From da629a74cbbe9cf76a85e7c26e2b8ac991b9f549 Mon Sep 17 00:00:00 2001
From: Arnaud Fontaine <arnaud.fontaine@nexedi.com>
Date: Mon, 7 Nov 2016 17:00:32 +0900
Subject: [PATCH] Revert "erp5: Update tesseract to 3.04.01 because of
compilation error."
This reverts commit 292be60ff42edc2ce57f2e6701ad283c18e9772b because of
a regression in tesseract > 3.02:
https://github.com/tesseract-ocr/tesseract/issues/360#issuecomment-258775303
Related Unit Tests failures:
* test_PDFDocument_asTextConversion (testDms.TestDocument)
* test_MonochromeImageResize (testDms.TestDocument)
* test_CMYKImageTextContent (testDms.TestDocument)
=> AssertionError: 'ERP5 is a free software\n\n' != 'ERPS is a free software.\n\n'
Tested versions:
* 3.04.01 (leptnonica: 1.73): NOK
* 3.03.03 (leptonica: 1.71): NOK
* 3.03-rc1 (leptonica: 1.71): NOK
* 3.02.01 (leptonica: 1.69): OK
* 3.02 (leptonica: 1.71): OK
---
component/leptonica/buildout.cfg | 7 ++++--
.../leptonica-1.69-zlib-include.patch | 22 +++++++++++++++++++
component/tesseract/buildout.cfg | 9 ++++++--
.../tesseract-3.00-gcc-4.7-build.patch | 10 +++++++++
.../tesseract/tesseract-3.01-remove-bom.patch | 15 +++++++++++++
5 files changed, 59 insertions(+), 4 deletions(-)
create mode 100644 component/leptonica/leptonica-1.69-zlib-include.patch
create mode 100644 component/tesseract/tesseract-3.00-gcc-4.7-build.patch
create mode 100644 component/tesseract/tesseract-3.01-remove-bom.patch
diff --git a/component/leptonica/buildout.cfg b/component/leptonica/buildout.cfg
index 1b18493a6..5bc3e6868 100644
--- a/component/leptonica/buildout.cfg
+++ b/component/leptonica/buildout.cfg
@@ -11,10 +11,13 @@ extends =
[leptonica]
recipe = slapos.recipe.cmmi
-url = http://www.leptonica.org/source/leptonica-1.73.tar.gz
-md5sum = 092cea2e568cada79fff178820397922
+url = http://leptonica.googlecode.com/files/leptonica-1.68.tar.gz
+md5sum = 5cd7092f9ff2ca7e3f3e73bfcd556403
configure-options =
--disable-static
+patch-options = -p1
+patches =
+ ${:_profile_base_location_}/leptonica-1.69-zlib-include.patch#cff3dc942075190939b407c38e0d3201
environment =
CPPFLAGS=-I${zlib:location}/include -I${libjpeg:location}/include -I${libpng:location}/include -I${libtiff:location}/include -I${webp:location}/include -I${giflib:location}/include
LDFLAGS=-L${zlib:location}/lib -Wl,-rpath=${zlib:location}/lib -L${libjpeg:location}/lib -Wl,-rpath=${libjpeg:location}/lib -L${libpng:location}/lib -Wl,-rpath=${libpng:location}/lib -L${libtiff:location}/lib -Wl,-rpath=${libtiff:location}/lib -L${webp:location}/lib -Wl,-rpath=${webp:location}/lib -L${giflib:location}/lib -Wl,-rpath=${giflib:location}/lib
diff --git a/component/leptonica/leptonica-1.69-zlib-include.patch b/component/leptonica/leptonica-1.69-zlib-include.patch
new file mode 100644
index 000000000..7f907611d
--- /dev/null
+++ b/component/leptonica/leptonica-1.69-zlib-include.patch
@@ -0,0 +1,22 @@
+This patch is originally taken from:
+
