Autoconf changes for 2.01

git-svn-id: https://tesseract-ocr.googlecode.com/svn/trunk@110 d0cd1f9f-072b-0410-8dd7-cf729c803f20

ChangeLog

@@ -30,3 +30,12 @@ July 02 2007 - V2.00
 	  Added UNLV regression test capability.
 	  Fixed problems with copyright and registered symbols.
 	  Fixed extern "C" declarations problem.
+August 27 2007 - V2.01
+	  Fixed UTF8 input problems with box file reader.
+	  Fixed various infinite loops and crashes in dawg code.
+	  Removed include of config_auto.h from host.h.
+	  Added automatic wctype encoding to unicharset_extractor.
+	  Fixed dawg table too full error.
+	  Removed svn files from tarball.
+	  Added new functions to tessdll.
+	  Increased maximum utf8 string in a classification result to 8.

Makefile.am

@@ -12,6 +12,8 @@ dist-hook:
 # added using EXTRA_DIST. $(distdir)/tessdata would in
 # theory suffice.
 	rm -rf `find $(distdir) -name CVS`
+	rm -rf `find $(distdir) -name .svn`
+	rm -rf `find $(distdir) -name .deps`
 # Also remove extra files not needed in a distribution
 	rm -rf `find $(distdir) -name configure.ac`
 	rm -rf `find $(distdir) -name acinclude.m4`

ReleaseNotes

+Tesseract release notes August 27 2007 - V2.01
+Fixed UTF8 input problems with box file reader.
+Fixed various infinite loops and crashes in dawg code.
+Removed include of config_auto.h from host.h.
+Added automatic wctype encoding to unicharset_extractor.
+Fixed dawg table too full error.
+Removed svn files from tarball.
+Added new functions to tessdll.
+Increased maximum utf8 string in a classification result to 8.
+
 Tesseract release notes July 17, 2007 - V2.00
 
 First release of the International version.

ccmain/tessedit.cpp

@@ -54,6 +54,21 @@
 #include "danerror.h"
 #include "globals.h"
 
+/*
+** Include automatically generated configuration file if running autoconf
+*/
+#ifdef HAVE_CONFIG_H
+#include "config_auto.h"
+// Includes libtiff if HAVE_LIBTIFF is defined
+#ifdef HAVE_LIBTIFF
+#include "tiffio.h"
+#endif
+#endif
+
+#ifdef GOOGLE3
+#include "third_party/tiff/tiffio.h"
+#endif
+
 //extern "C" {
 #include          "callnet.h"    //phils nn stuff
 //}
@@ -68,8 +83,6 @@ EXTERN BOOL_EVAR (tessedit_write_vars, FALSE, "Write all vars to file");
 EXTERN BOOL_VAR (tessedit_tweaking_tess_vars, FALSE,
 "Fiddle tess config values");
 
-EXTERN INT_VAR (tweak_ReliableConfigThreshold, 2, "Tess VAR");
-
 EXTERN double_VAR (tweak_garbage, 1.5, "Tess VAR");
 EXTERN double_VAR (tweak_ok_word, 1.25, "Tess VAR");
 EXTERN double_VAR (tweak_good_word, 1.1, "Tess VAR");
@@ -82,6 +95,7 @@ EXTERN double_VAR (tweak_NonDictCertainty, -2.5, "Tess VAR");
 EXTERN double_VAR (tweak_RejectCertaintyOffset, 1.0, "Tess VAR");
 EXTERN double_VAR (tweak_GoodAdaptiveMatch, 0.125, "Tess VAR");
 EXTERN double_VAR (tweak_GreatAdaptiveMatch, 0.10, "Tess VAR");
+EXTERN INT_VAR (tweak_ReliableConfigThreshold, 2, "Tess VAR");
 EXTERN INT_VAR (tweak_AdaptProtoThresh, 230, "Tess VAR");
 EXTERN INT_VAR (tweak_AdaptFeatureThresh, 230, "Tess VAR");
 EXTERN INT_VAR (tweak_min_outline_points, 6, "Tess VAR");
@@ -142,8 +156,6 @@ int init_tesseract(const char *arg0,
 
   start_recog(configfile, textbase);
 
-  ReliableConfigThreshold = tweak_ReliableConfigThreshold;
-
   set_tess_tweak_vars();
 
   if (tessedit_use_nn)           //phils nn stuff
@@ -333,6 +345,7 @@ void set_tess_tweak_vars() {
     RejectCertaintyOffset = tweak_RejectCertaintyOffset;
     GoodAdaptiveMatch = tweak_GoodAdaptiveMatch;
     GreatAdaptiveMatch = tweak_GreatAdaptiveMatch;
+    ReliableConfigThreshold = tweak_ReliableConfigThreshold;
     AdaptProtoThresh = tweak_AdaptProtoThresh;
     AdaptFeatureThresh = tweak_AdaptFeatureThresh;
     min_outline_points = tweak_min_outline_points;

ccmain/tessedit.h

@@ -25,15 +25,6 @@
 #include                    "pgedit.h"
 #include          "notdll.h"
 
