fanout strategy. A new dashboard
surfaces every run with filters, expandable per-copy status, cancel,
and retry-failed-copies. WS pushes keep everything live.
Backend
- copies field on POST /run, capped by new pipeline_max_copies setting
(default 50, hard cap 1000). PipelineRun.parent_run_id chains retries.
- SlicerPipelineUpdate accepts target_kind (specific_printer /
printer_class), target_model_class, fanout_strategy.
- Eligibility matcher branches: class-targeting enumerates matching
Printer rows, runs per-printer checks via a status_lookup closure,
returns printer_reports[]. New issue kinds: no_class_matches,
class_not_set.
- _pick_assignments distributes copies per strategy:
- max_parallel: target_model set, printer_id None — scheduler picks
- round_robin: copy i → eligible[i % N], fixed printer_id
- fill_one_first: all copies pinned to eligible[0]
All three reuse the slice-once path through slice_dispatch.enqueue.
- New routes:
- GET /pipeline-runs (paginated, filterable by pipeline + status)
- POST /pipeline-runs/{id}/retry-failed (creates child run with
copies = failed+cancelled count, parent_run_id set)
- Cancel cascades to all N queue entries (only pending/queued)
- _roll_up_run_status computes run-level status from per-job statuses;
introduces partial_failure for "some completed, some failed".
- ws_manager.broadcast_to_user emits pipeline_run_updated on every
state transition with the full materialised response.
Frontend
- Pipeline editor: target_kind radio + class picker (filtered to
installed models) + fanout-strategy radio. Read-only row shows
"X1C · Round robin" for class pipelines.
- RunWithPipelineModal: copies number input bounded by
settings.pipeline_max_copies. Accepts class-targeted pipelines.
- Settings → Workflow → Queue & Dispatch: new "Slicer Pipeline limits"
card with the max-copies input.
- New /pipelines/runs dashboard page (sidebar entry, gated on
pipelines:read). Two-filter dropdown, 25-per-page pagination, per-row
expandable to job list, Cancel + Retry-failed buttons.
- useWebSocket case for pipeline_run_updated invalidates both
pipeline-runs-all and pipeline-runs/{id} query keys.
Adds a global local_login_enabled setting plus a per-provider
is_autologin flag on OIDCProvider so operators who run their own SSO
enabled, or if the calling admin has no UserOIDCLink — either would
lock everyone out. App-layer invariant: at most one provider can carry
is_autologin; setting it on one clears it on every other.
/auth/advanced-auth/status surfaces both new fields so the LoginPage
decides UI in one query. The env-var bypass flips the reported
local_login_enabled back to true so the SPA matches what the route
will accept.
Continue Auto-Drying while a print is running on capable hardware.
New Settings > Print Queue > "Continue drying while printing" toggle
(default OFF). Extends _check_auto_drying in print_scheduler.py to
evaluate running printers when supports_drying_while_printing(model,
firmware) returns true. Strict allowlist verified per Bambu wiki
release notes for "Print While Drying" / "printing while filament is
drying": H2D 01.03.00.00+, H2C/H2S/P2S/H2D Pro 01.02.00.00+, X2D/A2L
01.01.00.00+, X1C 01.11.02.00+. P1*, A1, A1 Mini, X1 (non-C), X1E
intentionally excluded. Mid-print drying temperature is capped at
max(40, preset_temp - 5) to protect spools from heat damage inside the
hot enclosure during a print, matching Bambu's own "lower drying
temperature during printing" guidance.
Rotate-spool toggle in the drying popover is now disabled when any tray
in the targeted AMS has filament threaded into the feed tube
(tray.state === 11). The whole AMS rotates as one mechanism, so a
single loaded slot locks the entire unit. Previously the toggle was
always clickable and the firmware rejected with dry_sf_reason=[3]
(ConsumableAtAmsOutlet) after the click. The first cut keyed on the
printer-level tray_now but missed the H2D's typical post-print state
where tray_now resets to 255 while filament stays in the tube — the
per-tray state field reports it correctly. Submission also clamps
rotateTray off so a stale-true state from a previous AMS can't leak
through.
