Oscar Smith

use Differentiation Interface

1

Very WIP, but this starts us on the move to DifferentiationInterface. Remaining TODO items: - [ ] FunctionWrapper support - [ ] resizing - [ ] "AutoBackend" (that uses the...

* [ ] Support autodiff=ADTypes * [ ] Deprecate autodiff=true * [ ] Fix/remove? FunctionWrappers * [ ] coloring/sparsity? * [ ] https://github.com/SciML/OrdinaryDiffEq.jl/pull/2482

redo of https://github.com/SciML/OrdinaryDiffEq.jl/pull/2463 since the branch conflicts got pretty nasty.

requires https://github.com/SciML/SciMLBase.jl/pull/800 and this isn't quite ready yet.

and add `NonAdaptiveController` which can be specified as an equivalent of `adaptive=false` but which does not replace it. @Theozeud can you take a look at this and see if it...

add `finite_difference_jvp`

2

Add the pushforward operation with implementation taken from jacobian but simplified. This is my first time touching this code, so let me know if there's anything obviously wrong here.

Make `add --weak` automatically create extension entries

6

It would be great if there were an easy way to add weakdeps from the repl rather than editing the Project.toml file.

Bad error message for missing end

1

Right now the error for missing an `end` in otherwise good code is ``` ERROR: LoadError: ParseError: # Error @ /home/oscardssmith/.julia/dev/OrdinaryDiffEq/src/derivative_utils.jl:766:5 end end # └ ── Expected `end` ``` The...

remove the lazy init and set isfresh-false on init

4

as discussed in https://github.com/SciML/LinearSolve.jl/pull/533

use LinearSolve precs setup

9

This updates NonlinearSolve to use the new `precs` interface in LinearSolve.