Update output of value_gradient! (#96)

* Update output of value_gradient!

* Fix value_gradient call output in alpha test.

Author: pkofod

test/alphacalc.jl

@@ -16,8 +16,7 @@
             df = NLSolversBase.OnceDifferentiable(f,g!,x)
 
             xtmp = copy(x)
-            phi_0 = NLSolversBase.value_gradient!(df, x)
-            grtmp = NLSolversBase.gradient(df)
+            phi_0, grtmp = NLSolversBase.value_gradient!(df, x)
             p = -grtmp
             dphi_0 = dot(p, grtmp)
 
@@ -42,8 +41,7 @@
             df = NLSolversBase.OnceDifferentiable(f,g!,x)
 
             xtmp = copy(x)
-            phi_0 = NLSolversBase.value_gradient!(df, x)
-            grtmp = NLSolversBase.gradient(df)
+            phi_0, grtmp = NLSolversBase.value_gradient!(df, x)
             p = -grtmp
             dphi_0 = dot(p, grtmp)
 

test/initial.jl

@@ -7,8 +7,7 @@
     x = [-1., -1.]
     df = NLSolversBase.OnceDifferentiable(f,g!,x)
 
-    phi_0 = NLSolversBase.value_gradient!(df, x)
-    grtmp = NLSolversBase.gradient(df)
+    phi_0, grtmp = NLSolversBase.value_gradient!(df, x)
     p = -grtmp
     dphi_0 = dot(p, grtmp)