Backend: supports_drying_while_printing in printer_manager.py covers
display names and internal SSDP/MQTT codes (O1D, O1E/O2D, O1C/O1C2,
O1S, N6, BL-P001, N7, N9). New print_drying_enabled boolean in
settings schema. Frontend: toggle on SettingsPage, gate + clamp on
PrintersPage drying popover using existing amsData cache. i18n: 3 new
keys x 11 locales, no English fallback. Tests: 7 cases on the gate
matrix (TestSupportsDryingWhilePrinting), 4 cases on the scheduler
mid-print path (TestMidPrintDrying), 9 cases on the rotate gate state
transitions. Full backend pytest -n 30 green (4251/4251), ruff clean,
frontend npm run build clean, i18n parity 5355 leaves per locale.
- Introduced CostCenter and related models for managing print costs and budgets.
- Updated PrintArchive and PrintQueueItem models to include cost_center_id and estimated_cost.
- Implemented budget reservation logic in finance services to validate and manage print costs.
- Enhanced ArchiveService and BackgroundDispatchService to handle cost center information during print jobs.
- Added wallet transaction handling for print charges, including partial charges based on filament usage.
- Created finance billing and budget services to manage user wallets and budget reservations.
- Ensured user finance defaults are created upon user registration, including wallets and private cost centers.
- Updated print scheduler to validate budget before processing print jobs.
New setting "Auto-add unknown RFID spools" under Settings -> Filament -> Filament Tracking,
default ON for back-compat. When turned off, the backend stops auto-creating an inventory
record for an unknown RFID tag and instead broadcasts an unknown_tag WS event that pops
a global confirmation modal in the Bambuddy UI showing the printer / AMS-X label / slot /
material / colour. Add or Cancel; no nag on every MQTT push.
Backend
- Module-level _unknown_tag_last_broadcast dict dedupes per (printer, slot, tag). Set is
committed AFTER ws_manager.broadcast() returns so a crashed broadcast doesn't poison
the dedup and permanently silence the slot.
- Empty-slot MQTT push clears that slot's entry, so remove+reinsert reliably re-prompts.
- Successful matches via get_spool_by_tag / find_matching_untagged_spool / create_spool
also clear the entry so a future tag swap re-prompts.
- Tray data (tray_type, tray_color, tray_sub_brands, tray_count) shipped in the WS payload
directly so the modal renders the real material / colour instead of relying on the
React Query cache that lags the WS event by several seconds.
- Two new endpoints back the modal's confirm action:
POST /api/v1/inventory/spools/from-slot (INVENTORY_UPDATE)
POST /api/v1/spoolman/spools/from-slot (FILAMENTS_UPDATE)
Both look up the slot's tray data server-side and create + auto-assign atomically.
- Spoolman /from-slot now raises HTTP 500 when the slot-assignment INSERT fails instead
of returning success while the DB rolled back the binding.
- sync_ams_tray gained an optional auto_add_unknown_rfid kwarg (default True so existing
callers are unaffected); auto-sync and both manual sync routes thread the setting.
Frontend
- useUnknownTagPrompt hook listens for the unknown-tag CustomEvent, reads the tray fields
out of the event detail, and feeds a single-modal queue. No long-lived dismissed set;
the backend dedup handles spam suppression.
- UnknownSpoolModal wraps the existing ConfirmModal with a material + colour-swatch
preview block.
- Mounted in Layout.tsx alongside useSponsorPrompt so SpoolBuddy kiosk / login / setup
routes are excluded.
- getAmsLabel moved to utils/amsHelpers.ts; ConfigureAmsSlotModal.tsx and PrintersPage.tsx
both import the shared version (canonical AMS-A / HT-A / External labels).
- AppSettings TS interface gained spoolman_enabled, auto_add_unknown_rfid, spoolman_url
so the runtime cast in the hook is no longer needed.
