OrcaSlicer's "Send job" flow sat on "Synchronizing device information…"
until it gave up, even though FTP upload worked when the user clicked
"Send job anyway". The virtual printer's MQTT server gated all incoming
command handling on `f"device/{self.serial}/request" in topic` — if the
slicer's cached serial for the VP didn't exactly equal the VP's computed
self.serial (model prefix + per-VP serial_suffix), every get_version,
pushall, and project_file publish was silently dropped. Nothing was
logged past the initial "MQTT publish to …" line, so the slicer never
received a push_status or get_version response on its subscribed
device/{serial}/report topic and hit its sync timeout. Responses were
also unconditionally published on device/{self.serial}/report, so even
when the inbound check happened to pass, replies targeted a topic the
slicer wasn't listening on if its serial had drifted.
Both directions are now serial-adaptive:
- `_handle_publish` accepts any authenticated publish on a
`device/*/request` topic and extracts the serial from the topic itself
rather than comparing against self.serial.
- A per-connection `_client_serials` dict tracks the serial the slicer
actually uses, populated from the first SUBSCRIBE or PUBLISH seen on
each connection and cleared on disconnect/stop.
- `_send_status_report`, `_send_version_response`, `_send_print_response`
now take an optional `serial` parameter (defaulting to self.serial)
so every outgoing publish — including the periodic 1-second status
push — targets the topic the slicer subscribed to.
- The version response's embedded `module[].sn` fields now also carry
the client's serial so the payload is internally consistent with the
topic.
- When the client's serial differs from self.serial an INFO log records
the adaptation so it's visible in future support bundles.
The working case (slicer's cached serial equals self.serial, as in my
own H2D-1 Proxy setup) is bit-for-bit identical to the old behavior —
the new check is strictly more permissive and only affects cases the
old code silently dropped.
Regression tests cover:
- `_extract_serial_from_topic` valid/invalid topic shapes
- mismatched-serial publish → handler runs, response topic and sn field
both use the client's serial
- non-`/request` topics → still rejected
- pushall → status_report routed to the client's subscribed topic
- `_client_serials` cleared on stop()
detect_current_branch() was reading .git/HEAD from settings.base_dir,
which points at the data volume (DATA_DIR=/app/data in Docker) and
never contains .git. The repo is at /app, so the lookup always failed
and the code fell through to the GIT_BRANCH env-var → "main" fallback.
The SpoolBuddy device was therefore checking out `main` regardless of
which branch Bambuddy itself was running.
The old subprocess-based implementation had the same bug but it was
masked: the stock Docker image has no `git` binary, so `git rev-parse`
raised FileNotFoundError, the except clause swallowed it, and the
fallback kicked in. Swapping to filesystem reads exposed the wrong
lookup path.
Add a module-level _APP_DIR constant (parents[3] of the module file,
same depth as config.py uses for its own _app_dir) and read `.git/HEAD`
from there. A regression test plants a decoy .git in the data dir and
asserts we still pick up the real one from the app root.
Per your NO GIT WRITES rule, nothing is staged or committed.
Follow-up to the asyncssh migration. asyncssh.connect() internally
calls getpass.getuser() for ~/.ssh/config host matching, regardless
of the explicit `username=` passed for the remote login. Under an
arbitrary Docker PUID with no /etc/passwd entry, getpass.getuser()
raises "No username set in the environment" (OSError in Python 3.13+,
previously a bare KeyError).
Fix: set LOGNAME=bambuddy, USER=bambuddy, HOME=/app in the Dockerfile.
getpass.getuser() tries env vars before pwd.getpwuid(), so the lookup
never touches the passwd database and works for any PUID the operator
picks — no helper code, no image rebuild for different UIDs.
Also pass config=[] to asyncssh.connect() so it does not try to load
~/.ssh/config (whose default path needs a resolvable home directory).
An earlier draft of this fix added a Python helper that caught the
KeyError and injected LOGNAME at module import. That was both more
code than needed and broken on Python 3.13, which wraps the KeyError
in an OSError the helper didn't catch — so the module import itself
crashed, producing a 500 on /spoolbuddy/devices/{id}/update. Reverted
in favour of the one-line ENV fix.
