v0.10.37

ForwardDiff v0.10.37

Diff since v0.10.36

Merged pull requests:

• Update how_it_works.md (#660) (@thomvet)
• Bump actions/checkout from 3 to 4 (#663) (@dependabot[bot])
• Update for Documenter.jl v1 (#683) (@hyrodium)
• Broaden promote_rule for Dual to include AbstractIrrational (#687) (@ElOceanografo)
• Restrict StaticArrays to compatible versions (#688) (@andreasnoack)
• Bump actions/cache from 3 to 4 (#692) (@dependabot[bot])
• Bump codecov/codecov-action from 3 to 4 (#693) (@dependabot[bot])
• Make isless, <, >, <=, and >= consistent with == and isequal (#695) (@devmotion)
• Bump julia-actions/setup-julia from 1 to 2 (#699) (@dependabot[bot])
• Add support for StaticArrays >= 1.7 (#703) (@devmotion)
• Remove redundant DEFAULT_CHUNK_THRESHOLD (#704) (@sjkelly)
• Stop precomputing chunk sizes (#708) (@KristofferC)
• Apply #656 to master (#711) (@devmotion)
• Backport bug and test fixes to release-0.10 to prepare for new 0.10 release (#712) (@devmotion)
• Fix method ambiguity for Base.TwicePrecision (#715) (@devmotion)

Closed issues:

• MethodError: no method matching getrf!(::CuArray{ForwardDiff.Dual{ForwardDiff.Tag{ODEFunction (#612)
• Broken external link (#649)
• Correctly forming nested dual numbers. (#671)
• ForwardDiff fails to compute correct derivative (#676)
• Error requiring Symbolics from Optimization (#685)
• promote_rule ambiguity with AbstractIrrational and ForwardDiff.Dual (#686)
• Working with anonymous functions (#694)

v0.10.36

ForwardDiff v0.10.36

Diff since v0.10.35

Closed issues:

• Using a Pkg extension to reduce package load time caused by StaticArrays (#629)
• AD in-place instead of broadcast (#645)
• Is the mutating code the problem here? How could I debug this? (#646)

Merged pull requests:

• avoid defining one arg hash (#631) (@KristofferC)
• enable dependabot for GitHub actions (#635) (@ranocha)
• Fix two typos (#638) (@goggle)
• Bump codecov/codecov-action from 1 to 3 (#641) (@dependabot[bot])
• Bump actions/checkout from 2 to 3 (#642) (@dependabot[bot])
• Bump actions/cache from 1 to 3 (#643) (@dependabot[bot])
• Add test of second-order derivative of t -> abs(t)^2 (#650) (@devmotion)
• floatmin for nested Duals sharing the same tag (#656) (@chriselrod)
• example in docs (#657) (@cossio)

v0.10.35

ForwardDiff v0.10.35

Diff since v0.10.34

Closed issues:

• type instabilities when passing a function to jacobian! (#516)
• Highly invalidating method: promote_rule(::Type{R}, ::Type{ForwardDiff.Dual{T, V, N}}) where {R<:Real, T, V, N} (#620)
• The Hessian of the LinearAlgebra.jl dot is incorrect (#622)
• Jacobian returns zero matrix (#623)

Merged pull requests:

• Forbid offsets (#586) (@mcabbott)
• Add Invalidations.yml [ci skip] [skip tests] (#596) (@ranocha)
• Version 0.11 (#613) (@mcabbott)
• Support specializing on functions (#615) (@j-fu)
• Remove if VERSION ≥ v"1.4" etc. (#616) (@mcabbott)
• Fix minor typo in advanced.md (#621) (@thomvet)
• Turn StaticArrays into an Extension Package (#625) (@ChrisRackauckas)
• make StaticArrays dependency into an extension on v0.11-dev (#628) (@ChrisRackauckas)

v0.10.34

ForwardDiff v0.10.34

Diff since v0.10.33

Closed issues:

• Building ranges with Duals (#385)
• Automatic ChainRules compatibility (#579)
• Incorrect behavior of max(d::Dual, ::Float64) when partials are -Inf (#603)

Merged pull requests:

• Add link to ForwardDiffChainRules.jl (#608) (@gdalle)
• Backport #605 (dual exponent) to release-0.10 (#614) (@mcabbott)

v0.10.33

ForwardDiff v0.10.33

This release was yanked in JuliaRegistries/General#73304, i.e. ] up will not install it.

Changes here should be in 0.11, see #613

Diff since v0.10.32

Closed issues:

• support for the determinant function (#197)
• Derivative of matrix inverse for diagonal matrix is not correct? (#490)
• Incorrect gradient with cascaded reshape, linear equation solver operations (#506)
• Derivative lost in 3-arg mul (#536)
• Inconsistent behavior in assignment to sparse matrices of Dual numbers (#542)
• The Hessian of x-> dot(x,H,x) should be H+H', not zero (#551)
• ==(x,y) is true when only partials(y) have NaNs (#600)
• • Operator not defined for ForwardDiff.Dual{...} (#601)
• ldexp does not maintain type of Float32 arguments (#604)

Merged pull requests:

• Change == to ignore measure-zero branches (#481) (@mcabbott)
• Expand explanation of Preferences (#598) (@devmotion)
• dual exponent (#605) (@ptiede)

v0.10.32

ForwardDiff v0.10.32

Diff since v0.10.31

Merged pull requests:

• DEFAULT_CHUNK_THRESHOLD is configurable via Preferences.jl (#582) (@lassepe)
• check for iszero(partials(x)) in pow (#592) (@chriselrod)

v0.10.31

ForwardDiff v0.10.31

Diff since v0.10.30

Merged pull requests:

• Define SpecialFunctions.gamma_inc for ForwardDiff.Dual (#587) (@devmotion)
• Update README.md (#588) (@mcabbott)

v0.10.30

ForwardDiff v0.10.30

Diff since v0.10.29

Closed issues:

• Forward AD for mod2pi() (#113)
• Calculating gradient wrt to logpdf/loglikelihood of Gamma distributions (#283)
• realmin and realmax not defined (#306)
• derivative(xlogx, 0.) gives 0 (should be -Inf) (#365)
• Define prevfloat/nextfloat for Dual Numbers (#371)
• minor: link to upgrade guide on main README.md is broken, needs updating (#454)

Merged pull requests:

• Add definition for logabsgamma (#585) (@devmotion)

v0.10.29

ForwardDiff v0.10.29

Diff since v0.10.28

Closed issues:

• Spurious type promotions due to Float64 coefficients in the diff rules (#566)

Merged pull requests:

• Differentiation of eigen(::Symmetric) (#575) (@KnutAM)

v0.10.28

ForwardDiff v0.10.28

Diff since v0.10.27

Merged pull requests:

• Support complex output in derivative (#583) (@antoine-levitt)