- SpoolmanSettings.tsx gets a new toggle row in the Filament Tracking card, visible in
both built-in and Spoolman branches; auto-save + toast already wired.
Reporter @thenewguy runs an engineering farm with one AMS per material
(PLA, ASA, Nylon, PVB, HIPS) — Bambuddy's single global ams_humidity_fair
threshold (default 60%) was driving both the queue / ambient auto-drying
trigger AND the hourly humidity alarm uniformly, which is wrong for
multi-material setups where Nylon wants <10% and PLA is fine at 60%.
Drying RUN parameters were already per-filament via drying_presets;
this commit adds the missing per-filament TRIGGER.
New setting ams_humidity_thresholds — JSON map of filament-type to
threshold percent with a "default" key for unknown / unmapped types.
Empty / unset → both consumers fall back to ams_humidity_fair so the
upgrade is silent.
Resolver lives in PrintScheduler.resolve_humidity_threshold(trays,
thresholds, fallback) — picks the lowest (most-restrictive) threshold
across all loaded tray types, matching the conservative-params strategy
_get_conservative_drying_params already uses for temp / hours. Empty
tray slots contribute no constraint; all-empty AMS falls through to the
"default" key. Filament names normalized to uppercase base (so
"PLA Basic" / "pla basic" both map to PLA).
Two consumer sites rewired through the same resolver so the scheduler
and the alarm path can never disagree about whether an AMS is "too
humid":
- print_scheduler.py::_check_auto_drying — per-AMS humidity comparison
for start / stop / skip decisions.
- main.py AMS sensor / alarm worker — hourly humidity alarm notifier.
UI: new table in Settings → Workflow → Auto-Drying, below the existing
Drying Presets table. Default row + 8 default filament types
(PLA / PETG / TPU / ABS / ASA / PA / PC / PVA) pre-filled from the
current ams_humidity_fair value so the editor starts sensibly.
Input pattern: draft-on-edit / commit-on-blur (transient humidityDrafts
state per row). onChange only updates the draft; onBlur (and Enter)
parses + clamps to [5, 95] + commits. Empty value on blur clears the
override and falls back to default. Caught mid-PR via a typing test:
the naive per-keystroke clamp snapped "3" → 5 before the user could
type the second digit of "30".
Setting is in the public _UI_PREFERENCE_FIELDS allowlist (same rationale
as drying_presets and ams_humidity_fair — non-sensitive integer map,
no SETTINGS_READ permission required for badge-color rendering).
New PrinterSensorHistory table + 60s recorder + GET/DELETE /printer-sensor-history
route gated behind a new PRINTER_SENSOR_HISTORY_READ scope (separate from AMS). UI
adds a 10x10 LineChart icon on each heater tile - click body opens the existing
target-temp popover unchanged, click icon opens a HeaterHistoryModal mirroring the
AMSHistoryModal shape (kind toggle + 6h/24h/48h/7d range + current/avg/min/max +
recharts line for value + dashed target). Read-only X1C/P2S chamber tile finally
gets an interaction. Retention configurable via printer_sensor_history_retention_days
(default 30, sibling of ams_history_retention_days). 8 new i18n keys translated in
all 11 locales, parity green. 4 backend + 6 frontend tests added; full pytest -n 30
6226/6226, vitest 2176/2176, ruff/eslint/build all clean.
The 24h session cap from the M-2 audit finding was hard-coded, so the
"Remember Me" checkbox could only control storage location, never
duration. Add session_max_hours setting (default 24, max 720) honoured
at all four token-issuance sites: plain login, 2FA TOTP/email, 2FA
backup, OIDC.
- backend/app/core/auth.py: SESSION_MAX_HOURS_HARD_CEILING + resolver
that clamps to [1h, 720h] and falls back to 24h on missing/blank/
unparseable. DB errors propagate — the login transaction must abort
on a broken DB rather than silently extend or shrink the lifetime.
- backend/app/api/routes/auth.py, mfa.py: all four sites read the
resolved value instead of ACCESS_TOKEN_EXPIRE_MINUTES directly.
