Unit test failing in toast3: Pointing weights do not agree

[ickc]

Since around 5 days ago, I'm seeing Pointing weights do not agree in the toast3 branch.
(which should probably narrow down the changes to these 3 commits:
54132fb...9ecc50e).

One example is in
https://github.com/ickc/so-software-environment/actions/runs/8248173751/job/22557911328#step:7:1815,
which basically is complaining

Pointing weights do not agree: [ 1 0 -1] != [ 1 0 1]...

Any ideas why?
Thanks.

[tskisner]

Yep, I'm actively working on that since yesterday. I have been double-checking the HWP stokes weights (which were verified on real data) against what I get from the math document which describes our pointing model. Should have a fix today.