Follow-up to the previous commit that swapped the `ssh`/`ssh-keygen`
subprocesses for asyncssh. asyncssh.connect() internally calls
getpass.getuser() to resolve the *local* username for ~/.ssh/config
host matching, regardless of the explicit `username=` we pass for the
remote login. Under an arbitrary Docker PUID with no /etc/passwd
entry, getpass.getuser() tries LOGNAME/USER/LNAME/USERNAME (all unset
in python:3.13-slim) and falls back to pwd.getpwuid(), which raises
KeyError. asyncssh rewraps that as "Unknown local username: set one
of LOGNAME, USER, LNAME, or USERNAME in the environment" — which
surfaced in the UI as "ssh connection failed: no username set in the
environment".
Fix is two-part:
- _ensure_local_username_env() runs at module import. If getpass
.getuser() already works, or any of LOGNAME/USER/LNAME/USERNAME is
set, it is a no-op. Otherwise it sets LOGNAME=bambuddy so asyncssh
can proceed. Native installs are untouched.
- asyncssh.connect() is now called with config=[] to skip the
default ~/.ssh/config load, which relies on a resolvable home
directory that may not exist under arbitrary Docker PUIDs.
Three new unit tests cover the env-var fallback, including the case
where the operator has set USER but the passwd lookup still fails.
Commit 67749565 eliminated ssh-keygen from the SpoolBuddy remote-update
flow, but the update path still shelled out to the OpenSSH `ssh` client
for every command. Like ssh-keygen, the `ssh` binary calls
getpwuid(getuid()) during startup and aborts with "No user exists for
uid <N>" when the container runs under an arbitrary PUID that isn't in
/etc/passwd (python:3.13-slim only ships a root entry, so any
`user: "1000:1000"` compose setup trips the same error).
detect_current_branch() had a related problem: when the git repo is
bind-mounted into the container, .git exists inside Docker, so the code
tried to run `git rev-parse`. Git isn't in the image, so the subprocess
silently fell back to the GIT_BRANCH env var — and if git ever were
added, it could hit the same getpwuid trap.
The entire update path is now subprocess-free:
- _run_ssh_command uses asyncssh (pure-Python, built on the already
installed cryptography library). Connection errors map to rc=255 to
match `ssh`'s convention; asyncio.timeout handles the timeout path.
- detect_current_branch reads .git/HEAD directly (handling git-worktree
`gitdir:` pointer files too), keeping the same GIT_BRANCH → "main"
fallback chain.
- shutil and the inline `import subprocess` are gone from the module.
Regression tests assert that neither keypair creation, branch
detection, nor command execution spawns any subprocess. Native installs
are unaffected.
The SpoolBuddy remote-update flow shelled out to `ssh-keygen` to create
its update keypair on first use. Inside the Docker container the process
runs under an arbitrary PUID that is not listed in /etc/passwd, so
ssh-keygen aborted at the getpwuid() home-directory lookup with
"no user exists for uid 1001" and the update button failed.
Generate the ed25519 keypair in-process via the `cryptography` library
(already a dependency) and serialize it in OpenSSH format. No subprocess,
no /etc/passwd lookup. Native installs are unaffected.
Added a regression test that asserts no subprocess is spawned during
keypair creation so this can't come back.
Two bugs surfaced while investigating camera reconnect behaviour in #925.
The camera page briefly displayed "Reconnecting attempt 6 of 5" before
giving up, because the attempt counter could be incremented to the
maximum while the reconnect banner was still rendering. The displayed
value is now clamped to the configured maximum.
Every failed ffmpeg spawn logged the full ~20-line ffmpeg version,
configuration, and lib* banner, producing hundreds of lines of noise
per failed camera click (one reported click produced 555 log lines
across 30 retries). A new _summarize_ffmpeg_stderr helper strips the
banner and caps output at the last 10 meaningful lines, applied at
all three stderr log sites (immediate-failure, stream-ended,
read-timeout). Covered by unit tests for empty input, banner
stripping, line cap, blank-line filtering, and banner-only input.