- backend/app/schemas/settings.py, routes/settings.py: schema field
with ge=1 le=720 + int coercion in _build_settings_response.
- frontend/src/pages/SettingsPage.tsx: half-width card at top of
Settings -> Users left column with 24h/7d/30d presets, custom input,
and a yellow warning when value > 24h.
- frontend/src/i18n/locales/*.ts: 8 new keys per locale, real
translations in all 11 (en/de/es/fr/it/ja/ko/pt-BR/tr/zh-CN/zh-TW).
- backend/tests/integration/test_session_policy.py: 15 tests across
resolver clamping, login JWT exp end-to-end, settings API round-trip.
Already-issued tokens keep their original expiry; the new setting only
affects future logins.
close_all_connections() only disposes the engine's connection pool —
asyncio tasks like print_scheduler.run() and the smart-plug snapshot
loop wake on their 30 s cadence and lazily reopen pool connections
holding RowExclusiveLock on print_queue / smart_plug_energy_snapshots.
The restore's DROP TABLE ... CASCADE pass needs AccessExclusiveLock on
every public table, producing an AB/BA deadlock that rolls back the
entire restore transaction.
Reproduced 2026-06-09 restoring a native install's backup into a fresh
Docker+Postgres deploy:
asyncpg.exceptions.DeadlockDetectedError: deadlock detected
Process X waits for AccessExclusiveLock on relation 109940
Process Y waits for RowExclusiveLock on relation 110182
Fix:
- Layer 1: pause print_scheduler / smart_plug_manager /
notification_service / background_dispatch via their existing stop
affordances before close_all_connections(), with a 1.0 s sleep for
in-flight loop iterations to release sessions. Restore handler
already requires a container restart on success, so the paused
services come back via the next lifespan startup.
- Layer 2: prepend SET LOCAL lock_timeout = '10s' to the begin-block
in _import_sqlite_to_postgres so any reactive writer (per-printer
MQTT, hourly AMS history recorder) that slips through the pause
window fails fast and visibly instead of producing a new deadlock.
Reporter wanted to slice via the Bambu Studio sidecar but open files
locally in OrcaSlicer. preferred_slicer drove both the in-app
SliceModal sidecar selection AND the desktop "Open in Slicer" URI
handoff, so picking one forced the other.
New open_in_slicer setting (str | None) drives only the desktop URI;
null inherits from preferred_slicer so existing installs behave
identically. Storage in the existing app_settings key/value table;
GET normalises the "None" string back to null mirroring the
default_printer_id convention.
Frontend: Settings -> Slicer card adds a second dropdown ("Open in
Slicer" with "Same as API slicer" / Bambu Studio / OrcaSlicer);
ArchivesPage, MakerworldPage, ModelViewerModal switch desktop-URI
call sites to open_in_slicer ?? preferred_slicer. MakerworldPage's
"Slice in {{slicer}}" label additionally branches on useSlicerApi
so the label matches what the button actually dispatches.
Bambuddy's project_file MQTT payload hardcoded "nozzle_offset_cali": 2 (skip),
giving users on H2D / H2D Pro / H2C / X2D no way to control the same toggle
BambuStudio exposes. Critical for diamond-nozzle setups that must keep the
calibration off.
start_print() now takes a nozzle_offset_cali kwarg; the value is encoded as
1 (run) or 2 (skip) and gated on is_dual_nozzle so single-nozzle machines
always send 2 even if a stale flag arrives. The kwarg threads through
printer_manager, both background_dispatch sites, and print_scheduler so
every dispatch path respects the per-item setting.
print_queue gains a nozzle_offset_cali column (DEFAULT TRUE, is_sqlite()
branch for Postgres BOOLEAN). Settings default key default_nozzle_offset_cali
defaults to TRUE to match BambuStudio. Schemas updated across print_queue,
library FilePrintRequest, archive ReprintRequest, settings.
