fix for issue 331

git-svn-id: https://tesseract-ocr.googlecode.com/svn/trunk@476 d0cd1f9f-072b-0410-8dd7-cf729c803f20
LittleDW · Sep 29, 2010 · 653abd9 · 653abd9
1 parent dbed3e0
commit 653abd9
Show file tree

Hide file tree

Showing 10 changed files with 67 additions and 23 deletions.
diff --git a/ccutil/Makefile.am b/ccutil/Makefile.am
@@ -12,7 +12,7 @@ include_HEADERS = \
     mainblk.h memblk.h memry.h memryerr.h mfcpch.h \
     ndminx.h notdll.h nwmain.h \
     ocrclass.h ocrshell.h platform.h qrsequence.h \
-    secname.h serialis.h stderr.h strngs.h \
+    secname.h serialis.h stderr.h strngs.h scanutils.h \
     tessclas.h tessdatamanager.h tessopt.h tordvars.h tprintf.h \
     unichar.h unicharmap.h unicharset.h unicity_table.h \
     varable.h
@@ -24,7 +24,7 @@ libtesseract_ccutil_la_SOURCES = \
     elst2.cpp elst.cpp errcode.cpp \
     globaloc.cpp hashfn.cpp \
     mainblk.cpp memblk.cpp memry.cpp ocrshell.cpp \
-    serialis.cpp strngs.cpp \
+    serialis.cpp strngs.cpp scanutils.cpp\
     tessdatamanager.cpp tessopt.cpp tordvars.cpp tprintf.cpp \
     unichar.cpp unicharmap.cpp unicharset.cpp \
     varable.cpp

diff --git a/ccutil/Makefile.in b/ccutil/Makefile.in
@@ -74,9 +74,9 @@ libtesseract_ccutil_la_LIBADD =
 am_libtesseract_ccutil_la_OBJECTS = ambigs.lo basedir.lo bits16.lo \
 	boxread.lo ccutil.lo clst.lo debugwin.lo elst2.lo elst.lo \
 	errcode.lo globaloc.lo hashfn.lo mainblk.lo memblk.lo memry.lo \
-	ocrshell.lo serialis.lo strngs.lo tessdatamanager.lo \
-	tessopt.lo tordvars.lo tprintf.lo unichar.lo unicharmap.lo \
-	unicharset.lo varable.lo
+	ocrshell.lo serialis.lo strngs.lo scanutils.lo \
+	tessdatamanager.lo tessopt.lo tordvars.lo tprintf.lo \
+	unichar.lo unicharmap.lo unicharset.lo varable.lo
 libtesseract_ccutil_la_OBJECTS = $(am_libtesseract_ccutil_la_OBJECTS)
 libtesseract_ccutil_la_LINK = $(LIBTOOL) --tag=CXX $(AM_LIBTOOLFLAGS) \
 	$(LIBTOOLFLAGS) --mode=link $(CXXLD) $(AM_CXXFLAGS) \
@@ -294,7 +294,7 @@ include_HEADERS = \
     mainblk.h memblk.h memry.h memryerr.h mfcpch.h \
     ndminx.h notdll.h nwmain.h \
     ocrclass.h ocrshell.h platform.h qrsequence.h \
-    secname.h serialis.h stderr.h strngs.h \
+    secname.h serialis.h stderr.h strngs.h scanutils.h \
     tessclas.h tessdatamanager.h tessopt.h tordvars.h tprintf.h \
     unichar.h unicharmap.h unicharset.h unicity_table.h \
     varable.h
@@ -306,7 +306,7 @@ libtesseract_ccutil_la_SOURCES = \
     elst2.cpp elst.cpp errcode.cpp \
     globaloc.cpp hashfn.cpp \
     mainblk.cpp memblk.cpp memry.cpp ocrshell.cpp \
-    serialis.cpp strngs.cpp \
+    serialis.cpp strngs.cpp scanutils.cpp\
     tessdatamanager.cpp tessopt.cpp tordvars.cpp tprintf.cpp \
     unichar.cpp unicharmap.cpp unicharset.cpp \
     varable.cpp
@@ -402,6 +402,7 @@ distclean-compile:
 @AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/memblk.Plo@am__quote@
 @AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/memry.Plo@am__quote@
 @AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/ocrshell.Plo@am__quote@
+@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/scanutils.Plo@am__quote@
 @AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/serialis.Plo@am__quote@
 @AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/strngs.Plo@am__quote@
 @AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/tessdatamanager.Plo@am__quote@

