mirror of
https://github.com/openglow-org/forgefirm.git
synced 2026-10-16 01:29:55 -07:00
1e28eb92cd83e74feaa56ee8d7f1e6c13784bc5a
733
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 ScottW514
|
1e28eb92cd |
Catalog: setup.factory-return leaves
setup.factory-return leaves for forgectrl's tests/host_guard.sh (d21e200): POST /restore/factory-return with the token and no confirm is 400 and names confirm, confirm=0 is 400, GET /update/status shows no job running, GET /slots lists the archives (run only where /usr/sbin/ffboot is not on the host), and GET /setup carries the confirmed call; confirm=1 is never sent. The catalog goes from 98 to 97. It was the only test in the "the operator" section of suite/setup.py, and called no helper of its own. tests/test_setup_suite.py drops its id and test_the_factory_return_never_runs_the_return. src/update.* stays covered by the update tests, src/ui/wizard.js by the wizard and mirror tests, and src/main.c by the tests that already name it. test_setup_suite.py passes (31 tests). The coverage lint on the tree manifest finds no uncovered path and no host-tested row problem across the 97 tests; its one finding is setup.check-lens's empty src/lensfind.* entry. pyflakes is clean on both files. The full forgetest suite was not run. No HOST_TESTED row: the pinned workflow does not run host_guard.sh. Fingerprints: the deletion changes suite/setup.py's shared text, so the 34 tests that read it move again: the 9 left in setup.py, setup_dark 9, setup_burns 2, bedsize 1, exthost 9, extcall, extcat, extsender and extwizard 1 each. The pending pin bump from c274f19 already owes 32 of them; setup.machine-name and exthost.platform are the two it adds, as before. releases/ is unchanged. |
||
|
ScottW514
|
33e8c6e43f |
Catalog: logs.level-settings leaves
logs.level-settings leaves for forgectrl's tests/logs_test.c (9317b54): the six levels taken and verbose and 7 refused, port 70000 refused, udp and tcp taken and tls refused, a server name with spaces refused, and a remote level changed since boot listed as configured against effective with pending_reboot raised, and lowered once they match. The settings route's refusal of a value its validator rejects, with the file left byte for byte, is tests/host_routes.sh's bounds case. The catalog goes from 99 to 98. It was the last test in suite/logs.py and called no helper of its own, so nothing else goes. logs.tree-tail-export and logs.routing stay: they are the only catalog covers of forgectrl src/fflog.*, and the pinned workflow runs neither fflog_e2e.sh nor the new log tests. setup.record-export keeps its requirement on logs.tree-tail-export. The coverage lint on the tree manifest finds no uncovered path and no host-tested row problem across the 98 tests; its one finding is setup.check-lens's empty src/lensfind.* entry, which the forgectrl pin bump clears. pyflakes is clean on logs.py. The full forgetest suite was not run. Fingerprints: the deletion changes suite/logs.py's shared text (the two separator lines), so logs.tree-tail-export and logs.routing move; no other module imports it. The pending forgectrl pin bump from c274f19 already owes both, so this commit adds none to it. releases/ is unchanged. |
||
|
ScottW514
|
71556b2480 |
Catalog: the setup gate, the advisories, the extensions' consent and cloud-off leave
Four catalog tests leave for the host proofs that carry them; the catalog goes from 103 to 99. - setup.gate-blocks-controllers, setup.advisories-rehash, setup.extensions-consent and setup.cloud-disabled-surface. forgectrl's tests/host_setup.sh (91d30f1) runs the four HTTP proofs against the built daemon with the controller up: each reason the gate holds on and the live fan-count close and reopen; the first-run advisories, their hashes, and an acceptance that takes the press back; the extensions' advisory, its five refusals, and the consent; and every cloud selection refused while cloud mode is off, POST /mode included. builtin_test.c holds the cloud sweep, sha256_test.c the digest, and settings_dir_test.c the search bit the consent adds to the data directory. - gate_is_open and _GATE_COVERS go with the gate test, ADVISORY_DOCS and the hashlib import with the rehash; nothing else calls them. The record builders, REQUIRED_STEPS, SAFETY_PHRASE, Restore and request stay. - The requires that named them: _FIRST_RUN_REQUIRES held only the two deleted ids and goes, so setup.first-run-flow requires nothing from it and setup.first-run-page requires setup.first-run-flow alone; exthost.service keeps exthost.platform. - Coverage: HOST_TESTED rows for forgectrl src/sha256.* (tests/sha256_test.c) and src/builtin.* (tests/builtin_test.c), both run by forgectrl's build.yml at the pinned revision. docs/advisories/** stays behavioral: setup.first-run-flow, which accepts them, covers safety-and-risk.md, licenses.md, privacy.md and cloud-service.md, and exthost.service, which turns extensions on over it, covers extensions.md. - tests/test_setup_suite.py drops the four ids from the id list and the takeover list, the gate test's precheck assertion (the method now names the intake prechecks only), and CloudDisabledSurfaceTests, with the imports only that class used; the docstring no longer describes it. test_setup_suite.py passes (32 tests). The coverage lint on the tree manifest finds no uncovered path and no host-tested row problem across the 99 tests; its one finding is setup.check-lens's empty src/lensfind.* entry, which the forgectrl pin bump clears. pyflakes is clean on the four files. The full forgetest suite was not run. Fingerprints: deleting the four functions changes suite/setup.py's shared text, so 35 of the 99 tests' implementation hashes move: every remaining test in setup.py (10), and the modules that import it at module level, setup_dark (9), setup_burns (2), bedsize (1), exthost (9), extcall (1), extsender (1), extwizard (1), and extcat (1, through its import of exthost). The next forgectrl pin bump already owes 33 of them; exthost.platform and setup.machine-name are the two it adds. releases/ and TEST-AUDIT.md are unchanged. |
||
|
ScottW514
|
7172bba92b |
Catalog: the report channel, the event stream and what-changed leave
Three catalog tests leave for the host proofs that now carry them; the catalog goes from 106 to 103. - cooling.report-channel. forgectrl's tests/host_routes.sh (e69d967) runs the channel against a live daemon: a run report with the controller's secret writes exhaust 65535; no secret, a wrong secret and a short secret are 403 and the duty file is unchanged; an accepted idle while the ring still reads running writes 32768; the secret dies with the controller. cool_report_test.py and super_test case L own the secret, and cooling.fans-quiet-after-motion owns the live M8/M9 return to the idle duty. The forged Host: localhost was the same secret refusal. - events.stream. events_test.c owns the hello, the replacement, and motors.released; host_routes.sh owns the live cap (three peers held, a fourth 503, GET /settings 200 beside them, closing one admitting the next). forgectrl.lease and exthost.events still cover src/events.*, and _EventStream stays for forgectrl.lease. - setup.what-changed. setup_test holds tray -> laser.focus and fan -> airflow under the image's table beside tube -> laser.floor; host_routes.sh holds /wiz/changed (403, 400) and that every step the menu names is one GET /wiz lists. src/setup.* and src/wiz.* stay covered by the other setup tests, and help.js, panel.js and index.html by camera.key-read. preserved() and flag_of() go with it; recommended() stays. tests/test_setup_suite.py drops setup.what-changed from the id list, the takeover list and the covers lookup, and its flag_of case keeps only the recommended() assertion. test_setup_suite.py passes (36 tests). The coverage lint on the tree manifest finds no uncovered path across the 103 tests; its one finding is setup.check-lens's empty src/lensfind.* entry, which the forgectrl pin bump clears. pyflakes on the four files reports only the unused t0 in cooling.py that HEAD already has. The full forgetest suite was not run. Fingerprints: the cuts change the shared text of suite/cooling.py, suite/forgectrl.py and suite/setup.py, so 56 of the 103 tests' implementation hashes move (cooling 12, setup 26, forgectrl 5, and exthost 13 through its import of suite/forgectrl.py). The next forgectrl pin bump already owes 54 of them; exthost.platform and setup.machine-name are the two it adds. releases/ is unchanged. |
