Only define one dlib target in CMake and shared vs. static is determined by BUILD_SHARED_LIBS variable. (#1138)

* Changed cmake so that there is only the dlib target and it isn't forced
to be static or shared, instead, the build type will toggle based on the
state of CMake's BUILD_SHARED_LIBS variable.

* Make CMake build dlib statically whenever DLIB_IN_PROJECT_BUILD==true,
regardless of the state of BUILD_SHARED_LIBS.  This means projects that
use dlib by saying add_subdirectory(dlib) will always statically link to
dlib, unless DLIB_IN_PROJECT_BUILD is explicltly set to false and
BUILD_SHARED_LIBS set to true.

* Change types of tensor's size-related members, so that integer overflow does not happen so easily (see #1148)

* Fix missing typecast

functions.

* Added support for variadic Python functions in find_max_global.

* Add test for find_{min,max}_global on variadic functions.

see what happened.

Fixed pybind11 not doing the overload resolution correctly on the cnn_face_detector's operator() in Python.

didn't have complete coverage and would let incorrectly typed numpy arrays
through.

* Added set_dnn_prefer_smallest_algorithms()

* Added description for set_dnn_prefer_smallest_algorithms();w

* Changed architectures to algorithms.

runtime, even with very bad parameter selections.

- Moved all uses of PYBIND11_MAKE_OPAQUE to a common header to avoid possilbe ODR violations.
- Added python binding for make_bounding_box_regression_training_data().

of a tensor changed.  I.e. if you ran mini-batches with different sample sizes
you would get an assert triggering.  This has been fixed.