Update to 2.00, provided by Rumko on pkgsrc-users.

July 02 2007 - V2.00
	  Converted internal character handling to UTF8.
	  Trained with 6 languages.
	  Added unicharset_extractor, wordlist2dawg.
	  Added boxfile creation mode.
	  Added UNLV regression test capability.
	  Fixed problems with copyright and registered symbols.
	  Fixed extern "C" declarations problem.

7 files changed, 67 insertions, 73 deletions

diff --git a/graphics/tesseract/Makefile b/graphics/tesseract/Makefile
index 3102c7a1479..2528accfd84 100644
--- a/graphics/tesseract/Makefile
+++ b/graphics/tesseract/Makefile
@@ -1,7 +1,7 @@
-# $NetBSD: Makefile,v 1.1.1.1 2007/05/18 06:39:27 wiz Exp $
+# $NetBSD: Makefile,v 1.2 2007/07/28 01:02:14 wiz Exp $
 #
 
-DISTNAME=	tesseract-1.04b
+DISTNAME=	tesseract-2.00
 CATEGORIES=	graphics
 MASTER_SITES=	http://tesseract-ocr.googlecode.com/files/
 
@@ -13,7 +13,6 @@ PKG_DESTDIR_SUPPORT=user-destdir
 
 GNU_CONFIGURE=	yes
 USE_LANGUAGES=	c c++
-WRKSRC=		${WRKDIR}/tesseract-1.04
 
 post-build:
 	${SED} -e "s,@PREFIX@,${PREFIX}," ${FILESDIR}/tesseract.sh \
diff --git a/graphics/tesseract/PLIST b/graphics/tesseract/PLIST
index 29f995cfb63..2346d3cbd36 100644
--- a/graphics/tesseract/PLIST
+++ b/graphics/tesseract/PLIST
@@ -1,7 +1,9 @@
-@comment $NetBSD: PLIST,v 1.1.1.1 2007/05/18 06:39:27 wiz Exp $
+@comment $NetBSD: PLIST,v 1.2 2007/07/28 01:02:14 wiz Exp $
 bin/cntraining
 bin/mftraining
 bin/tesseract
+bin/unicharset_extractor
+bin/wordlist2dawg
 include/tesseract/adaptions.h
 include/tesseract/adaptive.h
 include/tesseract/adaptmatch.h
@@ -37,7 +39,6 @@ include/tesseract/clst.h
 include/tesseract/cluster.h
 include/tesseract/clusttool.h
 include/tesseract/cmndwin.h
-include/tesseract/cnTraining.dsp
 include/tesseract/const.h
 include/tesseract/context.h
 include/tesseract/control.h
@@ -52,6 +53,7 @@ include/tesseract/debugwin.h
 include/tesseract/djmenus.h
 include/tesseract/dlltest.cpp
 include/tesseract/dlltest.dsp
+include/tesseract/dlltest.vcproj
 include/tesseract/docqual.h
 include/tesseract/drawedg.h
 include/tesseract/drawfx.h
@@ -114,9 +116,11 @@ include/tesseract/labls.h
 include/tesseract/linlsq.h
 include/tesseract/listio.h
 include/tesseract/lmedsq.h
+include/tesseract/lookdawg.h
 include/tesseract/lsterr.h
 include/tesseract/mainblk.h
 include/tesseract/makechop.h
+include/tesseract/makedawg.h
 include/tesseract/makerow.h
 include/tesseract/matchdefs.h
 include/tesseract/matchtab.h
@@ -129,7 +133,6 @@ include/tesseract/memryerr.h
 include/tesseract/mergenf.h
 include/tesseract/metrics.h
 include/tesseract/mf.h
-include/tesseract/mfTraining.dsp
 include/tesseract/mfcpch.cpp
 include/tesseract/mfcpch.h
 include/tesseract/mfdefs.h
@@ -188,6 +191,7 @@ include/tesseract/quadratc.h
 include/tesseract/quspline.h
 include/tesseract/ratngs.h
 include/tesseract/rect.h