||
|
ScottW514
|
8e1fe5090d |
Harnesses: the null-sink cases other tests prove leave
The grblHAL harnesses keep what only a running controller shows; the cases a host unit test in grblHAL-glowforge proves, and the log and reply wording, leave. - laser_lifecycle_test: test_arm_waits_for_the_fans goes. laser_arm_test case A4 (grblHAL-glowforge) proves the fan grace stretching the arm's acknowledgment budget, on the same arm path. The docstring names it. - xy_mode_test: case_x32, case_invalid and the log checks go; the default, x16 and the low tick stay. xy_scale_test holds each mode's tick and the not-a-mode parse. - planner_blocks_test: one restart, at $398 1000; the 400 pass and the fresh-store probe go. - ctlport_test: test_poll_lf and test_crlf go (the CR LF poll stays); in test_sender_out the network turn-away goes (serial_test holds it), and the busy gating, the grant and the outlives-client part stay. test_port_dark is whole. - mcode_test: the timeout case runs on the host build's 5 s GFMCODE_WAIT_S (grblHAL-glowforge's CI configures it; the image keeps 30 s) and still shows the hold, at 5.1 s; 5 s outlasts the 3 s test_dark watches the wait; test_table refuses one form per parser branch. test_timeout_holds and test_dark stay. - envelope_test: test_unset and test_bounds go (lens_home_test holds the clamp); test_armed and the open, apply and close cases stay. - manual_home_test: the out-of-range offset block goes, a repeat of gfhome_clamp_home_mm (lens_home_test); the modal-M3 FIRE check and every release and energize case stay. - cool_report_test: one non-hex secret and one CR LF injection stay; runner-malformed and the exact header-set checks go. A bad secret is never sent and a good one is. - tray_test: test_offset keeps 20 mm; test_errors keeps a bad P and a bad axis; the frame shift, envelope, barrier, persistence, port op, unreferenced lens and camera home stay. Proof: with grblHAL-glowforge's matching commit, the eleven harnesses its CI runs pass under Linux against the null-sink build: nine on a build with a 1 s M-code wait, and mcode_test and tray_test, the two that wait at an M-code, on the 5 s build CI configures (-DGFMCODE_WAIT_S=5.0). pyflakes reports nothing in the nine changed files. Push this before grblHAL-glowforge's commit. Acceptance: none; host harnesses only. |
||
|
ScottW514
|
520d62ac05 |
forgetest tests: the confirmed trims of the tool's own unit tests
The cases that tested a mock, a timer, a wording or a table, or that
replayed a judgment a host test already holds, leave forgetest/tests.
What catches a false PASS, keeps the bench safe, or guards the gate
stays.
- test_baseline: the fake sysfs tree is a mixin with no test_ methods,
so ExtensionFreeTests and BaselineModeTests no longer re-run
BaselineTests' 25 cases (133 runs, now 45). CoolPublishTests, the
lamp-fade cases and TransientNotLeftoverTests go; PositionDeadbandTests
keeps a reading past the band and an unreadable reading, both never
excused. Both latch-relock cases stay.
- test_fixture: MdnsTests, the three timing cases, the probe's log
wording, and nine prompt-routing variants go. A live test never runs
unattended, the arm-press opt-in, the short-press edge, the presence
press handing over the arm press, and the arm notice dropping at the
press stay (35 cases, now 21).
- test_cloud_suite: three false-PASS guards stay: the latch open at a
dark print's button, a lid stop that is not edge-driven, the kernel
refusing the resume. The guards for the latch unlocking in the hold
and a paused print resuming left with their catalog tests in
|
||
|
ScottW514
|
8536ae7076 |
laser_stream_test: the analog copies, the in-run levels and the curve run leave
The analog dose model runs on no machine; the density sessions beside these prove the same rules on the shipped model. - Sessions J and K (M5 at idle, the next job's first cut) run under density only. - Session C (churn, analog) goes; its pad budget (no runaway pad across cycle churn) moves onto session G, density-churn, which already held FIRE, termination and the gaps on the same job. - Session H (a level change inside a run costs no byte) goes: every density power byte is already held at full duty by check_density; the byte count was a cost, not an energy rule. - Rule 19's end-to-end curve run goes: laser_arm_test.c case P pins the curve's mapping. The rolloff sessions keep the curve in force. - The constants only those sessions read go (CURVE_DEFAULT, curve_density_for, CURVE_S, JOB_CURVE, JOB_LEVELS); the header's rules 11, 16 and 19 say what the harness now holds. Proof: laser_stream_test passes whole on a null-sink build of the grblHAL-glowforge working tree under Linux (23 sessions). pyflakes is clean. Acceptance: none. The harness runs in grblHAL-glowforge's CI. |
||
|
ScottW514
|
09cf9242a9 |
laser_lifecycle_test: the sessions the unit tests and the stream prove leave
Each case below repeated a rule another test in CI already proves, read assertion by assertion next to it in the 2026-10-13 test audit: - test_job_window: laser_arm_test.c case I (S changes, and now M5 then M3, inside the window never re-prompt; df7087b) and laser_stream_test session K (M2 closes the window, the next job re-arms). - test_hold_grace: test_resume_after_grace makes the same assertion. - test_verdict_blocks_arm, test_arm_waits_for_engine_ack, test_arm_proceeds_on_a_late_ack: laser_arm_test.c cases C, A2, A3 and cooling_test.c's arm-time gate. test_arm_waits_for_the_fans stays. - test_sender_change: laser_arm_test.c case H; the disarm on a client change is unconditional. - test_jog_does_not_extend_window: laser_arm_test.c case S10. - test_verdict_fail_tier_ends_job: laser_stream_test rule 25 and laser_arm_test.c case V2. - test_verdict_pause_resumes_without_press: laser_stream_test rule 24, which now counts one arm across the pause, and case V1. - test_sender_change_mid_job: merged into test_sender_change_holds_and_rearms, which takes its "held inside the move" position check. - test_sender_change_then_reset: the window is closed and the latch locked by the sender change itself (case H); a reset from Hold is not energy. The header's rules 1, 2, 5, 11, 12 and 14 keep their numbers and name the tests that now prove them. Proof: laser_lifecycle_test passes whole on a null-sink build of the grblHAL-glowforge working tree under Linux (17 sessions), and laser_stream_test with it. pyflakes is clean; no helper was left without a caller. Acceptance: none. The harness runs in grblHAL-glowforge's CI. |
||
|
ScottW514
|
280ceee849 |
laser_stream_test: rule 24's pause holds the window open: one arm
Rule 24 proved the pause tier holds lit to the stop, masks FIRE while held, never writes the latch and resumes lit on the clean verdict. What only laser_lifecycle_test's verdict-pause session held was that the window stayed open through it: the resume armed nothing new. Rule 24 now counts the arms, so that session can leave. - verdict-rehold: exactly one "laser armed" across the pause, the re-hold and the resume. Proof: laser_stream_test passes whole on a null-sink build of the grblHAL-glowforge working tree under Linux. pyflakes is clean. Acceptance: none. The harness runs in grblHAL-glowforge's CI. |
