Skip to content

Releases: SciML/DiffEqBase.jl

Fix Juno Rendering Regression

06 May 12:57
Compare
Choose a tag to compare

This fixes the regression in the rendering of the solution types in Juno due to the AbstractArrayOfVector changes. Also a small refactor setting up DiffEqDiffTools.jl

Hotfix complex numbers initial dt

05 May 15:21
Compare
Choose a tag to compare

There was a bug introduced due to units handling that caused the unitless norm of complex numbers to be complex, causing an error in initial dt comparisons. This hotfix fixes that specific issue, and is being put out quickly since two issues immediately opened for it!

Test solution bug fix

04 May 05:36
Compare
Choose a tag to compare

Fixes DiffEqDevTools and DelayDiffEq's tests

ApproxFunDiffEq Ready

01 May 23:37
Compare
Choose a tag to compare

Movement of certain constant function declarations.

Steady State Solution Hotfix

30 Apr 05:22
Compare
Choose a tag to compare

Needed for SteadyStateDiffEq.jl tests to pass for its registration.

DAE Hotfix

30 Apr 03:19
Compare
Choose a tag to compare

Error in the type ordering for DAEs found by Sundial's tests. Tests still fail until OrdinaryDiffEq and StochasticDiffEq update, but this change is what's needed for Sundials tests to pass (locally checked)

Stable Release

30 Apr 00:57
Compare
Choose a tag to compare

This marks the end of the restructuring changes. This means it includes the change of the solution type to AbstractVectorOfArray, moves out the noise processes and uses the new noise process interface, adds the "refined" problem types, adds mass matrices, finalizes the callback interface, finalizes the linear/nonlinear interface, and throws in some general patches. This is marked as stable as these have been in the list of changes since the beginning, and now are complete. The only other known future changes are

  • linear operator types
  • problem conversions/promotions

The first is a new future related to PDEs and thus it's DifferentialEquations 3.0 material, and will be non-invasive to add. Problem conversions/promotions would be nice, but not a priority right now so maybe 3.0 material. Should be non-disruptive too when that finally comes. So this sounds like a good stable place to set 1.0.0!

API Overhaul

11 Apr 04:59
Compare
Choose a tag to compare

RODEs
NoiseProcess integrators
No test problems
Analytic definitions
EulerHeun and RKMil interpretation
noise_rate_prototype and noise_prototype
save_everystep
saveat number
save indices
save_start
saveat -> no everystep
Callbacks in problems

Function plotting, MonteCarloProblem, Parameter Interface

07 Apr 07:57
Compare
Choose a tag to compare

Retcodes and Callback Initialize

31 Mar 04:21
Compare
Choose a tag to compare
v0.13.0

implement retcodes