make hessians of scaled interpolants work

[ajwheeler]

fixes #627

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 87.96%. Comparing base (ba5b9ec) to head (a625a34).
Report is 1 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #628      +/-   ##
==========================================
+ Coverage   87.74%   87.96%   +0.21%     
==========================================
  Files          29       29              
  Lines        1910     1911       +1     
==========================================
+ Hits         1676     1681       +5     
+ Misses        234      230       -4     

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[bclyons12]

This successfully fixes #627 , both in the example given and in my full code. Note that you may want to remove the rand or seed it in the test to avoid potentially irreproducible results in the future.

[ajwheeler]

Interestingly, this test fails for u = reshape(collect((1.0 : (N^2))), N, N). (edit: to be clear, I think this is numerical, and may not fall into the category of being a "bug").

[mkitti]

Interestingly, this test fails for u = reshape(collect((1.0 : (N^2))), N, N). (edit: to be clear, I think this is numerical, and may not fall into the category of being a "bug").

Would it pass with an increased tolerance for isapprox?

[mkitti]

Could we bump the version to 0.16.1?

[ajwheeler]

Sorry for the spam, but for anyone reading this in the future, u = reshape(collect((1.0 : (N^2))), N, N) has a condition number of roughly 1e19 for N = 10, so it's no surprise that the test fails for that input. (I should have realized this...)