-// Includes libtiff if HAVE_LIBTIFF is defined
-#ifdef HAVE_LIBTIFF
-#ifdef GOOGLE3
-#include "third_party/tiff/tiffio.h"
-#else
-#include "tiffio.h"
-#endif
-#endif
-
                                  //progress monitor
 extern ETEXT_DESC *global_monitor;
 
@@ -46,10 +37,6 @@ int init_tesseract(const char *arg0,
 void recognize_page(STRING& image_name);
 void end_tesseract();
 
-#ifdef _TIFFIO_
-void read_tiff_image(TIFF* tif, IMAGE* image);
-#endif
-
 //handle for "MODES"
 void extend_menu(RADIO_MENU *modes_menu,
                  INT16 modes_id_base,         //mode cmd ids offset

ccmain/tesseractmain.cpp

@@ -37,6 +37,25 @@
 #include "tfacep.h"
 #include "callnet.h"
 
+/*
+** Include automatically generated configuration file if running autoconf
+*/
+#ifdef HAVE_CONFIG_H
+#include "config_auto.h"
+// Includes libtiff if HAVE_LIBTIFF is defined
+#ifdef HAVE_LIBTIFF
+#include "tiffio.h"
+#endif
+#endif
+
+#ifdef GOOGLE3
+#include "third_party/tiff/tiffio.h"
+#endif
+
+#ifdef _TIFFIO_
+void read_tiff_image(TIFF* tif, IMAGE* image);
+#endif
+
 #define VARDIR        "configs/" /*variables files */
                                  //config under api
 #define API_CONFIG      "configs/api_config"

ccutil/Makefile.am

@@ -8,7 +8,7 @@ include_HEADERS = \
     nwmain.h ocrclass.h ocrshell.h platform.h secname.h serialis.h \
     stderr.h strngs.h tessclas.h tprintf.h varable.h \
     mfcpch.cpp scanutils.cpp scanutils.h unichar.h \
-    unicharmap.h unicharset.h
+    unicharmap.h unicharset.h boxread.h
 
 lib_LIBRARIES = libtesseract_ccutil.a
 libtesseract_ccutil_a_SOURCES = \
@@ -16,4 +16,4 @@ libtesseract_ccutil_a_SOURCES = \
     elst2.cpp errcode.cpp globaloc.cpp hashfn.cpp mainblk.cpp \
     memblk.cpp memry.cpp ocrshell.cpp serialis.cpp strngs.cpp \
     tprintf.cpp varable.cpp unichar.cpp tessopt.cpp \
-    unicharmap.cpp unicharset.cpp
+    unicharmap.cpp unicharset.cpp boxread.cpp

ccutil/host.h

@@ -45,17 +45,6 @@
 #ifndef   __HOST__
 #define   __HOST__
 
-/*
-** Include automatically generated configuration file if running autoconf
-*/
-#ifdef HAVE_CONFIG_H
-#include "config_auto.h"
-#if defined(MOTOROLA_BYTE_ORDER) || defined(WORDS_BIGENDIAN)
-#define __MOTO__  // Big-endian.
-#endif
-#endif
-
-
 /******************************************************************************
  **                                IMPORTANT!!!                                                                                                                 **
  **                                                                                                                                                                              **

configure.ac

@@ -7,7 +7,7 @@
 # ----------------------------------------
 
 AC_PREREQ(2.50)
-AC_INIT(tesseract, 2.00, theraysmith@gmail.com)
+AC_INIT(tesseract, 2.01, theraysmith@gmail.com)
 AC_REVISION($Id: configure.ac,v 1.4 2007/02/02 22:38:17 theraysmith Exp $)
 AC_CONFIG_AUX_DIR(config)
 AC_CONFIG_SRCDIR(ccmain/tesseractmain.cpp)

image/imgtiff.cpp

@@ -23,6 +23,17 @@
 #else
 #include          <unistd.h>
 #endif
+
+/*
+** Include automatically generated configuration file if running autoconf
+*/
+#ifdef HAVE_CONFIG_H
+#include "config_auto.h"
+#if defined(MOTOROLA_BYTE_ORDER) || defined(WORDS_BIGENDIAN)
+#define __MOTO__  // Big-endian.
+#endif
+#endif
+
 #include          "fileerr.h"
 #include          "imgerrs.h"
 #include          "img.h"

makemoredists

@@ -4,6 +4,7 @@ ver=`ls -1rt *[0-9].tar.gz |tail -1`
 ver=${ver%.tar.gz}
 for l in eng deu fra ita spa nld
 do
-	tar chozf $ver.$l.tar.gz tessdata/$l.*
+	chmod 644 tessdata/$l.*
+	tar --group root --owner root -chozf $ver.$l.tar.gz tessdata/$l.*
 done
 tar chozf $ver.exe.tar.gz $winlist

tessdata/configs/Makefile.am

 datadir = @datadir@/tessdata/configs
-data_DATA = inter makebox box.train unlv
-EXTRA_DIST = inter makebox box.train unlv
+data_DATA = inter makebox box.train unlv api_config
+EXTRA_DIST = inter makebox box.train unlv api_config

tessdata/configs/api_config

+tessedit_zero_rejection T