Dev Call Dec 11, 2017 2300 UTC

CraigElder · December 11, 2017, 11:00pm

SITL

Copter

Rover

removing the reversed check boxes in MP’s RC calibration screen for Rover at least:

Issues & PRs

*problem with create()

CraigElder · December 23, 2017, 9:20am

Attendee count (max): 20

Pre-agenda:

UTC23:08 Quick announcements

UTC23:10 - copter and Randy

Leonard continues to poke away at new loiter
Making some additional changes to add another parameter to allow lean angle to be set when vehicle is not moving
Lean angle acceleration vs velocity
New flight mode changes have been made
Another PR for renaming the flight modes
[=#0000ff]https://github.com/ArduPilot/ardupilot/pull/7389[/]
Might re-order the Copter.h file
Like Rover.h
Tridge will get around to do Plane at some stage
[=#0000ff]amilcar_personal[/]: What about sub ??
[10:18 AM] To [=#0000ff]Weekly devcall[/]: I am planning on doing Sub
[10:18 AM] (Channel) [=#0000ff]amilcar_personal[/]: Nice
[10:18 AM] To [=#0000ff]Weekly devcall[/]: (need to ask Jacob first!
[10:18 AM] To [=#0000ff]Weekly devcall[/]: )
Apologies to people whose PRs break
Peter will help rebasing things that break

UTC23:19 - Rover and Randy

UTC23:22 - partners

a little behind on partner’s renewals
He’s getting to them!
8 have renewed
Most non-renewals don’t respond
The one feedback from non-renewal is because the partner left the industry
Rob’s partner payment
Should have come from GSoC?
No, payment is to project
Funding for individuals that go to mentor summit
Which Jaime did receive
No payment to individual mentors
Need to make sure the GSoC payments got made
Need visibility on SPI accounts!
Accounts tend to lag be a few months
We had $53,000 back in August
Should have gone up significantly since then thanks to our partners
Some donated more than the minimum - thanks very much!
Hex
Lightware
We have voted to use money to get people down to the developer conference in February in Canberra
A very, very valuable thing

UTC23:35 - ::create call

Hard to reproduce
Autotest server reproduced it once
And Peter’s laptop is reliable on it
Peter’s initial diagnosis seems to be correct
Uninitialised values off the stack being copied into bss-variables
Tridge thinks it would affect STM32 as well
May or may not be getting away with it ATM
Tridge wants to discuss with Lucas
This stuff is supposed to help with static analysers
May need to do a large PR to revert the create stuff if we can’t nut out how to fix the problem
Large effort which got a lot of review
Didn’t understand the way ::create works
Not a compiler bug
Looks to just be flawed
We seem to be lucky usually that the stack is filled with zeroes
Not in copter release
Not in Plane release
In the pending Rover release
Not in Sub release

UTC23:35 - sounds for PX4

A set of sounds specific for ArduPilot
Merge Solo sounds?
Contribution of the month?
Ringtone definitions in Chibios
Tone editor in SkyViper
Can type in a new set of tones and hit play
[=#0000ff]https://jschall.github.io/[/]
Contribution-of-the-month-style award
Starting in January
Competition closes 31st January
Don’t touch
Battery failsafe
Arming tone
Other things that are supposed to be annoying
Luis to work on which ones
Can we make it easy for people to do their own?
If we make this too easy companies will start adding their own sounds
We lose branding
More happy customers!
[=#0000ff]Jacob[/]: and diagnosing things will be slightly more complicated.
[11:01 AM] (Channel) [=#0000ff]Jacob[/]: for someone outside of the company
[=#0000ff]https://github.com/ardupilot/MAVProxy/blob/master/MAVProxy/modules/mavproxy_misc.py#L223[/]
[=#0000ff]https://github.com/ArduPilot/ardupilot/blob/master/libraries/AP_Notify/ToneAlarm_PX4.cpp#L38[/]
[=#0000ff]http://mavlink.org/messages/common#PLAY_TUNE[/]
[=#0000ff]David_SpektreWorks[/]: Is there any safety worry if we are changing the arming sounds? This will have to be well documented
Watch out for copyright infringement!
Luis will do the legwork and get an email and announcements together

UTC00:05 - Votes / Bugmaster

Perceived conflict of interest
Funding committee should have handled the vote
Hindsight is 20/20
Do we actually have a funding committee?
Probably should have been tasked with creating the two votes rather than Francisco having to do it
Nudge on the funding committee to review proposals
Four members who nominated
https://docs.google.com/forms/d/e/1FAIpQLScDFStUkuYJXmJukn4aXSgewBjDeedCGHR2EPIB7tSpeVS8eg/viewform
Congratulations to Luis, Olivier, Pierre
Tridge has adjusted the mailing list membership
Now we have a funding committee
Nothing improper has been done here
It would just have been better to have someone else create the votes
There was a week were someone else could have stepped in and say someone else should have run the vote
(Channel) [=#0000ff]J[/]: I think the margin for the second part of the vote is too close to call. And we have at least one person against that did not vote
Should have had the vote open for the duration of the call
[11:22 AM] To [=#0000ff]Weekly devcall[/]: I didn’t raise the issue of the vote being closed because the results were in-line with what I would’ve voted for anyway.
The certainty to the ArduPilot project on outgoing expense is no way near as much of an issue as the change in income is to Francisco
We need to do better in future
And we’ve spent a lot of time on this devcall for this small
Randy to update Wiki
Votes should span a devcall
Funding committee to open funding votes
Votes should be on agenda for dev call

UTC00:30 - Resistors and compasses

UTC00:42 - Jani is working with Shenzhen manufacturers association now