Applied patch to refix issue 331

git-svn-id: https://tesseract-ocr.googlecode.com/svn/trunk@1064 d0cd1f9f-072b-0410-8dd7-cf729c803f20

ccstruct/blread.cpp

@@ -50,9 +50,8 @@ bool read_unlv_file(                    //print list of sides
   name += UNLV_EXT;              //add extension
   if ((pdfp = fopen (name.string (), "rb")) == NULL) {
     return false;                //didn't read one
-  }
+  } else {
-  else {
+    while (tfscanf(pdfp, "%d %d %d %d %*s", &x, &y, &width, &height) >= 4) {
-    while (fscanf (pdfp, "%d %d %d %d %*s", &x, &y, &width, &height) >= 4) {
                                  //make rect block
       block = new BLOCK (name.string (), TRUE, 0, 0,
                          (inT16) x, (inT16) (ysize - y - height),

ccutil/Makefile.am

@@ -18,8 +18,9 @@ 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 \
-    nwmain.h object_cache.h qrsequence.h secname.h sorthelper.h stderr.h tessdatamanager.h \
+    nwmain.h object_cache.h qrsequence.h secname.h sorthelper.h stderr.h \
-     tprintf.h unicity_table.h unicodes.h universalambigs.h
+    scanutils.h tessdatamanager.h tprintf.h unicity_table.h unicodes.h \
+    universalambigs.h
 
 if !USING_MULTIPLELIBS
 noinst_LTLIBRARIES = libtesseract_ccutil.la
@@ -34,15 +35,11 @@ libtesseract_ccutil_la_SOURCES = \
     elst2.cpp elst.cpp errcode.cpp \
     globaloc.cpp indexmapbidi.cpp \
     mainblk.cpp memry.cpp \
-    serialis.cpp strngs.cpp \
+    serialis.cpp strngs.cpp scanutils.cpp \
     tessdatamanager.cpp tprintf.cpp \
     unichar.cpp unicharmap.cpp unicharset.cpp unicodes.cpp \
     params.cpp universalambigs.cpp
 
-if EMBEDDED
-include_HEADERS += scanutils.h
-libtesseract_ccutil_la_SOURCES += scanutils.cpp
-endif
 
 if MINGW
 AM_CPPFLAGS += -I$(top_srcdir)/vs2008/port -DWINDLLNAME=\"lib@GENERIC_LIBRARY_NAME@\"

ccutil/scanutils.cpp

@@ -19,12 +19,9 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-#ifdef EMBEDDED
-
 #include <ctype.h>
 #include <stdarg.h>
 #include <stddef.h>
-#include <inttypes.h>
 #include <string.h>
 #include <limits.h>
 #include <stdio.h>
@@ -35,6 +32,11 @@
 #include "scanutils.h"
 #include "tprintf.h"
 
+// workaround for "'off_t' was not declared in this scope" with -std=c++11
+#ifndef off_t
+typedef long off_t;
+#endif  // off_t
+
 enum Flags {
   FL_SPLAT  = 0x01,   // Drop the value, do not assign
   FL_INV    = 0x02,   // Character-set with inverse
@@ -215,19 +217,48 @@ double strtofloat(const char* s)
   return minus ? -f : f;
 }
 
+static int tvfscanf(FILE* stream, const char *format, va_list ap);
+
+int tfscanf(FILE* stream, const char *format, ...)
+{
+  va_list ap;
+  int rv;
+
+  va_start(ap, format);
+  rv = tvfscanf(stream, format, ap);
+  va_end(ap);
+
+  return rv;
+}
+
+#ifdef EMBEDDED
+
 int fscanf(FILE* stream, const char *format, ...)
 {
   va_list ap;
   int rv;
 
   va_start(ap, format);
-  rv = vfscanf(stream, format, ap);
+  rv = tvfscanf(stream, format, ap);
   va_end(ap);
 
   return rv;
 }
 
-int vfscanf(FILE* stream, const char *format, va_list ap)
+int vfscanf(FILE* stream, const char *format, ...)
+{
+  va_list ap;
+  int rv;
+
+  va_start(ap, format);
+  rv = tvfscanf(stream, format, ap);
+  va_end(ap);
+
+  return rv;
+}
+#endif
+
+static int tvfscanf(FILE* stream, const char *format, va_list ap)
 {
   const char *p = format;
   char ch;
@@ -250,7 +281,7 @@ int vfscanf(FILE* stream, const char *format, va_list ap)
   enum Bail bail = BAIL_NONE;
   int sign;
   int converted = 0;    // Successful conversions
-  unsigned long matchmap[((1 << CHAR_BIT)+(LongBit()-1))/LongBit()];
+  unsigned long matchmap[((1 << CHAR_BIT)+(CHAR_BIT * sizeof(long) - 1))/ (CHAR_BIT * sizeof(long))];
   int matchinv = 0;   // Is match map inverted?
   unsigned char range_start = 0;
   off_t start_off = ftell(stream);
@@ -278,7 +309,8 @@ int vfscanf(FILE* stream, const char *format, va_list ap)
             flags |= FL_SPLAT;
           break;
 