||
|
ScottW514
|
8149a44f93 |
laser_stream_test: session B's FIRE-gap check moves to its density twin
Session B ran JOB_M3_TERM under the analog model; session F runs the same job under density and checked its span, density and termination, but not check_fire_gaps (no stepless run carrying FIRE), which B alone held. F now runs it, and B goes: analog never runs on a machine, and the analog ladder keeps its own termination and gap checks. - check_fire_gaps on density-full, its worst gap in the PASS line. - Session B's run and checks go; the JOB_M3_TERM comment names F. Proof: the harness passes whole on a null-sink build of the grblHAL-glowforge working tree under Linux, density-full with a worst fire gap of 270 ticks. pyflakes is clean. Acceptance: none. The harness runs in grblHAL-glowforge's CI. |
||
|
ScottW514
|
c91f6daa9e |
laser_stream_test: rule 14 runs under the density model
Rule 14 (a level set while the stream is idle still cuts at that level) ran only under the analog model, which no machine runs: its check counted fire ticks per analog duty. Under the shipped density model the level is the FIRE density at full duty, and nothing proved a stale level there. - Session I runs JOB_IDLE_S under DENSITY_CONF_FLOORED and judges it with rule 18's check_density: every power byte at full duty, and each move's span at its own S's density (23/127, 46/127, 81/127), not the one before it. The stream still ends dark. - The header's rule 14 says the level is judged as the FIRE density. Proof: the harness passes whole on a null-sink build of the grblHAL-glowforge working tree under Linux (densities 0.1812, 0.3624, 0.638 against 0.1811, 0.3622, 0.6378). A control judging the same stream against the levels in reverse order fails. pyflakes is clean. Acceptance: none. The harness runs in grblHAL-glowforge's CI. |
||
|
ScottW514
|
f8dc21b309 |
forgetest: motion.button-hold-resume leaves the catalog
motion.button-hold-resume held on the machine that the button holds a travel job and the next press resumes it to its target. The host proves it now: laser_lifecycle_test rule 7 runs the same job (M5, G91, G1 X40 F300) through the toggle and lands it within 0.05 mm of +40 ( |
||
|
ScottW514
|
a0a4edb322 |
laser_lifecycle_test: a travel job takes the button toggle and lands on its target
Rule 7 proved the button's pause/resume toggle on an armed job only, and nothing checked that the move resumed from the hold reaches its target. The catalog's motion.button-hold-resume held both on the machine with a travel job; it can leave the catalog once the host does. - test_button_pause_resume, after its armed half, runs the catalog's job: M5, G91, G1 X40 F300 (8 s). It must reach Run without asking for the arm press; a press holds it part-way; a second press resumes it to Idle within 0.05 mm of +40. - Rule 7's line in the header and the case's docstring say so. Proof: the case passes on a null-sink build of the grblHAL-glowforge working tree under Linux (the armed half as before; the travel job held at 8.747 mm and resumed to 39.999 mm). A control, the same job held and soft-reset instead of resumed, ends at 8.747 mm, which the landing check fails. pyflakes is clean. Acceptance: none. The harness runs in grblHAL-glowforge's CI. |
||
|
ScottW514
|
67ac8f4064 |
forgetest: events.button-telemetry leaves the catalog
events.button-telemetry held on the machine that one physical press reaches the event stream as pressed then released through its 25 Hz switch poll, and the telemetry event's fields. The host proves the rules now: - the press between two 5 Hz gathers, idle snapshot, scripted switch word: forgectrl events_test test_short_press (d8d07fc); - the telemetry field string, and that a press the machine is waiting for (arming, armed, leased, a button wait) is no event: events_test test_button. The physical button reaching the switch word stays in the catalog: setup.check-switches. - evmore.py, the test's own module, goes with its import in suite/__init__.py; nothing else imported it. - Its covers (forgectrl src/events.*, main.c, status.*) stay covered: events.stream holds all three, and other catalog tests hold main.c and status.*. No HOST_TESTED row is needed. Proof: the coverage lint reports 0 uncovered paths and 0 host-tested row problems over 107 tests (its one failure is setup.check-lens's src/lensfind.* entry, which the forgectrl pin bump clears); the gate self-check refuses an empty artifact cleanly. The unit modules that load the whole catalog pass under Linux: test_artifact, test_campaign, test_cooling_order, test_fixture, test_manifest, test_operator. Acceptance: the catalog hash moves; no remaining test's fingerprint moves (the module goes whole). |
||
|
ScottW514
|
0350b943d7 |
forgetest: laser.py's unused motion imports, and the two helpers no test calls
laser.py imported wait_state, wait_state_text, wait_left_state and
drain_text from motion.py and used none of them. wait_state and
drain_text stay in motion.py, which uses them; wait_state_text and
wait_left_state had no caller anywhere, so they go.
Proof: pyflakes no longer reports laser.py's four unused imports and
adds no finding. The coverage lint reports 0 uncovered paths and 0
host-tested row problems over 108 tests (its one failure is
setup.check-lens's src/lensfind.* entry, which the forgectrl pin bump
clears); the gate self-check refuses an empty artifact cleanly; the
forgetest host suite passes under Linux (539 tests).
Acceptance: the shared text of laser.py and motion.py moves, and with
motion.py's the tests of every module that imports it (motion, laser,
tray, updlock, and bedsize through homeoff). Against the image on the
bench that adds no run:
|
||
|
ScottW514
|
35add7cad0 |
forgetest: laser.power-floor leaves the catalog; laser.py's dead names go
laser.power-floor read $$ off the controller and checked $35 is the floor, that it lands at PWMSAR 12 or above, and that $31 is 0. The host proves all of it now: laser_stream_test rule 18 reads $$ on a fresh controller and after the arm ( |
||
|
ScottW514
|
d7379d9fc0 |
laser_stream_test: rule 18 reads $31 and the floor off the controller
Rule 18 proved a typed $35 is overwritten at the arm and the ladder
renders through the floor key, but took the shipped settings on faith:
the harness mirrors $31 = 0 and the board's 10 % floor as constants. The
catalog's laser.power-floor read them off the machine; it can leave the
catalog once the host does.
- The floor-derived session reads $$ on the fresh controller (no floor
key: the shipped default) and again after the arm overwrote the typed
$35. Both reads must show $31 = 0 (the floor, not $31, sets the bottom
of the S range) and $35 = the board's floor, and that floor must land
at PWMSAR_FLOOR_MIN (12) or above.
- A set floor key is laser_arm_test.c case J's (12.5 honored, an
out-of-range key falling back to 10, 0 honored). forgectrl keeps no
default for laser_floor_density: an absent key is the controller's
own floor, so there is no second default to compare.
Proof: laser_stream_test passes whole on a null-sink build of the
grblHAL-glowforge working tree under Linux ("$31=0 and $35=10 (PWMSAR
12) fresh and after the arm"). A control session with a floor key of 5
reads back $35 = 5 both times through the same parse, which rule 18
would fail. pyflakes is clean.
Acceptance: none. The harness runs in grblHAL-glowforge's CI.
|
||
|
ScottW514
|
21f53e4e83 |
forgetest: laser.verdict-cut leaves the catalog; the stream rules hold its claim
laser.verdict-cut held, on the machine, that a stale cooling verdict
mid-cut holds the job dark under the open window with the latch
untouched, and that the clean verdict resumes it lit with no press.
