fix: enable support for complex ODEProblem again

[oameye]

Checklist

• Appropriate tests were added
• Any code changes were done in a way that does not break public API
• All documentation related to code changes were updated
• The new code follows the
  contributor guidelines, in particular the SciML Style Guide and
  COLPRAC.
• Any new documentation only uses public API

Additional context

resolves #3763

[oameye]

Is their a chance this change can be backported to v9? As of v9.73 complex ODEProblem solving is broken.

[ChrisRackauckas]

Yeah @AayushSabharwal could setup the backport.

[oameye]

Awesome :)

[AayushSabharwal]

This implementation will cause all tunables and constant parameters to be stored as complex values if anything is complex. I'm not sure if that's what we want.

[ChrisRackauckas]

I think that's right due to the promotion rules?

[AayushSabharwal]

I don't particularly mind it either way, just making sure we're on the same page. The operating point [x => 0.5im, par => 5] will cause par to promote to 5.0 + 0.0im.

[ChrisRackauckas]

Yes I think that's expected, because the derivative will also have non-trivial complex values, so at least the tunables needs to change to complex.

[ChrisRackauckas]

Can you do the backport?

[AayushSabharwal]

Sure

[oameye]

bump for the backport