-          case '0' ... '9':
+          case 0: case 1: case 2: case 3: case 4:
+          case 5: case 6: case 7: case 8: case 9:
             width = (ch-'0');
             state = ST_WIDTH;
             flags |= FL_WIDTH;
@@ -542,6 +574,7 @@ int vfscanf(FILE* stream, const char *format, va_list ap)
   return converted;
 }
 
+#ifdef EMBEDDED
 int creat(const char *pathname, mode_t mode)
 {
   return open(pathname, O_CREAT | O_TRUNC | O_WRONLY, mode);

ccutil/scanutils.h

@@ -19,14 +19,25 @@
 #ifndef TESSERACT_CCUTIL_SCANUTILS_H_
 #define TESSERACT_CCUTIL_SCANUTILS_H_
 
-#ifdef EMBEDDED
-
 #include <stdint.h>
 #include <stddef.h>
 #include <stdio.h>
-//#include <klibc/extern.h>
 #include <sys/stat.h>
 
+/**
+ * fscanf variant to ensure correct reading regardless of locale.
+ *
+ * tfscanf parse a file stream according to the given format. See the fscanf
+ * manpage for more information, as this function attempts to mimic its
+ * behavior.
+ *
+ * @note Note that scientific floating-point notation is not supported.
+ *
+ */
+int tfscanf(FILE* stream, const char *format, ...);
+
+#ifdef EMBEDDED
+
 // Attempts to parse the given file stream s as an integer of the base
 // 'base'. Returns the first successfully parsed integer as a uintmax_t, or
 // 0, if none was found.

classify/clusttool.cpp

@@ -46,7 +46,7 @@ History:	6/6/89, DSJ, Created.
 uinT16 ReadSampleSize(FILE *File) {
   int SampleSize;
 
-  if ((fscanf (File, "%d", &SampleSize) != 1) ||
+  if ((tfscanf(File, "%d", &SampleSize) != 1) ||
     (SampleSize < 0) || (SampleSize > MAXSAMPLESIZE))
     DoError (ILLEGALSAMPLESIZE, "Illegal sample size");
   return (SampleSize);
@@ -72,7 +72,7 @@ PARAM_DESC *ReadParamDesc(FILE *File, uinT16 N) {
 
   ParamDesc = (PARAM_DESC *) Emalloc (N * sizeof (PARAM_DESC));
   for (i = 0; i < N; i++) {
-    if (fscanf (File, "%s", Token) != 1)
+    if (tfscanf(File, "%s", Token) != 1)
       DoError (ILLEGALCIRCULARSPEC,
         "Illegal circular/linear specification");
     if (Token[0] == 'c')
@@ -80,14 +80,14 @@ PARAM_DESC *ReadParamDesc(FILE *File, uinT16 N) {
     else
       ParamDesc[i].Circular = FALSE;
 
-    if (fscanf (File, "%s", Token) != 1)
+    if (tfscanf(File, "%s", Token) != 1)
       DoError (ILLEGALESSENTIALSPEC,
         "Illegal essential/non-essential spec");
     if (Token[0] == 'e')
       ParamDesc[i].NonEssential = FALSE;
     else
       ParamDesc[i].NonEssential = TRUE;
-    if (fscanf (File, "%f%f", &(ParamDesc[i].Min), &(ParamDesc[i].Max)) !=
+    if (tfscanf(File, "%f%f", &(ParamDesc[i].Min), &(ParamDesc[i].Max)) !=
       2)
       DoError (ILLEGALMINMAXSPEC, "Illegal min or max specification");
     ParamDesc[i].Range = ParamDesc[i].Max - ParamDesc[i].Min;
@@ -119,7 +119,7 @@ PROTOTYPE *ReadPrototype(FILE *File, uinT16 N) {
   int SampleCount;
   int i;
 
-  if ((Status = fscanf (File, "%s", Token)) == 1) {
+  if ((Status = tfscanf(File, "%s", Token)) == 1) {
     Proto = (PROTOTYPE *) Emalloc (sizeof (PROTOTYPE));
     Proto->Cluster = NULL;
     if (Token[0] == 's')
@@ -129,7 +129,7 @@ PROTOTYPE *ReadPrototype(FILE *File, uinT16 N) {
 
     Proto->Style = ReadProtoStyle (File);
 
