Use a CachingOptimizer for the MIP solver and test Cbc

Project.toml

@@ -1,13 +1,14 @@
 name = "Pavito"
 uuid = "cd433a01-47d1-575d-afb7-6db927ee8d8f"
 repo = "https://github.com/jump-dev/Pavito.jl.git"
-version = "0.3.3"
+version = "0.3.4"
 
 [deps]
 MathOptInterface = "b8f27783-ece8-5eb3-8dc8-9495eed66fee"
 Printf = "de0858da-6303-5e67-8744-51eddeeeb8d7"
 
 [compat]
+Cbc = "0.9, 1"
 GLPK = "0.15, 1"
 Ipopt = "0.8, 0.9, 1"
 JuMP = "0.22, 0.23"
@@ -16,11 +17,12 @@ MathOptInterface = "0.10.3, 1"
 julia = "1.6"
 
 [extras]
+Cbc = "9961bab8-2fa3-5c5a-9d89-47fab24efd76"
 GLPK = "60bf3e95-4087-53dc-ae20-288a0d20c6a6"
 Ipopt = "b6b21f68-93f8-5de0-b562-5493be1d77c9"
 JuMP = "4076af6c-e467-56ae-b986-b466b2749572"
 MINLPTests = "ee0a3090-8ee9-5cdb-b8cb-8eeba3165522"
 Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
 
 [targets]
-test = ["GLPK", "Ipopt", "JuMP", "MINLPTests", "Test"]
+test = ["Cbc", "GLPK", "Ipopt", "JuMP", "MINLPTests", "Test"]

src/MOI_wrapper.jl

@@ -335,24 +335,25 @@ end
 # utilities:
 
 function _mip(model::Optimizer)
-    if isnothing(model.mip_optimizer)
-        if isnothing(model.mip_solver)
-            error("No MIP solver specified (set `mip_solver` attribute)\n")
-        end
-
-        model.mip_optimizer =
-            MOI.instantiate(model.mip_solver, with_bridge_type = Float64)
-
-        supports_lazy =
-            MOI.supports(model.mip_optimizer, MOI.LazyConstraintCallback())
-        if isnothing(model.mip_solver_drives)
-            model.mip_solver_drives = supports_lazy
-        elseif model.mip_solver_drives && !supports_lazy
-            error(
-                "MIP solver (`mip_solver`) does not support lazy constraint " *
-                "callbacks (cannot set `mip_solver_drives` attribute to `true`)",
-            )
-        end
+    if model.mip_optimizer !== nothing
+        return model.mip_optimizer
+    end
+    if model.mip_solver === nothing
+        error("No MIP solver specified (set `mip_solver` attribute)\n")
+    end
+    model.mip_optimizer = MOI.Utilities.CachingOptimizer(
+        MOI.Utilities.UniversalFallback(MOI.Utilities.Model{Float64}()),
+        MOI.instantiate(model.mip_solver, with_bridge_type = Float64),
+    )
+    supports_lazy =
+        MOI.supports(model.mip_optimizer, MOI.LazyConstraintCallback())
+    if model.mip_solver_drives === nothing
+        model.mip_solver_drives = supports_lazy
+    elseif model.mip_solver_drives && !supports_lazy
+        error(
+            "MIP solver (`mip_solver`) does not support lazy constraint " *
+            "callbacks (cannot set `mip_solver_drives` attribute to `true`)",
+        )
     end
     return model.mip_optimizer
 end

test/runtests.jl

@@ -6,6 +6,7 @@
 
 using Test
 
+import Cbc
 import GLPK
 import Ipopt
 import MathOptInterface
@@ -15,6 +16,14 @@ const MOI = MathOptInterface
 include("MOI_wrapper.jl")
 include("jump_tests.jl")
 
+# !!! info
+#     We test with both Cbc and GLPK because they have very different
+#     implementations of the MOI API: GLPK supports incremental modification and
+#     supports lazy constraints, whereas Cbc supports copy_to and does not
+#     support lazy constraints. In addition, Cbc uses MatrixOfConstraints to
+#     simplify the copy process, needing an additional cache if we modify after
+#     the solve.
+
 @testset "MOI" begin
     TestMOIWrapper.runtests(
         MOI.OptimizerWithAttributes(
@@ -28,6 +37,14 @@ include("jump_tests.jl")
     )
 end
 
+@testset "Cbc" begin
+    TestMOIWrapper._run_moi_tests(
+        false,  # mip_solver_drives
+        MOI.OptimizerWithAttributes(Cbc.Optimizer, MOI.Silent() => true),
+        MOI.OptimizerWithAttributes(Ipopt.Optimizer, MOI.Silent() => true),
+    )
+end
+
 @testset "JuMP" begin
     TestJuMP.runtests(
         MOI.OptimizerWithAttributes(