PrintModal renders the new toggle only when the selected printer is dual-
nozzle (printer-mode: nozzle_count===2; model-mode: DUAL_NOZZLE_MODELS).
SettingsPage default-print-options row + QueuePage bulk-edit tri-state both
hide unless any registered printer is dual-nozzle. Labels reuse the existing
settings.default* keys so the only new i18n strings are
settings.defaultNozzleOffsetCali / Desc and queue.bulkEdit.nozzleOffsetCali
- real translations in all 11 locales.
#1429 (reported by @TrickShotMLG02, confirmed by @Mape6 on a flat single-LAN
that rules out subnet / mDNS-reflector theories): with the physical printer
off the slicer's "Send" landed in Bambuddy's archive; once the printer
powered on every subsequent "Send" went straight to the printer's SD card
and bypassed Bambuddy. Bundle analysis: mape6-before showed clean FTP
receive + archive lines, mape6-after had zero FTP attempts to Bambuddy
once the printer was online.
Cause: mqtt_bridge.py::_resolve_client encoded _target_ip_uint32_le /
_vp_ip_uint32_le ONLY on client-identity change and early-returned on
every refresh tick when the same client object was still bound. If
target_client.ip_address was empty at first bind (DB row stale, or client
constructed before SSDP refresh filled it in), the encoding stayed None,
the net.info[*].ip rewrite block was skipped, the cache filled with the
real printer IP, sticky-key preservation kept the poisoned net value
alive across every subsequent incremental push, and the slicer followed
the leaked IP. Only Bambuddy-restart-with-printer-off cleared it — the
workaround both reporters independently arrived at. Same shape on
multi-NIC printers (X1C, H2D Pro): the rewrite only matched entries
whose ip equalled _target_ip_uint32_le, so a secondary interface IP
Bambuddy never saw would leak through unchanged.
Bridge fix:
- _resolve_client calls a new _refresh_ip_encoding() on every refresh
tick, even when client identity is unchanged; self-heals once
ip_address becomes valid.
- _refresh_ip_encoding() sweeps the existing _latest_print_state when
encoding becomes valid for the first time. Without the sweep,
sticky-key preservation keeps the pre-arm poisoned cache alive
forever — incremental pushes that don't include net carry the bad
value forward.
- _rewrite_net_info_ips() rewrites EVERY non-zero net.info[].ip entry
that doesn't already equal the VP IP, not just entries matching
_target_ip_uint32_le. Multi-NIC printers stop leaking secondary
interfaces. Zero-IP placeholders are left alone so "active interface"
detection still works.
- INFO logging on encoding arm/update and on cache sweep so future
bundles directly answer "did the rewrite fire?".
Mode wire-value rename (#1429 follow-up, separate confusion source):
- Both reporters' support bundles showed mode: immediate while the UI
said "Archive"; @TrickShotMLG02 quoted: "I have no idea why it says
immediate in the support-info.json file. In the webui the printer is
set to archive". UI button "Archive" had always saved immediate, and
"Queue" had always saved print_queue. Canonical wire values are now
archive / review / queue / proxy, matching the button labels 1:1.
- New normalize_vp_mode() + VP_MODE_* constants in
models/virtual_printer.py; manager.py normalises on construction so
a legacy row read pre-migration still dispatches correctly.
- core/database.py::run_migrations rewrites existing virtual_printers
and settings rows; idempotent (re-runs are no-ops); identical SQL
under SQLite and Postgres.
- API routes accept both legacy and canonical on input, normalise
before storage. GET /settings/virtual-printer normalises on read so
the frontend's mode-button highlight works for stale legacy values.
- Three frontend VP components (VirtualPrinterSettings,
VirtualPrinterCard, VirtualPrinterAddDialog) switched click handlers
and type aliases to canonical; each got its own normalizeMode()
helper so a stale-cached settings payload still highlights the right
button. Two pre-existing `printer.mode === 'queue' ? 'review'`
legacy mappings in VirtualPrinterCard were the source of a test
failure caught mid-implementation where the new canonical 'queue'
was being mis-aliased back to 'review' and hiding the auto-dispatch
+ force-color-match toggles.
mode handler is NOT the dispatch bug: manager.py::_archive_file (the
handler for archive mode) doesn't dispatch to the physical printer.
