dotnet: Remove reference to dotnet release command

- Currently not implemented...

Add abseil patch

- Add patches/absl-config.cmake

Makefile: Add abseil-cpp on unix

- Force abseil-cpp SHA1 to 45221cc
  note: Just before the PR #136 which break all CMake

Makefile: Add abseil-cpp on windows

- Force abseil-cpp SHA1 to 45221cc
  note: Just before the PR #136 which break all CMake

CMake: Add abseil-cpp

- Force abseil-cpp SHA1 to 45221cc
  note: Just before the PR #136 which break all CMake

port to absl: C++ Part

- Fix warning with the use of ABSL_MUST_USE_RESULT
  > The macro must appear as the very first part of a function
    declaration or definition:
    ...
    Note: past advice was to place the macro after the argument list.
  src: dependencies/sources/abseil-cpp-master/absl/base/attributes.h:418
- Rename enum after windows clash
- Remove non compact table constraints
- Change index type from int64 to int in routing library
- Fix file_nonport compilation on windows
- Fix another naming conflict with windows (NO_ERROR is a macro)
- Cleanup hash containers; work on sat internals
- Add optional_boolean sub-proto

Sync cpp examples with internal code
- reenable issue173 after reducing number of loops

port to absl: Python Part

- Add back cp_model.INT32_MIN|MAX for examples

Update Python examples

- Add random_tsp.py
- Run words_square example
- Run magic_square in python tests

port to absl: Java Part

- Fix compilation of the new routing parameters in java
- Protect some code from SWIG parsing

Update Java Examples

port to absl: .Net Part

Update .Net examples

work on sat internals; Add C++ CP-SAT CpModelBuilder API; update sample code and recipes to use the new API; sync with internal code

Remove VS 2015 in Appveyor-CI

- abseil-cpp does not support VS 2015...

improve tables

upgrade C++ sat examples to use the new API; work on sat internals

update license dates

rewrite jobshop_ft06_distance.py to use the CP-SAT solver

rename last example

revert last commit

more work on SAT internals

fix

ortools/constraint_solver/routing_enums.proto

@@ -23,9 +23,12 @@ package operations_research;
 // First solution strategies, used as starting point of local search.
 message FirstSolutionStrategy {
   enum Value {
+    // See the homonymous value in LocalSearchMetaheuristic.
+    UNSET = 0;
+
     // Lets the solver detect which strategy to use according to the model being
     // solved.
-    AUTOMATIC = 0;
+    AUTOMATIC = 15;
 
     // --- Path addition heuristics ---
     // Starting from a route "start" node, connect it to the node which produces
@@ -72,14 +75,20 @@ message FirstSolutionStrategy {
     // optional nodes (with finite penalty costs).
     BEST_INSERTION = 7;
     // Iteratively build a solution by inserting the cheapest node at its
-    // cheapest position; the cost of insertion is based on the the arc cost
+    // cheapest position; the cost of insertion is based on the arc cost
     // function. Is faster than BEST_INSERTION.
     PARALLEL_CHEAPEST_INSERTION = 8;
+    // Iteratively build a solution by constructing routes sequentially, for
+    // each route inserting the cheapest node at its cheapest position until the
+    // route is completed; the cost of insertion is based on the arc cost
+    // function. Is faster than PARALLEL_CHEAPEST_INSERTION.
+    SEQUENTIAL_CHEAPEST_INSERTION = 14;
     // Iteratively build a solution by inserting each node at its cheapest
-    // position; the cost of insertion is based on the the arc cost function.
+    // position; the cost of insertion is based on the arc cost function.
     // Differs from PARALLEL_CHEAPEST_INSERTION by the node selected for
-    // insertion; here nodes are considered in their order of creation. Is
-    // faster than PARALLEL_CHEAPEST_INSERTION.
+    // insertion; here nodes are considered in decreasing order of distance to
+    // the start/ends of the routes, i.e. farthest nodes are inserted first.
+    // Is faster than SEQUENTIAL_CHEAPEST_INSERTION.
     LOCAL_CHEAPEST_INSERTION = 9;
 
     // --- Variable-based heuristics ---
@@ -100,8 +109,13 @@ message FirstSolutionStrategy {
 // descent, they will try to escape local minima.
 message LocalSearchMetaheuristic {
   enum Value {
+    // Means "not set". If the solver sees that, it'll behave like for
+    // AUTOMATIC. But this value won't override others upon a proto MergeFrom(),
+    // whereas "AUTOMATIC" will.
+    UNSET = 0;
+
     // Lets the solver select the metaheuristic.
-    AUTOMATIC = 0;
+    AUTOMATIC = 6;
 
     // Accepts improving (cost-reducing) local search neighbors until a local
     // minimum is reached.
@@ -116,8 +130,9 @@ message LocalSearchMetaheuristic {
     // Uses tabu search to escape local minima
     // (cf. http://en.wikipedia.org/wiki/Tabu_search).
     TABU_SEARCH = 4;
-    // Uses tabu search on the objective value of solution to escape local
-    // minima
-    OBJECTIVE_TABU_SEARCH = 5;
+    // Uses tabu search on a list of variables to escape local minima. The list
+    // of variables to use must be provided via the SetTabuVarsCallback
+    // callback.
+    GENERIC_TABU_SEARCH = 5;
   }
 }