-
-
Notifications
You must be signed in to change notification settings - Fork 30
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Migrate to SciMLBase v2 #189
Conversation
One last test failure and it looks legitimate. |
@ChrisRackauckas I think it's a dependency catch-22. This pr requires SciMLBase v2.6, so I need to update the compat entry, but when I do I get the error
since LinearAlgebra is fixed to v0.0.0 in the sysimage of Julia v1.9.3, and SciMLBase v2.6 requires LinearAlgebra v1.6 to meet the new General Registry requirements. This registry change seems like a developer inconvenience. Am I missing something? |
@ChrisRackauckas this post mentions the issue if calling |
88cb21a
to
24fce57
Compare
The solution is that we need to just get rid of the subpackages and make them extension packages. Since this is a breaking change we might as well fold it into the other breaking changes here. This will make maintenance a lot easier since now we don't need to independently release. |
Okay that took a bit but 🎉 ! |
This pr updates Integrals.jl to the new interface in SciMLBase v2, and needs SciML/SciMLBase.jl#535 before merging.
It is a big rewrite since all of the calls to the algorithms needed to be rewritten to use the new interface correctly. And the deprecated interface still works! (There may be an edge case when the user specifies
nout=1
, but otherwise I checked all the old tests pass.)This pr also improves how the IntegralsCuba.jl and IntegralsCubature.jl packages call the C-libraries so that the integrand can use arrays of any shape.
The AD was also updated to the new interface, including a bugfix for nested forward-mode AD, and there are opportunities to extend this pr by adding user-provided jacobians to
IntegralFunction
andBatchIntegralFunction
.This pr may fix many of the open issues. I think it addresses the following:
#169
#186
#164
#67
#86
#76
#40
And possibly others.
This should get the main branch up and running again.