86 lines
4.9 KiB
Markdown
86 lines
4.9 KiB
Markdown
# Step 10 fix: Final real native/MIDI package acceptance
|
|
|
|
## Objective and reviewed prerequisites
|
|
|
|
Read the latest plan continuation and CURRENT report_7_fix/report_8_fix/
|
|
report_9_fix/report_9_fix_controls. Real native codecs, target subset maps,
|
|
atomic native/MIDI MCP service and explicit CC requests are implemented and
|
|
reviewed. The previous MIDI-only closure was retracted; this is real native
|
|
software acceptance, not a fallback-only delivery.
|
|
|
|
Complete final tests/build/docs/evidence for the actual supported versions and
|
|
formats: Eclipse 4.0.1 `.griff`, Ray5 4.1.0 `.briff`, SD3 3.4.4/SL-DFH `.sd3p`,
|
|
and MIDI. `agp`/`sdp` remain accepted request aliases, not mislabeled file bytes.
|
|
|
|
No commits, delegation, later prompts, unrelated features, audio analysis or
|
|
new plugin host.
|
|
|
|
## Safety and application status
|
|
|
|
A live Cubase15 process with an unnamed project was observed (PID 41836 at the
|
|
check). Do NOT manipulate/close/save that user project, send global keyboard/
|
|
mouse input, or launch competing audio hosts/drivers for acceptance.
|
|
No software test may be falsely reported as actual audible or DAW import
|
|
acceptance. Record live plugin/Cubase checks as UNRUN if safe operator validation
|
|
is unavailable; distinguish that certification limit from implemented native
|
|
serialization and exact native content checks.
|
|
|
|
## Owned work
|
|
|
|
- Final directly necessary native/package/stdio regressions, if any gap remains.
|
|
- Fix only actual task-caused bugs with focused tests; do not weaken existing
|
|
assertions or add cosmetic abstractions.
|
|
- README, current architecture/profile/format/compatibility docs and plan status
|
|
for actual native/MIDI behavior; remove stale current blocked-layout/MIDI-only/
|
|
unregistered-writer claims, preserve clearly labeled historical reports.
|
|
- Update `prompts/report_10.md` with a new CURRENT native implementation/software
|
|
acceptance section; keep prior MIDI-only acceptance as labeled history inside
|
|
the same single plain triple-backtick fence.
|
|
|
|
## Required final acceptance
|
|
|
|
1. Run full current pytest, Ruff lint/format, strict mypy and pip check.
|
|
2. Build final wheel and sdist from the actual final source.
|
|
3. Run the EXACT persisted final wheel outside the checkout with real stdio,
|
|
not only editable source or an earlier build. Verify imported module and
|
|
JSON resources actually come from the installed noneditable wheel.
|
|
4. Use an original minimal test template for ordinary installed acceptance.
|
|
Generate default all-part four-artifact output, legacy-alias native-only,
|
|
actual-name native-only, MIDI-only, palm-muted lead/bass, CC1/native points
|
|
and explicit reset, missing-template fallback and disabled-fallback errors.
|
|
5. Independently reopen generated MIDI/XML/SD3 saved bytes and compare exact
|
|
requested musical notes, velocities, physical fingering, techniques,
|
|
tempo/meter/end, channels, controller events, artifact metadata and warnings.
|
|
6. Run genuine read-only local-template acceptance with ORIGINAL generated
|
|
music against:
|
|
C:\Users\g3\Documents\Toontrack\Superior Drummer 3\sd3-demo.sd3p
|
|
Preserve source bytes/identity/mtime/hash and all non-song instrument state.
|
|
Keep generated private-kit-state evidence ignored, outside source/tests/
|
|
distributions; never upload it or ship the template.
|
|
7. Verify failures/collisions/corruption do not publish partial sets or overwrite
|
|
existing output. Clean only owned test workspaces and child processes.
|
|
8. Confirm packages include six real profile resources and the actual codecs,
|
|
no private native templates/factory music/manuals/executables/credentials.
|
|
9. Validate README JSON examples against live typed schemas; document exact
|
|
current text vocabulary, technique/CC limits, native aliases, target defaults,
|
|
optional operator template requirement, setup/client configuration and
|
|
trusted-root/no-overwrite/fallback behavior accurately.
|
|
10. Record the final source/build hashes/versions, pass/fail/skip counts, actual
|
|
outputs and source preservation. No full-library/all-plugin-version/
|
|
arbitrary-technique/live-playback compatibility claim.
|
|
|
|
Additional locally verified Riffer facts, if useful for acceptance:
|
|
all 1,089 installed riff XMLs contain conventional Key values including Cm,
|
|
A#m and other major/minor spellings; 1,049 loop-enabled examples use loopEnd in
|
|
PPQ480 ticks (e.g. 4 bars 4/4 -> 7680). These are schema facts, not test music;
|
|
do not copy any factory sequence into fixtures.
|
|
|
|
## Required report
|
|
|
|
When complete, write your end-of-step report to ./prompts/report_10.md including: what was implemented, test results (pass/fail with details), any issues encountered, and whether you're ready for the next step
|
|
|
|
One plain triple-backtick fence for the entire report, current native acceptance
|
|
first and prior MIDI-only report labeled historical. State clearly which native
|
|
software outputs are implemented/verified and which live application checks were
|
|
not performed. Do not rewrite history as if the earlier interim was approved.
|