Every part of that is a rule the host now proves:
- dark while held: laser_stream_test rules 24 and 27, check_hold_dark on
every verdict hold's stationary stretch, and check_fire_gaps (
|
||
|
ScottW514
|
ef15c55577 |
laser_stream_test: the verdict's holds ship no FIRE while held
Rules 24 (the pause tier, held and held again) and 27 (a stale verdict) asserted the deceleration into each hold ran lit to the stop, the stretch between the holds dark, and the cut lit after the clear, but not that the stream is dark inside a hold: the central claim of the catalog's laser.verdict-cut, which can leave the catalog once the host proves it. Rule 21 proved it only for a plain feed hold. - check_hold_dark: no FIRE tick inside a hold's stationary stretch, its edges aside (one step pulse straddles each edge: rule 21's margin, HOLD_EDGE_TICKS, 120 ticks at x8 scaled to the mode: 480 at the x32 default). - Rule 24 checks both of its holds; rule 27 checks its one hold. - check_fire_gaps runs on both sessions, as on rule 21's. - The header's rules 24 and 27 say so; their PASS lines report the held stretches. Proof: laser_stream_test passes whole on a null-sink build of the grblHAL-glowforge working tree under Linux: verdict-rehold dark while held for 22620 and 22622 ticks, verdict-stale for 22621, check_fire_gaps clean on both. A negative control on check_hold_dark: a dark 5000-tick hold passes, and one FIRE tick inside it fails. pyflakes is clean. Acceptance: none. The harness runs in grblHAL-glowforge's CI. |
||
|
ScottW514
|
571f702113 |
forgetest: the emission witness samples a rapid after its M5; m5-rapid-dark leaves
The rule that a rapid after an M5 ships no FIRE is the host's (laser stream rule 16). What only the machine shows is LASER_ON staying off with FIRE clear while the chain can fire; kernel.fire-line's dark sample runs with the chain unarmed. laser.m5-rapid-dark held that sample in a test of its own, with its own arm and press. The emission witness, which every campaign runs, now holds it. - The square ends at M5 with no program end, so the window stays open. - The dark point: the LASER_ON count reads 0 after the M5, the window still armed, and the SoC lock line released (laser_locked false: the engine's armed flag can read true after the latch has locked, and a dark reading with the latch locked is the unarmed case). All three are checked before anything else is sent. - Only then is the rapid sent: G0 X20, a 2.5 s dwell, G0 X-20, a 0.5 s dwell, the program end. Sending it after the dark point keeps the mid-square dwell's own zero from being taken for the dark point. - The test samples until the window closes (a dwell reports Idle, so Idle cannot end it), with the controller's machine X beside each sample. - Every sample from the dark point until the window closes reads no emission, and the HV current stays at its idle reading (HV_DARK_MAX); the reports show the head out at +20 mm and back at the start inside that stretch, and the latch reads released on the sample at +20 mm. - The square still has to show emission and an HV rise, judged on the samples before the M5; the beam detector is judged there too. - The job-based disarm is timed from the head being back: under 10 s, where the idle grace is 60 s. - The rapid's list and its sampler live inside the test, and the deletion takes only m5-rapid-dark's own lines (decorator through body), so laser.py's shared text is byte-for-byte unchanged. - live_fire_drills.py's m5dark note and the bench README say the emission witness now carries the drill's check. Proof: the fingerprint parts of laser.py before and after: the shared text unchanged, every other test's own text unchanged. In the controller, glowforge_laser.c keeps the latch unlocked from the job's first laser-on through M5/M3 toggles and relocks it at program end or after the laser_disarm_s spindle-off grace (60 s by default), so the dark point and the rapid fall inside the window. The coverage lint reports 0 uncovered paths over 110 tests (its one failure is setup.check-lens's src/lensfind.* entry, which the forgectrl pin bump clears); pyflakes adds no finding; the forgetest host suite passes under Linux. Acceptance: only laser.emission-witness moves, and as an always-required test it must pass on the bench in the next campaign before a release can inherit its PASS. |
||
|
ScottW514
|
6a434bbf77 |
forgetest: nineteen tests a host test already proves leave the catalog
The catalog is for what needs the machine or an operator. Each test below asserted software rules that a host test in CI already proves, read assertion by assertion next to that host test before it left; the record of that reading is kept with the audit. Safety rules (each approved for deletion by the owner on that reading): - laser.disarm-in-hold: laser_lifecycle_test test_resume_after_grace. - laser.arm-wait-lid: laser_arm_test E-G, laser_lifecycle_test test_lid_open_in_wait. - laser.port-dark: ctlport_test test_port_dark (port jogs, a long jog and its cancel, four injected words, the panel ops: no FIRE tick under an open window with M3 modal). - motion.respawn-gate: super_test A and D. - motion.interlock-cancel-home: the lifecycle loop cases, laser_arm_test G. - motion.lid-policy-hold: laser_lifecycle_test test_lid_policy_hold. - cloud.lid-during-button-wait, cloud.dark-print, cloud.verdict-refuse, cloud.paused-lid-cancel: test_machine_lid_button.py. - cooling.fan-duty-readback: cool_fanwrite_test. - cooling.intake-one-fan: cool_intake_test A-E. - exthost.hold-pause-tier: holdkeep_test, run_test.py, cool_flow_test P, holds_test. Other rules: - homing.cloud-offsets: manual_home_test [own-offsets], test_gfhome_motion; the real session stays in cloud.mode-switch. - exthost.core-range: version_test, install_test. - exthost.operator-destinations: run_test.py, netrules_test, install_test, extpkg_test. - exthost.lifecycle: run_test.py's ext.shutdown read before the stop. - exthost.mcode: the null-sink mcode_test, forgectrl mcode_test, sdk_test, install_test. - motion.soft-limits: z_envelope_test's homed cases, which now assert no step before the alarm, and envelope_test. - cooling_intake.py, extcore.py, extdest.py, extlife.py and extmcode.py go whole; homeoff.py keeps only the camera-home helpers that cloud.mode-switch, setup.check-envelope and their meta-tests use. - The helpers only the removed tests used go with them. - exthost.armed-freeze still runs the cloud suite's dark-print helper, whose latch checks still run; its description says so. It now requires cooling.floor-and-warm-up, the dark print's own prerequisite, in place of the removed cloud.dark-print. - HOST_TESTED: forgectrl src/holds.* (holds_test) and src/mcode.* (mcode_test), whose only catalog home left with these tests. - The meta-tests that replayed the removed tests go: test_cooling_intake_suite.py, the cloud replay cases and their fixture. The one that catches a false PASS stays and runs the dark print as armed-freeze does: a latch open at the button fails. - Text that named a removed test (exthost.motion-job's description and comment, test_baseline's docstring) no longer does. Proof: the coverage lint on a manifest from the recipe pins reports 0 uncovered paths and 0 host-tested row problems over 111 tests (its one failure is setup.check-lens's src/lensfind.* entry, which the forgectrl pin bump clears); pyflakes adds no finding; the gate self-check refuses an empty artifact cleanly; the forgetest host suite passes under Linux (547 tests); the catalog loads with 111 tests after armed-freeze's description edit. Acceptance: the catalog hash moves, so the next image opens a new campaign. Every remaining test in laser.py, motion.py, cloud.py, cooling.py, exthost.py, homeoff.py and bedsize.py moves its fingerprint (their modules' shared text changed); exthost.armed-freeze moves by its requires. |
||
|
ScottW514
|
bef1bf9902 |
z_envelope_test: a move refused past the bed ships no step
The homed cases asserted that a move past X max, Y max or X min draws ALARM:2, and the harness's own description said "before any motion", but nothing checked the motion. The catalog's motion.soft-limits held that half on the machine, by the kernel counters; it leaves the catalog with this harness as its proof. - Every session mirrors the shipped stream (GFSINK_DUMP) into its work directory; steps() counts the tick bytes with an X, Y or Z step bit, as ctlport_test.py does. - Each refused move is judged by the step count across it: unchanged, or the case fails with the ticks that shipped. - The move inside the bed must ship steps: the control that shows the count sees motion at all. Proof: z_envelope_test passes on a null-sink build of the grblHAL-glowforge working tree under Linux, every case including the three refusals before any motion and the control. Acceptance: none. The harness runs in grblHAL-glowforge's CI. |
||
|
ScottW514
|
7246707df5 |
forgetest: the gate test leaves the step lists to forgectrl's setup_test
setup.gate-blocks-controllers asserted on the machine which steps the image's table requires (GET /wiz's missing list) and recommends (each not run yet, never the gate's parts or the optional steps). Both are the setup table's decisions over a record, software with no hardware in them, and forgectrl's setup_test now proves both on the image's own table: the required steps in order (10b), the missing list emptying when they are done (4b), and the recommended steps in order, each not run yet (10b). The catalog keeps what needs the machine: the supervisor holding and spawning the controller, the process, and the Grbl port. - The two missing-list checks and the recommended-list check leave the test's body; its description says where they are proven. - RECOMMENDED_STEPS and NEVER_RECOMMENDED go with them (no test read them any more), and the meta-test case that only compared them. Proof: the forgetest host suite passes under Linux (test_setup_suite with the case removed among it); the coverage lint is unchanged. Acceptance: every setup.py test (the constants were the module's shared text), setup.gate-blocks-controllers among them. |
