Remove code that is no longer needed

The code in ccutil/hashfn.h was needed for some old compilers. Now that we support MSVC >= 2010 and compilers that has good support for C++11, we can drop this code.

As a result of this file removal, we now use:
  std::unordered_map
  std::unordered_set
  std::unique_ptr
directly in the codebase with '#include' for the needed headers.

Author: amitdo

ccutil/Makefile.am

@@ -18,7 +18,7 @@ include_HEADERS = \
 
 noinst_HEADERS = \
     ambigs.h bits16.h bitvector.h ccutil.h clst.h doubleptr.h elst2.h \
-    elst.h genericheap.h globaloc.h hashfn.h indexmapbidi.h kdpair.h lsterr.h \
+    elst.h genericheap.h globaloc.h indexmapbidi.h kdpair.h lsterr.h \
     nwmain.h object_cache.h qrsequence.h sorthelper.h stderr.h \
     scanutils.h tessdatamanager.h tprintf.h unicity_table.h unicodes.h \
     universalambigs.h

ccutil/hashfn.h

@@ -57,9 +57,9 @@ struct RadicalStrokedHash {
 };
 
 // A hash map to convert unicodes to radical,stroke pair.
-typedef TessHashMap<int, RadicalStroke> RSMap;
+typedef std::unordered_map<int, RadicalStroke> RSMap;
 // A hash map to count occurrences of each radical,stroke pair.
-typedef TessHashMap<RadicalStroke, int, RadicalStrokedHash> RSCounts;
+typedef std::unordered_map<RadicalStroke, int, RadicalStrokedHash> RSCounts;
 
 // Helper function builds the RSMap from the radical-stroke file, which has
 // already been read into a STRING. Returns false on error.

ccutil/unicharcompress.cpp

@@ -22,7 +22,8 @@
 #ifndef TESSERACT_CCUTIL_UNICHARCOMPRESS_H_
 #define TESSERACT_CCUTIL_UNICHARCOMPRESS_H_
 
-#include "hashfn.h"
+#include <unordered_map>
+
 #include "serialis.h"
 #include "strngs.h"
 #include "unicharset.h"
@@ -236,17 +237,19 @@ class UnicharCompress {
   // encoder_ is the only part that is serialized. The rest is computed on load.
   GenericVector<RecodedCharID> encoder_;
   // Decoder converts the output of encoder back to a unichar-id.
-  TessHashMap<RecodedCharID, int, RecodedCharID::RecodedCharIDHash> decoder_;
+  std::unordered_map<RecodedCharID, int,
+                           RecodedCharID::RecodedCharIDHash>
+      decoder_;
   // True if the index is a valid single or start code.
   GenericVector<bool> is_valid_start_;
   // Maps a prefix code to a list of valid next codes.
   // The map owns the vectors.
-  TessHashMap<RecodedCharID, GenericVectorEqEq<int>*,
+  std::unordered_map<RecodedCharID, GenericVectorEqEq<int>*,
               RecodedCharID::RecodedCharIDHash>
       next_codes_;
   // Maps a prefix code to a list of valid final codes.
   // The map owns the vectors.
-  TessHashMap<RecodedCharID, GenericVectorEqEq<int>*,
+  std::unordered_map<RecodedCharID, GenericVectorEqEq<int>*,
               RecodedCharID::RecodedCharIDHash>
       final_codes_;
   // Max of any value in encoder_ + 1.

ccutil/unicharcompress.h

@@ -1211,7 +1211,7 @@ double LSTMTrainer::ComputeCharError(const GenericVector<int>& truth_str,
 // Computes word recall error rate using a very simple bag of words algorithm.
 // NOTE that this is destructive on both input strings.
 double LSTMTrainer::ComputeWordError(STRING* truth_str, STRING* ocr_str) {
-  typedef TessHashMap<std::string, int, std::hash<std::string> > StrMap;
+  typedef std::unordered_map<std::string, int, std::hash<std::string> > StrMap;
   GenericVector<STRING> truth_words, ocr_words;
   truth_str->split(' ', &truth_words);
   if (truth_words.empty()) return 0.0;

lstm/lstmtrainer.cpp

@@ -21,9 +21,10 @@
 #ifndef TESSERACT_TEXTORD_BBGRID_H_
 #define TESSERACT_TEXTORD_BBGRID_H_
 
+#include <unordered_set>
+
 #include "clst.h"
 #include "coutln.h"
-#include "hashfn.h"
 #include "rect.h"
 #include "scrollview.h"
 
@@ -364,7 +365,7 @@ template<class BBC, class BBC_CLIST, class BBC_C_IT> class GridSearch {
   // An iterator over the list at (x_, y_) in the grid_.
   BBC_C_IT it_;
   // Set of unique returned elements used when unique_mode_ is true.
-  TessHashSet<BBC*, PtrHash<BBC> > returns_;
+  std::unordered_set<BBC*, PtrHash<BBC> > returns_;
 };
 
 // Sort function to sort a BBC by bounding_box().left().

textord/bbgrid.h

@@ -46,7 +46,7 @@ const int kMinLigature = 0xfb00;
 const int kMaxLigature = 0xfb17;  // Don't put the wide Hebrew letters in.
 