diff --git a/ccutil/scanutils.cpp b/ccutil/scanutils.cpp
@@ -47,7 +47,7 @@ enum Ranks {
   RANK_LONG = 1,
   RANK_LONGLONG = 2,
   RANK_PTR      = INT_MAX // Special value used for pointers
-  RANK_PTR      = 3 // Special value used for pointers
+//  RANK_PTR      = 3 // Special value used for pointers
 };
 
 const enum Ranks kMinRank = RANK_CHAR;
@@ -214,19 +214,48 @@ double strtofloat(const char* s)
   return minus ? -f : f;
 }
 
+static int tess_vfscanf(FILE* stream, const char *format, va_list ap);
+
+int tess_fscanf(FILE* stream, const char *format, ...)
+{
+  va_list ap;
+  int rv;
+
+  va_start(ap, format);
+  rv = tess_vfscanf(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 = tess_vfscanf(stream, format, ap);
+  va_end(ap);
+
+  return rv;
+}
+
+int vfscanf(FILE* stream, const char *format, ...)
+{
+  va_list ap;
+  int rv;
+
+  va_start(ap, format);
+  rv = tess_vfscanf(stream, format, ap);
   va_end(ap);
 
   return rv;
 }
+#endif
 
-int vfscanf(FILE* stream, const char *format, va_list ap)
+static
+int tess_vfscanf(FILE* stream, const char *format, va_list ap)
 {
   const char *p = format;
   char ch;
@@ -253,6 +282,7 @@ int vfscanf(FILE* stream, const char *format, va_list ap)
   int matchinv = 0;   // Is match map inverted?
   unsigned char range_start = 0;
   off_t start_off = ftell(stream);
+  double fval;
 
   // Skip leading spaces
   SkipSpace(stream);
@@ -414,7 +444,7 @@ int vfscanf(FILE* stream, const char *format, va_list ap)
                 break;
               }
 
-              double fval = streamtofloat(stream);
+              fval = streamtofloat(stream);
               switch(rank) {
                 case RANK_INT:
                   *va_arg(ap, float *) = static_cast<float>(fval);
@@ -539,7 +569,9 @@ 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);
 }
+#endif
diff --git a/ccutil/scanutils.h b/ccutil/scanutils.h
@@ -19,27 +19,34 @@
 #ifndef SCANUTILS_H
 #define SCANUTILS_H
 
-#ifdef EMBEDDED
-
 #include <stdint.h>
 #include <stddef.h>
 #include <stdio.h>
-#include <klibc/extern.h>
 #include <sys/stat.h>
 
+// 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 that scientific floating-point notation is not supported.
+// This variant is used to ensure correct reading regardless of locale.
+int tess_fscanf(FILE* stream, const char *format, ...);
+
+#ifdef EMBEDDED
+
+#include <klibc/extern.h>
+
 // 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.
 uintmax_t streamtoumax(FILE* s, int base);
 
 // 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 that scientific loating-point notation is not supported.
+// Note that scientific floating-point notation is not supported.
 int fscanf(FILE* stream, const char *format, ...);
 
 // 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 that scientific loating-point notation is not supported.
+// Note that scientific floating-point notation is not supported.
 int vfscanf(FILE* stream, const char *format, va_list ap);
 
