Add test for ErrorException when calling solution computed with dense=false

Author: PerezHz

test/many_ode.jl

@@ -34,11 +34,10 @@ import Logging: Warn
         δt = (_abstol/q0T[1].coeffs[end-1])^inv(_order-1)
         @test TaylorIntegration.stepsize(q0T, _abstol) == δt
 
-        sol = (@test_logs (Warn, max_iters_reached()) taylorinteg(
-            eqs_mov!, q0, 0.0, 0.5, _order, _abstol, nothing))
-        @test_logs((Warn, max_iters_reached()),
+        sol = @test_logs((Warn, max_iters_reached()),
             @inferred(TaylorSolution{Float64, Float64, 2},
                 taylorinteg(eqs_mov!, q0, 0.0, 0.5, _order, _abstol, nothing)))
+        @test_throws ErrorException sol(0.25)
         tv = sol.t; xv = sol.x
         @test length(tv) == 501
         @test isa(xv, SubArray)

test/solution.jl

@@ -10,11 +10,9 @@ import Logging: Warn
     xv = rand(2,2)
     psol = Taylor1.(rand(2,1),2)
     nsteps = 2
-    # @show @code_warntype TaylorIntegration.build_solution(tv, xv, psol, nsteps)
     sol2 = TaylorIntegration.build_solution(tv, xv, psol, nsteps)
     @test sol2 isa TaylorSolution{Float64, Float64, 2}
     @test string(sol2) == "tspan: (1.0, 2.0), x: 2 Float64 variables"
-    # @show @code_warntype TaylorIntegration.build_solution(tv, Vector(xv[1,:]), Vector(psol[1,:]), nsteps)
     sol1 = TaylorIntegration.build_solution(tv, Vector(xv[1,:]), Vector(psol[1,:]), nsteps)
     @test sol1 isa TaylorSolution{Float64, Float64, 1}
     @test string(sol1) == "tspan: (1.0, 2.0), x: 1 Float64 variable"