Skip to content

20220317 Dev Meeting

Hamish Willee edited this page Mar 16, 2022 · 6 revisions

Agenda

General:

Notes

  • common.xml: remove wip from AIS_VESSEL #1817

  • VTOL - define MAV_TYPE for belly-sitting non tiltrotor #1808

    • Hamish concerned about reserved values being given any meaning by anyone as it means they can't be changed.
    • Julian interprets this like Don - they are reserved for VTOL and it gives you some degree of future proofing. Non-breaking to show VTOL behaviour. Only consequence is to ground station, that has to manage change of ID.
    • James is between two extremes: not particularly harmful, but law of unintended consequences means that it would be better not to.
    • AO - hamish to update message and let Don know.
    • HW Done. Also created https://github.com/mavlink/mavlink/pull/1818
  • common: remove option for relative speed change #1812 - can the relative-speed param be removed from MAV_CMD_DO_CHANGE_SPEED

    • Yes, this would be a good cleanup.
    • AO: Hamish to merge.
  • Component information protocol - move into standard #1806 - translations/dynamic update.

    • Component message with type and url does not allow scope for fallback URL (simply not enough space to work with).
    • Agree that getting everything from generalinformation metatdata with multiple parsing is a pain, but it does address this case.
    • Beat approach generally approved with following comments:
      • Deprecate COMPONENT_INFORMATION and add (replace) it with <wip> COMPONENT_META or COMPONENT_METADATA. Message just has URL and CRC for component metadata.
      • Add the dynamic message (wip)
      • make other changes required in QGC, PX4.
    • Possibly update the general metadata schema
      • remove the CRC for translations.
      • document that CRC should only be supplied for static files
      • remove the component_information_basic data from metadata IFF @julianoes confirms thsi should be done (dev call thinks the separate message is useful, and doesn't mind the duplication - so this is a call for the implementers to decide)
  • Rename and document parameter encoding protocol bits (Associated docs update)

    • This went in, as did associated docs. Also created RFC for Tridge proposal.
    • AO Julian - PX4 to start emitting the flag.
  • FYI only Enums for FTP follow on

    • Approve. Somewhat ambivalent but good to encourage new contributors. Also several people indicated they would rather use these than the magic numbers.
  • Batteries with more cells -

    • Current approach is not scalable - we can't keep adding cells.
    • Need a new battery messages design. Can live in parallel for a while.
    • Cells largely just accumulated to give total voltage - not generally needed. Some people interested in the cells for fault detection - but they only need to indicate problem with a particular cell, so might not need all data. Might just need cumulative fault data or whatever.
    • AO Propose RFC for new battery. Get buy in from Michael du Breuil as good person affected by current design.

Follow ups

Attendees: JulianO, Hamish, JamesP, Beat.

Clone this wiki locally