+include/tesseract/reduce.h
 include/tesseract/rejctmap.h
 include/tesseract/reject.h
 include/tesseract/render.h
@@ -262,6 +266,7 @@ lib/libtesseract_classify.a
 lib/libtesseract_cutil.a
 lib/libtesseract_dict.a
 lib/libtesseract_display.a
+lib/libtesseract_full.a
 lib/libtesseract_image.a
 lib/libtesseract_main.a
 lib/libtesseract_textord.a
@@ -270,7 +275,19 @@ lib/libtesseract_viewer.a
 lib/libtesseract_wordrec.a
 libexec/tesseract
 share/doc/tesseract/README
+share/tessdata/configs/box.train
+share/tessdata/configs/inter
+share/tessdata/configs/makebox
+share/tessdata/configs/unlv
 share/tessdata/confsets
+share/tessdata/deu.DangAmbigs
+share/tessdata/deu.freq-dawg
+share/tessdata/deu.inttemp
+share/tessdata/deu.normproto
+share/tessdata/deu.pffmtable
+share/tessdata/deu.unicharset
+share/tessdata/deu.user-words
+share/tessdata/deu.word-dawg
 share/tessdata/eng.DangAmbigs
 share/tessdata/eng.freq-dawg
 share/tessdata/eng.inttemp
@@ -279,8 +296,48 @@ share/tessdata/eng.pffmtable
 share/tessdata/eng.unicharset
 share/tessdata/eng.user-words
 share/tessdata/eng.word-dawg
+share/tessdata/fra.DangAmbigs
+share/tessdata/fra.freq-dawg
+share/tessdata/fra.inttemp
+share/tessdata/fra.normproto
+share/tessdata/fra.pffmtable
+share/tessdata/fra.unicharset
+share/tessdata/fra.user-words
+share/tessdata/fra.word-dawg
+share/tessdata/ita.DangAmbigs
+share/tessdata/ita.freq-dawg
+share/tessdata/ita.inttemp
+share/tessdata/ita.normproto
+share/tessdata/ita.pffmtable
+share/tessdata/ita.unicharset
+share/tessdata/ita.user-words
+share/tessdata/ita.word-dawg
+share/tessdata/nld.DangAmbigs
+share/tessdata/nld.freq-dawg
+share/tessdata/nld.inttemp
+share/tessdata/nld.normproto
+share/tessdata/nld.pffmtable
+share/tessdata/nld.unicharset
+share/tessdata/nld.user-words
+share/tessdata/nld.word-dawg
+share/tessdata/spa.DangAmbigs
+share/tessdata/spa.freq-dawg
+share/tessdata/spa.inttemp
+share/tessdata/spa.normproto
+share/tessdata/spa.pffmtable
+share/tessdata/spa.unicharset
+share/tessdata/spa.user-words
+share/tessdata/spa.word-dawg
+share/tessdata/tessconfigs/batch
+share/tessdata/tessconfigs/batch.nochop
+share/tessdata/tessconfigs/matdemo
+share/tessdata/tessconfigs/msdemo
+share/tessdata/tessconfigs/nobatch
+share/tessdata/tessconfigs/segdemo
 share/tesseract/phototest.tif
-@dirrm share/doc/tesseract
 @dirrm share/tesseract
+@dirrm share/tessdata/tessconfigs
+@dirrm share/tessdata/configs
 @dirrm share/tessdata
+@dirrm share/doc/tesseract
 @dirrm include/tesseract
diff --git a/graphics/tesseract/distinfo b/graphics/tesseract/distinfo
index 3e8ed942a24..c81aa1eb5a4 100644
--- a/graphics/tesseract/distinfo
+++ b/graphics/tesseract/distinfo
@@ -1,9 +1,5 @@
-$NetBSD: distinfo,v 1.1.1.1 2007/05/18 06:39:27 wiz Exp $
+$NetBSD: distinfo,v 1.2 2007/07/28 01:02:14 wiz Exp $
 
