Merge branch 'master' of github.com-egorpugin:tesseract-ocr/tesseract

src/ccstruct/boxread.cpp

@@ -27,14 +27,13 @@
 #include "helpers.h"        // for chomp_string
 #include <tesseract/unichar.h>        // for UNICHAR
 
+#include <climits>          // for INT_MAX
 #include <cstring>          // for strchr, strcmp
 #include <fstream>          // for std::ifstream
 #include <locale>           // for std::locale::classic
 #include <sstream>          // for std::stringstream
 #include <string>           // for std::string
 
-#include <limits.h>         // for INT_MAX
-
 namespace tesseract {
 
 // Special char code used to identify multi-blob labels.

src/ccutil/errcode.h

@@ -51,11 +51,7 @@ class TESS_API ERRCODE {           // error handler class
 
 constexpr ERRCODE ASSERT_FAILED("Assert failed");
 
-#if defined __cplusplus
-# define DO_NOTHING static_cast<void>(0)
-#else
-# define DO_NOTHING (void)(0)
-#endif
+#define DO_NOTHING static_cast<void>(0)
 
 #define ASSERT_HOST(x) (x) \
   ? DO_NOTHING \

src/ccutil/host.h

@@ -20,6 +20,7 @@
 
 #include <tesseract/platform.h>
 
+#include <climits>
 #include <limits>
 
 /* _WIN32 */
@@ -36,11 +37,6 @@
 #endif  // _WIN32
 
 #ifndef _WIN32
-#  ifdef __cplusplus
-#    include <climits>
-#  else /* C compiler*/
-#    include <limits.h>
-#  endif /* __cplusplus */
 #  ifndef PATH_MAX
 #    define MAX_PATH 4096
 #  else

src/ccutil/serialis.cpp

@@ -22,8 +22,8 @@
 
 #include "helpers.h"  // for ReverseN
 
+#include <climits>    // for INT_MAX
 #include <cstdio>
-#include <limits.h>         // for INT_MAX
 
 namespace tesseract {
 

src/classify/intproto.h

@@ -173,7 +173,7 @@ enum IntmatcherDebugAction {
 #define PPrunerMaskFor(I) (1 << PPrunerBitIndexFor(I))
 
 #define MaxNumClassesIn(T) (T->NumClassPruners * CLASSES_PER_CP)
-#define LegalClassId(c) ((c) >= 0 && (c) <= MAX_CLASS_ID)
+#define LegalClassId(c) ((c) >= 0 && (c) < MAX_NUM_CLASSES)
 #define UnusedClassIdIn(T, c) ((T)->Class[c] == nullptr)
 #define ClassForClassId(T, c) ((T)->Class[c])
 #define ClassPrunersFor(T) ((T)->ClassPruner)

src/dict/matchdefs.h

@@ -20,7 +20,8 @@
 
 #include <tesseract/unichar.h>
 
-#include <cstdio>
+#include <climits> // INT16_MAX
+#include <cstdint> // int16_t
 
 namespace tesseract {
 
@@ -28,7 +29,6 @@ namespace tesseract {
   and the maximum class id for any matcher. This must be changed
   if more different classes need to be classified */
 #define MAX_NUM_CLASSES   INT16_MAX
-#define MAX_CLASS_ID    (MAX_NUM_CLASSES - 1)
 
 /** a CLASS_ID is the ascii character to be associated with a class */
 using CLASS_ID = UNICHAR_ID;
@@ -44,67 +44,6 @@ using PROTO_ID = int16_t;
   The feature id ranges from 0 to N-1 where N is the number
   of features in a character description. */
 using FEATURE_ID = uint8_t;
-#define NO_FEATURE      255
-#define NOISE_FEATURE   254
-#define MISSING_PROTO   254
-#define MAX_NUM_FEAT    40
-#define MAX_FEATURE_ID    250
-
-/** a RATING is the match rating returned by a classifier.
-  Higher is better. */
-using RATING = float;
-
-/** a CERTAINTY is an indication of the degree of confidence of the
-  classifier.  Higher is better.  0 means the match is as good as the
-  mean of the matches seen in training.  -1 means the match was one
-  standard deviation worse than the training matches, etc. */
-using CERTAINTY = float;
-
-/** define a data structure to hold a single match result */
-typedef struct
-{
-  CLASS_ID Class;
-  RATING Rating;
-  CERTAINTY Certainty;
-}
-
-
-MATCH_RESULT;
-
-/** define a data structure for holding an array of match results */
-typedef MATCH_RESULT SORTED_CLASSES[MAX_CLASS_ID + 1];
-
-/*----------------------------------------------------------------------------
-          Public Function Prototypes
-----------------------------------------------------------------------------*/
-/**
-  all feature matchers that are to be used with the high level
-  classifier must support the following interface.  The names will, of
-  course, be unique for each different matcher.  Note also that
-  FEATURE_STRUCT is a data structure that is defined specifically for
-  each feature extractor/matcher pair.
-*/
-
-/* misc test functions for proto id's and feature id's */
-#define IsValidFeature(Fid) ((Fid) < MAX_FEATURE_ID)
-#define IsValidProto(Pid) ((Pid) >= 0)
-
-#if defined(__STDC__) || defined(__cplusplus)
-# define _ARGS(s) s
-#else
-# define _ARGS(s) ()
-#endif
-
-/* matchdefs.c */
-int CompareMatchResults
-_ARGS ((MATCH_RESULT * Result1, MATCH_RESULT * Result2));
-
-void PrintMatchResult _ARGS ((FILE * File, MATCH_RESULT * MatchResult));
-
-void PrintMatchResults
-_ARGS ((FILE * File, int N, MATCH_RESULT MatchResults[]));
-
-#undef _ARGS
 
 } // namespace tesseract
 

unittest/Makefile.am

@@ -1,5 +1,8 @@
 # Makefile for Tesseract unit tests
 
+# Be less noisy by default. Can be overridden with `make V=1`.
+V = 0
+
 # Absolute path of directory 'langdata'.
 LANGDATA_DIR=$(shell cd $(top_srcdir) && cd .. && pwd)/langdata_lstm