+http://leptonica.googlecode.com/issues/attachment?aid=560001000&name=zlib-include.patch&token=m2sugSYxB4xwAuNgrKXyHTxBYNg%3A1337345966091
+
+To fix the following issue with leptonica:
+
+http://code.google.com/p/leptonica/issues/detail?id=56
+
+diff -Nurd -x'*~' leptonica-1.68.orig/src/pngio.c leptonica-1.68/src/pngio.c
+--- leptonica-1.68.orig/src/pngio.c 2011-02-01 00:41:12.000000000 -0500
++++ leptonica-1.68/src/pngio.c 2011-07-09 09:17:17.000000000 -0400
+@@ -108,6 +108,10 @@
+
+ #include "png.h"
+
++#ifdef HAVE_LIBZ
++#include "zlib.h"
++#endif
++
+ /* ----------------Set defaults for read/write options ----------------- */
+ /* strip 16 bpp --> 8 bpp on reading png; default is for stripping */
+ static l_int32 var_PNG_STRIP_16_TO_8 = 1;
diff --git a/component/tesseract/buildout.cfg b/component/tesseract/buildout.cfg
index e8d297910..d81e19c84 100644
--- a/component/tesseract/buildout.cfg
+++ b/component/tesseract/buildout.cfg
@@ -14,8 +14,13 @@ parts =
[tesseract]
recipe = slapos.recipe.cmmi
-url = https://github.com/tesseract-ocr/tesseract/archive/3.04.01.tar.gz
-md5sum = 645a21effcf2825a3473849d72a7fd90
+url = http://tesseract-ocr.googlecode.com/files/tesseract-3.01.tar.gz
+md5sum = 1ba496e51a42358fb9d3ffe781b2d20a
+patch-options =
+ -p1
+patches =
+ ${:_profile_base_location_}/tesseract-3.00-gcc-4.7-build.patch#ca80db3ec489c547b03f3ee48879c1b1
+ ${:_profile_base_location_}/tesseract-3.01-remove-bom.patch#2e691858cb492b7c17d23bf0912b3d24
pre-configure =
libtoolize -f -c
aclocal -I ${libtool:location}/share/aclocal -I config
diff --git a/component/tesseract/tesseract-3.00-gcc-4.7-build.patch b/component/tesseract/tesseract-3.00-gcc-4.7-build.patch
new file mode 100644
index 000000000..76446109f
--- /dev/null
+++ b/component/tesseract/tesseract-3.00-gcc-4.7-build.patch
@@ -0,0 +1,10 @@
+--- tesseract-3.00/viewer/svutil.cpp.old 2012-04-11 09:34:26.168608940 +0200
++++ tesseract-3.00/viewer/svutil.cpp 2012-04-11 09:34:46.108565692 +0200
+@@ -21,6 +21,7 @@
+ // thread/process creation & synchronization and network connection.
+
+ #include <stdio.h>
++#include <unistd.h>
+ #ifdef WIN32
+ #include <windows.h>
+ #include <winsock.h>
diff --git a/component/tesseract/tesseract-3.01-remove-bom.patch b/component/tesseract/tesseract-3.01-remove-bom.patch
new file mode 100644
index 000000000..c50254a65
--- /dev/null
+++ b/component/tesseract/tesseract-3.01-remove-bom.patch
@@ -0,0 +1,15 @@
+The patch below removes a utf-8 BOM mark.
+
+Avoid touching it as the BOM is invisible, and copy/pasting might not work.
+
+It is needed because old compilers treat the BOM as garbage instead of
+whitespace.
+
+--- tesseract-3.01/ccutil/strngs.h.orig 2012-05-24 15:13:22.743808379 +0200
++++ tesseract-3.01/ccutil/strngs.h 2012-05-24 15:16:54.468858282 +0200
+@@ -1,4 +1,4 @@
+-/**********************************************************************
++/**********************************************************************
+ * File: strngs.h (Formerly strings.h)
+ * Description: STRING class definition.
+ * Author: Ray Smith
--
2.30.9