Upper bound diffeqbase and patch bandind error

[avik-pal]

We were relying on the AbstractArray{<:AbstractArray} for specifying initial conditions. This is the last version to support that. We will transition to a VectorOfArray from v6

[github-actions]

Benchmark Results

	master	45374f4...	t[master]/t[45374f4...]
Simple Pendulum/IIP/MultipleShooting(10, Tsit5; grid_coarsening = false)	1.8 ± 0.061 ms	1.83 ± 0.056 ms	0.98
Simple Pendulum/IIP/MultipleShooting(10, Tsit5; grid_coarsening = true)	3.31 ± 0.091 ms	3.36 ± 0.078 ms	0.986
Simple Pendulum/IIP/MultipleShooting(100, Tsit5; grid_coarsening = false)	0.052 ± 0.0033 s	0.0522 ± 0.0034 s	0.996
Simple Pendulum/IIP/MultipleShooting(100, Tsit5; grid_coarsening = true)	1.06 ± 0.0076 s	1.11 ± 0.0068 s	0.95
Simple Pendulum/IIP/Shooting(Tsit5())	0.177 ± 0.0034 ms	0.232 ± 0.0064 ms	0.763
Simple Pendulum/OOP/MultipleShooting(10, Tsit5; grid_coarsening = false)	4.12 ± 0.19 ms	4.15 ± 0.16 ms	0.993
Simple Pendulum/OOP/MultipleShooting(10, Tsit5; grid_coarsening = true)	7.25 ± 3.9 ms	7.32 ± 4 ms	0.99
Simple Pendulum/OOP/MultipleShooting(100, Tsit5; grid_coarsening = false)	0.127 ± 0.0088 s	0.128 ± 0.0022 s	0.999
Simple Pendulum/OOP/MultipleShooting(100, Tsit5; grid_coarsening = true)	4.49 ± 0.00045 s	4.73 ± 0.001 s	0.951
Simple Pendulum/OOP/Shooting(Tsit5())	0.697 ± 0.026 ms	0.792 ± 0.037 ms	0.88
time_to_load	8 ± 0.031 s	7.96 ± 0.032 s	1.01

Benchmark Plots

A plot of the benchmark results have been uploaded as an artifact to the workflow run for this PR.
Go to "Actions"->"Benchmark a pull request"->[the most recent run]->"Artifacts" (at the bottom).

[avik-pal]

Don't merge I will look into the NLLS, it is most likely LM that is causing all the trouble. LM seems to be robust but also works worse than TR in practice