||
|
ScottW514
|
ed1322518b |
forgetest: the lint takes python -m unittest given a test's file or module
forgectrl's CI runs its dev-server mock test as `python3 -B -m unittest -v tests/test_devserver_mock.py`, and the workflow judge counted python -m unittest only as `unittest discover`: a HOST_TESTED row for that test was refused as not run. _unittest_names: an argument to python -m unittest (not discover) runs the test when it is the test's file (resolved against the step's working-directory) or a dotted name whose module is that file (tests.test_x, tests.test_x.Case.test_one); a -k pattern's value and an argument with a shell variable are not names. Proof: host tests: the CI's own command line, a module, a module with a class and a method; not a module beside it, nor a -k pattern naming it. The manifest tests (34) pass, and the forgetest host suite passes under Linux. Against forgectrl's pinned workflow a row for tests/test_devserver_mock.py is accepted; every earlier verdict on the real workflows is unchanged. Acceptance: none. forgetest is dev-only and in no fingerprint; the lint runs in CI, not in a campaign. |
||
|
ScottW514
|
8003d6f18c |
forgetest: the burns cover only what the burns run
BURN_COVERS named src/accel.* and src/wizcalc.*, carried over from the sheet test the burns replaced. The burns' path is wizlive.c, then wiz_lens_rehome() and lenshome_run(); neither file is on it, and the focus burn holds the lens window as found rather than measuring stops. An accelerometer or dark-wizard change re-required both burns for nothing. Both files stay covered: src/accel.* by the cooling tests, src/wizcalc.* by the setup checks that run it (lens travel among them). Proof: the coverage lint on a manifest from the recipe pins reports 0 uncovered paths, as before. test_setup_burns, test_manifest, test_campaign and test_artifact pass under Linux, and the forgetest host suite passes under Linux. Acceptance: setup.focus and setup.floor (BURN_COVERS is setup_burns.py's shared text, so both fingerprints move; the catalog hash changes with the covers). |
||
|
ScottW514
|
a9c794be34 |
forgetest: a path a host test proves needs no catalog test
The coverage lint failed CI on any behavioral path no catalog test covered, so every software change pulled a new or wider acceptance test in with it, though the catalog is for what needs the machine or an operator. A path whose behavior a host test in CI proves now has a row in HOST_TESTED (manifest.py): the component, a glob, the host test, and the CI workflow that runs it, the last two as "<repository>:<path>" - a file of a component in the manifest (its tests/ and .github/ files are listed there), or "forgefirm:" and a path in this repository (the null-sink harnesses, which grblhal-glowforge's workflow runs). - coverage_report counts a host-tested path as covered. Fingerprints are unchanged: a path no catalog test covers is in no fingerprint, so a change there makes no acceptance run necessary. - host_row_problems reports a row whose glob selects no behavioral file; a host test or workflow that does not exist (a component file missing from the manifest, a "forgefirm:" file missing from the checkout, the dev-only forgetest, no repository named); a workflow that should be readable and is not; one that does not check out the test's repository; and one with no step that runs the test. - The workflow is read at the pinned revision: read_source takes a component's blob from the bare repo manifest-from-tree.py fetched for its source URL (the checkout's .manifest-cache, or --sources), a "forgefirm:" file from the checkout. With no sources (an image) it is not read and the lint says so; with sources, a read that fails is a problem, never a pass. - workflow_steps reads the workflow's steps by indentation (uses, run, working-directory, with), comments dropped, whole lines and from " #"; a quoted scalar loses its quotes, a quote inside a command stays. checkout_path finds where the test's repository is checked out (its own: the checkout naming no repository; another: the checkout of <owner>/<name>, the name from the component's source URL). - workflow_runs looks at what a step's run commands execute, split into simple commands (;, &&, ||, |, newlines, then/do/else; a backslash continuation joined), with sudo and its flags, env, time, exec and variable assignments passed over, each path resolved against the step's working-directory and compared with the test's workspace path: the command itself (a script run directly, a C test's built binary <dir>/<stem>); the script an interpreter runs (python, bash, sh) or a for loop's list, by path or a glob matching the whole path segment by segment; unittest discover starting in the test's own directory with a pattern its name matches (test*.py unless -p); pytest given the test or its own directory, for a name it collects. An argument to anything else (echo, cat, a build target), a step's name, a comment, a trigger's paths filter, or a path that only ends like the test's does not count. - The lint prints the row problems and fails --enforce on them, as on an uncovered path or an empty covers entry; --json carries them as host_tested, with workflows_read (false when a workflow that should have been read was not). It reads HOST_TESTED when it runs. - The table starts empty: rows come with the tests that leave the catalog. Proof: host tests: a host-tested row takes its paths out of the report and moves no fingerprint; good rows report nothing; each kind of bad row is reported with its reason; a workflow that runs one harness and names another only in its paths filter, a comment line and an inline comment runs the first and not the second; workflow_runs on a built binary, sudo and a binary, a script, a harness under another checkout, pytest on a directory, a for-loop glob and unittest discover under a step's working-directory, and not under another one; and not on a paths entry, an inline comment, a glob or a discover that the path only ends like (tools/tests/..., scripts/bench/tests/...), a longer name, another file, a step's name, or a commented line in a block; not on echo with the stem or the binary's path, a build target, cat of a script, unittest discover from a directory above the test or with a pattern its name misses, or pytest on a directory above it; on discover with -p '*_test.py', pytest given the file, a script run directly and by bash, and sudo -E with assignments continued onto the next line; checkout_path for its own and another repository and none; repo_name from a source URL; a component's workflow read from a bare repo's blob, judged, not judged without the sources, and a blob that cannot be read reported; the command line with a workflow it cannot read reports workflows_read false and the read failure, and fails --enforce. The manifest tests (34) pass on Windows and under Linux; the forgetest host suite passes under Linux. Against the recipe pins and their fetched workflows: rows for forgectrl's tray_test, forgeext's holdkeep_test, sandbox_test (under sudo) and run_test.py (sudo, env and assignments over three lines), the laser_stream_test harness (grblhal-glowforge's checkout of forgefirm), gfhardware's test_machine_lid_button (a for-loop glob under its working-directory), gfutilities' test_frame (pytest on tests) and forgetest's own test_manifest (unittest discover under forgefirm/forgetest) pass; rows for forgectrl's fflog_e2e.sh, the bench script fire_test.py and kas/source-bundle.yml (only in forgetest-ci's paths filter) are refused. The lint on that manifest reports 0 uncovered paths and 0 host-tested row problems. Acceptance: none. forgetest is dev-only and in no fingerprint; the lint runs in CI, not in a campaign. |
||
|
ScottW514
|
1191389a2b |
forgetest: no press is ever held; the arm notice goes down at the press
- The presence press at the Ready gate polled forgectrl's /status over HTTPS every 0.25 s plus the request's own time, so a short press fell between two reads and the operator held the button for seconds. It is read as an edge from the switches' input events (hw.ButtonEdges, no grab), opened before the notice goes up: a press of any length counts. Events read and not yet looked at are kept, so a short press whose release arrives in the same read gives the release to the next wait. With no switches device (the host harnesses) the status is polled at 50 ms. - An operator's arm notice (arm_press, press_now) stood until the test's next step, over a press already taken. A watcher takes it down the moment the controller puts the button out at the press; a notice put up meanwhile is left alone. Proof: host tests: a press and release written back to back on a FIFO standing in for the switches' device pass the Ready gate in well under a second with the status never showing the button down (the first version dropped the release with the press and waited 10 s: the test caught it); the arm notice stays while the button is dark or lit and goes down when it goes out, and a notice put up meanwhile stays. The fixture (35), setup (87), manifest (26) and operator (22) host tests pass. Acceptance: these are the harness's own paths; every live and operator test with a press runs through them. |
||
|
ScottW514
|
4c5dde2549 |
forgetest: the burns, the lens and the intake checks follow forgectrl's review fixes