The "files end up on the printer's SD card" symptom was the IP-leak
from the bridge cache. The mode rename is purely clarity / support-
bundle accuracy.
Two attacker-controlled strings were being joined to library_dir with no
resolve + containment check in the project ZIP import endpoint:
- linked_folders[*].name from the request's project.json
- per-entry zf.namelist() paths from the ZIP itself
An absolute path in either field collapsed the join (Path("/lib") / "/etc"
becomes Path("/etc") because pathlib discards the left side when the right
is absolute) and the next write_bytes landed wherever the attacker chose.
Adjacent finding from the routes audit: GET /archives/{id}/photos/{filename}
had NO validation on filename and FileResponse-served arbitrary paths -
the DELETE counterpart at least gated on the photos membership check.
Adjacent finding from the services audit: ArchiveService.attach_timelapse
wrote archive_dir / filename where filename ultimately came from a printer's
FTP listing (compromised-printer threat model) or the /timelapse/select
query param. A malicious printer that exposes a directory entry with ..
segments could write the timelapse outside the archive directory.
New backend/app/utils/safe_path.py::safe_join_under(parent, *parts) is the
single source of truth: rejects empty / null-byte / absolute parts up-front,
joins under parent, resolves both sides, asserts is_relative_to. Returns the
resolved canonical path on success, raises HTTPException(400) on escape, or
PathTraversalError when http=False (for service-layer callers that need to
match a non-HTTP return contract).
Wired into the import vectors, both archive photo handlers, and the
attach_timelapse service. The full audit sweep inspected every Path/Name
join in backend/app/api/routes/ AND backend/app/services/ - 25 route-layer
sites + 8 service-layer sites confirmed safe and tagged with
# SEC-PATH-OK: <reason> so future audits trust the inline guard at a glance.
Fifth CI backstop test_route_path_arithmetic_is_safe_joined_or_marked
AST-walks both layers and fails the build on any <dir-like>/<bare variable>
join that doesn't either route through safe_join_under or carry the marker.
The services layer is in scope because it receives values verbatim from the
routes AND from external sources Bambuddy has no control over (the printer
FTP-listing case above).
SECURITY.md gets a fifth rule + a fifth row in the CI test mapping table;
the rule now names the printer FTP-listing case explicitly so future
services-layer audits set the right expectation.
--------------
fix(library): suppress warning storm when bulk-uploading ZIPs of empty/stub STL files
Uploading a ZIP of stub or empty STL files (e.g. the 24-byte
"solid test\nendsolid test" shape) produced one WARNING per file in
stl_thumbnail.py::generate_stl_thumbnail. The warnings were technically
correct - trimesh returns a valid Mesh with zero vertices, the safeguard
matches, and the function returns None so the library entry is still
created without a thumbnail - but the volume turned a successful upload
into thousands of WARNING lines in the journal.
Two changes:
1. The per-file "Failed to load STL or empty mesh" message in
stl_thumbnail.py is now logger.debug instead of logger.warning. It's
a per-file content observation, not an actionable error; the caller
already handles None correctly. The branch now catches the rare
"large enough but trimesh still can't parse it" case, visible in
debug logs without spamming production.
2. New module constant MIN_USABLE_STL_BYTES = 200 (smallest binary STL
with one triangle is 134B, smallest ASCII ~150B; 200 is a safe floor
below any real STL). The three thumbnail call sites in library.py
(extract_zip_file, single-file upload, _backfill_external_stl_thumbnails)
pre-skip files below this size before calling generate_stl_thumbnail.
Stubs never enter the trimesh pipeline at all.
Behavior is unchanged for real STLs: any file >=200 bytes runs through
the existing pipeline, MAX_VERTICES still triggers simplification at
100k vertices for the 256x256 thumbnail render, large files still get
thumbnails.
