merged

README.md

-# dlib C++ library ![Travis Status](https://travis-ci.org/davisking/dlib.svg?branch=master)
+# dlib C++ library [![Travis Status](https://travis-ci.org/davisking/dlib.svg?branch=master)](https://travis-ci.org/davisking/dlib)
 
 Dlib is a modern C++ toolkit containing machine learning algorithms and tools for creating complex software in C++ to solve real world problems. See [http://dlib.net](http://dlib.net) for the main project documentation and API reference.
 

dlib/image_saver/save_jpeg.cpp

@@ -10,14 +10,15 @@
 #include "../pixel.h"
 #include "save_jpeg.h"
 #include <stdio.h>
+#include <sstream>
+#include <setjmp.h>
+#include "image_saver.h"
+
 #ifdef DLIB_JPEG_STATIC
 #   include "../external/libjpeg/jpeglib.h"
 #else
 #   include <jpeglib.h>
 #endif
-#include <sstream>
-#include <setjmp.h>
-#include "image_saver.h"
 
 namespace dlib
 {

dlib/test/active_learning.cpp

@@ -150,7 +150,7 @@ namespace
 
             // When we pick the best/front ranked element then the active learning method
             // shouldn't do much worse than random selection (and often much better).
-            DLIB_TEST(test_rank_unlabeled_training_samples(samples, labels, max_min_margin, 25, true) >= 0.97);
+            DLIB_TEST(test_rank_unlabeled_training_samples(samples, labels, max_min_margin, 35, true) >= 0.97);
             DLIB_TEST(test_rank_unlabeled_training_samples(samples, labels, ratio_margin, 25, true) >= 0.96);
             // However, picking the worst ranked element should do way worse than random
             // selection.

dlib/test/max_cost_assignment.cpp

@@ -120,7 +120,7 @@ namespace
             T true_eval = assignment_cost(cost, assign);
             assign = max_cost_assignment(cost);
             DLIB_TEST(assignment_cost(cost,assign) == true_eval);
-            assign = max_cost_assignment(matrix_cast<char>(cost));
+            assign = max_cost_assignment(matrix_cast<signed char>(cost));
             DLIB_TEST(assignment_cost(cost,assign) == true_eval);