Several fixes (#2572)

* repared serialization after update process; fixes #2545

* non-stratified folds in python could omit some data instances

* Makefile: fixes for older makes on windows; clean R-package too

* make cub to be a shallow submodule

* improve $(MAKE) recovery

.gitmodules

@@ -10,3 +10,4 @@
 [submodule "cub"]
 	path = cub
 	url = https://github.com/NVlabs/cub
+	shallow = true

Makefile

@@ -16,6 +16,20 @@ endif
 
 ROOTDIR = $(CURDIR)
 
+# workarounds for some buggy old make & msys2 versions seen in windows
+ifeq (NA, $(shell test ! -d "$(ROOTDIR)" && echo NA ))
+        $(warning Attempting to fix non-existing ROOTDIR [$(ROOTDIR)])
+        ROOTDIR := $(shell pwd)
+        $(warning New ROOTDIR [$(ROOTDIR)] $(shell test -d "$(ROOTDIR)" && echo " is OK" ))
+endif
+MAKE_OK := $(shell "$(MAKE)" -v 2> /dev/null)
+ifndef MAKE_OK
+        $(warning Attempting to recover non-functional MAKE [$(MAKE)])
+        MAKE := $(shell which make 2> /dev/null)
+        MAKE_OK := $(shell "$(MAKE)" -v 2> /dev/null)
+endif
+$(warning MAKE [$(MAKE)] - $(if $(MAKE_OK),checked OK,PROBLEM))
+
 ifeq ($(OS), Windows_NT)
 	UNAME="Windows"
 else
@@ -215,6 +229,7 @@ endif
 clean:
 	$(RM) -rf build build_plugin lib bin *~ */*~ */*/*~ */*/*/*~ */*.o */*/*.o */*/*/*.o #xgboost
 	$(RM) -rf build_tests *.gcov tests/cpp/xgboost_test
+	cd R-package/src; $(RM) -rf rabit src include dmlc-core amalgamation *.so *.dll; cd $(ROOTDIR)
 
 clean_all: clean
 	cd $(DMLC_CORE); $(MAKE) clean; cd $(ROOTDIR)
@@ -229,8 +244,7 @@ pypack: ${XGBOOST_DYLIB}
 	cd python-package; tar cf xgboost.tar xgboost; cd ..
 
 # create pip installation pack for PyPI
-pippack:
+pippack: clean_all
-	$(MAKE) clean_all
 	rm -rf xgboost-python
 	cp -r python-package xgboost-python
 	cp -r Makefile xgboost-python/xgboost/
@@ -241,8 +255,7 @@ pippack:
 	cp -r rabit xgboost-python/xgboost/
 
 # Script to make a clean installable R package.
-Rpack:
+Rpack: clean_all
-	$(MAKE) clean_all
 	rm -rf xgboost xgboost*.tar.gz
 	cp -r R-package xgboost
 	rm -rf xgboost/src/*.o xgboost/src/*.so xgboost/src/*.dll
@@ -264,13 +277,11 @@ Rpack:
 	cp xgboost/src/Makevars.in xgboost/src/Makevars.win
 	sed -i -e 's/@OPENMP_CXXFLAGS@/$$\(SHLIB_OPENMP_CFLAGS\)/g' xgboost/src/Makevars.win
 
-Rbuild:
+Rbuild: Rpack
-	$(MAKE) Rpack
 	R CMD build --no-build-vignettes xgboost
 	rm -rf xgboost
 
-Rcheck:
+Rcheck: Rbuild
-	$(MAKE) Rbuild
 	R CMD check  xgboost*.tar.gz
 
 -include build/*.d

python-package/xgboost/training.py

@@ -235,8 +235,7 @@ def mknfold(dall, nfold, param, seed, evals=(), fpreproc=None, stratified=False,
             idx = np.random.permutation(dall.num_row())
         else:
             idx = np.arange(dall.num_row())
-        kstep = int(len(idx) / nfold)
+        idset = np.array_split(idx, nfold)
-        idset = [idx[(i * kstep): min(len(idx), (i + 1) * kstep)] for i in range(nfold)]
     elif folds is not None and isinstance(folds, list):
         idset = [x[1] for x in folds]
         nfold = len(idset)

src/gbm/gbtree_model.h

@@ -75,6 +75,7 @@ struct GBTreeModel {
       }
       trees.clear();
       param.num_trees = 0;
+      tree_info.clear();
     }
   }