------------
fix(stl-thumbnail): silence matplotlib first-import noise (writable cache + font_manager log level)
On first STL upload, three matplotlib-internal log lines surfaced:
WARNING [matplotlib] /opt/claude/.config/matplotlib is not a writable directory
INFO [matplotlib.font_manager] Failed to extract font properties from NotoColorEmoji.ttf
INFO [matplotlib.font_manager] generated new fontManager
The writable-dir warning fired because Bambuddy's $HOME isn't writable for
matplotlib's default config path; matplotlib fell back to /tmp/matplotlib-XXX
which lost the font cache on every host reboot, so font_manager rebuilt it
each cold start - producing another batch of INFO lines.
Fix is two small additions in stl_thumbnail.py before the matplotlib import:
1. New _configure_matplotlib_cache() sets MPLCONFIGDIR to
settings.base_dir/.cache/matplotlib (mkdir if missing) so the cache
persists across container restarts and the writable-dir warning never
fires. Respects an externally-set MPLCONFIGDIR so operators who chose
their own path aren't overridden. Best-effort with a debug fallback if
settings can't be imported or the mkdir fails.
2. logging.getLogger("matplotlib.font_manager").setLevel(WARNING) at module
import demotes the per-font INFO scan that fires when font_manager
builds its cache cold. Real font warnings (>= WARNING) still surface.
3 new tests: font_manager logger at WARNING after module import;
_configure_matplotlib_cache creates the directory under base_dir and sets
MPLCONFIGDIR; an externally-set MPLCONFIGDIR is preserved verbatim.
5516 backend tests green, frontend gates clean.
notify_missing_spool_assignments_on_print_start queried only the legacy
SpoolAssignment table. In Spoolman mode that table is empty -- bindings
live in spoolman_slot_assignments -- so every used tray was flagged
missing, firing a false-positive notification on every print.
- spool_assignment_notifications.py: the assigned-tray set is now the
union of SpoolAssignment + SpoolmanSlotAssignment rows. Union-only,
so legacy-mode behavior cannot regress.
- settings.py: the Spoolman toggle cleared SpoolAssignment on switch-on
but never cleared SpoolmanSlotAssignment on switch-off. Added the
symmetric clear so stale Spoolman rows can't leak into a later
internal-mode session and mask a real missing-assignment warning.
Adds 3 notification tests + 1 mode-switch integration test. An audit
of the remaining SpoolAssignment consumers confirmed usage_tracker,
spool_tag_matcher and routes/inventory are correctly internal-mode-only.
Reporter @maziggy followed the Energy Tracking wiki literally - "create a
key with Write Settings permission, PATCH /api/v1/settings with
{energy_cost_per_kwh: ...}" - and hit:
{"detail":"API keys cannot be used for administrative operations"}.
Triage showed three independent drifts:
1. Wiki listed nine fictional API-key permissions (Read Printers / Write
Settings / Admin / ...) but the UI only ever exposed four toggles
(Read Status, Manage Queue, Control Printer, Allow Cloud Access).
There was no Write Settings toggle to tick.
2. Even if it had existed, the backend hard-denies SETTINGS_UPDATE for
every API key via _APIKEY_DENIED_PERMISSIONS - intentional protection
because PATCH /settings can rewrite SMTP/LDAP/MQTT credentials and the
HA access token. Wider surface than any documented use case needs.
3. So the wiki had been promising a workflow that was never deliverable.
Fix: introduce a narrowly-scoped door rather than relax the deny list.
- New column can_update_energy_cost (default FALSE - existing keys
never silently gain settings-write capability on upgrade).
- New route POST /api/v1/settings/electricity-price accepting
{"energy_cost_per_kwh": <float >= 0>}. Field name matches what the
wiki already documented so the HA rest_command example needs only a
URL+method change, not a payload change.
- Custom dependency require_energy_cost_update() bypasses
_APIKEY_DENIED_PERMISSIONS for this one route for API keys with the
flag set. JWT users still go through standard SETTINGS_UPDATE.