 // Create a file at the specified path. See the creat manpage for more 

diff --git a/ccutil/serialis.cpp b/ccutil/serialis.cpp
@@ -79,7 +79,7 @@ DLLSYM void serialise_FLOAT64(FILE *f, double the_float) {
 DLLSYM double de_serialise_FLOAT64(FILE *f) {
   double the_float;
 
-  if (fscanf (f, "%lg", &the_float) != 1)
+  if (tess_fscanf (f, "%lg", &the_float) != 1)
     READFAILED.error ("de_serialise_FLOAT64", ABORT, NULL);
   return the_float;
 }

diff --git a/classify/clusttool.cpp b/classify/clusttool.cpp
@@ -87,7 +87,7 @@ PARAM_DESC *ReadParamDesc(FILE *File, uinT16 N) {
       ParamDesc[i].NonEssential = FALSE;
     else
       ParamDesc[i].NonEssential = TRUE;
-    if (fscanf (File, "%f%f", &(ParamDesc[i].Min), &(ParamDesc[i].Max)) !=
+    if (tess_fscanf (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;
@@ -289,7 +289,7 @@ ReadNFloats (FILE * File, uinT16 N, FLOAT32 Buffer[]) {
     Buffer = (FLOAT32 *) Emalloc (N * sizeof (FLOAT32));
 
   for (i = 0; i < N; i++) {
-    NumFloatsRead = fscanf (File, "%f", &(Buffer[i]));
+    NumFloatsRead = tess_fscanf (File, "%f", &(Buffer[i]));
     if (NumFloatsRead != 1) {
       if ((NumFloatsRead == EOF) && (i == 0))
         return (NULL);

diff --git a/classify/ocrfeatures.cpp b/classify/ocrfeatures.cpp
@@ -167,7 +167,7 @@ FEATURE ReadFeature(FILE *File, FEATURE_DESC FeatureDesc) {
 
   Feature = NewFeature (FeatureDesc);
   for (i = 0; i < Feature->Type->NumParams; i++) {
-    if (fscanf (File, "%f", &(Feature->Params[i])) != 1)
+    if (tess_fscanf (File, "%f", &(Feature->Params[i])) != 1)
       DoError (ILLEGAL_FEATURE_PARAM, "Illegal feature parameter spec");
 #ifndef __MSW32__
     assert (!isnan(Feature->Params[i]));

diff --git a/classify/protos.cpp b/classify/protos.cpp
@@ -377,7 +377,7 @@ void ReadProtos(register FILE *File, CLASS_TYPE Class) {
 
   for (Pid = 0; Pid < NumProtos; Pid++) {
     Proto = ProtoIn (Class, Pid);
-    fscanf (File, "%f %f %f %f %f %f %f\n",
+    tess_fscanf (File, "%f %f %f %f %f %f %f\n",
       &Proto->X,
       &Proto->Y,
       &Proto->Length,

diff --git a/config/config.h.in b/config/config.h.in
@@ -239,6 +239,10 @@
 #define GRAPHICS_DISABLED
 #endif
 
+#ifndef DISABLE_NLS
+#define USE_NLS
+#endif
+
 /* config_auto.h: end */
 #endif
 
diff --git a/tessdata/configs/Makefile.in b/tessdata/configs/Makefile.in
@@ -215,8 +215,8 @@ target_alias = @target_alias@
 top_build_prefix = @top_build_prefix@
 top_builddir = @top_builddir@
 top_srcdir = @top_srcdir@
-data_DATA = inter makebox box.train unlv ambigs.train api_config kannada box.train.stderr logfile
-EXTRA_DIST = inter makebox box.train unlv ambigs.train api_config kannada box.train.stderr logfile
+data_DATA = inter makebox box.train unlv ambigs.train api_config kannada box.train.stderr logfile digits
+EXTRA_DIST = inter makebox box.train unlv ambigs.train api_config kannada box.train.stderr logfile digits
 all: all-am
 
 .SUFFIXES: