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.
-----
Give each test worker its own scratch folder in the #3025 and #3029 tests (issue #3025)
Both modules wrote into one shared folder under settings.base_dir and deleted
it after every test. Under xdist, one worker's cleanup removed files another
worker was still serving, so tests failed at random with a 404. The folder name
now includes the process ID.
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 commands carry the external spool as -1 in the flat ams_mapping
(the firmware rejects 254/255 there) and the real target only in
ams_mapping2. We captured only the flat list, so external-spool prints
looked unmapped and the usage tracker's position-based fallback
charged them to the first loaded AMS tray.
- Resolve external spools from ams_mapping2 when capturing a
project_file (dual-nozzle keeps 254/255, single-nozzle -> 254)
- Tracker: an explicit -1 no longer falls back to a positional tray;
a mapping naming no tray for any used slot defers to tray_now
- Keep the #1822 H2S tray_now override working with resolved mappings
An app opened from the sidebar signs in inside Bambuddy's iframe, but every
page sent frame-ancestors 'none', so the browser refused to show
/connect/authorize there. That path now gets 'self', like the streaming
overlay: every ancestor must be Bambuddy itself, so foreign pages still
cannot frame it.
Minimal OAuth 2.0 authorization-code flow with PKCE (S256): admins register
an app with one exact callback URL (Settings > API Keys > Connected Apps);
/connect/authorize asks for consent once and returns a single-use, 60 s code
bound to app, callback and challenge; POST /api/v1/connect/token swaps it,
with the client secret, for the user's identity and permissions. Codes and
secrets stored hashed, exchanges rate-limited per client and IP, no redirect
before the callback is validated, API keys cannot authorize, refused while
auth is disabled. i18n for all 15 locales.
-----
fix(db): upgrading from 0.2.4.0 or older no longer crashes at startup
The #2974 failure-reason conversion ran before the #1378 migration that adds
print_log_entries.failure_reason, so older databases stopped with "no such
column: failure_reason". It now skips a table without the column, only runs
where a legacy label exists, and on SQLite rebuilds archive_fts first, since
archives created before that index existed trip "database disk image is
malformed" when updated.
POST /queue/batches accepts external_source + external_ref; both are
returned on every batch and filterable on GET /queue/batches. The pair is
unique (index uq_print_batches_external), so a retried create answers 409
instead of queueing the same order twice. Migration covers SQLite and
PostgreSQL.
The support bundle gives each printer's connection diagnostic 15 s and
discarded the whole result on overrun, recording only "timed_out". A
14-printer farm's bundle carried that marker for every printer and
nothing else, so it could not say which check was slow.
run_connection_diagnostic now keeps an optional progress dict current
(finished checks + the step in flight). On timeout the snapshot records
stalled_in, elapsed_s and the checks that completed.
The concurrent-dispatch tests went red on the Docker shard with
"expected all 6 printers to be uploaded to concurrently, but the
high-water mark was 4", on a scheduler that was dispatching all six
correctly.
peak == 6 is the claim that the sixth dispatch reaches its upload
before the first one finishes, and the dispatches do not arrive
together: each runs a preamble of database work first. So the
assertion was a race between that spread and a fixed 0.15 s sleep.
Measured here the spread is ~14 ms; on the runner it passed 150 ms.
Padding the sleep would only move the threshold and slow every test
that uses it.
_UploadRecorder(assemble=True) now holds each call until every
upload the pass launched has arrived, reading len(_inflight), which
is filled synchronously at launch and is therefore the batch size.
That states the property the peak assertions are about, directly and
with no time in it, and it ends sooner than the sleep it replaces -
the file runs 7.0s to 5.9s. _BATCH_DEADLINE_SECONDS bounds it so a
dispatch that really has gone serial fails on its peak assertion
instead of hanging to the pytest timeout.
test_check_queue_returns_without_awaiting_the_uploads keeps the
sleep, with the reason recorded next to it: it reads the rows while
the uploads are open, and an assembled batch releases as soon as it
is complete, which would let the dispatches flip those rows
mid-assertion.
The test engine also takes the application's own _set_sqlite_pragmas
listener rather than a copy, so the file is opened the way the
running system opens one - WAL, synchronous = NORMAL, a 15 s busy
timeout. SQLite's defaults spend an fsync per commit and lock the
whole file, which made a dispatch's preamble cost more than the
upload it precedes. That is what turned the previous commit's move
to a file-backed database into a visible failure.
Verified in both directions rather than by a green run: with two
printers' preambles delayed by a second, the old recorder reports
exactly the "high-water mark was 4" the runner saw and the shared
library test reports 3 == 4, while the assembling recorder passes
the same scenario. Without the induced skew, ruff is clean, the full
backend suite is green at 12224 passed, and ten consecutive runs of
the file pinned to one core show no failures.
The scheduler's concurrent-dispatch tests failed on CI and passed
locally, on a commit that touched nothing but a text file. Six tests
in tests/unit/test_scheduler_concurrent_dispatch.py went red with
every queue item logged as "Status set to 'printing'" and four of
them read back as "pending", alongside "cannot commit transaction -
SQL statements in progress" and a PrintArchive that could not be
refreshed.
The fixtures built their farm on sqlite+aiosqlite:///:memory:, and
SQLAlchemy backs an in-memory SQLite with a StaticPool: one DBAPI
connection handed to every session, with nothing keeping them apart.
That was harmless while check_queue awaited its uploads inline,
because only one session was ever live at a time. Under the
refillable upload pool the uploads run as concurrent background
tasks with a session each, so their transactions interleave on that
single connection - a sibling session's close() rolls back another's
flushed-but-uncommitted UPDATE, and a commit() landing while another
session still holds a cursor raises the commit error above. Whether
the interleaving lands badly comes down to core count and
interpreter version, which is why a 30-core box on 3.13 stayed green
and a 4-vCPU runner on 3.11 did not.
The four fixtures now put the database in the test's own tmp_path,
which gets an AsyncAdaptedQueuePool and a connection per session -
what the application itself runs with (_resolve_pool_kwargs in
backend/app/core/database.py: pool_size 20, max_overflow 200). So
the harness is being brought in line with production rather than
having its assertions relaxed; nothing in the scheduler changes,
and the concurrency under test was correct throughout.
Checked against the failure mode rather than against a green run: an
isolated repro of the same shape - six writer sessions and one
reader session closing mid-transaction - yields all-pending on the
in-memory engine and all-printing on a file. The new risk is real
SQLite write contention on one file, so the file was run twelve
times pinned to two cores with no failures and no lock errors.
tests/unit/test_scheduler_busy_reasons_3018.py carries the same
harness on an in-memory engine, but dispatches to a single printer,
so it has no second session to race; left as is.
The Print / Schedule dialog's filament mapping is where the colour a slice
asked for is compared against the colour actually loaded, and only the
left-hand side of that comparison had a swatch. The slot, and every slot in
its dropdown, was text -- and the text cannot be trusted: a slot's colour name
is resolved from the Color Catalog or, failing that, from hue, so a
third-party beige is announced as "Orange". A "Color mismatch" warning then
gives no way to tell a real mismatch from two names for the same hex without
opening the printer card in another tab, which on a farm swapping twenty or
thirty non-Bambu colours between machines is a check made many times a day.
Each slot now carries its colour and its hex, and the slot whose colour is
exactly the one the slice asked for is ticked. This works for a slot bound to
an inventory spool and for one configured through Configure Slot or on the
printer itself: the second kind has no inventory row behind it, and the
printer's own tray colour is then what draws. A bound spool contributes what a
tray record cannot -- SlotSpoolIdentity gains extra_colors and effect_type, so
a two-tone or glittery spool draws as itself rather than as its base colour.
The same treatment goes to the filament-override picker used for model-based
assignment. It is the same choice on the other dispatch path, and leaving it
text-only would have made one decision read two ways.
Both controls stop being <select>s to do it, because an <option> renders text
and nothing else. SlotPicker keeps what the select gave for free -- arrow,
Home/End, Enter and Escape keys, listbox semantics, and the border colouring
that encodes match, same-type-different-colour and not-loaded -- and is
portaled with position:fixed so it is not clipped by the dialog's own scroll
container, flipping above the row when there is no room below.
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.
-----
Give each test worker its own scratch folder in the #3025 and #3029 tests (issue #3025)
Both modules wrote into one shared folder under settings.base_dir and deleted
it after every test. Under xdist, one worker's cleanup removed files another
worker was still serving, so tests failed at random with a 404. The folder name
now includes the process ID.
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 commands carry the external spool as -1 in the flat ams_mapping
(the firmware rejects 254/255 there) and the real target only in
ams_mapping2. We captured only the flat list, so external-spool prints
looked unmapped and the usage tracker's position-based fallback
charged them to the first loaded AMS tray.
- Resolve external spools from ams_mapping2 when capturing a
project_file (dual-nozzle keeps 254/255, single-nozzle -> 254)
- Tracker: an explicit -1 no longer falls back to a positional tray;
a mapping naming no tray for any used slot defers to tray_now
- Keep the #1822 H2S tray_now override working with resolved mappings
fix(#1898): keep Telegram link previews, and keep the outcome prompt's failures its own
Four follow-ups to the post-print outcome confirmation merged in #3047.
