mirror of
https://github.com/Rulez777/FramePicoTracker.git
synced 2026-10-17 12:01:16 +02:00
26 lines
2.0 KiB
Markdown
26 lines
2.0 KiB
Markdown
# Useful continuation tasks
|
|
|
|
## Reproducible offline setup
|
|
|
|
Resolve missing imports against the exported source, document Python and native dependency versions, and produce component build instructions. Do not assume the broad historical requirements are a lockfile. The native UI needs json-c, EGL and desktop OpenGL as well as its other listed libraries. Fetch third party code only into a separate development checkout with upstream notices intact.
|
|
|
|
## Source export configuration
|
|
|
|
Replace deliberately non-working private identity placeholders with an explicit recipient configuration mechanism that retains all comparison gates. Refactor the three blocked legacy C candidates only if needed for a concrete research question. Do not replace exact identity checks with wildcard admission. Preserve the frozen original210 provenance.
|
|
|
|
## Safe offline regressions
|
|
|
|
Port selected historical tests with genuinely synthetic or rights cleared fixtures. Explain which bug each test catches. Do not publish captured packets or calibration arrays as convenient test constants.
|
|
|
|
## Optical marker visibility
|
|
|
|
Read the latest probe result and the prepared longer exposure candidate. Establish why there are insufficient consistent markers before changing geometry or widening rejection guards. A device run needs a separately prepared and authorized bounded procedure, observed settings and a complete restoration record.
|
|
|
|
## Timing, identity and fusion
|
|
|
|
Use verified clock domains, exposure time and per boot mappings. Prove which tracker produced each observation before publication. Validate axes and bias from permitted measured inputs. Do not treat the packet reference's dongle measurements as live onboard Roy evidence.
|
|
|
|
## Body roles and controller coexistence
|
|
|
|
Attempt these after one tracker has stable identity and pose quality. Keep the radio owner exclusive and restoration verifiable. Measure the normal controllers and trackers together; a source implementation is not a coexistence result.
|