Closed avik-pal closed 3 months ago
Attention: Patch coverage is 61.29032%
with 12 lines
in your changes are missing coverage. Please review.
Project coverage is 30.03%. Comparing base (
3746cda
) to head (7ff95fc
). Report is 9 commits behind head on master.
Files | Patch % | Lines |
---|---|---|
src/problems/bvp_problems.jl | 68.42% | 6 Missing :warning: |
src/remake.jl | 50.00% | 3 Missing :warning: |
src/solutions/ode_solutions.jl | 33.33% | 2 Missing :warning: |
src/scimlfunctions.jl | 66.66% | 1 Missing :warning: |
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
I am happy with the current state of this. To summarize
nlls
trait to BVProblem -- we try to infer it but err on the side of caution and set it to Nothing if we can reliably figure it out. It is the job of the upstream library to figure out how to deal with this caseoriginal
to the ODESolution. I agree with @ChrisRackauckas that we probably don't want to unnecessarily copy the code into a BVPSolution since original might be useful for ODEs as well. resid
not for IVPs but that is probably not a good enough reason to copy so much code.
Checklist
Additional context
The main compilation cost for BVDiffEq is coming from solving NLLS problems. Having this allows us to have a type stable path for NLLS BVPs.