 /* static */
-SmartPtr<LigatureTable> LigatureTable::instance_;
+std::unique_ptr<LigatureTable> LigatureTable::instance_;
 
 /* static */
 LigatureTable* LigatureTable::Get() {

training/ligature_table.cpp

@@ -23,16 +23,17 @@
 #define TRAININGDATA_LIGATURE_TABLE_H_
 
 #include <string>
+#include <unordered_map>
+#include <memory>
 
-#include "hashfn.h"
 #include "util.h"
 
 namespace tesseract {
 
 class PangoFontInfo;  // defined in pango_font_info.h
 
 // Map to substitute strings for ligatures.
-typedef TessHashMap<string, string, StringHash> LigHash;
+typedef std::unordered_map<string, string, StringHash> LigHash;
 
 class LigatureTable {
  public:
@@ -61,7 +62,7 @@ class LigatureTable {
   // corresponding ligature characters.
   void Init();
 
-  static SmartPtr<LigatureTable> instance_;
+  static std::unique_ptr<LigatureTable> instance_;
   LigHash norm_to_lig_table_;
   LigHash lig_to_norm_table_;
   int min_lig_length_;

training/ligature_table.h

@@ -688,7 +688,7 @@ void FontUtils::GetAllRenderableCharacters(const vector<string>& fonts,
 // Utilities written to be backward compatible with StringRender
 
 /* static */
-int FontUtils::FontScore(const TessHashMap<char32, inT64>& ch_map,
+int FontUtils::FontScore(const std::unordered_map<char32, inT64>& ch_map,
                          const string& fontname, int* raw_score,
                          vector<bool>* ch_flags) {
   PangoFontInfo font_info;
@@ -704,7 +704,7 @@ int FontUtils::FontScore(const TessHashMap<char32, inT64>& ch_map,
   }
   *raw_score = 0;
   int ok_chars = 0;
-  for (TessHashMap<char32, inT64>::const_iterator it = ch_map.begin();
+  for (std::unordered_map<char32, inT64>::const_iterator it = ch_map.begin();
        it != ch_map.end(); ++it) {
     bool covered = (IsWhitespace(it->first) ||
                     (pango_coverage_get(coverage, it->first)
@@ -722,7 +722,7 @@ int FontUtils::FontScore(const TessHashMap<char32, inT64>& ch_map,
 
 
 /* static */
-string FontUtils::BestFonts(const TessHashMap<char32, inT64>& ch_map,
+string FontUtils::BestFonts(const std::unordered_map<char32, inT64>& ch_map,
                             vector<pair<const char*, vector<bool> > >* fonts) {
   const double kMinOKFraction = 0.99;
   // Weighted fraction of characters that must be renderable in a font to make

training/pango_font_info.cpp

@@ -23,9 +23,9 @@
 #include <string>
 #include <utility>
 #include <vector>
+#include <unordered_map>
 
 #include "commandlineflags.h"
-#include "hashfn.h"
 #include "host.h"
 #include "pango/pango-font.h"
 #include "pango/pango.h"
@@ -203,15 +203,15 @@ class FontUtils {
   // corresponding character (in order of iterating ch_map) can be rendered.
   // The return string is a list of the acceptable fonts that were used.
   static string BestFonts(
-      const TessHashMap<char32, inT64>& ch_map,
+      const std::unordered_map<char32, inT64>& ch_map,
       std::vector<std::pair<const char*, std::vector<bool> > >* font_flag);
 
   // FontScore returns the weighted renderability score of the given
   // hash map character table in the given font. The unweighted score
   // is also returned in raw_score.
   // The values in the bool vector ch_flags correspond to whether the
   // corresponding character (in order of iterating ch_map) can be rendered.
-  static int FontScore(const TessHashMap<char32, inT64>& ch_map,
+  static int FontScore(const std::unordered_map<char32, inT64>& ch_map,
                        const string& fontname, int* raw_score,
                        std::vector<bool>* ch_flags);
 

training/pango_font_info.h

@@ -31,8 +31,8 @@
 
 #include <string>
 #include <vector>
+#include <unordered_map>
 
-#include "hashfn.h"
 #include "host.h"
 #include "pango_font_info.h"
 #include "pango/pango-layout.h"
@@ -210,7 +210,7 @@ class StringRenderer {
   Boxa* page_boxes_;
 
   // Objects cached for subsequent calls to RenderAllFontsToImage()
-  TessHashMap<char32, inT64> char_map_;  // Time-saving char histogram.
+  std::unordered_map<char32, inT64> char_map_;  // Time-saving char histogram.
   int total_chars_;   // Number in the string to be rendered.
   int font_index_;    // Index of next font to use in font list.
   int last_offset_;   // Offset returned from last successful rendering