Fix JET tests

[devmotion]

As preparation for #1195 (see #1195 (comment)).

JET had found two obvious errors (undefined T and incorrect function call missing a semicolon), unsafe function calls due to ReverseDiff (loaded in the test) pirating fill (could return a TrackedArray; fixed by using fill! with an explicitly constructed Array), and many potentially undefined variables in the Fminbox implementation. For now I just unrolled the first iteration, but possibly it could be refactored in some way.

[github-actions]

Benchmark Results (Julia vlts)

Time benchmarks

	master	863dc43...	master / 863dc43...
multivariate/solvers/first_order/AdaMax	0.487 ± 0.0058 ms	0.489 ± 0.0059 ms	0.997 ± 0.017
multivariate/solvers/first_order/Adam	0.486 ± 0.0058 ms	0.487 ± 0.006 ms	0.998 ± 0.017
multivariate/solvers/first_order/BFGS	0.26 ± 0.0056 ms	0.261 ± 0.0055 ms	1 ± 0.03
multivariate/solvers/first_order/ConjugateGradient	0.168 ± 0.0037 ms	0.167 ± 0.0033 ms	1 ± 0.03
multivariate/solvers/first_order/GradientDescent	1.54 ± 0.0082 ms	1.55 ± 0.008 ms	0.998 ± 0.0074
multivariate/solvers/first_order/LBFGS	0.228 ± 0.0053 ms	0.227 ± 0.0051 ms	1 ± 0.033
multivariate/solvers/first_order/MomentumGradientDescent	2.18 ± 0.013 ms	2.2 ± 0.013 ms	0.988 ± 0.0085
multivariate/solvers/first_order/NGMRES	0.438 ± 0.0078 ms	0.431 ± 0.0082 ms	1.02 ± 0.026
time_to_load	0.397 ± 0.0085 s	0.395 ± 0.0043 s	1 ± 0.024

Memory benchmarks

	master	863dc43...	master / 863dc43...
multivariate/solvers/first_order/AdaMax	0.34 k allocs: 7.16 kB	0.34 k allocs: 7.16 kB	1
multivariate/solvers/first_order/Adam	0.34 k allocs: 7.16 kB	0.34 k allocs: 7.16 kB	1
multivariate/solvers/first_order/BFGS	0.36 k allocs: 15.5 kB	0.36 k allocs: 15.5 kB	1
multivariate/solvers/first_order/ConjugateGradient	0.362 k allocs: 14.2 kB	0.362 k allocs: 14.2 kB	1
multivariate/solvers/first_order/GradientDescent	2.09 k allocs: 0.0759 MB	2.09 k allocs: 0.0759 MB	1
multivariate/solvers/first_order/LBFGS	0.341 k allocs: 14.7 kB	0.341 k allocs: 14.7 kB	1
multivariate/solvers/first_order/MomentumGradientDescent	2.45 k allocs: 0.0818 MB	2.45 k allocs: 0.0818 MB	1
multivariate/solvers/first_order/NGMRES	1.56 k allocs: 0.117 MB	1.56 k allocs: 0.117 MB	1
time_to_load	0.153 k allocs: 14.5 kB	0.153 k allocs: 14.5 kB	1

A plot of the benchmark results has been uploaded as an artifact at .

[codecov]

Codecov Report

❌ Patch coverage is 59.30233% with 35 lines in your changes missing coverage. Please review.
✅ Project coverage is 87.42%. Comparing base (67320a2) to head (863dc43).

Files with missing lines	Patch %	Lines
src/multivariate/solvers/constrained/fminbox.jl	58.33%	35 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1209      +/-   ##
==========================================
- Coverage   87.80%   87.42%   -0.38%     
==========================================
  Files          45       45              
  Lines        3517     3548      +31     
==========================================
+ Hits         3088     3102      +14     
- Misses        429      446      +17     

☔ 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.