-    if ((fscanf (File, "%d", &SampleCount) != 1) || (SampleCount < 0))
+    if ((tfscanf(File, "%d", &SampleCount) != 1) || (SampleCount < 0))
       DoError (ILLEGALSAMPLECOUNT, "Illegal sample count");
     Proto->NumSamples = SampleCount;
 
@@ -173,7 +173,7 @@ PROTOTYPE *ReadPrototype(FILE *File, uinT16 N) {
         Proto->Distrib =
           (DISTRIBUTION *) Emalloc (N * sizeof (DISTRIBUTION));
         for (i = 0; i < N; i++) {
-          if (fscanf (File, "%s", Token) != 1)
+          if (tfscanf(File, "%s", Token) != 1)
             DoError (ILLEGALDISTRIBUTION,
               "Illegal prototype distribution");
           switch (Token[0]) {
@@ -245,7 +245,7 @@ PROTOSTYLE ReadProtoStyle(FILE *File) {
   char Token[TOKENSIZE];
   PROTOSTYLE Style;
 
-  if (fscanf (File, "%s", Token) != 1)
+  if (tfscanf(File, "%s", Token) != 1)
     DoError (ILLEGALSTYLESPEC, "Illegal prototype style specification");
   switch (Token[0]) {
     case 's':
@@ -290,7 +290,7 @@ FLOAT32* ReadNFloats(FILE * File, uinT16 N, FLOAT32 Buffer[]) {
     Buffer = reinterpret_cast<FLOAT32*>(Emalloc(N * sizeof(FLOAT32)));
 
   for (i = 0; i < N; i++) {
-    NumFloatsRead = fscanf(File, "%f", &(Buffer[i]));
+    NumFloatsRead = tfscanf(File, "%f", &(Buffer[i]));
     if (NumFloatsRead != 1) {
       if ((NumFloatsRead == EOF) && (i == 0)) {
         Efree(Buffer);

classify/cutoffs.cpp

@@ -68,7 +68,7 @@ void Classify::ReadNewCutoffs(FILE *CutoffFile, bool swap, inT64 end_offset,
     Cutoffs[i] = MAX_CUTOFF;
 
   while ((end_offset < 0 || ftell(CutoffFile) < end_offset) &&
-         fscanf(CutoffFile, "%" REALLY_QUOTE_IT(UNICHAR_LEN) "s %d",
+         tfscanf(CutoffFile, "%" REALLY_QUOTE_IT(UNICHAR_LEN) "s %d",
                 Class, &Cutoff) == 2) {
     if (strcmp(Class, "NULL") == 0) {
       ClassId = unicharset.unichar_to_id(" ");

classify/featdefs.cpp

@@ -265,13 +265,13 @@ CHAR_DESC ReadCharDescription(const FEATURE_DEFS_STRUCT &FeatureDefs,
   CHAR_DESC CharDesc;
   int Type;
 
-  if (fscanf (File, "%d", &NumSetsToRead) != 1 ||
+  if (tfscanf(File, "%d", &NumSetsToRead) != 1 ||
     NumSetsToRead < 0 || NumSetsToRead > FeatureDefs.NumFeatureTypes)
     DoError (ILLEGAL_NUM_SETS, "Illegal number of feature sets");
 
   CharDesc = NewCharDescription(FeatureDefs);
   for (; NumSetsToRead > 0; NumSetsToRead--) {
-    fscanf (File, "%s", ShortName);
+    tfscanf(File, "%s", ShortName);
     Type = ShortNameToFeatureType(FeatureDefs, ShortName);
     CharDesc->FeatureSets[Type] =
       ReadFeatureSet (File, FeatureDefs.FeatureDesc[Type]);

classify/mastertrainer.cpp

@@ -38,6 +38,8 @@
 #include "shapetable.h"
 #include "svmnode.h"
 
