ArduPilot EU Dev Call 2024-09-25

Attendees (max): 11

UTC0701
Peter: Approved

  • We need Tridge to merge this.

UTC0708
Andy: We could use -Os to shrink the binary.
P: Let’s try that.

  • MergeOnCIPass

UTC0711
Andy: Approved


UTC0713
George: Not relevant to the test, but this feature will cause reduction in gains, if there is a false-positive land detection mid-flight. We’ve had one or two of these.
Andy,P: That’s correct.


UTC0721
Andrew: This was caused by an alignment of 2-3 different bugs and it made the CANARD bug surface.
P: We could ask a user of dronecan ESCs to test.
Bob: I can do it.
A: We also need Michel to release Mission Planner in tandem.
P: Why 100us delay?
A: It’s just enough to work, without lowering the bandwidth significantly.

  • For H7 boards unfortunately we can fit only up to 128 CAN messages. This makes the bug more pronounced.

P: Testing on HW would be very sensible.


UTC0733
A: Future feature: We could add a param setbit, getbit in MAVProxy.
P: MergeOnCIPass


UTC0738
A: Looks good
P: We need Randy to approve.


UTC0743
A: Seems reasonable.
P: We need Randy to approve.


UTC0744
B: Other places were using allow_throttle_while_disarmed(). Brought it here as well.
A: There is a case of users having safety enabled but exclude the starter motor. This change would mean that they will get running throttle.
P: We should merge this and if someone complains, we add an OPTION bit.
A: MergeOnCIPass


UTC0752
P: Rhys will agree with Randy on where to place the new scripts.


UTC0803
A: I’m happy to approve it.
Andy: We’ll wait for Bill to approve, then merge.


UTC0808
P: Still failing tests.
Revo Mini grew by 500B.
Some work needed to be done still


UTC0814
P: Needs testing.


UTC0815
A: I need to talk to Sid about this when he’s available again.

  • The Invensense driver now does a lot more SPI transactions and this has increased the CPU load a lot.
  • Hopefully Sid and Andy will agree to something in the next few days.

UTC0820
P: There might be a context-related bug in the autotest.
B: I’ll take another look.


UTC0829
P: A bit unusre what the _rsem is for.
A: Probably to prevent scripting calls corrupt data.
P: Do we want the automatic scripting backend enable?

  • Do we need to explicitly create a new binding for the video stream information? Can’t the new MAVLink lua module provide something similar? This costs quite a bit of flash.
  • Also needs Randy’s approval.

UTC0848
MergeOnCIPass


UTC0849
A: The Pixhawk6X ccache test is failing, even though the change shouldn’t affect it.

  • The Pixhawk6X binary grows by > 500B, which is very suspicious.
    P: Perhaps -02 is leakiing to other places, potentially through the new header include?