NLopt.jl Versions Save

NLopt v1.0.2

Diff since v1.0.1

Merged pull requests:

• Rethrow errors instead of returning FORCED_STOP (#194) (@odow)
• Avoid exception handling for verifying algorithm validity (#212) (@ararslan)
• Prep for v1.0.2 (#213) (@odow)

Closed issues:

• Exception thrown by objective function is not rethrown by nlopt (#156)
• problem (#187)
• Linking to preinstalled nlopt binaries (#196)
• Call NLopt with vectors (#207)
• GPU Capabilities? (#209)

NLopt v1.0.1

Diff since v1.0.0

Merged pull requests:

• Fix MOI.get for MOI.ConstraintFunction (#208) (@odow)
• predeclare global used in extension (#211) (@palday)

Closed issues:

• memory allocation? (#205)
• In Julia 1.10, NLoptMathOptInterfaceExt errors on load (#210)

NLopt v1.0.0

Diff since v0.6.5

Closed issues:

• "nlopt failure" should be recoverable (#14)
• Ctrl-C exception in REPL (#68)
• Problems with LN_PRAXIS, GN_DIRECT and others... (#70)
• Initial step size documentation incorrect (#79)
• Signal a new starting point in nested optimization (#80)
• NLopt returns BoundsError even though model is properly defined as MathProgBase model (#101)
• MethodError for opt.initial_step (#132)
• issue with pmap: document? (#146)
• Add vector constraints (#163)
• LN_COBYLA steps out of bounds (#166)
• COBYLA reports solution out of bounds (#183)
• Quits if initial point does not satisfy nonlinear constraints (#185)
• Problem setting the Local Optimizer (#189)
• Maxtime and compilation (#190)
• ReInitCache() error (#197)
• Should support for MathOptInterface be moved to an extension? (#203)

Merged pull requests:

• Fix copying of Opt objects (#184) (@lcontento)
• Fix typo when adding vector of constraints (#191) (@odow)
• Fix return value of ListOfModelAttributesSet in MOI_wrapper.jl (#192) (@odow)
• Throw better error when querying initial_step (#193) (@odow)
• Add license headers (#195) (@odow)
• update CI badge (#199) (@stevengj)
• Move MathOptInterface support to extension (#204) (@palday)

NLopt v0.6.5

Diff since v0.6.4

Closed issues:

• algorithm termination by maxeval results in Exception (#13)
• Build failure on OSX when conflicting Homebrew taps exist (#57)
• @NLconstraint, JuMP (#65)
• Nelder-Mead termination (#86)
• with nlopt-2.4.2, algorithm_name() fails in Julia 0.6 (#89)
• precompile error (#92)
• Invalid NLopt arguments (#96)
• Segfault NLopt 0.3.6 in Julia 0.6.2 (Repeatable) (#99)
• NLopt Linear Inequality Constraints (#103)
• GN_ORIG_DIRECT segmentation fault (#104)
• NLopt ignoring stopval, tolerance (#107)
• Tolerance-Value dependence of Output (#124)
• Find multiple local minima? (#129)
• returns :XTOL_REACHED but doesn't modify the input values (#136)
• Pass gradient cost function in JuMP (#139)
• Can not run test example in Julia (#149)
• NLopt doesn't catch ForcedStop for some optimizers (#158)
• JuMP.value and JuMP.solution_summary are broken when using NLopt.jl solvers (#169)
• Run several optimizations in parallel (#180)
• Solve FORCED STOP issues (#181)

Merged pull requests:

• Remove trailing colon in function definitions in README.md (#172) (@pierre-haessig)
• Reenable copytest (#176) (@blegat)
• Simplify with VariablesContainer (#177) (@blegat)
• Remove useless line in test (#178) (@blegat)
• Switch from AppVeyor to Github actions (#179) (@blegat)
• Update support to MOI v1.0 (#182) (@odow)

NLopt v0.6.4

Diff since v0.6.3

Closed issues:

• Specifying subsidiary optimization algorithm for augumented Lagranian when using JuMP (#126)
• Unknown status FAILURE (#160)

Merged pull requests:

• Implement local_optimizer for MOI interface (#173) (@blegat)
• Handle missing result status in MOI wrapper (#174) (@blegat)
• Breaking changes for MOI 0.10.0 (#175) (@blegat)

NLopt v0.6.3

Diff since v0.6.2

Closed issues:

• remove_constraints!(opt) doesn't work; chkn() not defined (#133)
• Minimizers are Float32 instead of Float64 (#148)
• Dead link in README (#164)
• JuMP tutorial to NLopt doesn't work (#165)
• NLopt with JuMP puts out-of-bounds value into objective function before optimizing. (#170)

Merged pull requests:

• Fix #133, remove_constraints!(opt) doesn't work; chkn() not defined (#159) (@sjvrensburg)
• Fix typo in JuMP example (#162) (@aquohn)
• Remove leftover hessian code (#167) (@blegat)
• Remove useless line (#168) (@blegat)
• Remove useless evaluation at zero (#171) (@blegat)

NLopt v0.6.2

Diff since v0.6.1

Closed issues:

• Issues with Global optimization (#151)
• sslcerts error when installing NLopt (#152)
• Passing parameters and second derivatives (#154)

Merged pull requests:

• new algorithm-parameter API (#153) (@stevengj)

NLopt v0.6.1

Diff since v0.6.0

Closed issues:

• Compatibility with JuMP 0.19? (#130)

Merged pull requests:

• Add MOI wrapper (#150) (@blegat)

NLopt v0.6.0

Diff since v0.5.1

Closed issues:

• build failiure on OSX 10.9 (#6)
• Segfault in LN_SBPLX and LN_NELDERMEAD (#64)
• segfault when setting equal upper and lower bounds and constraints (#81)
• Build failure on Travis Linux v0.6 without sudo (#94)
• Build error (#98)
• LoadError: cannot assign variable Base.count from module Main (#119)
• Tutorial example gives answer "type Opt has no field numevals" (#120)
• Use of NLopt for complex numbers (#121)
• erf not defined (#125)
• forced stop when minimizing x^2 (#127)
• Cannot find :FORCED_STOP location (#135)
• curl error when building (#137)
• NLopt build error (#138)
• Yet another build error (#143)

Merged pull requests:

• rename get_initial_step in README.md (#123) (@magerton)
• Install TagBot as a GitHub Action (#140) (@JuliaTagBot)
• add Project.toml (#141) (@JeffBezanson)
• make Callback_Data mutable (#142) (@JeffBezanson)
• switch to NLopt_jll (#145) (@stevengj)

fix __precompile__ deprecation on 1.0