+#include "scanutils.h"
+
 namespace tesseract {
 
 // Constants controlling clustering. With a low kMinClusteredShapes and a high
@@ -360,8 +362,8 @@ bool MasterTrainer::LoadFontInfo(const char* filename) {
     fontinfo.name = font_name;
     fontinfo.properties = 0;
     fontinfo.universal_id = 0;
-    if (fscanf(fp, "%1024s %i %i %i %i %i\n", font_name,
+    if (tfscanf(fp, "%1024s %i %i %i %i %i\n", font_name,
-               &italic, &bold, &fixed, &serif, &fraktur) != 6)
+                &italic, &bold, &fixed, &serif, &fraktur) != 6)
       continue;
     fontinfo.properties =
         (italic << 0) +
@@ -397,7 +399,7 @@ bool MasterTrainer::LoadXHeights(const char* filename) {
   int total_xheight = 0;
   int xheight_count = 0;
   while (!feof(f)) {
-    if (fscanf(f, "%1023s %d\n", buffer, &xht) != 2)
+    if (tfscanf(f, "%1023s %d\n", buffer, &xht) != 2)
       continue;
     buffer[1023] = '\0';
     fontinfo.name = buffer;
@@ -441,13 +443,13 @@ bool MasterTrainer::AddSpacingInfo(const char *filename) {
   char uch[UNICHAR_LEN];
   char kerned_uch[UNICHAR_LEN];
   int x_gap, x_gap_before, x_gap_after, num_kerned;
-  ASSERT_HOST(fscanf(fontinfo_file, "%d\n", &num_unichars) == 1);
+  ASSERT_HOST(tfscanf(fontinfo_file, "%d\n", &num_unichars) == 1);
   FontInfo *fi = &fontinfo_table_.get(fontinfo_id);
   fi->init_spacing(unicharset_.size());
   FontSpacingInfo *spacing = NULL;
   for (int l = 0; l < num_unichars; ++l) {
-    if (fscanf(fontinfo_file, "%s %d %d %d",
+    if (tfscanf(fontinfo_file, "%s %d %d %d",
-               uch, &x_gap_before, &x_gap_after, &num_kerned) != 4) {
+                uch, &x_gap_before, &x_gap_after, &num_kerned) != 4) {
       tprintf("Bad format of font spacing file %s\n", filename);
       fclose(fontinfo_file);
       return false;
@@ -459,7 +461,7 @@ bool MasterTrainer::AddSpacingInfo(const char *filename) {
       spacing->x_gap_after = static_cast<inT16>(x_gap_after * scale);
     }
     for (int k = 0; k < num_kerned; ++k) {
-      if (fscanf(fontinfo_file, "%s %d", kerned_uch, &x_gap) != 2) {
+      if (tfscanf(fontinfo_file, "%s %d", kerned_uch, &x_gap) != 2) {
         tprintf("Bad format of font spacing file %s\n", filename);
         fclose(fontinfo_file);
         delete spacing;

classify/normmatch.cpp

@@ -263,7 +263,7 @@ NORM_PROTOS *Classify::ReadNormProtos(FILE *File, inT64 end_offset) {
 
   /* read protos for each class into a separate list */
   while ((end_offset < 0 || ftell(File) < end_offset) &&
-         fscanf(File, "%s %d", unichar, &NumProtos) == 2) {
+         tfscanf(File, "%s %d", unichar, &NumProtos) == 2) {
     if (unicharset.contains_unichar(unichar)) {
       unichar_id = unicharset.unichar_to_id(unichar);
       Protos = NormProtos->Protos[unichar_id];

classify/ocrfeatures.cpp

@@ -167,7 +167,7 @@ FEATURE ReadFeature(FILE *File, const FEATURE_DESC_STRUCT* FeatureDesc) {
 
   Feature = NewFeature (FeatureDesc);
   for (i = 0; i < Feature->Type->NumParams; i++) {
-    if (fscanf (File, "%f", &(Feature->Params[i])) != 1)
+    if (tfscanf (File, "%f", &(Feature->Params[i])) != 1)
       DoError (ILLEGAL_FEATURE_PARAM, "Illegal feature parameter spec");
 #ifndef _WIN32
     assert (!isnan(Feature->Params[i]));
@@ -198,12 +198,12 @@ FEATURE_SET ReadFeatureSet(FILE *File, const FEATURE_DESC_STRUCT* FeatureDesc) {
   int NumFeatures;
   int i;
 
-  if (fscanf (File, "%d", &NumFeatures) != 1 || NumFeatures < 0)
+  if (tfscanf(File, "%d", &NumFeatures) != 1 || NumFeatures < 0)
-    DoError (ILLEGAL_NUM_FEATURES, "Illegal number of features in set");
+    DoError(ILLEGAL_NUM_FEATURES, "Illegal number of features in set");
 
-  FeatureSet = NewFeatureSet (NumFeatures);
+  FeatureSet = NewFeatureSet(NumFeatures);
   for (i = 0; i < NumFeatures; i++)
-    AddFeature (FeatureSet, ReadFeature (File, FeatureDesc));
+    AddFeature(FeatureSet, ReadFeature (File, FeatureDesc));
 
   return (FeatureSet);
 

viewer/scrollview.cpp

@@ -775,6 +775,7 @@ void ScrollView::Image(struct Pix* image, int x_pos, int y_pos) {
   size_t size;
   pixWriteMem(&data, &size, image, IFF_PNG);
   int base64_len = (size + 2) / 3 * 4;
+  y_pos = TranslateYCoordinate(y_pos);
   SendMsg("readImage(%d,%d,%d)", x_pos, y_pos, base64_len);
   // Base64 encode the data.
   const char kBase64Table[64] = {