- baseline: the lid lamp fades to a new level (the kernel's smooth trigger), so the hand-back reads it until it holds the wanted level twice in a row, for up to 3 s, before it is a leftover. motion.deadman failed its hand-back reading 232 against 236 0.05 s after the supervisor re-applied 236 to a respawned controller. - setup.check-lens and setup.focus: grblHAL reads the lens window and the edge only at its start, and their restore came after the wizard had restarted the controller on the test's values. Each test now waits for the step to end, puts the settings back, and restarts the controller (stop, start, running and verified with a new pid) on every exit path; cloud mode restarts nothing. - setup.focus and setup.floor: the served program ends with G0 Z<edge>, G0 X0 Y0, G92.1, M2; setup.floor requires setup.focus and checks the scrap's thickness against the lens reach; the burns cover lenshome. - setup.check-intakes: the two assertions that could not fail are gone; an intake-count prompt fails the test unanswered (the bench reference's intakes both turn). The precheck's refusal names no setting. - setup.check-airflow: the machine's count rules (intake 2 turning writes 2; silent on a one-fan machine, the count is not written). Shared text: setup_dark.py's precheck string (all setup_dark tests, setup_burns, setup.check-envelope), setup_burns.py's imports, covers, docstring and preview check, cooling_intake.py's docstring; every test they move covers paths forgectrl changed in the same work. Proof: the forgetest host suite passes (548), with new host tests for the lamp's fade, the restarts on pass, fail and abort, the floor's Focus refusal and reach, the tail without G92.1, the intakes prompt, and the airflow count rules; pyflakes is clean on the changed modules. Acceptance: these are the tests. |
||
|
ScottW514
|
005ff8008b |
forgetest: the coolant flow calibration is a required step; the gate's missing list
forgectrl requires cooling.flow after preferences, machine, and cloud, and GET /wiz lists the required steps the gate waits for as missing. - REQUIRED_STEPS is the form steps plus cooling.flow, in the gate's order; REQUIRED_FORM_STEPS are the form steps. cooling.flow leaves RECOMMENDED_STEPS. - setup.gate-blocks-controllers expects "setup required: preferences, machine, cloud, cooling.flow" on a record with no step done, with all four named in missing, and an empty missing list once they are done and the gate is open. - setup.first-run-flow walks the form steps alone: its seed keeps the checks done, the calibration among them, so the gate's reason narrows from the three form steps to open. Proof: the setup host tests (74) and the manifest tests pass; the host test of record_with_required_steps keeps the four. Acceptance: these are the tests. |
||
|
ScottW514
|
748d28686d |
forgetest: the setup burns need two witnesses and record the beam sensor
setup.focus and setup.floor follow forgectrl: a burn counts on the tube current and the kernel's LASER_ON samples, and its result records the beam sensor's rise and beam_sensor_seen, never requiring a rise (a head may give no output from the sensor). The check reads that both are recorded and logs them; the descriptions name the two witnesses. Proof: the setup host tests (74) and the manifest tests pass, two of them new: a burn whose sensor gave no output passes, and a result that does not record the sensor fails, each with the settings put back. Acceptance: these are the tests. |
||
|
ScottW514
|
ecbfe64cef |
forgetest: update.slots-and-signature reads the apply job's log lines
The foreign-signature drill's refused apply now also checks the forgectrl log, from the size it had before the apply: "update: writing the uploaded archive into slot <x>" and "update: the apply failed: archive is not signed". The edit sits inside the test's own function. Proof: the manifest tests pass and the module imports clean; on the bench reference with the logging forgectrl, the same drill by hand logged both lines. Acceptance: this is the test; it covers src/update.c already. |
||
|
ScottW514
|
a729197a8f |
forgetest: the first-run flow checks the machine facts see the head
setup.first-run-flow, after the machine step, checks that GET /wiz's machine facts read the head as present with a hardware id. The facts had read every head as not detected (forgectrl looked for "hw_id=" where the module prints "id="); the bench reference always has a head, so the check holds on every campaign. Proof: the setup host tests and the manifest tests pass; on the bench reference with the fixed forgectrl, GET /wiz reads the head as present with hw_id 044c. Acceptance: this is the test; it covers src/wiz.* already. |
||
|
ScottW514
|
ed52e752ee |
forgetest: the setup required up to the intake fans; the burns replace the sheet
The catalog follows forgectrl's setup (openglow-org/forgefirm#4): the gate is the advisories with the press, the account, the required steps (preferences, machine, cloud), and the intake fan count; every other step is recommended; the setup sheet is retired for two self-contained burns and a dark lens travel check; the override file is gone. - setup.gate-blocks-controllers installs four records: without consent, without the account, and without the three required steps the controller is gated with the reason naming the cause; with them done the gate opens, the controller runs, and the other steps are recommended as not run. Clearing cool_intake_fans closes the gate live with its reason, and putting it back as found reopens it. - setup.override-until-reboot is deleted with its helpers, and forgetest.init no longer writes the override file at boot. - setup.first-run-flow follows the reason from the three required steps to open; setup.first-run-page walks the choice screen after them; setup.what-changed reads the one-level map with the gate unmoved. - New setup.check-intakes (the count by the tachs equals the machine's, the setting put back as found) and setup.check-lens (the window written equals the stops found or the fallback, and the settings and /status hold it; both keys restored). A precheck refuses the intake and airflow checks while cool_intake_fans is unset, so neither can leave it unset in the middle of a campaign. - setup_burns.py replaces setup_sheet.py: setup.focus and setup.floor answer the gantry stops, the home, the scrap and the thickness, and keep the preview, emission and window checks; motion.tray starts the focus burn for its tray-out refusal. Proof: the forgetest host suite (529 before the required-steps edits, and the full suite after them) and the manifest load pass; pyflakes reports no undefined names. The coverage lint is clean against forgectrl's working tree; against the current pin, setup.check-lens's src/lensfind.* glob selects no file until the forgectrl pin moves to the commit that adds it. Acceptance: these are the tests themselves; the campaign runs them with cool_intake_fans set on the bench. |
||
|
ScottW514
|
dfd32d5fd1 |
Release v0.0.8
FORGEFIRM_RELEASE is 0.0.8, and releases/v0.0.8 carries the release notes and the acceptance artifact that forgetest exported on the bench reference. The artifact is from campaign c-20261010224522-439d on the pins-only dev image 20261010224118: authorized, 128 of 128 tests satisfied, 48 run on that image and 80 inherited. scripts/acceptance-gate.py against that build's release manifest gives AUTHORIZED, with inputs identical to the bench image.v0.0.8 |
||
|
ScottW514
|
35221de791 |
forgetest: every machine action a test takes is declared
The queues route a test by its kind and its declared actions: an auto test runs unattended with or without the bench fixture, and an operator test whose actions the fixture covers runs unattended only while it does. exthost.events opens and closes the lid through Context.act but was declared auto with no actions, so the unattended queue ran it with no fixture on the network, and it waited 180 s for a person who was not there: the fx29 campaign (image 20261010213702) ended on "lid open was not seen on the machine within 180 s" with forgefixture.local silent to mDNS and 172.16.1.171 silent to ping. - exthost.events: operator, actions lid, with the operator's steps. - events.button-telemetry: the same defect with the button; operator, actions button, with the operator's steps. - setup.check-switches: declares interlock beside lid and button; the check works the interlock loop on a Pro. test_fixture's new CatalogTests case reads every catalog test's source and fails a test that calls Context.act on a channel it does not declare, or that is auto and calls it at all. It found these three and no other. Fingerprints: the three decorators are each test's own text, so these three move and no other (diffed across the change, 128 tests). Proof: the whole host suite passes (509 tests) and pyflakes finds no undefined name. RoutingTests already prove the routing these declarations feed: an operator test the fixture covers moves to the unattended queue, one it does not stays attended, and with no fixture the routing is the old one. Acceptance: the change is the three tests' declarations; with the fixture up each runs unattended through it as before, and with the fixture down each stays in the attended list instead of stranding the unattended queue. |
||
|
ScottW514
|
46b6df1c2d |
Pin forgectrl 0.1.30 on the pushed head
- forgectrl c274f19 (0.1.29 -> 0.1.30): a machine built with one intake fan. cool_intake_fans 1 publishes the second intake as not fitted and never judges it, and the setup's airflow check writes the count it finds (openglow-org/forgefirm#3). Proof: bitbake -c fetch forgectrl from this pin succeeds (2 tasks). The tree manifest from the pins holds forgectrl at c274f19 (218 files), and the coverage lint over it finds no uncovered path and no empty entry (128 tests). Acceptance: cooling.intake-one-fan and setup.check-airflow gate the change, with every test whose coverage names src/cool.*, src/wizdark.*, src/wizcalc.*, src/main.c or src/ui/**. The forgectrl and forgetest trees of this pin passed cooling.fan-gate-trips, cooling.intake-one-fan and setup.check-airflow bind-mounted on image 20260928194539; the campaign runs on the image built from these pins. |
||
|
ScottW514
|
16a8ae9ba3 |
forgetest: cooling.intake-one-fan, and setup.check-airflow reads the intake fan count
forgectrl c274f19 adds cool_intake_fans: at 1, a machine built with one
intake fan, the second intake is published not fitted and never judged,
and the setup's airflow check writes the count it finds.
