Fix oscillating velocity for differential drive

[icros-personal]

Also attempt to add unit tests to exercise this code, but they are Windows-specific at the moment (needs fix!)

[mjansen4857]

I’m pretty sure that module acceleration being able to go negative is important to the trajectory generation. I am not 100% on that though. I’ll have to do some testing with this change.

EDIT: never mind I just realized it’s not. Swerve kinematics won’t give a negative. This change should probably be made to the forward accel pass as well. Plus the c++, python, and GUI versions.

Also, I think adding the ability to actually do PPLib unit tests is better to do in a separate PR.

[icros-personal]

I've removed the unit test and build logic out and tried to make comparable changes to C++ / Python / GUI code. I haven't tested them, however - I'm not sure how to build them, actually.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 84.53%. Comparing base (1027141) to head (6d2a646).
Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #1052   +/-   ##
=======================================
  Coverage   84.53%   84.53%           
=======================================
  Files          95       95           
  Lines        9730     9730           
=======================================
  Hits         8225     8225           
  Misses       1505     1505           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[mjansen4857]

Once the CI finishes you should be able to download and try the GUI build. That uses the same algorithm as PPLib so as long as that works I’m not too worried about it. The issue wouldn’t be noticeable in the GUI though since that still just uses swerve kinematics for diff drive.

[mjansen4857]

Have you run this locally or something to verify this is the fix for the issue? It sounds right to me but just want to make sure there isn’t more to it.

[mjansen4857]

To fix the failing c++ builds you need to change this:
units::math::abs(accelStates[m].speed())

to this:
units::math::abs(accelStates[m].speed)()

[icros-personal]

I've run it locally in Java to make sure it works there. I hadn't tried it in any other form. I hadn't previously tried it on the forward pass, but it was positive there so I don't expect it to change.

[icros-personal]

I tried downloading the Windows GUI build but couldn't figure out how to persuade it to let me install it while unsigned.
From PowerShell, using Add-AppPackage -AllowUnsigned, it reports: "Add-AppPackage : Deployment failed with HRESULT: 0x80073D2C, The package deployment failed because its publisher is
not in the unsigned namespace."

[mjansen4857]

You can’t install the msix version, that can only be uploaded to the Microsoft store. Just download the windows zip.

[icros-personal]

You can’t install the msix version, that can only be uploaded to the Microsoft store. Just download the windows zip.

Ah, I see. I missed that artifact. Yes, the GUI seems to work much as before, I don't really see any difference with or without (as you predicted). The Java version is fixed for me. I haven't tried either Python or C++.