-SHA1 (tesseract-1.04b.tar.gz) = 263a65e462ed864c4da115cdcb3f3e78613de485
-RMD160 (tesseract-1.04b.tar.gz) = 5e9c70d4435a59157f0af6503a57b02a4a74350e
-Size (tesseract-1.04b.tar.gz) = 2899276 bytes
-SHA1 (patch-ae) = c22f254b73fb9bbd02cf8ef7b4ccbea475afd5df
-SHA1 (patch-ag) = 581ec7ac0528bb28fddb3fbaa35a87bb1835a82e
-SHA1 (patch-ah) = 22987d8523631c5c6e8b2fb5096ff87c5bc13124
-SHA1 (patch-ai) = e219077d2acf0652a9bf6418d3f8ce4e11782ed5
+SHA1 (tesseract-2.00.tar.gz) = 3051d90b74a255db791ce0d7a6dc3e12203727b9
+RMD160 (tesseract-2.00.tar.gz) = 96deb8315861e0e0c423fc2d12c9a69b776cb456
+Size (tesseract-2.00.tar.gz) = 1072362 bytes
diff --git a/graphics/tesseract/patches/patch-ae b/graphics/tesseract/patches/patch-ae
deleted file mode 100644
index 04ad0e37cfb..00000000000
--- a/graphics/tesseract/patches/patch-ae
+++ /dev/null
@@ -1,16 +0,0 @@
-$NetBSD: patch-ae,v 1.1.1.1 2007/05/18 06:39:27 wiz Exp $
-
---- cutil/globals.h.orig	2006-06-17 00:17:07.000000000 +0200
-+++ cutil/globals.h
-@@ -43,9 +43,11 @@ extern int acts[MAXPROC];        /*actio
- extern int debugs[MAXPROC];      /*debug flags */
- extern int plots[MAXPROC];       /*plot flags */
- extern int corners[4];           /*corners of scan window */
-+extern "C" {
- extern int optind;               /*option index */
- extern char *optarg;             /*option argument */
-                                  /*image file name */
-+}
- extern char imagefile[FILENAMESIZE];
-                                  /* main directory */
- extern char directory[FILENAMESIZE];
diff --git a/graphics/tesseract/patches/patch-ag b/graphics/tesseract/patches/patch-ag
deleted file mode 100644
index 280764380a6..00000000000
--- a/graphics/tesseract/patches/patch-ag
+++ /dev/null
@@ -1,15 +0,0 @@
-$NetBSD: patch-ag,v 1.1.1.1 2007/05/18 06:39:27 wiz Exp $
-
---- cutil/tordvars.h.orig	2007-05-16 21:33:53.000000000 +0000
-+++ cutil/tordvars.h
-@@ -45,8 +45,8 @@ extern int similarity_enable;           
- extern int similarity_debug;                /* Level of debug output */
- extern int write_raw_output;                /* Text before context */
- extern int write_output;                    /* Text file output */
--//extern "C" { extern int display_ratings; }  /* Show the ratings */
--extern int display_ratings;                 /* Show the ratings */
-+extern "C" { extern int display_ratings; }  /* Show the ratings */
-+//extern int display_ratings;                 /* Show the ratings */
- extern int show_bold;                       /* Use bold text */
- extern int display_text;                    /* Show word text */
- extern int display_blocks;                  /* Show word as boxes */
diff --git a/graphics/tesseract/patches/patch-ah b/graphics/tesseract/patches/patch-ah
deleted file mode 100644
index ced73824249..00000000000
--- a/graphics/tesseract/patches/patch-ah
+++ /dev/null
@@ -1,13 +0,0 @@
-$NetBSD: patch-ah,v 1.1.1.1 2007/05/18 06:39:27 wiz Exp $
-
---- ccutil/debugwin.cpp.orig	2006-06-16 22:17:04.000000000 +0000
-+++ ccutil/debugwin.cpp
-@@ -229,7 +229,7 @@ DEBUG_WIN::DEBUG_WIN(                   
-   length += sprintf (command + length, "trap \"\" 1 2 3 13 15\n");
-   length +=
-     sprintf (command + length,
--    "/usr/bin/X11/xterm -sb -sl " INT32FORMAT " -geometry "
-+    "/usr/X11R6/bin/xterm -sb -sl " INT32FORMAT " -geometry "
-     INT32FORMAT "x" INT32FORMAT "", buflines, xsize / 8, ysize / 16);
-   if (xpos >= 0)
-     command[length++] = '+';
diff --git a/graphics/tesseract/patches/patch-ai b/graphics/tesseract/patches/patch-ai
deleted file mode 100644
index 45150d3da32..00000000000
--- a/graphics/tesseract/patches/patch-ai
+++ /dev/null
@@ -1,14 +0,0 @@
-$NetBSD: patch-ai,v 1.1.1.1 2007/05/18 06:39:27 wiz Exp $
-
---- configure.orig	2007-02-02 21:37:43.000000000 +0100
-+++ configure
-@@ -7083,7 +7083,8 @@ else
-   if test "$cross_compiling" = yes; then
-   ac_cv_func_fork_works=cross
- else
--  cat >conftest.$ac_ext <<_ACEOF
-+  cat confdefs.h >conftest.$ac_ext
-+  cat >>conftest.$ac_ext <<_ACEOF
- /* By Ruediger Kuhlmann. */
-       #include <sys/types.h>
-       #if HAVE_UNISTD_H