cooling.intake-one-fan, in a suite module of its own, proves both halves
on a machine with two intake fans with the setting alone. With
cool_intake_fans 1 and an 8 s grace, a session at the cut fan profile
must publish intake_2 as not fitted with no floor while it reads a real
speed, run OK, and judge intake_1 ok at the intake floor; with the intake
floor at its legal maximum it must trip AIRFLOW (hold, fire blocked) with
intake_1 TRIPPED and intake_2 still not fitted, and the fault must end
with the session; restored, every fan reads ok at its floor. It starts
only from a machine with both intakes judged, requires
cooling.fan-gate-trips, and puts the count, the intake floor and the
grace back as found on every exit path.
setup.check-airflow now reads the count from the result: 2 needs all
four fans at least 1000 rpm, 1 needs intake 2 reported {"fitted": false}
and the other three ok; either way the setting must read what the check
found, and the count is put back with the five floors.
Fingerprints: the new test has a module of its own, so no cooling test
moves; the setup.check-airflow edit is inside its function. Diffed across
the change: 127 -> 128 tests, one added, setup.check-airflow the only
one moved.
Proof: the touched host tests pass (test_cooling_intake_suite 4,
test_setup_dark 10, test_cooling_suite 25) and pyflakes is clean.
test_cooling_intake_suite replays the test against a scripted engine,
and two broken engines fail it with the settings put back: one that
ignores the setting, and one that takes intake 1 out with intake 2.
test_setup_dark's new AirflowCheckTests take a two-fan and a one-fan
result, and fail a count the setting does not match and a result with
no count. On the bench reference, image 20260928194539 (dev) with these
files and forgectrl c274f19 bind-mounted, cooling.fan-gate-trips,
cooling.intake-one-fan and setup.check-airflow pass with a clean
baseline before and after each (forgectrl's commit has the readings).
|
||
|
ScottW514
|
72907f311a | Update README | ||
|
ScottW514
|
3873013383 |
v0.0.7 release notes: the text as revised after publishing
The notes were rewritten after v0.0.7 was published: the extensions paragraph and every item under "Also new". The release page on GitHub now carries the same text, followed by its "Installing your own build" section. The v0.0.7 tag keeps the notes as first published. |
||
|
ScottW514
|
2ea501f975 |
Release v0.0.7
FORGEFIRM_RELEASE is 0.0.7, and releases/v0.0.7 carries the release notes and the acceptance artifact that forgetest exported on the bench reference. The artifact is from campaign c-20260929000044-e1ea on the pins-only dev image 20260928194539: authorized, 127 of 127 tests satisfied, 7 run on that image and 120 inherited. scripts/acceptance-gate.py against that build's release manifest gives AUTHORIZED, with inputs identical to the bench image.v0.0.7 |
||
|
ScottW514
|
1bc15fe70a |
forgetest: theme.css from forgectrl at its new pin
The pin to forgectrl 4b39663 brings theme.css's rules for the Extensions tab and a package's page in the panel's theme (94d8af5), which forgetest's shared copy did not have, so check-ui-vendor would fail forgetest-ci. The copy is forgectrl's file at 4b39663, byte for byte; the rules only add classes forgetest's page does not use. Proof: scripts/check-ui-vendor.py, fetching forgectrl's files from GitHub at the pinned revision, reports the shared UI files identical. |
||
|
ScottW514
|
66b39c6f2b |
Pin forgectrl 0.1.29, forgeext 0.1.1, grblHAL-glowforge 0.1.20 on the pushed heads, and meta-openglow in the kas lock
The work since the last pins, pushed to each component's default branch, is what the recipes now fetch: - forgectrl 4b39663 (0.1.28 -> 0.1.29): the manual home offset and the gantry stops by name in the panel, the Extensions tab and ui.background, the Grbl sender keep-out, the head camera for local viewers and live video on a package's page, the display units in the bridge's self answer, the crumb tray's switch and offset, the Lens card's fields saved, the listener's cipher order (ChaCha20-Poly1305 first), and the camera pipeline (coherent capture buffers on the GPU path, the two-pass NV12 render, the frame gate, the /cam/status timing, one switch fd). - forgeext 2d0391e (0.1.0 -> 0.1.1): ui.background, sender.keep_out and live video in the SDK, a package removed and installed again started from the new install, and the same ffx archive for the same input. - grblHAL-glowforge b34636f (0.1.19 -> 0.1.20): the manual home offset, the gantry stops by name, the controller port's sender out and in with the keep-out, and the crumb tray's M103 P0 and P1. - kas lock: meta-openglow f1f7f58 -> f012d8e, kernel TLS with the NEON crypto and the TX records in cache lines of their own (patch 0016). Proof: bitbake -c fetch of forgectrl, forgeext, grblhal-glowforge, linux-fslc and kernel-module-glowforge from these pins succeeds (10 tasks), with kas moving layers/meta-openglow to f012d8e. The tree manifest from these pins holds forgectrl at 4b39663 (217 files), forgeext at 2d0391e (90) and the driver at b34636f (177), and the coverage lint over it finds no uncovered path and no empty entry (127 tests). Acceptance: the same trees, through the externalsrc overlay and the mirrored meta-openglow at f012d8e, are image 20260928182434, on which all 127 tests are satisfied (the unattended campaign and the attended tests the operator ran). The meta-openglow move is a platform change; the image built from these pins alone is next. |
||
|
ScottW514
|
83686fc712 |
forgectrl.tls-records: a window another client connected in is measured again
The test judges each connection by the kernel's TLS counters and the crypto engine's interrupt, and both are the machine's, not the connection's. On image 20260928175056's campaign the test before it restarted forgectrl, the operator's panel (six HTTPS connections from the bench PC) reconnected in the same second as aes-1.3#1, and the kernel counted 7 sessions each way where the test wanted exactly 1; the connection itself chose AES-128-GCM, carried the page byte for byte, and counted no decrypt error. _fetch now also reads the machine's accepted TCP connections (PassiveOpens, shared by both families); a case whose window saw any connection but its own is measured again, up to 30 times after a pause drawn from 0.1 to 0.9 s (so no steady poller, such as a page polling this suite once a second, stays in step with it), and fails if every window was contested. The judgments are unchanged: exactly one session each way for a cipher the kernel seals, none for CBC, the engine's interrupt counting the AES records and not the ChaCha20 ones. Proof: pyflakes clean. On the bench reference, image 20260928175056, with an openssl client connecting to :443 every 0.25 s for the first 6 s of the run and the drill polling this suite once a second: the image's test fails (desktop-1.2 counted 2 sessions each way), the fixed test passes with contested windows measured again (at most 4 attempts a case), and passes again without the extra clients (at most 4). Acceptance: the change is forgectrl.tls-records itself; tlsrec.py holds no other test, so no other fingerprint moves. |
||
|
ScottW514
|
f47ddad839 |
exthost.platform: the frozen window opens once the freeze has landed
The test took its first reading of the frozen probe's cpu.stat the moment cgroup.events said frozen 1, and wanted 0 us over the next 1.5 s. The kernel reports a task frozen before its last way into schedule() is charged to the group, so a reading at the event can come before that charge lands. On image 20260928170058's campaign the frozen probe was charged 16 us and the test failed. The first reading now comes 0.2 s after the event; the window and its judgment (0 us) are unchanged. Proof: pyflakes clean. On the bench reference, a spinner under cpu.max 5% in a cgroup of its own, frozen 520 times with the first reading taken both ways: read at the event (polled every 50 ms, as the test does), the next 0.5 s was charged 4 times (13 to 20 us); read after the 0.2 s settle, never. Acceptance: the change is exthost.platform itself, inside its own function, so no other test's fingerprint moves. |
||
|
ScottW514
|
eb9e61d3c6 |
motion.soft-limits: past the envelope the home armed, and stopped before it fails