Telegram: link previews were switched off for every message rather than
only for the outcome prompt, so a print_complete template carrying
{finish_photo_url} lost its photo preview whenever the photo was too
large to attach. _send_telegram now takes link_preview, and only the
prompt turns it off, as the Slack unfurl change already did.
Archives: Reset left the new Unconfirmed filter on, so the list stayed
narrowed and the button seemed to do nothing.
Print start: when the external-print check hit a failed statement, it
rolled back the caller's whole transaction, which expired the printer
and the just-created archive; on an async session the next read of
either raises, and the start notification, energy reading and timelapse
baseline were skipped. The check's reads now run in a savepoint, and a
failed flag write reloads the archive and printer after its rollback.
Print complete: a failed outcome prompt left the notification session
needing a rollback, so the per-user print email sent on it next failed
too. The dispatch now rolls back on failure.
An app opened from the sidebar signs in inside Bambuddy's iframe, but every
page sent frame-ancestors 'none', so the browser refused to show
/connect/authorize there. That path now gets 'self', like the streaming
overlay: every ancestor must be Bambuddy itself, so foreign pages still
cannot frame it.
Minimal OAuth 2.0 authorization-code flow with PKCE (S256): admins register
an app with one exact callback URL (Settings > API Keys > Connected Apps);
/connect/authorize asks for consent once and returns a single-use, 60 s code
bound to app, callback and challenge; POST /api/v1/connect/token swaps it,
with the client secret, for the user's identity and permissions. Codes and
secrets stored hashed, exchanges rate-limited per client and IP, no redirect
before the callback is validated, API keys cannot authorize, refused while
auth is disabled. i18n for all 15 locales.
-----
fix(db): upgrading from 0.2.4.0 or older no longer crashes at startup
The #2974 failure-reason conversion ran before the #1378 migration that adds
print_log_entries.failure_reason, so older databases stopped with "no such
column: failure_reason". It now skips a table without the column, only runs
where a legacy label exists, and on SQLite rebuilds archive_fts first, since
archives created before that index existed trip "database disk image is
malformed" when updated.
POST /queue/batches accepts external_source + external_ref; both are
returned on every batch and filterable on GET /queue/batches. The pair is
unique (index uq_print_batches_external), so a retried create answers 409
instead of queueing the same order twice. Migration covers SQLite and
PostgreSQL.
The support bundle gives each printer's connection diagnostic 15 s and
discarded the whole result on overrun, recording only "timed_out". A
14-printer farm's bundle carried that marker for every printer and
nothing else, so it could not say which check was slow.
run_connection_diagnostic now keeps an optional progress dict current
(finished checks + the step in flight). On timeout the snapshot records
stalled_in, elapsed_s and the checks that completed.
Follow-ups after merging PR #3128 (with the #2990 preview work):
- PDF preview failed on every browser without
Map.prototype.getOrInsertComputed (Chrome < 145, Firefox < 144,
older Safari): "This file cannot be previewed" for any PDF. Load
pdf.js's legacy build, which bundles the polyfills for page and
worker, and bundle the worker through Vite (?worker&url) so
build.target lowers its class static block for Safari 16.0-16.3.
The browser-baseline check only scanned .js output and never saw
the copied .mjs worker; it scans .mjs too.
- PDF grid thumbnails only existed after someone opened the preview.
Page one is now rendered server-side with PDFium (pypdfium2, new
dependency, prebuilt wheels for every shipped platform) on upload,
ZIP extraction and external-folder scans; Generate Thumbnails
backfills PDFs as well. Renders are serialised behind a lock
(PDFium is not thread-safe), run off the event loop, and scale
from the page size so a huge MediaBox cannot allocate a huge
bitmap. Unreadable PDFs land without a thumbnail and keep the
browser fallback.
- A large STEP file takes over a minute to mesh in the browser and
showed only a spinner. STEP loads now show "Converting STEP
model... N s" and a note that large files can take a minute or
more, in all 15 locales.
- Drop the two occt-import-js "externalized for browser
compatibility" build warnings (path/crypto are only required in
its Node branch); any other externalization still shows.
The concurrent-dispatch tests went red on the Docker shard with
"expected all 6 printers to be uploaded to concurrently, but the
high-water mark was 4", on a scheduler that was dispatching all six
correctly.
peak == 6 is the claim that the sixth dispatch reaches its upload
before the first one finishes, and the dispatches do not arrive
together: each runs a preamble of database work first. So the
assertion was a race between that spread and a fixed 0.15 s sleep.
Measured here the spread is ~14 ms; on the runner it passed 150 ms.
Padding the sleep would only move the threshold and slow every test
that uses it.
_UploadRecorder(assemble=True) now holds each call until every
upload the pass launched has arrived, reading len(_inflight), which
is filled synchronously at launch and is therefore the batch size.