- General PATCH /settings remains denied for API keys - flipping the
narrow flag does NOT widen general settings-write access. Pinned by
test_patch_settings_still_denied_with_energy_flag.
Frontend: fifth "Update electricity price" toggle on the create-API-key
card + amber "Energy" badge on existing keys with the flag set. Three
new i18n keys across all 8 locales (German translated, English fallbacks
elsewhere).
* feat(auth): proxy OIDC provider icons server-side (#1333)
Strict img-src CSP blocked external OIDC icon hosts on the login page.
Loosening CSP was rejected via the MakerWorld precedent, so icons are
proxied: admin sets icon_url, backend fetches and caches the bytes in a
deferred BLOB column, the SPA renders from a same-origin
/api/v1/auth/oidc/providers/{id}/icon endpoint.
The Clear Plate button (and 4 other features on the Printers page) read
their state from /settings, which requires SETTINGS_READ. Granting that
permission also adds the Settings nav item and leaks SMTP/LDAP/MQTT
credentials — exactly what users were trying to avoid by giving an
operator only printers:clear_plate.
New /settings/ui-preferences endpoint returns a curated, opt-in subset
of non-sensitive fields. Matches the existing /default-sidebar-order
precedent. PrintersPage switched to the new endpoint; admin pages still
use /settings for full access.
feat(spoolman-inventory): squashed feature work for rebase onto dev
Squashed all commits from feature/spoolman-inventory-ui onto a single commit
to enable a clean rebase onto dev. Original per-commit history preserved at
backup tag backup/spoolman-inventory-ui-prerebase-20260507-105721.
chore(i18n): extend parity gate to all locales with strict/info tiers
Previously the script only inspected en/zh-CN/zh-TW, leaving de/fr/it/ja/pt-BR
drift invisible. Now locales are auto-discovered from src/i18n/locales/, and a
STRICT list (de, zh-CN, zh-TW — currently in parity) gates CI while the rest
report informationally until their drift is caught up. ja notably has 27 real
placeholder bugs worth fixing before promotion to strict.
feat(spoolman-inventory): squashed feature work for rebase onto dev
Squashed all commits from feature/spoolman-inventory-ui onto a single commit
to enable a clean rebase onto dev. Original per-commit history preserved at
backup tag backup/spoolman-inventory-ui-prerebase-20260507-105721.
Restoring a settings backup ZIP appeared to succeed but the user found
settings reverted to defaults, most printers/archive rows missing, and
~1 GB of archive files on disk with only 1 row in the database. Same
shape as #668 (closed in March without an actual fix — that user
happened to make it work by rolling back to a stable release, which
masked the bug).
Cause: the live DB runs in WAL mode. Anything the fresh container wrote
between startup and the restore call (seed_default_groups, init_db
migrations, heartbeat writes) sits in bambuddy.db-wal with valid
checksums, and engine.dispose() doesn't checkpoint it. FastAPI's
dependency injection keeps the route handler's own `db: Depends(get_db)`
session checked out across engine.dispose() (per SQLAlchemy docs,
dispose only closes pooled connections, not checked-out ones), so the
WAL inode is held open through the whole restore. After shutil.copy2
rewrote the main DB inode in place, SQLite's WAL recovery on the next
init_db() re-applied the stale frames on top of the restored content,
partially clobbering it with fresh-install state.
Initial fix attempt of deleting -wal/-shm/-journal sidecars before the
copy was insufficient (verified experimentally) — the still-open
request session reads the unlinked sidecars via held fds and bleeds
the WAL state back into the new file when it eventually closes.
Real fix: replace shutil.copy2 with SQLite's online backup API
(src_conn.backup(dst_conn)). The page-by-page protocol opens both DBs
as proper SQLite connections, acquires the right locks, and routes
new pages through the destination's own WAL. Concurrent open sessions
see their own transactional snapshot until they close (transaction
isolation) but can't corrupt the restored state.
maziggy