After a home the driver's X/Y envelope reaches to the far edges the bed check measured (envelope_x_mm, envelope_y_mm, held by gfhome_clamp_envelope_mm to 50 mm up to the travel plus 30 mm), and to $130 by $131 only on a machine that was never measured. The test aimed its refusals at $130 + 5 and $131 + 5, which holds only on an unmeasured machine. On the bench reference, measured at X 508 and Y 288.925, image 20260928153805's campaign sent G90 G1 X500 F600, the controller rightly ran it, and the failed check ended the test before its soft reset, so the head drove the whole 476.6 mm while the restores were refused with 409 for 40 s; the jog check after it would have run a 500 mm jog the same way. The test now takes the far edges as the driver does and sends each move and the jog 5 mm past them, and it stops a move that was not refused (the soft reset, now ahead of the checks) and a jog that was not refused (the jog cancel) before it fails. The judgment is unchanged: ALARM:2 with no kernel motion, error 15 for the jog, a move inside the bed runs. Proof: pyflakes clean. On the bench reference, image 20260928153805 with this suite bind-mounted, the test passes: the far edges 508 and 288.925, G90 G1 X513, Y293.9 and X-1 each refused with ALARM:2 and 0.0 mm of kernel motion, the jog refused with error 15, the move inside the bed run, the head handed back where it was found, the baseline clean. Acceptance: the change is motion.soft-limits itself; its own function and description carry it, so no other test's fingerprint moves. |
||
|
ScottW514
|
55e789bc34 |
camera.frame-health: five seconds of stream, none withheld, and its timing
forgectrl now publishes a stream frame only once the frame after it has come back clean, withholds both neighbors of a frame the receiver flagged, and reports health.withheld and a per-stage timing block in /cam/status. camera.frame-health keeps its snapshot burst and adds five seconds of MJPEG: at least 30 frames dequeued, none flagged, none withheld, a withheld count and a timing block present while the viewer watches, and capture end to publication above 0 and under 250 ms. Proof: the test fails on image 20260928000227's forgectrl, which reports neither the withheld count nor the timing block, and passes on forgectrl with the frame gate (75 frames, none withheld, 67.3 ms capture to publication, on the bench reference). The unit suite (500 tests), pyflakes, and the coverage lint (0 uncovered paths) pass on the host. |
||
|
ScottW514
|
11031ad87f |
GnuTLS with kernel TLS, and forgectrl.tls-records
The gnutls bbappend builds GnuTLS with --enable-ktls and installs /etc/gnutls/config with ktls = true, so after the handshake the kernel seals and opens forgectrl's HTTPS records (the kernel side is meta-openglow's CONFIG_TLS and patch 0016). It also backports GnuTLS dc016daf: 3.8.4 hands the kernel the record sequence number where a TLS 1.2 ChaCha20-Poly1305 connection's IV belongs, so every such connection failed the kernel's first decryption. forgectrl is the only program on the image that links GnuTLS. forgectrl.tls-records, in its own module (suite/tlsrec.py), reads the result from the outside with openssl s_client on loopback: a desktop offer (AES-GCM first) gets ChaCha20-Poly1305 over TLS 1.3 and 1.2 and the kernel takes both directions' keys; an AES-128-GCM-only offer, three times over each protocol, gets it, the kernel takes its keys, and the CAAM's job-ring interrupt counts the records; a TLS 1.2 CBC-only offer connects and stays in GnuTLS (the control for the kernel's counters); every copy of the panel page equals the plain-HTTP copy byte for byte, with no decrypt error; the kernel's drivers are rfc7539(chacha20-neon,poly1305-neon) and ctr-aes-caam with ghash-ce. Proof: on the image before these fixes the test failed on both bugs it names (the TLS 1.2 ChaCha20 page arrived empty with a decrypt error; the AES-GCM pages arrived damaged); on image 20260927224418 it passes. The unit suite (500 tests) and the coverage lint (0 uncovered paths) pass on the host. |
||
|
ScottW514
|
cd84786fc5 |
tray_test.py and motion.tray: the crumb tray's mode on the null-sink controller and on the machine
- scripts/bench/tray_test.py, run by the driver's CI: M103 P1 and P0 move Z and its envelope by tray_offset_mm on the lens step grid, tell the sender, write and remove the marker beside the shared config, and ship no step; a missing P, P2, P0.5, P-1, and an axis word are errors that change nothing; the switch behind a move waits for the move; the port's tray op with a sender connected, its count exact, refused out of form, during a program, and at a package's M-code; the mode across a soft reset and a killed controller, a start over the marker taking the lens reference in the tray-out frame; the offset's range; an unreferenced lens; a camera home in both modes handing the runner the same park. The kernel-idle half of the barrier needs the pulse device: a null sink is always idle. - scripts/bench/ctlport_test.py: the dark case carries tray out and in under an open armed window. - forgetest: motion.tray in its own module, suite/tray.py, so no other test's fingerprint moves. It switches the mode both ways through the panel's route and through M103 from a Grbl client and judges each switch: /status's mode, Z, and reach moved by the offset's grid, the port's state, the marker, the reference and its source unchanged, the client told, the kernel's counters unmoved. A jog to the Z the lens read before a switch is refused in the other mode and taken back in the mode found; a setup card started with the tray out is refused in the operator's words; the panel page carries the switch. The mode goes back as found in a finally. It does not restart the controller, which would drop the X and Y reference: persistence is tray_test.py's. - The bench registry and README list tray_test.py. Proof: tray_test.py ALL PASS against the driver with the mode, and six driver copies each missing one part of it fail. The driver's whole CI harness list passes. The forgetest unit suite: 500 tests OK. The coverage lint over the working trees: nothing uncovered, 126 tests. On the bench reference, with the binaries hot-deployed on the fx18 image: motion.tray PASS, and FAIL against the driver whose anchor lost its source word on the second switch. Acceptance. motion.tray is new; the suite and bench files are in no other test's fingerprint. |
||
|
ScottW514
|
6e71ac86f0 |
forgetest: exthost.sender-keep-out, camera.lid-privacy for local viewers, and ctlport_test's keep-out
- suite/extsender.py (new module, so no other test's fingerprint moves): exthost.sender-keep-out. A reference package whose service claims the Grbl sender out as it starts, installed with the sender.keep_out grant (and refused without it), on an idle GRBL-mode machine with no sender connected. It checks the holder and the ext: lease of the extension kind, the host's claim file fresh past forgectrl's grace, a sender from the machine's LAN address turned away with the message and one from the loopback admitted, the operator's release (lease gone, network welcomed, the package reading released), a new claim after a restart, and the stale end with the notice after a disable; everything is put back as found. - camera.lid-privacy: the head camera answers the panel's token and the extension host's header with the lid open, and refuses the cloud client's form (no token, no header) and a stream for no local viewer; the lid camera stays refused to everyone. covers gains src/auth.*. - scripts/bench/ctlport_test.py: test_sender_out on the null-sink controller build (Sender(host=) and closed, own_address(), generation()). Proof: forgetest's unit tests pass; the coverage lint passes against the working trees' file lists. Against the current pins it reports src/senderout.* as selecting nothing until forgectrl's pin moves to the revision that adds it, which lands with this. On the bench reference with the suite mounted: exthost.sender-keep-out PASS (first run found the check racing the host's first write; the test now waits for the file and holds past the grace), and camera.lid-privacy passed every check of its body (its hand-back failed only on the controller gated by the privacy advisory's Revision 5 before it was accepted). ctlport_test.py passes in the driver's CI. |
||
|
ScottW514
|
1727b819cb |
homing.manual and the harnesses: the gantry stops by name
The parts the manual home rests against are the gantry stops in homing.manual's title and comments, in manual_home_test.py's docstrings, and in a comment of the baseline's hand-back. No check changes. Acceptance: homing.manual's own text changed, inside its function, so no other test's fingerprint moves. manual_home_test.py passes against the driver. |