That states the property the peak assertions are about, directly and
with no time in it, and it ends sooner than the sleep it replaces -
the file runs 7.0s to 5.9s. _BATCH_DEADLINE_SECONDS bounds it so a
dispatch that really has gone serial fails on its peak assertion
instead of hanging to the pytest timeout.
test_check_queue_returns_without_awaiting_the_uploads keeps the
sleep, with the reason recorded next to it: it reads the rows while
the uploads are open, and an assembled batch releases as soon as it
is complete, which would let the dispatches flip those rows
mid-assertion.
The test engine also takes the application's own _set_sqlite_pragmas
listener rather than a copy, so the file is opened the way the
running system opens one - WAL, synchronous = NORMAL, a 15 s busy
timeout. SQLite's defaults spend an fsync per commit and lock the
whole file, which made a dispatch's preamble cost more than the
upload it precedes. That is what turned the previous commit's move
to a file-backed database into a visible failure.
Verified in both directions rather than by a green run: with two
printers' preambles delayed by a second, the old recorder reports
exactly the "high-water mark was 4" the runner saw and the shared
library test reports 3 == 4, while the assembling recorder passes
the same scenario. Without the induced skew, ruff is clean, the full
backend suite is green at 12224 passed, and ten consecutive runs of
the file pinned to one core show no failures.
The scheduler's concurrent-dispatch tests failed on CI and passed
locally, on a commit that touched nothing but a text file. Six tests
in tests/unit/test_scheduler_concurrent_dispatch.py went red with
every queue item logged as "Status set to 'printing'" and four of
them read back as "pending", alongside "cannot commit transaction -
SQL statements in progress" and a PrintArchive that could not be
refreshed.
The fixtures built their farm on sqlite+aiosqlite:///:memory:, and
SQLAlchemy backs an in-memory SQLite with a StaticPool: one DBAPI
connection handed to every session, with nothing keeping them apart.
That was harmless while check_queue awaited its uploads inline,
because only one session was ever live at a time. Under the
refillable upload pool the uploads run as concurrent background
tasks with a session each, so their transactions interleave on that
single connection - a sibling session's close() rolls back another's
flushed-but-uncommitted UPDATE, and a commit() landing while another
session still holds a cursor raises the commit error above. Whether
the interleaving lands badly comes down to core count and
interpreter version, which is why a 30-core box on 3.13 stayed green
and a 4-vCPU runner on 3.11 did not.
The four fixtures now put the database in the test's own tmp_path,
which gets an AsyncAdaptedQueuePool and a connection per session -
what the application itself runs with (_resolve_pool_kwargs in
backend/app/core/database.py: pool_size 20, max_overflow 200). So
the harness is being brought in line with production rather than
having its assertions relaxed; nothing in the scheduler changes,
and the concurrency under test was correct throughout.
Checked against the failure mode rather than against a green run: an
isolated repro of the same shape - six writer sessions and one
reader session closing mid-transaction - yields all-pending on the
in-memory engine and all-printing on a file. The new risk is real
SQLite write contention on one file, so the file was run twelve
times pinned to two cores with no failures and no lock errors.
tests/unit/test_scheduler_busy_reasons_3018.py carries the same
harness on an in-memory engine, but dispatches to a single printer,
so it has no second session to race; left as is.
The Print / Schedule dialog's filament mapping is where the colour a slice
asked for is compared against the colour actually loaded, and only the
left-hand side of that comparison had a swatch. The slot, and every slot in
its dropdown, was text -- and the text cannot be trusted: a slot's colour name
is resolved from the Color Catalog or, failing that, from hue, so a
third-party beige is announced as "Orange". A "Color mismatch" warning then
gives no way to tell a real mismatch from two names for the same hex without
opening the printer card in another tab, which on a farm swapping twenty or
thirty non-Bambu colours between machines is a check made many times a day.
Each slot now carries its colour and its hex, and the slot whose colour is
exactly the one the slice asked for is ticked. This works for a slot bound to
an inventory spool and for one configured through Configure Slot or on the
printer itself: the second kind has no inventory row behind it, and the
printer's own tray colour is then what draws. A bound spool contributes what a
tray record cannot -- SlotSpoolIdentity gains extra_colors and effect_type, so
a two-tone or glittery spool draws as itself rather than as its base colour.
The same treatment goes to the filament-override picker used for model-based
assignment. It is the same choice on the other dispatch path, and leaving it
text-only would have made one decision read two ways.
Both controls stop being <select>s to do it, because an <option> renders text
and nothing else. SlotPicker keeps what the select gave for free -- arrow,
Home/End, Enter and Escape keys, listbox semantics, and the border colouring
that encodes match, same-type-different-colour and not-loaded -- and is
portaled with position:fixed so it is not clipped by the dialog's own scroll
container, flipping above the row when there is no room below.
maziggy