The underlying "camera service stops accepting connections after
prolonged uptime" behaviour in the X1C firmware is still under
investigation — these two fixes are independent of that root cause.
Two related LDAP authentication changes.
Fix: POSIX primary group membership was ignored. authenticate_ldap_user
only searched for posixGroup entries via memberUid (supplementary
groups). A user's primary group — referenced by the gidNumber attribute
on the user object matching gidNumber on a posixGroup — was never
resolved, so users whose role came from their primary group landed
without the expected permissions. The authenticator now runs a second
search for posixGroup entries whose gidNumber matches the user's
primary gidNumber, then dedupes DNs case-insensitively before passing
the list to resolve_group_mapping (LDAP DNs are case-insensitive by
spec).
New feature: ldap_default_group setting. Settings → Authentication →
LDAP → Advanced has a new "Default group" selector. When an LDAP user
authenticates but is not listed in any mapped LDAP group, they are
assigned to this fallback group instead of being left with no groups
(and therefore no permissions). A warning is logged each time the
fallback is applied so admins can spot missing group assignments.
Empty setting preserves the old behavior.
Tests: added 4 mocked authenticate_ldap_user tests covering primary
gidNumber lookup, dedupe of overlapping memberUid+primary gid matches,
case-insensitive DN dedupe, and the guard when a user entry has no
gidNumber attribute. Also extended the existing parse_ldap_config tests
to cover the new default_group field.
Backend: ldap_service.py (primary group + dedupe + default_group
field), schemas/settings.py (schema field), api/routes/auth.py
(fallback wiring in _provision_ldap_user / _sync_ldap_user).
Frontend: LDAPSettings.tsx default-group dropdown in the Advanced
collapsible, api/client.ts type field, new i18n keys in all 7 locales
(defaultGroup, defaultGroupNone, defaultGroupHint).
The on_ams_change auto-sync callback set locations for new spools but
never called clear_location_for_removed_spools(), leaving stale locations
that caused double-booked slots. Also pass synced_spool_ids in the
single-printer sync route to match the sync-all endpoint behavior.
Users can authenticate against an LDAP/AD server with configurable
server URL, bind DN, search base, and user filter. Supports StartTLS
and LDAPS — plaintext is not allowed. Both Active Directory (memberOf)
and POSIX groups (memberUid) are mapped to BamBuddy groups on each
login. Auto-provisioning creates local accounts on first LDAP login.
Local admin accounts remain as fallback when LDAP is unreachable.
Password management is disabled for LDAP users.
When auto_archive was disabled but the print was dispatched by BamBuddy
(queue/reprint), on_print_start discarded the expected print entry and
returned early. The archive was never promoted to _active_prints, so at
completion archive_id and ams_mapping were both None — all tracking paths
failed silently. Now detects expected prints before the auto-archive
early-return and falls through to the normal promotion path. Also injects
the stored ams_mapping into the usage tracker session for printers where
MQTT request topic subscription fails (P1S, A1).
When multiple smart plugs were assigned to the same printer, only the
first plug's automation triggered. All automation paths (print start
auto-on, print complete auto-off, queue auto-off, scheduler power-on)
now iterate every plug linked to the printer. Also fix queue auto-off
hardcoded to Tasmota instead of using the correct service for the plug
type.
Queue status update (printing → completed) failed silently when SQLite
was locked by another writer, leaving ghost jobs permanently stuck in
printing status. Add run_with_retry() for SQLite lock retries and split
runtime tracker into per-printer commits to reduce lock hold time.
The dev mode probe (ams_filament_setting to ext slot) fired on every
auto-reconnect, which destabilized some firmware MQTT brokers (A1/P1)
causing a reconnect-probe-disconnect feedback loop. Now caches the
probe result across reconnects and only probes once on first connect,
with a 5s delay to let the session stabilize.
When auto-archive was off, archive_id was None at print completion so
the entire 3MF tracking path was skipped. AMS remain% fallback also
failed on printers reporting remain=-1. Now searches library files and
previous archives by filename to locate the 3MF without an archive,
and captures the AMS slot-to-tray mapping at print start so it's
available at completion regardless of archive state.
Per-model start/end G-code snippets configurable in Settings (Workflow
tab). Queue items get "Inject G-code" toggle — scheduler injects
snippets into a temp 3MF copy before FTP upload. Supports Farmloop,
SwapMod, AutoClear, Printflow 3D and similar bed-clearing systems.
Original files are never modified.
Add timeout and retry to the developer mode probe. After a keep-alive
timeout, paho auto-reconnects but the session can be half-broken: the
printer sends status but ignores commands. The probe had no recovery —
one unanswered probe permanently blocked retries. Now times out after
10s with one retry; two consecutive failures force-close the socket for
a clean reconnect.
Move H2S from the drying-unsupported blocklist to the firmware-gated
list, requiring minimum firmware 01.02.00.00 (released end of March
2026). Both remote AMS drying and queue auto-drying now work on H2S.
Bambuddy can now use an external PostgreSQL database via the
DATABASE_URL environment variable. SQLite remains the default.
Dialect-aware helpers handle upserts, PRAGMAs, FTS (FTS5 vs
tsvector+GIN), backup/restore, and health checks. All migration
blocks use savepoints to prevent Postgres transaction poisoning.
Backups are always portable SQLite format regardless of backend.
Cross-database restore imports SQLite backups into PostgreSQL
with automatic boolean/datetime conversion, NOT NULL default
filling, and FK constraint handling.
REST/Webhook smart plugs can now fetch power and energy data from
individual URLs instead of requiring all values in a single status
response. Each value falls back to the shared Status URL when no
separate URL is set, preserving backward compatibility. Added power
and energy multipliers for unit conversion (e.g. 0.001 for Wh→kWh).
Race condition in _update_state: dev mode probe released GIL via MQTT
publish between raw_data overwrite and vt_tray list restoration, letting
the event loop iterate over raw dict keys (strings) instead of spool
dicts. Affects A1, P1, and X1Plus firmware that don't send the fun field.
Fix: normalize vt_tray dict→list before raw_data assignment, restore
preserved fields before any GIL-releasing work, add defensive guard in
printer_state_to_dict.
Single-nozzle printers (X1C, P1S, A1) report tray_now=254 for external
spool, but BambuStudio sends ams_id=255 (VIRTUAL_TRAY_MAIN_ID) in the
print command's ams_mapping2 field. Bambuddy was passing 254 as-is,
causing firmware to target AMS tray 0 instead of external spool —
resulting in 07FF_8012 "Failed to get AMS mapping table" or prints stuck
at heatbed heating when an AMS is connected but empty.
Map external spool to ams_id=255 for all non-H2D printers. H2D
dual-nozzle printers retain 254 (deputy) / 255 (main) distinction.
When multiple AMS spools match the same type/color criteria, an optional
setting now prefers the spool with the lowest remaining filament. This
helps consume partial spools before starting new ones. Sorting applies
to all matching paths: queue scheduler, print modal, and multi-printer
mapping. Unknown remain values (-1) sort to end.
Admins can now filter the Statistics page by user via a new
stats:filter_by_user permission. A user dropdown appears in the stats
header showing all users plus "No User (System)" for prints without
attribution. The filter applies to all stats widgets, failure analysis,
and CSV/Excel exports. Backend validates the permission on all 4 stats
endpoints, returning 403 if the filter is used without authorization.
Stagger option now available when printing directly to multiple printers,
not just in queue mode. Prints are automatically queued with staggered
start times using group size/interval from Settings. New "Require
plate-clear confirmation" setting lets farm users disable per-printer
plate confirmations so queued prints start automatically on finished
printers.
Also fixes settings API type parsing for require_plate_clear (boolean),
stagger_group_size and stagger_interval_minutes (integer) — without this,
saved values returned as strings would cause the settings toggle to
always show enabled and trigger a permanent save loop.
Printers with no AMS hardware (P1S/P1P with only external spool)
rejected print commands with "Failed to get AMS mapping table"
because use_ams was always sent as true. Now auto-sets use_ams=false
when all filament slots map to external spools or are unmapped.
H2D-series excluded since they use use_ams for nozzle routing.
Auto-off now has a "Keep Enabled" toggle that keeps it active between
prints instead of disabling after each use (one-shot). Useful for HA
accessories like BentoBox filters that should always power off after
prints. Default behavior (one-shot) is unchanged.
The stale connection fix (6ac375a) force-closed the socket to trigger
paho reconnect, but paho's disconnect callback redundantly broadcast
state changes, causing rapid connected/disconnected cycling in the UI.
Now sets a flag before the force-close so _on_disconnect skips the
redundant broadcast. Also relaxed keepalive from 15s to 30s — the
aggressive interval caused spurious disconnects on transient hiccups.
Moved _disconnection_event.set() to the top of _on_disconnect so no
early-return filter can prevent it from unblocking disconnect() callers.
Removed leftover debug payload dump.
The daemon now collects CPU temp, core count, load average, memory/disk
usage, OS info, and system uptime every heartbeat using stdlib-only reads
from /proc and /sys. Stats are sent as a JSON blob in the heartbeat
payload, stored in a new system_stats TEXT column, and displayed in a
new "System" tab in SpoolBuddy Settings with color-coded usage bars.
Assigning or unassigning a spool now broadcasts a spool_assignment_changed
event to all connected WebSocket clients. The frontend handles this event
by invalidating the spool-assignments and slotPresets caches, so other
open tabs update automatically without a page reload.
BambuStudio converts virtual tray IDs (254/255) to -1 in the flat
ams_mapping and relies on ams_mapping2 for external spool routing.
Bambuddy was passing raw 254/255 in the flat array, which H2D firmware
rejects with "Failed to get AMS mapping table".
- Convert external tray IDs to -1 in flat ams_mapping (match BambuStudio)
- Fix ams_mapping2 for external trays: each virtual tray is its own AMS
unit with slot_id 0, not a shared unit differentiated by slot
- Fix main/deputy nozzle ID comment (255=main, 254=deputy per BambuStudio)
- Add 8 unit tests for start_print() mapping construction
External folder scan generated thumbnails for 3mf/stl/gcode files but
skipped image files. Added IMAGE_EXTENSIONS check with
create_image_thumbnail() to the scan loop.
Some printers (e.g. H2D) only send {id, state} in incremental MQTT
tray updates — no tray_type, tray_color, or other fields. When
filament is unloaded (state changes from 11 to 10), the old tray
data persisted indefinitely because the merge logic only updates
fields present in the incoming message.
Backend:
- Detect tray state != 11 without tray_type in incremental updates
and clear stale tray data (bambu_mqtt.py)
- Expose tray `state` field via REST API and WebSocket broadcasts
(printer.py schema, printers.py route, printer_manager.py)
- Include AMS tray content in WebSocket dedup key so load/unload
transitions trigger broadcasts (main.py)
Frontend:
- Add `state` to AMSTray interface (client.ts)
- Show configure/assign buttons for state=10 (spool present, not
loaded) but hide for state=9 (truly empty) on AMS/HT slots
- Hide fill level bar on empty slots
Tests:
- 5 new tests covering state-based clearing, preservation, reload,
and idempotency
When a Bambu Lab spool is detected in the AMS but no tag match exists,
check for an untagged inventory spool with matching material/subtype/color
before creating a new entry. Links the RFID tag to the existing spool
(data_origin="rfid_linked") to prevent duplicate inventory entries.
15 backend unit tests covering keypair management, branch detection,
SSH command execution (success/failure/timeout), and full update flow
with error handling. 4 new frontend tests for the Updates tab (version
display, buttons, SSH setup, update available state).
BambuStudio connects to undocumented proprietary ports 2024-2026
on A1/P1S models during the print flow. The proxy wasn't forwarding
these ports, causing connection refused (RST) and triggering the
access code dialog instead of printing. MQTT and FTP worked fine —
port 2024 was the sole blocker.
Added transparent TCP pass-through proxies for ports 2024-2026,
following the same pattern as the existing FileTransfer (6000) and
RTSP (322) proxies. Silently ignored on models that don't use them.
When targeting a specific printer, the scheduler's power-on-wait loop
created new MQTT clients on each attempt. Each new client re-tried the
request topic subscription, which some brokers (e.g. A1) reject by
disconnecting. This caused a 170s thrash loop leaving the connection
fragile, so the eventual print command silently failed to reach the
printer.
Cache request topic support per serial number at the class level so
new client instances inherit the knowledge and skip the subscription.
When a printer shuts down it sends a final MQTT message with
tray_exist_bits=0 and power_on_flag=false. The tray_exist_bits
clearing code processed this all-zero value, wiping every AMS
slot's filament data. On reconnect, the auto-unlink check saw
empty tray data (no color, no type) and deleted all spool
assignments as "fingerprint mismatch".
Fix: skip tray_exist_bits slot clearing when power_on_flag is
false. Defaults to true when absent for backwards compatibility.
Adds 3 regression tests covering shutdown preservation, genuine
removal still working, and missing power_on_flag fallback.
Replace fixed 30-second debug log collection with an interactive
3-step flow: start logging, reproduce the issue, stop & submit.
Users now control timing instead of racing a countdown.
Backend: split _collect_debug_logs() into POST /start-logging and
POST /stop-logging endpoints; add debug_logs field to submit request.
Frontend: 3-step progress indicator with elapsed timer, pulsing
active state, and 5-minute auto-stop. Updated all 7 locale files.
Add a "Rotate spool during drying" checkbox to the manual drying popover
for AMS 2 Pro and AMS-HT units. The firmware-level rotate_tray field was
already sent (hardcoded to false) — this makes it user-configurable.
The checkbox defaults to unchecked and resets each time the popover opens.
Firmware silently disables rotation if filament is currently loaded.
When all matching printers were busy and a job was queued with ASAP
timing, the scheduler immediately fired a "Job Waiting for Filament"
notification even though the job was just waiting for a printer to
finish — no user action required.
Added _is_busy_only() check to skip the waiting notification when the
only reason is "Busy". Notifications still fire for actionable reasons
(missing filament, offline, wrong color). Also renamed the default
notification template title to "Queue Job Waiting" and updated
descriptions across all 7 locales.
The closed-source bambu_networking DLL validates TLS connection parameters
and rejects connections where the certificate doesn't match the printer's
real BBL CA certificate. The TLS-terminating proxy presented Bambuddy's
own certificate, causing X1C/X1 prints to silently fail after verify_job.
Switch to transparent TCP proxying for FTP, FileTransfer, Camera, and FTP
data — only MQTT remains TLS-terminated (required for IP rewriting). The
slicer now gets end-to-end TLS directly with the printer's real certificate.
Changes:
- SlicerProxyManager uses TCPProxy for FTP (990), FileTransfer (6000),
Camera (322), and pre-listens on FTP data ports (50000-50100)
- Only MQTT (8883) uses TLSProxy for IP rewriting
- Remove debug logging from MQTT and FTP proxy code
- Fix install.sh missing AmbientCapabilities=CAP_NET_BIND_SERVICE
- Update module docstring, migration docs, README proxy description
- Add tests verifying transparent proxy architecture
When the slicer and printer are on different VLANs, Bambu Studio could
not send prints through the proxy because the printer's real IP leaked
through MQTT payloads, the bind protocol forwarded the real printer's
identity, file transfer and camera ports were not proxied, and FTP
data connections raced the TLS handshake on zero-byte uploads.
- Rewrite IP addresses in MQTT PUBLISH payloads (string + integer)
with proper packet framing and cross-chunk buffering
- Respond to bind/detect with VP identity via BindServer
- Add TLS proxies for port 6000 (file transfer) and 322 (RTSP camera)
- Buffer slicer FTP data during printer connection setup
- Advertise configured VP name in SSDP proxy
- Add cross-subnet SSDP wildcard listener for VPN setups
- Register UserEmailPreference model in models/__init__.py
- Add 11 unit tests for MQTT rewrite, IP conversion, SSDP name
maziggy